Professional Documents
Culture Documents
865206
AEROACOUSTICS:
ACOUSTIC WAVE PROPAGATION;
AIRCRAFT NOISE PREDICTION;
AEROACOUSTIC INSTRUMENTATION
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Edited by
Ira R. Schwartz
NASA Ames Research Center
Moffett Field, California
Assistant Editors:
Henry T. Nagamatsu
General Electric Research and Development Center
Schenectady, New York
Warren C. Strahle
Georgia Institute of Technology
Atlanta, Georgia
Volume 46
PROGRESS IN
ASTRONAUTICS AND AERONAUTICS
Copyright 1976 by
American Institute of Aeronautics and Astronautics
Preface x/x
Tables of Contents
(Companion Volumes 43, 44, 45)
Volume 43
Preface xix
Preface x/x
Volume 45
Preface xix
VOLUMES EDITORS
1. Solid Propellant Rocket Martin Summerfield
PRINCETON UNIVERSITY
Research. 1960
Ernst Stuhlinger
NASA GEORGE C. MARSHALL SPACE
FLIGHT CENTER
J. M. Sellen Jr.
SPACE TECHNOLOGY LABORATORIES
James S. Farrior
LOCKHEED MISSILES AND SPACE
COMPANY
1964
Charles J. Mundo
INSTITUTE OF NAVAL STUDIES
COMSAT LABORATORIES
Satellite Technology.
1974
William G. Schmidt
CML SATELLITE CORPORATION
Warren C. Strahle
GEORGIA INSTITUTE OF TECHNOLOGY
Warren C. Strahle
GEORGIA INSTITUTE OF TECHNOLOGY
Warren C. Strahle
GEORGIA INSTITUTE OF TECHNOLOGY
Warren C. Strahle
GEORGIA INSTITUTE OF TECHNOLOGY
Until recently, little progress had been made towards finding solutions to
aeroacoustic problems, partly because of their many complexities and partly
because of the limited research programs to support technological develop-
ment. As the emphasis on aeroacoustics research and noise suppression
technology has increased during the past several years, significant progress
has been made towards achieving the goals of the aircraft industry and the
rural communities. Recent advancements in technology have made it
possible for jet engine noise to be reduced to levels that will permit many of
the present large commercial aircraft to satisfy the FAR-36 noise
requirements. The goals for future commercial aircraft, however, are to ad-
vance technology in order to permit the reduction of the FAR-36 noise
requirement by approximately 10 EPNdB per decade. Much work,
therefore, still must be done to develop the scientific and technological tools
required for substantial advancements in aerodynamic noise abatement.
More emphasis must be placed in areas of research that will provide a better
understanding of the mechanisms of engine and airframe noise generation,
propagation, and attenuation, including flight effects. However,
technological breakthroughs in noise abatement need not be paced by com-
plete scientific understanding of the mechanisms of the phenomenon in-
volved. We may find and implement very successfully an efficient method of
suppressing jet and core engine noise and yet not understand completely the
very complex mechanisms that modify and suppress the turbulence modes
of noise-producing sources. Such an event, for example, would be com-
parable to the discovery of successful methods of controlling boundary-
layer flows by the use of vortex generators and other devices over 30 years
ago without having a complete understanding at that time of the boundary-
layer flow mechanisms involved in the aerodynamic phenomenon.Thus, this
recently increased effort in aeroacoustics research and technology must be
continued to meet present and future environmental noise requirements
being demanded by Congress, the FAA, and a critical public. Accordingly,
the Second AIAA Aeroacoustics Specialists Conference was held at Hamp-
ton, Va.,March 24-26, 1975, to assess the most recent accomplishments in
aeroacoustic research and technology for different flight vehicles and to
stimulate further the growth of new ideas and understanding. A broad spec-
trum of aeroacoustic research and engineering programs is being performed
in many countries, universities, government laboratories, and industries, as
reflected by the technical sessions of the Conference, at which 108 papers
were presented. Of these, 98 papers are published in four volumes of the
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
The progress made during the past 1 Vi to 2 years, as reported in the four
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Volume 43. Two of these studies used vortex ring approaches, and the other
was concerned with the Navier-Stokes equations as a starting point. All of
these approaches have provided additional knowledge and insights into
various aspects of jet noise. They definitely should be continued.
With regard to supersonic jet noise, studies of the mechanisms of noise
source generation and transmission have provided useful data to our reser-
voir of knowledge and understanding. Further work in this area is en-
couraged.
Several investigations dealing with flight effects which were conducted in
wind tunnels, and also by utilizing aircraft in flight, have indicated the im-
portance of separating the aircraft component noise sources and hence
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
isolating the jet noise. Comparison of static and flight noise levels indicated
a forward arc amplification of the flight noise levels. The question that must
be answered is whether the amplification of noise is due to jet noise
phenomenon or perhaps an interaction of airframe and core engine noise.
In the area of jet noise suppression, various promising suppressor con-
cepts were investigated, the results of which are discussed in Volume 43. The
Swirling Flow Jet Noise Suppressor that was investigated by Schwartz, using
a full-scale turbojet engine, produced significant noise reduction with
minimal thrust losses. Results of previous studies of the swirling flow jet
noise suppressor, using a full-scale turbofan engine as reported in Volume
37 of this Series, indicated that even greater beneficial suppression effects
could be obtained by swirling the jet in a turbofan engine. The potential ad-
vantages that this swirl jet noise suppressor offers in comparison with other
known suppressors are very impressive. Further research on full-scale,
higher thrust jet engines are continuing.
The aircraft engine core noise problem has been receiving greater em-
phasis by researchers, as reflected in seven research type papers in Volume
43. An additional paper in Volume 46 deals with prediction methods of core
noise. Since engine core noise includes combustion noise sources, and tur-
bine and duct noise sources, the identification of the mechanisms of noise
generation and transmission becomes very complex. The increased interest
in engine core noise is due to the fact that core and jet noise have become the
threshold of dominant aircraft noise in present transport engines. The
coupling of upstream core noise sources with jet noise sources is known to
produce significant increases in jet noise levels. One or two possible
mechanisms could cause core noise, and herein is the difficult problem. It is
the opinion of most researchers that one of the potential mechanisms is the
direct noise that is radiated from the turbulent combustion in the primary
combustor and transmitted through the turbine, passing out the nozzle into
the far field. The other mechanism that is considered to cause core noise is
the noise that is emitted from hot spots being convected through the tur-
bine. The question that arises is: Which one of these mechanisms, or
perhaps both mechanisms, is responsible for engine core noise? The other
question confronting researchers is: What are the coupling mechanisms of
core engine noise and jet noise?
In order to solve any of the important jet noise or core engine noise
problems, experiments must be conducted in engine-type apparatus so that
all the important aerodynamic and thermodynamic parameters, and hence
the resultant interactions, will be present in the flow model. Although the
research emphasis has increased in both core engine noise and jet noise areas
during the last few years, additional emphasis must be placed in these areas
so as ultimately to attain practical solutions.
In an assessment of the status of fan noise research, as presented at the
Conference and in Volume 44, it was concluded that one of the most
significant achievements in recent years that could potentially advance our
knowledge in this area was the observation that flight noise from a fan was
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
lower than those levels expected or obtained from static testing. The cause
of this discrepancy has been attributed to the poor inflow conditions at the
inlet of the engine on static stands. Therefore, rotor-stator interaction noise
data previously taken in static testing are considered suspect and apparently
incorrect. Experiments with screens to remove turbulence appeared to be
successful. The consensus is to test the configuration in a wind tunnel where
the forward velocity will provide a sufficiently clean inflow to simulate
flight conditions. Once the inflows are smooth, the rotor-stator interaction
noise, which apparently was not the dominant source in most of the
previous tests, can be evaluated.
The recent progress in duct acoustic research is reflected by fourteen
papers in Volume 44 which cover essentially three areaslinear duct
acoustics, the characteristics of absorbent materials, and nonlinear duct
acoustics. Most of the emphasis in the area of the propagation in ducts was
placed on computational methods to attain effective and economical
modeling of the propagation in variable geometry and in hardwall or soft-
wall ducts. New computation methods were developed to reduce computer
processing and storage requirements. Baumeister suggests a wave envelope
technique to reduce the processing and storage needs and successfully ap-
plies it to the optimum, segmented, acoustic duct liner design. Quinn uses a
finite difference technique in combination with conformal mapping to com-
pute the optimum multi-section duct liner design similar to Baumeister's ap-
proach. Kapur and Mungur propose acoustic finite element approaches as a
promising new approach to solve acoustic duct problems. Although similar
methods have been succesfully applied to structures of complex design, their
applicability to acoustic problems requires development of so-called func-
tional equivalents to acoustics problems.
of the waves which inhibit nonlinear transfer of energy into the lower har-
monics.
The two experimental papers on rotor noise describe work at speeds less
than the high tip speed limit but approaching it. For example, a research
paper by Leverton investigates rotor noise for the case of hover, which is
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
limited in the tip Mach number at a maximum of 0.68. His results indicate
that the rotational noise increases as approximately a V 2 law. Also, the
rotational noise is essentially independent of the thrust except for the higher
harmonics. The author concludes that profile drag or blade thickness is the
principal mechanism as the noise source and not the fluctuating loads.
The other experimental paper on rotor noise by Harris and Lee also
describes a low blade tip speed Mach number investigation in which
rotational noise and vortex noise are studied for a range of parameters. Em-
phasis in this investigation was placed on studying the effects of four
parameters: 1) the blade pitch setting (variation of thrust and blade
loading), 2) the number of blades, 3) the rotor shaft tilt angle, and 4) the tip
speed. The authors contend that, within some of the limits of the
parameters, their results agree well with theory; however, for others there is
no agreement. They emphasize the fact that there is a critical need to extend
the experiments from the low tip Mach number regime to the higher tip
Mach numbers. These experiments may provide better modeling for the
theories to determine whether it is thickness that generates the blade slap or
whether it is vortex interaction with the blades that produces the noise
source mechanisms.
Another major area of aeroacoustics research which was reviewed and
assessed at the Conference was VTOL or V/STOL Aircraft Noise. The
recent progress in this area was reflected by eight papers, presented in
Volume 45, which cover a broad spectrum of topics: 1) scrub-
bing/interaction noise, 2) upper surface blown flap turbulence and noise, 3)
acoustic pressures on aircraft surfaces related to powered-lift surfaces, 4)
shielding effects, and 5) forward velocity effects on jet wing/flap interaction
noise and also on under-the-wing externally blown flap noise. Most of this
research was directed at developing better methods for the prediction of
noise generation and propagation, at determining the effects of pressure
fluctuation on surfaces and interior noise, and at developing methods for
suppressing aircraft noise. Specific problems associated with VTOL or
STOL, which also exist in other aircraft and engine noise control programs,
include: Inlet design and wall treatment for noise suppression, prediction of
fan noise and development of methods for control, engine core and com-
bustion noise prediction and control, flap-interaction noise prediction and
control, and prediction of airframe noise.
It is apparent that the various experimental research programs conducted
in these V/STOL problem areas have improved in quality by utilizing
sophisticated test techniques, instrumentation, facilities, and imaginative
models. The results of these tests should provide the basis for improved
modeling of the flow mechanisms and hence allow the development of better
methods for the prediction and calculation of VTOL and STOL noise. For
example, the experimental investigation by Fink (Volume 45) which
examined the aeroacoustic mechanism that produces externally blown flap
(EBF) scrubbing noise, has provided concepts of the noise-producing
mechanisms that will be useful to develop prediction methods for EBF noise
estimation. Scrubbing noise appeared to come from weak surface loading
fluctuations that were coherent along the scrubbed span. The opinion is that
these loading fluctuations were induced by converted large-scale vortex
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
The increased activity by researchers in the airframe and airfoil noise area
is illustrated by the fact that fourteen papers were presented at the Con-
ference, all of which are included in Volume 45. Airframe noise is defined as
the noise generated by an aircraft's nonpropulsive components in flight.
Thus, airframe noise is attributed to turbulent flows interacting with aircraft
surfaces. For advanced aircraft, airframe (nonpropulsive) noise could be
nearly equal to noise from propulsion systems during the final stages of lan-
ding approach. Future noise regulations are estimated at 10 PNdB below
Federal Air Regualtion 36 (FAR-36). Thus, if airframe noise sources exceed
the FAR 36-10-dB level, this federal regulation noise level cannot be
satisfied by reducing propulsion noise sources alone. In fact, a number of
recent measurement programs on different types of aircraft have indeed
determined that airframe noise attains a level within 10 dB of the overall air-
craft power-on noise levels for landing approach conditions. Therefore, air-
frame noise must be considered as a potential obstacle for commercial air-
craft to overcome in order to satisfy proposed government noise regulations
successfully.
Significant progress has been made during the two years since the first
AIAA Aeroacoustics Conference to identify acoustically important air-
frame components and to develop noise prediction methods for each com-
ponent. It became apparent from these research programs that there was a
great dependence of the radiated noise spectrum on the aerodynamic con-
figuration of the aircraft. Further, these investigations established a need to
develop a better or more detailed understanding of the different parameters
and aeroacoustic mechanisms responsible for airframe noise and the
feasibility of reducing principal sources of this nonpropulsive noise. Ac-
cordingly, an assessment of recent research progress and state-of-the-art in
the airframe noise area, as reflected by the fourteen papers mentioned in the
preceding paragraph, revealed significant progress on most aspects of this
aeroacoustics problem.
Noteworthy is the progress being made to develop prediction methods for
the calculation of airframe radiation noise into the far field. Noise
generation by the interaction of airflows and surfaces has been researched
over many years; however, studies of the far-field noise generated by an air-
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Ira R. Schwartz
Editor
Associate Editors
February 1976
AIAA 2nd AEROACOUSTICS CONFERENCE
March 24-26,1975 Hampton, Va.
Conference Committee
General Chairman
Ira R. Schwartz
NASA Ames Research Center
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
A. L. Abrahamson*
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Abstract
I. Introduction
Wind-Velocity Gradients
When air flows uniformly over a large, solid, and flat surface,
there is a smooth transition from zero velocity at the surface of the
solid to the freestream flow velocity some distance away. A sound
wave in such a boundary layer propagates so that its rays follow
curved instead of straight paths. In the case of sound propagating
in an upwind direction, as depicted in Fig. 1, this gives rise to a
shadow zone some distance from the source. From Huygen ? s
principle (that all points on a wavefront act as secondary sources),
SOUND PROPAGATION OVER LONG DISTANCES
Shadow Zone
llllli:!0pll;li!l:
Receiver i;
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
it may be seen that diffraction occurs into the shadow zone, and,
instead of a sharp cutoff, the sound intensity decays exponentially
into the shadow. In the case of sound propagation downwind, a
downward bending will occur, and, apart from following a slightly
longer path, the effect on the intensity of direct sound reaching the
receiver will be slight. Sound directed toward the surface will be
reflected, however, and contained in multiple hops to augment direct
sound at the receiver. In the presence of a nonmonotonic gradient,
focusing may occur, and direct sound intensity downwind from the
source may be amplified substantially.
Temperature Gradients
in sound waves being bent upward (Fig. 1), whereas a positive tem-
perature gradient will result in sound waves being bent downward.
Turbulence
Nonturbulent Region
Turbulent Region
Receiver
Fig. 2 Effect of atmospheric turbulence on low-level sound propa-
gation.
SOUND PROPAGATION OVER LONG DISTANCES !
Turbulent Region
Receiver
RMS Sound errain Wind
Pressure With Scattering and Velocity
Level at
Surface ____Ground Absorption Profile
- _ _ _ No_Sjcatteringj3r
____Ground Absorption
DISTANCE
Fig. 3 Effects of wind refraction, scattering, and ground absorption
on sound propagation: downwind approach.
8 A. L. ABRAHAMSON
Two flight paths were chosen, from due east and due west, to
coincide with the direction of the prevailing wind. Three noise
measurement stations were used for each flight path, with distances
between stations ranging from a minimum of 7300 ft to a maximum
of 12, 700 ft. Terrain in each direction was reasonably flat and
homogeneous, with grassland and tidal mudflats to the east, and
ploughed fields to the west.
10 Frequency 63 Hz /
Sound Propagation Upwind /
o
o
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Best-Fit Line
(for this point set)
Model Line
(from Figure 6)
0 I
0 1 2 3 4
Excess Atmospheric -Attenuation (dB)/lOOO ft
Fig. 4 Effect of wind velocity on excess atmospheric attenuation
from measured one-third-octave data: upwind.
Frequency 63 Hz
10
Sound Propagation Downwind
o
<D
CQ
<w-l
>>
CJ
Best-Fit Line
(for this point set)
Model Line
(from Figure 7)
0 \
0 1 2 3 4
Excess Atmospheric Attenuation fdB)/lOOO ft
Fig. 5 Effect of wind velocity on excess atmospheric attenuation
from measured one-third-octave data: downwind.
10 A. L. ABRAHAMSON
5 0 *
0>
a
0 Mean A = 0. 06
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
0 G
100 1000 10,000
Frequency in Hertz
1.0
Intercept (I3) vs Frequency (f)
o o
ex
(L)
O Mean B = 0. 36
0
100 1000 10,000
Frequency in Hertz
Fig. 6 Results of regression analysis on excess atmospheric ab-
sorption for sound propagation: upwind.
.2
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
V. Analysis of Results
tances. Since the test was concerned with propagation over very
long distances, noise levels only marginally higher than the pre-
vailing ambient noise were involved. Extraction of acoustic propa-
gation data under these conditions required the assumption of a
moderately stationary natural ambient.
The model derived from the regression analysis for angles less
than 2 reduced to the following:
E = Au+ B
E = Cu + D
C = 0.06
D = 0.36
SOUND PROPAGATION OVER LONG DISTANCES 13
f = frequency, Hz
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
2 or less, a simple power law with distance from the source was
found to represent "excess" sound attenuation reasonably.
Although little other data exist for sound propagation over dis-
tances in the range 8000 to 50, 000 ft, a cautionary note should be
sounded in the indiscriminate application of the model contained
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
References
P. E. Tubb*
Abstract
Nomenclature
17
18 P. E. TUBB
AD
0.940
0.1 ^p
UL
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Introduction
Overflights of supersonic aircraft give rise to transient
pressure waveforms at ground level known as sonic booms. The
predicted pressure signature at ground level for a steady at-
mosphere has a characteristic N shape as shown in Fig. 1. The
main parameters characterizing the waveform are its peak over-
pressure Ap, its rise time to, and its duration T, which for a
supersonic transport (SST) have typical measured values of
2 psf, 300 msec, and 1 msec, respectively.1 Experimental data
such as those presented in Ref. 2 indicate that the predicted
signatures are essentially correct with respect to gross fea-
tures of shape and duration; however, measured rise times are
typically 1 to 10 msec, which is of the order of 1000 times
greater than the theoretical Taylor value of 2.7 jasec for a
shock with an overpressure of 1 psf. 3 [in Taylorfs theory,
the steepness of a shock front is governed by a balance between
nonlinear steepening and diffusion due to viscosity and heat
conduction. He derived the formula Ji ~ (Ui-Us)"1 cm, where Ji
is the shock front thickness in centimeters, Ui is the velocity
of the gas entering the shock, and (Ui-U2) is the particle
velocity in centimeters per second behind the shock in a shock-
fixed reference frame. Shock thickness and rise time are
equivalent concepts in that the rise time is the time required
20 P. E. TUBB
the anomalously long rise time and the distorted shape of sonic
booms. Crow^?5 and Pierce^ have attributed the spiking and
rounding to turbulence in the planetary boundary layer. The
mechanism proposed by Crow was the interaction of the weak
shocks with turbulence to produce high-frequency scattered
waves that arrived behind the shocks and distorted the wave
structure. Pierce, on the other hand, attributed the distor-
tion to rippling of the wavefront due to local focusing and
defocusing by large turbulent eddies. He argues that focus-
ing will produce folding of the wavefront, resulting in a
double shock (spiking), whereas defocusing will reduce the
strength of the shock (rounding). Ribner et al.7 independently
put forth a similar argument to explain the spiked and rounded
booms based on earlier experiments^ on focusing and defocusing
of sound rays by jet flows. They argued that local regions in
large-scale (e.g., 200 ft) eddies of atmospheric turbulence
were somewhat jetlike: the experiments of Ref. 8 suggested a
focusing effect (spiked boom) when the sonic boom propagated
counter to the jet flow, and a defocusing effect (rounded boom)
when the boom propagated with the jet flow. Very realistic
spiked and rounded booms were produced when these notions were
tested in the- laboratory with air jets and simulated sonic
booms from a shock tube.7
Experiment
Method
Two main elements were necessary for this investigation:
l) turbulence, and 2) a shock wave. Turbulent flow was gener-
^- Oscilloscope
Ai
Fig. h Schematic of experimental setup
22 P. E. TUBE
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
i
i-
Fig. 5 Relation of shock tube, duct, and micro-
phones (typical measurements).
ated by placing obstructions at the exit of a wind-tunnel duct,
and a cons tant-temperature anemometer was used to record typi-
cal values of the mean flow velocity U and the root-mean-square
axial component of the turbulent velocity u f . A weak shock,
produced by a shock tube, was passed at right angles through
the turbulent flow and the resulting pressure signature re-
corded using two microphones connected to an oscilloscope
equipped with a Polaroid camera. The rise times of these
shocks then were measured and compared to the rise times of
shocks recorded in the absence of turbulent flow. A schematic
of the experimental arrangement is shown in Fig. k, and some
typical measurements relating the positions of the duct, shock
tube, and microphones are given in Fig. 5.
THE RISE TIME OF A WEAK SHOCK 23
Experimental Considerations
form to rise from 10f0 to 90fo of its peak overpressure (Fig. 2).
Thus, the experimentally measured rise time r will be equal to
t0 only if the pressure rise is linear, but the error will
usually be less than that involved in trying to measure t0
directly.
Effects of Mean Flow. As mentioned earlier, the experi-
ments of Ref. 7 demonstrated that an N wave passing through
the velocity gradients of a jet blowing counter to or with
THE RISE TIME OF A WEAK SHOCK 25
1
I I
/Y
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
1 I
I
Duct
3" 4
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
(a.) 3 0. D. Tu-tin
O. D. A l u m i n u m Tubin
I- Wooden Dowels
binj
m
7_ 3 O - 0 . A l u m i n u m Tutbin V___l" x l" Pine Strapping
2O
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
3O
36 Occurrence
90
85
8O
J-I
: :: 5
,
2O rWD
35 40 45 SO 55 6(.
15
10 *o i\?,
\
\
' I
5 I f>
VE
WI ^
c> 5 IO 15 20 25 '30" 35
T^usecJ
Fig. 11 Distribution of reference rise time (rr)
and rise time after passage through turbulence
(T^) for shocks produced using "red zip" diaphragms<
32 P. E. TUBE
% increase in rms
rise time 30 136
Abstract
Introduction
When an acoustic wave propagates through a moving inhomoge-
neous medium containing random velocity and temperature fluctu-
ations, the acoustic energy is scattered, and wave amplitude and
phase show random fluctuations. Theoretical descriptions of this
phenomenon have been derived by a number of investigators over the
past 30 years.
35
36 M. N. HUAN6
Continuity Equation
+ (S/dx.Mpu.) - 0 (1)
Momentum Equation
(c)/St)(pu.)
1 + (d/dx.)(pu.u.)
l = t.. . (2)
3 3 i 3 >3
Energy Equation
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
where
There are six unknown variables (p, u., P, T), but Eqs. (1-3)
represent only five equations. [Equation (2) represents three equa-
tions for i = 1, 2, 3).] The sixth equation is the equation of state
for an ideal gas:
P = pRT (5)
38 M. N. HUANG
Using Eq. (8) and the gas law to eliminate p from Eq. (1), the
modified continuity equation becomes
The basic flow quantities can be written for the combined sound
and turbulent field in the following form:
p = p + p! + p", u. = u. + u! + uV
'2
Since in the atmosphere P T ~ pu ( , the ratio of turbulent, fluctu-
ating pressure to the mean pressure is second order in turbulent
Mach number. Density fluctuations may result from both velocity
and temperature fluctuations independently of each other. Plotkin
showed that density fluctuations due to velocity variations are
second order in 61 . Those due to thermal turbulence are first order
in &2' In any given situation, it is difficult to determine which mech-
anism is producing the density perturbation. In this study, the
larger-order, thermal effect always is assumed. Pressure tur-
bulence cannot exist independently from velocity turbulence and is
characterized by M|. For the present case, where ii| ^ 0, pressure
turbulence also may be characterized by MM^-, where M is the Mach
number of the mean flow. This follows from computing (u^ + u| ) %.
An additional condition required for neglecting pressure turbulence is
thus that M be small.
Applying d/dx. and D/Dt to Eqs. (15)and (16) and combining the
results, which reduce to
SOUND SCATTERING FROM TURBULENCE 41
where
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
+ u"u f 1 /C 2 (20)
J i,ijJ o
In writing Eq. (20), the results of assuming uniform mean flow and
neglecting higher-order terms of 6 in Eqs. (13) and (14) have been
used. Further simplification is achieved if the turbulent flowfield
is assumed incompressible (i.e., u? . = 0). In this analysis, the
term involved in u{ ^ is retained so that the eventual solution dis-
plays the effects of compressibility.
x. = x. - u.t, =t (21)
1 1 1 '
gives
42 M. N. HUANG
u"c>u'/3xc>x
... _ ... , _. , / d x i -f u"r (22)
J ^ 3 J i i jJ
To derive an expression for the scattered sound field, we expand
TT and u'/into series of the parameter 6, i. e. ,
flowfield, and fti , V- . are associated with the scattered wave due to
the turbulent flowfieicL To be consistent with the wave equation (22),
which is valid only to the first order of 6, terms containing the sec-
ond or higher order of 5 in Eqs. (23) and (24) are dropped. With
the neglect of higher-order terms, this method of perturbation be-
comes the single scattering or "Born" approximation.
(26)
oi
/at) -f c2o (Siro/&xi > = o (28
2 2
(3V (T f /T)TT + V u'
li o ,i oj i, j
+ uf V (30)
J 01, j
lated to the acoustic velocity in the classical way by PC^. The same
result can be found by eliminating Tf from Eqs. (29) and (30) and
solving for V . in the far-field position.
TT ( x.,
x.,0 -
= A eexp^
x p i o ) t -- i K .x
. .1 (33)
0\ I / ! O 01 iJ
i
where eA . = K
^ /
./K .
Ol Ol O
-a
The relation between KQ and K^ is derived by utilizing SnelPs
law and using the boundary condition, i. e. ,
- M (f - r T )] | d 3 r T (41)
where Q(r,t) represents the terms on the right-hand side of Eq. (26).
The factor | r - F T | - M ( r - r T ) is due to changing from a
moving frame to a fixed frame.
r /\ /\
A
r - rf iK r - r f
A A A .<">
e o e o <L__ e
iK X
si i (44)
r - rr r - rr
where 9 is the angle between the incident wave K Q and the scattered
wave vector K g . Inserting expressions Eqs, (43) and (44) into Eq.
(42) and performing integration by parts in the limiting case of
large r and small scattering volume, the following equation is
obtained:
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
P r A A ,_ A / ^ AAA A2
f
X \ I (K K )(T /T + 2e .u!/C ;) + v(K . - K .)(K: .K .- K 7)
jv ! L v s M
O oj j oM
si OI oj sj o
/ -A- ~l r ^ A A ~ | SA
x u r /C 1C exp i(K
v - K )xM d r T (45)
i o oJ L si 01 iJ
From Eq. (17), the sound scattered pressure P and If are re-
lated b y s i
VPC <47>
Therefore, Eq. (45) gives the scattered pressure in the fixed
frame system, and the underlined term is due to the compressibility
of turbulent flow. By the definition of acoustic intensity as shown
in Tatarski,2 Eq. (48) is written:
I =(pCjV2)(Tn7*) (48)
SOUND SCATTERING FROM TURBULENCE 47
and
U!(?,t) = \ dt.(K) exp il K - r - oj"]
where dfc and d\|f. are the spectral densities of temperature and ve-
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
and
U!(r, t) = ^ dt.(K) exp i f K . r - K ut - uu(t - |r - r T | /C )~| (49)
Using the above relations, and inserting Eq. (45) into Eq. (48)
produces
I = { TT pC 3 A2V/[ 4r^l - Mcos(9 -0) } 1 } (K K )2
L o L \ / J J o s
X
[ n^ + ^oiViJ^b] (50)
scatterers to the energy received bv the same set per unit volume.
That is a - tr2/! V, where V =fy d^r = V [1 - Mcos(9 -fl)]""1*.
_ 3 2
If the incident acoustic intensity is I0 = pC 0 A /2, the effective
scattering cross section per unit scattering volume per unit solid
angle becomes
where
v
oi(xi' *> = Gi<x2)exp i
D2 TT /Dt2 - C 2 TT .. = C 2 (T'TT
v ./f)7 . + 2!" v(uJV . .)7 .
1 o ,11 o o.i ,i L j 01, j ,1
+ u' 5 0TT 1 (60)
2 1, 2 o,iJ
The incompressibility of turbulent flow is assumed in the preceding
equation. Again using Eq. (41), the scattered pressure IT is ob-
tained by introducing expressions (56) and (57) into the right-hand
side of Eq. (60).
Discussion
The structure of atmospheric turbulence can be investigated by
using sound scattering as a tool. Consequently, the more precisely
the calculations of scattered sound intensity are obtained, the more
valuable the available information to find the turbulent character-
istics becomes. The principal objective of this analysis is to pro-
vide accurate prediction of scattered sound energy. Although the
scattering cross section in this work gives the same result as that of
Ref. 1, a different approaching method is used. The restrictions
on the incident-wave propagation angle and receiver angle in Ref. 1
are sin20 > > 2Mcos0 and sin2(9 -0)> > 2Mcos(0 -0). In this
analysis, the angles are less confining, i. e. , 0 > sec""*(l + M) and
and (9 -0) > sec -1 (1 + M).
i I _
2 2 2
[2
L d(u! . . V .c) u!/^x l /C \ v(61)
;
3 01 3 i 03 j iJ o J
where $ = V . The preceding equation can be solved for $ by
using Eq. (41)' again. Finally, the acoustic velocity can then be de-
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
P. L. Chow
Wayne State University, Detroit, Mich.
and
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Abstract
Introduction
51
52 P. L. CHOW, C. H. LIU AND L. MAESTRELLO
Governing Equations
, V) ' = 0 (2)
where
fu V 0 ypV \
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
, V) = 0 u V pV .
V-
\ V 0 u V/
/Y(V u) 0 (vP) A
0 V u (Vp)
V 0
7*
(vG) -y
and y is the ratio of constant specific heat.
Formulation of Problems
problem (3) and (4), one needs to determine the Green's matrix
G, which satisfies the following adjoining equation:
+
f n G (t,s;r,p ' )<!> M(p') ds dp'
J/n J/
0 z -0 ^ 1 ^
+
/O ^z'>0 G4"(t,s,r,r')F(s,r') ds dr' (7)
t=0 ~ ^
(10)
V = A($,V) + B S ) (12)
where the operators A and B are defined by Eqs. (3) and (4),
respectively. A^($) is similar to A($) except for changing
the element (l/p)V to (-p/<p>2) , and B^($) is the same as
B($) with the element (-l/p2)Vp replaced by (2p/<p>3)V<p>
- (l/<p>2)Vp. Then an application of the smoothing method
yields the following matrix integrodifferential equation:
In view of (11) and (12), Eq. (13) can be written more explic-
itly as
{9/3t (t,r)
+ <A($,V)(t,r")Go(t,s;r,r"')B($)(s,r"')>
+ <B($)(t,r)Go(t,s;r",r')A($,V)(s,r"')>
+ <B($)(t,r)Go(t,s;r,r')B(i)(s,r')> (16)
= G = 0 (17)
t=0 z=0
The symbols A*, B* designate the adjoint operators of A
and B, respectively. As demonstrated previously, the half-
space problem [Eqs. (3) and (4)] can be reduced to a full-space
problem [Eq. (8)]. For each elemented solution, the averaged
problem [Eq. (7)] in a half-space is also reducible to a full-
space problem by a symmetrical extension of the functions A,
B, F, R, G0 about z = 0 and the method of images. For the
sake of brevity, however, the details are not presented.
and
- IX
(22)
z=0
(23)
where use has been made of (21) . An equation similar to (22)
can be obtained after a Fourier transform in p.
Special Cases
Case 1: Uniform Mean Flow
u (z) - + V<p>
z p
TURBULENCE
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
1- yp V
dz o
- V<p>
dz
- 2 r}
3P" 0 (27)
where
!% c V < Pr> - i dz
, 2 + o
L /JRzz 3z
a g 11 3az T <rp > d ; .
dz
2
- /R & ^ d? ' = (28)
In the derivation of Eq. (28), use is made of the fact that the
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
A= 1/2
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Z=
a = l, (3 = 0.01, k = 10, o = 0
0 10
x.
Fig. 2 Amplitude of wave in shear flow.
2r
ko = 1
A=
a/2
a = l, (3 = 0.1, k = 10,
0 10
x .
Fig. 3 Amplitude of wave in shear flow.
62 P. L. CHOW, C. H. LIU AND L. MAESTRELLO
2r
AA = (pp
/ *
)
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
a = l, (3 = 0.01, k = 10
0 1 2 3 4 5
x
Fig. 4 Amplitude of wave in shear flow with turbulence.
2r
ko = 1
A = (p'ff*)1/2
a = l, 3 = 0.1, k = 10,
0 10
Fig. 5 Amplitude of wave in shear flow with turbulence.
SCATTERING OF COHERENT SOUND WAVES 63
0- kx
(rad)
a = l, (3 = 0.01, k = 10, 0 =
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
-3
2 4 6 8 10
x
Fig. 6 Deviation of phase from plane wave in shear flow.
ka T
z=3
0 = 1, p = o.01, k = 10,
0 - kx z=2
(rad)
0 2 4 6 8 10
x
Fig. 7 Deviation of phase from plane wave in shear flow with
turbulence.
For the purpose of comparison, the amplitude for the
acoustic wave in the mean flow is plotted in Figs. 2 and 3 for
3 = 0.01 and 0.1, respectively. The effects of turbulence on
the wave amplitude A at different heights are shown in
Figs. 4 and 5. Near the boundary, there exists a strongly
damped layer because of the attenuation effect by turbulent
scattering. Away from the wall, the wall intensity is ampli-
fied by the mean flow and reaches a maximum and then decays
along the downstream as a consequence of random scattering,
which also shifts the points of maxima along the curves. In
Figs. 6 and 7, the deviations of phase from that of the plane
64 P. L. CHOW, C. H. LIU AND L. MAESTRELLO
Concluding Remarks
Appendix
Correlation Matrix R
/R R R -\
/ P1P2 P P
1 2 P
1 U 2\
R R R - (Al)
P P P
12 1P2 P
1U2
\ R-
K- R- R- -
SCATTERING OF COHERENT SOUND WAVES 65
where
R R R R (A2)
uu yx yy yz
\ R R R
\ zx zy zz //
References
9
Maglieri, D. J., "Some Effect of Airplane Operations and
Atmosphere on Sonic Boom Signature," The Journal of The Acous-
tical Society of America, Vol. 39, No. 3, March 1966, pp. 536-
542.
Alan R. Wenzel
Abstract
Introduction
67
68 A. R. WENZEL
Basic Equations
2 2
L = V +k ,
L
i = -2(kou -lkoVj)
and kg = (JO/CQ.
Coherent Wave
where
in = p(s)ds ,
1
0
, = r t(s)ds ,
/
v p(s) =
V2
l(6ij "s"2sisj)ST'(s)
(The prime denotes differentiation.) In introducing this nota-
tion, we have assumed that the turbulence velocity field is
incompressible as well as isotropic. ^-2 The function T(S) is
the normalized longitudinal velocity correlation of the turbu-
lence. In the derivation of Eqs. 10 and 11, terms of order
e^, as well as higher-order terms in (kQ&i)~l and (kQ^)"-'-*
were neglected.
where now
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
kQ + ia (13)
and
e\^
* ^-n (o^
I^ ^i + vv^
-^o /) C14}
v-1
-^/
Incoherent Wave
q(r) -
+(r!/rt)u.(rt)](rT)~1exp(-art)drt . (15)
(In deriving Eq. 18 we have made use of the fact that correla-
tions between scalar and velocity fields vanish for isotropic,
incompressible turbulence.) The double integral in Eq. 18 can
be evaluated approximately by making the coordinate transfor-
mation 5 = r^ + r^1, n = r^ - r^', and noting that, for
r >>
^1 2> t^ie resulting n integration can be extended to
infinity without introducing significant error. The trans-
formed integral is then easily evaluated, after which we obtain
This result is the same as that obtained from the Born approx-
imation and shows that, when the propagation distance is small,
74 A. R. WENZEL
References
3
Yura, H. T., "Physical Model for Strong Optical - Amplitude
Fluctuations in a Turbulent Medium," Journal of the Optical
Society of America^ Vol. 64, Jan. 1974, pp. 59-67.
4
de Wolf, D. A., "Saturation of Irradiance Fluctuations Due to
Turbulent Atmosphere," Journal of the Optical Society of Amer-
ica3 Vol. 58, April 1968, pp. 461-466.
SOUND PROPAGATION IN A TURBULENT MEDIUM 75
5
Brown, W. P., Jr., "Fourth Moment of a Wave Propagating in a
Random Medium," Journal of the Optical Society of America,
Vol. 62, Aug. 1972, pp. 966-971.
6
Sancer, M. I. and Varvatsis, A., "Saturation Calculation for
Light Propagation in the Turbulent Atmosphere," Journal of the
Optical Society of America, Vol. 60, May 1970, pp. 654-659.
7
Frisch, U., "Wave Propagation in Random Media," in Probabil-
istic Methods in Applied Mathematics,, A. T. Bharucha-Reid,
editor, Academic Press, New York, 1968, p. 114.
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
8
Wenzel, A. R. and Keller, J., "Propagation of Acoustic Waves
in a Turbulent Medium," The Journal of the Acoustical Society
of America3 Vol. 50, Sept. 1971, pt. 2, pp. 911-920.
Abstract
Introduction
77
78 RIBNER, LAM, LEUNG, KURTZ, AND ELLIS
Oblique Angles
Fig. 5 Wave steepening assumed in the
computer generation of thunder.
MODEL OF LIGHTNING THUNDER PROCESS 83
Results
| 100ms _>)
Pressures in
2 <*-*^ Arbitrary Units
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
-1 H-
Fig. 6 Computer-generated "thunder" (upper
trace) and pressure-time signature (lower
trace4) of an example (not the same) of
real thunder (hand traced).
Conclusions
References
John P. Raney*
Abstract
Introduction
93
94 J. P. RANEY
[AIRCRAFT MEASURED
MODELING FLYOVER
DATA DATA
AIRCRAFT
MANEUVER
DATA
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
/ INPUT
X TO"B"
WYLE
DATA
BASE
B
NEF
w
X Y NEF
PROGRAM
N L PROGRAM
E
NOISE
GRID NOISE
GRID
PRINT
GENERAL L NEF
CONTOUR R CONTOURS
PROGRAM C
ANOPP Development
Prediction Methodology
Potential Users
PROPAGATION
JET NOISE
BLOWN FLAPS
FAN/COMPRESSOR NOISE
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
ROTOR/PROPELLER NOISE
NASA
FAA
DOD
Engine
manufacturers
Airframe
manufacturers
EPA
Airport managers
Consultants
98 J. P. RANEY
(SINGIE EVENT)
AF-BBN
DOT-WYLE
AIRCRAFT DATA BASE: EFFECTIVE NOISE LEVEL (EPNL, ETC.) VS SLANT RANGE
STRAIGHT LINE SEGMENTS DESCRIBE AIRCRAFT OPERATIONS
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
AIRCRAFT DATA BASE: NOISE LEVEL (dBA, PNdB, ETC.) VS RANGE AND DIRECTIVITY
SMOOTH, CURVING AIRCRAFT TRAJECTORY
ESTIMATES TIME HISTORY OF AIRCRAFT NOISE LEVEL AND PROVIDES REFINED
ESTIMATE OF COMMUNITY NOISE ENVIRONMENT
GENERAL KNOWLEDGE OF ACOUSTICS AND AIRCRAFT NOISE REQUIRED FOR USE
Fig. 5 ANOPP level II.
the needs of various user groups and to provide a self-
contained systematic means for validating and improving the
state of the art of aircraft noise prediction.
PROPAGATION
PROPAGATION
ANOPP Availability/Accessibility
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Summary
References
E. L. Zwieback*
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Abstract
The evolution of the noise measurement techniques reflect
the varied reasons for the measurements and also reflect the
advancements in electronic instrumentation. Current techniques
are described by a number of standard procedures, not com-
pletely definitive. The results of measurements by various
organizations may be, in some cases, noise levels which are not
comparable. Described are basic typical procedures utilized in
making comprehensive measurements. Differences in procedures
are indicated. General limitations include variations in
the f l i g h t effects of noise sources, in the atmospheric propa-
gation of noise, in the recorded noise signal near the ground,
and in data processing methods.
Introduction
Aircraft flyover noise measurements have been conducted
for approximately 30 years; the measurement objectives, equip-
ment, and procedures have grown increasingly complex to satisfy
a wide variety of requirements. Current measurement objectives
are*detailed acoustic source levels for research and develop-
ment; flyover noise levels to demonstrate compliance with
aircraft noise certification and operator specifications,
levels for development of noise-abatement operations, levels
over a wide range of aircraft operations for community noise
exposure forecasting, and monitored levels for airport oper-
103
104 E. L. ZWIEBACK
ONE-THIRD
OCTAVE-BAND
SPL (dB)
ARP 796 1965 SAE measurement of aircraft exterior noise in the field. X X
FAR part 36 1969 Noise standards: aircraft type certification. X X X
R507 1970 ISO recommendation. Procedure for describing aircraft X X X
noise around an airport.
R1761 1970 ISO recommendation. Monitoring aircraft noise around X
an airport. 3>
11
PUC title 4, 1970 California aeronautics noise standards X X X
0
subchapter 6 5
^
ICAO annex 16 1971 International standards and recommended practices: X X X n
-H
aircraft noise.
IEC 179 1965 IEC recommendation, precision sound level meters X ~n
r~
IEC 225 1966 IEC recommendation, octave, half-octave and third-octave X 0
band filters intended for analysis of sounds and vibrations. <
n
SI. 11-1966 1966 American national standard specification for octave, half- X 73
octave, and third-octave band filter sets. o
ARP 1080 1969 SAE frequency weighting network for approximation of X 11
CO
perceived noise level for aircraft noise. m
SI. 4-1971 1971 American national standard specification for sound level X
meters.
ARP 866 1964 SAE standard values of absorption as a function of tempera- X
ture and humidity for use in evaluating aircraft flyover
noise.
ARP 865A 1969 SAE definitions and procedures for computing the perceived X
noise level of aircraft noise.
ARP 1071 1972 SAE definitions and procedures for computing the effective X
perceived noise level for flyover aircraft noise.
<-i
108 E. L. ZWIEBACK
Measuring Equipment
AIRCRAFT OPERATIONS
DATA REPRODUCER
DATA ACQUISITION AND FLYOVER
AND PROCESSOR
RECORDING SYSTEM NOISE
DATA
RECORDING SYSTEM
RETROREFLECTOR
PORTABLE OPERATIONS
CALIBRATION FACILITY
AND CABLE ,, 1 ^ WEATHER STATION
-4*._JjL_
Q--^_ ^
.
NOISE TEST CENTRAL
RADIO COMMUNICATIONS
Measuring Techniques
Special airplane operating procedures have been developed
for noise measurements to acquire the variety of separate types
of test data required and also, in part, to m i n i m i z e high-cost
aircraft f l i g h t operations. To appreciate the magnitude of
aircraft noise test costs, the costs sometimes are indicated
in terms of the f l i g h t time required for testing. Compre-
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
AVERAGING ON
"PRESSURE" BASIS FOR
N INTEG PERIODS
IE
ADJUSTMENT OF INPUT
NOISE FOR
CALCULATION OF AMBIENT NOISE
CORRECTION TO _L
INCLUDE SYSTEM CONVERSION OF INPUT
FREQUENCY RESPONSE, NOISE TO CALIB SOUND
MIC PRESS. RESPONSE PRESS. LEVELS
ANDSPL REFERENCE JL
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
LEVEL CALCULATION OF
PNL VALUES
JL
CALCULATION OF TONE
CORRECTIONS AND
PNLT VALUES
CALCULATION OF
DURATION CORRECTION
AND EPNL VALUE
MOLECULAR 20
ABSORPTION, aMOL
(dB/1000 FT)
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
4 6 8 10 12 14
ABSOLUTE HUMIDITY (gm/m3)
Noise Recording
Recording noise signals accurately near the ground surface
requires considerable care if the test program objective is to
obtain noise data free of ground surface effects. On the
contrary, most measurement programs do not include this object-
ive, and the results are aircraft noise data combined with
particular ground surface effects. For hard reflecting
surfaces, the surface effects are quite large but fairly
straightforward. These surface effects are a function of sound
wavelength and multipath interactions. Straightforward adjust-
ments are possible assuming that the angles of incidence are
114 E. L. ZWIEBACK
ONE-THIRD 20
OCTAVE-BAND
LEVEL,
10
^^H
REFLECTION
^
DIFFERENCE -i r
(dB) 0
(DIFFERENCE
BETWEEN NOISEE SOURCE HEIGHT: 400 FT
-10
TOTAL AND
MICR OPHONE HEIGHT: 4 FT
DIRECT SPL's)
NOISlE SOURCE ANGLE: 65 DEC
-20 HARC) REFLECTING SURFACE
WHIT E NOISE SPECTRA
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
rv
ONE-THIRD 50 ^
OCTAVE-BAND \ AMBIENT
\ 1 <
SPL
40 \
(dB RE 2 x 10-5 \
k:^
N/M2)
30 TOTAL AMB IENT NO SE LEVE LS^y. ^ L/2-IN.-D A
FOR RECOR DING
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
MICROPh ONE
20 SYSTEM -\
*- h-J
10 /
l-IN.-DIA MICROPHONE ' ^^^^ ^-.-^^
5
50 125 250 500 IK 2.5K 5.0K 10K
GEOMETRIC MEAN FREQUENCY (Hz)
Fig. 7 Typical low-amplitude noise levels for aircraft
flyovers.
Data Processing
Data processing of the analog recordings of the transient
acoustic signals into integrated noise measures for reference
aircraft and weather conditions requires a number of different
types of data operations. One of the fundamental operations
involves the frequency filtering and conversion into root-mean-
square (rms) values of the analog acoustic signals. Filtering
may be done with analog or digital techniques. Significantly,
the specific shape of the one-third octave filters, always
nonideal, w i l l vary within a standard envelope. The nonideal
filter shapes can interact with the noise spectra character-
istics involving discrete frequencies, and the interaction can
be compounded by doppler-shift effects to produce tone levels
influenced by filter-band edge crossings. The narrow-band
spectral time history of flyover noise signals (during a land-
ing approach) shown in Fig. 8 illustrates the concept of the
doppler-shift/fiIter-band interaction. Standard band edge
frequencies of some of the one-third octave filter bands are
indicated in the figure.
116 E. L. ZWIEBACK
1/3-OCT
SOUND
PRESSURE LEVEL
(dB)
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
2
1/3 OCTAVE-BAND SIGNAL DATA PROCESSING
NOISE LEVELS
CHANGE WITH
AVERAGING TIME
CHANGE (dB/SEC)
0 10 20 30 40 50
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
FLYOVER DIRECTIVITY
ANGULAR RATE CHANGE
AT OVERHEAD DURING DATA
2
McPike, A. L., "Recommended Practices for Use in the Measure-
ment and Evaluation of Aircraft Neighborhood Noise Levels,"
Paper 650216, 1965, Society of Automotive Engineers, Warrendale,
Pa.
^Committee A-21, "Comparison of Ground-Runup and Flyover Noise
Levels," AIR 1216, A p r i l 1972, Society of Automotive Engineers,
Warrendale, Pa.
5
Committee A-2I, "Measurements of Aircraft Exterior Noise in
the Field," ARP 796, 1965, Society of Automotive Engineers,
Warrendale, Pa.
6
"Part 36 Noise Standards: Aircraft Type Certification,"
Federal Aviation Regulations, 1969, Federal Aviation Admin-
istration, Department of Transportation, Washington, D.C.
7
Bishop, D. E. and Galloway, W. J., "Community Noise Exposure
Resulting from Aircraft Operations: Acquisition and Analysis
of Aircraft Noise Performance Data," Rep. AMRL-TR-73-l07,
Sept. 1974, Aerospace Medical Research Laboratory, Wright-
Patterson Air Force Base, Ohio.
8
"Aircraft Noise," International Standards and Recommended
Practices, Annex 16 to the Convention on International C i v i l
Aviation, 1971, International C i v i l Aviation Organization,
Montreal, Canada.
^Zwieback, E. L., "Flyover Noise Testing of Commercial Jet
Airplanes," Journal of Aircraft, Vol. 10, No. 9, Sept. 1973,
pp 538-545.
10
Auzolle, S. and Hay, J., "Method of Measurement and Analysis
of Noise of an Aircraft in Flight," Tech. Transl. TTF-14, 058,
Dec. 1971, NASA, Washington, D.C.
11
"Aircraft Noise Measurement Systems," Request for Proposal
WA5R-4-0095, Dec. 1973, Federal Aviation Administration,
Washington, D.C.
12
M i l l e r , R. L. and Oncley, P. B., "The Experimental Determin-
ation of Atmospheric Absorption from Aircraft Acoustic F l i g h t
Tests," Contractor Rep. CR-I89I, Nov. 1971, NASA, Washington,
D.C.
AIRCRAFT FLYOVER NOISE 121
13
Smith, C. M., "Atmospheric Attenuation of Aircraft Noise, M
Rep. HSA-HAD-R-GEN-214, Sept. 1973, Hawker Siddley Aviation,
Ltd., Hatfield Herts, England.
14
Committee A-21, "Acoustic Effects Produced by a Reflecting
Plane," Proposed AIR 1327, A p r i l 1973, Society of Automotive
Engineers, Warrendale, Pa.
15
Stouder, D. J., "Use of Digital Averaging Techniques for
the Analysis of Aircraft Flyover Noise," Proceedings of
Inter-Noise 74, 1974, Washington, D.C., pp 137-140.
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
16
"Aircraft Sound Description System," Federal Aviation
Administration Order, Aug. 1973, Federal Aviation Adminis-
tration, Department of Transportation, Washington, D.C.
17
"Noise Assessment for Land-Use Planning," Circular
II60AN/86, 1974, International C i v i l Aviation Organization,
Montreal, Canada.
18
Seykra, C. A., Storey, W. C., and Yates, R., " V a l i d i t y of
Aircraft Noise Data," Journal of the Acoustical Society of
America (Abstract only), Vol. 55, No. X6, Suppl., A p r i l 1974,
pp S48.
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
*
Paul B. Oncley
MAN-Acoustics and Noise, Inc., Seattle, Wash.
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Abstract
Sound transmitted or received near a surface is modified
by an interference pattern that is determined both by the geom-
etry and by the acoustic constants (complex impedance and prop-
agation constant) of the surface. The classifical treatment,
assuming a point source and single frequency, has been expanded
by Howes and by Hoch and Thomas to include extended sources and
various types of filters including one-third octave. It is not
adequate, however, to deal with the frequent cases of propaga-
tion at near-grazing incidence. The analyses of Rudnick and of
Ingard are examined and linked with the Hoch-Thomas equations.
Several more recent approaches are discussed briefly, but for
experimental verification better information on the soil con-
stants is required. The available data are discussed, and
methods for improving the data are suggested. Advantages and
questions associated with microphones on the surface are given,
and it is shown why the foregoing analyses have been more suc-
cessful in extrapolation and noise prediction than in correct-
ing experimental data. Experimental and theoretical results
are compared.
I. Introduction: Classic Theory
The analysis of sound propagation close to Earth's surface
is of great practical importance and presents many difficulties
because of the irregularities and inhomogeneities both in
Earth's surface and in the air immediately above the interface.
123
124 P, B. ONCLEY
= [D2 + (hs-hr)2]%
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
SOURCE
RECEIVER
zsin* - [1 - (k/k9)2cos2i^
R _ __________^ /2)
p
zsin* + [1 - (k/k2)2cos2^
f0 - Dc/4hshr (3)
X- IU 1 1 i 1 1 1 i ] 1 i 1 1 1 1 i 1 1
' ' i 111
3
'A A
""^N/^^\
/ \ /
I: o / \ V
1/
00
0-
/ \i i
5 ? -10 - \/\ / i
S r^
*~ SJ-20 -
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
30
T 2
!
0 100 10 00 i0,000
Frequency (Hz)
Fig. 2 Calculated pressure level deviation from field
conditions; r=200 ft, hs=hr=7.07 ft,is surface
equivalent to concrete, is surface equivalent to
Quietone.
rfh
Jf Dw df
C(T) = (5)
VF) df
The time lag T is the sound transit time across Ar, or
T = Ar/C. For noise that is relatively uniform across the
band, w(F) may be replaced by a constant WQ so that Eq. (5)
becomes
sin [>T(f-f )]
C(T) COSTTT (fk+O
D a (6)
PREDICTION OF GROUND EFFECTS 127
Values of Af/f-j are 0.707 for octave bands and 0.2316 for
one-third octaves, giving a = 0.725 and 3 = 6.325 for one-third
octave measurements or, in terms of k.Ar (k. = 27rf./C),
AN = 10 log 1Q {1 + S2 +
+ 2S[sin(0.1158k 1 Ar)/(0.1158k i Ar)]cos(1.00668k i Ar)} (8a)
[1 + (r1/r2)|Rp|exp(ikAr + 16)]
AN = 10 log 1(J {1 + S 2 | Q | 2 +
i r^
Here | Q | e is equated to the reflection coefficient
1>6
Rp|e . When constant proportion bandwidth filters (such as
one-third octave) are used, Eq. (8) is modified similarly:
N = 10 log 1Q {1 + S
+ 2S|Qi|[sin(aAr/X.)/(aAr/Xi)]cos[e(Ar/X.) + 6-]} (10)
The subscripts i indicate that |Q-j|, x-j, and 6-j are evalu-
ated at the geometric center frequency of the respective filter
bands.
where
and
r> o
12kr [l -
9
(1-Rp) cos > z
F(w) can be expanded to give
F(w) = - [-1
2w (2wr (2w)
Although Rudnick's complicated expressions were formidable
in 1947, they are handled readily by modern digital computers.
Moreover, it readily can be seen that the coefficient Q in e-
quations such as (9) and (10) actually describes the entire
reflected signal, and for the spherical wave analysis we can
define
Thus the previous equations can be used directly for the analy-
sis of one-third octave band data and where extended source is
considered.
It is useful to consider the behavior of the "numerical
distance11 w and the "boundary loss factor" F(w). For example,
w is directly proportional both to propagation distance r2 and,
through k = 2-rrf/c, to frequency. For near-grazing cases,
cos2\|; approaches 1 and (1 - RD)2 approaches 4, and for most
130 P. B. ONCLEY
0.004
M
Fig. 3 Graph of the reflection function F(w) as
a function of the numerical distance w. Magnitude
of F(w) vs magnitude of w; phase of w as a para-
meter.
PREDICTION OF GROUND EFFECTS 131
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
surface.
I l T i l l
2 Re(z)
0
o
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
T- O
+> ~<-
<a
T3
0)
Q_
-6
-8
IM(z)
+5
QQ
s o
-5
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
References
, J. W. (Lord Rayliegh), The Theory of Sound, Dover,
New York 1945, Vol. 2, p. 78.
2
Howes, W. L., "Ground Reflection of Jet Noise," TN-4260,
April 1968, NACA.
3
Hoch, R. and Thomas, P., "Influence des reflexions sur les
spectres de pression acoustics des jets," 1st Collogue
d'Acoustique, 1968, Toulousse.
4
Rudnick, I., Journal of the Acoustical Society of America, 19,
348 (1947). ~~~
Sommerfeld, A., Annalen der Physik, 28, 665 (1909); also 81,
1135 (1926). ~ ~
6
Van der Pol, B., Physica 2_, 843 (1935).
Norton, K. A., Proceedings of the Institute of Radio Engi-
neers, 24^ 1367 (1936); also 25, 1203 (1937).
o
Ingard, U., Journal of the Acoustical Society of America, 23,
329 (1951). ~
g
Lawhead, R. B. and Rudnick, I., Journal of the Acoustical
Society of America 23_, 541; also Z3, 546 (1951).
10
Delaney, M. E. and Bazley, E. N., Journal of Sound and Vibra-
tion 13_, 269 (1970).
]1
Wise, H., The Bell System Technical Journal 8, 662 (1929).
PREDICTION OF GROUND EFFECTS 137
12
Piercy, J. E. and Embleton, T. F. W . , "Effect of Ground on
Near-Horizontal Sound Propagation," Automotive Engineering Con-
gress , 1974, Paper 740211, Society of Automotive Engineers.
13 Pao, S. P. and Evans, L. B., Journal of the Acoustical Socie-
ty of America 49_, 1069 (1971).
Aylor, D., Journal of the Acoustical Society of America 51,
197; also 5j_, 414 (1972).
Wenzel, A. R., Journal of the Acoustical Society of America
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
55_, 956 (1974); also Oncley, P. B., Journal of Sound and Vibra-
tion J3_ (1), 27 (1970).
16
Walker E. J. and Doak, P. E., "Effects of Ground Reflection
on the Shape of Sonic Bangs," 5th Internation Congress on
Acoustics, 1965, Paper L-55, Liege.
Abstract
Background
139
140 J. D. REVELL, G. J. HEALY AND J. S. GIBSON
ratio aircraft (an F106B delta wing,^ and the Jetstar-^) seem
to indicate lower noise levels than earlier predictions would
imply.3>13 ^e present paper describes recent developments
in prediction methods which appear to be useful. Reference 10
should be consulted for other methods.
j drag component
Kj, D
tr
dB ft/
-i I f -i 11 rP l f u ? l f \ 7 l / \
^ ^-ll^-H-il I /-^1/r^-l (4)
f
St " <Str)VTE(J)/deq(J) (5)
J/z
S(x) = SQn(x) exp - as lx-1 (8)
144 J. D. REVELL, G. J. HEALY AND J. S. GIBSON
s /?
SQ(x) = x D / K; x < 1
A =AC
Trg D V C DTT =AC
DgSw
d = A ,b = Ac S /b (12)
TTg TTg/ TTg Dg W TTg
80
00
o
a.
C/)
70
DQ
LU WING PROFILE DRAG
60 .FREQUENCY 222 Hz
PROFILE DRAG
O
O CONTROLS SPECTRUM
I I I I
Q
FLTTEST (HEALY, REF. 3)
SE 50
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
^THEORY
LU
40
50 100 200 500 1k 2k 5k 10k
FREQUENCY, Hz
v t /V (2TiARf(dvtxob)) (13)
(QASPL)vtx = 6 3 + 1 0 log
Qvtx = < I-
(14)
70
OQ
o
Q_
C/D
Q 60 S
1
LU WING
< 50 PROFILE DRAG-
FREQUENCY O
120 Hz 0
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Q i I
CL THEORY I
1
5 40 A FLIGHT TEST 11
LU
0 FLIGHT TEST 13
O
30
o
50 100 200 500 1k 2k 5k 10k
FREQUENCY, Hz
Fig. 3 Theory vs gliding flyover test data for "clean 11 wing
DC-3.
CD 100
"O <MEAUSRED (RUN A-1)
-FAN FUNDAMENTAL
Q.
CO
90
o Qn
80
p
6 PREDICTED
DC (GROUND PREDICTED
/
f 70 I REFLECTED) (FREE-FIELD)
\ / \
. I i i I i il__________I i , I i
60
20 40 70 100 200 400 700 1k 2k 4k 7k
FREQUENCY, Hz
Fig. 4 C-5A Galaxy flyover test data vs theory for "clean"
configuration.
Comparisons will next be made for large-aspect-ratio wings
with landing gear plus flap extension. Figures 5 and 6 present
results of freefield calculations which compare favorably with
C-5A test data7 for flap angles of 16 and 40, respectively,
with landing gear down. Results of recent calculations with
the current method differ negligibly from Ref. 9, from which
Figs. *5 and 6 were taken. The current method of evaluating
landing gear effects, compared to that of Ref. 9, has resulted
in a reassessment of the spectral composition of the low-
frequency part of the spectra shown in Figs. 5 and 6. Pre-
viously, this portion of the spectrum was attributed entirely
to the profile drag of the wing with deflected flaps, whereas
the present analysis indicates that the landing gear produces
equal noise. These predictions also indicate that, at high
frequency, the one-third-octave SPL values are dominated by
the induced drag noise (above 125 Hz at the 40 flaps angle,
CL = 1.5, and a>ove 200 Hz at the 16 flap angle, CL = 0.96).
100 I
CD LANDING GEAR - INDUCED DRAG
T3
CONTROLS SPECTRUM CONTROLS SPECTRUM-
35 90
GQ
LU
80
<
O 70
O
Q EXTRAPOLATED-
- PRESENT THEORY
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
DC
I O dvtxob = 0.006 AERO NOISE
h- 60 (REF.7)
LJLJ A dvtxob = 0.007 1 I I
FLIGHT TEST RUN A-3 (GIBSON, REF. 7)
50
20 40 70 100
200 400 700 1k 2k 4k 7k
FREQUENCY, Hz
Fig. 5 C-5A Galaxy flyover test data vs theory for 16 (40%)
flaps, landing gear extended.
\ r^
en luu
/^
~fa.
"0
-^^
^
^^^
_i
kT^Xj i v \ZH
Sssj A/ N
90
S LA NDING
^H
' 0
<
- irMDUCEC
GE AR
CO ^_^D RAG '^V^
^ r-r n
LU 80 \ji\j NTROLS
> SP ECTRUM
"* c ONTRO LS
-NVSo
< O PECTRLJM~
NLv
vv <J
6
70
PRESENT THEORY
FAN FUNIDAIV ENTAL
^[ V<3
>^^
EXTF APC iLATEC
f*~
QC
AERC)NO ISE-^" 0
V dvtxob = 0.008 \V
E 60 (REF 7) V
O dvtxob = 0.007 \
LU s
%
FLIGHT TEST RUN A- 5 ( G H3sorSI, REF. 7)
o
20 40 70 100
200 400 700 1k 2k 4k 7k 10k
FREQUENCY, Hz
Fig. 6 C-5A Galaxy flyover test data vs theory for 40 (100%)
flaps, landing gear extended.
03 110
PREDICTED MEASURED
(FREE-FIELD) (RUN A-2)
Q_
CO A
100 FAN FUNDAMENTAL
CO
90
o N
*. ^N
P 80 PREDICTED/ <V NN
G (GROUND REFLECTED) ESTIMATED
OC
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
I
LU
70
20 40 70 100 200 400 700 1k 2k 4k 7k
D
FREQUENCY, Hz
Fig. 7 C-5A Galaxy flyover test data vs theory for "clean"
wing, landing gear extended.
CO
CL
CO
CD
LU
<
CJ
O
Q
OC
z
o
500 1k 2k 5k 10k
FREQUENCY, Hz
Fig. 8 Effects of landing gear deployment on JetStar with
flaps retracted: theory vs flight data.
CD
T3
Q.
CO
<
00
0
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
0
Q
CC
- ENGINE NOISE
- ONESECOND AVERAGE (SEE TEXT)
G CO
Eo
IZ
60
Concluding Remarks
References
2
Smith, D. L., Paxson, R. P., Talmadge, R. D., and Hortz,
E. R. , "Measurements of the Radiated Noise from Sailplanes,"
TM-70-3-FDDA, July 1970, Air Force Flight Dynamic Laboratory.
o
Gibson, J. S. , "Recent Developments at the Ultimate Noise
Barrier," Paper 74-59, Aug,. 1974, International Council of
the Aeronautical Sciences.
14
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
20
Schloemer, H. H., "Effects of Pressure Gradients on
Turbulent Boundary Layer Wall Pressure Fluctuations," Journal
of the Acoustic Society of America, Vol. 42, No. 1, pp 93-113,
July 1967.
REVIEW OF THEORY AND METHODS FOR TURBINE
NOISE PREDICTION
D. C. Mathews,* R. T. Nagel,+ and J. D. Kester*
Pratt & Whitney Aircraft, East Hartford, Conn.
Abstract
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
I. Introduction
155
156 D. C. MATHEWS, R. T. NAGEL AND J. D. KESTER
COMPRESSOR CORE
NOISE ENGINE
NOISE
PEAK
NOISE
PNdB
early 1960's of the more efficient turbofan cycle engines, jet exhaust noise was
reduced markedly but at the expense of increased whine from the fans. Second-
generation high-bypass-ratio turbofans, such as the JT9D, CF-6, and RB.211,
provided even lower levels of jet exhaust noise and introduced new features to
provide significant reductions in fan noise. As a result of the fan and jet noise
improvements provided by these high-bypass turbofan engines, it was revealed
that noise from the turbine made a significant contribution to the engines' far-
field noise signature.
bypass engines such as the JT9D required that acoustical linings be installed in
the tailpipes of these engines to improve the noise signature. Based on the trends
illustrated by Fig. 2, it is obvious that turbine noise will continue to be a signifi-
cant factor in the noise signature of future aircraft engines. It is this trend that
motivates the need for accurate and proven methods to predict turbine noise.
Methods to predict noise from engine components fall into two basic cate-
gories: empirical and analytical. Analytical methods are possible only when the
aerodynamic and aeroacoustic characteristics of the source generation, propaga-
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
tion, and radiation processes are recognized and understood to the extent that
they can be described by reliable mathematical models. Although the develop-
ment of an analytical prediction method is the ultimate goal for any prediction
system development effort, this goal seldom is realized fully in practice. A
variety of elements are required to model analytically the propagation of fan
noise, including the modal propagation concepts introduced by Tyler and
Sofrin, the functions developed by Sears^ to calculate the response of airfoils
to gusts, and the models introduced by KajP' ^ and Mani^ to allow the calcula-
tion of noise propagation through upstream and downstream vane rows. Even
though relatively massive technical efforts have been undertaken to develop
analytical prediction methods for fan and jet noise, the methods in current use
are not purely analytical but are a blend of analytical and empirical techniques.
The importance of the turbine as a noise source was recognized relatively
recently, and, therefore, turbine noise prediction methods have not received as
much attention as fan and jet noise sources. Our survey of the literature on tur-
bine noise prediction methods has disclosed that few analytical models of tur-
bine noise problem elements are reported, and that most prediction methods are
purely empirical. Rather than focusing on details of the unsteady aerodynamic
processes that are at the heart of the physics of turbine noise generation, these
methods tend to focus on the thermodynamic parameters that are used to de-
scribe the mean properties of flow through the turbine. Typical steady-state
parameters that often are utilized include mass flow, temperature, pressure ratio,
turbine speed, and, in some cases, overall work extraction.
it is obviously unwarranted for any new designs except those that are merely
aerodynamically "scaled" models. In cases where noise predictions are required
for studies of "paper engines" or to develop parametric trend studies, only mean
flow information is available. Consequently, there will be a continuing require-
ment for prediction systems that embody this concept. However, it should be
recognized that the accuracy of these prediction systems is limited by this very
basic assumption.
3) Data base orientated. It is common that the most complete and accur-
ate turbine acoustic data available to a manufacturer for use in an empirical pre-
dection system are those measured from its own products. Because noise is a
factor in the competitive sale of engines, it is common that these data are not re-
leased publicly. Consequently, it is difficult to establish the extent to which a
prediction procedure developed by one manufacturer applies to predict accurate
levels for designs made by others. With this background, some of the current
procedures will be discussed in the following section to illustrate the similarities
and differences encountered in state-of-the-art turbine noise predictions.
Rolls-Royce Method
nents. Following procedures developed previously for fans and compressors, the
discrete tones were assumed to be generated by the interaction of rotor blade
wakes with downstream stator vanes and interactions of stator vane wakes with
downstream rotor blades. The broadband noise was assumed to be due to ran-
dom unsteady lift fluctuations on the airfoils induced by freestream turbulence.
The noise characteristics are related primarily to the blade inlet relative velocity.
Other parameters, such as density, turbulence intensity, blade incidence, solidity,
interstage spacing, and number of stages, were assumed merely to modify these
characteristics. Smith and Bushell acknowledge that an accurate correlation of
data which accounts for effects of all the preceding parameters is extremely dif-
ficult to develop, and therefore offer correlations based on only a few of these
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
parameters. The correlation for the final-stage fundamental tone SPL (dB) at
the peak far-field angle (100-ft sideline distance) is given by the following rela-
tion:
The turbine tone levels from full-scale engines did not increase with speed
as rapidly as those observed from tests on model turbine rigs, with f j (Vre],)
from model data following roughly a 30 log (Vrej) behavior and full-scale engine
data following approximately 10 log (Vrep. Similarly, the broadband correla-
tion from Ref. 6 is given by
oJT9D
QJT8D
AJT3D Oo
w
ou^
MAX SPL AT HIGHLY LOADED TURBINE
PEAK ANGLE
NORMALIZED BY SIZE,
NO. OF STAGES, 5dB A
SPACING,
DAD D
TEMPERATURE 8 ?
LIGHTLY LOADED TURBINES
the inlet axis for both turbine rigs and engines. The spectrum is assumed to be
the same at all far-field angles, but a distinction is made between the spectrum
from full-scale engines and that from a turbine rig.
The turbine noise procedure developed at Pratt & Whitney Aircraft (P&WA)
predicts peak angle one-third-octave band levels, polar directivity patterns, and
the one-third-octave band spectral characteristics of turbine noise. The proce-
dure was based on a correlation of data from JT9D, JT8D, and JT3D engines,
corrected for turbine mass flow and number of stages. The observed effects of
blade/vane spacing,7 together with turbine inlet temperature and turbine stage
work, also are included to obtain a good data collapse when plotted against the
turbine wheel speed as shown in Fig. 3. The dominant parameters in this proce-
TURBINE NOISE PREDICTION 161
dure are the turbine stage work and turbine speed. The need for a stage work
term in the correlation became apparent during development of the procedure.
Specifically, it was found that turbine noise from the highly loaded JT9D tur-
bine was approximately 10 dB higher than that from the more lightly loaded
JT3D and JT8D engine turbines, even after normalizing by the parameters sug-
gested in Ref. 6. This lack of "collapse" is shown in Fig. 4. The same data,
when normalized by an additional stage work term, correlated much better as
shown in Fig. 3. It is difficult to assess the influence of the number of stages on
turbine noise levels. One approach is to assume that turbine noise levels are
directly proportional to the number of stages in the low-pressure turbine. This
was assumed in the P&WA method and is included in the correlation of Fig. 3.
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
A turbine noise prediction method for interim use in the NASA Aircraft
Noise Prediction Program (ANOPP) was selected in early 1974 by Krejsa and
Valerino? The method basically is that derived by Smith and Bushell,6 but the
levels were adjusted to provide agreement with the turbine noise data from addi-
tional turbofan engines that were available. The empirical technique predicts
far-field turbine noise levels, directivity, and one-third-octave band spectra. The
procedure considers the broadband and discrete tone components of turbine
noise separately, and assumes that both components are related to the inlet rela-
tive tip velocity of the turbine last stage, the primary mass flow, and the local
speed of sound at the turbine exit. The effects of stator/rotor spacing on the
discrete tone level are also considered.
The correlation equation of Ref. 8 for turbine tone levels was noted by
NASA to overpredict results from the P&WA JT8D engine by approximately 10
dB. It was assumed in Ref. 8 that this difference was attributable to the com-
mon flow exhaust geometry of this engine. It therefore was recommended in
Ref. 8 that, for turbofan engines with the primary nozzle exit plane upstream of
the fan nozzle exit, a correction factor of -10 dB be applied to the tone level pre-
diction. It also was assumed in Ref. 8 that both tone and broadband compon-
ents have spectra shapes that are centered at the fundamental blade passage fre-
quency calculated for the last stage. The spectra shape assumed for turbine
broadband noise was based on fan broadband noise characteristics.
General Electric Methods
The General Electric Company (G. E.), under contract from the FAA,9"11
recently developed both an analytical prediction technique for discrete frequency
162 D. C. MATHEWS, R. T. NAGEL AND J. D. KESTER
tion model and the predictions were compared with in-duct measurements from
a G.E. single-stage turbine rig, good agreement was obtained, particularly with
regard to rotor/stator spacing effects. The predicted trends, however, are inade-
quate for describing the far-field noise from multistage turbines, since proce-
dures were not provided to account for the effects of attenuation through down-
stream blade rows, propagation from nozzles, or radiation effects through ex-
haust shear layers. Therefore, it was necessary to resort to empirical methods
for predicting turbine noise from multistage full-scale engines.
G. E. Empirical Prediction. Relationships are presented in Ref. 11 which
relate turbine noise levels to various steady-state turbine operating parameters.
One such expression was produced by correlation of data from several G. E. en-
gines and is given by
This expression predicts the maximum one-third-octave band SPL of tones plus
broadband at a sideline distance of 200 ft and at an angle of 120 from the inlet
axis, the assumed angle of maximum noise radiation. The correlation parameter
of turbine ideal work extraction normalized by the inlet enthalpy AT/T also can
be expressed in terms of the turbine pressure ratio by
7-1/7
AT/T=l-(l/Prr n (4)
where
Pr = turbine pressure ratio
V^p = blade tip speed of the dominant stage, fps
A = core nozzle exit area, ft
7 = ratio of specific heats
band noise. Good agreement with G.E. engine data was reported when using
this expression.
Tone x x X X ;H
Rolls-Royce6 Broadband x x X
D>
0
Tone and m
Pratt&Whitney 12 broadband x x x x x x x rt*
Tone x x X X X
C_,
8
NASA Broadband x x x
good agreement with the experimental findings in Ref. 7. The -20 log (spacing)
dependence used by Rolls-Royce" in Eq. (1) was based on earlier fan noise ex-
perience. Since the time of development of the earliest turbine noise prediction
method by Smith and Bushell, it has been suggested by several investigators
(see Table 1) that a turbine work parameter would be required to "collapse"
turbine noise data adequately from engines with significant differences in load-
ing characteristics. Some procedures, however, do not incorporate a work term.
A fairly recent development in turbine noise technology was the recog-
nition of the effects of fan duct length on observed far-field turbine noise
characteristics. After the tones have propagated through the primary engine
tailpipe, they must pass through the turbulence in the fan shear layer. In this
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
process, some of the acoustic energy in the tones is scattered to other angles and
frequencies. These phenomena were discussed in more detail by Mathews and
Peracchio12 and are discussed in Sec. V of this paper. NASA and P&WA in-
clude this effect by using an empirical fan duct length correction. The latter
method makes use of data obtained on a turbofan engine tested with three dif-
ferent length fan ducts. Kazin et al. also have observed this phenomenon and
present an empirical method for predicting the decrease in tone energy due to
scattering. This method, however, is not included explicitly in the G. E. turbine
noise prediction procedure.
APPROACH POWER
120 FROM INLET AXIS
TURBINE NOISE
3 4 5 6 7 8 9 10
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
FREQUENCY KHz
Fig. 5 Typical JT8D-109 engine noise spectra.
MEASURED DATA PREDICTED LEVELS
-JT8D-109 ENGINE D GENERAL ELECTRIC
A NASA
PRATT & WHITNEY
ROLLS ROYCE
sured values over a range of engine speeds. The levels shown in Fig. 6 represent
the sum of the tone and the broadband SPL at the peak radiation angle. It is
seen that both measured and predicted turbine noise levels increase gradually
with rpm and generally tend to level off at high engine speeds. The previously
described prediction systems envelope the measured data with a band approxi-
mately 7 dB wide.
It is difficult to evaluate, from the single comparison in Fig. 6, all of the
shortcomings or strong points of the various procedures examined, nor is this
the intent of the current paper; Fig. 6 does reveal, however, the general range of
uncertainty that potential users of any procedures should expect. Since the pro-
cedures are empirical in nature, each is subject to the inherent shortcomings dis-
cussed in Sec. II, and, therefore, the 7-dB variation shown in Fig. 6 is not sur-
TURBINE NOISE PREDICTION 167
MEASURED DATA
- JT8D-109 ENGINE
PREDICTED DIRECTIVITY
0 GENERAL ELECTRIC
6 NASA
PRATT & WHITNEY
__. ROLLS ROYCE
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
prising. If data from an engine with a highly loaded turbine had been used in the
comparisons, the scatter in the predictions probably would have exceeded the
7-dB band shown, since two procedures (see Table 1) do not include a turbine
work term.
MEASURED DATA
- JT8D-109 ENGINE
PREDICTED SPECTRA
-D-GENERAL ELECTRIC
-A-NASA
PRATT & WHITNEY
APPROACH POWER,
120 FROM INLET AXIS
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
veloped for the complex mechanisms of noise generation by turbine rotor and
stator rows, propagation through downstream stages and exhaust nozzles, and
radiation through the turbulent exhaust flow to the far field. Development of
these models will involve both the formulation of analytical expressions to de-
scribe turbine noise generation, propagation, and radiation phenomena and the
research tests necessary to verify the accuracy of the models. In the remainder
of this section, the elements required for each of these models are described,
together with a discussion of the status of available models.
Narrow-band spectral analysis (see Fig. 8) reveals that tones are the domi-
nant feature of turbine noise. It therefore is clear that a model to predict the
generation of tones at the fundamental and harmonics of turbine blade passing
frequency must be given highest priority. Because of the similarities between the
turbine and fan/compressor design features, it is possible to postulate that blade/
vane interaction noise-generation processes are similar in each. Therefore, some
of the findings from fan/compressor noise research may be applied to model
turbine tone generation.
In fans and compressors, it has been found that a major source of noise
generation, called "interaction" noise, is attributable to the surface pressure
fluctuations on rotor airfoils as they "chop" through the wakes from upstream
vanes, and on downstream vanes as wakes from upstream blades pass by. These
sources both produce noise at blade passing frequency and its harmonics. Such
processes clearly are present in turbines, but with some notable differences due
to basic differences between the aerodynamics of turbines and compressors.
It is not uncommon for the pressure drop across one turbine stage to be as
large as the pressure rise across six to eight compressor stages. This is possible
because the boundary layers on turbine airfoils are subjected to favorable pres-
sure gradients and are therefore much less likely to separate than in the case of a
compressor. Because of these favorable gradients, a much higher degree of aero-
dynamic tolerance is associated with turbine stages. Therefore, the blade and
vane configurations are much different from those of compressors. Turbine
170 D. C. MATHEWS, R. T. NAGEL AND J. D. KESTER
blades typically are highly cambered to achieve the required stage work. This
leads to much steeper axial pressure gradients along the airfoils and to wake
characteristics that are different from those for compressor stages. It also is pro-
bable that the surface pressure response to perturbations in the incoming flow is
much different in turbines than for compressors. Research tests are required to
quantify these aerodynamic differences between compressors and turbines.
Many of the concepts that have been developed to predict noise propaga-
tion in compressors also apply to turbines, but some new factors are introduced
because of the higher levels of aerodynamic loading associated with turbines.
The most significant factor in noise propagation in both turbines and compres-
sors is the concept of "acoustic spinning modes." Experience with fan noise has
shown that the pressure field associated with an isolated subsonic rotor does not
propagate. However, wheen stationary vanes are introduced either upstream or
downstream, noise at blade passing frequency may propagate. Propagation oc-
TURBINE NOISE PREDICTION 171
curs when acoustic interaction patterns are generated which "spin" at super-
sonic speeds even when the rotor speed may be well below sonic. * Depending
on the number of blades and vanes used in a turbine row, interaction patterns
are generated which may or may not propagate. Patterns that do not propagate
are said to be "cutoff." Cutoff patterns attenuate rapidly near the rotor and do
not carry significant amounts of acoustic energy to the far field. Cutoff typically
is achieved by use of a large number of stationary vanes upstream and down-
stream of a rotor stage. Fans typically require about twice as many vanes as
blades to achieve cutoff. In turbines that operate at elevated temperatures and
lower tip Mach numbers, cutoff may be achieved with fewer vanes. The concept
of cutoff has been used widely to reduce noise levels from fans,but it has not
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
SCATTERING BY
EXHAUST TURBULENCE
REFRACTION THROUGH
VELOCITY AND TEMPERATURE
GRADIENTS
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
SPL-dB
3 4 5
FREQUENCY-KHz
Fig. 11 Effects of fan duct configuration on
turbine noise.
10. A turbine noise radiation model must address the effects of jet flow on the
nozzle termination acoustic impedance and be developed to apply to a wide
range of exhaust geometries. For example, common flow nozzles exist on many
bypass engine installations in which the primary exhaust nozzle exit plane is lo-
cated well upstream of the final jet nozzle that passes both fan and primary flow.
The calculation of nozzle termination impedance for this type of geometry
would be extremely difficult and has not been attained to date.
pass through two turbulent mixing regions: one at the shear interface between
the primary and fan streams which involves mixing of different temperature
streams, and one at the fan-ambient shear interface. The turbulent eddies con-
vected in these regions provide "discontinuities" in the flow from which turbine
tone scattering has been observed.10-12 This scattering is responsible for chang-
es in both spectral and directivity characteristics of the radiated noise. Whereas
the bulk of the acoustic energy at the primary nozzle exit is contained in dis-
crete tones, significant spectral "haystacks" of broadband noise are observed in
the far field in a frequency range about each tone, as shown in Fig. 8. The scat-
tering mechanism is discussed in the following paragraphs.
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
As a tone passes through the turbulent region of the exhaust shear layers,
variations in the speed of propagation of the sound waves are induced. These
changes may be produced by temperature variations that cause the sound speed
to fluctuate, or by random refraction and convection of the sound by the tur-
bulent eddy. Both effects may cause random phase variations along a wavefront
and give rise to both spatial and frequency scattering of the tone. -22
Advanced turbine noise prediction procedures will require models for the
radiation phenomena described previously. Further experimental research is
needed, however, to verify the relationships that exist between eddy size, con-
vection velocity, and tone frequency, so that accurate models can be obtained.
ferences in the correlation parameters used. Also, each procedure employs dif-
ferent empirical data to reflect the measured data base that uniquely is available
to the particular agency or company. Among the more significant differences
noted among the procedures was the inclusion or lack of a term to account for
effects on noise of turbine work extraction. Evidence was presented to suggest
that such a term is required.
used because extensive turbine noise data were available, and because data from
this engine were not used in formation of any of the empirical predictions. Re-
sults of the comparisons of measured vs predicted noise levels, directivity, and
spectra show that none of the procedures achieved "perfect" agreement. In
general, these comparisons suggest that there is considerable room for improve-
ment in most of the procedures before they can be considered reliable. The sug-
gested direction for improvement of these procedures is not to acquire more tur-
bine noise data but rather to conduct the experimental and analytical research
necessary to develop correlation parameters that are related more closely to the
physical processes of noise generation, propagation, and radiation.
From the results of our survey, it appears that much more research is re-
quired before reliable analytical models of turbine noise are available. Work
conducted toward the formulation of analytical models should provide better
insight into the physical processes involved. This insight should lead to the selec-
tion of more fundamental correlation parameters for improved empirical pre-
diction systems. Because of the importance of turbine noise in the design of
future engines, the development of improved turbine acoustic technology must
be given a high priority in future acoustic research and development programs.
References
^Kazin, S. B., et. al., "Core Engine Noise Control Program, Vol. I - Identifi-
cation of Component Noise Sources," Rept. DOT-FA72WA-3023, May 1974,
Federal Aviation Administration.
10
Kazin, S. B., et. al., "Core Engine Noise Control Program, Vol. II - Iden-
tification of Noise Generation and Suppression Mechanisms^' Rept. DOT-FA72
WA-3023, May 1974, Federal Aviation Administration.
1
1 Kazin, S. B., et. al., "Core Engine Noise Control Program, Vol. Ill - Pre-
diction Methods," Rept. DOT-FA72WA-3023, June 1974, Federal Aviation
Administration.
13
Silverstein, A., Katzoff, S., and Bullivant, W. K., "Downwash and Wake
Behind Plain and Flapped Airfoils," Rept. 651, 1939, National Advisory Com-
mittee for Aeronautics.
176 D. C. MATHEWS, R. T. NAGEL AND J. D. KESTER
^ Mathews, D. C. and Nagel, R. T., "Inlet Geometry and Axial Mach Number
Effects on Fan Noise Propagation," AIAA Paper 73-1022, Oct. 1973.
18
Morse, P. M. and Ingard, K. U., Theoretical Acoustics, McGraw-Hill, New
York, 1968, p. 471.
Levine, H. and Schwinger, J., "On the Radiation of Sound from an Un-
flanged Circular Pipe," Physics Review, Vol. 73, 1948.
* +
R.E. Motsinger and J.J. Emmerling
General Electric Company, Cincinnati, Ohio
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Abstract
Nomenclature
177
178 R. E. MOTSINGER AND J. J. EMMERLING
2
OASPL = overall sound pressure level, re 20 yN/m
P = total pressure, psf (N/m2)
SPL = sound pressure level for one-third-octave band center
frequency, re 20 yN/m2
T = total temperature, R (K)
V = velocity, fps (m/sec)
6 = ratio of total pressure atmosphere or reference
pressure, P/PQ
Subscripts
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
a = air
b = flame speed
c = combustor
d = directivity factor
e = equivalent
f = frequency
L - laminar
N = nozzle
R = reference
T = turbine
0 = atmospheric or freestream
3 = combustor inlet station
4 = combustor exit station
5 = turbine exit station
I. Introduction
10
9
e o
CQ -10
n3
-20
o T63 (250 HP)
9 n T58 (1500 HP)
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
-40
o 100 500 1000 5000 10000
FREQUENCY, Hz
Fig. 1 Combustor noise spectra for turboshaft engines.
70%
humps, one peaking at 125-160 Hz and the other at 630 Hz. The
lower-frequency hump stems from distortion of jet noise by the
first ground-reflection reinforcement. The higher-frequency
hump is attributed to combustor noise with a varying level of
fan and jet noise contribution, depending on the power setting.
It is not possible to compare the full spectral distribution
from this turbofan with that from the turboshaft engines be-
cause of the contribution of other noise sources from this
engine. Note, however, that the peak frequency (630-Hz
one-third-octave band) occurs in the same octave band (500-Hz)
as the distribution shown in Fig. 1.
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Literature Review
more attention has been devoted to the first item, both ana-
lytically and in combustion component testing, and this mecha-
nism, therefore, forms the basis of the correlations described
in this paper. On the other hand, the results from consider-
ation of the second item quantitatively predict the location
of the frequency of peak noise generation (which already has
been noted experimentally in Fig. 1) and are reported qualita-
tively to check reasonably well with "measured (inferred)"
core engine sound power; consequently, this approach is one
requiring further research.
Acoustic Sound
power pressure
Reference level level Spectrum Directivity
5
Bushell .. D D
Marshall . D D D
Ho and Tedrick D D
0
Motsinger D ... D D
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
j 10
Grande T ... ...
Arnold T
12
Gerend ... D D D
14
Dunn ... D ...
1S
Strahle T ... D D
16
Swan ... ... D .
17
Abdelhamid T D D
1Q
18
Plett . . T D D
19
Chiu T . ...
20
Chiu T,D D ...
7 81 OAPWL
~4-1;)~J [~J U+F T~ [V4 ] [De]
7 23 OAPWL
~14-l;t{~ [~2] U+ FJ4 [V:J [ De1 n
0
3:
c:o
12 b OASPL
d
. .. -0
;;0
fT1
0
.......
e
~ARJ2 ~J3 [m~~;/~ .. . . .. n
14 c OASPL -1
.......
0
:z
15 OAPWL .. . ... ~LrH
1+F
~ 2.68 1.3~
\;j 'til [D~84J
a205.5 + 10 log [BNB T]
b[-15 for annular, -6 for can or can-annular] 5 dB
c76 + correction for combustor type, engine directivity
d200-ft sideline, 110 0 from inlet
eat I-meter radius 00
w
184 R. E. MOTSINGER AND J. J. EMMERLING
o N
B E turbine transmission coefficient
The fifth and sixth equations, from Refs, 12 and 14, are
very similar in terms of the functions and parameters used, and
they are based upon correlation of engine data. The details
included in the original references are more extensive in Ref.
14; for example, the effect of source motion on far-field
directivity pattern (source convection factor in flight) is
included. Again, both references differentiate among engine
types .
where
a = 0, fuel lean primary zone
a = 1, fuel rich primary zone
COMBUSTION NOISE PREDICTION 185
PQ 150
I
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
J
140
co
130
O
o
120
100
10
_ ]EXPECTED VALUE
h+^ n % *
v x<
1
>
\
S *'''" V
v N
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
O \
-5 JLpr
<>
o J85 - 4c0 Hz
n TF34 - 4oo Hz
-10 \ A GE4 - 400 Hz
<> F1O1 COMPONENT - 400 Hz
0 CF6 COMPONENT - 400 Hz
X ,T64 - 400 Hz
-15
60 80 100 120 140 160 180
ANGLE FROM INLET, (DEGREES)
Fig. 5 Combustor noise directivity.
and the levels decrease more rapidly both fore and aft than the
correlation proposed in Ref. 3. The expected-value line shown
on the figure is recommended. Margin of error is indicated by
the data scatter.
where
K = 4 8 for turbofans
= 56 for turboshafts
= 64 for turbojets
m = core engine airflow rate, Ibm/sec
a
Spectrum
Directivity
References
?
Kazin, S.B., and Paas, J.E. , "NASA/GE Quiet Engine A?
Acoustic Test Results, 11 CR-12175, Oct. 1973, NASA.
2
Pickett, G.F., "Turbine Noise Due to Turbulence and Tem-
perature Fluctuations," Eighth International Conference on
Acoustics, July 23-31, 1974, London.
10
Grande, E., "Core Engine Noise," AIAA Paper 73-1026,
Oct. 1973.
Swan, W.C. and Simcox, C.D., "A Status Report of Jet Noise
Suppression as Seen by an Aircraft Manufacturer," First Inter-
national Symposium on Air Breathing Engines, June 18-22, 1972,
Marseille.
194 R. E. MOTSIN6ER AND J. J. EMMERLING
195
196 DISCUSSION: AIRCRAFT NOISE PREDICTION
account for the fact that this noise must be carried through
the turbine and exhaust nozzle before it is radiated to the
far field. In order to understand this phenomenon better, a
good understanding of the flow through the turbine and exhaust
nozzle and the effect of this flow on the combustion noise
transmission are required. Again, basic thermodynamics and
fluid mechanics are an inseparable part of the acoustic pre-
diction problem.
Turbine noise. Like combustion noise, turbine noise
presently is predicted by empirical formulas which account for
only the gross variables of the problem. The few analytical
models which have been attempted use concepts similar to fan
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
D. L. Mart lew*
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Abstract
Introduction
203
204 D. L. MARTLEW
DOWNSTREAM
REVERBERATION
CHAMBER
General Characteristics
of Hartmann Generators
130
GAP CUP DEPTH
111
15
10
CUP DIAMETERS f Y If
PRESSURE RATIOS 2-2-3.4
120
100
A 6 10
FREQUENCY - KHz
Fig. 9 Spectra of "in-duct"
noise generator
212 D. L. MARTLEW
TL = Li - Ls + 10 logic K
power between the test duct on one side of the source and the
reverberation chamber on the other. Thus, an additional
assumption has to be made when using the change in the
difference between reverberation chamber levels as a measure
of attenuation, namely that the energy split does not change
between the hard and soft walled situations. Obviously this is
a difficult assumption to prove,but confidence in its
correctness has built up with experience and no discrepancy
has been encountered which has cast doubt upon it. More
recently, after cross checks using sources in the reverberation
chambers, the very close agreement in measurements made with
the quite different sound sources has provided strong
circumstantial evidence in favor of the measurement technique
used with the in-duct source.
TRAVERSING MICROPHONE
10dB
._!_
FIXED MICROPHONE
FIXED MICROPHONE
100
DOWNSTREAM REV. CHAMBER
4 6 10
FREQUENCY - KHz
Fig. 11 Effect of flow on
spectrum shape
THE USE OF HARTMANN GENERATORS 215
RELATIVE FREQUENCY
0-6 0-8 1-0 12 1-4 1-6
or
dB
^ACTUAL ATTENUATION
-10
dB
-20
Development of a Reverberation
Chamber Noise Source
OASPL-dB
160
RATIO
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
, SPL-dB
PRESSURE RATIO 2-8:1
BANDWIDTH 10 Hz
FREQUENCY - KHz
120
100
4 6 8 10
FREQUENCY - KHz
ATTENUATION - d B
30 50 dB OA5PL
MACH No. +0-3
89/. PERFORATE FACING
20 - 30 r
1-5" CELL DEPTH
10- 20 - 30
0 L 10 - 20 - 30
0L 10 - 20
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
0L 10
1 2 3 4 5 6 7 8 10
FREQUENCY - KHz
@ IN- DUCT NOISE GENERATOR SPECTRUM AS IN FIG. 9
(B) ONE 2" GENERATOR IN REV. CHAMBER. SPECTRUM SIMILAR TO
Conclusions
References
R. A. Kantola*
General Electric Company, Schenectady, NY
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Abstract
Outdoor acoustic facilities are used widely for jet noise
experiments because of their relatively low cost when compared
to anechoic chambers. However, many problems do exist with
outdoor facilities. It is the intent of this paper to de-
scribe a new facility at the General Electric Corporate Re-
search and Development (CRD) Center which minimizes these draw-
backs. The unique combination of capabilities that this out-
door facility offers includes hemispherical microphone cover-
age, permanently installed microphones, acoustically treated
ground plane, and real-time data processing. This facility is
intended primarily for high-temperature jet noise research and,
as such, has a silenced burner capable of operation to 2000R.
These features result in a facility with minimal ground reflec-
tions and a short turnaround time so that testing in short pe-
riods of favorable weather is possible. Most important, these
features do not compromise the capability of the facility to
obtain high quality jet noise measurements on both cold and
heated jets.
223
224 R. A. KANTOLA
Introduction
Nomenclature
Facility Description
PERFORATED
HASTELLOY X
LINER
INLET AIR
Fig. 2 Burner-muffler construction.
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
ANGLE TO
THE JET
AXIS g
90 f=2500Hz
6 =90
frO
=90
SPL
(dB) MICROPHONE
80
* -MICROPHONE
SPEAKE
6' ^ACOUSTICAL PAD
(8" THICK)
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
70
SPLa20L06|0(l/R)
_L _L _L
4 6 8 10 12
MICROPHONE RADIUS (FEET), R
Fig. 4 Spherical divergence tests (loudspeaker acoustic source),
J
4
o o
3 n
0
A n D
2 >-
1 _ -, 5hJo ^j^Ljcr 0
D
nnn0
0- o
AdB D AAA2ft AA D
-1 ft A
DA n
-2 - n o
O
-3 A
o HARD PAD 1
D 4" FOAM LINING >
-4 _ TESTS
---POINT SOURCE PREDICTION * * . ^Vl'g'^
-5
WITH H A R D PAD '
D
-6
-7
-8 1
n
IO Z I0 5 10" 10
1/3 O C T A V E FREQUENCY ( H z )
Fig. 5 Ground reflection corrections for different ground
treatments.
230 R. A. KANTOLA
<<, 0 = 30*;
JET NOISE
Vj = l009fp$
530R
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
20
(dB)
140
120 _L_
-0.2 -O.I 0 O.I 0.2
LOG,0(Vj/A0)
130
O
o
120
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
NO -^00-0 *
V j ~ 600 FT/SEC
100
90
-1.2 -0.8 -0.4 0.4 0.8
Cff^b Vj=l9l9fps
170 QCP ^ V,S94.R
o o
O O fVs
/% ^O
7" I6
o 0
*~o
O
^ n
u ^J*= I5l7fpt Q
w
o
2* 150 0
0
0
/o /%
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
^ 140
Q_ 0 Tj=l903 p R q^
LU O ^*
0 0^ D = 2"
8 0 Q
^>)0^ T a =486R
ro
LINED PAD
0 ^***b
Vj = 744fps %^ = 0 P
120 0 ^ L= 1990 R
no 1 0 1 1 1 1
-2
-10
-20
j = I 5 l 7 f p s , Tj = l85lR
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
-30 D V j = I489fp$, T j = l 5 0 9 R
o D = 2"
o Ta=486R
-40 LINED PAD
-50 -
-60 1 I 1 1 J
- 2 - 1 0 1 2
LOG,0(fD/Vj)
Fig. 10 Effect of jet heating at VjMBOO fps.
I U 1 .X
90
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
o o
D %\% 0
80 8
^Vj6H4o0fo
| Vj=605fps
0 Q
Vj=59lfps
70 i i i i i I I 1 1 1 1 1 1
30 60 90 120
ANGLE TO JET AXIS, 9 (DEGREES)
JET
H APERTURE
CHAMBER A CHAMBER B
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
TOTAL
o 2 JE APERTURE
2" JE C H A M B E R
l"j CHAMBER
Vj - 990 999 fps
Tj = 512'
BARE P,
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
120 i-
INTERACTION
NOISE
110
PWL
(dB)
O X / D =13, d 0 / D = 8
D X/D =14, d 0 / D = 8
100 D=l"
Vj =992 f p s
Tj=5llR
B A R E PAD
ABSORBENT CHAMBER
90 I I
10
1/3 OCT. B A N D - F R E Q U E N C Y (Hz)
the jet flow, and the minimum aperture sizes were determined.
Results from both the aperture-alone tests and with the absor-
bent chamber in place are shown in Fig. 14. The range of sizes
was very close to the relation (dQ-D)/x = 0.5, where D is jet
diameter, dg is aperture diameter, and x is axial distance from
jet exit. This corresponds to a truncated cone starting at the
jet lip and flaring out with a half-angle of 14.
From these measurements, it is found that it is virtually
impossible not to generate interaction noise, except when the
orifice is located very close to the nozzle exit. Where does
this noise appear in the measured far-field spectrum? When
far downstream, for example when x/D > 6, the interaction spec-
trum peaks are separated well enough to be identified. An exam-
ple of this is shown in Fig. 15, where data obtained for two
axial locations of the same orifice plate are shown. The in-
teraction noise is seen to dominate in the low frequencies.
The decrease in PWL at the high frequencies also can be seen
clearly as the axial distance of the orifice plate from the
nozzle exit is increased. When the orifice plate is closer
to the nozzle, the interaction peak frequency becomes closer
to the jet noise peak frequency, as would be expected from a
Strouhal number consideration. In these cases, the separation
of the jet noise and the interaction noise becomes more diffi-
242 R. A. KANTOLA
140 i
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
0=
Vj = 9 9 0 - 9 9 9 f p s
Tj = 5llR
120
B A R E PAD
ABSORBENT CHAMBER
IIOl
0 10 15 20 25
X/D
Fig. 17 Decrease of measured acoustic power with aperture
location.
-10 r-
-12
-14
-16
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
-18
Vj = 990-999 fps
Tj=5llR
-20 BARE PAD
ABSORBENT CHAMBER
-22
-24 I J
0 5 10 15 20 25
X/D
Fig. 18 Axial source distribution (% OAPWL/X/D).
Or
S -10
0.419
0=1"
= 990-9991 ps 0.838
-20
j = 5llR
BARE PAD
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
ABSORBENT CHAMBER
-30
10 15 20
X/D
References
^Kantola, R. A., "Outdoor Jet Noise Facility, A Unique Ap-
proach," AIAA Paper 75-530, 1975.
2
Hoch, R. G., Duphochel, J. P., Cocking, B. J., and Bryce, W.
D., "Studies of the Influence of Density on Jet Noise," Jour-
nal of Sound and Vibration, Vol. 28, No. 4, June 1973, pp.
649-668.
3
Ahuja, K. K., "Correlation and Prediction of Jet Noise," Jour-
nal of Sound and Vibration, Vol. 29, No. 2, July 1973, pp. 155-
iw.
4
01sen, W. A., Gutierrex, 0. A., and Dorsch, R. G., "The Ef-
fect of Nozzle Inlet Shape, Lip Thickness, and Exit Shape and
Size on Subsonic Jet Noise," TMS-68-1182, 1973, NASA.
OUTDOOR JET NOISE FACILITY 245
Abstract
Introduction
247
248 Y. KADMAN AND R. E. HAYDEN
WINDOW
MICROPHONE\ I!
-^E
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Tunnel Circuit
General
The Nozzle
the external surface of the nozzle was covered with the same
foam used in the acoustical treatment of the test chamber.
In the smaller BBN tunnel, it had been found that the con-
cept of a stagnation plate collector was successful for achiev-
ing low noise and minimum recirculation at the expense of
reduced tunnel aerodynamic efficiency. In this design, the jet
impinges on a rigid flat plate and is sucked out around the
stagnation plate perimeter, which avoids the noisy edges of a
bellmouth-type collector. The comparison with other concepts
follows .
252 Y. KADMAN AND R. E. HAYDEN
The side and top passages of the cowl continue down to the
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
The four lined exhaust ducts discharge into the fan plenum
chamber. The double-inlet, double-width centrifugal fan is
centered in a cavity-backed, foam-lined room. The lining pre-
vents the buildup of reverberant energy near the fan. The fan
has no inlet boxes and rests on a concrete slab, which is iso-
lated from the rest of the structure. The fan is surrounded
with screens, which evenly distribute the flow being sucked
from the exhaust ducts.
Isolated from the fan itself, the fan exhaust box is lined
with foam or fiberglass. This lining is covered with screens
or perforated metal, at the locations where the flow velocities
are high. As in the rest of the tunnel, cavities back most of
the lining, which greatly enhances its sound absorption capa-
bility.
\?
)0 fps
r*
/x2J
^
/
-150
"^-,
V
x^S.
'^^
-70
~^>x J
X
X <>^ X^
\
-80
I 100-/
\- \
\
8 \
\
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
-90
\
\
Aerodynamic Performance
TUNNEL WALL
ANECHOIC
TREATMENT
DETAIL
(NOT TO SCALE)
1 2 4 8 1011.5
DISTANCE FROM TUNNEL CENTERLINE(ft)
Acoustic Performance
'V
^ 10ft FROM NOZZLE FACE
l\
/-~^ \ 40 fps
\K ^
V^ r-57 N
III
r90
XN
X
\
S
\
^^.XX
' 1
/
X \ s^^^
^
1 -^
oy ^^: ^ -^
S^ X
^^^
NO FL
..^ "Xx
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
-\
6 3 5 63 1 5 2 50 500 1000 2000 40OO 8OOO 16,000 31
so
3BN HIGH SPEED FREE JET
tfIND TUNNEL -28" x 40" NOZZL
^X
:t
<
\
O
60
\ ^^ s
0
v -< 175
~^
\ r^ 5 * ^ - ~^^^
^
^
S 50
N
^^ ^
40 s r 82
\
-^^
30
\
31.5 63 125 250 500 1000 2000 4000 8OOO 16,000 31,500
ONE-THIRD OCTAVE BAND CENTER FREQUENCY ( H z )
2 -60
Fig. 5 Tunnel noise vs col-
lector placement. A IN ft 2
R IN ft
U IN fps
SPLINdB(re0.002/ibor)
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
DIMENSIONLESS FREQUENCY,^
200 Hz. It was established that the fan and diesel noise are
not a factor in the flow noise. The background noise compari-
son with other tunnels is shown in Fig. 5-
High lift devices may deflect the tunnel jet and increase
the background noise levels.1 By extending the nozzle with a
tilted section, we measured the effects of the main-jet flow
deflection by angles up to 10. We did not detect any changes
in either the acoustic or aerodynamic performance of the tunnel,
a fact that bore out the validity of the concept of large sepa-
ration between the nozzle and the collector.
Summary
References
8
Maestrello, L., "UTIAS Air Duct Facility for Investigation of
Vibration Noise Induced by Turbulent Flow Past a Panel," TN 20,
1958, University of Toronto, Institute of Aerospace Studies.
258 Y. KADMAN AND R. E. HAYDEN
9
Hanson, C . E . , M The Design and Construction of a Low-Noise,
Low-Turbulence Wind Tunnel," TR79611-1, 1969, Acoustics and
Vibration Laboratory, Massachusetts Institute of Technology,
Cambridge, Mass.
10
Ver, I.L., "Acoustical Modeling of the Test Section of the
NASA Langley Research Center Full-Scale Wind Tunnel," Rept.
2280, 1971, Bolt Beranek and Newman Inc.
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
CORRECTION OF OPEN-JET WIND-TUNNEL MEASUREMENTS
FOR SHEAR LAYER REFRACTION
Roy K. Amiet*
Abstract
Nomenclature
259
260 R. K. AMIET
Subscripts
Introduction
Theoretical Development
WAVEFRONT
Fig. 1 Acoustic source beneath plane zero-thickness
shear layer.
262 R. K. AMIET
by the line SCO, which, below the shear layer, makes an angle
Q T with the shear layer, and an angle 9 above the shear layer.
The change from Q T to 9 as the sound passes through the shear
layer is a result of refraction by the shear layer. The fluid
densities above and below the shear layer are assumed to be the
same. (This assumption can readily be eliminated.) There is
little change in density across the shear layer of the UTRC
Acoustic Research Tunnel. The Mach number M is assumed uniform
below the shear layer and zero above it.
If the shear layer had not been present so that the uni-
form Mach number M continued out to infinity, the sound on
reaching the former position of the shear layer would continue
to propagate rectilinearly, following the dashed line in Fig. 1
rather than the solid line. Thus, the sound heard at position
0 in the presence of the shear layer would be heard at position
A or B in the absence of the shear layer. If one wishes to
correct the data to an equal sideline distance, point A is
used, whereas if one wanted to correct the data to an equal
radial distance from the source, point B is used.
here unless one were able to separate out the near- and far-
field parts of the measurement.
r 2 2 ~l
~l 11/2
where - [d ~M COS 9} -COS Q\
P
A f h f . Q6+(
, , V| ,N/-l l / 2 r 3fl , , i n/. 31 1/ 2
-J [sm0 + ( T -l)C J
IE 2
8?
o 0
20 40 60
100 120 140 160 180
o -2 MEASURED A N G L E S
CM
-4
-6
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
10 -
8-
E
t 2
160
140
uj 120
_i
O
< 100
o
HI
o 80
r
r .
(6)
ro
CT)
00
-f
2
CO
JN ii
\ >
X
I
00
ro
ii
CO
Q o> 2
o
Q
SHEAR LAYER CORRECTION 269
Experimental Results
M = 0.27 |
- M = 0
180
effort has been made along these lines, but significant prob-
lems are involved. When a sound source is placed in a stream,
one cannot expect that the radiation pattern of the source will
remain unchanged even if one has no shear layer. Thus, to have
an idea of the amount of shear layer refraction, the source
directivity must be measured inside the shear layer and com-
pared to that outside the shear layer.
Conclusions
80 O EXPERIMENT
THEORY
70
Q>
LU 60
< 50
Q
UJ
040
30
20
10
0 10 20 30 40 50 60 70 80 90
MEASURED ANGLE, # m
= t + (Mx-a-)/Co/32 (Al)
or
cos$'
costf = -M] (A5)
dx0/d = (A7)
spreading in xy
SHEAR LAYER CORRECTION 275
Pt = Ae i M -t> (AID)
SOUND RAY, n t
y@ _
(A13)
O--MX
sincft
l-Msin<cos\// o~ - M x 0
sin</> sin\// /3 2 z
I- M sine/) cos \l/ cr-Mxo
z /.2
0/
"
where Eq. (A^) was used to evaluate x^h.
(dz/d^'j , = h/ (Al8)
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
Combining Eqs. (A20) and (A22) then gives Eq. (3) for P .
B
For correction to equal sideline positions rather than
equal radius, the ratio of P to P is
A C-
P
A/PC-= h/y,
Combining this with Eqs. (A20) and (A21) gives Eq. (2).
a;
Equation (3), -when combined with Eq. (Bl), gives the cor-
rected press-ore PB in terms of the measured angle ^ as
giving
P oc (I/COS0') (/32cosm+M)(l-Mcosm)
PB oc (i//32o-J) (l-
where
x' = r cos#' y' = r sin$'
References
search Center.
Michael J. Rudd*
Abstract
Introduction
281
282 M. J. RUDD
and
and
p d . t ) = ~ 1/4TTS / ( 3 2 p / 3 t 2 ) d 3 q
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
2
r (I,co)
P = co"L L L /16TT2s2(l-M cosG)4 p2[o3/(l-M cos6)] (8)
x y z c c
Laser Shadowgraph
Since the wavelength of light does not appear, this result also
can be obtained by geometrical optics. Similarly, let us inte-
grate over the shadowgraph along the y direction. Then the
spatial spectrum |$(,m)|2 becomes |$(,0)|2, since the inte-
gration acts as a low-pass filter. In turn, this quantity is
equal to the area under the normalized spatial correlation
function in the y direction multiplied by the power spectrum in
the x direction, i.e., L |$()|2. Then, the power spectrum of
the shadow bands in the x direction is
provided that the convection velocity does not vary too rapidly
with frequency; i.e., U varies slightly over a narrow fre-
quency band. Now, the term $(00) is related to the density
spectrum by a constant, i.e.,
BEAM
EXPANDER-
-^x
PHOTO-
DIODE
0-
A
LASER
SLIT
NOZZLE
Fig. 1. Laser shadowgraph system.
288 M. J. RUDD
Broadband Measurements
Cross-Correlation Measurements
Conclusions
References
7
Roe, G.E., "An Optical Study of Turbulence,11 Journal of Fluid
Mechanics, Vol. 43, September 1970, pp. 607-635.
8
Townsend, A.A., "The Interpretation of Stellar Shadow-Bands as
a Consequence of Turbulent Mixing," Quarterly Journal of the
Royal Meteorological Society, Vol. 91, January 1965, pp. 1-6.
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
EFFECTS OF TRANSDUCER FLUSHNESS ON FLUCTUATING
SURFACE PRESSURE MEASUREMENTS
Richard D. Hanly*
Abstract
291
292 R. D. HANLY
Introduction
O-RING
SEAL
WIND
TUNNEL
WALL
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
VARIABLE HEIGHT
TRANSDUCER, T1
FLUSH TRANSDUCER, T2
l \ . . /WIND TUNNEL
f:( WALL
BOLT
SHAFT
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
10,-4
n-5
O
FLUSH
y/6 = 0.00090
y/6 = 0.00045
10,-6
= 1.68
= 4.43 in.
10 -7
10 -3 10-2 10,-1 1.0 10
f/U 0
Fig. 3 Effects of transducer protrusion on power spectral
density.
296 R. D. HANLY
10~J r
= 1.68
= 4.43 in.
/y/5 = -0.00045
10 -4
y/5 = -0.00090
10 -5
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
10 -6
10"
10 -3 10 -2 1.0 10
f5/uO
Fig. 4 Effects of transducer submergence on power spectral
density.
10~3 r
y/5 = 0.0113
y/5 = 0.00452
/\ V /5 =
0-00226
"
io-33 10~ 1.0 10
(Fig. 4). When the PSDs for protruding II (Fig. 3) and those
for submerged TI (Fig. 4) are compared, it is evident that the
curves for negative y/6 show much less spread than do those
for positive y/6, as is to be expected from knowledge of the
Cp curves. The submerged PSD shapes, however, when com-
pared to that of the flush position, show a higher degree of
uniformity than do the protruding PSD shapes, indicating that
all frequency components of the signal are affected to nearly
the same extent in the submerged case but not in the protruding
case.
Coherence
/FLUSH = 1.68
,y/5 = 0.00045 = 4.43 in.
,y/5 = 0.00090
/y/6 = 0.00226
Downloaded by Boeing Company on July 15, 2013 | http://arc.aiaa.org | DOI: 10.2514/4.865206
1.0 r
FLUSH
y/5 = -0.00045
10-3
6 - y/5 = +0.00045 / /
f6/Uo
1.0 p
.8 -
y/5 = -0.00090
.6 -
.4 - y/6 = -0.00226
I _J
10- 10- 1.0 10
IU
I ^00
References
303
304 DISCUSSION: AEROACOUSTIC INSTRUMENTATION