What was the main focus of the NASA conference?

The main focus was on aircraft engine noise reduction, particularly the Quiet Engine project.

What is the Quiet Engine program?

The Quiet Engine program is an experimental initiative aimed at demonstrating effective noise reduction techniques in full-sized aircraft engines.

What historical context is provided regarding noise reduction research?

The document notes that the British pioneered jet engine noise reduction research in the early 1950s, followed by intensive programs in the U.S. involving NASA's predecessor, NACA.

What are some sources of engine noise discussed in the document?

Sources of engine noise include the fan and compressor noise from the inlet and fan discharge duct, as well as turbulent mixing noise from the primary jet at the rear of the engine.

Aircraft Engine Noise Reduction — NASA-SP-311 (1972)

Q: Who participated in the conference?

Participants included representatives from NASA Headquarters, NASA Langley and Flight Research Centers, General Electric Company, Boeing Aircraft Company, and Pratt & Whitney Aircraft Division.

Document

NASA SP-_,t!

CASE

AIRCRAFT ENGINE

NOISE REDUCTION

A conference held at LEWIS RESEARCH CENTER CLEVELAND, OHIO MAY 16-17, 1972

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

NASA SP-311

AIRCRAFT ENGINE

NOISE REDUCTION

The proceedings of a conference held May 16-17, 1972, at the NASA Lewis Research Center, Cleveland, Ohio Prepared at Lewis Research Center Scientific and Technical ln[ormation Office 1972 NATIONAL AERONAUTICS AND SPACE ADMINISTRATION Washington, D.C.

For sale by the National Technical Information Servtce Springfield, Virginia 22151 Price $6.00 FOREWORD The Lewis Research Center is the National Aeronautics and Space Administration's principal field installation for research and development of advanced aerospace-propulsion and power- generating systems. More specifically, a substantial part of the Center's activities is devoted to progress in the technology of aircraft propulsion. Work includes such diverse areas as com- ponents, controls and other aspects of installation, V/STOL and low-cost engine designs, and - in particular - noise reduction.

The results of our work are published as NASA reports and as articles in the technical journals. In addition, an occasional technical conference assists us in communicating more directly with others in the engineering fraternity. Accordingly, this con- ference - in which we are joined by NASA Headquarters, NASA Langley and Flight Research Centers, General Electric Com- pany, Boeing Aircraft Company, and Pratt & Whitney Aircraft Division - is held to present to you the results of recent and current work.

Bruce T. Lundin Director iii CONTENTS Page °..

lU FOREWORD ...................................

I. INTRODUC TION II. FAN NO]BE AND PERFORMANCE James J. Kramer, Melvin J. Hartmann, Bruce R. Leonard, Ill. FAN NOISE SUPPRESSION Charles E. Feller, John F. Groeneweg, Edward J. Rice, IV. JET NOISE Uwe H. yon Glahn, Vernon H. Gray, Eugene A. Krejsa, V. THE QUIET ENGINE PROGRAM VI. QUIET ENGINE DESIGN HIGHLIGHTS VII. QUIET ENGINE NACELLE DESIGN VIII. QUIET ENGINE TEST RESULTS Carl C. Ciepluch, Frank J. Montegani, Mike J. Benzakein, IX. QUIET ENGINE DEMONSTRATION Harry E. Bloomer .............................. ?.15 X. APPLICATIONS TO AIRCRAFT PROPULSION SYSTEMS Newell D. Sanders, W. Harry Close, Arthur A. Medeiros, XI. STOL NOISE SOURCES AND FAN NOISE TREATMENT XII. FLAP NOISE Robert G. Dorsch, Paul L. Lasagna, Domenic J. Maglieri, and William A. Olsen ............................. 9.59 XIII. DESIGN INTEGRATION AND NOISE STUDY FOR A LARGE STOL AUGMENTOR WING TRANSPORT XIV. SONIC INLET DEVELOPMENT FOR TURBOFAN ENGINES V I. INTRODUCTION Newell D. Sanders This conference is a progress report on some of the principal NASA programs for the reduction of aircraft engine noise. A major emphasis of the conference is focused on the Quiet Engine project.

The first portion of the conference is devoted to noise technology: fan noise and performance, fan noise suppression, and jet noise. This tech- nology applies to a broad spectrum of airplanes including new subsonic air- planes, the older subsonic airplanes in service today, STOL airplanes of the future, and future supersonic transports.

The second portion of the conference reports on the Quiet Engine pro- gram and demonstrates the noise reductions achieved. The Quiet Engine is an experimental engine incorporating many low-noise features. It was as- sembled for NASA by the General Electric Company. The engine experi- ment has three purposes: (1) to demonstrate the effectiveness of noise reduction techniques in a full-sized engine, a proof of concept approach; (2) to measure the performance penalties accompanying the use of noise- reduction techniques and to find methods of reducing these penalties; (3) to uncover obstacles to still further reductions of engine noise. Also as part of the Quiet Engine program, the Boeing Company, under NASA contract, designed and built experimental sound-absorbing mufflers to surround the engine and reduce the noise even more.

Following the Quiet Engine portion of the conference, possible applica- tions of low-noise technology to future airplanes and to the airplanes of to- day are discussed. Finally, a brief progress report on research for quiet STOL airplanes is given.

Historically, the British pioneered in research for the reduction of jet engine noise in the early 1950's. Later, the aircraft and engine industry in the United States together with the NACA (predecessor of NASA) engaged in an intensive program with the same objective. These programs resulted in the multitubed nozzles as shown in figure I-1. This photograph shows one of the nozzles that was tested in the Lewis wind tunnel. Nozzles

like this were used on many airliners. More complicated nozzles (fig. I-2)

in combination with ejector shrouds were also used. The moderate noise suppression from these nozzles did not stem the rising tide of public objec- tion to aircraft noise. With the advent of Sputnik and the Space Age, re- search on noise declined. The Lewis Research Center dropped noise re- search entirely.

The phenomenal growth of air traffic in the early 1960's intensified the noise problem. Recognizing the seriousness of the situation, President Johnson in 1966 directed appropriate Government agencies to plan programs for noise abatement. At that time, NASA began formulating new programs directed at the noise problem.

The first major new program was the Acoustically Treated Nacelle Pro- gram managed by the Langley Research Center. In that program, Douglas and Boeing built experimental mufflers to silence the fans on DC-8 and 707 airplanes. The program demonstrated significant noise reductions for air- planes approaching an airport, but noise reductions on takeoff were small.

Results of that program were fully reported in a conference in 1969 at the NASA Langley Research Center.

Some of the sources of engine noise are shown in figure I-3. The noise which is emitted from the inlet and from the fan discharge duct is generated principally by the fan and, to a lesser extent, by the compressor. The noise emitted by the primary jet at the rear of the engine is generated by the tur- bulent mixing of the jet with the surrounding air. This turbulent-mixing noise is responsible for the tremendous roar associated with jet airplanes.

Less important but not insignificant are turbine noises which are emitted at the rear of the engine.

A tape-slide demonstration of these noises has been prepared to show the basic elements that make up the noise from fan and jet engines. It is obvious from this demonstration that an aircraft noise reduction of only 3 decibels is barely noticeable. A 10-decibel improvement is clearly no- ticeable. A 20-decibel reduction is a large improvement, and the public would applaud the aviation community for such a reduction. Even a 30- decibel reduction may be sought in special cases as, for example, the STOL airplane.

Goals for airplane noise research were stated approximately 1 year ago

in the joint DOT-NASA Civil Aviation Research and Development Study re-

port (the CARD study). The recommended goal was a 10-decibel noise reduction per decade. Figure I-4 shows some noise levels related to this goal. The noise level for the DC-8 at the FAR-36 (Federal Air Regulation 36) takeoff point is approximately 116 PNdB (perceived noise decibels).

The certification limit for new airplanes of the same gross weight is 104 EPNdB (effective PNdB). The new wide-bodied jet airplanes equal or better this value. Thus, the first 10-decibel reduction has been achieved. As you will hear later in this conference, the Quiet Engine is close to 90 EPNdB, which is below FAR-36 minus 10 decibels. The next level, FAR-36 minus 20 decibels, is below demonstrated noise levels for airplanes of this class at takeoff.

Figure I-1 8-LOBE SUPPRESSOR AND EJECTOR ON AIRCRAFT Figure I-2 TURBOFAN-ENGINE NOISE EMISSION AIRFLOW " "_'__ _ -- .... ,,_\ .... _ - INLET " PRIMARY FAN DISCHARGE JF DUCT Figure I-3 CS-63Z07 NOISE REDUCTION GOALS TAKEOFF;FOURENGINES EPNdB 116 -- 104 -- FAR 36 FAR 36 FAR 36 - 10 - 20 Figure I-4 II. FAN NOISE AND PERFORMANCE James J. Kramer, Melvin J. Hartmann, Bruce R. Leonard, Jack F. Klapproth., and Thomas G. Sofrin** For an engine like the Quiet Engine, which is designed for long-range subsonic cruise at conventional takeoff and landing, the fan noise controls the engine noise. This discussion includes the generation process of fan noise and what can be done to control it by design changes in the fan.

In discussing the interplay between the aerodynamics and acoustics of fans, a family of fans are referred to which were tested as part of the Quiet Engine Program. The design characteristics of these fans are tabulated in table II-1. For simplicity in terminology, the fans are lettered A to G.

TABLE II-1. - FAN DESIGN CHARACTERISTICS Fan Pressure Tip speed, Number ft/sec ratio of stages A 1160 1.5 B 1.5 C 1.6 D 1.4 E 1107 1.5 F , 1.6 G 1000 1.45 Fans A, B, and C were furnished by the General Electric Company as part of the Quiet Engine Program. They were designed and built by General *General Electric Company.

* * Pratt & Whitney Aircraft Company.

Electric, tested for aerodynamic performance, and then tested for acoustic performance at Lewis Research Center. Fans A and B are single-stage, low-speed fans designed for a pressure ratio of 1.5. Fan C is a single- stage, relatively high-speed fan designed for a slightly higher pressure ratio. Fans D and E were designed, built, and tested at Lewis. They were designed as single-stage fans with pressure ratios of 1.4 and 1.5. Fan F was designed and built by Pratt & Whitney. That fan is now installed in the Lewis fan noise test facility, and data are being generated with it at this time. Fan G isthe only two-stage fan that was tested as part of the Quiet Engine Program. It was designed for low speed at 1000 feet per second and a pressure ratio of 1.45. It is a modified version of the inner panel of the fan used on the TF-39 engine.

The first and most important thing a fan must do in order for it to be an integral part of a useful, viable engine system is to produce pressure rise.

The fan requirements necessary for a quiet engine are discussed in the next section.

FAN REQUIREMENTS The Quiet Engine Program is directed toward providing a fan engine for long-range CTOL aircraft. These engines must be economical, quiet, and clean and must provide good operating margins and growth potential. The aerodynamic requirements of the fans are established by the engine cycle conditions. It is necessary to determine the optimum fan flow, pressure ratio, and operating characteristics for the particular thrust, cycle pressure ratio, and temperature of interest. Some of the considerations and areas of interest can be shown by the following trend curves.

In figure II-1 the trends of specific fuel consumption and fan pressure ratio with bypass ratio are indicated at two levels of cruise Mach number.

For a cruise Mach number of 0.80, the specific fuel consumption continues to decrease with increasing bypass ratio. For this cruise Mach number, fan pressure ratios of 1.4 to 1.6 are considered (crosshatched area). This is the range of fan pressure ratios studied in the Quiet Engine Program. To indicate the effect of cruise Mach number on these parameters similar curves are shown for a cruise Mach number of 0.95. At this higher cruise

Machnumber, fan pressure ratios of 1.9 to 2.1 wouldbe of interest (cross-

hatchedarea).

It would appearthat evenhigher bypassratios may be desirable for

lower fuel consumption;however, the enginesbecomevery large. This is

illustrated in figure II-2; the lower-pressure-ratio fan enginesare asso-

ciated with lower fan jet velocities and, as shownin the sketch, are very

large. At the higher pressure ratios, substantially higher fan jet velocities

occur but the fans are relatively smaller and of lower bypassratios. The

higher-pressure-ratio fans may require two or more stages. The low-

pressure-ratio fans can be single stage, but they require a large nacelle and

installation drag becomesa large fraction of the thrust. The large fans

must be driven by the relatively small-diameter turbine. This may require

increasing the number of turbine stagesto achievethe necessarypower from

the small core flow. This increasedengineweight must be consideredas a

penalty for the very-high-bypass-ratio enginesystem.

The fan must also be suitable to meet all the required engineoperating

conditions. This can be illustrated by referring to a fan performance map

such as shownin figure II-3. Flow curves are given for three rotational

speeds. The highest speedline shownincludesthe cruise operatingpoint.

Abovethe cruise operatingpoint, the fan encountersstall andbecomesun-

stable. Similar pressure ratio flow curves are indicated at the speedre-

quired for takeoff andat a lower speed. The fan must producethe desired

pressure andflow alongthe cruise andtakeoff operatingconditions The

positions of the takeoff andcruise operatingpoints dependon the cruise

Machnumber andthe enginecycle. Theseoperatingconditionsshouldbe in

the region of high fan efficiency. They must also be sufficiently removed

from the stall line to allow engineacceleration andtransient operation. The

margin betweenthe operatinglines andthe fan stall must be sufficiently

large to take into accountinlet flow distortions, as shownschematicallyin

figure II-4. Crosswindsandinlet flow distortion causethe stall line to be

movedtoward the operatinglines. There are a number of additional reasons

for building or selecting the fan design to havea high stall margin. These

includethe eventualperformancedegradationwhich results from blade for-

eign object damageandprovision of a thrust growth potential. Enginethrust

growth requires an increase in either fan flow or pressure ratio. Fans de-

signed with large stall margins can be more easily uprated with minor hard- ware changes.

Having established the overall requirements, the fan must be selected to meet the requirements of pressure ratio, efficiency, and operating mar- gins. Fan pressure ratio is established primarily by the rotational speed and aerodynamic loading, as shown in figure II-5. On this calculated graph the rotational speed is given as rotor blade tip speed. Increasing the blade tip speed at a given aerodynamic loading results in a substantial increase in fan pressure ratio. However, at the high speeds the rotor blades must op- erate at a higher relative Mach number, as indicated on the lower scale.

The blade shape must be selected to be suitable for the flow Mach number level used. Increasing the aerodynamic loading, or the lift of the airfoil sections, results in increasing the fan pressure ratio at any given speed.

However, at some level of loading, the airfoils reach a "stalling condition" and the machinery cannot operate. The limit region is shown by the "shad- ing" and is not a clearly defined limit as it depends on details of blade de- sign. The levels of aerodynamic loading parameter indicated are for the rotor blade tip region, which usually becomes the critical airfoil section.

Rotor tip elements usually initiate fan stall. Thus, to achieve a high stall margin, it is necessary that the rotor blade tip aerodynamic loading be held somewhat below the limit region shown.

Fan pressure ratios from 1.4 to 1.6 are considered for the Quiet Engine Program. This would indicate that blade tip speeds of 1000 to 1600 feet per second should be considered. The higher speeds would be at lower aerody- namic loading and potentially should lead to a larger stall margin. The relative flow Mach number to the blading for the high-speed fan would ap- proach 1.6. The fans designed for the Quiet Engine Program incorporated low-noise features, as shown in figure II-6. These fan performance points are shown on the background plot as solid symbols. Fans A and B were de- signed for about ll50-foot-per-second tip speed, and fan C was designed for 1550-foot-per-second tip speed. The open symbols represent a number of experimental fans designed primarily to investigate aerodynamic perform- ance over a range of rotor tip speeds from 1000 to 1800 feet per second.

These experimental fans are of relatively high aerodynamic loading, with pressure ratios ranging from 1.5 to about 2.2. Fan C and one experimental fan are at substantially lower levels of aerodynamic loading.

Peak efficiency for these fans is indicated in the upper portion of fig- ure II-6. The band of data indicates that efficiencies of 0.88 to 0.84 can be achieved over this range of designs. The blade shapes were selected for the blade Mach number and loading level of the design. Fans A, B, and C of the Quiet Engine Program are included in this efficiency band. Fan A has a slightly higher efficiency than fan B.

Measured stall margins for fans A and B are shown on the bar chart in figure II-7; the bars are located at about the design speed. The stall margin for fan C was not measured above 90-percent speed. The stall margin for fan B was estimated to be at least 22 percent, as shown. Stall margins for some of the experimental stages are shown. The aerodynamic loading of the experimental fans is generally somewhat higher than that of the Quiet Engine Program, and the stall margins are generally slightly less. Most of these fan stall margins were measured with initially built hardware, and some stall margin improvement may be available through small blading changes.

For example, two low-speed highly loaded stages, as indicated by the solid bars, were improved to 16- and 17-percent stall margin by stator reset and redesign. No particular trend of stall margin is intended by this figure. In long-range CTOL aircraft, it is probably desirable to have a fan stall mar- gin of the order of 20 percent. In the range between 1100- and 1600-foot- per-second blade tip speed, where most of the fans have been investigated, a 20-percent stall margin appears feasible.

The fan requirements are determined by the engine cycle and applica- tion. The fan rotational speed and aerodynamic loading level must be selec- ted to meet the performance and operational requirements of the engine and the application. Fans A, B, and C of the Quiet Engine Program, which in- corporate low-noise features, are shown to have efficiency and stall margin trends comparable to a rather wide range of experimental fans.

FAN NOISE FUNDAMENTALS Fan noise is a prominent component of aircraft powerplant noise. As shown in figure II-8, it is generated internally by the fan components, prop-

agatesthrough the inlet andfan dischargeducts, andpropagatesto the

groundfrom both inlet anddischargeduct planes. Additional important en-

gine noise sourcesare indicatedandinclude let mixing noise from fan and

primary streams, turbine noise, and core compressor noise radiated for-

ward through the fan.

During an airplane flyover, noisesfrom theseenginesourcespeakat

different instants andcontribute to a complicatedhistory of the flyover noise

event. Sinceall enginenoise sourcesare also presentduring static ground

noise tests, special meansmust be usedto separatefan noise so that its

characteristics canbe studied. Someof the special test rigs usedin fan

noise programs will be described. Throughtheir use, supplemented by

analytical methods, a fairly clear picture of fan noise hasemerged.

Fan noise radiates in a complicateddirectivity pattern from the inlet

anddischargeducts. At eachangular position in the far field the noise is

convenientlydescribedby meansof its spectrum, or frequencydistribution

of acoustic energy. Figure II-9 gives two illustrative spectra of forward-

radiated fan noise. In the top half of this figure, which is typical of

subsonic-blade-tip-speedsingle-stage fan operation, a broad baseof con-

tinuous sound, called broadbandnoise, exists over a wide frequencyrange.

This noise is generatedby random pressure fluctuations in the fan and

sounds,generally, like jet mixing noise. Prominent features of the spec-

trum are the two sharppeaksor "spikes" at fan bladepassagefrequency

andits harmonic. Thesediscrete tones correspondto fan whineandare

generatedby periodic pressure fluctuations on the fan airfoil surfaces.

At supersonicfan bladetip speeds,noise of a different character ap-

pears. It is exhibitedin figure II-9 as a closely spacedseries of discrete

frequency spikes separatedby shaft rotation frequencyandhas a distinctive

sounddescribedvariously as deep,rich, raspy, or buzz saw. Multiple-pure-

tone (MPT) noise, combination-tonenoise, andbuzz saware the usual de-

scriptive terms. In brief, it results from the rotating pattern of shock

wavesfrom the fan bladeswhentheir tip speedis supersonic. Becauseof

the nonlinear propagationbehavior of shockwaves, slight variations in the

shock structure from bladeto blade, causedinitially by small manufacturing

deviations, are amplified with distance from the fan into a pattern having

markedirregularities. Sincethis irregular pattern turns with the rotor, its

frequencyis that of shaft rotation speedandall multiples thereof rather

thanbeing restricted to blade passagefrequencies, which wouldresult from

theoretically identical blading.

These noise spectral features maybe incorporated in a subjective

rating of the noise exposureby meansof units suchas PNdBin a manner

describedpreviously. But a better understandingof fan noise details for

purposesof noise reduction can only be reachedif the processesthat pro-

ducethe spectral components are lookedinto still more closely.

Three processes are involved in fan noise, as illustrated in figure

II-10: generation, duct propagation, andradiation. Understandingof fan

noise is facilitated by examiningthe features of theseprocesses, andnoise

reduction conceptsare mademore feasible to developby separatingthe

overall problem into theseareas.

Basically, fan noise is generatedat the rotating bladesand stationary

vanes. Both broadbandanddiscrete bladepassagenoise result from fluc-

tuatingblade surface pressures. These fluctuations arise from inflow

velocity variations into a stagedueto the turbulent flow structure andto

airfoil wakesfrom upstream stages. At supersonicfan blade tip speeds,

additional noise is producedby the blade shockwavepatterns. Thesegen-

eration processesare more fully reviewedin this section. It will serve to

point out here that the essential generationproblems are aerodynamicrather

than acoustic in nature. While the basic performance of a fan isunderstood,

analytical description of the flow details is far from complete. And it is

exactly thesedetails, in particular the unsteadyfeatures, that are respon-

sible for the infinitesimal fraction of the mechanicalpower convertedto

acoustic energy. Several theories for unsteadyblade airloads exist, but to

date they havenot beencritically comparedin the context of fan geometry.

More importantly, no comprehensivemeasurementsof the airfoil fluctuating

surface pressure distributions haveyet been madein a fan to checktheory.

A rich andimportant field of aerodynamicresearch is here awaitingmuch

neededanalytical andexperimental exploration.

As will presently be described, both discrete-frequency blade passage

noise andmultiple-pure-tone noise generatedby the fan aerodynamicloads

take the form of coherent, periodic acousticwavepatterns in the fan inlet

andexit ducts. Thesepatterns propagatein spiral pathsto the free ends, from which they then spread out and radiate to free space. The process of duct propagation has received a great deal of attention because it is subject to control both by sound-absorbing duct lining and by the selection of fan rotor and stator geometry. Duct lining is the subject of a subsequent paper and will not be reviewed here. Propagation in a hard-wall duct also has features which can be exploited for noise reduction. It is well known that a pipe, as exemplified by a "speaking tube," is a good conductor of sound.

These tubes usually involve the equivalent of a pulsating piston or diaphragm at the transmitting end, which vibrates to and fro in a synchronous man- ner. In a fan or compressor the excitation or driving force on the duct is fundamentally different. The sound does not pulse back and forth in an um- brella fashion; instead there is a rotating pattern of circumferentially alter- hating high- and low-pressure ridges extending radially from the centerline to the walls. This spinning excitation pattern may be visualized by recog- nizing qualitatively that the pressure very near the face of the rotor alter- nates periodically as the blades spin by a fixed point. This spinning pattern, containing B (where B is the blade number) cycles of pressure variation around the circumference, provides the excitation and differs markedly from the pattern of a synchronous piston-like structure.

Whereas a synchronously pulsing pattern always propagates axially in a duet, a rotating pattern behaves in a more complex manner. It has been learned that the circumferential velocity component with which the pattern sweeps the wall must actually be supersonic to allow it to propagate in a spiral path along and out the duet. If the pattern spins more slowly, such that its circumferential wall speed is subsonic, it will not truly propagate.

Instead, the magnitude of the pressure fluctuations will decay exponentially with axial distance from the source plane. Thus, the rotating pressure field of a subsonic-tip-speed fan rotor decays inside the inlet and the discharge duet and does not contribute to the noise heard from the powerplant. How- ever, the interactions between rotor blades and stator vanes or other cir- cumferential disturbances produce other patterns of noise at blade passage frequencies. It will be shown that in many eases interaction noise is pro- duced in patterns that spin at high supersonic speeds. These patterns spiral easily through the fan ducting and are responsible for the noticeable discrete frequencies present in the spectra of fan noise. Controlling the blade-wane

number ratio makes possible bringing the circumferential wall speed of the

interaction patterns below sonic speed, causing them to decay inside the powerplant, and thus effectively eliminating this source of noise. Unfortu- nately, there is sufficient flow distortion due to other influences affecting the fan so that there remains a considerable level of blade passage noise after blade-vane interaction has been eliminated. Many details of duct propaga- tion remain to be explored. These include primarily the effects of varying duct contours and airflow gradients. Methods for analyzing the propagation of fan broadband noise are considerably more involved since the processes are basically random and statistical considerations must be employed.

The last stage in fan noise transmission involves radiation into the far field of the patterns spinning at the duct faces. Comparatively little control of this part of the noise chain is possible, but studies of radiation from the spinning duct patterns have led to understanding of the complex directivity patterns of fan noise. On the axis there is a minimum level, which rises and fails in a series of lobes in an arc around the engine. Angular locations of these peaks and valleys have been linked to the details of the pattern shapes at the engine inlet and exit ducts. Since the radiation directivity pat- terns govern both far-field peak levels and duration in an overflight, con- sideration of this phase of fan noise emission completes the tracing of the important fan noise processes from their source to the ground observer.

Special facilities have proved indispensable in fan noise studies. These include both rigs for running isolated components and special instrumenta- tion systems for recording and processing test information. Figure II-11 presents representative examples of the variety of equipment used.

Figure II-11 shows the inlet to a 52-inch-diameter single-stage research fan in an outdoor noise test facility that allows far-field surveys to be made of inlet and discharge noise. Prominent in the foreground is a probe micro- phone system with axial, radial, and circumferential traversing capability.

It is by means of such traverse equipment that the complicated spinning pat- terns of fan interaction and combination-tone noise are mapped.

Anechoic chambers play an important part in all major noise studies, and fan noise programs have been no exception. The NASA Langley Research Center anechoic chamber is also shown in figure II-11, with a small-scale compressor installed. In addition to allowing weather-free operation the echo-free environment has proved instrumental in permitting accurate fan noise direetivity patterns to be measured without contamination by wind or thermal and ground reflection influences.

Studies of multiple-pure-tone noise involve details of the blade shock wave patterns at supersonic speeds. To visualize these patterns, optical methods such as schlieren photography have been employed in supersonic cascade wind tunnels, as shown in figure II-11. Work is underway to con- duct similar optical programs in special rotating fan rigs.

The effects of forward flight speed upon fan noise generation, propa- gation, and radiation are being explored (fig. II-11) in an acoustic wind tun- nel in the open throat of which is installed a miniature working model of a fan-powered nacelle driven by a high-pressure air turbine. Again, anechoic construction of the working section enables free field effects to be reliably measured. In spite of its size the small (4.2 in. diam) fan can be driven at supersonic tip speeds, and multiple pure tones are clearly produced.

Discrete frequency noise at blade passage frequency and its harmonics was one of the most conspicuous noise characteristics of early turbofan powerplants. These fans employed inlet guide vanes, and it was found that the cutting of the vane wakes by the rotor blades was a major noise source.

Current high-bypass-ratio engines dispense with inlet vanes, but the shed- ding of wakes from the rotor blades into the downsteam exit stators can be a similar source of blade passage interaction noise. How these interaction patterns are generated is described in the following paragraphs.

Figure II-12 portrays in a developed view a row of moving blades shed- ding wakes into the vanes of a downstream stator. As a wake passes by a stator vane the effective angle of attack and the velocity change, producing a fluctuating lift distributed over the vane surface. The life fluctuates with a base frequency equal to the blade passage rate and constitutes what is called a dipole source of sound. In the illustration, equal numbers of blades and vanes are employed, so that when one vane is about to receive a blade wake all the other vanes are at the identical point in their cycles. As shown, the stator vanes would thus produce an array of sources pulsating in unison. In practice, however, the numbers of blades and vanes are different, so that while one vane would be on the verge of intercepting a blade wake, another vane might just have recovered from a wake passage. Generally, there is a sequential phasing of the wake interaction events around the stator assembly.

This phasing forms a pattern that sweeps circumferentially around the vane array rather than in a synchronous, umbrella-like oscillation. The num- ber of complete cycles in the pattern has been derived analytically in terms of blade and vane numbers, and the existence of such patterns has been fre- quently confirmed on a variety of rigs and engines.

However, the plausibility of this interaction pattern effect is most di- rectly perceived by use of a simple optical analogue called the Moir_ effect.

If the stator assembly is represented by an array of radial spokes drawn on stationary background and the rotor blading is similarly represented on a clear sheet of plastic, when the two patterns are overlaid the interference of light and dark regions will produce a pattern of resultant intensity that sug- gests the blade-vane interaction effect. Consider now two such arrays, one containing 46 spokes and the other having 48. The resulting two cycles (48 - 46 = 2) of intensity variation are conspicuous when photographed. In a live demonstration, one wheel can be turned slowly about a common concen- tric axis. The interference pattern will be observed to spin comparatively rapidly. In fact, if the 46-spoke pattern is turned, the interference pattern will turn 23 times (46/(46 - 48) = 23) as fast as the simulated rotor in the opposite direction. The number of lobes or cycles of variation in the inter- ference pattern and its rotational speed can be changed by changing the num- ber of vanes in the stator. Parenthetically, this does not change the blade passage frequency, just the associated acoustic mode structure. Consider now the Moir_ pattern for 46 rotor blades in conjunction with 60 simulated stator vanes. Although the pattern is not as well defined as the first case, 14 interference cycles can be counted from photographs. When the plastic 46-spoke rotor overlay is turned, the interference pattern will be observed to turn more slowly than in the first example. In fact, it turns at a multiple of rotor speed given by the expression 46/(46 - 60) _ -3, the minus sign in- dicating a backward motion.

It will be recalled that the requirement for a spinning acoustic pattern to propagate in a duct is that its tip speed at the wall be supersonic. These optical diagrams make it quite clear that supersonic interference patterns can be generated by interaction effects when the rotor itself is subsonic.

They also show how the pattern can be slowed down by increasing the number

of stator vaneswithoutchangingrotor speed. Whenthe interference pat-

tern itself drops belowsonic speed,it no longer propagatesaxially from

the generating zone, andnoise dueto this interaction is essentially elimi-

nated. Thesefeatures of interaction noise havebeenexploitedin the design

of modern turbofanpowerplants. While major interaction noise sources

havebeeneliminated, discrete frequencynoise still exists becauseof non-

uniform inflow. However, there has beena markedimprovementover early

turbojet andturbofandiscrete frequencynoise.

Sourcesof fan broadbandnoise are suggestedin figure II-13. As with

discrete interaction noise, it is generatedby airload fluctuations onthe

blade andvanesurfaces. In this case, thesefluctuations are randomin time

instead of periodic. Turbulencein the air entering the rotor, turbulence

from the rotor impactingthe stator, andboundary-layerturbulent fluctua-

tions all constitute random broadbandnoise sources. Quantitativeunder-

standingof this category of noise is in a primitive stagecomparedwith the

discrete frequencycasebecauseof the inherently greater complexity asso-

ciated with turbulenceandrandom processesgenerally.

Simplified analysesof onepart of the problem, the interaction of a rotor

with incomingturbulence, haveprovided insight into the shapeof the broad-

band spectrum. If the scale of the turbulenceis small comparedwith the

blade chord, the correlation betweenlift fluctuations occurring on neighbor-

ing bladeswill be low, andthe resulting noise-against-frequencydistribu-

tion will be relatively flat. Onthe other hand, if the axial scale of inflow

turbulence is large, correspondingto long streaks, several blades will suc-

sessively cut through eachsuchnonuniformity, generatingbursts of noise at

the blade passagefrequency. The resulting spectrum will contain peaks

centeredaroundblade passagefrequencies, andthe peakswill becomepro-

gressively sharper as the axial scale of the turbulenceis increased. Con-

sequently, in controlling broadband noise, it is clearly important to ensure that the flow passages are designed to produce the smoothest possible flows.

As described previously, combination-tone or multiple-pure-tone noise is associated with the shock waves produced by rotors operating at super- sonic tip speeds. It is helpful to consider two cases: first, an ideal rotor containing perfectly spaced identical blades; and second, an actual rotor

assembly incorporating small blade-to-blade deviations in shapeand orien-

tation.

Figure II-14 portrays the shockwavestructure attachedto the leading

edgesof an ideal supersonicrotor. Onthe right side of the figure is a rep-

resentation of the pressure-time history that wouldbe detectedby a probe

microphonein the inlet duct. A repetitive sawtoothpattern results as the

successionof shockwavespassesby. In a relatively short distance from

the rotor the amplitude of these shockswill haveattenuated,becauseof

nonlinear effects, to what is called an acoustic disturbanceor Machwave.

This symmetric wavepattern propagatesin a spiral path out the inlet since

its circumferential wall speedis supersonic. It wouldbe detectedin the

far field as a sharpdiscrete noise at blade passagefrequencyandits har-

monics.

In practice, whathappensis significantly different. Actual rotors con-

tain small blade-to-blade differences dueto manufacturingdeviationsand

service wear. Thesevariations are usually small; this hasbeenconfi.rmed

by movingan inlet duct probe microphoneclose to the blade leading-edge

plane of test rotors. In this very close location the shockwavestructure of

a normal productionrotor is extremely uniform. Correspondingly, the

spectrum of the noise is clearly dominatedby blade passageharmonics.

However, there are smai1variations in shockstrength from oneblade to

another. The importanceof these normally negligible deviationsis that

shockwavebehavior is nonlinear: the higher amplitude shockspropagatea

little faster than their lower amplitudeneighbors. Consequently,the uni-

form pattern existing very close to the rotor becomeswarpedas distance

from the rotor is increased. Despite the small magnitudeof the initial

shockamplitudes, there is sufficient difference in their nonlinearbehavior

to producea markednonuniformity in amplitudeand spacingof the pattern

within a short axial distance of the rotor.

Figure II-15 portrays the phenomenon just described. To the right of

the figure is shownthe pressure-time trace recorded by a probe microphone

placeda few chord lengths aheadof the rotor. Here, the pressure irregu-

larities are conspicuous;in manycasesthere is no visible evidenceof blade

periodicity. The pattern repeats faithfully with every turn of the rotor, so

that the spectrum of the resulting soundwill havea fundamentalfrequency of

shaft rotation speedrather thanblade passagefrequency. Sincethe pattern

contains manysharp irregularities, a large number of harmonicsresult, giving rise to the multiple-pure-tone (MPT) noise descriptive term.

It might be supposed that it would be relatively simple to eliminate MPT noise by sufficiently close control of rotor construction, but this has not proved possible. Even if a sufficiently perfect rotor could somehow be pro- duced, unequal blade wear in service would soon cause enough irregularity to develop MPT noise. Nor would it be desirable to have such a perfect rotor; its sound would be the shrill whine of blade passage frequency, which is much more disturbing than the distributed tonal quality of MPT or combination-tone noise. Reduction of this noise component is being explored by designing blading to reduce the strength of all the blade shocks and by means of sound-absorbing wall liner constructions. Wall lining has proved quite successful in reducing multiple-pure-tone noise.

An understanding of these basic principles of fan noise has been used to improve significantly the characteristics of modern turbofan engines now entering service and planned for future use. Inlet guide vanes have been eliminated and the spacing between rotating and stationary blade rows has been increased. The well-known effect of increased rotor-stator spacing is shown in figure 11-16. Figure II-17 compares the fan geometry of a modern high-bypass-ratio siugle-stage fan with an earlier generation low-bypass- ratio two-stage turbofan engine. The noise reduction features include re- moval of inlet guide vanes, reduced fan tip speed, elimination of the second- stage fan and its interstage stator, increased rotor-stator spacing, and selection of stator vane number to eliminate interaction effects.

This review of fan noise fundamentals has identified the sources of the several types of fan noise and has indicated measures that can be taken to reduce noise in powerplant design. However, the problem of predicting ac- tual levels of noise produced by a specific design configuration contains many uncertain elements. Two general methods are available. Noise levels may be calculated from the results of theoretical analyses of the generation, duct propagation, and radiation phases; or they may be obtained empirically from scale-model test data. The latter method is usually more reliable.

There are so many steps in the theoretical calculations and these involve un- confirmed or doubtful assumptions at several stages that the reliability of the 2O

calculatedendproduct is questionable. This argument does not indicate that

the theoretical aspects of noise generation should be ignored. Theory has suggested many useful concepts for experimental evaluation, several of which are currently in use. But a great deal of theoretical work and experi- mental verification remains to be done before reliable noise level predic- tions can be calculated on a purely theoretical basis.

Figure II-18 portrays the alternative theoretical and empirical methods used to predict one type of fan noise - blade-vane interaction noise at blade passage frequency. Both processes start with given information about the fan geometry and operating conditions. The van and blade numbers and the rotor speed are such that interaction noise propagates. On the left part of the diagram are shown three sources of this noise: impacting of velocity de- fects from the rotor into the stator, perturbation of the stator vane loads by the passage of the rotor blade potential fields, and the effect upon the rotor blade loading due to its cutting the upstream potential field of the stator vanes. These load fluctuations must be calculated at several spanwise loca- tions. From the rotor fluctuations can be calculated the acoustic field gen- erated in the duct, and its propagation through the duct can be determined.

The acoustic field generated by the stator has to pass upstream through the rotor before reaching the inlet ducting. Calculation of the transmission pro- cess through the rotor is a highly involved, completely unchecked procedure.

The rotor and transmitted stator fields combine in the inlet to produce a re- sulting pattern at the inlet face. This pattern involves significant radial variations that add to the complexity of the far-field directivity pattern, which is the final stage of the calculation. The outcome of such calculation procedures is subject to considerable uncertainty.

On the other hand, test results from an appropriate model of the fan geometry are relatively straightforward to obtain. Normalizing the data yields fairly reliable predictions of blade passage noise as a function of op- erating parameters. Eventually, as more data on a greater variety of con- figurations are compiled and as theoretical aspects of the processes are confirmed, it will be possible to combine both theoretical and empirical me- thods to establish reliable prediction methods for new fan configurations.

FAN AERODYNAMIC-ACOUSTIC DESIGNINTERACTIONS So far the fan component aerodynamic requirements and the acoustic characteristics have been considered separately. Quite often, the features desirable for acoustics are not compatible with best aerodynamic perform- ance.

Detailed design trade studies must be made in order to obtain a fan and engine configuration which best balances the conflicting requirements. To be meaningful, these trades must be assessed on the basis of the total sys- tem aspects. Not only the fan component, but also the acoustic performance must be evaluated by the subjective reaction of the ground observer to the aircraft flyover. The engine performance should be measured by the overall aircraft economics. The evaluation requires consideration of all the ele- ments which contribute - with the problem being complicated by strong in- teractions between these elements. For example, the appropriate noise evaluation requires an accounting of (I) The noise source, such as the fan (2) Propagation of the noise through the acoustically treated ducts (3) Radiation from the engine to the ground observer (4) The subjective response of the observer The interactions between these elements can be strong enough that a lower fan source noise, for example, does not necessarily mean a lower noise as measured by the observer.

In order to establish the proper overall perspective, it is necessary to consider all the successive elements, including the interaction between the elements. The design compromises between some of the more important aerodynamic and acoustic fan design parameters are considered herein - in particular, (1) Rotor-stator spacing (2) Number of rotor blades (3) Vane-blade ratio (4) Fan tip speed (5) Blade design Rotor-Stator Spacing The first item of rotor-stator spacing is illustrated in figure II-19. The spacing is defined as the separation between the trailing edge of the rotor and the leading edge of the stator. This separation is measured as a frac- tion of the rotor blade chord. The acoustic effect of this spacing is obtained by testing at several different spacings with the resultant impact on the noise shown in the right part of the figure. The maximum aft tone-corrected per- ceived noise level (PNLT) is plotted as a function of spacing-to-chord ratio for takeoff and approach. A continuing reduction in noise with increased spacing is shown, with the effect being more important to the approach con- dition than to the takeoff. In this case, increasing the spacing from approxi- mately 1/4 chord, which would be usual for normal turbomachinery design practice, to two chords reduces the aft PNLT by 6 to 8 dB.

Large spacings between the rotor and stator obviously tend to increase the engine length and weight. However, by adjusting the engine design such that some of the core compression stages are moved from the high-pressure spool to the low-pressure spool and placed ahead of the frame, the penalty associated with the large spacing can be minimized.

The aerodynamic effects of increased spacings can be illustrated by using results of recent NASA fan component tests as shown in figure II-20.

Generally, aerodynamic studies are conducted in separate facilities from noise tests; so the aerodynamic effects of blade row spacing on the same fan on which these acoustic data were obtained cannot be shown. Spacing has been investigated in a fan designed for a pressure ratio of 1.5 at about ll00-foot-per-second rotor blade tip speed. This places it in the range of a moderately high aerodynamic loading. The 53 rotor blades are of high as- pect ratio. The spacing shown in the upper half of the figure is 31 rotor blade chord lengths; 112 stators were used. Considerable wall curvature is required in the region of the blades to achieve the desired area change or annulus area contraction over this stage. The stage was also tested with other reduced stator spacings. The blading is shown at a spacing of 1 rotor chord in the lower part of the figure.

The trend of measured peak fan efficiency with blade row spacing is shown in figure II-21. In the range from 1 to 31blade chord spacings, about 2 percentage points in efficiency are lost. This aerodynamic penalty may be somewhat different for various blade designs. But the general trend and penalty must be traded off against the indicated noise reduction for in- creased blade row spacing.

To summarize, increasing the spacing between the rotor and the stator permits a significant reduction in the source noise of about 6 to 8 PNdB for an aerodynamic penalty of slightly less than 1 percent in fan efficiency. Ex- perience has shown that low-noise designs cannot be achieved without some penalties, and this trade must be considered a relative bargain. Conse- quently, all advanced Quiet Engines incorporate a wide spacing between the rotor and the stator.

Rotor Blade Number The second parameter influencing the aerodynamic-acoustic fan char- acteristics is the number of rotor blades. In the Quiet Engine Program, two low-speed fans with similar aerodynamic characteristics such as tip speeds, solidifies, and pressure ratios but differing primarily in the number of blades were tested. These fans are shown schematically in figures II-22 and II-23. Both of these fans were built and tested aerodynamically and acoustically in full size. The advantages of fan A, with the large number of blades, were a lower noise over the full-speed range; higher efficiency; a shorter fan which still satisfied the blade spacing ratios discussed pre- viously; and a lighter weight - particularly when considering the total weight of the fan blade, the disk, and the fan containment.

Fan B, with fewer blades, has advantages from the standpoint of fewer parts, potentially lower manufacturing costs, and a higher stall margin for a given tip speed. When confronted with the problem of choosing between the two fans, the better performance and lower noise of fan A more than offset the cost advantages of the low-speed fan. Adequate stall margin can be ob- tained with the high number of blades by proper choice of fan tip speed and aerodynamic design.

Vane-Blade Ratio The third parameter is described as the ratio of the number of stator vanes to the number of rotor blades, or the vane-blade ratio. (The effect of this ratio on the rotor-stator interference patterns and the desirability of maintaining an appropriate relation between the number of stator vanes and rotor blades has been discussed in a previous section of this paper.)

Noise characteristics of a typical fan were measured by using two dif- ferent stators aerodynamically matched to the rotor discharge conditions, as shown in figure II-24. The acoustic characteristics shown in the figure illustrate the increase in noise that can occur when the number of stator vanes is reduced from 108 - which satisfies the desirable acoustic relation - to 76. Increases in noise of more than 5 PNLT are observed at the low speeds where the cutoff phenomenon occurs, with the penalty reducing to about 2 PNLT at the takeoff condition.

The combinations of the desirability for a large number of rotor blades along with the desirability of a vane-blade ratio of about 2 leads to a large number of thin, narrow stator vanes, which impacts on both the aerodynamic performance and the aeromechanical characteristics of the vane.

Keeping the ratio of stator to rotor blades high has resulted in critical stator range problems. A fan which was designed for a pressure ratio of 1.5 at a rotor blade tip speed of 1000 feet per second is shown in figure II-25. In this case, 24 rotor blades were used. The gap corresponds to 2 rotor blade chords, and 64 stators were used. Thus the stator vanes are rather high- aspect-ratio (4.1) blades. The performance of this fan is shown by the dashed lines in figure II-26. Stage pressure ratio increases to about 1.5 and then is flat as the flow rate is decreased. Fan peak efficiency is slightly above 87 percent. However, reducing the flow rate only a small amount resulted in stall. The stall margin is less than 7 percent. It was also necessary to remove a small amount of flow from the stator corner to achieve this performance and stall margin.

A redesigned stator with 50 vanes, an aspect ratio slightly below 3, and a slight change in wall curvature at the hub was also tested (fig. II-25). The performance for the redesigned stator is shown as the solid lines in figure II-26. Pressure ratio and efficiency levels are about the same as for the

previous design. However, the flow cc.uld be reduced a good deal more be-

fore fan stall was observed. And in this case, slit suction from the stator hub was not necessary. The stall margin has been increased 9 percent by this stator redesign. In this fan the stator must operate with high aerody- namic loading and is responsible for stall. Small changes in vane number and aspect ratio and the local wall curvature had a relatively large effect on stall margin. However, even the reduced number of stators is over twice the number of rotor blades.

The increase in stator chord as the number is decreased also is bene- ficial in providing a structurally desirable stator. Even though stator vanes are supported at both the inner and outer walls, midspan dampers could be required to avoid serious wane flutter, particularly if a very large number of short-chord stators were considered necessary from a fan noise standpoint.

The combination of the aerodynamic and acoustic design requirements forces the fan towards the aerodynamic loading limits on the stator as well as pushing the design towards the aeromechanical limits.

Fan Tip Speed The fourth parameter to be discussed is fan tip speed. From the stand- point of the unsuppressed fan component, the data shown previously indicated the lowest noise fans are generally the subsonic fans. However, the best untreated fan component does not imply the best low-noise engine system.

For example, use of low fan speeds introduces the need for more stages on the low-pressure spool - leading to larger, heavier, and costlier engines.

A portion of the weight and cost advantage of a high-speed engine can be used to introduce more sound-absorbing panels to reduce the noise measured by the observer to the same level as that of the subsonic fan. Consequently, the best low-noise propulsion system cannot be a priori assumed to be the lowest noise untreated fan component. The choice can be made only after a total system evaluation is completed.

The effects of tip speed on noise were evaluated in the Quiet Engine Pro- gram by designing a fan with a 1550-foot-per-second tip speed. The acoustic comparisons are shown in figure II-27. Tip speed effects are obtained by comparing fans B and C - each with 26 blades. The fan noise levels are of concern essentially at two points - approach and takeoff. For approach power settings, the fan operates at about 60-percent speed in the region of 1.2 pressure ratio. The noise levels at approach are essentially identical.

The takeoff power settings are at the highest pressure ratio points shown.

The difference in maximum perceived noise level at this condition between B C is about fans and 3_ PNdB. A portion of this difference is due to the fact that fan B is operating at a lower pressure ratio. The acoustic chal- lenge is essentially that of reducing the noise penalty due to tip speed at this takeoff point.

Of particular interest on the high-speed fan is the trend at takeoff of a reduction in maximum perceived noise level as speed and pressure ratio are increased - which is contrary to the usual correlations. The maximum noise is associated with a partial-speed, or off-design, operating condition.

In fact, even at takeoff the fan is operating at an aerodynamic off-design - usually at about 90 percent of the aerodynamic design speed. A natural hy- pothesis is that the noise-generating phenomena are related to the aerody- namic flow fields illustrated in figure II-28. The design condition has the shock at the entrance to the blade channel with a relatively weak bow shock propagating upstream. At off-design, the passage shock is forced out of the channel, resulting in a strong shock pattern propagating upstream of the rotor. The reduction in noise as the design speed is approached occurs concurrently with the reduction in strength of the upstream shock pattern.

To explore the potential of reducing the noise associated with supersonic operation of the blading, a modification was made to a scale-model version of fan C. The blade channel area was opened to permit the shock to move back toward the channel entrance at the takeoff speed and reduce the strength of the upstream propagating wave patterns.

The acoustic impact of this modification is shown in figure II-29. The basic fan C characteristic shows the strong multiple pure tones (MPT) in the 500- to 1500-hertz region. The modified blade substantially reduced the MPT noise, by as much as 10 decibels over the entire MPT spectrum. How- ever, the blade passing tone was increased by about 5 decibels.

The acoustic energy obtained by integrating around the forward angles of the fan showed an 8-decibel reduction, or a reduction to about 20 percent of its original value. When the blade passing tone (BPT) is adjusted for its greater annoyance, the subjective evaluation decreases this to a 3-PNdB re- duction.

Associated with the increase in blade channel area was an increase in the effective blade camber. The effect on aerodynamic performance of this change was to raise the pressure ratio at which peak efficiency occurs to a region well above the engine operating line. Although the peak efficiency levels were actually increased at all speeds, the efficiency on the engine operating line was improved only to 90 percent - or takeoff speed. Above this speed, and in particular at the critical altitude cruise, an unacceptable efficiency penalty of about 4 percent was observed.

Studies of blade shapes particularly suited to minimize the aerodynamic penalties and operated with a weak forward shock pattern have been initiated.

These blade shapes would be particulary useful where relatively high blade speeds are utilized to achieve a rather low pressure ratio, that is, at low aerodynamic loading. Blade shapes with strong passage shocks have been repeated in figure II-30 along with blades designed to achieve a weaker for- ward shock configuration. In this case the blade is shaped so that the first passage shock is a weak oblique shock. This is followed by a second oblique shock near the blade trailing edge when the first shock is essentially at- tached to the blade leading edge and the forward shock configuration is very weak.

Blades for this shock configuration were designed into a fan with a pres- sure ratio of 1.5 at 1600 feet per second. This is a somewhat higher speed than used in fan C. The initial model of this fan resulted in a design speed efficiency of nearly 85 percent, as shown in figure II-31. The peak effi- ciency region and operating line would be well away from the stall line. The measured stall margin was at least 17 percent. It would appear the low blade losses were achieved by this blade design. However, the rotor had blade corner vibrations. This indicates than in addition to aerodynamic and noise considerations the new fan blade designs must also consider the aero- elastic effects.

The noise generated by this fan was not evaluated, but the aerodynamic data obtained indicate that the desired shock configuration was obtained.

In summary, a number of designparameters are being evolvedwhich

impact on bothacoustic andaerodynamicperformance. This work has al-

ready led to significant reductions in noise levels for the newhigh-bypass-

ratio enginesnowbeing introducedinto service.

A few of these parameters havebeendiscussedin somedetail. These,

as well as others, needto be further exploited - not only to assure that much

quieter configurations are obtained,but also to obtain the datafor engine

tradeoff studies. Thesestudiesare necessaryto permit selection of the

total systems whichachieve the desired noise suppressionat a mininum eco-

nomic penalty.

PERFORMANCE RESULTS WITH LARGE-SCALE FAN FACILITY A quiet fan is needed in order to have a quiet engine. As discussed previously, the fan must meet certain aerodynamic requirements. How- ever, it is also a noise generator. Although much theoretical acoustic analysis has been done, fan designs are heavily dependent on empirical data.

Consider now what an inefficient noise generator a quiet fan really is.

Shown in figure II-32 are data points for various low-speed and high-speed fans that have been tested. The percentage of fan shaft horsepower that radiates as noise, or sound power, is plotted against fan pressure ratio.

For the lowest noise fans that are of interest, less than 0. 005 percent of the shaft horsepower radiates as noise. This means that for a 20 000- horsepower fan, less than 1 horsepower radiates as noise.

In order to obtain the necessary large-scale fan acoustic data for em- pirical correlations and extrapolation, NASA built the test facility shown in figure II-33. Initially, the research fan was located 50 feet from the building wall and 19 feet above grade to minimize ground reflections and the effects of the ground on inlet flow. The 10- by 10-Foot Supersonic Wind Tunnel drive motors were utilized to power the fan. Fortunately, when the tunnel drive system was designed, the motors were double-ended, to provide for a time when another facility might be added. This resulted in a costs saving by eliminating the need for drive motors in the fan test facility.

The motors in the building drive the fan through a speedincreaser and

a 50-foot-long drive shaft. A portion of the building wall wastreated with

6 inches of polyurethanefoam to minimize acoustic reflections that the mi-

crophonesmight pick up. An openarea as free of reflecting surfacesas

possible was chosenfor the test site, as shownin figure II-34.

The microphoneswere placedat 10 ° intervals, generally, on 100-foot

radii centered on the fan. The forward microphoneswere located closer to

the fan to clear the drive motor building, andthe data from these micro-

phonesare corrected to a 100-footradius. The microphonesare at 19-foot

elevation: the sameelevationas the fan centerline. The fan inlet (fig.

II-35) is free of obstructions suchas bearings, struts, guidevanes, and

instrumentation, in order to minimize inlet flow distortion. However,

someof the inlet air flows over the supportstructure andcausessome

inlet flow distortion.

Inlet flow distortion causesfluctuating forces onthe rotor bladesas

they pass throughthe distorted flow region. This, in turn, results in an in-

crease in the noise generatedby the fan, which manifestsitself primarily in

the bladepassagetone (BPT). A narrow-band spectrumis shownin figure

II-36 to illustrate the effect of inlet flow distortion, in terms of BPT, onthe

soundpressure level. The maximumsoundpressure level of the discrete

toneat blade passagefrequencyfor takeoff speedis very pronounced. An

analysis of the noise revealedthat the discrete tone was higher thanex-

pected.

A decision was madeto movethe fan farther from the building to pro-

vide relief from the inlet flow problem. The shaft was extendedso that the

fan was 100feet from the building. The fan was turned aroundso that the

inlet extendedwell beyondthe support structure in order to obtain an unob-

structed, or clean, inlet (fig. II-37). The fan was nowdriven from the rear

andflow wasdischargedtoward the building wall. The clean inlet with less

inlet flow distortion resulted in a considerablereduction in the blade pas-

sagetone (fig. 11-38). However, the low-frequency broadbandnoise was

greatly increased. This was causedby the dischargeairflow scrubbingover

the bearingsand shaft supportsandbeingturned by the wall of the drive

building. This low-frequencynoise was unacceptable;and since it was im-

practical to movethe fan farther from the building wall andthere wasno

readily apparent way to reduce the scrubbing noise, it was necessary to ac- cept the best compromise available. Consequently, the fan was turned around to the final version - 100 feet from the building wall and discharging away from the building wall, as shown in figure II-39. The inlet obstruc- tions were also minimized. In order to do this, the pedestal was separated to allow undisturbed flow from below and the support structure was stream- lined (fig. II-40). This reduced the broadband noise, as shown in figure II-41. The blade passage tone is not as low as desired. However, it is the best that can reasonably be done with the facility and no further changes are contemplated.

The family of fans described early in this paper has been tested in sup- port of the Quiet Engine Program. Some of the more interesting aspects of the basic data that have been obtained during the evaluation and selection process of these fans are discussed in the following paragraphs.

In figure II-42, some of the data from fan D are shown in terms of narrow-band spectra. The data are shown as sound pressure level as a function of frequency at 50 ° from the inlet at takeoff and approach speeds.

Illustrated in the figure are broadband noise, some evidence of multiple pure tones (MPT), and blade passage tones (BPT) with harmonics.

In figure II-43, the same fan D data are shown in terms of 1/3-octave frequency bands instead of narrow bands. At the takeoff speed (square sym- bols) the broadband, MPT, and BPT are apparent. The overall power levels can be integrated and the noise components separated as a function of speed.

Total noise (fig. II-44) is a summation of the components of broadband, BPT, and MPT. The total noise increases with increasing speed; and MPT comes in at a tip speed of just below 900 feet per second and climbs fast, while the BPT peaks and drops off. This represents an interchange of ener- gy from BPT to MPT with increasing" speed.

To obtain subjective noise ratings of fans, 1/3-octave plots are used and weighted levels of human response in the various frequency bands are fac- tored in to get the perceived noise level expressed in terms of PNdB. For fan D this resulted in a maximum level of 95.5 PNdB at takeoff speed at a 1000-foot altitude and at 50 ° from the inlet (fig. II-43).

This procedure can be applied for every angle with a resultant plot of perceived noise as a function of angular position, as shown in figure II-45 for fan D. The noise reaches a maximum at 50 ° from the inlet and then peaks again at 110 °. The level for fan D is essentially the same for front and rear.

From data such as have been presented, the various types of subjective noise ratings which are used to evaluate the relative quietness of engines can be calculated. Three of these are perceived noise level (PNL); tone- corrected PNL, or PNLT, which penalizes the engine for excessive BPT; and effective perceived noise level (EPNL), which takes into account the duration of the noise as the plane passes overhead. EPNL is the value used in FAR-36. The results of these calculations are shown in table II-2. The TABLE II-2. - FAN D NOISE RATING [Pressure ratio, I. 4; tip speed, 1107 ft/sec. ] Takeoff Noise rating Approach power power 104.4 107.1 Maximum perceived noise level, PNdB Maximum tone-corrected 105.1 108.6 perceived noise level, PNdB 102.6 103.1 Effective perceived noise level, I)NdB values in the table are the maximum noise that a person on the ground would hear during takeoff of a four-engine airplane 3.5 nautical miles from brake release and, for the same airplane, on approach 1 nautical mile from touch- down.

On the basis of the basic data for all the fans, one was selected for the Quiet Engine. A plot of the data for the low-speed, single-stage fans is shown in figure II-46 in terms of maximum noise levels as a function of fan pressure ratio. It is apparent that fan A is the lowest noise fan. Fortu- nately, fan A is also the best from aerodynamic and mechanical standpoints.

When the single-stage, high-speed fan data are included, it is apparent that the noise level for the supersonic fan is considerably higher than that for low-speed fans (fig. II-47). However, this may not be as bad as it first appears. (It is discussed in another paper.)

Another way to look at the difference between low- and high-speed fans is how the noise varies with angle from the inlet. Figure II-48 shows a noise directivity plot for fans A and C similar to that previously shown for fan D. The noise level peaks at 50 °, falls off, and then reaches another peak at 120 ° (similar to fan D). Both front and rear peaks are essentially equal. High-speed fan C differs from low-speed fan A primarily in the higher front-end noise level.

Examination of narrow-band spectra of sound pressure level as a func- tion of frequency at 50 ° from the inlet for takeoff (fig. II-49) shows that fan A is similar to fan D with broadband, BPT, and harmonics noise evident.

Fan C, however, is dominated by MPT, which masks the BPT. This phe- nomenon is caused by nonuniform shocks emanating from the leading edge of the rotor blades at high speeds.

CONCLUDINGREMARKS The fact that there are tradeoffs between the aerodynamics and acous- tics in the fans is not surprising. The design of every engine component in- volves some compromise. The basic knowledge of fan noise generation is not sufficiently precise to permit quantitative estimates or predictions of fan noise. Consequently, a largely empirical approach is used. The facilities which are used to obtain fan noise empirical data have been discussed and some of the results shown.

The empirical work can be summarized through the use of several charts, the first of which (fig. II-50(a)) shows the data that have been ob- tained on several single-stage, low-speed fans. The maximum perceived noise level (in PNdB) is shown plotted against fan pressure ratio for fans of

Quiet Enginesize which wouldproducea total of 90 000poundsof takeoff

thrust (four engines). The measurednoise is extrapolatedto a simulated

1000-footflyover.

The data group together well for the single-stage, low-speedfans which

cover the pressure ratio range of 1.4 to 1.6. Before deciding ona single-

stagefan for the Quiet Engine, two-stage-fandata were also examined{fig.

II-50(b)). Thesedata include, on the extreme left, that from a modified

TF-39 enginewhich was run as part of the Quiet EngineProgram andalso

two-stagefan data collectedfrom various low-bypass-ratio enginesnowin

commonservice onnarrow-bodies jet transports. The datafrom these op-

erational, two-stage, low-bypassratio engineswere obtainedwith fans that

were runningat fairly high speedsandhadclosely spacedinlet guide vanes,

as opposed to the datapoint on the extreme left, which represents a widely

spaced,two-stagefan operatingat low speed. Consequently,the correla-

tion bandaroundthesedata points probably is to a great extentfortuitous,

althoughthe fans havesomerelation to eachother whenplottedagainst

pressure ratio.

In figure H-50(c) are shownsomedata points for single-stage, high-

speedfans andagainthesedata points tend to group together. The data are

somewhat higher than for the single-stage, low-speedfans andsomewhat

lower than for the two-stagefans.

The collection of data points in figure II-50(c/ represents the basis from

which to work in the Quiet EngineProgram for subsoniccruise, CTOL air-

planes. Other applications of propulsion systemswith low-noise technology

built into them are of interest andare being examined. In order to do that

effectively, our knowledgehasto be extendedin two directions: to the

higher pressure ratios andthe lower pressure ratios. In figure II-50(d),

someopensquareshavebeenaddedwhich represent plannedtests of single-

stage, high-speedfans. Obviously, their actual noise outputis estimated

becausethe tests are just nowbeing planned;however, the hardwareis be-

ing built. The noise estimates shownare basedon empirical correlations so

that it shouldcomeas no surprise that they fall right in the databand.

Work in the low-pressure-ratio range is also beingextendedandis

shownin figure II-50(e) by the opensquares. Thesefans are designedfor

various pressure ratios that are a part of the technologyprogram in support

of externally blownflap STOL airplanes. Again, the noise levels are esti-

mated and may be different from those shown in the figure.

In general, from the data shown in figure II-50(e) for 90 000 pounds of thrust (four engines), a variety of fans can be selected which produce noise levels in the range of 100 to 120 PNdB. Selection of a particular fan depends on the fan configuration and the design pressure ratio required for a specific operation and that, in part, depends on the type of mission to be performed.

CTOL LONG RANGE CRUISE 2.5 _ ,,_\\\\\, Mc

2.o ",,,,2\\_ o.95

FAN PRESSURE RATIO I. 5 "w, litflfsllfiif//i/fltlll .

1.o I l

.9-- S PECIFIC FUEL .7 CONSUMPTION iiil/ll//lll/////ll// i/i/iyli ._0

I

I I

.5 4 8 12 CS-63060 BYPASS RATIO Figure II-1 FAN ENGINE SIZE _ _ FAN PRESSURE RATIO I--

i I I I i I

0, i000 1400 1800 JET VELOCITY, FTISEC CS-63049 Figure 11-2 FAN PERFORMANCE MAP STALL--....

LINE _A _OPERATING FAN PRESSURE RATIO _CRUISE KEOFF WEIGHTFLOW CS-63050 Figure II-3 DISTORTION PERFORMANCE STALL /-- ACCOUNTSFOR _'X_" DISTORTION /_/"'- OPERATING LINE FAN PRESSURE RATIO "----CRUISE \ -TAKEOFF WEIGHTFLOW CS-63051 Figure II-4 FAN PRESSURE RATIO AERODYNAMIC LOADING PARAMETER 2.0!

2.2 FAN 1.8 PRESSURE RATIO 1.4 I I I I I.C 500 1000 1500 2000 2500 ROTORBLADETIP SPEED,FTISEC I I I ,,I .8 1.2 1.6 2.0 CS-63134 RELATIVEMACH NUM BER Figure II-5 COMPARATIVE PERFORMANCE OF LOW-NOISE FANS DESIGN SPEED EFFICIENCY ,-A rC

• i i _ i q_

AERODYNAMICLOADING 2.6 -- PARAMETER//

S/\ o

FAN PRESSURE 1.1_-,-,:':'.':_.__ _C_ 3 RATIO 1.0 ] I t I J 800 I000 1200 1400 1600 1800 ROTOR BLADE TIP SPEED, FTISEC C5-63132 Figure II-6 FAN STALL MARGIN r--i EXPERIMENTAL STAGES QUIET ENGINE PROGRAM FANS LOW SPEED, HIGHLY LOADED STAGES STALL 20-- MARGIN,

H

1000 1200 1400 1600 ROTOR BLADE TIP SPEED, FT/SEC Figure II-7 CS-63053 MAJOR POWERPLANT NOISE SOURCES

f

TURBINE

f

FAN FAN INLET& DISCHARGE COMPRESSOR FAN & PRIMARY JETMIXING FAN DISCHARGE FAN INLET _ NOISE LEVEL PRESSOR " CS-63394 APPROACHING--"OVERHEAD "" DEPARTING Figure 1I-8 FAN NOISE SPECTRA AT SUB- AND SUPERSONIC TIP SPEEDS SUBSONIC

I _ ._BLADE PASSAGE FREQUENCY

SOUND PRESSURE LEVEL SUPERSONICf MULTIPLE PURETONES dB I I FREQUENCY iAi _RoADBAND NOISE T, 0 2 4 6 8 I0 FREQUENCY, kHz cs-63zz3 Figu re II-9 PHASES IN EMISSION OF FAN NOISE GENERATION ON BLADE AND VANE SURFACES --_...

"_..'-- _.

_. -- -.

/_f" _ " RADIATION PRODUCES 1RECTIVITYPATTERN _- FAR FIELDDIRECTIVITYPATTERN cs-_3393 Figure II-I0 FAN NOISE RESEARCH FACILITIES CS-63212 Figure II-11 GENERATION OF DISCRETE BLADE PASSAGE NOISE BY PERIODIC WAKE CUTTING INFLOW ROTORBLADES_ .,_ _ _ I- BLADE I •o--'_-'._, "_._. __ , WAKES STATOR VANES _;.

VELOCITY DEFICITIN WAKES PRODUCES PERIODICALLY FLUCTUATING AIR ANGLE TO VANES INDUCING LIFT FLUCTUATIONS ON VANES FORMINGMOVING PRESSURE PATI'ERNS IN DUCT cs-_32_1 AND RADIATING TO FAR FIELD AS NOISE Figure II-12

BROADBAND NOISEGENERATION DUE TO RANDOM

BLADE LOADVARIATIONS

TURBULENCE ININLET AIRFLOWF_;./,_.e _,, _ _., _, _,c, __ ,-*_,,.,,-_

& WALL BOUNDARY LAYERS I L ,_'i ~ _" _' ~,7_" Z_'_-'_,;_._- _._- "_ ,v

,.pr

AvE

LADE I_//-_%// BOUNDARY

o,o

_Jj//I// LAYER IF ee'_"," • ,s,J_ .r" ...... BLADE WAKE / I JJ sSSTALL_ / L'VANE & BLADE STATORVANES CS-63230 Figure II-13 MULTIPLE PURE TONE NOISE AT SUPERSONIC TIP SPEEDS-IDEALIZED WAVE PATTERN INCOMING AIRFLOW ..... I ,.i MICROPHONE PI _k_k_.l DISCRETE NOISE -_" .... '_ _ - - -TIME WAVEFORM _ACH WA_ CS-6323Z FAN BLADETIPS, DEVELOPED VIEW TIP VELOCITY > = 1100FTISEC Figure II-14 MULTIPLE PURE TONE NOISE AT SUPERSONIC TIP SPEEDS- ACTUAL WAVE PATTERN INCOMING AIRFLOW L ,. MICROPHONE P I_ MULTIPLE PURE TONE

/

TIP VELOCITY > z 1100 FTI SEC FANBLADE TIPS, DEVELOPED VIEW CS-63ZZ9 Figure I1-15 SPACING ON FAN EXIT GUIDE VANE EFFECT OF FAN DISCRETE NOISE CHANGE IN NOISE LEVEL, A dB I I 1 I t ,..

4O 8O 120 160 2OO ROTOR-VANE SPACING IN PERCENT OFROTOR BLADE AXIALCHORD LENGTH CS-632l 3 Figure 11-16 COMPARISON OF FAN GEOMETRY A--HIGH TIP SPEED INLET_ _.

Y

TWO-STAGE LOW BYPAS S RATIO LOW TIP SPEED /- INCREASED SPACING SINGLE-STAGE FAN .-d--- ACOUSTIC 1 TREATMENT NO INLET GUIDE VANES LOWER JET VELOCITY CORRECT NUM,E_/ OF GUIDE VANES SINGLE-STAGEHIGH BYPASS RATIO Figure II-17 FAN NOISE PREDICTION SEQUENCES FORWARD RADIATED BLADE PASSAGE TONE NOISE DESIGN AND OPERATING FEATURES

/ i l

"LAOEWAKES I r _LADE PO_NT,AL 15TATOR POTENT,AL l

'NTO STATO. I LE.ELD O_ STATOR I IF,ELOON ROTOR

F FLUCTOAT,NG } FLOCTUAT,NG

.l

| STATOR LIFT j ROTOR LIFT

IFORMULATE AND RUN1

MODEL TESTS j

l 1 1 IN DUCT IN DUCT

TRANSMISSION THRU 1 BLADE ROW I L__ -(, f_ MODES IN INLET I PROPAGATION OF

IRAO'A''ON OF NO'SE1

TO FAR FIELD 1 COMPUTE BLADE PASSAGE PREDICT BLADE PASSAGE NOISE AT RECEIVER NOISE AT RECEIVER Figure II-18 SPACING EFFECTS ON NOISE PNLT ¢____EOFF PROACH WEIGHT} BOOSTSTAGESMINIMIZE LENGTH PENALTY I i I J I 2 SPACING, ROTOR CHORDS CS-63136 Figure II-19 LOW NOISE FANS- STATOR SPACING DESIGN 3.5 CHORDS 1 CHORD CS-63057 Figure II-20 SPACING EFFECTS ON EFFICIENCY .86-- .84 EFFICIENCY . 82 .80

• 78 J I t

SPACING, ROTOR CHORDS Figure II-2] CS-63047 ROTOR BLADE NUMBERS-FAN A ADVANTAGES LIGHTER HIGHER EFFICIENCY SHORTER LOWER NOISE CS-63056 Figure II-22

ROTORBLADE NUMBERS- FAN B

ADVANTAGES FEWER PARTS HIGHER STALL MARGIN CS-63048 Figure II-23 VANE/BLADE RATIO EFFECTS I LSI6

TT

INCREASED NUMBER OF 5dB _ STATOR VANES T ATOR VANES e REDUCES STALL PNLT • J O 76 MARGIN ,I_ l I I 12]1108 e INCREASESVANE I I VIBRATION 60 70 80 90 I00 PERCENT DESIGN SPEED CS-63137 Figu re II-24 4'7 DESIGN OF LOW-NOISE FANS-NUMBER OF STATOR VANES CS-63055 Figu re II-25 AERODYNAMIC PERFORMANCE OF LOW NOISE FANS-NUMBER OF STATOR VANES STATOR VANES EFFICIENCY -_ 50 ?, .8 _ I ASM : 9_ PRESSURE RATIO

I

160 170 180 WT FLOW, LBISFC CS=63054 Figure II-26 TIP SPEED EFFECTS FAN CON1PONENT FAN (NUMBER II0-- OF BLADES) C126) 100--

1000 FT

SIDELINE

MAX PNL,

dB

gC- A 1401 / ,,aO¢" B 126) I I I I J 1.2 1.3 1.4 1.5 1.6 I.I FAN PRESSURE RATIO CS-63131 Figure II-2?

SPEED SHOCK PATTERNS HIGH OFF DESIGN DESIGN

/

F LOW ROTATION ROTATION CS-63130 Figure II-28 BLADE GEOMETRY EFFECTS HIGH SPEED FAN C dB PNdB, 200FT, dB I BLADE FWD SPL, MAX. FWD [] C 161 123 O C MOD. 153 120 200 FT SIDELINE SPL, 110 "_ dB 7O I I I 1 I I I 50 I00 500 I000 5000 I0000 FREQUENCY, Hz CS-63135 Figure II-29 HIGH SPEED BLADE SHAPES HIGHPRESSURERATIO LOW PRESSURERATIO

/ Jw

FLOW ROTATION ROTATION CS-C_31Zcl Figure II-30 FAN PERFORMANCE DESIGN SPEED, 1600 FTISEC .9 EFFICIENCY .7 I 1.7-- _-_PERATING LINE 1.6-- FAN / 1.5-- PRESSURE RATIO 1.4--

I I

1.3 130 140 1.50 160 WT FLOW, LB/SEC cs-_,sos_ Figure11-31 FAN SOUND POWER GENERATION .O_ - •015 •011] -- SOUNDPOWER FAN SHAFTPOWER' %

°° o; FANA

i

1 1 1 I 1

1.6 1.1 1.2 1.3 1.4 1.5 FAN PRESSURE RATIO CS-_3228 Figure II-32 FAN NOISE TEST FACILITY SHORT SHAFT- FRONT DRIVE Figure II-33 C5-()321!

FAN NOISE TEST AREA Figure II-_ FAN INLET Figure 11-35 NARROW BAND SPECTRUM TAKEOFF SPEED

Qo° 110 --

105-- I _r SHORT SHAFT- PRESSURE lEVEL, dB 95 -- 8_ I I - 45O0 500 1500 2500 3500 FREQUENCY,Hz CS-63226 Figure II-36

Figure I]-37

COMPARISON OF NARROW BAND SPECTRA TAKEOFF SPEED !

il FSHORT SHAFT-

.oF-<?oo

I_ FRONTDRIVE SOUND 100l--.- PRESSURE _VEL, dB 9.5 __, REARDRIVEij

9o F. ,,",.,__

s5 1 I I 500 1500 2500 35OO 4.5OO FREQUENCY,Hz CS-63225 Figure II-38 Figure II-39

FAN SUPPORT PEDESTAL

Figure II-40 NARROW BAND SPECTRA WITH SEPARATED PEDESTAL TAKEOFF SPEED 1lO- af-SHORT SHAFT- I_' FRONTDRIVE 105-- _ _LONG SHAFT- SOUND 10£- FRONTDRIVE PRESSURE , Ill _VEL, dB 95 ..... /I _REAR DRIVE

\ ,.,,"1

90_ , ...-I----.__ l 85t v''- I _1 I 5OO 1500 25O0 35OO FREQUENCY,Hz cs-63zz4 Figure II-41 FAN D NARROW BAND SPECTRA 1.4 PRESSURE RATIO; 1107 FT/SECTIP SPEED IIC- loo 9C 8O SOUND TAKEOFF SPEED PRESSURE 70 -- _ -____J LEVEL, dB 100 r- APPROACH SPEED 8O 7O 0 2000 4000 6000 8000 I0000 FREQUENCY,Hz CS-63227 Figure 11-42 FAN D ONE-THIRD OCTAVE SPECTRA 1.4 PRESSURERATIO; 1107 FT/SEC TIP SPEED; PNL : 95. 5 PNdB; I000 FT FLYOVER 105 - _ _o_ 95- SOUND PRESSURE LEVEL, dB 65 _ I I I I I I I 50 IOO 200 500 I000 2000 5000 I0 O0O20 000 FREQUENCY,Hz C,S- _', 32Z ] Figure II-43 FAN D NOISE COMPONENTS 1.4 PRESSURE RATIO; 1107 FTISEC TIP SPEED ,,f.,>.--o TOTAL 155-- _J_BROADBAND 150 /P'y,,/P'"_ BLADE-PASSAGE SOUND 145 _ TONE POWER 140-- LEVEL, dB _MULTIPLE PURE 135130-- -- /TONE

123 I I I I

600 700 800 900 i000 1100 ROTORTIP SPEED,FTISEC cs-63z14 Figure II-44 5'1

FAN D PERCEIVED NOISELEVEL AT TAKEOFF

1.4 PRESSURERATIO; 1107R/SEC TIP SPEED; 1000_ FLYOVER 80--

NOISE

LEVEL,

PNclB

lO-- 60--

I 1 I I

40 80 120 160 ANGLEFROMINLET, DEG CS-i, _.2 £ 5 Figure 11-45 COMPARISON OF NOISE LEVELS- LOW SPEED FANS 1000FTFLYOVER 110 -- MAXIMUM 10(] A NOISE LEVEL, PNdB

_

I I i I I

8O 1.0 1.2 1.4 1.6 FAN PRESSURE RATIO CS-63218 Figure 11-46 HIGH-SPEED FANS AND NOISE LEVELS- LOW-

COMPARISON OF

]000FT FLYOVER 110-- HIGH-SPED FAN-'_Y'_ C MAXIMUM NOILEvEL PNdB SE 10090,! _ E B__

I

80 i i i I I

1.6 1.0 1.2 1.4 CS-63ZZO FAN PRESSURE RATIO Fiqure II-41 FAN PERCEIVED NOISE DIRECTIVITY 1000FTALTITUDE II0 F- FAN NOISE LEVEL, PNdB 70-- INIFT -_EXHAUST

60 I I I I I I I I I

0 20 40 60 80 I00 120 140 160 ANGIF FROM INLET, DEG CS-63216 Figure II-48 NARROW-BAND SPECTRA AT TAKEOFF SPEED SOUND I I I I PRESSURE 110 -- _VEL, dB 70 I I I I I 0 2000 4000 6000 8000 I0 000 FREQUENCY,Hz CS-63233 Figure 11-49 FAN NOISE 90 000 POUND TAKEOFF THRUST; 1000 FOOT FLYOVER 125 - MAXIMUM NOISE LEVEL, ,NGLE- STAGE . _'i_!,'. SINGLE-STAGE, PNdB 'i. LOW-SPEED FANS LOW-SPEED FANS I I I I I I I I I 1 I I (b) (a) CS-63397 CSo63399 125 -- IWO- TWO STAGE FANS FANS 115 -- SINGLE-STAGE SINGLE-STAGE AAXIMUM HIGH-SPEED HIGH-SPEED FANS NOISE FANS SINGLE-STAGE LEVEL, SINGLE-STAGE LOW-SPEED PNdB LOW-SPEED FANS FANS

1°, I

I I I I I I I 1 I I I I I 1.0 1.4 1.6 1.8 2.0 2.2 2.4 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 FAN PRESSURE RATIO cs-_>_39_ FAN PRESSURE RATIO cs-6339;, (d) (c) 125 -- }o TWO-STAGE FANS SINGIF- STAGE HIGH-SPEED FANS MAXIMUM SINGLE-STAGE NOISE LOW-SPEED LEVEL, FANS PNdB I I I I I I I 0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 FAN PRESSURE RATIO C5-fi3398 (el Figure I1-90 III. FAN NOISE SUPPRESSION Charles E. Feiler, John F. Groeneweg, Edward J. Rice, Edward B. Smith,* and Roger H. Tucker** A significant reduction in noise can be achieved, even for low-noise fans, by the use of acoustic suppression. Acoustic suppression may be at- tained by using sonic inlets or acoustically treated surfaces, which are discussed in this paper. Acoustically treated surfaces suppress noise by converting acoustic energy to heat through friction and turbulent dis- sipation.

Suppression requirements and the limits to suppression are treated first. Then the important suppressor parameters and how they influence suppressor design are shown by drawing on theoretical studies and labora- tory experiments. Finally, data from full-scale fan suppressors are pre- sented, and estimates of the performance penalties associated with suppres- sion are made.

The internal noise sources of a turbofan engine are shown in figure III-1. It is important to recognize that acoustic linings (suppressors) can affect only noise originating within the engine and not that originating outside the engine. The important sources to consider are the fan, the turbine, and the compressor. Noise from these rotating elements propagates out the en- gine inlet and the fan and core exhaust nozzles. The relative level of noise from these sources in an unsuppressed engine is shown in figure III-2. The exact relation of these levels will depend on the particular engine cycle; but, in general, fan noise is the dominant source for turbofan engines and, as shown in this figure, it determines the total engine noise. At an appreciably lower level is the core engine noise, dominated by the turbine noise, and then perhaps at a still lower level is the external jet noise.

The same engine with internal noise suppression might appear as shown in figure III-3. The inlet has treatment on three splitter rings and on the *General Electric Company.

**The Boeing Company.

outer cowl wall andan extendedcenterbody. The fan exhaust duct has treat-

ment on the inner and outer cowl walls and on a single splitter ring. Simi- larly, treatment is located in the inlet of the compressor and in the core ex- haust duct.

With this suppression added, the relative noise levels might be as shown in figure III-4. The fan and core noise levels are shown reduced to the jet noise level, and the total noise is therefore equally determined by the three sources shown. It is important to recognize here that the total noise can be reduced about 5 PNdB more by further reducing fan and core engine noise.

At that point the total noise is solely determined by jet noise and no benefits can be obtained by more acoustic suppression. The jet noise thus deter- mines, or limits, the amount of suppression that can usefully be employed.

It is a noise floor that is ultimately reached. In practice, there are other noise sources or noise paths that may set floors to limit the useful suppres- sion at levels above the jet noise floor. For example, a possible floor might be set by the scrubbing noise of the high-velocity air over the sup- pressor and cowl surfaces, or by radiation of noise through the engine casing. In any case, it is important to note that there is a finite amount of suppression to be achieved by acoustic treatment.

NOISE SPECTRUM CHARACTERISTICS In order to determine the suppression design requirements, the fan noise spectrum characteristics must be known. These are examined for a low-speed fan in figure III-5. At a far-field measurement position the typi- cal 1/3-octave-band sound pressure level spectrum shown in this figure is composed of fan noise and jet noise. The fan noise consists of a broadband spectrum with superimposed tones which are equal to the blade passi_Jg fre- quency (and multiples) of the fan rotor. Fan noise suppression can be at- tained in the frequency range from about 400 hertz to 10 kilohertz since the fan noise is higher than the jet noise in this range.

In figure III-6, the far-field sound pressure level spectrum produced by a high-speed fan is compared to the previously shown low-speed fan noise spectrum. This spectrum is typical for high-speed fans at the "takeoff" power setting in the front quadrant of the fan. At low power settings and in the aft quadrant at takeoff power, the spectra produced by the high- and low- speed fans are very similar. The noise component added by the high-speed rotor inthe forward quadrant ranges from 100 hertz to the blade passing frequency fundamental and is the result of a series of "tones" which are multiples of the rotor shaft rotational frequency. These tones, commonly referred to as multiple pure tones (MPT) or "buzz saw" noise, are an addi- tional fan noise element requiring suppression. The suppression of this noise is discussed in the section FULL-SCALE FAN SUPPRESSION TESTS.

The primary concern of noise suppression is to reduce the annoyance of fan noise. The typical forward-radiated spectra are plotted in figure III-7 as the annoyance in NOY units as a function of frequency. The NOY weight- ing tends to emphasize the energy in the 3150- to 4000-hertz range, which consequently emphasizes this frequency in terms of required suppression.

The weighted jet noise spectrum is now a very flat "floor."

Required attenuation spectra are plotted in figure III-8. These were obtained by determining the sound attenuation necessary at each frequency to bring the attenuated spectra down to a constant annoyance level. This annoyance level was set by the peak in the weighted jet noise spectrum shown in figure III-7. Again, the larger low-frequency fan noise produced by the high-speed fan is peculiar to the forward quadrant at "takeoff" power set- tings only. The conclusion from this figure is that the acoustic treatment designed and placed both in the fan inlet and the exhaust must cover a very wide frequency range and produce peak suppressions in far-field sound pres- sure levels close to 30 decibels.

In order to accomplish maximum suppression in terms of APNdB, three important acoustic treatment design parameters must be considered.

These are shown in figure III-9. The level of peak sound pressure level attenuation is important and is determined by the maximum required atten- uation of curves such as those shown in figure III-8. The frequency where peak suppression is obtained is important since the maximum contribution to the NOY-weighted spectrum usually require s the maximum attenuation (fig. III-7). Finally, the bandwidth of the attenuation spectrum should be made as wide as possible in order to attenuate effectively the very wide fan spectrum shown in figure III-8. The important treatment design parameters which are used to design acoustic treatment which achieves these estab- lished fan noise suppression requirements are discussed in the next sec- tion.

SUPPRES SION PARAMETERS An enlarged view of a section of the exhaust suppressor of an engine is shown in figure III-10. Some of the more important parameters are shown: the duct height H is the distance between lined walls; L is the liner length.

The wall impedance is an important parameter which is treated in some de- tail later in this section. The environmental parameters include the steady- flow Mach number M and the incoming noise spectrum.

Sound propagation theory is available for such a lined duct, but the de- tails of this theory cannot be covered here. However, some of the more important results of the theory are presented. The theoretical parameters are listed at the bottom of the figure. These include the frequency param- eter fH/c in which f is the sound frequency, H is the height between the lined surfaces, and c is the speed of sound.

The frequency parameter can be related to the duct height in sound wave lengths H/_. Other important parameters are the treated-duct length- height ratio L/H, the steady-flow Mach number M, and the wall impedance.

The initial sound pressure profile at the duct entrance is a very strong pa- rameter in determining the sound attenuation since it determines the modal content within the duct. The various modes are attenuated at different rates, and the overall attenuation is thus governed by the modal content. Of all the possible modes, one is of particular interest since it damps slower than the rest. This is referred to as the least-attenuated mode. In fairly long ducts, this will be the only remaining mode and the attenuation will be determined primarily by this mode.

Another initial condition is a plane pressure wave at the lined duct en- trance. This will excite several modes within the lined duct.

Spinning modes were mentioned in the preceding paper. With these, several pressure maxima and minima occur around the circumference of the duct. This entire pressure pattern may be spinning around the duct axis.

This pressure pattern may be produced by the spinning rotor or by the rotor-stator interaction. In general, these modes will damp more rapidly than the axisymmetric modes.

The first result to be considered is the effect of the frequency param- eter on the peak sound attenuation (fig. III-11(a)). The ordinate of this fig- ure is the peak attenuation normalized by the duct length-height ratio. This normalization collapses all the various curves for different L/H ratios onto one curve. The abscissa is the frequency parameter fH/c. The peak attenuation is the maximum possible attenuation obtainable with a particular geometry (L and H) at a given sound frequency. This sound attenuation is obtained at a unique wall impedance called the optimum impedance, which is discussed later in this section.

In the typical attenuation spectrum presented in figure III-8, the peak attenuation and the frequency at which the peak occurs can be associated with the parameters shown here.

The important point in figure III-11(a) is that at high values of the fre- quency parameter only small attenuations can be obtained unless very large duct L/H values are used. To increase the attenuation, the duct height H must be made smaller, which leads to splitters in large ducts. A reduction in duct height decreases the frequency parameter and increases the atten- uation. A practical range of operation for a turbofan engine might be in the frequency parameter range of 0.75 to 2.

For a better picture of what the frequency parameter really means we can select several sample points along this curve and show some specific examples. This is done in figure III-11(b). For these sample inlets, the duct diameter is 6 feet, the frequency for peak attenuation is 2000 hertz, and a peak attenuation of 20 decibels is considered.

The frequency parameter with a single splitter ring is quite large due to the large distance between treated surfaces. To get the 20-decibel atten- uation, a large duct length must be used. If two splitter rings are used, the duct height and frequency parameter are decreased. A considerable reduc- tion in suppressor length is achieved over the single splitter case.

As three and four splitter rings are used, the length of treatment is progressively shortened; however, more and more of the inlet area is blocked with these splitter rings. The Mach number will be increased, and thus the losses increased, unless the outer cowl is expanded to accommo- date the flow.

As the treated length is shortened in these examples, more rings are required. It is interesting to see the relative area of suppression required for this variety of inlets. This is shown in figure III-12. In this figure, the relative suppressor area is plotted against the number of splitter rings. All areas are normalized by the area required with wall treatment only. As the number of splitter rings is increased, the total suppressor area is substan- tially reduced. Again it should be recognized that as the wetted area of sup- pression is reduced the inlet area blockage is increased and thus the Mach number is increased. Also as the number of splitter rings is increased, the inactive area is increased due to leading and trailing edges. In any real case the number of splitter rings should be selected so that the necessary attenuation is obtained with the minimum pressure loss.

In figure III-13, the effect of Mach number on peak attenuation is shown.

The zero Mach number curve is the same as on figure III-11(a); however, two additional curves are shown. The top curve is for a Mach number of -0.4. The negative sign implies that the sound is propagating in a direction opposing the steady flow. This is the case found in the fan inlet. The bot- tom curve is for a Mach number of 0.4. Here the sound propagates in the same direction as the steady flow. Positive Mach numbers thus imply the case found in the exhaust duct.

The most important point in figure III-13 is that the peak attenuation is a function of Mach number only for values of frequency parameter below about 1.5. Low values of frequency parameter, of course, imply very low frequencies or small duct heights. In this case the sound will attenuate faster in the inlet than in the exhaust duct.

For values of frequency parameter above i. 5, the peak attenuation is not a function of Mach number and the approximate equation shown in the figure can be used. The attenuation is proportional to duct length divided by duct height squared. This emphasizes the strong effect of duct height on at- tenuation. Doubling the length only doubles the attenuation, while decreas- ing the height by a factor of 2 increases the attenuation by a factor of 4.

In this section, theoretical results have been discussed to show trends and relationships between the several variables. The theory has been kept as simple as possible so as not to complicate the picture any more than necessary. In figure III-14, some data are compared with theoretical re- sults. The peak sound attenuation normalized by duct length-height ratio is plotted against frequency parameter. The limits of plane wave theory are indicated by the dashed lines. This plane wave description implies that the initial pressure wave at the duct entrance is uniform across the duct. The plane wave excites several modes within the lined portion of the duct. The normalization by L/H does not collapse all of the peak attenuation curves for various values of L/H onto a single curve. Instead several curves are obtained, and the band indicated encloses these curves over a wide range of duct length-height ratios.

Superimposed upon these theoretical results are experimental peak at- tenuations from several sources. These include engine, fan, and duct data.

The important point is that the experimental data and the theoretical curves follow the same trend with frequency parameter. The attenuation magni- tudes agree fairly well with the theory also.

Important ingredients in the propagation analysis are the properties of the soft walls of the passage. A schematic cross section through a soft wall is shown in figure IlI-15. The wall structure consists of a porous face sheet, shown with just one opening for simplicity. Behind the porous face sheet are resonant cavities. Pressure fluctuations, including acoustic dis- turbances, near the wall cause gas to flow in and out of the opening with velocity V. During the process, energy in acoustic form is removed from the duct by transforming it into disordered gas motion which eventually re- sults in heating of the gas.

For purposes of analysis, the response of the soft wall to sound is char- acterized by impedance, which is the ratio of the pressure to velocity (fig.

IH-16). Impedance has two components. The part corresponding_to pres- sure and velocity variations which are in phase is termed resistance. This is related to the dissipation of acoustic energy by flow through the porous wall. Resistance is inversely proportional to porosity.

The part of impedance which corresponds to pressure and velocity vari- ations which are out of phase is termed reactance. Two mechanisms con- tribute to reactance. One is the inertia of the gas moving in and out of the wall. The other is the compliance of the gas in the cavity as compression

andexpansionoccur during inflow andoutflow. This compressibility is a

large factor in the reactanceandis inversely proportional to cavity volume

or depth.

In addition to geometric parameters such as porosity and cavity volume which influence impedance, environmental conditions also affect the acoustic response of the wall. The spectral levels and frequencies are important along with the steady-flow field as indicated by the mean velocity profile at the wall. Models have been developed, at least for simpler wall construc- tions, which relate impedance to both geometric and environmental param- eters. Experimental methods of measuring impedance under realistic en- vironmental conditions are also being developed and applied.

In the following figures these two components of impedance are referred to in nondimensional form arrived at by dividing by the impedance of free air.

Now consider what analysis indicates about the values of resistance and reactance which will maximize the dissipation of acoustic energy, that is, maximize the attenuation of sound as it propagates down the duct. In figure 111-17, the desired, or optimum, impedances are shown versus the frequency parameter fH/c. For a fixed passage height, theory indicates that as the frequency at which we want to maximize attenuation is increased, resistance should be increased and reactance decreased. In physical terms this means that at low frequencies, wall structures must be thicker and more porous, while for high-frequency tuning they must be thinner and less porous.

The situation shown in figure III-17 is for no steady flow in the duct, M = 0. Fixing the frequency parameter at a particular value, for example fH/c = 1, and varying the passage Mach number leads to the optimum impe- dance shown in figure III-18. As before, negative Mach numbers refer to in- let cases and positive Mach numbers to exhaust cases. Maximizing inlet at- tenuation requires thinner, less porous wall constructions. As the exhaust Mach number increases, optimum resistance and reactance tend toward zero, which leads to the specification of thicker, more open structures.

Thus, the propagation theory specifies desired wall impedance values.

Models and experiments must be used to translate these wall impedances in- to actual geometric constructions.

7O ACOUSTIC LINER DESIGN In figure III-19 four types of lining construction are illustrated. The lining in the upper left position is a single-layer lining formed with a per- forated plate bonded to a honeycomb cellular structure. The perforated plate is the resistive impedance element, while reactive impedance is due to both the perforated plate and the air volume in the honeycomb cells.

The lining in the upper right position is also a single-layer lining; how- ever, the facing sheet is a layer of porous material. This material may be woven or fibrous, metallic or nonmetallic. The lining in the Quiet Engine nacelle is an example of this type of construction. In this application, the porous facing sheet is laminated from glass cloth impregnated with polyi- mide resin.

Both the perforated plate and porous layer lining construction can be de- signed for equivalent acoustic performance.

A bulk absorber is shown in the lower left corner. Both resistive and reactive impedance results from the porous material under the perforated facing sheet. The porous material may be either metallic or nonmetallic fibers or foams. The facing sheet generally has high porosity and is nor- mally used as a protective cover rather than as a significant contributor to lining impedance.

The last example is a double-layer lining formed with porous outer and inner facing sheets and two honeycomb structures. With this construction, the two layers are designed to attenuate noise at two frequencies. Two seg- ments of this type of lining construction are located in the Quiet Engine nacelle fan duct.

A multilayer lining is shown in figure M-20. Like double-layer acous- tic linings, the purpose of this construction is to provide peak attenuations at multiple frequencies. This example shows a perforated sheet covering an X-shaped core structure. The walls of the core are perforated to form re- sistive elements between the adjacent cells. This construction is used in the Quiet Engine fan frame and compressor inlet duct.

In the next three figures, the merits of particular lining constructions with regard to their impedance properties and acoustic performance are compared. In figure III-21, calculated impedance properties of single- and double-layer linings are compared with the optimum impedance. The ex- ample is based on designing for the least attenuated acoustic mode. The upper group of curves shows that the optimum normalized resistance in- creases with increased frequency parameter fH/c. The dot-dash curve shows the resistance of a single-layer lining designed for optimum attenua- tion at an fH/c of 0.75. Its resistance also increases with frequency pa- rameter but at a slower rate. The dashed line is the resistance of a double- layer lining. This lining has been designed to be optimum at fH/c values of 0.75 and 1.5, which is illustrated by the intersection of the double-layer resistance curve with the optimum at these two points.

The lower group of curves shows that optimum reactance decreases with increasing frequency parameter. The single-layer lining has the oppo- site trend and is near optimum only in the region of the design point. The reactance of the double-layer lining is similar to the single layer at values of fH/c below the design point. Above the design fH/c the reactance in- creases and then decreases to the optimum value at an fH/c of 1.5.

In figure III-22 the calculated acoustic performances of these two linings are compared. No difference is noted in 1/3-octave attenuation be- tween the single- and double-layer linings below the peak at a fH/c of 0.75.

Above this point, the single-layer lining attenuation decreases continuously.

The double-layer lining attenuation also decreases after the first peak and then increases to peak at its second design point at an fH/c of 1.5.

In this example, the increased bandwidth of the double-layer lining over the single layer is achieved with no increase in lining wetted area and only a 16-percent increase in lining thickness.

Double-layer acoustic linings can broaden the attenuation spectrum over that of a single-layer lining. However, in low-noise applications, use of only one double-layer construction may still not provide satisfactory atten- uation over an adequate frequency range. An alternate is to combine sev- eral lining constructions to provide sufficient bandwidth, as shown in fig- ure III-23. The curve shows sound attenuation versus frequency for three single-layer linings marked A, B, and C, with peaks at three different fre- quencies. Note that the total attenuation spectrum is much broader than any of the individual components.

This combination of different wall constructions can be done in two ways, as shown in the inserts on the hgure. The first sketch illustrates a series combination with similar liningsopposing each other onthe two walls.

The second method is parallel .construction, as illustrated in the second sketch. In this case, the opposite walls have differentlining constructions.

This parallel combination is desirable when a single splitter is used since the thinner high-frequency linings can be located on this splitter to minimize ring thickness and to reduce aerodynamic losses. The thicker low-frequency liningsare then located on the duct walls. Current propagation models are based on ducts with the same lining construction on opposite walls; conse- quently, series combination of liningsis better understood. However, itis anticipatedthat the parallel combination can be an effective design option after propagation models are established for optimizing liningswith dissimi- lar construction on opposite walls.

EX PERIMENTAL METHOD S One method of evaluating the performance of acoustic treatment in a fan duct is to measure the noise reduction in a test which simulates the condi- tions in the fan passages as closely as possible. The airflow through the duct, the incident sound pressure level spectrum, and the duct geometry are readily simulated. Simulation of mode structure is also desirable but dif- ficult to achieve without the use of a fan. Figure IH-24 shows a schematic of such an acoustic flow duct that provides this simulation except for mode structure. Two reverberation chambers are connected by a rectangular test duct which can be treated selectively on each of the duct walls. Splitters can be tested also. A high-intensity noise source is located in one chamber with a microphone which measures the input spectrum level. A microphone in the other chamber detects the acoustic signal with and without the acoustic treat- ment in place in the duct. Airflow direction can be reversed to simulate both inlet and exhaust configurations. By varying the airflow rate and input fre- quency the performance of the acoustic treatment configurations can be mea- sured at Mach numbers and sound pressure levels typical of a full-scale fan.

An alternate approach to this measurement scheme is shown in figure III-25. As before, a steady airflow can be passed through the duct with Mach numbers equivalent to those found in typical engine flow passages.

The height between treated walls can be varied to simulate engine duct geometry. For the case shown here, the noise source is downstream of the treatment test section and therefore simulates the inlet mode. By moving the source downstream of the facility noise muffler and upstream of the treatment section, the exhaust mode is set up. The noise reduction pro- duced by the treatment is measured by traversing acoustic probes which measure the sound pressure level distribution between the treated walls and permit the calculation of sound energy reduction by the treatment.

These facilities provide a means to evaluate treatment effects such as duct geometry and wall impedance that can be correlated to give design pro- cedures or used to guide theoretical studies. They can also be used to eval- uate novel wall treatment constructions that are not amenable to any theory.

COMPARISONS BETWEENTHEORY AND EXPERIMENT An example of a measured and calculated sound attenuation spectrum is shown in figure III-26. The measurements were obtained as described pre- viously in the reverberation chamber facility at a Mach number of 0.28, and it can be seen that the theoretical curve predicts the main features of the measured attenuation spectrum.

Another example of experimental duct data is shown in figure III-27.

These data were taken in the acoustic flow duct facility (fig. III-25). These data, with repeat points, were acquired with pure tones and consequently exhibit more experimental scatter than in the previous example. Again the main features of the experimental data are reflected by the theoretical pre- diction.

In order to calculate the theoretical duct attenuation accurately, the acoustic impedance components are a required input. The effects of sound pressure level and grazing flow must be incorporated in the impedance model. In order to determine the accuracy of the theoretical impedance model, measurements of the impedance components of a single-layer treat- ment in the presence of grazing flowhave been made in a flow duct similar to

the oneshown in figure III-25. A comparison of the calculated impodance

(for a particular impedance model) with the measured impedance at a Mach number of 0.35 is shown in figure III-28. The acoustic resistance with flow is seen to be very flat with frequency and agrees well with the calculations.

The calculated reactance agrees with the measured data at low frequencies but not at high frequencies. Measurements will be relied on in this case and can become the basis for modifying the impedance model.

These comparisons indicate that the acoustic impedance components can be calculated and measured under simulated engine flow conditions, and that the noise reduction capability of acoustic treatment can be effectively tested in laboratory flow duct tests.

FULL-SCALE FAN SUPPRESSION TESTS Duct test facilities provide a useful tool for screening suppressor ma- terials; however, the noise environment cannot truly simulate that found in the turbofan engine ducts. The Lewis full-scale fan test facility was used to test full-scale suppressors.

Figure III-29 shows a cutaway view of this facility. Shown are the 6-foot-diameter rotor and the stators. The inlet suppressor shown has three splitter rings. The fan is driven by the shaft passing through the inlet.

The exhaust duct suppressor has a lined inner and outer cowl with one lined splitter ring.

An inlet which was tested in this facility is shown in figure III-30. This suppressor is 6 feet in diameter and is made up of cylindrical sections. The outer cowl is lined, as are both sides of the three splitter rings. A parallel combination of wall constructions are used for the rings. The outer side of each splitter ring is tuned for high frequency, while the inner sides are tuned for a lower frequency.

Typical noise data obtained with this inlet suppressor on fan A are shown in figure III-31. The sound pressure level spectra at 50 ° to the inlet for both the unsuppressed and suppressed cases are shown. The dominant blade passage tones are reduced to levels which are barely distinguishable in the suppressed spectrum. The difference between the suppressed and un- "/5 suppressed spectra, that is, the attenuation spectrum, is shown at the bottom of the figure. A maximum attenuation of about 18 decibels occurs at blade passage frequency.

For comparison, the same inlet suppressor was put on high-speed fan C and the results are shown in figure III-32. The unsuppressed spectrum is characterized by strong multiple pure tones which are roughly the same level as the blade passage tone at 2000 hertz. Theblade passage tone is nearly indistinguishable in the suppressed spectrum. Since the suppressor was designed for a low-speed fan with a spectrum similar to fan A, its bandwidth was not wide enough to realize uniform attenuation throughout the multiple pure tone range. Although substantial attenuation (about 19 dB) was observed at 800 hertz, the multiple pure tone content at 500 hertz was not reduced as much. Another section of suppressor tuned in that range would have been desirable.

The variation of perceived noise levels with angle is shown in figure III-33 for low-speed fan A. As before, open symbols are for unsuppressed data, while solid symbols refer to the suppressed case. The fan was oper- ating at takeoff speed and the levels correspond to a 1000-foot flyover. About 12 PNdB of suppression was realized in the inlet and about 6 PNdB in the ex- haust. For comparison we can superimpose the same type of data for fan C, as shown in figure III-34. It can be seen that the magnitudes of the per- ceived noise reductions (unsuppressed minus suppressed) were about the same both in the inlet and exhaust as they were for fan A. However, treat- ment acting over a greater bandwidth is required to bring fan C noise down to levels approaching those of suppressed fan A.

There is some evidence that cowl treatment alone can be very effective in reducing the multiple pure tone noise of high-speed fans like fan C. The results of such a test on fan C are shown in figure III-35. Two attenuation spectra corresponding to the cases with and without splitters are given.

Attenuations in the multiple pure tone range, 500 to 1600 hertz, were essen- tially the same whether or not the splitters were present. This result is not predicted by the analyses discussed previously. Planned suppressor tests on fan C will further explore the characteristics of multiple pure tone atten- uation.

Our experience with suppressors on five full-scale fans is summarized in figure III-36. The maximum perceived noise levels shown are for 90 000 pounds thrust generated at takeoff speed for a 1000-foot flyover. Inlet levels are shown on the left and exhaust levels are on the right. Unshaded bars are unsuppressed levels, shaded regions represent suppressed levels, and the heavy black line is the estimate of fan jet noise. Reductions of 12 to 18 PNdB were realized in the inlet with jet noise a limiting factor only for fan E. On the exhaust side, reductions range from 4 to 10 PNdB. Jet noise was clearly limiting fan E levels. The overall suppressor performance in the exhaust was generally less satisfactory than in the inlet. The possibility of noise being generated by flow through lined exhaust passages operating at higher Mach numbers may be a factor limiting exhaust suppressor performance.

In this connection, some effects of passage geometry on suppressor performance in the aft direction are shown in figure III-37. These data, which were obtained from tests on full-scale fan D, show how splitters can affect noise. Aft sound power spectra are shown for three configurations.

The configurations are represented schematically by a cross section through the fan axis. The base case, shown with triangular symbols, is the hard cowl with no splitters. When hard splitters are added, as shown by the open circular symbols, the noise is raised over a broad frequency range. With all passage surfaces soft, the solid circular symbols, the levels are reduced to the lowest values measured except in the small frequency range around 600 hertz where the suppressed levels are not quite as low as those exist- ing without suppression. This result suggests that noise generation proces- ses occur in the lined passages, especially in the fan exhaust. While sup- pression reduces the noise, it may not remove enough acoustic energy at all frequencies to overcome generation effects.

CORE NOISE SUPPRESSION In the late 1960's, a program was sponsored by NASA Langley Research Center to reduce fan noise from the Pratt & Whitney JT3D engine. Flight tests with the engine in a quiet nacelle (fig. III-38) reduced fan noise to the point where noise from other sources became significant. For exam- ple, at approach power settings, fan noise was reduced below the level of turbine noise. A subsequent program was established by NASA Langley to design, fabricate, and demonstrate acoustic treatment of the turbine noise.

The tailpipe lining shown in the schematic Jconsisted of brazed stainless- steel perforated sheet/honeycomb linings installed on the tailpipe wall and both sides of a ring installed in the tailpipe. The lining was tested on a JT3D engine housed in a quiet nacelle from the 1969 program.

Figure III-39 shows the treated tailpipe. The tailpipe lining shown in the photograph was analytically designed based on aerodynamic and acoustic measurements obtained within a production tailpipe. The 19-inch-long lining was designed for maximum attenuation of the fourth-stage turbine tone at 5500 hertz at an approach power setting. The linings required for this application had low open areas and thin core thicknesses. For example, the lining on the tailpipe wall and the opposing ring was a 3.44-percent perfor- ated plate with a honeycomb core depth of 0.18 inch. The ring inner lining was a 1.85-percent perforated sheet with a 0.10-inch core depth.

Test results for the lining installed in a JT3D engine are shown in figure III-40. In this figure is shown the I/3-octave spectrum measured for the treated and untreated tailpipe at a 200-foot radius at a polar angle of 120 ° from the inlet centerline at the design speed. This is the angle at which the fourth-stage turbine tone has maximum sound pressure level. The shaded area represents the attenuation accomplished by the treated tailpipe. Ap- proximately Ii decibels of suppression were obtained in the 5000-hertz I/3- octave band containing the fourth-stage turbine tone. The second- and third- stage turbine tones, located in the 8000-hertz band, were attenuated 7 deci- bels at this angle. The reduction in perceived noise at this angle was 3.9 PNdB.

This program demonstrated that the same lining design and analysis techniques described previously for fan noise suppression can be success- fully applied to reducing turbine noise.

INSTALLATION CONSI DERATIONS To complete the discussion, there are several other aspects of a suppressor that should be mentioned. These are shown in figure III-41 and categorized as installation considerations. In the inlet, there is concern about anti-icing of the splitter ring leading edges and their support struts and about the ability of these members to withstand foreign object damage.

In the core exhaust, materials and bonding methods must be suitable for the high temperatures that are present. There are also several more general factors that enter into treatment selection. For example, materials and bonding techniques must be selected with a view to their sonic fatigue life. The construction technique must provide a way for water and other contaminants to drain from the backing cavities of the treatment. Still another factor is the pressure loss associated with splitters and with por- ous or perforated surface materials.

For this purpose, splitters are located by potential flow analysis so that they follow streamlines, and care is taken in the shaping of leading- and trailing-edge surfaces. Some of these considerations are discussed in another paper, which describes the nacelle for the Quiet Engine.

Finally, three standard factors - maintainability, cost, and weight - must be considered.

One question of obvious interest is the performance penalty associated with acoustic treatment. In figure III-42, we have made an estimate of the pressure loss as a function of noise reduction. The estimated total pressure loss in percent is plotted against the estimated reduction in perceived noise level. The estimates are for a value of the frequency parameter between 1 and 2. Curves are shown for three values of the flow Mach number over the surfaces. It can be seen that the losses increase with flow Mach number.

For small noise attenuation the losses are also small; however, as the amount of attenuation increases above about 10 PNdB, the losses increase sharply. This is due to the decrease in treatment effectiveness as treat- ment length is increased to give the large attenuation required.

The data points shown on the curves are Boeing estimates for the Quiet Engine nacelle. At the cruise and takeoff conditions the Mach numbers in the engine are about 0.6 and 0.48, respectively. This suppressor was de- signed for about 15 PNdB of attenuation so that these estimates are in reasonable agreement with the curves shown.

It should be emphasized that these curves are only estimates and should not be applied in any general sense. Improvements in suppressor technology should lead to a reduction in these losses.

SUMMARY To summarize this discussion, significant noise reductions can be rea- lized by acoustic treatment for all the engines currently under consideration.

The amount of reduction will depend on the particular engine cycle and its use. A floor is set to suppression by the jet noise; however, in some prac- tical cases, noise sources other than jet noise seem to limit the benefits of treatment. Work is needed to identify these sources.

The principal suppressor parameters and their relationships have been discussed. These relationships have been used to formulate a suppressor design methodology which has been applied to the design of several full- scale suppressors.

Tests of these suppressors on several full-scale fans have demonstrated noise reductions of the order of 10 PNdB. The amount of suppression in sev- eral instances seems to have been limited by reaching noise floors that are not clearly at the estimated jet noise level but are not far above it. It should also be remembered that the fans involved in these tests were designed for as low a noise output as currently possible and therefore the noise to be sup- pressed is not as far above the jet noise floor as it might be in some other fan stages.

In addition to identifying noise floors, two other areas need attention.

The first is attenuation by the fan exhaust suppressors, which seems in most cases to be less than that of the inlet. This, as suggested, may be related to noise generation by the higher velocity air in the fan exhaust duct scrub- bing over the splitter and cowl surfaces. The second is the effectiveness of the outer cowl treatment alone on multiple pure tones. We need to under- stand why this treatment alone is as effective as it is. This question relates to whether a high-speed fan can become as quiet as a low-speed fan without a substantial increase in the amount of treatment.

In the future, substantial improvements in suppressor efficiency should occur as understanding of the mechanisms of suppression improves. In this way, reduced amounts of treatment may be possible with no loss in noise reduction.

8O INTERNAL NOISE SOURCES INLET FAN DISCHARGE

NO, SE-_, _ FAN NO_SE_ \

_ .......... l-, _ r--- TURBINE NOISE -_ COMPRESSOR NOISE -_ Figure III-I cs-_3z_,_ TYPICAL UNSUPPRESSED SOURCE NOISE

5_B

RELATIVE LEVEL ..

L

FAN CORE JET TOTAL C5-6]190 Figure III-2 ENGINE WITH SUPPRESSION Figure 111-3 CS-63329 TYPICAL SUPPRESSED SOURCE NOISE (-"--'I UNSUPPRESSED [_ SUPPRESSED 7- 5 PNdB _t_ RELATIVE \\\\x NOISE r--_ I ,:_:: LEVEL FAN CORE JET TOTAL CS-63180 Figure 111-4 NOISE SPECTRUM REQUIRING SUPPRESSION -LOW SPEED FAN SOUND tOmB PRESSURE LEVEL _ JETNOISEJ_

i1

III I I Ill I I I

40 I00 i000 I0O00 FREQUENCY. Hz Figure III-5 CS-_3186 NOISE SPECTRA REQUIRING SUPPRESSION- LOW AND HIGH SPEED FANS

H,OH SPEED FAN-._ b'

T ,%'

_oo_ • _,__

SOUND PRESSURE LEVEL JET NOISE _I

11 I I

111 I i[

40 100 FREQUENCY, Hz CS-63176 Figure 111-6 SUBJECTIVE INFLUENCE ON SUPPRESSOR ATTENUATION GOAL ,%, i', ,%_P/I ',, ANNOYANCE, NOYS • .-, _OW S PEED FAN_

___T_ _,-,7 _- --_'%_7" T_

100 1000 10 000 FREQUENCY,Hz CS-63191 Figure 111-7 REQUIRED ATTENUATION SPECTRA- LOW AND HIGH SPEED FANS 30 -- HIGH SPEEDFAN-, SOUND ATTENUATION, dB

S

200 1000 10 000 FREQUENCY. Hz Figure 111-8 CS-63181 PARAMETERS FOR TYPICAL ATTENUATION SPECTRUM 30 A dBpEA K --_ 20- / i _ _BANDWIDTH AI-[ENUATION, dB 10- J fPEAK o I I II _ _ _1 200 1000 10000 FREQUENCY,Hz CS-63188 Figure Ili-9 SUPPRESSION PARAMETERS IMPORTANT PARAMETERS: fH/c _H/X, UH, M, WALL IMPEDANCE, INITIAL SOUND PRESSURE PROFILE CS-63332 Figure III-10

EFFECT

OF FREQUENCY PARAMETER ON PEAK ATTENUATION (a)LEASTATFENUAEDMODETHEORY _ M :0 20- PEAK I0 AI-rENUATION _L_ UH 8

_B_

k-o-_-j

4 PRACTICAL

2 J I II , , %

.2 .4 .6 .8 l 2 4 6 FREQUENCY PARAMETER, fHlc CS-63193 (b)FOR SAMPLE INLETS: D : 6 FT, f : 2000Hz, AdB : -20 40- -_ rPRACTICAL NUMBER OF RANGE RINGS PEAK 20-- -- ,-____4_ ATTENUATION 10 i- ] _ X -_.

PER IJH 8 _ ,,- '--_---_ p

tUN ) 6_

(-,, _B_ 4- !

__ . _± ...-

.2 .4 .6 .8 1 2 4 6 FREQUENCY PARAMETER, fl-I/c CS-63196 Figure III-ll NUMBER ON REOUIRED EFFECT OF SPLITTER RING SUPPRESSOR AREA DUCTDIANtETER = 6 IT, f = 2000 Hz .8 .6 .4 RELATIVE SUPPRESSOR AREA .2 .1 .08 .06-- ,1 0 2 3 CS-_3Z7 l NUMBEROF SPLITTERS Figure III-12 MACH NUMBER EFFECT ON PEAK ATTENUATION MACH 40 - _ NUMBER ._ _.4 (INLET)

\\

IE×'AUSTL'>..\

ATTENUATION 10- PER L/H PEAK DUCT) -"" 2 I I i I 'l H_ _ "% .2 .4 .6 .8 1 2 4 6 FREQUENCY PARAMETER,fHIc CS-63[97 Figure III-15 PEAK SOUND ATTENUATIONS FROM THEORY AND EXPERIMENT \ A ENGINE 20 1" DATA 10 \ [] FAN \ O DUCT 8_\\ 0 \\ SOUND ATTENUATION _R UH

A dB_ 4- o b_o_,'

t-_-j

\ %.

2 - "" _"_7":"-- LIMITS OF PLANE \ ,, WAVETHEORY \ \ \ \ l.,,l I I I"1 J .8 I 2 4 6 8 10 FREQUENCY PARAMETER, fH/c CS-63195 Figure III-14 SOFT WALL CROSS SECTION r , i,7 //-POROUS FACING SHEET Figure III-15 CS-63267 SOFT WALL PROPERTIES P _;__v IMPEDANCE V _ REACTANCE: RESISTANCE: Voo DISSIPATION BY p i-- INERTIAOF GAS FLOW THROUGH _'i -- //// MOVINGIN AND OUT OFWALL POROUSWALL _] [ COMPRESSIBILITY I POROSITY f ',_" OF GAS IN CAVITY CAVITY VOLUME (DEPTH) ENVIRONMENTAL PARAMETERS: SPECTRUM FLOWFIELD Figure II1-16 cs-6326_ THEORETICAL IMPEDANCE FOR MAXIMUM ATTENUATION M=0 RESISTANCE

0 I I

O REACTANCE .5 1 2 5 FREOUENCY PARAMETER,fHIc CS-63264 Figure III-17 MACH NUMBER EFFECTS ON OPTIMUM IMPEDANCE fH/c= I E_ INLET.f,.EXHAUS_ RESISTANCE 1 0t I I I I I i I I I I 0- REACTANCE-2 -1 --i i_1 i _ I I I I I -, 0 ,5 DUCTMACH NUMBER CS-63265 Figure III-18 LINING MATERIALS AND CONSTRUCTION PERFORATED PLATE- POROUS LAYER-HONEYCOhIB HONEYCO_AB BULK ABSORBER DOUBLE LAYER Figure III-19 cs-6_33o MULTILAYER WALL CONSTRUCTION Figure III-20 IMPEDANCE CONTROL BY CONSTRUCTION M=0.3 SINGLE LAYER 2 DOUBLE LAYER RESISTANCE I "" 10 _ • DESIGNPOINT .... I I OPTIMUM-' I I / 1- j- J REACTANCE 0 ---- MUM ] 5 .75 1.0 1.5 2.0 FREQUENCY PARAMETER, fHlc cs-631Bz Figure III-21 ATTENUATION BANDWIDTH COMPARISON SINGLELAYER DOUBLE LAYER / /,_ /I/ ii

_- / \ / /

S.OUND

ATTENUATION,

dB

10 -.

I I I I • 05 .75 1.0 1.5 2.0 FREQUENCY PARAMETER, fHlc Figure III-22 CS-G3_B3 METHODS OF INCREASING SUPPRESSION BANDWIDTH C SERIES PARALLEL SOUND ATTENUATION 600 I000 2000 400O 6000 10 000 FREQUENCY, Hz Figure III-23 cs-631_B FLOW DUCT TESTS USING REVERBERATION CHAMBERS AIR /-NOISE SOURCE AIR "_ MICROPHONE SOUND -- AIR "L-----_EXHAUST DUCT FLOWJ_ INLET Figure 111-24 CS-63177 FLOW DUCT TEST USING SOUND MEASUREMENTS IN THE DUCT ,,-TEST SECTIONACOUSTIC NOISE SOURCE i TREATMENT l INLET MODE -7 _//////////////////A / /L--AIR MUFFLER : MICROPHONE PROBES_I Figure III-25 CS-6327Z FLOW DUCT ATTENUATION SPECTRUM USING REVERBERATION CHAMBER 10- THEORY _ _ 0 EXPERIMENT SOUND 6 AII'ENUATION, O_ _, M = 0.28 dB 4 m 0 I I I I I I 2O00 4OOO 6OO0 8OOO FREQUENCY, Hz Figure III-26 CS-63198 FLOW DUCT ATTENUATION SPECTRUM USING SOUND MEASUREMENTS IN DUCT SOUND THEORY ATTENUATI ON, 0 EXPERIMENT dB M=0.4 01 I I I I 2000 4000 600O FREQUENCY, Hz CS-63194 Figure III-27 MEASUREMENTS WITH GRAZING FLOW IMPEDANCE M=0.35 n RESISTANCE1.0 RESISTANCE MODEL r-i , I i I .5is / REACTANCE MODEL 1.0-- i / o / .5- O / I / o 0-- REACTANCE /0 / ".5 _///o i I -i0, _ I 3OOO • 1000 20O0 FREQUENCY, Hz CS-63263 Figure IIl-_ FULL-SCALE FAN WITH SUPPRESSION CS-63179 Figure III-_ INLET DUCT WITH ACOUSTIC TREATMENT CS-63200 Figure III-30 SUPPRESSOR PERFORMANCE ON LOW SPEED FAN A TAKEOFF SPEED; I00FOOT RADIUS llO SOUND I00 PRESSURE LEVEL, dB 90

i1"w'l' -T.... T-T'II _ J i I I

8O SOUND ATTENUATION, dB I000 lO000 CS-63187 FREQUENCY, Hz Figu re 111-31 SUPPRESSOR PERFORMANCE ON HIGH SPEED FAN C TAKEOFF SPEED; 100 FOOTRADIUS 120 - . /_500 PRESSURE 100 SOUND 110- j__PRESSED LEVEL, dB go _w \_SUPPRESSED ,,I i i lil I i I II 8O 2O AI-I'ENUATION, 10 SOUND f "_'l J_"__ oB 0 i_1 ill I I l il 40 lO0 1000 10000 FREQUENCY, Hz cs-63_89 Figure III-32 FAN PERCEIVED NOISE DIRECTIVITY- LOW SPEED FAN A TAKEOFF SPEED, 1000 FOOTFLYOVER 110F _-UNSUPPRESSED NOISE gO LEVEL, PNdB 80

6olOl I I Ill I I i

O 20 40 60 80 100 120 140 160 ANGLE FROM INLET, DEG Figure III-33 cs-63184 9'I FAN PERCEIVED NOISE DIRECTIVITY -FANS A AND C TAKEOFF SPEED, 1000FOOTFLYOVER 11OF If HIGH SPEED FAN C NOISE LEVEL, PNclB 70 -gf SOLID SYMBOLS- SUPPRESSED , INLET _ EXHAUST O0 Ul I I Ill I I I 0 20 40 60 80 I00 120 140 160 ANGLE FROM INLET, DEG CS-63! 92 Figure III-M SUPPRESSOR PERFORMANCE ON HIGH SPEED FAN C TAKEOFF SPEED; I00FT RADIUS WITH INLET SPLITTER RINGS1 -- _ _. _ FINLET WALL 50o _ TREATMENT ONLY SOUND i0 ATTENUATION, dB 0 t I000 I0000 FREQUENCY, Hz Figu re III-35 cs-63445 SUPPRESSOR PERFORMANCE ON FULL-SCALE FANS TAKEOFF SPEED; 90 000 LB THRUST; 1000FOOTFLYOVER [_UNSUPPRESSED FAN JET--, I SUPPRESSED / I 110 - EXHAUST I 100- -::-- NOISE LEVEL, PNdB 90- _\\\x B A ¢ 80 FAN: D E B A C D E Figure III-36 cs-63174 EFFECTS OF PASSAGE GEOMETRY ON SUPPRESSOR PERFORMANCE 1_-- , HARD // "

_

POWER LEVEL, ]30 dB 125 -(_ _

I_I I I III _ i _iI .J

12010 100 1000 10 000 2U 000 FREQUENCY, Hz C5-63175 Figure III-37 JT3D TURBINE SUPPRESSOR r 1,,f-ACOUSTIC a__,/_, TREATMENT CS-63331 Figure III-38 TURBINE NOISE SUPPRESSOR Figure III-39 JT3D TURBINE NOISE SUPPRESSION APPROACH SPEED; 200 FOOT RADIUS TREATMENT TURBINE TONES ! %.

0 NACELLE -- Q NACELLE & TAILPIPE SOUND PRESSURE 85 LEVEL, dB 8O

i I I iI i a

10 OOO FREQUENCY, Hz CS-f_3173 Figure III-40 TREATMENT INSTALLATION CONSIDERATIONS / / \ / /-- ANTI-ICING GENERAL: HIGH TEMPERATURE -_ / SONIC FATIGUE LIFE z_ FOREIGNOBJECTDAMAGE WATERDRAINAGE PRESSURE LOSS MAINTAINABILITY COST WEIGHT CS-63333 Figure III-41 ESTIMATED PRESSURE LOSS DUE TO SUPPRESSION 1 < fHIc <2 10 -- BOEING NACELLE 0.6/.5 / /

/

PRES S URE EST IMATES ///_/4 LOSS,

A2 5-

;/

p P

__1

O 5 10 15 20 SOUND AI-I'ENUATION, APNdB CS-634Z4 Figure III-42 IV. JET NOISE Uwe H. yon Glahn, Vernon H. Gray, Eugene A. Krejsa, Robert Lee,* and Gene L. Minner Jet noise can be the critical factor in establishing the noise factor for an engine. The jet-noise floor is determined primarily by the jet exhaust ve- locity. At the high exhaust velocities of an SST engine, the ability to meet the noise standards in Federal Air Regulation 36 (FAR-36) is difficult even with the best available exhaust noise suppressors. If the noise level re- quirement is lowered 10 or 20 decibels below FAR-36, as ha_ been advocat- ed in the Civil Aviation Research and Development Policy Study, a major breakthrough in noise suppression must occur or a variable-cycle engine must be used. Operation of many older conventional CTOL aircraft, such as the DC-8, also present noise problems because their noise levels, due to their high exhaust velocities, are near the present FAR-36 sideline require- ment at takeoff. On the other hand, advanced CTOL engines, because of their low exhaust velocities, can achieve noise levels at least 15 decibels below FAR-36 without noise suppressors. For further reductions the jet noise level could be lowered by the use of exhaust noise suppressors. It should be noted, however, that experience has shown that jet noise suppres- sion becomes increasingly more difficult with low jet velocities.

For practical applications, the jet noise must be predictable together with a means of assessing attainable suppression levels.

The specific problem areas and progress made in understanding jet noise and its suppression discussed in this paper are (1) Subsonic velocity power law (2) Jet density effect (3) Supersonic jet noise (4) Jet noise correlation (5) Quiet engine noise measurements and prediction (6) Jet noise suppressors General Electric Company.

(7) Flight effects on suppressors (8) Thrust reverser noise The validity of the Lighthill eight-power law with respect to the variation of noise level for low subsonic velocities will be examined. The effect of jet density on subsonic jet noise is then considered. Having examined the major problem areas of subsonic jets, the noise of supersonic jets is briefly dis- cussed. Then a NASA correlation for the peak sideline overall sound pres- sure level (OASPL) is presented. The available noise prediction techniques as applied to NASA's Quiet Engine is then considered.

The acoustic and aerodynamic performance of a number of jet noise suppressors then is discussed. This discussion will also include the effect of flight velocity on the performance of jet noise suppressors.

The last item concerns an aspect of jet noise that has been largely ig- nored to date, namely, thrust reverser noise.

STATUS OF JET NOISE With the introduction of turbojet powered aircraft, jet noise became a major problem. More current turbofan engines, having larger flow areas and lower exhaust velocities, had less jet noise, and internally generated noise was the major problem. Much effort at present, notably in the Quiet Engine and STOL Programs, is devoted to reducing internal noise by fan de- sign and acoustic absorbers. The effectiveness of these measures has been limited by a jet-noise floor, even at exhaust speeds below 1000 feet per sec- ond, which are characteristic of turbofan engines. As future noise regula- tions become more restrictive, the jet-noise problem will become even more important. It is therefore necessary to have a basic understanding of low-velocity jet noise.

The main concern with noise disturbance has been with peak levels for a fixed observer. It has been standard practice (fig. IV-1) to present jet noise correlations in terms of the maximum 200-foot sideline overall sound pres- sure level (OASPL). The noise level is normalized on jet density pj and area A. Jet exit velocity V is the primary determinant of the noise and is used as the abscissa.

A standard procedure for prediction of jet noise of engines at exhaust

speeds above 1000 feet per second was established by the Society of Automo- tive Engineers Aerospace Information Report-876. The curve (representing sound pressure squared) in the range of 1000 to 2000 feet per second varies with about the eighth power of velocity. This result compares favorably with the classical Lighthill theory, which predicts that the jet noise total sound power varies as V 8.

It was expected that one should be able to extend the SAE correlation curve to lower jet velocities. However, a large amount of the early engine data exhibited higher noise levels and lower velocity dependence. It was suspected that this different behavior might be caused by noises from sources other than the jet, increasing in relative importance as jet velocity was reduced. This suspicion was supported by jet noise data from NASA's first quiet fan rig. The fan had noise suppressors that eliminated some noise from internal sources. The results (circle symbols in fig. IV-1) show the same trend as the SAE curve, although the level is shifted downward a few decibels. This shift probably results from overcorrection for pj and will be discussed later.

The General Electric Company has provided jet noise data (fig. IV-2) taken for fans A, B, and C in full-scale and B and C in scale-model versions.

These data are in satisfactory agreement with the curve fit of the quiet fan (QF 1) results (fig. IV-1), which are approximately an eighth power curve.

The deviation at low velocity is believed due to internal noise.

Small scale model jets have long been used to study jet noise, assuming that these studies were representative of full-scale engine jet-noise behavior.

Experience with model jets (fig. IV- 1) showed the data lying in a broad un- certainty band from slightly above the engine experience band down to some- what below the extrapolated SAE curve. It is believed that jet noise in this velocity range varies directly with V 8 and that internal suppressible noises have caused the deviations.

In order to determine if an engine shows results similar to those for fans, the effect ol muffling a J-65 turbojet engine was studied (fig. IV-3).

A long acoustically treated inlet was placed on the engine in order to reduce the inlet-radiated noise to a negligible level. An exhaust duct liner (fig. IV-4) was designed to remove noise in the frequency range of interest for jet noise, with peak attenuation at 250 hertz (large holed surface) and 1000 hertz (small holed surface). The liner shown served as the inner sur- face of the annular exhaust muffler. The outer boundary of the annular flow passage was covered with the same type of liner. This was a research tool and not flightworthy hardware.

The engine was run with both hard-wall and lined-wall exhaust ducts.

Typical 1/3-octave frequency spectra of sound power at 50 percent of design speed are shown in figure IV-5. Comparison of the lined- and hard-wall spectral results shows a significant noise reduction by the liner over the frequency range of interest for jet noise. As can be seen the maximum at- tenuation occurred near the design points of 250 and 1000 hertz.

The peak-sideline OASPL as a function of velocity (fig. IV-6) is also higher for the hard configuration than for the lined configuration. The data for the lined configuration follow the V8 curve down to about 700 feet per second. These data indicate that there is a significant amount of low- frequency internally generated noise in an engine. With adequate internal noise suppression the pure jet noise obeys the V 8 law.

Some insight into the effect of jet density on jet noise can be obtained from data taken at different temperature levels since jet density is inversely proportional to jet temperature. The data used in this presentation cover a temperature range from ambient to about 2500 ° R. These data are taken from reports published by NACA-NASA and Boeing. In addition, some re- cent unpublished data from the NASA Lewis hot-jet facility (fig. IV-7) is used. In this facility pressurized air is supplied from the Center's central air supply. A muffler is located downstream of the supply line valve to re- move any valve and upstream noise. Just downstream of the muffler is a preheater. This preheater is made up of five annular combustors that can heat the air to 1400 ° F. The final section of the rig is a J-85 afterburner that can heat the air an additional 2000 ° F. The nozzle is an adjustable con- vergent nozzle. For the data used herein, the nozzle was set at 11 inches.

The data from this facility and the published data will be presented in terms of the peak sideline overall sound pressure level, referenced to 200 feet. As indicated in figure IV-8, the peak-sideline OASPL is normal- ized by the exit nozzle area times the jet density squared. Lines represent- ing data at several temperature levels are shown. The jet density is largest for the lower temperature data and decreases as the temperature increases.

Also shown is the curve of jet noise from SAE AIR 876. The data segregate by jet temperature level with the higher temperature data coinciding with the SAE curve.

A similar plot is shown in figure IV-9, except that only the nozzle area is used to normalize the OASPL. On this plot the data fall together, and no trend with temperature or density is apparent. The data in figures 157-8 and IV-9 indicate that normalizing by density squared overcorrects for density effects. At the lower velocities the relation between OASPL and velocity is about an eight-power law. At higher jet velocities (greater than 1500 ft/sec) the curve bends over, and a three-power law is approached. At these higher velocities, the flow is still subsonic.

The measured peak-sideline OASPL for supersonic jets is shown in fig- ure IV- i0 together with the subsonic curve. As the jet Mach number in- creases above I. 0, the OASPL for the supersonic jets increases to values above the subsonic jet curve. This increase is probably due to shock- turbulence interaction or the fact that for supersonic jets the mixing region becomes much longer than that of a subsonic jet. The data taken are at sev- eral temperatures, and as the temperature increases the difference between the supersonic and subsonic jet decreases. It is evident that supersonic jet noise is influenced by both jet temperature and jet exhaust Mach number.

Correlation parameters have been developed that include these preced- ing factors in what otherwise can be classed as a modified Lighthill correla- tion. The correlated data are shown in figure IV-11. The ordinate consists of the peak-sideline OASPL normalized by the expanded-flow area, jet Mach number, and shock parameter F', which is a function of the jet velocity, Mach number, and the ambient speed of sound as follows: 7.5(M - 1) 3 l !

=1+i.

0135 + (M-1)3]11+ 2(M-1)4]11+ 0.05(V) 8 M-3_ where M is the jet Mach number, V is the jet velocity, and a is the am- bient speed of sound. The abscissa is a dimensionless velocity and Mach number parameter. Also shown in this figure is the subsonic jet noise curve.

It is apparent that good correlation between the subsonic jet noise data (shown by the plain symbols) and that for supersonic jets (tailed symbols) has been achieved over the entire range of conditions noted. R should be pointed out that the proposed correlation was developed without includingthe jetdensity as a parameter. R remains to be seen if other convergent-nozzle data from worldwide sources also correlate after being properly screened for internalnoise effects.

In order to account for changes in the ambient speed of sound a and in the ambient density p, the ordinate in this figure should be expanded to in- clude these parameters. Thus, the ordinate should include the additional term -10 log pa 3. The ordinate can be normalized to avoid negative values.

Inclusion of these terms does not alter the basic correlation, but itdoes change the absolute value of the ordinate.

Because the peak-sideline OASPL is closely related to the sound power, a similar correlation can be developed for totalsound power data.

The preceding OASPL correlation applies to simple nozzles, so the question arises as to how well the noise for fan-jetengines that have both core and bypass jetexhausts can be predicted. The NASA Quiet Engine falls in this class.

Some of the noise characteristics of the Quiet Engine are shown in fig- ure IV-12. Typical 1/3-octave band frequency spectra are shown of the sound pressure level measured at takeoffspeed at 120° to the fan inletdirec- tion for the engine and the fan. At this angle, both internaland external sources axe strong noise contributors.

Typically for an engine of this size, the jetnoise is primarily contained in the low-frequency part of the spectrum (up to 1000 Hz). The internally generated noise dominates the spectrum at higher frequencies, notably at the blade passing frequency and itsharmonics. There is also some internally generated noise that contributes to the low-frequency part of the spectrum.

By internallygenerated noise is meant allsound propagating from inside the engine.

A significant difference at low frequency between the data taken for the quiet fan and the Quiet _.ngine is shown in figure IV-12. This difference is a result of added jet noise generated by the high-speed core flow, which was not present in the fan, and there may be some additional contribution from internal sources.

In general, jet noise is not a limiting item to the Quiet Engine systems

noise at takeoff power while the aircraft is in flight. There are several

things that havebeen learnedaboutjet noise as a result of the Quiet Engine

Program. It was foundthat, whenthe fan ducts are fully suppressed,the

exhaustjet noise from the fan stream doesbehavelike anypure jet noise

oughtto behave, namely, the acoustic power andvelocity relation obeysthe

classic eighth power law.

In figure IV-13 the 200-footsideline maximumOASPL is plotted as a

function of exhaust jet velocity. The data are based on all the full-scale and scale-model fans that have been tested at Lewis and at General Electric under the Quiet Engine Program. No engine data are included in this plot.

The OASPL is taken from 50 to about 1000 hertz.

The data correlate quite well with velocity raised to the eighth power.

An empirical equation developed by General Electric for the straight line is shown. Both the velocity exponent and the proportionality constant agree well with previous results on this subject reported by Lewis, and also with other scale-model jet data. The data shown are for fan jets discharging at essentially ambient jet density. Whether the density term used in the em- pirical equation is indeed a significant correlating parameter cannot be tested. At General Electric the retention of the density term to the first power for jet correlation work is still favored.

Correlation procedures adequate for predicting a single-stream pure jet noise may not be adequate for predicting the exhaust noise of a turbofan engine. This is illustrated in figure IV-14 in which a comparison of the measured and predicted sideline OASPL is made as a function of engine thrust for engine A. The prediction method follows that developed in the previous figure and again is shown in equation form in figure IV-14. Both the fan exhaust and core jet noise are predicted and summed, assuming no interaction effect between the two exhaust streams.

At high power settings, the measured noise level appears to be some- what lower than the predicted value, due probably to a favorable interaction effect between the fan and core streams. At low power settings, the meas- ured exhaust noise exceeds that predicted using the eighth power law. This suggests that the exhaust noise from the core engine must have included noise sources other than the externally generated jet noise and that these are low-frequency noise sources. The J-65 engine core noise suppressor experiment described earlier supports this hypothesis.

Internally generated low-frequency noise, without flight effect, can add several perceived noise decibels (PNdB) to the total systems noise on en- gine A at approach power when the fan noise is fully suppressed.

It is apparent that jet noise prediction based on single-stream pure jet noise data is not fully workable for predicting turbofan engine exhaust noise because of internal noise considerations. A strictly empirical method based on correlation of exhaust noise from several turbofan engines has been de- vised at General Electric. Empirical fits of spectral and directivity data ob- tained from the CF-6, CJ-805, and TF-39 engines led to the formulation of a set of prediction equations that appears to apply quite well also to the NASA Quiet Engine. A comparison of predicted and measured spectra for engine A at the 200-foot sideline, 120 ° takeoff power is shown in fig- ure IV-15. The spectral contents above 100 hertz are accurately predicted.

Below 100 hertz strong ground-reflection nulls in the test data appear to cause some discrepancy, but the impact on the PNdB unit is insignificant.

Predicted spectrum based on the SAE method is also shown. The SAE procedure tends to overpredict the exhaust noise of turbofan engines at takeoff power.

It should be emphasized that these empirical correlation methods are appropriate for engines whose internal noise, including low-frequency sources, has not been suppressed. Such internal noise can and should be suppressed for a certain application (e. g., quiet STOL engines). Low fre- quency core noise suppression represents an area requiring further effort.

The difference between the predictions using SAE procedures and the data for the Quiet Engine is probably due, in part, to the coannular-bypass engine flow. The effect of the fan flow is to reduce the noise generated by the core flow.

The precise means of quantifying the noise reduction due to the relative velocity is not completely known. The noise of several coannular jets was measured, and the noise attenuation effect due to the bypass flow was indi- cated. As shown in figure IV-16, as the bypass velocity approaches about six-tenths of the core velocity, the noise attenuation reaches a maximum.

Beyond this value, the total noise production increases and eventually be- comes greater for the combined jet than for the single jet, the attenuation then being negative. This figure is applicable only to the specified area ratio. Further work is being done here at Lewis and by industry in order to obtain a basic understanding of this effect.

Jet Noise Suppression A common type of experimental jet noise suppressor is shown schemat- ically in figure IV-17 and consists of a multielement or mixer nozzle com- bined with an ejector. The mixer nozzle serves the function of breaking up a single large jet into many small jets. These multiple jets provide a mutual shielding effect on each other, thus reducing the noise level. The mixing of the small jets with the surrounding air results in a rapid jet veloc- ity decay so that the ejector exhaust velocity is reduced with a consequent reduction in noise level. Finally, the center frequency of the noise spec- trum of the individual small jet is much higher than that for a single large jet, making it much easier to attenuate the jet mixing noise with a relatively thin acoustic liner.

The ejector serves as a surface to which the liner can be attached as well as providing thrust augmentation during takeoff. In order to minimize performance losses in cruise, the ejector can be retracted and the mixer nozzle elements stored within the confines of the engine nacelle. Develop- ment of a variable-geometry ejector to minimize thrust losses for cruise can also be considered.

Several types of mixer nozzles that could be used with an ejector for noise suppressors are shown in figure IV-18. These nozzles include arrays of multitubes, trapezoids, and lobes among other shapes. Similar multi- element nozzles have been investigated in the past, and more recently in the SST noise-reduction program by General Electric and Boeing.

In order to evaluate further the benefits derivable from ejector-mixer nozzle suppressors, a current Lewis program in which a Boeing multitube nozzle is being tested under contract with and without acoustically lined ejectors.

In this program one multitube suppressor nozzle design is being tested at both small- and full-scale; in both cases an ejector is added, first with hard walls and then with a variety of acoustically fined walls. In addition, the same series of tests are being made with single-tube elements of the III large- and small-scale nozzles. This should help to establish better scaling relations and to optimize the design of linings.

The full-scale J-75 engine static test rig is shown in figure IV-19.

Part of the concrete pad and some of the microphone stations shown. The arrangement of the 37 tubes in the suppressor nozzle is also shown in the figure. These tubes are each 4.3 inches in diameter. The outside line was added to indicate the walls of a hexagonal ejector.

One sampling of preliminary data from the J-75 tests is shown in fig- ure IV-20. The top curve shows the sound pressure level spectrum for a standard conical nozzle at a pressure ratio of 2.4 and at the directivity angle of maximum noise (140 ° from the inlet axis). The peak sound pressure level of 119 decibels occurs at low frequencies because of the large nozzle diam- eter of 28 inches. The lower curve shows the results when the suppressor nozzle with 37 tubes is used in place of the standard nozzle. In some of the lower frequency bands, as much as a 21-decibel reduction in sound pressure level is obtained, and the remaining peak is shifted to frequencies above 1000 hertz.

In figure IV-21 the two previous curves are repeated and, in addition, the results obtained for the 37-tube suppressor nozzle with a hard-wall ejector are shown. The effect of the hard-wall ejector on sound pressure level is minor except for a slight attenuation at the high frequencies.

Sound pressure level values after application of an acoustically soft liner to the ejector are shown in figure IV-22. This liner is made of stainless-steel honeycomb sandwich panels with perforated facing sheets.

The sound pressure level was attenuated as much as 11 decibels at the de- sign center frequency of 1600 hertz. This higher range of frequencies was not affected appreciably by the suppressor nozzle alone, as was shown by the previous curves.

To interpret the performance of these suppression devices, compari- sons should be made on the basis of perceived noise levels which account for the human annoyance factors at the various frequencies. These 200-foot- radius noise measurements are extrapolated to the 0.35-nautical-mile side- line distance of FAR-36, and the perceived noise suppression values are given at the right end of the various curves in APNdB. The values are ref- erenced to the standard nozzle as zero and are for the 140 ° angle.

The suppression of 19 PNdB for the softwall ejector is not the minimum

sideline suppression, however, as shown in figure IV-23. The suppression, at 140 ° from the inlet axis, corresponds to the peak perceived noise level for the standard nozzle, but the softwall ejector peaks at about 110 °. The difference between the two peaks is the minimum sideline suppression that an observer would experience, or about 12 PNdB in this case. Effective perceived noise levels (EPNdB) penalize for the time duration of peak noise and will be worse for the standard nozzle than with the ejector, by another 1 to 2 decibels.

This minimum sideline suppression is shown in figure IV-24 in APNdB as a function of jet velocity, or pressure ratio. The suppressions are less at the lower jet velocities, but the improvement due to the lined ejector is evident down to quite low velocities, as is seen by the difference between the top curve and the other two. Jet velocities below 2000 feet per second are of interest because advanced transports will likely be operating in this jet velocity range, in order to achieve low overall noise levels.

In figure IV-25 the changes in thrust from the standard nozzle are shown as a function of jet velocity and pressure ratio. The 37-tube suppressor nozzle had about a 5-percent thrust loss, while, due to thrust augmentation of the ejector, the combination of suppressor and ejector resulted in a 2- to 6-percent gain in thrust. It should be emphasized that these data are not for flight hardware nor for flight speed conditions and are preliminary.

NASA is also conducting research on jet noise suppressors. Shown in figure IV-26 is a nozzle design that shows some promise for noise reduction at supersonic exhaust conditions. The nozzle is a convergent-divergent nozzle that operates in an overexpanded condition. The design of the nozzle is such that the jet is divided into lobes. An ejector is placed around the nozzle. The convergent portion is a standard conical nozzle. The divergent portion consist of eight plates separated by V-gutters. A step area increase exists at the throat. A low pressure exists in the base cavity formed by this area step. This low pressure causes the flow to overexpand. The flow at- taches to the plates and is divided among the plates by the V-gutters. A strong shock structure results, and the velocity rapidly decreases.

Some of these effects can be seen in figure IV-27, which shows the axial Mach number distribution along the jet centerline and along the plates. The Mach number is higher along the plate region, which indicates that the flow

has expanded into this region and, in fact, is evenconcentratedin this re-

gion. Shocksexist just upstream of the endof the V-gutters. Downstream

of these shocks, the Machnumber decreasesandthe flow becomessubsonic

in about3 or 4 diameters, andthe normally long supersonicmixing region

has beeneliminated.

The noise reduction characteristics of this nozzle are shownin fig-

ure IV-28. The decibelreduction per percent thrust loss is plotted against

nozzle pressure ratio. The particular nozzle tested was optimized for

supersonicexhaustconditions, andno noise reduction was obtainedat sub-

sonicvelocities. At t typical operating point andat a pressure ratio of 3.5,

a 12-decibelnoise reduction anda 9-percent thrust loss were obtained. The

initial tests on this nozzle were madeusing a small-scale cold facility. Re-

cently, the nozzle hasbeentested on the Lewis hot jet facility, andthe re-

sults are similar to those from the cold tests.

Jet noise from high velocity jets is of interest becauseof possible appli-

cations to advancedsupersonictransport enginedesigns. Oneimportant

questionis howaircraft forward velocity influencesthe noise suppression

characteristics. Flight tests havebeenconductedby NASAonthe F-106

airplane (fig. IV-29), which wasmodified to be poweredby the J-85 engines

for the noise test. Various types of multielement suppressorswere tested

bothunder static andflight conditions. Someof theseare shownin the fig-

ures IV-30 andIV-31. The cylindrical ejector nozzle is usedas the refer-

encebaseline nozzle for the suppressor nozzles. The suppressor nozzles

consist of a plug nozzle, 12-chute, 32-spoke, and64--spoke suppressor noz-

zles. GeneralElectric participated in the flight program andin the data

reduction.

In figure IV-32 is a comparisonof PNdBsuppressionas a function of jet

exhaustvelocity under static andunder flight conditionsfor three suppressor

nozzles (12 chute, 32 spoke, and64 spoke). The solid line is the peak-to-

peaksuppressionin PNdBunder static condition ona 300-foot sideline be-

tweenthe suppressor andthe baseline nozzle. The dashedline is the sup-

pression achievedin flight (altitude, 300-foot; aircraft speed, approx.

220knots).

Several observationsmay be drawn from the datain figure IV-32:

(1) Multielement suppressors generally havean increasing amountof sup-

pression as the number of elementsin the designincreases. Thus, the

64 spoke is better than the 32 spoke, which in turn is better than the 12 chute.

This fact is fairly well known. Unfortunately the suppressor performance loss also increases with increasing number of elements.

(2) The amount of suppression decreases with decreasing velocity. This holds true for both static and flight conditions. The probable reason for this is that reduction in mean shear, which is responsible for noise reduction, is dependent on the effectiveness of the flow induction process that takes place between adjacent elements in the multielement suppressor. As the velocity is reduced, the pumping capability is reduced, and hence suppression is less. However, the tendency for lowered noise reduction with drop in veloc- ity might have also been partly due to the presence of internal noise in the engine as external jet noise is lowered by the suppressor.

(3) In the higher velocity range, the amount of suppression under flight conditions is about comparable with that obtained under static conditions. At the lower velocities (<1500 ft/sec), noise suppression in flight appears to be significantly poorer than that obtained under static conditions when compared on the basis of equal jet exhaust velocity. No satisfactory explanation for this trend exists at present.

In figure IV-33 a typical comparison of the measured flyover PNdB as a function of time characteristics between the baseline nozzle and a 32-spoke suppressor nozzle is shown. Time duration characteristics essentially re- flect the noise directivity characteristics of the noise source as well as the airplane speed and the observer's distance. The top is the baseline nozzle, and the bottom is the 32 spoke. It is seen that the noise characteristics of the suppressor nozzle appear to rise and fall more rapidly relative to time than those of the baseline nozzle whose noise seems to linger over a longer duration. From the standpoint of effective perceived noise computation, the longer duration noise is more annoying and therefore less favorable. The duration corrections for the two nozzles are calculated and shown on the right side of the plot. On the baseline nozzle 6 PNdB is subtracted to change from PNdB to EPNdB; on the 32-spoke nozzle, 9 PNdB would be subtracted.

This particular comparison suggests, of course, that the AEPNdB associ- ated with suppression is greater than the APNdB number.

The observations made with reference to figures IV-30 and IV-31 should not be taken to represent complete generalization of the problem. Both at Lewis and at General Electric data have been obtained that do not always

conform to the trend just described, suggestingthat the problem is really

more complex. Nevertheless, the results shownin the last two figures, al-

thoughfragmentary in nature, indicate that flight effect may havea substan-

tial impact onthe suppressoracoustical performance, perhapsin a way that

cannotbe predicted without considerableadditional experiencefrom flight-

test results. The conclusiondrawn is that development of jet noise suppres-

sors for future supersonictransport enginesmust require flight tests of the

suppressordesignsduring the early stageof the development cycle.

Thrust Reverser Noise Thrust reversing (fig. IV-34) is used to shorten the landing distance for both conventional and STOL aircraft. In addition, reversing the core jets may be used in flight to steepen the approach flight path.

Lewis recently initiated studies of thrust reverser noise. Target-type reversers have so far been used because of their simplicity, and because they can reverse both circular-nozzle and slot-nozzle flows. Shown in fig- ure IV-35 are types that have been tested, namely, a V-gutter target and a semicylindrical target.

The noise directivity of a semicylindrical reverser is shown in fig- ure IV-36 as a polar plot of the OASPL as a function of the angular position 8. Looking first at the jet noise from the 2-inch circular nozzle alone, it is apparent that the jet has a pronounced directivity. The maximum OASPL is 107 decibels at an angle of 160 ° . The minimum, toward the upstream direc- tion, is 12 decibels less than the maximum. The reverser noise, in con- trast, is nearly uniform in all directions and is everywhere louder than the bare jet maximum by 1 to 6 decibels. Toward the upstream direction, the reverser is about 17 decibels louder.

Study of the noise spectra for the nozzle and reversed jets (fig. IV-37) shows that the SPL for the cylindrical reverser peaks over a broad band at higher frequencies than the bare nozzle. This peaking at higher frequencies adds to the effective perceived noise penalty, but these high frequencies at- tenuate in the atmosphere faster than those of the bare nozzle.

From the preceding discussion, it is obvious that thrust reversers gen- erate more noise than nozzles and direct it more strongly toward the critical sideline andflyover points. To present this problem in more specific terms,

the experimental data havebeenscaled up to the case of a CTOLaircraft of

300000poundsgross weight using fan-jet reversers on four Quiet Engines.

Figure IV-38 showsthe perceived noise level distribution alongthe 0.35-

nautical mile sideline, not including extra groundattenuation. It is apparent

that achievingnoise levels belowabout 100PNdBfor this examplewill be

difficult.

Other methodsfor reducingnoise from reversers are being studied.

For example, cascadereversers may be quieter thantarget reversers and

perhaps canmakeuse of acoustically treated surfaces. In addition, the use

of acoustical doors or shields to redirect reverser noise awayfrom the side-

lines may be feasible.

CONCLUDING REMARKS The main conclusions reached by this discussion may be summarized as follows: 1. At low subsonic jet exhaust velocities, jet noise varies as the veloc- ity to the eighth power.

2. At high subsonic exhaust velocities (above 2000 ft/sec), we reaffirm that jet noise approaches a variation with velocity to the third power.

3. Use of the jet density squared, as in SAE AIR 876, overcorrects den- sity effects on jet noise.

4. Subsonic and supersonic jet noise levels can be predicted from a sim- ple empirical correlation.

5. Although the present Quiet Engine jet noise can be estimated within its geometry constraints, the understanding of such details as noise mecha- nisms and geometry considerations for fan-jet engines continues to require further work.

6. Use of multielement nozzles and acoustically lined ejectors can sig- nificantly suppress jet noise at intermediate jet exhaust velocities. Further work is required at both low exhaust velocities (less than 1000 ft/sec) and high exhaust velocities (greater than 3000 ft/sec) to achieve needed suppres- sion levels.

7. Forward velocity effects in flight can cause some attenuation reduc- tions with suppressors. Thus, wind tunnel or flight tests should be made before committing suppressor configurations to the hardware stage.

8. In view of proposed reductions in the FAR-36 noise levels, thrust reversers can present new noise problem areas, particularly for advanced CTOL and STOL aircraft.

ll8 JET NOISE EXPERIENCE 200FEETSIDELINE 180- • .-_>.-.

- ,-" iLJ ENGINE PEAK EXPERIENCE-'_SAE AIR-816 SIDELINE 140 _MODEL JET .,_" OASPL - \EXPER_ ._'_" I0 LOG p_A 120 - \ 100 _'EXTRAPOLATED SAE /.,4"IP'"-QUIET FAN 1 DATA 80 I I I 400 1000 2000 4000 JETVELOCITY,FIISEC CS-63310 Figure IV-I FAN JET NOISE 200FEETSIDELINE DO n FAN 0 D o B FULL SCALE oC PEAK SIDELINE o B } SCALEMODEL ^_ 80- OASPL " C j_L CURVE FiT -1OLOG A 70- I// I I I I I J 60, 400 500 000 700 800 9OO Iu00 JET VELOCITY, FTISEC CS-633 ]. 1 Figure IV-2 J-65 ENGINE WITH SUPPRESSORS INSTALLED Figure IV-3 CS- 632(_ 1 J-65 EXHAUST SUPPRESSOR CS-63259 Figure IV-4 LEVEL J-65 SPECTRA OF SOUND POWER HARD AND LINED DUCTSAT 50%SPEED 140-- o HARD DUCT [] LINED DUCT SOUND POWER 1_ LEVEL dB I I I I I I I 1--50_ 100 200 400 1000 2000 _ 10 000 FREQUENCY,Hz Figure IV-5 cs-63z47 J-65 NORMALIZED MAXIMUM OASPL 200 FT SIDEUNE 150-- o HARD DUCT []LINEDDUCT 140-- PEAK SIDELINE 130-- OASPL 2 -I0LOG(pjA) o o_ =v8 120-- 0 []

°J

[] I I I I I 11_ 0 n 6O0 8O0 1000 1200 1400 1600 JET VELOCITY,FTISEC C5-63307 Figure IV-6 HOT JET FACILITY £ CS-& :,258 Figu re IV-7 COMPARISON OF SUBSONIC JET NOISE DATA WITH SAE PREDICTION TEMPERATURE, 160 -- OR 150 "" 140 /_./1460 PEAK- 130 SAE AIR 876/'--. _..'/" SIDELINE j,- OASPL " / 660 _ 1 O p2,A 120 - o NASA, FAN log J -- ,/ ./ -- --NACA 110 F (-,,_#/ AMBIENT ----- BOEING L_A{K>, " 'eBS>' .... BOEING _ .- _.,n:_O_M BIEN T ..... BOEING 901" I I I I 500 1000 1500 2000 2500 JET VELOCITY, FTISEC CS-63243 Figure IV-8 SUBSONIC JET PEAK-SIDELINE OASPL VARIATION WITH JET VELOCITY 120- 110 - S REFERENCE PEAK- 11.0 NASA o 1150-1650 SIDELINE 2.87 BOEING 90' o_ o l_O OASPL 2.87 BOEING . _f ¢ 2400 -10 log A 13.0 NASA 80 r"[t_C z_ AMB.

2.06 NASA NASA 4.0

. _" o o60

72. O* NASA

7o ,_ _'"F_ D

I _-_- v AMB.

60 I l I l I [ _= " " , I J .8 1.0 1.5 2 2.5x103 .4 .6 JET VELOCITY, FTISEC CS-63254 Figure IV-9 COMPARISON OF SUPERSONIC JET NOISE WITH SUBSONIC JET NOISE O0 0 O0

13° F %

12o - o °_ o%'/ REFERENCE _'I_IAM., JET PEAK- SIDELINE Ii0-- OR OASPL oS IN. TEMP., -I0 log A NACA o j o 4 660 BOEING [] 2.87 1400 BOEING I00 -j <> 2.87 2460 / '-SUBSONIC DATA CURVE

9o I 1 1 1 1

I 1.5 2 2.5 3 3.5xi03 JETVELOCITY,FTISEC CS-63306 Figure IV-IO CORRELATION OF PEAK-SIDELINE OASPL 120- REFERENCE PEAK SIDELINE OASPL j o 660 4.0 NACA -I0 LOGA M3F' ,._ o 1460 2.87 BOEING 10(3- / o 2460 2.87 BOEING TAILED SYMBOLS DENOTE SUPERSONICJET 9o I I I I I i0-1 10 0 lo I _0 2 10 3 10 4 CS-63255 Figure IV-11 QUIET ENGINE A NOISE AT TAKEOFF SPEED o FAN A 120° [] ENGINE A SOUND PRESSURE LEVEL I I 1 I I I I 4O 100 2OO 4O0 1000 2OOO 4O0O 10000 FREQUENCY,Hz Figure IV-12 CS-63248 CORRELATION OF FAN JET NOISE FROM ALL SUPPRESSED QEP FANS TESTED II0-- o FULLSCALE FAN A [] FULLSCALE FAN B / _" FULLSCALE FAN C 100 _ <> SCALE MODEL FAN B o SCALE MODEL FAN C _ 200'SIDELINE MAX ^ OASPL - 10LOG pjA w __°//" OASPL = "_133.5 + 80 LOG V + / 10LOG pi A J

7o I J I l I i I

400 500 600 800 1000 FAN EXHAUST VELOCITY,FTISEC CS-63339 Figure IV-13 MEASURED AND PREDICTED JET NOISE FOR ENGINE A 110 -- I00 -- 7_T A 200FT SIDELINE ENGINE A MAX OASPL - I0 LOGpjA, dB __ "-PREDICTION BASEDON: i I OASPL = -133.5 + 80 LOG V + 10 LOG pjA

I I I I

60 80 100 THRUST, PERCENT CS-63338 Figure IV-14

ENGINE EXHAUSTNOISEPREDICTION

INCLUDES JET AND CORE NOISE; ENGINE A; 1.20°; TAKEOFF POWER

100--

---..... SAE 870 PREDICTION

O O "--...

90 --0__'_

200 FT SIDELINE

SOUND PRESSURE

" 0 - '_GE PREDICTION -"'b,._ _'- LEVEL, _TEST DATA (MINZNER METHOD_ -

dB

80 ENGINE A

I I I Ill I I 1 i I i I,I

I00 I000 FREQUENCY, Hz CS-63342 Figure IV-15 CO-ANNULAR JET NOISE ATTENUATION B O SOUND O PRESSURE LEVEL AI-rENUATION, ABp _ AdB -5 I 1 I I I -1O O • 25 . .50 .75 1.00 I. 25 VELOCITYRATIO, VBp/V c Figure IV-1.6 CS-63309 MIXER-TYPE SUPPRESSOR NOZZLE WITH EJECTOR CS-63337 Figure IV-17 TYPICAL MULTIELEMENT NOZZLES SQUARE-EN DED TRA PEZOI DAL MULTITU BE NOZZLE ANNULAR BY PA SS-TY PE ROUND-ENDED TRAPEZOIDAL NOZZLE WITH 8 CORE TUBES NOZZLE. ALTERNATE LOBES CANTED 100 OUTWARD FROM CS-633] 3 NOZZLE CENTERLINE.

Figure IV-18 NOISE TESTS WITH J 75 ENGINE SUPPRESSOR NOZZLELINEDEJECTOR F EXHAUSTNOZZLE r MUFFLER AROUND , & EJECTOR / ENGINE & INLET / REMOTE TEST INSTALLATION HEXAGONAL EJECTOR BELLMOUTH C5-b_]74 31-TUBE SUPPRESSOR NOZZLE ARRAY Figure IV-19 J-75 ENGINE WITH 57-TUBE SUPPRESSOR NOZZLE PRESSURE RATIO,2.4 281N.

120- NOZZLE-_ K 4.3 IN.

STANDARD ", __T.

/p. j-_ - / % / " _ JL

_..," "- / E I_"

SOUND PRESSURE 110, "_" ,'F _ I_ LEVEL AT 200FT ,, , SOR NOZZLE 90 i I I I I I 1 50 100 200 500 10O02000 50O010000 1/3-OCTAVE BAND CENTER FREQUENCY, Hz Figure IV-20 cs-63246 J-75 ENGINE WITH SUPPRESSOR AND HARD WALL EJECTOR PRESSURERATIO, 2.4 NOZZLE",, IN.

110 _ 37-TUBE """ SOUND PRESSURE SUPPRESSOR-, .--, -...

LEVELAT 200 FT ,_, "_.2 -.

& 140 o ANGLE, dB

].oo'_ _ -_

9O I I I I I I I % 100 200 500 1000 2000 5000 10000 1/3-OCTAVEBAND CENTER FREQUENCY,Hz CS-63250 Figure IV-21 J-75 ENGINE WITH SUPPRESSOR AND SOFT WALL EJECTOR PRESSURERATIO, 2.4 130 6.3% OPENAREA-, 2. I IN.

STANDARD F_T

• ,2o NOZZ, , I

I ", 3_ SOUND PRESSURE ii0 -;/ 37-TUBE -.... /_PNdB LEVELAT 200 Fr "-. ,' AT O. 35 & 1400ANGLE, SUPPRESSOR -_.--.#:... . / • " ..... - N.MI.

liB i00 ( 90 - SUPPRESSOR WITH / u_, 14 HARD WALL EJECTOR--" Xo 19 8o I I I I I I I 50 100 200 500 1000 2000 5000 10000 CS-63312 1/3-OCTAVEBAND CENTER FREQUENCY,Hz Figure IV-22 VARIATION OF PNL ALONG 0..35NAUTICAL MILE SIDELINE J-75ENGINE; PRESSURERATIO, 2.4 /- MINIMUM SIDELINE 110-- / SUPPRESSION I 100 I// ANDARDNOZZLE 12dB PERCEIVED I). 19dB NOISE LEVEL, PNdB W-TUBE SUPPRESSOR 90 --}- 10dB I _ WITH SOFTWALL EJECTOR I I I I I "°1 -1000 O 1000 2000 3000 4000 DISTANCEALONG 0.35-N. MI. SIDELINE, FT I I I I I I 6O 80 100 120 140 155 ANGLEFROMJETINLET, DEG CS-633Z6 Figure IV-23 MINIMUM PNL SUPPRESSION AT 0.35 NAUTICAL MILE SIDELINE J-Y5 ENGINE WITH 37 TUBE SUPPRESSOR 12[- SUPPRESSOR WITH 10_ SOFTWALL EJECTOR..cr _ SUPPRESSOR PERCEIVED NOISE LEVEL SUPPRESSION, A PNdB 4_/ ./_._ WITH HARD F___/ WALL EJECTOR _1_ I l i I I 1000 1200 1400 1600 1800 2000 IDEAL JETVELOCITY, V i, FT/SEC I 1 I I I 1.4 1.6 2.0 2.2 2.4 PRESSURERATIO CS-63328 Figure IV-24 THRUST ON J-75 ENGINE RELATED CHANGES IN TO STANDARD NOZZLE -6- O -4-- 3/-TUBE SUPPRESSOR ALONE CHANGE -2-- IN 0 THRUST, .-.--.43 PERCENT +2-- +4- WITHEJECTOR _i 1 31-TUBE SUPPRESSOR I I +6 1000 1200 1400 1600 1800 2000 IDEAL JETVELOCITY, Vj, FTISEC I I I I I 1.4 1.6 2.0 2.2 2.4 PRESSURERATIO CS-63327 Fiqure IV-25 OVEREXPANDED MULTILOBED CONVERGENT DIVERGENT NOZZLE CS-63260 Figure IV-26 MACH NUMBER DISTRIBUTION ALONG PRIMARY PLATE AND NOZZLE CENTERLINE 2. 4 [-- o CENTERLINE I MACH NUMBER ,4,

[ " vi' q

I t

0 ! 2 3 4 AXIAL DISTANCE FROMTHROAT IN DIAMETERS, x/D Figure IV-27 CS-63237 PERFORMANCE CHARACTERISTICS OF OVEREXPANDED MULTILOBED CONVERGENT-DIVERGENT NOZZLE 2°0 V • 12dB REDUCTION /d' I. 5 -- WITH 99_ REDUCTION IN dB PER PERCENT 1.0 -- THRUST LOSS • 5 --

l _...a' I I

0 I 2 3 4 PRESSURE RATIO CS-63Z39 Figure IV-28 FLIGHT JET-NOISE RESEARCH CS-63257 Figure IV-29 F-106 NOISE SUPPRESSION NOZZLES CS-63344 Figure IV-30 CYLINDRICAL EJECTOR NOZZLE CS-63345 Figure IV-31 JET NOISE SUPPRESSION-STATIC AND FLIGHT REFERENCED TO CYLINDRICALEJECTOR NOZZLE,_O FT SIDELINE OR ALTITUDE

o t- /32 spoKE

PERCEIVED I I i J i r_ I J NOISE LEVEL I_ 1500 2000 2500l_ 1500 2000 2500 SUPPRESSION,A PNdB I0 F /_ --o-- FUG_ --o-- STATIC 50_P// _SP_E I i I I JETEXHAUSTVELOCITY, cs-63343 SEC Figure IV-32 TIME DURATION CHARACTERISTICS 300 FT ALTITUDEFLYOVER;MAXIMUM THRUST REFERENCE NOZZLE DURATIONCORRECTION (EPNL - PNLT) RELATIVE 10 dB PNLT I -6 dB (LESS FAVORABLE)

_-- lSEC

32 SPOKE PLUG NOZZLE RELATIVE PNLT 10 dB -g dB (MOREFAVORABLE)

L-- 1 SEC A

CS-63340 Figure IV-33 THRUST REVERSERS CS-63252 Fig u reIV-34 TARGET-TYPE THRUST REVERSERS SEMICYLINDER V-GUI-rER CS-63253 Figure IV-35 THRUST REVERSAL NOISE DIRECTIVITY 2-INCH NOZZLEAND CYLINDRICAL REVERSER; JETVELOCITY, 960 FIISEC 270° RALL SPL RADIUS), dB

no / ' [:_'" i i ',FLOW. l _,_o

, u 120 II0.-- 100i_90 I00 II _- _ _-_ -_ _ '0 120 *_u t \ X_'NOZZCE_II<.._/-'1'7/ / \ \ k ONLY-" i-_--_'-_'// f ANGLE FROM_ _....._ ;i;_/"-REFER SER

",,. 90°

CS-_3236 Figure IV-N PRESSURE LEVEL SPECTRA SOUND FOR REVERSER AND NOZZLE AT ANGLE OF MAXIMUM SIDELINE NOISE CYLINDRICAL I00 SOUND REVE____w_ _ PRESSURE j:_o/_ .cr,._-_'Q" "_--o- -c}..o_o.._ LEVEL 90 - /2_ - ,p'_7. - - ---o-.a.,_ AT 10H, dB 8O _o_/p..._,- x_ NOZZLE ONLY ".a I I I I I I 200 .500 1000 2000 5000 10000 20000 THIRD-OCTAVE-BAND CENTER FREQUENCY, Hz CS-63244 FigureIV-37 THRUST REVERSAL NOISE FOR 5OO OOO LB CLASS TRANSPORT QUIETFANENGINES;0.35-N. MI. SIDELINE PERCEIVED NOISE LEVEL, 90 PNdB 8O I I I I I I I J -8 -6 -4 -2 O 2 4 6x 103 -10 DISTANCEBEHINDPLANEON 0.35-NAUTICAL-MILE SIDELINE, FT Figure IV-38 cs-63238 THE QUIET ENGINE PROGRAM Vo Newell D. Sanders The first objective of the Quiet Engine program is to demonstrate noise reduction technology in an engine. At the inception of the program, in the mid-1960's, the goal was a 15- to 20-decibel reduction relative to the noises produced by the DC-8 and 707 airplane engines.

The first obstacle to reaching that goal was the high jet noise at takeoff.

Recall that the Acoustically Treated NaceUe Program with DC-8 and 707 air- planes produced only minor noise gains at takeoff because of the jet noise.

The most effective way of obtaining very low jet noise is to choose an engine with low jet velocity. And low jet velocity is achieved with a high-bypass- ratio engine.

Estimates of noise reductions accompanying increased bypass ratios are shown in figure V-1. This figure was first prepared by NACA in 1954 and was used to advocate the development of low-noise bypass engines in the United States. At that time the only bypass engine was the British Con- way. These early calculations were confirmed and extended to higher by- pass ratios by Pratt & Whitney as shown in figure V-2. The JT3D, which powers 707 and DC-8 airplanes, the JT9D, and an experimental engine are shown on the graph. At that time the JT9D design was not firm. This graph indicates that a bypass ratio of 5 is required to bring the noise level 15 dec- ibels below that of the JT3D fan technology curve.

The fan noise (fig. V-2) increases with increasing bypass ratio and dominates above a bypass ratio of approximately 1.0. The reduction of fan noise to the level of the jet noise requires the application of several tech- niques, some of which are discussed in papers II and HI. One of the tech- niques is to reduce fan speed as shown here (fig. V-2). Much of this de- crease resulted from other fan changes such as the elimination of inlet guide vanes, the elimination of the second fan stage, and increased spacing between the rotor and the stator. Reducing fan speed is beneficial, however, and if the tip speed is lowered below sonic speed, shock noises can be avoided.

In 1967, under NASA contract, Allison and Pratt & Whitney studied quiet engine designs incorporating high bypass ratios for low jet noise and low-speed, low-noise fans. These studies indicated that, at bypass ratios near 5 or 6 and with low noise fans, noise reductions in the order of 15 to 20 decibels relative to DC-8 and 707 airplane engines were possible.

Concurrently, under NASA contract, the Douglas Aircraft Company studied the application of these engine designs to the DC-8 airplane. The engine characteristics resulting from the study are shown in table V- 1.

TABLE V-1. - RESULTS OF ENGINE AND AIRPLANE STUDIES New engine JT3D Bypass ratio 5.0 1.41 Cruise thrust, (35 000 ft, M = 0.82), lb Takeoff thrust, lb 23 350 18 000 Nacelle weight, lb Thrust specific fuel consumption, lb/hr-lb 0.61 0. 830 Retrofit cost, S/airplane 5 000 000 The JT3D characteristics are shown for comparison. The study showed that, although the Quiet Engine nacelle was nearly 1500 pounds heavier than the JT3D nacelle, the improved fuel consumption more than offset the effects of the weight increase, and the airplane range was extended. In addition, the payload was not changed, the takeoff roll was shortened, and the initial cruise altitude was raised. Everything looked favorable except the cost.

The estimated retrofit cost was $5 million per airplane. For new airplanes, the situation is different. The cost of the quiet engine is favorable in com- petition with other, potential new high-bypass engines.

Following the Douglas, Pratt & Whitney, and Allison studies the Quiet Engine project was initiated.

The General Electric Company in 1968 entered into a fixed-price con- tract to build two experimental Quiet Engines using derated CF6 cores. One engine has a low-speed fan running at a tip speed of 1160 feet per second, and the other engine has a fan running at the high speed of 1550 feet per sec- ond. The two engines are expected to show the relative advantages of fans operating at low tip speeds with high lift coefficients in comparison with fans operating at high tip speeds with low lift coefficients.

The low-speed engine has been completed, tested by General Electric, delivered to Lewis, and tested at Lewis. The high-speed engine has been assembled and is now being tested by General Electric.

Sound absorbing ducts for the fan inlet and outlet along with other nacelle parts have been built by the Boeing Company for the Quiet Engine. This nacelle has been assembled with the engine and is now being tested at Lewis.

Design details of the Quiet Engine and its performance are described in subsequent papers.

NOISE REDUCTION FOR A BYPASS ENGINE

Z

TOTAL POWER LEVEL 12-- REDUCTION, dB 16-- UNMIXED JET MIXED JET

24 I I 1 1 I I

I

0 .4 .8 1.2 1.6 2.0 2.4 2.8 BYPASS cs-63933 MASS FLOW RATIO, PRIMARY Figure V-I EFFECT OF BYPASS RATIO ON TAKEOFF NOISE EQUAL TAKEOFF THRUST _, /-JT3D FAN TECHNOLOGY /(1400 FT/SEC TIP SPEED) -.- / .....-- .......

-10 RELATIVE ____- /JTgD FAN TECHNOLOGY NOISE -2O LEVEL, PNdB -- JT3D-3B _ JT!ZlE _ / -3O / JT9D "EXHAUST NOISE -401

I I I I I I I I

1 2 3 4 5 6 7 8 BYPASS RATIO Figure V-2 VI. QUIET ENGINE DESIGN HIGHLIGHTS Bernard L. Koff* This discussion highlights both components and engines designed, built, and tested by General Electric for the Quiet Engine Program. Figure VI-1 shows the overall development plan from component to turbofan engine testing. Two low-speed fans (cruise tip speed, 1160 ft/sec) and one higher speed fan (cruise tip speed, 1550 ft/sec) with lower aerodynamic loading were designed and manufactured. All three fans were built, instrumented, and tested in a component vehicle at the General Electric Large Fan Test Facility at Lynn, Massachusetts. Aerodynamic performance and operating characteristics, such as blade vibration, were evaluated with and without inlet distortion. Three screen patterns were used to simulate inlet- pressure distortion resulting from flow separation experienced in typical aircraft installations. After aerodynamic performance evaluation, all three fans were shipped to NASA-Lewis for acoustic testing. After testing fans A and B, fan A was selected for the low-speed turbofan engine.

Two turbofan engines (low-speed A and high-speed C) were built and shipped to Peebles, Ohio, for both acoustic and performance testing.

Flight-type inlet configurations, such as the thick lip DC-10 and the thin lip with blow-in doors, were evaluated against the standard bellmouth. The ef- fect of acoustic treated splitters, as well as wall treatment, was evaluated for both the fan inlet and exhaust duct. The discharge of the core engine was also treated to suppress the turbine noise. After evaluation, engine A was shipped to NASA-Lewis for further acoustic testing. Engine C is cur- rently undergoing performance and acoustic testing at Peebles, Ohio.

In figure VI-2 the turbofan demonstrators (low-speed A and high- speed C), both using the proven CF6 core engine operating at reduced speed and turbine temperature, are illustrated and compared.

Both fans are titanium, with fan C designed for lower aspect ratio to utilize advanced lightweight composites. As a result, fan C has longer *General Electric Company.

chord blading andis also slightly longer than fan A becauseof the two-chord

spacingbetweenthe rotor andbypassoutlet guidevanes.

EngineA hasa four-stage low-pressure turbine straddle-mountedbe-

tweenthe turbine midframe and rearframe. Engine C has a close coupled, highly loaded two-stage turbine, which is cantilever-mounted on the low- pressure shaft. This mounting substantially reduces the amount of surface area requiring cooling and also eliminates the need for a rear sump.

Figure VI-3 shows a Wpical performance map of bypass pressure ratio against fan corrected flow obtained in the Lynn Test Facility. The perform- ance of fan A in this case is shown at the design point at 100 percent cor- rected speed, including the stall line and efficiency islands. The pressure ratio, airflow, efficiency, and stall margin are tabulated and show excellent performance for this fan at high specific flow (88.3 percent efficiency at 42.5 lb/sec-ft 2 of annulus area).

A comparison of the three fans (two low speed and one high speed) is shown in figure VI-4. The fan A rotor has 40 tip shrouded titanium blades with two seals in the tip shroud and has the higher aspect ratio.

Fans B and C both have 26 titanium blades without shrouds. Fan C rotor blades are longer and also have the highest specific flow based on frontal area, as a result of having a lower radius ratio (0.36 compared with 0.47). The higher blade speed permits using a lower radius ratio without excessive aerodynamic loading. Fan C also has a higher bypass pressure ratio, which results in greater specific thrust (lb thrust/lb airflow) and therefore requires a smaller fan tip diameter and lower flow to produce the same engine thrust (68.3 in. against 73.4 in. and 915 lb/sec against 980 lb/ sec for the other fans).

All three fans have approximately the same vane-blade ratio (~21 ) and identical rotor to stator spacing in both bypass and core flow paths.

CORE ENGINE DESCRIPTION The single-spool 16-stage core compressor with variable inlet guide vanes and six variable stators is shown in figure VI-5. The design airflow is 139 pounds per second at 16.8 pressure ratio. The compressor operates

at 104 pounds per second flow and pressure ratio of 12.0 in the Quiet Engine

demonstrators. The horizontal split casing permits individual blade re- placement in the rotor spool. The eighth, ninth, and thirteenth stage bleed manifolds are shown and used for cabin, sump, and turbine cooling.

Figure V1-6 shows the annular stacked ring combustor and combined rear frame and outer casing. Thirty high-pressure atomizing dual-fuel nozzles are used for high- and low-range operation. High dome flow is also used for low smoke. The combustor has a design temperature rise of 1600 ° F and operates at a temperature rise of 1380 ° F in the Quiet Engine.

The two-stage high-pressure turbine is shown in figure VI-7. This turbine has moderate aerodynamic loading, a work extraction of approxi- mately 90 Btu per pound per stage and high efficiency (91.5 percent). The first-stage blade is cooled by convection with impingement and film cooling on the leading edge. The design turbine rotor inlet temperature is 2370 ° F and runs derated to 1970 ° F in the Quiet Engine demonstrators.

Figure VI-8 shows the core engine assembly prior to going horizontal at main engine buildup.

LOW-PRESSURE COMPONENTS - ENGINE A The low-pressure components for engine A use the tip shrouded fan and a four-stage modified CF6 turbine with the last-stage removed. The turbine was rematched by closing the stage 1 nozzle diaphragm area 6 percent.

The 40-blade-tip shrouded rotor assembly is shown in figure VI-9 and is a proven design concept introduced at General Electric in 1965. This fan configuration has high efficiency and rugged blading with excellent aero- elastic stability and low overall vibratory characteristics. The interlocks are located as shown and hold the blade tip sections in the design position during unsteady flow conditions. Individual replacement is made by remov- ing a shim below the dovetail and displacing the blade radially inward to disengage the shroud interlocks.

Figure VI-10 shows the fan A stator with bypass and core outlet guide vanes and splitter. The flow path surfaces are lined with acoustic panels using an aluminum perforated sheet and multiple degree-of-freedom diamond core treatment. The required porosity of the double diamond core is pro- vided by slots rather than holes to allow flexibility for compound forming.

This also eliminates the need for individual core tooling on each panel.

Drainage of condensate or fuel is also provided by circumferentially oriented channels formed by the core material combined with these forming slots and panel edge holes. An abradable material is used over the rotor tip shroud seals to allow the seals to wear in without overheating.

The tip shrounded four-stage low-pressure turbine rotor is shown in figure VI-11. This turbine is lightly loaded with a work extraction of 90 Btu per pound and a design point efficiency of 91.8 percent.

Figure VI-12 shows the main engine buildup with core, variable com- pressor stator actuation system, piping, and low-pressure fan and fan tur- bine components. The engine is shown in figure VI-13, mounted in the overhead thrust stand with the bellmouth inlet. An aft end view showing the separated fan and core exhaust is shown in figure VI-14 with two radial tra- versing acoustic probes at the entrance and the exit of the fan exhaust duct.

Figure VI-15 shows the performance test setup with the inlet rakes ahead of the fan and the traversing core smoke probe. The core smoke was measured at an SAE smoke number of 7 and well below the threshold of visibility, which is approximately an SAE smoke number of 25.

Figure VI-16 shows the engine in the test stand and the far field micro- phone test setup. There are sixteen 40-foot microphone towers in a 150 ° arc. The engine is shown in figure VI-17 with the thick-lip DC-10 type inlet and the acoustic probes in the fan exhaust duct; figure VI-18 shows the thin lip inlet with simulated blow-in doors. This inlet is 12 percent smaller in frontal area and 11 percent shorter than the thick lip inlet. The recovery at takeoff for this inlet was approximately 2 percent lower and at a somewhat higher radial distortion than the thick lip inlet.

LOW-PRESSURE COMPONENTS - ENGINE C The high-speed C engine was also built up for testing at Peebles and uses fan C with the new highly loaded two-stage low-pressure turbine. Fig- ure VI-19 shows the low radius ratio fan C assembly. This fan demonstrated good vibration characteristics with various kinds of inlet distortion without shrouds. The treated stator and frame incorporating acoustic panels is shown in figure Vl-20 with bypass and tandem core outletguide vanes to accommodate the high aerodynamic loading. Figure VI-21 shows the low- pressure combination turbine stator and frame that close couples the low- pressure and high-pressure turbines.

Figure VI-22 shows the two-stage highly loaded low-pressure turbine with 90- to 91-percent efficiencybased on recent engine test data. This turbine has equivalent efficiencyto the four-stage turbine with lessthan one- half the number of airfoils. The first stage of this two-stage turbine de- monstrated a high pitchlineloading (2JgAh/Up 2) of 1.6 and did most of the work extraction, while the second stage was lightlyloaded to reduce dis- charge swirl.

The main engine C buildup is shown in figure VI-23 before shipment to test, and figure VI-24 shows the high-speed engine in the test stand during performance testing.

SUPPERSUPPRESSEDINSTALLATIONS In addition to the basic performance and acoustic evaluation of the en- gines, a number of tests are being made with configurations to further sup- press engine noise.

Figure VI-25 illustrates the suppressed configurations that have been run on engine A and those planned for evaluation on engine C. The base configuration incorporates acoustic treatment extending from ahead of the rotor through the frame. The supersuppressed configuration tested includes an extended inlet with full-length treatment, a three-ring treated splitter in the inlet, extended aft treatment in the fan exhaust duct (inner and outer flow path), and a treated exhaust duct splitter. Treatment in both inner- and outer-core engine exhaust nozzles is also used to suppress the turbine noise. A number of tests with various combinations of treatment were made to identify the effectiveness of the separate sections of the treatment.

Some of these data will be discussed in following papers.

The hardware for the engine C supersuppressed configuration will be

evaluated this summer. A contoured bellmouth more representative of a

flight-type inlet will be used. A four-ring acoustic splitter will also be evaluated separately. A specially designed wall treatment to suppress the multiple pure tones characteristic of the high-speed fan will also be eval- uated. The aft treatment will include a contoured splitter with the duct Mach number reduced to improve the effectiveness of the treatment and to minimize flow scrubbing noise. Core engine exhaust nozzle treatment to suppress turbine noise is also used.

Planned engine tests will include a number of combinations that will permit evaluation of the effectiveness of the separate elements. Figure VI-26 shows the engine A on test with the acoustic treated splitters and outer casing. Figure VI-27 shows the aft acoustic treated splitter in the fan exhaust duct including the inner and outer flow path. The two rows of double-layer sheet/honeycomb acoustic panels for both inner and outer flow paths used in the core engine exhaust nozzle are shown in figure VI-28.

The acoustic and engine performance results of these engine configura- tions will be presented in subsequent papers.

SUMMARY Summarizing the component and engine testing: Three high perform- ance fans were designed, built, and tested, accumulating 444 hours. All three fans are both aerodynamically and mechanically suitable for direct in- corporation into engine applications. Two turbofan demonstrators to eval- uate both low- and high-speed fan systems were also designed, built, and tested (accumulating 158 hr). All components and engine systems are demonstrating high reliability.

These components have demonstrated advanced state-of-the-art in acoustics, aerodynamics, and mechanical design. The next step is to in- corporate this demonstrated technology and additional improvements into advanced flight systems with even lower noise and better performance.

NASA-GE QUIET ENGINE PROGRAM TURBOFAN ENGINE DEVELOPMENT ACOUSTIC/PERFORMANCE TESTING o INLET TYPES o FAN INLET/EXHAUST TREATMENT o CORE TREATMENT FAN A - TiP SHROUDED I 1160 It/sac U T I 1,5 P/P | FAN A ENGINE .465 R/R t NASA ENGINE ACOUSTIC TESTIN( NASA-LEWIS LYNN COMPONENT TEST AERODYNAMIC PERFORMANCE ACOUSTIC TESTING FAN B - UNSHROUDED 1160 ft,/tec UT 1.5 P/P .465 R/R i FAN C ENGINE FAN C - UNSHROUOED 1550 It/see UT 1.6 WP .36 R/R Figure VI-I NASA-GE QUIET ENGINES _-_-- '_ ENGINEA - LOW SPEED i_i, ix i , =--, , i,,' ENGINE C - HIGH SPEED Figure Vl-2 FAN A AERODYNAMIC PERFORMANCE i.71 /' PRESSURE RATIO 1.5 AIR FLOW 977 Ib/sec EFFICIENCY 88.3% STALL MARGIN 17% 1.4 i SPEED AIR FLOW/ANNULUS AREA 42.5 Ib/sec/ft 2 1.3 STALL _/;////_/.,./,/ FAN BYPASS PRESSURE _ EFFY."_. _2_-_z_i_ -- RATIO 1.1 ! _-<_ -_- ._o '/u l.ol _ _q...............

200 300 400 500 600 700 800 900 1000 TOTAL FAN FLOW, Ib/sec Figure Vl-3 QUIET ENGINE FAN DESIGN FANS A B C • AIRFLOI_. Iblsec 977 983 915 • BYPASS - PRESSURE RATIO l.S I.S 1.6 - EFFICIENCY, _, 88.3 87.0 84,0 • _,'_'_@I'6/,A F Itl,'Su.c-tt 2 33.3 33.5 36. D • ROTOR TiP DIA, in. 13.4 13.4 68,3 • BLADE/VANE RATIO 40190 26160 26160 • ROTOR " STATOR SPACING - BYPASS -- 2 ROTOR CHORDS -- - CORE -- 1-114 ROTOR CHORDS -- Figure Vl-4 COMPRESSOR STATOR ASSEMBLY 16 STG COMPRESSOR ROTOR Figure VI-5 COMBUSTION SECTION COMBUSTOR REkR FRkME AND OUI£R CkSING Figure Vl-6 TWO-STAGE HIGH-PRESSURE TURBINE 1ST STAGE NOZZLE DIAPHRAGM ROTOR ASSEMBLY L_IDSTAGE NOZZLE DIAPHRAGM Fig u re Vl-7 CORE ENGINE ASSEMBLY

i

Figu re VI-8 TIP SHROUDED FAN A ROTOR ASSEMBLY Figure VI-9 ACOUSTIC TREATED FAN A STATOR WITH BYPASS AND CORE OUTLET GUIDE VANES _r PERFORATED SHEET/DOUBLE DIAMOND MDO_CORE ACOUSTIC TREATr_ENT Figure VI-10

FOUR-STAGE LOW-PRESSURE TURBINE

ROTOR ASSEMBLY

Figure VI-11

MAINENGINE A ASSEMBLY

Figure VI-12

FRONT END OF ENGINE A IN TEST STAND AT PEEBLES, OHIO I Figure VI-13 OHIO AFT END VIEW OF ENGINE A IN TEST STAND AT PEEBLES Figu re VI-14 A PERFORMANCE TEST SETUP INCLUDING TRAVERSING

ENGINE CORE

SMOKE PROBE Figure VI-15 FAR FIELD ACOUSTIC MICROPHONE SETUP • !

Figure VI-16

ENGINE A THICKLIP INLET CONFIGURATION

4!i Figure Vl-17 ENGINE A THIN LIP BLOW-IN DOOR INLET CONFIGURATION Figure Vl-18 FAN C ROTOR ASSEMBLY Figure VI-19 ACOUSTIC TREATED FAN C STATOR WITH BYPASS AND TANDEM CORE OUTLET GUIDE VANES P

4 '

Figure VI-20

LOW-

TURBINE C MIDFRAME ANDFIRST-STAGE

PRESSURE NOZZLE DIAPHRAGM

/ Figure VI-21 TWO-STAGE LOW-PRESSURE TURBINE C ROTOR ASSEMBLY Figure VI-22

MAINENGINE C ASSEMBLY

Figure VI-23

ENGINE C TEST STAND AT PEEBLES, OHIO O"I Figure VI-24 NASA-GE SIMULATED SUPERSUPPRESSED ENGINE INSTALLATIONS ENGINEA Ill -_.__ I, .

ENGINEC Figure VI-25 NASA-BOEING INLET WITH ACOUSTICALLY TREATED SPLITTERS Figure VI-26 ENGINE A FAN EXHAUST DUCT SPLITTER WITH ACOUSTIC TREATMENT <INSTALLED) Figure VZ-27 ENGINE A CORE EXHAUST NOZZLE DOUBLE LAYER PERFORATED SHEET/HONEYCOmB ACOU STIC TREATMENT Figure VI-28 Vll. QUIET ENGINE NACELLE DESIGN M. Dean Nelsen* For nearly 2 years the Boeing Company has been under contract to Lewis Research Center in a program of developmental design, fabrication, and testing of acoustic suppressors for the fans of high bypass ratio engines.

The quiet fans which have been under study are products of the NASA Quiet Engine Program. The Fan Noise Suppression Program is aimed at the further reduction of fan noise emanating from engines with fans which have been designed to minimize noise. Specifically, the fans studied in the suppression program have been the Lewis fan D and the General Electric Company fan A, both low tip speed fans. The program has been both tech- nology and hardware oriented. Some of the specific objectives of this pro- gram include the development of the necessary technology and methodology for the design of inlets and fan ducts for the purpose of providing high fan noise attenuation. During 1969, the Douglas and Boeing companies flew nacelles designed to minimize the fan noise of existing JT3D engines during landing approach. The technology from this program was applicable to the two-stage fans of lower bypass ratio engines. In this case, the engine had not been designed to minimize noise. Thus, one of the objectives of the current Fan Noise Suppression Program is to adapt the technology generated during and since the previous flight test demonstrations and extend this tech- nology to the modern Quiet Engine. Of concern in the suppression of noise from large single-stage quiet fans is the influence of the noise suppressor on fan aerodynamic performance and the resulting possibility of additional noise generation. Therefore, an important objective of the Fan Noise Suppression Program is to ascertain these effects based on actual measurements on full- scale hardware. Four nacelles have been conceptually designed during this program: two for an engine installation, and two for a full-scale fan rig test. Of these four, two have been detail designed for scheduled testing.

A fan nacelle will be fabricated and delivered to Lewis in late 1972 for quiet * The Boeing Company.

fan D full-scale fan rig tests. The second nacelle is the Quiet Engine na-

celle now under test at Lewis. The remainder of this presentation is de- voted to an overview of the Quiet Engine nacelle, some of its design fea- tures, new technology aspects, and the expected performance.

The Quiet Engine nacelle has four primary objectives. First, it is a nacelle which may be installed on the Quiet Engine for the purpose of ground tests to provide detail information on suppressor technology and the inter- relations between the Quiet Engines and the nacelle. This Quiet Engine nacelle is designed with a goal of reducing peak fan noise by 15 PNdB (per- ceived noise decibels) in both the forward and aft quadrants. Although 15 PNdB cannot be measured on the Quiet Engine today because of jet and turbomachinery noise floors, the configuration is designed to provide long term flexibility such that noise floors can be studied and reduced without uncovering the fan noise. Finally, the nacelle must be representative of good aircraft nacelle design practices; that is, the nacelle must provide compatibility of internal and external aerodynamic lines. The inlet config- uration must be chosen to provide the capability of being anti-iced and must not have adverse effects on the proper operation of the engine. Such things as influence of crosswind is an important consideration. The fan duct choice should again be compatible with proper operation of the engine and be real- istic and representative of flight nacelles in the sense that additional hard- ware, such as thrust reversers, could be worked into the design.

The selected nacelle which has been designed and fabricated for the Quiet Engine is shown in figure VII-1. The nacelle excluding afterbody and pylon consists of a three-ring treated inlet, wrap cowl around the accessory section, and a single ring half-length treated fan duct. The nacelle design shown is intended for a pylon installation on an aircraft with a design cruise at Mach 0.82. The inlet rings and fixed treated centerbody are supported primarily from six forward radial struts and three aft stabilizing struts.

The wrap cowl section covers the fan case mounted accessories. The fan duct provides for two possible concepts in fan reverser design and is de- signed to provide access to the engine. The nacelle is basically of glass fiber construction with aluminum supporting structure. The external cowling is necessary to preclude flanking noise paths from the engine case accessory and associated structures.

With a fan noise design objective of 15 PNdB, the relation of total inlet length to the number of acoustically treated circumferential rings in the in- let is indicated in figure VII-2. A flight inlet must provide internal diffusion for minimization of external aer.odynamic drag during cruise plus providing for the necessary acoustic treatment to achieve the noise reduction. If the acoustic requirement were ignored, the inlet length which provides good balance between internal and external performance (with the absence of auxiliary "blow-in" doors) would provide an aerodynamic throat internally designed for the maximum corrected weight flow through the engine and ap- proximately a 25-percent elliptical contraction between the highlight and the throat. Just downstream of the throat diffusion begins, and the contours at that station are critical. In order to prevent boundary layer separation under static operation, the angle 0 must be very carefully controlled. If the hub-tip ratio of the Quiet Engine with fan A and the entrance flow angles to the fan are considered, the best aerodynamic design would be approxi- mately 78 inches long. A "no-ring" inlet with peripheral treatment and a 15-PNdB noise reduction requirement would have astronomical length. The insertion of two acoustically treated rings would require approximately a 90-inch inlet; the three-ring inlet would be equal in length to the best aero- dynamic design. A further increase in the number of acoustically treated rings would provide a longer inlet since the treatment thickness increases as flow passage size decreases. Consequently, to maintain the acceptable initial diffusion and maintain reasonable internal Mach numbers, a longer inlet is required. Thus, the selected three-ring design for the Quiet Engine inlet provides no compromise on inlet length and, consequently, no com- promise in external aerodynamics.

In figure VII-3, the treated area required within the inlet is related to the number of acoustically treated rings. A two-ring inlet with an unlined centerbody requires approximately twice the treated area of a three-ring design. Treating the centerbody for the two-ring design has tremendous leverage in the reduction of the required treated area. For the three-ring design, a lined centerbody still provides a 25-percent reduction in treated area. The selected inlet design, then, requires 353 square feet of acoustic treatment to provide a 15-PNdB noise reduction. Treated area, of course, is wetted area in an inlet, and a wetted area is skin friction and weight.

Therefore, internally, the chosen design represents a substantial reduction in internal pressure losses over designs with fewer rings. A detailed anti- icing analysis of the three-ring design indicated marginally adequate capa- bility within existing engine bleed availability. A further reduction in inlet treated area by the addition of a fourth ring would result in a design that probably could not be anti-iced.

The inlet acoustic treatment was designed realizing the severe effect that a boundary layer (or shear layer) has on lining attenuation. Figure VII-4 shows the theoretical relation of attenuation as a function of one- seventh power turbulent boundary layer thickness. The lower curve relates the dropoff in attenuation that one could expect for a lining design for the noted geometry and environmental conditions if the boundary layer thick- ness were ignored. Immersing a lining designed for the mean flow Mach number, in the absence of a boundary layer, could have disastrous results.

A typical outer cowl inlet boundary layer near the fan face is of the order of 1.5 inches during takeoff conditions. In this case, a mean flow design would lose about 40 percent in theoretically available attenuation. This loss in ex- pected attenuation has been observed in flow duct testing when one compares the test data with theory. The same phenomenon is not observed when the airflow and the acoustic energy are traveling in the same direction, such as in a fan duct. It has been shown theoretically in this program that, if the boundary layer is considered in the design of the inlet linings, a significant portion of the lost attenuation can be recovered by impedance adjustment.

The upper curve on figure VII-4 indicates the theoretically achievable atten- uation when the lining impedance is adjusted for the boundary layers indi- cated. Physically, the adjustment required is a reduction in resistive im- pedance with a corresponding increase in backing cavity depth. Thus, a key item of technology in inlet lining design is a thorough understanding of the boundary layer growth on all acoustically treated surfaces.

The configuration of the outer passage of the inlet is shown in figure VII-5. This passage is typical of all four of the flow channels in theinlet.

The acoustic rings were located on streamlines. A boundary layer analysis was made to evaluate the stability of the boundary layer at the ring locations as well as to aid in the lining design. The required attenuation of fan noise is quite broadband since the quiet fan has relatively low blade passage tone

amplitudeswith respect to the associatedbroadbandnoise. Three basic

single degree of freedom linings were selectedto provide the broadbandat-

tenuationfor a noise reduction of 15 PNdB. Thelinings were tunedto the

first, second,andthird harmonics on the blade passagefrequencyat ap-

proach power fan speed. The linings on eachwall of the channelwere seg-

mentedas shown. Each lining is opposed by a like design onthe opposite

wall, thus allowing a configuration readily adaptable to theoretical analysis.

Although there are only three basic linings in each channel, four sections of lining are provided in the outer channel in order to allow staggering of the linings on the treated rings. The staggering was accomplished to minimize the thickness of the rings and the associated aerodynamic performance losses. The next inward channel lining arrangement provides only three segments of lining, as can be seen on the inward side of the ring. The chosen face sheet material for the acoustic linings is a laminated glass fiber impregnated with a polyimide resin. The face sheet resistance is varied by the number of plies of glass fiber cloth in the laminate. The segments vary in length from 5 to 20 inches, and the backing cavity depth varies from 0.1 to nearly 0.75 inch. Since each segment of the lining requires a different acoustic impedance, both the face sheet resistance and backing cavity depth must be segmented.

In order to obtain a better idea of how a segmented lining, as shown in the previous figure, has been incorporated in the nacelle, a typical lining construction from the inlet is shown in figure VII-6. As mentioned pre- viously, the lining facing sheets are constructed of many layers of polyimide impregnated glass cloth. The layers of glass cloth are arranged with an- gular orientation to provide uniformity in the permeability of the structure.

As shown on this example, the transition between one segment of lining and another segment of lining provides a uniformly smooth aerodynamic surface with no loss in treated area. The 15 layers of cloth required for the thin lining must taper to the 7 plies for the thick lining. This transition can be accomplished in actual practice in approximately 0.5 inch. It is interesting to note that the thin lining facing sheet thickness is greater than the backing cavity depth. The polyimide facing sheet is bonded to a honeycomb cell structure and again bonded to an impervious backing sheet. This material was chosen for the nacelle design for numerous reasons. Some of these reasons are as follows: (1) The material has structural load carrying capability as it is used in the nacelle.

(2) The material has service test experience in airline usage.

(3) The material has a 400 ° continuous operating temperature capa- bility, thereby being compatible with engine compartment temperature.

(4) The most compelling reason is that glass cloth layup allows the seg- mentation of the acoustic linings with an uninterrupted structure. As seen from the photomicrograph (fig. VII-6), a uniform permeable structure re- suits.

All the acoustic linings in the nacelle were designed analytically. Test- ing was accomplished, however, to determine the acoustic impedance prop- erties of the materials which were considered. The special test setup used to determine the impedance properties under the environment of sound pres- sure level and grazing airflow is shown schematically in figure VII-7. This grazing flow impedance test apparatus was designed to accept acoustic pan- els on one wall of the test section. The testing section, with three walls unlined, has a 2 by 2 inch cross section to prevent the existence of acoustic modes other than the fundamental mode. Thus, the device is an acoustic wave guide. The airflow and sound are supplied in the same direction to minimize the effects of boundary layer on wave propagation. Within the test section, the wall opposite the test specimen is provided with a stationary flush-mounted microphone and a translating flush-mounted microphone.

When the data from the two microphones are compared, the complex propa- gation constants in the axial direction can be ascertained from the resulting axial phase and attenuation rates. Knowing these constants and that only one mode exists makes it possible to compute the complex impedance.

Typical data obtained from this impedance rig are shown in figures VII-8 and VII-9. Figure VII-8 shows the resistive impedance of a 12-ply polyimide face sheet at 1500 hertz as a function of the test section Mach number and sound pressure level. It is interesting to note that the normal- ized resistance is both a strong function of Mach number and sound pressure level. Likewise, the reactance of the 12-ply face sheet as a function of Mach number and sound pressure level is shown in figure VII-9. Knowledge of such environmental effects on installed impedance, as noted in these last two figures, is necessary in order to reliably predict attenuations analyt- ically.

The theoretical prediction of the attenuation is shown in figure VII-10 as a function of frequency for the Quiet Engine nacelle inlet outer channel de- scribed in the previous figures. The theoretical prediction was made as- suming no boundary layer in the inlet. The lining impedances had not been adjusted for boundary layer effects. The data scattered around the theo- retical curve are test results from flow duct measurements for the actual lining design in the Quiet Engine inlet. Of course, the flow duct measure- ments include the effects of boundary layer. Since the linings tested were adjusted for the presence of realistic boundary layers, most of the attenua- tion that would normally be expected to be lost if the boundary layer in the design had been ignored is recovered.

The inlet rings and treated centerbody are shown prior to assembly in figure VII-11. The rings were fabricated in two halves and mechanically joined prior to assembly. The treated portion of the centerbody was fabri- cated as a continuous 360 ° part.

The assembled inlet is shown in figure VII-12. The flight lip is re- movable to allow installation of a flight simulation bellmouth.

In figure VII-13 is a view of the inlet illustrating the relation of the inlet rings to the fixed centerbody and outer cowl. The six forward support struts incorporate provisions for acoustic and aerodynamic instrumentation probes.

Figure VII-14 shows the relation of the rings to the centerbody and cowl just upstream of the fan face. The fan A rotating spinner fits into the centerbody and is allowed to spin during engine operation. The Teflon rub seal is incorporated in the fixed centerbody to prevent a flanking noise path through the centerbody structure.

The fan duct lining configuration is shown in figure VII-15. In the same fashion as in the inlet design, the linings are segmented to provide the broadband attenuation. Likewise, the linings are staggered between the two channels to minimize thickness of the circumferential ring. A longer or thicker ring would severely affect the balance between internal and external aerodynamics. Thus, the low frequency lining, shown as number 2 on the diagram, was tuned to 2100 hertz, which is higher than the 1500-hertz first blade passage harmonic of the fan. Consequently, the fan duct design differs in concept from the inlet design in that a 15-inch section of double degree of freedom lining is provided on the inner and outer walls of the fan duct but not on the ring. This lining is number 4 on the diagram. Double degree of freedom linings are designed for best performance at two frequencies. The two frequencies selected for the fan duct lining are the first and third blade passage harmonics (1500 and 4500 Hz). The double degree of freedom lining was opposed by a single degree of freedom lining on the ring designed for 4500 hertz. The opposing linings were thereby a matched pair at 4500 hertz. At 1500 hertz (the double degree of freedom lining design frequency), the opposing lining is extremely large; thus, the 1500-hertz impedance for the double degree of freedom lining was chosen as if the channel size was twice its actual size. The attenuation spectrum of this combination could then be analytically predicted by two separate analyses - one for low fre- quencies, one for high frequencies.

The resulting theoretical attenuation spectrum prediction for the com- bination is shown by the lower curve in figure VII-16. The triangle data points near the predicted curve are test data for the same lining configura- tion as measured in the flow duct. Note the 1500- and 4500-hertz tuning and associated broadbandness this powerful combination of linings gives. The theoretical prediction of the expected fan duct attenuation spectrum is shown by the upper curve. The flow duct test data are shown by the circles. The maximum attenuation predicted and measured occurs in the 2000- to 4000- hertz regime. The true attenuation could not be measured in this frequency range because of the flow duct noise floor. However, the data points in that portion of the spectra represent the actual measurements. The test data show that the theoretical predictions were too conservative at the high fre- quencies.

Figure VII-17 shows the fan duct ring and outer wall assembly looking in the flow direction. The ring is supported by thin struts from the outer wall only. Note the uninterrupted acoustic skins from the leading to trailing edge of the ring, a total of 5 feet.

The theoretical predictions of expected acoustic performance of the nacelle are shown in table VII-1. The predictions are made on both polar and sideline for both takeoff and approach power settings. The fan noise polar predictions are near the 15 PNdB design goal and are probably con-

TABLE VII-1. - ESTIMATEDQUIET ENGINEA

NACELLE ACOUSTICPERFORMANCE

Peak noise reduction, a P owe r

Quadrant

APNdB setting Fan noise Complete engine Polar Sideline Polar Sideline 12 12 13 14 Forward Approach II 10 Takeoff 15 14 16 14 7 6 Aft Approach 8 6 Takeoff 13 10 Sideline data : approach, 370 ft; apolar data, 200 ft.

takeoff, 1000 ft.

servative since they do not account for the influence of segmentation on lin- ing performance, duct end impedance, and peak directivity indices greater than unity. The "Complete engine" columns are the predictions for the test stand. These predicted reductions are lower than for the fan by itself be- cause of the influence of jet and turbomachinery noise floors.

Table VII-2 summarizes the predicted inlet and fan duct pressure losses TABLE VII-2. - ESTIMATED QUIET ENGINE A NACELLE PRESSURE LOSSES Condition Fan duct pressure loss, Inlet pressure loss, percent percent Without With Without With treatment treatment treatment treatment 1.5 2.0 4.7 Cruise 0.4 1.1 3.3 Takeoff 1.6 .2 at takeoff and cruise for a nacelle without treatment and one with treatment.

The pressure losses in the fan duct are higher than the inlet in either case: however, inlet pressure losses affect installed performance more than fan duct losses.

The predicted effects of the pressure losses on the installed Quiet Engine A performance are shown for the treated and untreated nacelles in table VII-3. The 15-PNdB fan noise reduction is expected to result in a TABLE VII-3. - ESTIMATED QUIET ENGINE A NACELLE PERFORMANCE Condition Thrust loss, Takeoff specific fuel percent consumption increase, percent With Without treatment treatment With Without treatment treatment Cruise 8.7 2.2 7.3 1.9 Take off 6.5 1.5 5.3 1.3 5-percent takeoff thrust loss and a 5.4-percent cruise specific fuel con- sumption increase.

The treated nacelle is shown in figure VII-18 on the Lewis test stand.

The engine-nacelle combination is just beginning a series of comprehensive tests. Since the nacelle was designed both aerodvne_icaily and acoustically from an analytical basis, the test program should provide direct compari- sons between analysis and theory. In addition, the interactions between the nacelle and engine can be studied under laboratory conditions. Finally, the engine-nacelle combination provides an excellent test vehicle for the detail study and reduction of jet and turbomachinery noise.

QUIET ENGINE NACELLE SINGLE RING FAN DUCT THREE RING _RAP COWL INLET Figure Vll-1 EFFECT OF NUMBER OF INLET ACOUSTIC RINGS ON TOTAL INLET LENGTH EFFECT OF EFFECTOF _ TREATMENT THICKNESS.'-I TRANSITION TREATMENT LENGTH ON DIFFUSER LENGTH LENGTH'--/ TREATMENT I ,[--1 SELECTE! DES,GN DIFFUSER LENGTH-_ J IOO LENGTH--_ _,_ _ _,,_ l_, # _ J -e..

.,- 60 CENTER.LI ._ INLET BEST =,.,1 _ 40 AERO DESIGN 2O

,o I_,__N!!I_, I

NO I RINGS

,,NGS i R'NGS _'_t R,NGS I

I Figure VII-2 INLET CONFIGURATION TRADES ,_PNdB = 15 2.0- SELECTED DESIGN 353 FT2 .eC O j I--- )- i i.u

!,.ot

r- i.w _ i a¢ I i-- i Z Z I I I i z ! |= Z _ | -_ ,_ I I- _z!

r.t !

2 RING 3 RING Figure VII-3 ATTENUATION RECOVERY WITH LINING OPTIMIZED FOR SHEARED FLOW I00 I _. LINING OPTIMIZED PERCENT OF 60 _ MAXIMUM LINING OPTIMIZED ATTENUATION FOR MEAN FLOW RATE 40 NOTES: 1) INLET MACH NO. = 0.26 2) DUCT HEIGHT = 6 IN.

3) ANNULAR DUCT RADIUS RATIO = 0.82 4) FREQUENCY= 2000 HZ O0 i J i 0.5 1.0 1.5 BOUNDARY LAYER THICKNESS . INCHES ........

Figure VII-4 INLET ACOUSTIC LINING CONFIGURATION LINING CAVITY DEPTH LENGTH DESIGN NUMBER NO. PLIES fiNCHES} (INCHES) FREOUENCY(HZ) i® 12 .08 20 4500 I RANGE 14 .19 t 5 TO 10 3000 8 1 .74 i 8 t500 OUTER CHANNEL TYPICAL OF ALL CHANNELS Figure VII-5 TYPICAL LINING CONSTRUCTION iLASS FIBER/POLYIMIDE RESIN LAMINATE ACOUSTIC FACE SHEET i .12 (15 PLY)ACOUSTIC -j .10 CAVITY DEPTH TAPER FROM 15 PLY 7 PLY, ACOUSTIC FACE -.06 (7 PLY) ACOUSTIC .73 CAVITY DEPTH Figure VII-6 GRAZING FLOW IMPEDANCE TEST SECTION - HONEYCOMB CORE b -FACING SHEET TEST PANEL BACKING SHEET TYPICAL CROSS-SECTION--, MIKE OR TOTAL PRESSURE AIR FLOW PROBEHOLDER & SOUND Figure VII-7 GRAZING FLOW IMPEDANCE DATA 12 PLY POLYIMIDE RESISTANCE ISO0 Hz 2.G X .el z_ MACH 0.5 A 0___ i.u W 0 0_0' I.C Z

go

lio ' 13'o ' 1_o '

SPL_dB Figure VII-8 GRAZING FLOW IMPEDANCE DATA 12 PLY POLYIMIDE REACTANCE 1.5 1500 Hz X O MACH 0.0 O 1.0 Z _E 0.5 O Z 12o i_o I_o i_o SPL--dB Figure VII-9 INLE1 OUTER CHANNEL LINING TEST RESULTS QUIET ENGINE A NACELLE APPROACH POWER I I I --THEORY (NO SHEAR) I I I I FLOW DUCT MEASUREMENTS I Z (WITH SHEAR)_ I i _o 3o <( ® Z ® _

/

/ \

_ 20

, \

,) 3 4 5 6 7 8 9 10 FREOUENCY-KHZ Figure Vll-10 TREATED INLET COMPONENTS Figure Vll-]l ASSEMBLED INLET Figure Vll-12 FRONT VIEW or-.ASSEMBLED INLET Figure VII-13 REAR VIEW OF ASSEMBLED INLET NEAR FAN FACE FigureVII-14 FAN DUCT ACOUSTIC LINING CONFIGURATION LINING DEPTH LENGTH DESIGN NUMBER NO. PLIES (INCHES) (INCHES) FREQUENCY(HZ) I(_ 7 .2S (AVG.) 15 _(_L) 5 1.00 15 i_(_) 6 .60 15 il ®. 4 _ ts 9 .75 "DOUBLE DEGREE-OF-FREEDOM LINING (DDOF) Fiqu re VII-15 FAN DUCT LINING TEST RESULTS QUIET ENGINE A NACELLE APPROACH POWER 40, i : 1 t WITHIN FLOW DUCT,,, --- l THEORY NOISE FLOOR _. _ , _.,,u_ l I _-':_ _ "_ • FLOW DUCT / JlllJ_ _1__ _ MEASUREMENTS _ 3C /, I l . "_, f t ® ,_ T_OTAL I I k i

/ I_ \,o L

2(3 IO • _'_"-- "15" SDOF/DDOF l COMBINATIONI S I ONLYI 2 3 4 5 6 8 9 i0 FREQUENCY--KHZ Figure VII-16 FAN DUCT RING AND OUTER WALL ASSEMBLY Figure VII-17 TREATED NACELLE ON LEWIS TEST STAND : i i ¸ Figure VII-18 Vlll. QUIET ENGINE TEST RESULTS Carl C. Ciepluch, Frank J. Montegani, Mike J. Benzakein,* and Steven B. Kazin* The acoustic and aerodynamic test results obtained with the two NASA Quiet Engines are given in this presentation. Some of the test results ob- tained at the General Electric Company are reviewed first. This review in- cludes the performance of the untreated or baseline Quiet Engines. In addi- tion, test results are shown for various degrees and areas of engine acoustic treatment. Finally, the results obtained here at the Lewis Research Center when a flight-type, acoustically treated nacelle was added to one of the Quiet Engines is examined.

GENERALELECTRICTEST FACILITY The baseline Quiet Engines were initially tested at the General Electric Test Facility. Figure VIII-1 shows a schematic of the engine test facility at Peebles, Ohio. One of the primary functions of this facility is the measure- ment of engine noise. The field surrounding the engine is level and covered with a large aggregate crushed rock - a surface which is representative of an average between a hard reflecting surface and a completely absorbing sur- face. Placed around the engine are 16 microphone poles set at 10 ° incre- ments on a 150-foot arc centered at the engine inlet centerline. The micro- .4 phones are 40 feet above the ground while the engine centerline is 12_ feet above the ground plane. This height was chosen because the resulting re- flection pattern, caused by the interaction of the reflection from the ground and direct radiation to the microphone, is more representative of that which would be encountered in the actual flyover of the engine on an airplane.

The area behind the microphones is clear of obstructions for a sufficient distance so that no difficulties with reflections back to the microphones from behind will be encountered. The nearest structure being, in fact, the *General Electric Company.

36-inch-diameter scale model facility, which is about 175 feet behind the

microphone circle.

The signals from the individual microphones are lead underground to the control building on the left where they are recorded simultaneously on a multitrack high response tape recorder for later analysis.

Figure VIII-2 shows an aerial view of this same facility. The photograph is a view from right to left as related to the previously shown plat. The en- gine test stand is on the left.

The sound field for the engine pad extends from the lower left around to the center. The microphone stands at 40 ° , 50 ° , and 60 ° from the inlet are in the position for data recording. The remainder of the microphone stands are lying down. The scale model facility is in the center with the control room and the General Electric crosswind facilities behind the scale model.

Also contained within this facility is sufficient measuring equipment to analyze the performance of the engines on test. These instruments are read in the control room and at the main plant in Evendale (Cincinnati) by means of telephone lines where a direct computer link provides online performance data.

ENGINE A AERODYNAMIC PERFORMANCE A photograph of engine A on test at the General Electric Peebles Test Facility is shown in figure VIII-3. Engine A contains the 1160 foot per second, low tip speed fan; the fan is driven by a standard CF-6 engine core which has had one of the low pressure turbine stages removed. In table VIII-1 are pre- sented some of the original NASA design requirements along with measured and estimated engine capability. It can be seen that engine A meets all the takeoff design requirements. The measured thrust and specific fuel con- sumption for the sea level static takeoff condition meet or exceed the design goals. The good specific fuel consumption noted results from the higher than expected efficiency of the low tip speed fan. For the Mach 0.25 takeoff flight condition the aircraft is in the vicinity of the standard FAA takeoff noise measuring station. Thus the design requirements shown for this condition are important parameters for limiting engine noise levels. Estimated per- formance values are shown which were extrapolated from sea level static

TABLE VIII-1. - ENGINE A PERFORMANCE

Estimated Measured Design requirement Takeoff (sea level static): 22 000 22 000 Thrust, lb 0.360 0.356 Specific fuel consumption, lb/hr-lb Takeoff (Mach 0.25, sea level): 1030 (max.)

Fan tip speed, ft/sec Fan bypass airflow, lb/sec Core airflow, lb/sec 900 (max.)

Bypass jet velocity, ft/sec 1275 (max.)

Core jet velocity, ft/sec Cruise (Mach 0.82; 35 000 ft): Thrust, lb Fan tip speed, ft/sec 1.5 1.5-1.6 Fan bypass pressure ratio 5-6 6.1 Bypass ratio 0.645 0.66 Specific fuel consumption, lb/hr-lb 1775 (max.)

Turbine inlet temperature, OF data. It can be seen that engine A meets the fan tip speed, bypass jet, and core jet velocity requirements quite easily. The core jet velocity which is the dominant jet noise source is seen to be 1180 feet per second, or about 100 feet per second less than the design limit.

From the engine A cruise design requirements, which are also shown in table VIII-1, we can see that the fan tip speed at the cruise conditions is 1160 feet per second as previously indicated. The slightly higher than spec- ified engine bypass ratio does not adversely affect the engine performance.

The turbine inlet temperature, however, exceeded the design limit by some 145 °. Part of this problem resulted from a lower than predicted turbine inlet flow area. The turbine inlet temperature requirement was primarily established to operate the turbine at a conservative temperature level.

Since CF-6 turbine material and cooling technology have been certified for airline use at turbine inlet temperatures far in excess of 2000 ° F, the engine is still operating at a relatively conservative temperature. No decrement in engine reliability is therefore expected because of the higher turbine inlet temperature.

ENGINE A BASELINE ACOUSTICS For acoustic baseline testing of engine A, the configuration shown in figure VIII-4 was employed. The rotor and stator are spaced two true rotor chords apart; the rotor blades number 40 and the stator vanes number 90 for a vane-to-blade ratio of 2.25.

Acoustically absorbing panels were placed in the area of the fan and in the core engine inlet. The outer fan duct wall has 15.5 inches of treatment ahead of the rotor, 15 inches between the rotor and outlet guide vane, and 23 inches aft of the outlet guide vane. The treatment cross section is shown inset on the drawing. It is of a resonator type and is similar to the treat- ment which was flight qualified on the General Electric CF-6 engine. This type of material has demonstrated suppression over a relatively wide fre- quency band while displaying high flight reliability. The treatment is simi- lar to the core inlet; however, it is tuned to reduce the higher frequency noise generally associated with the compressor.

As previously noted, the fan was first tested acoustically as a component here at the Lewis Research Center and then in the demonstrator engine at the General Electric Peebles test site. Figure VIII-5(a) shows a comparison of the approach speed perceived noise directivity for the fan and engine on a 370-foot sideline. The angle shown is measured such that 0 ° is along the inlet axis and 180 ° is along the exhaust jet axis. This sideline distance is representative of the altitude achieved by most aircraft on the landing ap- proach when they pass over the FAA approach certification point. The two sets of data agree quite closely with the engine being slightly higher. This is as expected since the engine contains the core compressor, combustor, and turbine with their associated ducting as well, of course, as the core jet.

The data indicate that the fan at this power setting is producing just about the same level at the forward peak at 40 ° and the rear peak of 120 ° while the engine, because of its other noise sources, is slightly rear dominated.

The same type of data at takeoff thrust at a 1000-foot sideline is shown in figure VIII-5(b). This distance is representative of a typical altitude at- tained by a present-day four-engine aircraft on takeoff as the aircraft passes over the FAA certification point. The engine and fan component again are quite close. The difference between the two sets of data in the front angles, where the fan component is higher, is thought to be a result of the front end drive mechanism used at the Lewis Research Center. The engine core jet is now contributing significantly to the rear most angles (150 ° and 160°), which makes the engine data higher.

The part played by the core jet is clearly shown in figure VIII-6 on the 1/3-octave spectral comparison at 120 ° . For the most part, the noise above 400 hertz follows the same trend for both vehicles. However, below 400 hertz and particularly around 160 hertz the low frequency core jet noise makes a significant contribution.

The two peaks in the spectrum at 2 and 4 kilohertz are the fan rotor's blade-passage frequency and its second harmonic. These two key noise components are very close for the two vehicles.

ENGINEC PERFORMANCE Some results on engine C with the higher tip speed fan are now present- ed. In figure VIII- 7, engine C is shown on test at Peebles. Some of the per- formance results obtained on that vehicle are shown in table VIII-2. It is apparent that the takeoff sea level static design requirement of thrust and specific fuel consumption were adequately met. Extrapolating those results to the takeoff Mach 0.25 requirements shows that the fan tip speed goal of 1400 feet per second (max.) was met. The bypass and core jet velocities were also well below the design requirements. As shown in table VIII-2, the design requirements for performance at cruise were again met. The spe- cific fuel consumption for engine C is seen to be a little higher than what we previously had shown for engine A. This results from the lower efficiency of fan C.

18"( TABLE VIII-2. - ENGINE C PERFORMANCE Measured Estimated Design requirement Takeoff (sea level static): Thrust, lb 22 000 22 000 Specific fuel consumption, lb/hr-lb 0.37 0.37 Takeoff (Mach 0.25; sea level): Fan tip speed, ft/sec 1400 (max.) 1390 Fan bypass airflow, Ib/sec 695 Core airflow, Ib/sec Bypass jet velocity, ft/sec 900 (max.) 900 Core jet velocity, ft/sec 1275 (max.) 850 Cruise (Mach 0.82; 35 000 ft): 4900 ......

Thrust, lb 4900 Fan tip speed, ft/sec 1570 Fan bypass pressure ratio 1.5-1.6 ...... 1.66 4.5-6 ......

Bypass ratio 5.1 Specific fuel consumption, lb/hr-lb 0.66 ...... 0.662 Turbine inlet temperature, OF TURBINE CHARACTERISTICS Engines A and C, because of the large difference in the fan rotational speed, had two different low pressure turbine systems (fig. VIII-8). The aerodynamic characteristics of the two low pressure turbines are compared in table VIII- 3.

Engine A incorporated essentially the first four stages of the CF6-6 low pressure turbine. The shaft work was 89.50 Btu per pound and was moderately low. The turbine loading which is defined as the enthalpy drop divided by the square of the blade speed at the mean radius was relatively low (0.764). The turbine efficiency, which was estimated prior to the test at 0. 918, was measured at 0.902.

TABLE VIII-3. - LOW PRESSURE TURBINE AERODYNAMIC CHARACTERISTICS E ngine A B 4 2 Number of stages 89.50 95.6 Shaft work, Btu/lb 'A 2 0. 764 1. 035 Loading, gJ H/2Upitc h 0.918 0.903 Turbine efficiency predicted 0.902 0.90 Turbine efficiency measured Engine C incorporated a new turbine. Taking advantage of the higher rotational speed of the fan and the advances made in turbine aerodynamic technology, the number of turbine stages was reduced to two with a relative- ly high loading coefficient of 1. 035. The turbine diameter was also decreas- ed by about 8 percent. These changes resulted in an appreciable weight saving in the turbine area. The shaft work is slightly up because of the high- er fan pressure ratio on engine C. The turbine efficiency was predicted at 0. 902 and was measured at 0.90.

ENGINE C BASELINE ACOUSTICS Now some of the acoustic results obtained on baseline engine C are presented. In this engine configuration (fig. VIII-9) fan C has 26 blades, 60 vanes, and a two rotor-chord spacing between the rotor and the outlet guide vane.

As in engine A, the area around the fan and the engine core inlet was acoustically treated. The treatment along the outer wall being 15 inches ahead of the rotor, 20 inches between the rotor and stator, and 20 inches effective behind the stator. The treatment design is the same as was used in fan A.

A comparison of the approach speed perceived noise directivity at

370 feet between engines A and C is shown in figure VIII- 10(a). Both the high and low speed fans produce the same noise level at most angles - the excep- tion being at 120 ° where engine C is higher by about 11 perceived noise deci- bels (PNdB).

At takeoff speed (fig. VIII-10(b)) engine C is clearly noisier than engine A.

On a maximum sideline noise basis, engine A peaks at 120 ° with a level of 98.5 PNdB, and engine C peaks at 50 ° with a level of 101.7 PNdB. However, the front maximum shows engine C to be dominant.

The 1/3-octave spectra at 50 ° (fig. VIII-11) shows the majority of the noise difference to be contained between 250 and 2000 hertz. This noise is commonly referred to as multiple pure tones or "buzz saw" noise and is characteristic of high tip speed fans. Although this noise is front radiated, it does, in fact, contribute to the rear maximum as well in this baseline con- figuration and largely accounts for the difference between engines A and C at 120 ° .

The solution to this high front end radiation problem on engine C is a major goal of the engine C test program now underway at Peebles.

ENGINEA WITH ACOUSTIC TREATMENT In order to investigate the effect of fan acoustic treatment on the overall engine noise, a number of acoustic suppression configurations were tested on engine A (fig. VIII-12). In addition to the baseline acoustic treatment de- sign described previously, two additional treatment configurations were tested.

First was the configuration labeled duct wall treatment, which incor- porated 20 inches more acoustic treatment on the inlet duct outer wall and 35 inches more acoustic treatment on the exhaust duct inner and outer walls.

This treatment was of the multiple degree of freedom type and was similar to the treatment incorporated in the baseline.

Second was the configuration labeled as fully suppressed. Here the treatment on the inlet wall was extended 58 inches, and three cylindrical rings acoustically treated on both sides were incorporated. All the addi- tional inlet acoustic treatment was of the type as shown on the left side of figure VIII-14. A photograph of the fully suppressed inlet is shown in fig- ure VIII- 13.

The fully suppressed configuration also incorporated an additional 37 inch long splitter in the exhaust duct. This splitter was treated on both sides with l inch thick polyurethane foam covered by a perforated plate.

Some of the results obtained using the preceding two treatment config- urations are now summarized. Figure VIII-14(a) gives a comparison of the different configurations at the takeoff condition in terms of perceived noise levels as a function of angle. The duct wall treatment configuration provided a 3 to 4 PNdB decrease at the maximum front and aft angles compared to the baseline. The three-splitter inlet was quite effective in reducing the front quadrant noise (by an additional 7 PNdB), but the addition of the splitter in the exhaust duct provided only a 2 PNdB reduction in the aft quadrant.

Comparable results were obtained at the approach condition as shown in figure VIII-14(b). The three inlet splitters again provided an appreciable noise reduction while the exhaust duct splitter lowered the aft quadrant noise only by 2 to 3 PNdB.

It is interesting to examine the effectiveness of the three inlet splitters on the sound pressure level spectrum. Shown in figure VIII-15 is a plot of the third octave sound pressure level spectrum against frequency taken at 50 ° from the inlet axis at the approach condition. No reduction due to fan treatment from 0 to 500 hertz can be observed, the noise in those bands being controlled by core exhaust noise. However, the splitter provided an appreciable noise reduction not only at the fan blade passage frequency and its higher harmonics but across the whole fan noise spectrum. Inlet split- ters are therefore quite effective in reducing front end noise.

When the fan has been suppressed as just shown, the next major item of concern is the turbomachinery noise emanating from the low pressure tur- bine. In view of this, a turbine suppressor was developed (fig. VIII- 16) which covered both walls of the core nozzle with a double layer honeycomb resona- tor treatment. This design provides a sufficient suppression range to effec- tively reduce the relatively high frequency turbine noise. The treatment has an effective length of about 36 inches. The inset shows the design of this treatment which is constructed so as to withstand the high temperature en- viromnent in the nozzle.

The data shown in figure VIII- 17 were obtained on the engine with and without the turbine treatment while the fan was fully suppressed at a speed equivalent to the landing approach on the 150 foot measuring arc at 120 ° from the inlet. The data were analyzed with a 50-hertz narrowband pass filter.

This method of analysis divides the spectrum into smaller elements and al- lows a more detailed analysis, particularly of the tone content. Indicated in the figure are the blade passage tones of the first, third, and fourth low pressure turbine stages. There is no indication in either spectrum of a sec- ond rotor tone.

Reduction due to the treatment at the third and fourth rotor fundamentals is clearly evident. However, a problem exists at the frequency band sur- rounding the first rotor blade passing frequency. The acoustic treatment has had almost no effect on this noise. Further investigation of this phenom- enon is being pursued on the engine C program which is currently underway.

Fan A noise levels, shown previously, were reduced substantially by treatment of the fan inlet and exhaust ducts. It was felt that possibly with the fully suppressed configuration the noise radiating from the fan casing might make a contribution to the far-field acoustic signature. It was there- fore decided to wrap the fan casing with a 2-inch-thick layer of polyurethane foam covered by a 1/8-inch-thick lead vinyl sheet to cut down casing radia- tion (fig. VIII-18). The engine was tested in a fully suppressed configuration with and without the muffled casing and showed, however, no noticeable changes in far-field noise at any speeds.

LEWIS RESEARCH CENTER TEST FACILITY A significant milestone in the Quiet Engine Program was reached in December 1971 with the delivery of the low speed engine to Lewis for fur- ther testing. The major objective at Lewis was to determine how low a noise level could reasonably be reached when fan acoustic treatment is added to this engine. A flight-type, acoustically treated nacelle was built to make this evaluation. The nacelle acoustic treatment was tailored specifically to the low speed fan noise characteristics. Nacelle flow passages were care- fully designed in order to keep performance losses low. This combination of Quiet Engine and acoustic nacelle is the best one tested to date considering bothacoustics andaerodynamicperformance.

The Lewis enginenoise test facility is shownschematically in figure VIII-

19. The arrangementis very similar to the GeneralElectric facility with

microphonesevery 10° on a 150foot radius, but there are somedifferences.

The Lewis microphonesare at the sameheight as the enginecenterline (13 ft)

andthey start on the inlet axis andgo aroundto 160 ° . Also, the reflecting

planeis hard pavement. Engineoperationis controlled from the flight re-

search building where the noise instrumentation andanslysis equipmentis

located. The site is sufficiently far from the buildings so that reflections are no problem. A photograph of the site showing the engine in the thrust stand and some of the microphone poles is shown in figure VIII-20.

LEWIS RESEARCHCENTER QUIETENGINE TESTS So far the Lewis Quiet Engine has been tested in two basic configura- tions. The first, shown in figure VIII-21 is the baseline configuration, which is basically the same as that at Peebles. It has no acoustic treatment except what is built into the fan frame, and a bellmouth inlet is used.

The other configuration tested at Lewis was fully suppressed as shown in figure VIII-22. This consists of engine A mated with the Boeing acoustic nacelle. It has a flight inlet and also includes turbine acoustic treatment.

This is the NASA Quiet Engine.

A cross section of this configuration is shown in figure VIII-23. The mat- ing of the acoustic nacelle with the basic engine is shown. There is contin- uous treatment from inlet to exhaust in the fan duct outer wall, also from the inlet centerbody along the inner walls. There are three aerodynamically con- toured splitters in the inlet duct and one in the exhaust duet. The turbine treatment is indicated. The areas and weights of the acoustic treatment in the nacelle are noted in the figure. The weights shown are for the test hard- ware and can possibly be reduced by as much as 50 percent for flight hard- ware.

It is generally conceded that noise results from different facilities vary because of different facility flow patterns, instrumentation techniques, and calculation procedures. This is one of the reasons for retesting the base- line configuration at Lewis. The sideline perceived noise levels for baseline engine A from the Lewis and General Electric tests at takeoff are compared in figure VIII-24. In the rear quadrant the agreement is very good, but in the front quadrant there are some small differences which may be due to differ- ent inlet flow patterns. The Lewis baseline data, of course, provide the basis for assessing the performance of the acoustic nacelle.

In figure VIII-25 the baseline data and the nacelle data at takeoff are com- pared. These are static data. This means that the jet noise is higher than it would be in flight because there are no relative velocity effects. Two kinds of data are shown in figure VIII-25, predictions and engine test meas- urements. The predictions are based on the best information up to the time of the Lewis tests such as fan noise from the Lewis fan tests, jet noise and turbine noise from the General Electric tests, and nacelle performance pre- dictions from Boeing. When the baseline measurements are compared with the predictions, the actual engine is about 5 PNdB lower in the front quadrant than predicted. In the front quadrant at this condition the predictions indicate that the noise is dominated by the fan so that the predicted data shown are essentially Lewis fan test results. It is assumed that the measured engine data shown here are fan dominated. It is felt that since the engine has an unusually unobstructed inlet for a test stand that the difference shown is due to the absence of inlet flow distortion.

The effect of the nacelle at 50 ° on the measured data has been to reduce the noise (about 9 dB) to about 84 decibels. This 84-decibel level is con- sidered to be a floor. The predictions indicate that the floor is made up of both the suppressed fan and jet noise. The fact that some measurements were lower than the predictions might be due to the absence of inlet flow dis- tortion or it might be due to a misjudgment of the jet noise. Only further testing will explain the results.

The differences between measurements and predictions are not an indica- tion of poor nacelle performance. The reduced level is most probably a floor. No fan noise reduction can be demonstrated below that floor. Since the engine started out lower than expected to begin with, the only difference that can be demonstrated is less than predicted.

In the rear quadrant the measurements and predictions agree quite well.

It is believed that the suppressed data here represent a jet floor.

Figure VIII-26(a) shows spectra at the baseline front angle of 50 °. No suppression occurs below about 500 hertz which agrees with the understand- ing that this region is controlled by jet noise. The baseline engine data have a strong peak in the 2000 hertz band which is due to the blade passage tone.

The nacelle has removed all evfdence of the tone from the spectrum. This means that the tone has been reduced well below the floor that remains.

Otherwise, if the tone were reduced only to the floor, for example, it would combine with the floor noise and leave a bump in the spectrum. This is fairly clear evidence that the nacelle is working well. It is assumed that the remaining floor is jet noise.

The rear spectra are shown in figure VIII-26(b). A harmonic is evident in addition to the blade passage tone. Again the nacelle has removed all evi- dence of the tones. The general reductions are less than in the front quad- rant because of reaching the floor, believed to be jet noise.

In figure VIII-27 is shown the perceived noise directivity at approach.

Again in the front the measured values are something less than the predicted valdes. The difference is less, however, than for the case of takeoff. The fact that the difference is less here than at takeoff reinforces the belief that the reason is inlet flow distortion.

The floor reached by the nacelle in the front and rear agrees well with predictions. In the front the predictions indicate that both fan and jet noise are present. In the rear, fan, jet, and turbine noise are predicted to be contributing.

The spectra in the front at approach are very similar to those at takeoff and, consequently, are not shown. The rear at approach is very different, however, as shown in figure VIII-28. A strong tone in the 6300 hertz band is evident. The origin of this tone is not presently known, but it obviously needs to be identified so it can be reduced.

Some general conclusions can be drawn from these static data. First, the baseline engine is inherently quieter than originally believed, mainly be- cause of reduced front end noise. Second, there is every evidence that the acoustic nacelle is functioning as planned. The actual nacelle performance will only come from detailed in-duct acoustic testing. Further measure- ments are also necessary to identify the origin of noise floors that emerge.

ACOUSTIC TREATMENT PENALTIES In figure VIII-29 is presented the engine thrust penalty resulting from the addition of the acoustic treatment. Herein is plotted corrected engine thrust as a function of corrected fan speed. The upper curve was obtained from the data taken at the General Electric Company and represents the untreated en- gine A configuration. The lower curve represents engine A with the acous- tically treated nacelle, and these data were obtained at Lewis. At the take- off engine speed of 3260 rpm, the acoustic treatment reduces the engine thrust by about 5 percent. There are two primary reasons for this: one is a result of the additional wetted surface area introduced into the fan inlet and exhaust ducts due to the presence of the acoustic splitters, and the sec- ond is that the skin friction coefficient of the acoustic surface is higher than the usual smooth metal duct surface. The predicted thrust loss was 5 per- cent. The agreement here between measured and predicted values is quite good considering that the measured data were taken at two different facilities and that the differences are approaching measurement accuracies. In future testing at Lewis a more detailed evaluation of the thrust loss will be obtain- ed. The specific fuel consumption increase at the takeoff speed was meas- ured to be 6 percent. It is, therefore, apparent that the use of large amounts of acoustic treatment will penalize airplane economics due to both perform- ance losses and weight increase.

QUIET ENGINE FLYOVER NOISE STUDIES The static engine noise data tell only part of the story. To get the whole picture, the static data must be extrapolated to flight. This means taking into account relative velocity effects and discussing the results in terms of effective perceived noise decibels (EPNdB).

Some results of flyover studies for a fo_u" engine aircraft retrofitted with quiet engines and flying a path resulting from the retrofit are shown in figure VIH-30. The noise from individual sources associated with the engine is also presented.

Consider first takeoff. The first bar represents total engine noise com- puted from noise measurements at Lewis. The next bar shows the contribu- tion made by the fan only. The unsuppressed fan level was deduced from the baseline engine measurements. The amount of suppression shown repre- sents the predicted nacelle performance, which is not yet confirmed. This level then is the predicted fan noise contribution. The next bar is the jet noise derived from the General Electric engine tests and projected to flight.

Turbine noise is not considered since it does not show up at takeoff.

The total engine results in a flyover noise of 90 EPNdB. This is con- tributed to principally by fan noise so that further reduction in engine noise can be achieved by better fan suppression. Such reductions can only be mod- est, however, since the jet noise floor is close to being reached. Reduction of jet noise to benefit from further fan noise reductions can only be accom- plished by reducing the mean jet velocity such as with new cycle designs.

At approach it appears that better fan suppression will not reduce total engine noise unless something is done about what is calculated to be turbine noise. Jet noise is very low at approach.

The net potential impact of our results on the community noise problem is summarized in table VIH-4. A DC-8 airplane currently makes 116 and 118 EPNdB at takeoff and approach, respectively. These numbers compare with Federal Air Regulation 36 (FAR-36) limits of 104 and 106 for new air- craft of the same weight. Conservative extrapolation of the baseline Lewis tests to flight reduce these numbers to 97 and 98. Mating the engine with a tailored acoustic nacelle achieves values of 90 and 89, of the order of 15 EPNdB below current federal regulations.

TABLE VIII-4. - FLYOVER NOISE COMPARISON FOR FOUR ENGINE AIRCRAFT Takeoff Approach E P Nd B 116 118 DC- 8 104 106 FAR- 36 97 98 Baseline Quiet Engine A 90 89 Quiet Engine A with acoustic nacelle CONCLUDINGREMARKS From this discussion of the Quiet Engine test results the following con- clusions can be drawn. First, the low speed engine is basically a quieter engine than the high speed one in an untreated condition, primarily because it does not have the multiple pure tone noise content in the front end that is a characteristic of supersonic tip speed fan operation. But, this does not mean that the high speed engine can not be competitive from a noise stand- point, because additional fan acoustic treatment can be incorporated to re- duce the high speed engine noise levels to that of the low speed engine. The additional treatment, however, introduces an engine performance and weight penalty that will reduce the basic weight advantage of the high speed engine.

The extent of this penalty will be evaluated in future testing of the high speed engine and a final assessment of the merits of the high and low speed engines can then be made.

Using the low speed Quiet Engine technology without fan acoustic treat- ment on airplanes of the DC- 8 class should result in an airplane that betters FAR-36 noise regulations by as much as 7 or 8 EPNdB. Furthermore, it is believed that noise levels approaching FAR-36 minus 15 decibels can be achieved when an acoustic nacelle incorporating splitters is added to the en- gine. However, the acoustic treatment will result in a penalty in airplane economics. It is important to note that the treated noise level is almost 30 EPNdB below that of the current DC-8 and 707 type aircraft and it, there- fore, represents a substantial potential improvement in aircraft noise levels.

The question arises as to whether it is possible to reduce engine noise levels even further to perhaps FAR-36 minus 20 EPNdB. This will be a challenging task, especially if the impact on airplane economics must be minimized. New technology and changes in the engine cycle are needed to accomplish this. The fan is the component that needs attention first. It will be necessary to either reduce fan source noise or apply additional acoustic treatment. For the takeoff condition the jet noise floor will be encountered and this can be handled by increasing the engine bypass ratio in order to re- duce jet velocities and the resulting jet noise floor. For the approach condi- tion the aft turbomachinery noise floor needs to be suppressed.

In future Quiet Engine testing it is planned to investigate the turboma- chinery noise problem and also to explore for other noise floors such as

combustionandflow scrubbingnoises that may be encountered at very low

enginenoise levels.

GENERAL ELECTRIC PEEBLES ENGINE TEST FACILITY SCALEMODELTEST FACILITY i_ _ MICROPHONESEVERY 10 0 'x._ CONTROLBUILDING ENGI "r_.;_.,:.¢-';.;:..¢_:c'_,'-_" i_ "_.- -_-_",'- .......

-:.tq._ _;;e TREES _..:.d.":,<*i_::':'; i,'.;, ..... '" .....

CS-63283 Figure VIII-1 AERIAL VIEW OF GENERAL ELECTRIC PEEBLES TEST FACILITY (:S-- r' _ 38 I Figure VIII-2 ENGINE A ON TEST T '_<_ ),_,i, i, ¸ Figure VIII-3 QUIET ENGINE A BASELINE CONFIGURATION Figure VIII-4 CS-{_3 _22 ENGINE A AND FAN A PERCEIVED NOISE DIRECTIVITY (a) APPROACH SPEED,370FT Sl DELINE 9O PERCEIVED NOISE LEVEL, 8O PNdB

I I I I I I I I

m; 20 40 60 80 100 120 140 160 ANGLEFROM INLET, DEG CS-63154 (b) TAKEOFF SPEED, 1000FT SIDELINE 100-- PERCEIVED NOISE LEVEL, 8O PNdB o BASELINEENGINE o BASELINEFAN

6o I I I I [ I [ I

20 40 60 80 100 120 140 160 ANGLEFROMINLET, DEG Figure VIII-5 cs-63] 48 ENGINE A AND FAN A ½-OCTAVE BAND SPECTRA TAKEOFF SPEED, i000 FT SIDELINE 9£ 8C SOUND PRESSURE _VEL, dB o BASELINEENGINE_\ -- o BASELINEFAN "_o \ 4O I ,I I I I I[ 1 I I ,_ 60 100 200 400 600 1000 2000 4000 10 000 FREQUENCY, Hz CS-63151 Figure VIII-6 ENGINE C ON TEST CS-63380 Figure VIII-7 LOW PRESSURE TURBINES ENGINEC ENGINE A CS-63378 Figure Vlll-8 CS-t:3391 Figure VIII-9 BASELINE ENGINE PERCEIVED NOISE DIRECTIVITY (a) APPROACH SPEED, 370 FT SIDELINE

I-

0 20 40 50 80 100 120 140 160 ANGLE FROM INLET, DEG CS-;,3_84 (b) TAKEOFF S PE[O, ]000 FT SIDELINE

[ /I t l __ _1 I I I

70-.- 0 20 ZlO 60 80 100 120 lZlO 160 ANGLE FRO/'_ INLET, DE@ CS - ' 5 _'_4!

Figure VIII-IO 2 04 BASELINE ENGINE I/5-OCTAVE BAND SPECTRA TAKEOFFSPEED, 1000FT SIDELINE 90-- 80-- 70-- SOUND PRESSURE LEVEL, dB

i

3ol I I l I I __Llo]ooo

50 100 200 500 1000 2000 .5OOU FREQUENCY,Hz CS-83385 Figure VIII-11 ENGINE A FAN ACOUSTIC TREATMENT .... ";iiiiiiiiiiiiiiiiiiiiiiiiiiii_:.

r BASELINE c,s-_,3_24 DUCTWALL _ ISSS_ FULLYSUPPRESSED F-_ _ Figure VIII-12 TREATED INLET FOR ENGINE A CS-63379 Figure VIII-13 ENGINE A PERCEIVED NOISE DIRECTIVITY (a) TAKEOFF SPEED, 1000FT SIDELINE PERCEIVED 9(3 NOISE LEVEL, 80 PNdB

_'I ,, oucr WALL TREATMENT

7O --/ v FULLY SUPPRESSED

"I I I I I I I I

60_ 20 40 60 80 I00 120 140 160 ANGLE FROM INLET, DEG CS-63_53 (b) APPROACHSPEED, 370 FT SIDELINE I00-- PERCEIVED NOISE LEVEL, 8O PNdB 20 40 60 80 100 120 140 160 ANGLE FROMINLET, DEG Fioure VIII-14 CS-63147 ENGINE A i-OCTAVE BAND SPECTRA APPROACH SPEED, 370FT SIDELINE SOUND PRESSURE KVEL, _ o BASELINE dB _ _ DUCT WALL TREATMENT 50 v FULLY SUPPRESSED , , I i _l i , I , ._L__L_JJ 40 60 I00 200 400 600 I000 2000 4000 i0000 FREQUENCY, Hz Figure Vlll-15 CS-63149 ENGINE A WITH TURBINE TREATMENT \_l _'-........................ _ ....

CS-633Z3 Figure Vlll-16 EFFECT OF TURBINE TREATMENT ON ENGINE A 50 Hz NARROW BAND SPECTRA APPROACH SPEED, 150FT RADIUS 90-- UNTREATED ..... TREATED TURBINE STAGE TONES ,'_-.

SOUND 70-- FOURTHI'" i "'IFIRST PRESSURE LEVEL, _ A !THIRD/A dB 60--

I I I I

2 4 6 8 FREQUENCY, KHz c_-_._5 Figure Vlll-l/ QUIET ENGINE A WITH CASING TREATMENT -/" ENGINE CASING LEAD VINYL SHEE_ " " 2qN. POLYURETHANE FOAM Figure VIII-18 LEWIS ENGINE ACOUSTIC TEST FACILITY -:;;';I-LIGHT RESEARCH BLDG HANGAR APRON CONTROL ROOM, ,- ENGINE THRUST STAND NOISEINSTRUMENTATION ", /_-- _ __-- / -" [ PAVEMENT _ " -_. " --_. [ _1_ _-_ MICROPHONES AT -_. " ---_ m [ [ _I_- _>_ I0° iNTERVALS Figure Vlll-lg LEWIS ENGINE ACOUSTIC TEST SITE . "_7 ' Figure VIll-20 BASELINE QUIET ENGINE A Figure VIII-21 .........

QUIET ENGINE A WITH ACOUSTIC NACELLE Figure V111-22 QUIET ENGINE A WITH ACOUSTIC NACELLE NACELLEACOUSTICTREATMENT EXHAUST DUCT _ INLETDUCT AREA, SQFT t _ 666 WEIGHT, LB CS-63321 Figure VIII-23 A PERCEIVED NOISE DIRECTIVITY BASELINE ENGINE TAKEOFFSPEED, 1000FT SIDELINE PERCEIVED NOISE 80 LEVEL, n/ o LEWISDAT_ o__, GE DATA PNdB 7O I I I 1 I I I 1 6O 20 40 60 80 100 120 140 160 ANGLE FROM INLET,DEG Figure VIII-24 cs-c,3382

QUIETENGINEPERCEIVED NOISEDIRECTIViTY

TAKEOFF SPEED, I000 FT SIDELINE

PERCEIVED

NOISE

LEVEL,

PREDICTED MEASURED

PNdB

70 BASELINE • o ACOUSTICNACELLE • a 6O I I I I I I I 60 80 IO0 0 20 40 120 140 160 ANGLEFROMINLET, DEG Figure VIII-25 cs-_,3zsz QUIET ENGINE SOUND SPECTRA (a) ]/3-OCTAVE BANDS; TAKEOFFSPEED, 1000FTSIDELINE 80 -- o BASELINE [] ACOUSTICNACELLE SOUND PRESSURE 6O LEVEL, dB 5O 4O I 1 I I I IX_bl 100 200 500 1000 2900 5000 10 000 FREQUENCY,Hz CS-63276 (b) I/3-OCTAVE BANDS; TAKEOFF SPEED, I000 Fr SIDELINE 70-- SOUND PRESSURE 60 LEVEL, dB _ _ o BASELINE ",q c_ -- _ a AcousTIC \ \ • . NACELLE _, \ 4o, I I I I I I \_' 50 100 200 500 1000 2000 _00 IO 000 FREQUENCY, Hz Figure VIII-26 CS-63Z81 PERCEIVED NOISE DIRECTIVITY QUIET ENGINE APPROACH SPEED, 370 FTSIDELINE 100, 9O PERCEIVED NOISE 80 LEVEL PNdB 7O ACOUSTIC NACELLE • a I I I I I I i I 6O 20 40 60 80 100 120 140 160 ANGLEFROM INLET, DEG Figure VIII-27 cs-63z_o QUIET ENGINE SOUND SPECTRA ]J3-OCTAVEBANDS; APPROACH SPEED, 370FT SIDELINE 80 I_ 70-- SOUND PRESSURE 6O LEVEL /r--/-_ o BASELINE dB o . ASEU ........

50-- _ a ACOUSTICNACELLE 40' I I I I I I ] 50 I00 200 500 I000 2000 5000 I0000 FREQUENCY, Hz Figure VIII-28 cs-_3z79 EFFECT OF ACOUSTIC TREATMENT ON ENGINE A THRUST 24000- 22000 - _/ 20000- U CORRECTED 18000 NETTHRUST, NT V F/B, 16000 LB 14000 - ._/O "-ACOUSTICALLY 1000012000:7 -v_ TREATED NACELLE _ I I I I I ] 2200 2600 3000 3400 CORRECTED FAN SPEED, NIl/8, RPM -cs-63383 Fiqure VIII-29 QUIET ENGINE FLYOVER NOISE FOURENGINERETROFITrED AIRCRAFT 100- FAN FAN SUPPRESSION .., SUPPRESSION I EPNdB 80 -- Z C_ Z ,:E hi U.

TAKEOff APPROACH Figure VIII-30 CS-63278 IX. QUIET ENGINE DEMONSTRATION Harry E. Bloomer A demonstration was arranged for the conference attendees to compare the noise output of the Quiet Engine with treated nacelle to the noise output of the JT3D turbofan engine. (The design features of the Quiet Engine and Quiet Engine nacelle are presented in sections VI and VII. The test results are given in section VIII.)

The Quiet Engine is located on a test stand at the Lewis hangar apron as shown in figure IX-1. The attendees were initially positioned near the peak noise lobe of the forwa_'d quadrant to hear the fan noise and then in the rear quadrant to listen to the core jet and fan exhaust noise. The engine was operated at levels representing takeoff and approach. Attendees were visu- ally informed of the noise levels for each position and operating condition.

Through the cooperation of the Air Force 4950th Test Wing under the command of Colonel James Walsh, Wright-Patterson Air Force Base, Ohio, a C-135 airplane fitted with the military counterpart of the JT3D engine was furnished to participate in the engine noise demonstration. The aircraft was parked next to the Quiet Engine site. One engine was operated alter- nately with the Quiet Engine in order to provide a direct evaluation of the engine noise reduction achieved by the Quiet Engine. The JT3D engine was also operated at takeoff and approach settings.

The physical differences of the two engines can be compared briefly by examining figures IX-2 and IX-3, which are sketches of the JT3D and the Quiet Engine, respectively. In addition, table IX-1 provides pertinent per- formance information for the two engines.

TABLE IX-I. - ENGINE COMPARISON JT3D/TF33 Quiet Engine Take off: Thrust, lb 15 800 22 000 Bypass ratio 1.2 5.75 Core jet velocity, ft/sec Fan jet velocity, ft/sec Approach: Thrust, lb Bypass ratio 1.6 6.3 Core jet velocity, ft/sec Fan jet velocity, ft/sec QUIET ENGINE A WITH ACOUSTIC NACELLE Figure IX-I ENGINE INSTALLED IN CONVENTIONAL JT3D/TF33 UNTREATED NACELLE _- SHORT FAN DUCT COREJETVELOCITYAT / TAKEOFF1618FTISEC_ I I / J I i I IGV'S -"" I I I -.

: / : _ FAN JET VELOCITYAT ,I i TAKEOFF 1030 FT/SEC 2 STAGEFAN ], HIGH TIP SPEED k CLOSESPACED 0GV'S CS-63Z74 Figu re IX-2 NASA OUIET ENGINE INSTALLED IN QUIET NACELLE SINGLESTAGELOW ,-2 CHORDSPACED TIP SPEED FAN-n / 0GV'S I I ,- 3/4 LENGTHFAN DUCT I I NOIGV'S_ t I / ..___=..__ Ir._:_, _ COREJETVELOCITYAT " '_ TA?OFF 1177 FT/SEC -_ " ACOU J ! ................. i , | I I | !

L.FAN JETVELOCITY AT cs-_; 3z 7_ TAKEOFF 821 FT/SEC Figure IX-3 X. APPLICATIONS TO AIRCRAFT PROPULSION SYSTEMS Newell D. Sanders, W. Harry Close,* Arthur A. Medeiros, and Richard J. Weber Applications of the Quiet Engine technology to current and future air- planes are presented in this paper. The results of the Quiet Engine Pro- gram are briefly reviewed as follows. Most significantly, the goal of a 15- to 20-decibel reduction from Federal Air Regulation 36 (FAR-36) was ex- ceeded. These noise levels were calculated for a four-engine airplane in the DC-8 and 707 class using the measured noise and performance of the Quiet Engine on the test stand. The noise reduction at takeoff is 26 deci- bels; 3.5 decibels of this comes from the improved airplane climb to a higher altitude over the judging point for takeoff noise. The noise reduction on approach is even more, 29 decibels. Another significant development is the elimination of the distinctive fan tones.

The FAR-36 noise levels for new aircraft similar to the DC-8 airplane are 104 decibels at takeoff and 106 decibels on approach. The new, large, wide-bodied jets now in service meet or better the FAR-36 levels. Starting with 116 decibels in the early 1960's, the first decade goal of a 10- to 15- decibel reduction has been achieved with these new airplanes. The experi- ence with the Quiet Engine encourages us to believe that similar gains will be made in the next decade.

A comparison between the Quiet Engine and a modern new engine such as the JT9D, the CF6, or the RB211 is also of interest. A direct compari- son is difficult because the flight engines have moderate amounts of sound suppression in the nacelles. Comparable flight tests with the Quiet Engine have not been made. It is estimated, however, that the bare engine A is 6 effective perceived noise decibels (EPNdB) quieter than the best of the new engines bare.

*Department of Transportation.

USES OF QUIET ENGINE TECHNOLOGY The Quiet Engine as its exists today cannot be used on airliners; it is an experimental engine. The Quiet Engine uses a derated CF-6 engine core; the CF-6 engine produces twice the thrust of the Quiet Engine. Thus, the core is much oversized for the Quiet Engine, and this engine is much too heavy for airline use.

As part of the Quiet Engine Program, General Electric is making studies of engines that match the Quiet Engine fans to an advanced technology core. Figure X-1 shows cross sections of two engines: one using the low- speed fan A and the other using the high-speed fan C. Both engines have 22 000-pounds thrust at takeoff. The engine with fan A is longer than the one with fan C. It is larger in diameter, and has more turbine and booster stages.

Acoustic treatments for the two engines are shown in figure X-2. The upper half represents engine A and the lower half engine C. In both cases, acoustic treatment in the fan discharge duct matches the treatment that was used in the experimental Quiet Engine Program. It was pointed out in a pre- vious section that the back end noise dominated the Quiet Engine and the front end noise was suppressed more than need be. The treatments shown here were selected to give a balanced noise from the front and the rear.

For fan A this balance is obtained without splitter rings; the inlet is lengthened and lined with acoustic absorbers as shown. Fan C was noisier, and simple lengthening of the inlet would not give the balanced noise between the front and the rear. Therefore, a splitter ring was added to the inlet.

The inlet was shortened to give balanced noise from front to rear.

Significant characteristics of the engine installations are summarized in table X-1. Both engines produce 22 000 pounds of thrust at takeoff. En- gine A has the low-speed fan operating at a tip speed of 1160 feet per second; the fan pressure ratio is 1.49, and the bypass ratio is 6.8. Engine C has the high-speed fan operating at a tip speed of 1550 feet per second; the fan pres- sure ratio is 1.55; and the bypass ratio is 6.4. Surprisingly, engine A is only a little larger in diameter even though it has a higher bypass ratio.

Fan A, operates with a higher inflow velocity than does fan C; this higher TABLE X-I. - QUIET ENGINES USING ADVANCED CORES Engine A C 22 000 22 000 Takeoff thrust, lb 1160 1550 Fan tip speed (cruise), ft/sec 1.49 1.55 Fan pressure ratio (cruise) 6.8 6.4 Bypass ratio 68.7 68.3 Fan diameter, in.

900 920 Jet velocity (mixed), ft/sec Engine weight, lb 6900 6490 Engine plus installation weight, lb lb: Relative specific fuel consumption, Bare 0. 981 1.000 Installed 0.986 1.009 inflow velocity produces almost exactly the flow increase required for the higher bypass ratio.

The combinations of fan pressure ratios and bypass ratios give the de- sired low jet velocities (900 and 920 ft/sec). These values assume that the primary and fan flows are mixed before being discharged as a jet. These velocities give noise floors well below the suppressed fan noises.

The weight of the low-speed engine A is 3940 pounds bare. It is heavier than engine C as might be expected because of the additional turbine and fan stages; the difference is 440 pounds.

When sound treatment is added, the differential between the two engines is slightly less because of the additional splitter in the inlet of engine C.

Engine A with treatment weighs 6900 pounds, and engine C weighs 6490 pounds. The differential is reduced to 410 pounds per engine.

The specific fuel consumption of low-speed engine A is approximately 2 percent better than for high-speed engine C because of the higher fan effi- ciency. The extra losses caused by the splitter in the inlet of engine C widens the efficiency gap slightly.

Until now, it has been assumed that the engines will be used in a DC-8 class airplane, but such a retrofit is costly. A more likely application is in new trijet and twin-jet airplanes. In the study of a trijet airplane powered by the quiet Engine, the gross weight was 200 500 pounds. The corresponding FAR noise limits at takeoff and approach are 100 and 105 EPNdB, respec- tively. For engine A with suppression, the takeoff and approach noises are 91 and 92 EPNdB, respectively. These values are a little bit higher than values obtained with the experimental engine. The values are approximately 10 decibels below the FAR-36 levels. Engine C with the extra splitter in the inlet is 2 or 3 EPNdB noisier than engine A. It should be recalled that these noise levels are set by the aft end noise. The aft noise suppression, which was obtained in the experimental engine program, has been used, and the in- let suppression has been tailored to match. If future research lowers the back end noise another 5 decibels, the full end suppression that was demon- strated in the experimental engine program can then be used. In that event, the noise may decrease to 15 EPNdB below the FAR-36 level. In addition, distinctive fan tones will be suppressed completely.

The lower noise of engine A is obtained at extra initial cost and higher direct operating cost. Using the cost of one engine C and nacelle as the base for comparison, one engine A is expected to cost $73 000 more. This extra cost plus the net effects of the weight penalty, improved specific fuel con- sumption, extra maintenance, etc., are expected to cause the direct opera- ting cost of aircraft powered with engine A to be 1.4 percent higher than for aircraft using engine C.

The discussion thus far has been directed toward methods of reducing the noise from future airplanes. As indicated earlier, fitting completely new Quiet Engines to existing airplanes will be extremely expensive. Some less expensive ways to obtain significant noise reductions for these airplanes are now discussed.

QUIETINGOF EXISTING NARROW-BODIEDAIRCRAFT Noise reduction technology has been applied in the development of the new commercial aircraft, such as the DC-10, 747, and the L1011 and, of course, will be applied in newer aircraft. However, there are approxi- mately 1700 existing domestic commercial aircraft that still have a long and economically useful life that create noise levels considerably in excess of the FAR-36 requirement for new aircraft of similar weight.

The aircraft in this category include the 707 and DC-8 powered by the JT3D engine and the 727, 737, and DC-9 powered by the JTSD engine.

There is no reason that existing noise technology cannot be applied to these aircraft to reduce significantly the overall aircraft noise problem more quickly than merely waiting for these aircraft to be retired.

In recognition of this, the Department of Transporation and NASA have initiated a joint program to reduce noise levels of these existing aircraft.

Two approaches are being pursued: One requires no changes to the engines and uses a major amount of acoustic and jet suppression; the other involves engine modifications in addition to some degree of acoustic suppression.

Both approaches are discussed.

FAA has contracted with the Boeing Company for the design, fabrica- tion, and testing of a quiet nacelle for the 707. Shown in figure X-3 are cross-sectional views of the 707/JT3D production nacelle and of the FAA- Boeing retrofit nacelle. The two configurations indicate the approaches be- ing taken to provide certifiable hardware that will greatly reduce the annoy- ance of the turbofan powered commercial fleet of 707 and DC-8 airplanes.

The inlet cowl has been lengthened, and two inlet rings have been added to accommodate 70 square feet of acoustic treatment in the inlet. A new side cowl has been developed, and the fan ducts have been extended to three- quarters of the nacelle length to accommodate 171 square feet of acoustic treatment. These changes constitute the so-called "lower goal" configura- tion and are predicted to yield at least 15.5 EPNdB suppression on approach, 6 EPNdB at takeoff with cutback, and 4.5 EPNdB noise reduction on the side- line under FAR-36 procedures.

The plug nozzle shown in figure X-3 is also being added to gain an in- crement of jet noise suppression and is predicted to yield at least 3 EPNdB additional suppression at takeoff and on the sideline. This configuration is known as the upper goal nacelle under the FAA contract terms. Ground tests of the noise reduction nacelle with jet suppression will be conducted in September 1972. Flight tests are planned for February 1973.

There are likely to be more than twice as many domestic aircraft pow- ered by JT8D engines in the late 1970's and early 1980's as 707 and DC-8 aircraft powered by JT3D's. The 727, 737, and DC-9 aircraft that are pow- ered by the JT8D are short-range aircraft, but they are used almost as many hours per day as the long-range airplanes and thereby generate many more takeoffs and landings per day, each of which is a noise event in some community. Figure X-4 illustrates the peak perceived noise level of a 727- 200 airplane at a constant altitude of 370 feet for various thrust settings up to 100 percent takeoff thrust. The total noise level is indicated by the solid line; the constituents of this total noise level are indicated by the dashed lines. As with the JT3D engine, fan noise is the dominant noise source at approach thrust; however, the jet noise level is within 10 decibels of the total and is equal to the inlet noise level at approach power setting. As thrust is increased, it is evident that the jet noise continues to rise and rap- idly becomes the dominant noise source for this engine. Therefore, the means to reduce the noise levels generated by JT8D engines are somewhat different from those considered for the JT3D. First, it is evident that 15 EPNdB of noise reduction at approach thrust is not possible with this engine unless a significant amount of jet suppression can be achieved at the low thrust setting. At the low jet velocities associated with approach thrust, significant amounts of jet suppression are unlikely to be achieved. How- ever, at the higher exhaust velocities typical of the higher thrust settings, significant jet suppression can be anticipated and, in conjunction with inlet and fan-duct treatment, appreciable noise reduction can be affected.

In figure X-5 is illustrated a cross section of the 727,/JT8D upper goal nacelle being developed by Boeing for the FAA under another contract. The upper portion of the figure portrays the full suppression mode, and the lower portion illustrates the cruise mode.

As with the JT3D, extensive inlet and fan-duct treatment has been added to attenuate the forward and aft radiated fan noise. Two polyimide honey- comb treated inlet rings have been added as well as lining material on the

outer wall and inner fairing. Perforated-plate treatment has been added to

the existing long fan ducts, and the combined fan and core exhaust is chan- neled through a 20-lobe suppressor nozzle. Ambient air is entrained through the blow-in doors, and mixing is induced within the ejector shroud.

Brazed perforated-plate - honeycomb lining material is strategically placed in the ejector shroud and on the plug to attenuate some of the mixing noise and to work on the fan noise that propagates out the fan duct.

To minimize the losses associated with the multilobe nozzle during cruise, the centerbody will be extended as shown inthe lower portion of the figure. In this mode, the ejector doors are closed, the multilobe nozzle is not bounded in the central annulus of the flow, and a plug nozzle is now formed at the nacelle rear face.

This configuration will be flight tested later this year to verify the de- sign.

The acoustic performance of these nacelle and jet-suppression configu- rations compared with the current production aircraft is shown in table X-2.

TABLE X-2. - ESTIMATED NOISE LEVELS AND NOISE EXPOSURE AREAS Noise levels at Type of aircraft Single takeoff and FAR-36 conditions, landing land area EPNdB exposed to 90+ EPNdB, Takeoff Approach sq mile 707: Current 113 119.5 55.8 104 104 27.4 Nacelle-jet suppres- sion retrofit 727: Current 101 109.5 29.4 96 99.5 6.6 Nacelle-jet suppres- sion retrofit Noise levels are shown for takeoff and approach conditions in accordance with procedures prescribed by FAR-36. As previously stated, the 707/ JT3D nacelle and jet-suppression (or upper goal) configuration is anticipated 'l to produce at least a 9-EPNdB noise reduction on takeoff and a 15_-EPNdB noise suppression at the approach measuring point. A sideline noise reduc- tion of 71 EPNdB or more is also expected for this configuration. The land area exposed to 90 EPNdB or greater during one takeoff and landing by the current aircraft is 55.8 square miles. The 707/JT3D nacelle and jet- suppression retrofit configuration is anticipated to expose only 27.4 square miles for one takeoff and one landing.

The 727/JT8D production aircraft and nacelle and jet-suppression con- figuration noise levels are also shown in table X-2. An approximately 5-EPNdB noise reduction is anticipated at takeoff and a 10-EPNdB noise re- duction is expected for approach conditions. An 8-EPNdB sideline noise re- duction is also expected as a result of the nacelle-jet-suppression configura- tion. These reductions in noise levels will reduce the 90 EPNdB contour for one takeoff and landing operation from the current 29.4 square miles to 6.6 square miles.

As mentioned earlier, another option to achieve noise reduction in JT3D and JT8D powered aircraft is to modify the engine by incorporating several noise reduction features in addition to some degree of acoustic suppression.

This option is presently being studied by the Lewis Research Center in an- ticipation of establishing the required contracted effort.

A proposed modification for the JT3D engine is shown in figure X-6.

The portion of the engine shown above the centerline is the current produc- tion version of the JT3D; that below the centerline is the proposed JT3D modification. Several noise reduction features are incorporated into the modification. Inlet guide vanes have been eliminated, the fan diameter has been increased to permit a higher bypass ratio, and a single-stage fan has replaced the original two-stage fan. The single-stage fan permits increased blade to vane spacing to about two rotor chord lengths without significantly increasing engine length. In addition, the blade vane ratio is acoustically optimized and swept-fan exit guide vanes are incorporated. Acoustic treat- ment is provided at the inner and outer wall of the fan inlet and on both sur- faces of the two inlet rings. Acoustic treatment is also proposed at the inner

and outer wail of the fan discharge as well as both surfaces of the bypass

splitter. This is not necessarily the optimum acoustic configuration and several with and without splitters and inlet rings will be investigated.

In order to minimize the number of changes and, hence, the cost of changes, a matching or booster stage would be used between the new fan and the existing core, so that the only change required to the core is a resetting of the last stage of the low-pressure turbine.

A comparison of some design parameters for the present and modified JT3D is shown in table X-3. The more significant parameters are an in- TABLE X-3. - COMPARISON OF PRESENT AND MODIFIED JT3D Standard New front JT3D fan JT3D Fan tip diameter, in. 50.2 56.5 46O Sea level airflow, lb/sec 600 Bypass airflow, lb/sec 412 Bypass ratio 1.36 2.20 Fan pressure ratio 1.75 1.67 Core jet velocity at takeoff; ft/sec 1285 Inlet guide vane Yes No Fan stages 2 1 Fan tip speed, ft/sec: Takeoff 1423 1535 Cruise Cycle temperature, OF 1703 1740 crease in fan diameter of 6.3 inches, an increase in bypass ratio of 0.84, and a reduction in core jet velocity of 145 feet per second.

The fan modifications, the resultant decrease in core jet velocity and the acoustic suppression can produce significant noise reduction while achieving TABLE X-4. - ESTIMATED PERFORMANCE COMPARISONS OF 707/JT3D Percent Current New front fan change 17 300 12.2 15 430 Installed takeoff thrust per engine, lb -4.4 0.84 0.80 Cruise thrust specific fuel consumption, (lb/hr)/lb 1.9 145 745 148 485 Operating empty weight, lb 333 600 333 600 Maximum taxi weight, lb 9950 -12.3 11 350 FAR field length, ft -0.6 4860 4830 Range, n. mi.

performance gains. The estimated performance changes for the modified JT3D engine installed in the 707 airplane are shown in table X-4. The higher bypass ratio for the modified engine results in an increased installed thrust and an improved cruise thrust specific fuel 'consumption of 12.2 and 4.4 percent, respectively. But the operating empty weight of the 707 is increased by 1.9 percent primarily because of the larger fan diameter and the weight of acoustic treatment.

There are several ways in which the performance improvement and weight increase can be traded off to achieve improved aircraft performance.

In table X-4 it was assumed that the maximum taxi weight and the passenger and cargo payload were the same for both the current and modified aircraft.

With this assumption the FAR field length is reduced by 12.3 percent by the modification, but the range is decreased by 0.6 percent. Further studies are required to determine the best use of the engine performance improve- ments.

The noise reduction goals deemed possible with the engine refanning are shown in figure X-7 for both approach and takeoff of the 707. At approach the noise level at the FAR-36 measuring station could be reduced by about 18 EPNdB below that of the current 707, which is about 5 EPNdB below the FAR-36 requirements for new aircraft in this weight class. Similarly, at

takeoff the noise level at the FAR-36 measuringstation could be reduced

about 15EPNdBbelowthat of the current 707, which is about6 EPNdBbelow

FAR-36 newaircraft requirements.

Another techniqueof presenting the noise reduction effect is to showthe

land area exposedto somegiven noise level. The land area exposedto

90 EPNdBor greater for both takeoff andlandingis shownin figure X-8 for

boththe current and modified 707 aircraft. The refanned engine configura-

tion reduces this area from 55.8 to 13.7 square miles, or a 75-percent re- duction in affected land area. As mentioned previously, the corresponding land area for the 707 aircraft with the FAA nacelle is 27.4 square miles.

A similar approach can be used to reduce the noise produced by the JT8D engines; these engines power the three-engine 727, the two-engine 737, and the DC-8 aircrafts. Some of the pertinent design parameters for the current production JTSD and a JT8D refanned in a manner similar to JT3D are shown in table X-5.

TABLE X-5. - COMPARISON OF PRESENT AND MODIFIED JT8D Standarc New front JT3D fan JT3D Fan tip diameter, in.

40.5 47.4 Sea level airflow, lb/sec Bypass flow, lb/sec Bypass ratio 0.934 1.73 Fan pressure ratio 2.04 1.73 Core jet velocity at takeoff, ft/sec Inlet guide vanes Yes Yes Fan stages Fan tip speed, ft/sec: Takeoff Cruise Cycle temperature, OF As was the case for the JT3D new front fan configuration, the bypass ratio is increased by the use of a larger diameter single-stage fan in place of the production two-stage fan. This particular new front fan design as- sumes minimum turbine modifications and retains the existing turbine shaft speed, which results in high fan tip speed. Under these conditions, inlet guide vanes are retained and fan noise treatment is required to minimize the fan noise. This aspect of the design is being explored further to evaluate several promising alternatives that may lower the fan tip speeds and elim- inate the inlet guide vanes.

The primary objective, however, is to lower the core jet velocity by in- creasing the fan bypass flow. Installation restraints such as 727 center- duct maximum airflow and 737 ground clearance will be major factors in the final fan sizing. Costs associated with turbine modifications are the limiting factor on tip speed and they determine the practicality of eliminating the in- let guide vanes.

The noise benefit anticipated with the refanning of the JT8D is presented in figure X-9. Shown are 90-EPNdB contours for takeoff and landing of a 727 equipped with production JT8D's. All contours are for the maximum takeoff weight for which the 727 is certificated and for operation on a stand- ard FAA day. As shown, the land area exposed to 90 EPNdB or greater is reduced by refanning from 29.4 to 3.9 square miles, an 87-percent reduc- tion. The quiet nacelle previously described would reduce the exposed area to 6.6 square miles.

The pursuit of both the refanning configurations and the nacelle/jet sup- pressor configurations for the JT3D and JT8D engines will provide a broad range of avenues for retrofit decision making. Noise levels, performance, and nonrecurring and recurring costs must be evaluated and traded off to arrive at a final decision.

Some preliminary performance estimates that have been made of the nacelle and refan retrofit for the 707 and 727 aircraft are presented in table X-6. These early estimates indicate that both retrofit options reduce noise levels below FAR-36 requirements for aircraft of a similar weight class. The refan option provides improved performance and noise levels 3 to 6 EPNdB lower than the nacelle option, however, retrofit costs, which must also be considered, are higher for the reran option.

TABLE X-6. - PERFORMANCE AND NOISE COMPARISONS OF ACOUSTICALLY TREATED NACELLE AND NEW FRONT FAN RETROFIT Type of aircraft Performance change, percent Noise level, EPNdB Installed Cruise FAR Oper- Range takeoff thrust field ating Take- Ap- thrust sfc empty length off proach weight 707/JT3D: Nacelle -2.0 i.I 1.0 104 104 New front fan 12.2 -4.4 1.9 98 101 727/JT8D: Nacelle -3.0 3.4 1.0 96 99.5 New front fan 13.0 -4.3 2.5 92 96 TABLE X-7. - RETROFIT COST ESTIMATES Domestic Domestic fleet retrofit Type of Kit cost estimates, aircraft aircraft millions of dollars cost estimate, millions inventory a of dollars per ship set Nacelle New front Nacelle New front fan and and jet and jet fan and nacelle suppressor suppressor nacelle 707 and DC-8 460 0.6 to 0.8 1.2 to 2.0 ......................

727 660 .6 to 0.8 1.0 to 1.8 ......................

737 and DC-9: 450 .4to0.6 .8 to 1.4 ...........

1570 total ................... 860 to 1200 1600 to 2700 aEstimated for 1980.

Cost estimates of a very preliminary nature are shown in table X-7.

The estimated cost per aircraft for the nacelle retrofit varies from $400 000 to $600 000, whereas the refan cost estimates vary from $800 000 to $2 000 000.

These costs, translated into figures for retrofitting the entire projected domestic air fleet, would be estimated at an average of about a billion dol- tars for the nacelle against an average of approximately 2.1 billion dollars for the refan option. The approximately two to one ratio refan to nacelle estimated direct costs could be outweighed by the tangible performance ad- vantage of the refanning and the social benefit of the lower noise level pro- vided by refanuing.

It is premature at this time to arrive at any firm recommendation on the best type of retrofit. Performance and noise data must be obtained for both the nacelle retrofit and the engine refanning. Then these data can be used to perform the economic and system analyses necessary to permit sound con- clusions based on a rational evaluation consistent with each airline's par- ticular set of economic considerations.

The DOT/NASA program will pursue both options for all aircraft and will perform system economic studies that will support rule making and retrofit decisions.

FUTURE QUIET AIRCRAFT In the future advances in aerodynamics, structures, and propulsion systems can be anticipated. The NASA Advanced Transport Technology Program was initiated about a year ago to determine the kinds of improve- ments most beneficial for commercial transportation and to start building up a data base of advanced technology for use in airplanes of the 1980 decade.

To guide the program, comprehensive studies have been performed by Boeing, Convair, and Lockheed under contract to the Langley Research Center. Parallel studies of the propulsion system are being done for the Lewis Research Center by General Electric and Pratt & Whitney.

Figure X-10 illustrates a possible configuration for a 1980 airliner as suggested by these studies. One of the main features of the advanced plane

will probablybe use of the supercritical wing concept developed by Whitcomb

and his associates at Langley. The best way to capitalize on this and other advances is not yet determined. For example, the flight speed might be raised considerably higher than at present - perhaps nearly to Mach 1. Or the structural weight at lower speeds might be lowered. The most desirable mix of speed, range, and payload will be decided by the airplane companies and their customers. In any event, we can expect to face a continuing con- cern over environmental factors such as pollution and noise.

Figure X-ll shows the sideline noise produced by various turbofan engines of the type that might be suitable for high-speed airplanes. The cruise Mach number of 0.98 used for these data emphasizes the differences from the engines discussed previously. Increasing either fan pressure ratio or bypass ratio extracts energy from the core and so reduces the jet noise, which is predominately from the core. Eventually, however, a limit is reached as the bypass jet noise becomes more significant. Other factors not shown here have a large effect on the numerical results. For example, the higher turbine-inlet temperatures anticipated for the advanced engines require higher bypass ratios than those shown earlier for comparable noise goals.

Another difference between the advanced engines and those previously discussed is the fan pressure ratio. As will be seen later, higher pressure ratios become desirable at speeds near Mach 1. The dashed lines show that fan machinery noise must be suppressed to reduce the total engine noise be- low 100 decibels. A related problem is that these pressure ratios are in the region where we must choose between use of a one-stage or a two-stage fan.

Making the proper tradeoffs between fan efficiency, required suppression treatment, and overall engine weight is an important question requiring more research.

Figure X-12 shows how the same two engine design parameters affect airplane performance. Relative gross weight for a fixed range and payload is given as a function of bypass ratio and fan pressure ratio. The lowest weight is obtained at the highest pressure ratio. The optimum bypass ratio is 6, but a somewhat higher value is required to get the jet noise down to the level probably needed for a future airplane, of the order of 80 to 85 decibels, perhaps.

The effect of compromises in engine design for the sake of quieting the airplane is summarized in figure X-13. To reflect the economic penalties of the compromises, airplane performance is presented in terms of percent of return on investment (ROD.

The left curve represents the performance level that would be achieved with today's engine technology. As a reference point, changes in ROI will be used as a measurement, starting from a current-technology point at FAR-36.

By 1985 improvements in engine technology should shift the noise-ROI curve to the right as shown. The improvements include better designs of the rotating components that reduce the generation of machinery noise, lighter weight, such as through the use of composite materials, and higher turbine- inlet temperatures with reduced penalties associated with turbine cooling.

The overall improvement can be invested into gains in airplane economy or reductions in noise or both. At the upper right, for example, the ROI could be increased by some 7 percent over today's level by selecting a high- pressure ratio, low-bypass-ratio design. However, such an engine would be much too noisy. At the other extreme, engines could be obtained that ap- proach 20 decibels below FAR-36 with no penalty in ROI. The dashed por- tion of the curve is intended to suggest uncertainty about the amount of ma- chinery noise suppression that can be achieved in this time period without undue weight or pressure drop.

As a longer range goal, whose achievement is as yet unknown, the tech- nology would be advanced sufficiently to provide a benefit in ROI together with a reduction in noise. This is indicated by the third curve in the lower right corner of the figure. This discussion has been limited to engine tech- nology. In concluding this topic it should be noted that in the coming de- cades, improvements in airframe technologies and operating procedures such as curved approach paths will also help in reducing noise and increas- ing ROI.

Consider now a brief review of another type of future airplane. Figure X-14 shows an advanced supersonic commercial transport. Although it is not certain whether the United States will ever build this airplane, NASA, in order to keep the nation's options open, is preparing to start a program for advanced supersonic aircraft similar to the one for subsonic transports.

The goals of the program are generally the same: to assess the readiness of the various technologies needed, to identify profitable areas for further research, and to build up a data base for use if and when it is determined that it is desirable to develop such a vehicle. A basic premise of the pro- gram is that the country will not seriously contemplate developing a super- sonic transport unless we can reasonably predict, not merely high speed, but also profitable operation coupled with public acceptance. A major prob- lem that will receive attention in this program is that of excessive engine noise.

Figure X-15 presents some feeling for the magnitude of the problem.

Relative range is plotted against sideline noise, which is the most trouble- some point for a supersonic airplane. Engine noise is quite sensitive to the type of airframe considered. For this figure a high-wing-loading, fixed- wing configuration similar to the recent Boeing SST design was assumed.

The upper right point represents an afterburning turbojet (ABTJ) engine of the type used in that Boeing design. Sideline noise is greatly in excess of the current FAR limit of 108 decibels. One way to quiet this type of engine is to increase its size. The additional thrust then available is not required for takeoff, so instead it can be throttled back during takeoff, which reduces the noise. However, the larger propulsion system is heavier, which then hurts the range. Significant reductions in noise using this technique cause unacceptable losses in airplane performance. It may be possible to reduce these losses to some extent by replacing the afterburning turbojet by another type of propulsion system. Depending on the particular estimates of relative engine weights, installation drags, and so forth, the figure suggests that there is a benefit available through use of either duct burning turbofans (DBTF) or nonafterburning turbojets (dry TJ). Detailed studies of the en- gines in conjunction with the particular airplane of interest are necessary to confirm these estimates. However, at best, it appears that the alternative engines and oversizing are apt to be a costly way to achieve low noise levels.

The dashed line represents another possibility for low noise. The variable-cycle engine would combine the quietness of a high-bypass-ratio turbofan at takeoff with the efficient supersonic operation of an afterburning turbojet. If there were no weight or drag penalties incurred by this con- vertible engine, we might even obtain an improvement in range due to the

better subsonicfuel consumption of the turbofan. However, at the present

time, this line is more of a hope than a fact. Serious studies of variable- cycle concepts are just now being initiated.

A different approach to quieting of supersonic engines is shown in fig- ure X-16. Again, relative range is plotted against sideline noise. The solid line repeats the curve for the nonafterburning turbojet from the pre- vious figure. The possibility of installing mechanical noise suppression de- vices on the exhaust nozzle is now considered. The data points represent the reductions in noise that have been experimentally measured in the labor- atory and also show the losses in range due to the penalties in thrust caused by the suppressors. It is assumed that the thrust loss is imposed only dur- ing the takeoff process and that the suppressors are retracted with no loss during climb and cruise. The data points are generally in the vicinity of the dashed line that corresponds to a 1-decibel reduction for each percent of thrust penalty. Suppressors with this level of performance are considerably better than the simple oversizing technique. However, suppression devices by themselves cannot yet offer as much quieting as is required. That is, for an advanced airplane such as this, it will probably not be sufficient to just meet FAR-36. Secondly, this figure gives a somewhat misleading pic- ture of what has been accomplished to date with noise suppressors. The data points do not necessarily represent devices that are ready for flight application in terms of low weight, retractability, or durability. Further- more, many of the points were measured only at sea-level static conditions and actual performance in flight could be worse.

Further research on jet noise suppressors, perhaps in combination with improved engine cycles, should lead to a solution for the engine noise prob- lem of supersonic airplanes.

SUMMARY Estimates of the prospects for quiet airplanes in the future are now summarized. Figure X-17 shows noise estimates for five classes of air- planes. The heavy dashed line across the figure represents the FAR-36 noise certification levels at appropriate gross weights and operating condi- tions.

23 6

The SSTusing afterburning turbojets andno soundsuppressionis esti-

matedto produce129EPNdBon the sideline. Suppressiondevicesand op-

erating techniquescan reduce this noise markedly. It is believed, however,

that future noise certification levels for subsonicairplanes will be 10 deci-

bels lower than the present FAR-36 level andthat SST's will be required to

meet that level. This meansa noise level near 100decibels, a 30-decibel

reduction. A dual cycle engine, it is hoped, might meet this requirement.

The present fleet of 707's andDC-8's producesnoises nearly as high as

120decibels onapproach. Combinationsof soundabsorbingdevices and en-

gine modifications are expectedto give a noise near 100decibels. This

compareswith a FAR-36 value of 106decibels.

The newadvancedtechnologytransports (ATT) are being studiedwith

two noise goals in mind: oneis 10decibels andthe other is 20 decibelsbe-

low the FAR level of 106ontakeoff. Initial estimates indicate that the

10-decibel down, or 95-EPNdBgoal, canbe met andthat probably 90 EPNdB

can be reached. Further reduction to 20 decibelsdownfrom FAR-36 (i.e., 85 dB) will require a further advancein technology.

Newtrijets using newQuiet Enginesare expectedto meet FAR-36

minus 10 (i.e., 90 EPNdB)using the demonstratedQuiet Enginetechnology.

Modestimprovementin technologyare expectedto lower the back endnoise

another 5 decibelsandallow full use of the front end suppressiondemon-

strated in the Quiet Engine project. This reductiongives a level of

85 EPNdB. Future advancesin technologyare expectedto yield another

5 decibels to give noise levels of 80EPNdB.

The STOLairplane has a noise goal of 95 EPNdBalonga sideline

500feet from the runway. To comparethesenumbers with other values in

the figure, the goal hasbeen convertedto noise alonga sideline at 1500feet.

The goal stated this way is 80 EPNdB. It is apparentthat this value is below

anythingaccomplishedso far andthat it is on the level with future expecta-

tions for the Quiet Engine.

This promising future for greatly reducedairplane noise results from

a continuationof the present combinedefforts of the airplane industry, the

engineindustry, andof the Government.

QUIET ENGINES WITH ADVANCED CORES CS-63448 Figure X-!

ACOUSTIC TREATHENT FOR ENGINES A AND C ENGINE A ENGINEC cs-_', _447 Figure X-2 707'/JT3D NACELLE AND JET-SUPPRESSOR CONFIGURATION IMARY THRUST .,=..-- REVERSER FAN THRUST REVERSER ENGINE AND ACCESSORIES AND FAN DUCT 707 PRODUCTION NACELLE MODIFIED FAIRING PLUG JET NOSE DOME: SUPPRESSOR INLET IIMARY EXHAUST NOZZLE INLET SIDE COWL ACOUSTIC NOZZLE SLEEVE ACOUSTIC FAN OUCT FAN THRUST TREATMENT TREATMENT REVERSER JT3D QUIET NACELLE FAA PROGRAM CS-(, ",403 Figure X-3 FLYOVER PEAK PERCEIVED NOISE LEVEL-727-2OO 370 Fr ALTITUDE; THREE JT8D-7 ENGINES 130 -- TOTAL PEAK _N EXIT PERCEIVED NOISE LEVEL, ........ INLET PNdB 100 _ '-APPROACH THRUST I I I I I J 50 60 70 80 90 100 3O 40 PERCENT OF TAKEOFF THRUST Figure X-4 (:_ .... -;47 727/JT8D NACELLE AND JET-SUPPRESSOR CONFIGURATION AMBIENTAIRFLOW SUPPRESSION MODE STOWED IN SHROUD -- INDICATES ACOUSTICTREATMENT CRUISEMODE Figure X-5 REFAN NOISE REDUCTION FEATURES ]. NO INLETGUIDE VANES 2. HIGH BYPASS RATIO 3. SINGLE-STAGE FAN 4. TWO INLET RINGS 5. SWEPT FAN EXIT GUIDE VANES 6. INCREASED BLADE VANE SPACING 7. OPTIMUm4 BLADE VANE NO.

8. BYPASS SPLIIIER 9. ACOUSTIC WALL TREATMENT REFAN PRESENT NEW FAN JT3D J CORE

I

FigureX-6 RETROFIT NOISE REDUCTION GOAL FOR 707 REFAN 1201- 11( -- m m 100 -- EPNdB I I....- 9(]-- Z ::_L _...J I 8O TAKEOFF APPROACH Figure X-I 90 EPNdB NOISE CONTOURS FOR 707 TOTALAREA, SQ. MI.

JT3-D 55.8 REFAN 13.7 A AREA -42.1 REDUCTION = 75% fREFAN .... JT3-D / APPROACH_ _ TAKEOFF (WITHCUTBACK)

k I t i I I

-5 0 5 10 15 -15 -10 DISTANCEFROMBRAKERELEASE - MILES Figure X-8 90 EPNdB NOISE CONTOURS FOR 727 TOTAL AREA, SQ. MI.

JT8-D 29.4 REFAN 3. 9 A AREA -25.5 REDUCTION : 87% /rREFAN FJT8-D APPROACH TAKEOFF (WITH CUTBACK) I,, I I I I I -I0 -5 0 5 l0 15 20 DISTANCE FROMBRAKERELEASE - MILES CS-63305 Figure X-9 ADVANCED TECHNOLOGY TRANSPORT Figure X-IO CS-(} BZ94 OF ADVANCED ENGINES NOISE CHARACTERISTICS M= 0.98 - _JET NOISE 1___ ..__ _x_ -_:--. FANMACHINERY 0.25 N. MI. 100 SIDELINE NOISE, EPNdB 90 8O I I I 191 I \ J 2 4 6 8 I0 12 0 CRUISE BYPASS RATIO CS-(, _1('3 Figure X-II CYCLE DESIGN FOR LOW NOISE M =0.98 0.25-N MI SIDELINE .41- JET NOISE, EPNdB/// 1.3 -CRUISE FAN RELATIVE PRESSURE TAKEOFF 1.2 -- RATIO GROSS 1.5 WEIGHT

1 I

.. I

-f "-t--__ I'_.-'r _' I

1.0 12 14 2 4 6 8 10 CRUISE BYPASS RATIO C5-(_3102 Figure X-12 BENEFIT OF ADVANCED ENGINE TECHNOLOGY 20- UNCONSTRAINED /#i,,/, OPT. ENGINE, '_ i0 -- BPR _ 2, FPR _"3/ I CURRENT / /_- MID-80 s TECH.

NOISE /(SdB SOURCE SUPP., GOAL 0-- TECH. _ JCOMPOS. FAN, IMPROVED RELATIVE _ J TURB. MATERIALS TO FAR-36., -10 EPNclB y & COATINGS) / _ S_ / -20 / FURTHERADVANCED / / TECHNOLOGY

J 1 i J i

-3O -Z -2 0 2 4 6 8 A ROI, PERCENT Figure X-13 cs-_,_l o4 ADVANCED SUPERSONIC TRANSPORT 2 <McR <4 230 TO 400 PASSENGERS 3500 TO 5500 N. MI.

Figure X-14 CS-A _Z93 NOISE-RANGE TRADEOFF NO SUPPRESSION 1.1 Z_ VARIABLE BPR CYCLE

1. o_,,_%-_',,_-m,,,_,,_-_.,_ _

RELATIVE .8 -- RANGE _,, T, / _ INCREASING

-

.5 _ I I I I I 106 110 126 130 114 118 122 SIDELINENOISE, PNdB CS-(_3105 Figure X-15 OF NOISE SUPPRESSORS BENEFIT DRYTURBOJET FAPNdB/A%F N = 1 4000 - i 0 i ------_ 0 0 iOr--, __o.-_ O/

j _._>_ oo o/

3600 - ._ 0 / ENGINEOVERSIZING, I i RANGE, 3200 -- N MI NO SUPPRESSION 2800 - I I FAR 36 _ I I I I 1 108 112 116 120 124 128 SIDELINENOISE, PNOB cs-_31ol Figure X-16 SUMMARY OF NOISE LEVELS EPNdB 129_ 120

, , l ,zl

108 L ,,-FAR-36 _I- 1 7_ ,,I 1 106 '

tool J i_--_--7---r_,.-

I I

i , I t

951 I I I 90 90 I I STOL AT 85 L _ _] 85 --I_ 1.500' 80 l___J CS-&3335 Fig u re X- 17 XI. STOL NOISE SOURCES AND FAN NOISE TREATMENT Raymond J. Rulis So far the subject matter presented in this conference has been related to conventional takeoff and landing propulsion. Now we will turn to the ef- fects of noise on short takeoff and landing systems. A viable STOL system will, by its nature, provide service out of heavily congested areas. As a consequence, the environmental specifications that will be imposed on these systems will be very severe. Extremely quiet and pollution-free operation will be demanded. Noise goals for these systems will require the initiation of major development efforts.

Current NASA activity in the STOL area has been initiated largely in response to the joint DOT-NASA card study, which was mentioned previously.

The card study identified noise abatement and traffic congestion relief as the two highest priority needs in assessing national benefits related to avia- tion and aeronautic research and development. The propulsion programs that have been initiated at NASA in support of STOL are new when compared with the acitivity that is in support of CTOL systems. In fact, most of them have only been initiated in the last 1 to 11 years and are just beginning to yield results.

A very brief overview follows on STOL noise goals, noise sources, and their affect on engine design, and also a quick review of current NASA ac- tivities related to STOL systems.

No firm specifications exist at present for STOL systems that are com- parable to Federal Air Regulation 36 (FAR-36). However, a much used figure-of-merit for STOL sideline noise is 95 perceived noise decibels (PNdB) at a 500-foot sideline (fig. XI-1). The figure presents a comparison of CTOL and STOL noise for a 150 000-pound-gross-weight aircraft. The allowable CTOL noise at a specification sideline of 2100 feet is 103 PNdB.

Extrapolating this value to a 500-foot sideline results in an allowable noise value of approximately 124 PNdB, which is approximately 30 PNdB above the value of 95 PNdB for STOL.

Both the STOL noise goal and noise sources affect the propulsion sys-

tem design. As with CTOL systems, the noise sourcesare the same. They

are enginemachinery noise andjet exhaustnoise. In addition, eachSTOL

poweredlift system generatesan additional noise source peculiar to the lift

system itself. For example, the externally blown flap lift system hasflap

impingementnoise, the augmentorwing lift system hasaugraentornoise, andso on.

The blown-flap system (fig. XI-2) employsa turbofan enginewhosetotal

exhauststream is directed against a large flap system that redirects the

thrust vector. High Lift coefficients are generatedby the momentumchange

andby the inducedsupercirculation effect. The noise sources for this pow-

ered lift system are machinery noise, jet noise, andflap impingementnoise.

Figure XI-2 also showsan upper surface blowing installation. In this

case, the exhauststream is directed along the top of the wing andis turned

by the coandaeffect alongthe flap surface. The noise sourcesof this pow-

ered lift system are machinery noise, jet noise, andwing andflap scrubbing

noise.

The augmentorwing poweredlift system (fig. XI-3) takes all or a good

part of the total fan flow from a high-pressure-ratio fan andducts the air

throughthe wing onto the nozzle of the augrnentorflap. Someof the air is

also usedfor leading-edgeblowing. Noise sources for this system are

machinery noise, jet noise, andaugmentorflap noise.

Jet noise andhowit affects enginedesignis illustrated in figure XI-4.

In order to meet the 95-PNdBSTOLobjective, an exhaustvelocity, both fan

and core, of approximately 850feet per secondis required. This corre-

spondsto a high bypassfan design. Notethe values of CTOL fan designs for

comparison. In order to achieve the low STOL noise goal, the engine design is forced toward higher bypass fans than used in CTOL installations.

From the previous CTOL discussions it is apparent that the major ma- chinery noise source is the fan (fig. XI-5). This is also the case for STOL.

For STOL engine designs low-speed, single-stage fans are of interest for externally blown flap (EBF) powered lift systems, and the multistage fans are of interest for higher pressure ratio augmentor wing lift systems. The higher pressure ratios result in smaller internal ducting for the higher pres- sure ratio augmentor wing lift systems. To achieve a 95-PNdB sideline noise, substantial suppression is required for the EBF system and even greater amounts for the augmentor wing application. In addition to fan ma- chinery noise, data indicate that the suppression of other machinery noises particularly in the turbine, will be required if the low STOL noise goals are to be achieved.

STOL powered lift systems also generate additional noise sources (fig. XI-6). For the EBF system, a jet with a velocity of 850 feet per sec- ond, which by itself results in less than 95 PNdB noise, reacting with a flap results in a noise level of approximately 105 PNdB for takeoff. If STOL sys- tem noise is to be controlled by basic engine design, a fan engine with very low pressure ratios is required.

The effect of noise goals on the physical design of a propulsion system is illustrated in figures XI-7 and XI-8. Shown are two schematics represent- ative of EBF propulsion systems. Both have the same high-pressure-ratio core and are dimensionally scaled the same. The core jet has been designed to meet the STOL noise goal. The higher bypass engine has been configured with a sufficiently low-pressure-ratio fan to preclude the need for a velocity reducer. The other, a 1.35-pressure-ratio engine, does require a velocity reducer to lower EBF flap interaction noise. Note that the fans are designed for low noise; that is, rotor-to-stator spacing has been optimized. Both en- gines require substantial acoustic treatment as is shown in the inlet, fan exit, and core exit. For optimum cruise operation, variable-area nozzles are re- quired. The high bypass engine also employs a booster stage. Without this stage, a low cruise thrust value results. With the high bypass fans, the en- gine thrust design point becomes the cruise point instead of the takeoff point.

The augmentor wing propulsive lift concept requires an engine design that supplies high-pressure fan air to the augmentor nozzle. The high- pressure air passing through the wing nozzle results in noise levels as shown in figure XI-9. Values of suppressed and unsuppressed flyo'rer noise are shown. At high wing-slot pressure ratios, substantial noise reduction must take place to meet the STOL noise goals. Sideline noise, as shown, is not as severe as flyover noise.

A cutaway schematic of a two-stream engine suitable for augmentor wing powered lift systems is shown in figure XI- 10. This engine does not look substantially different from conventional high-pressure-ratio fans ex- cept for the fan inlet and the discharge air duct work. The multistage fan has been configured for noise with proper rotor-stator spacings. The core exhaust jet is designed for low velocity, hence, low noise. Inasmuch as

substantial fan noise reduction is required, a variable-area acoustic inlet is

shown. The inlet will provide near sonic flow conditions during takeoff and landing. Fan exit suppression is not required because the noise from fan- discharge air will be treated by the design of the augmentor wing proper.

A brief overview of two of the major NASA activities that have been in- itiated in support of STOL systems is now presented.

The first program is the QUESTOL Program. QUESTOL stands for quiet experimental STOL aircraft. In figure XI- 11 is an artist's concept of an EBF-powered lift aircraft for the QUESTOL Program. Note that it is a high wing, multifan jet aircraft. The program objectives are as follows: (1) Develop the technology base for turbofan STOL transport systems.

(2) Define the requirements and criteria for stability and control.

(3) Identify noise patterns.

(4) Develop data for operational criteria.

(5) Define guidance and navigational requirements.

(6) Investigate and validate promising powered lift concepts.

The QUESTOL Program will provide a test bed from which promising powered-lift concepts can be evaluated and, in addition, will provide the much needed system inputs required to enable the STOL system to become operational. The QUESTOL Program has one unique feature from a propul- sion viewpoint, namely, the planned use of an existing powerplant for the first aircraft configuration. The task of quieting an existing propulsion sys- tem to the STOL noise goals will be great. A leading candidate for the EBF STOL is the TF-34 engine. It is a 6 to 1 bypass ratio engine with a thrust rating of approximately 9200 pounds.

An extensive test program is underway using a TF-34 engine as an ac- oustic test bed (see fig. XI- 12). This engine requires large amounts of ac- oustic suppression for the fan and core. Tradeoffs of performance and noise will be evaluated. In addition, the velocity decayer shown will be evaluated with a large wing-flap system. The test program will permit a total system evaluation of noise to be made.

The second major program in support of STOL is the Quiet, Clean STOL Experimental Engine Program (QCSEE). While the QUESTOL Program pro- vides the required aircraft technology background for STOL, the QCSEE 25O Program will provide the technology required to develop viable quiet, clean propulsion systems.

A substantial amount of component technology is being applied in the first phase of this engine. The QCSEE Program is summarized as follows: (1) EBF development (fan pressure ratio, 1.15 to 1.35) (a) Small scale aerodynamics (b) Small scale acoustics (c) Large scale acoustics (d) Composite materials (2) Augmentor wing fan development (a) Small scale aerodynamics (b) Larger scale aerodynamics and acoustics (3) Propulsion system installation (a) Low-speed aerodynamics (fan pressure ratio, 1.15 to 1.5) (b) High-speed aerodynamics (fan pressure ratio, 1.15 to 1.5) (c) Sonic inlets (d) Thrust reversers (e) Externally blown fan wing interaction noise (f) Velocity decayers (g) Augmentor wing noise (4) QCSEE studies (5) Quiet turbofan STOL aircraft studies (6) QCSEE development You can readily see that we are concentrating heavily in the fan noise sup- pression area. The two powered-lift systems discussed require fans cover- ing a very large range of pressure ratios. These fans must be efficient, quiet, and able to tolerate severe airflow distortions. In addition to the fan technology, a substantial effort is being put forth to resolve the STOL pro- pulsion installation problems.

STOL and CTOL aircraft operate differently. STOL aircraft will ap- proach and take off with much higher angles of attack. In addition, runway layout will probably require that the STOL propulsion system have greater tolerance to crosswinds. In the overall program installation effects for both cruise and low-speed flight conditions will be covered.

Two study programs are included in the QCSEE Program. One is in di- rect support of the engine program. In it we shall be examining a variety of propulsion systems covering a wide range of fan pressure ratios (from 1.15 to 3.0). Results from parametric studies will be used to define and optimize the physical layouts of the more promising systems. The second study will cover the entire STOL system, with a sizeable effort going toward optimizing aircraft configurations. The two studies are being conducted in parallel and are closely integrated. Results from one are fed into the other, so that in the end we shall have a very good understanding of the propulsion needs for STOL.

It is expected that the study programs and the component technology ef- fort should enable us to begin the hardware development phase of the QCSEE Program by mid 1973.

In summary, the pronounced effect that the 95- PNdB sideline noise goal has on the design of the entire STOL system can easily by recognized - both for the propulsion system and the aircraft. Most of the problem areas asso- ciated with STOL propulsion systems have been defined and the required re- search begun to provide answers to the problems.

There is a great need for the two major programs QUESTOL and QCSEE to move forward rapidly to provide the technology base needed in order to permit the initiation of a viable and environmentally acceptable STOL system.

NOISE COMPARISON OF STOL AND CTOL GROSS WEIGHT, 150000 LB 125 - _ ,t-:" 12G - I-.- c_ 119- u_ Z .,,.-,, O O Z m I-,- I-- X PERCEIVED 1111 - O NOISE LEVEL, "', (,.9 Z ZZ PNdB 10; - __ F- "4 50OFT 2100 _ 500 _ SIDELINEDISTANCE CS-6] 135 (0.35MILE) Figure XI-I STOL SYSTEM BLOWN FLAP POWERED / '-'FAN FAN & _ \ JETNOISE_'\ L FLAP IMPINGEMENT NOISE NOISE _'- CORE & JETNOISE UNDER-THE-WING CONCEPT f-- JET& MACHINERY _'--__ ,/ NOISE J I' _ roWING SCRUBBING FAN NOISE-/_ _' .__ NOISE

--.N

OVER-THE-WING CONCEPT CS-63126 Figure XI-2 AUGMENTOR WING PROPULSION SYSTEM

'1 /'_'_'_:==__ '_-AUGMENTOR

J _'_JET & CORE NOISE NOISE Figure XI-3 CS-63128 JET NOISE GOAL EFFECT ON PROPULSION DESIGN 5OO-FTFLYOVER TOTALTHRUST, 97 000 LB; FOURENGINES 120- 110 - STOL ., "S _'_ JET

24 9 .,_'?,_ I

NOISE, 100- PNdB 6 1 g0 \,_'" 6 ByIAss 8O I I I I RATIO I 60O 800 1000 1200 1400 1600 CS-59064 EXHAUSTJET VELOCITY, FTISEC Figure Xl-4 UNSUPPRESSED FAN NOISE FOUR ENGINES - 90 000-LB TAKE-OFF THRUST; 500 FTSIDELINE 140-- MAXIMUM -- -_ SINGE-STAGE NOISE LEVEL HIGHSPED PNdB --_SINGII:-STAGE t LEWIS DATA LOW SPEED • OTHERS I I I 1 I I I I l.O 1.4 1.8 2.2 2.6 FANPRESSURE RATIO CS-63Z02 Figure XI-5 ESTIMATE OF MAXIMUM 5OO-FT SIDELINE FLAP NOISE DURING TAKEOFF 170 000-LB GROSS WEIGHT AIRPLANE; 94000-LB THRUST (FOUR ENGINES); 15 ° CLIMB ANGLE 110 - __,_,,_J_ PERCEIVED BYPASS RATIO: 13 _ NOISE LEVEL PNdB 95-- l_ '_' VCORE 500 600 700 800 c_O 1000 1100 CORE EXHAUST VELOCITY, FTISEC cs-63zss FigureXI-6 EXTERNALLY BLOWN FLAP PROPULSION SYSTEM UNBOOSTED BOOSTED FAN PRESSURERATIO L 15 1.15 BYPASS RATIO 24.3 23. 5 TAKEOFFTHRUST, LB WITHOUTBLEED WITH BLEED 25 6O0 33 9O0 CRUISETHRUST(UNINSTALLED) 23 500 33 000 (25 000 F-r, M = . 75) 5O00 COREJETVELOCITY,FT/SEC 840 FAN JETVELOCITY, FTISEC 500 Figure XI-7 EXTERNALLY BLOWN FLAP PROPULSION SYSTEM , , _ / FAN PRESSURE RATIO ................ 1.35 I TAKEOFF THRUST (LB) ................ 21 O00J BYPASS RATIO .................... 11.2J CRUISE THRUST (25 000FT,M = .15)........ 4420J CORE JETVELOCITY,FTISEC ............ 140J FAN JETVELOCITY,FTISEC ............. /I0

J

CS-63297 Figure XI-8 EFFECT OF PRESSURE RATIO ON AUGMENTOR WING NOISE GROSS WEIGHT, 170 000 LB; FOUR-ENGINES 120 -- _._ !._;-_:....

PERCEIVED

NOISE

SUPPRESSED - FLYOVER

LEVEL100 --

AT .500FT, PNdB 90-- SUPPRESSED- SIDELINE I I I I I I 8O 1.6 1.8 2.0 2.2 2.4 2.6 1.4 WING SLOT PRESSURERATIO CS-62963 Figu re XI-9 AUGMENTOR WING PROPULSION SYSTEM / - /' ,/ ,/ / }; - : /' ,,::' i:,,, ,.......

// _ :! i"_ .....

- ___ - ...... T _-.--. ' _ ,,

iI :::: ,,'. 7- ]

FAN PRESSURE RATIO ............... 2.4 TAKEOFF THRUST (LB) a ............... 17, 900 THRUST SPLITWING/EXHAUST ........... 83/17 alNCLUDES15% DUCTLOSS TO WING.

CS-(_3295 Figure XI-lO ARTIST'S CONCEPT OF EXTERNALLY BLOWN FLAP POWERED AIRCRAFT Figure XI-11 TF-3& ACOUSTIC TEST ENGINE )// " . 4"

t

CS-(_3/03 Figure XI-12 Xll. FLAP NOISE Robert G. Dorsch, Paul L. Lasagna,* Domenic J. Maglieri,** and William A. Olsen This paper reports the progress of NASA in-house research on the noise created by externally-blown-flap STOL airplanes, specifically, the noise associated with the powered lift obtained by flap blowing. The major part of the report concerns flap noise associated with lower-surface blowing. This portion is followed by a brief discussion of the flap noise related to upper- surface blowing.

LOWER-SURFACEBLOWING An externally-blown-flap STOL airplane with the engines located below the wing employs lower-surface flap blowing (fig. XII-1). Lift augmenta- tion is obtained by lowering the flaps directly into the fan-jet exhaust. The flap interaction noise data presented herein are for a double-slotted flap system similar to the type shown in figure XII-1.

In the Lewis tests (fig. XII-2) the exhaust of a 10 000-pound-force- thrust fan-jet engine with a bypass ratio of 6 was simulated by a 1/2-scale model of its bypass nozzle assembly supplied with pressurized ambient- temperature air. The nozzle assembly was pylon-mounted on a wing section having a 7-foot chord length and large double-slotted flaps for lift augmenta- tion. The span of the wing section was 9 feet. The core nozzle was 8 inches in diameter, and the outer diameter of the annular nozzle was approximately 23 inches.

Similar noise tests (fig. XII-3) were conducted at the NASA Flight Re- search Center (FRC) with a test facility which used a modified F-111B wing- and-flap system blown by the exhaust from a CF-700 turbofan engine. The *NASA Flight Research Center ** NASA Langley Research Center engine had a 12-foot-long acoustically treated inlet to suppress fan noise. A 9-foot-long acoustically treated tailpipe was installed aft of the engine to re- move machinery noise from the engine exhaust. The mixed fan and core stream exhausted from the tailpipe through a 22-inch-diameter convergent nozzle and was directed at the flap system. The wing was mounted on an X-Y table that could be actuated in either the longitudinal or transverse directions in order to vary the relative positions of the exhaust nozzle and flap system.

The noise tests at both Centers were conducted over a range of nozzle exhaust velocities and flap deflection angles.

Typical noise radiation patterns obtained from the Lewis 1/2-scale model are shown in figure XII-4. The exhaust velocity for the core was 765 feet per second and that for the fan was 582 feet per second. Overall sound pressure level is given for three flap position settings and for the noz- zle alone. The innermost set of data points represents the sound level for the nozzle alone. The triangular symbols give the sound level when the noz- zle is installed below the wing with the flaps in the retracted position. This arrangement is considerably louder at most angles than the nozzle alone.

The increased noise is caused by reflection of the nozzle noise and impinge- ment of the jet on the retracted flap system. When the flaps are lowered to the 10°-20 ° position, the sound level (square s}znbols) increases further as a result of greater impingement. And finally, the outermost set of data points (circular symbols) is the sound level _ith the flaps in the 30o-60 ° position. It is clear that there is a large increase in noise below the wing as the flaps are lowered into the jet exhaust. For example, directly below the wing at 90 ° from the engine inlet there is a 10-decibel increase in noise level when the flaps are lowered from the retracted to the 30o-60 ° position.

Similar trends can be seen in the sound pressure level 1/3-octave spec- tra (fig. XII-5). Here the sound pressure level a_ainst frequency at 85° from the inlet is shown for the same exhaust velocities. The flap noise spectra are broadband and similar to jet noise spectra. The strong increase in noise as the flaps are lowered is again readily apparent. At the 30°-60 ° setting the flap interaction noise completely dominates the nozzle noise.

The results with the FRC blown flap test facility were generally very similar to the Lewis results. The higher mixed exhaust temperature of the CF-700 engine did not have a significant effect on the flap noise radiation pattern.

Noise Scaling Laws Noise scaling laws are needed to scale the flap noise data up to the full- size blown-flap systems of STOL airplanes. Special blown-flap noise tests were conducted to establish the scaling laws.

For these tests the bypass nozzle assembly used on the 1/2-scale Lewis model was replaced by a 13-inch-diameter round convergent nozzle (fig. XII-6). In addition, a geometrically scaled small version of the 1/2- scale model was constructed. The small model (fig. XII-7) had a 2-inch- diameter nozzle and a wing chord length of 13 inches. The geometric scale factor was 6.5. This means that all 1/2-scale model dimensions are divid- ed by 6.5. The 1/2-scale and the 1/13-scale models were tested at the same nozzle pressure ratios and inlet temperatures. Thus, two sets of flap noise data were obtained that could be compared at the same exhaust veloci- ties. The data were taken at microphone radii of 50 feet for the large model and 10 feet for the small model.

The noise radiation patterns for the two models with 30 °- 60 ° flap settings are compared in figure XII-8 for three nozzle pressure ratios. At each nozzle pressure ratio the 1/2-scale data points are shown by the solid squares. The 1/13-scale data are shown as open circles. The small-model overall sound pressure level data were scaled up to the large-model data as follows: The noise level was assumed to be proportional to the square of the 6.5 scale factor. The difference in microphone distance was accounted for by using the inverse square law. This resulted in a 2.3-decibel correction, which was added to the 1/13-scale data. The comparison shows that nearly identical radiation patterns were obtained with the geometrically similar models.

The 1/3-octave spectra for the two models are compared at 85 ° from the inlet at a nozzle pressure ratio of 1.7 in figure XII-9. In order to make this comparison it is necessary to scale the frequency of the flap noise as well as the level. The 1/13-scale model spectra were scaled up to the 1/2- scale model data as follows: The frequencies were divided by the 6.5 scale

factor. And, the 2.3-decibel power level and microphone distance correc-

tion was added to the sound pressure levels at each scaled frequency. The spectra shown are corrected for ground effects. Again, the agreement be- tween the data from the two different-sized models is excellent.

The effect of nozzle exhaust velocity on flap noise below the wing is shown at 80 ° from the inlet in figure XII-10. Data are shown for the 1/2- scale model as solid symbols and for the 1/13-scale model as open symbols.

The data are shown for the three flap positions and for the nozzle alone.

The nozzle-alone data for both diameter nozzles show the well-established eighth-power velocity dependence. In contrast, the overall sound pressure levels for the flap noise vary with the sixth power of the velocity for both models.

The results shown in figures XII- 8 to XH- 10 indicate that noise meas- urements from geometrically similar acoustic test models can be used to predict the flap noise for full-size externally-blown-flap systems.

Preliminary noise estimates for a full-scale externally-blown-flap sys- tem that differs significantly from the preceding acoustic test model become more difficult. To assist in this type of scaling, research is being conduct- ed on the effects of geometric variations and differences in engine exhaust velocity profiles.

As an example, consider the problem of comparing the 1/2-scale Lewis flap noise data with the FRC data. Because of the different kinds of nozzle air supplies, the nozzle exhaust velocity profiles and decay rates were not the same. However, the data can be scaled by using the impingement veloc- ity profiles measured with aerodynamic rakes at the flap stations. As an example (fig. XH-11), consider two profiles having nearly the same peak velocity (775 ft/sec) at the 60 ° trailing flap station for the test conditions shown. The FRC flap impingement velocity profile is approximately 1.6 times the width of the Lewis profile. The FRC 60 ° trailing flap chord length is 1.5 times the length of the Lewis flap. This is nearly the same ratio.

Thus, an approximate scale factor of 1.5 can be used to scale the Lewis data to the FRC data.

The flap noise spectra for the two tests are compared at 40 ° from the engine inlet in figure XII-12. With a 60 ° flap deflection, the flap interaction noise peaks near this microphone location. The Lewis data have been scaled up to the FRC data by using the 1.5 scale factor and by accounting for the differences in microphone distance by the inverse square law. At frequen- cies above 500 hertz the agreement is very good. These are the frequencies which are important in determining the perceived noise level of an externally-blown-flap airplane. The data below 500 hertz were affected by ground reflections, causing the dips and peaks. When ground-effect correc- tions were applied, the low-frequency data were in much better agreement.

Both sets of data would give very similar noise estimates for full-scale air- plane flap noise.

Now consider a geometric variation in which the nozzle diameter is changed but the wing-flap size is kept the same. Results from small-scale tests, where the peak impingement velocity is fixed at 700 feet per second, are shown in figure XII-13. The variation in overall sound pressure level is given for nozzle diameters between 1 and 3 inches. The noise level is noted to vary directly as the first power of nozzle diameter rather than as diameter squared, which would be the case for geometric similarity.

Next consider the effect of changing engine nozzle location on the noise radiation patterns (fig. XII-14), all other parameters being held constant.

Overall sound pressure levels are plotted as a function of angle from the inlet for three positions of the nozzle. The angles of most interest are those positions directly under the aircraft, that is, 70 ° to 100 °. The solid circu- lar symbols are for the normal position of the nozzle. Moving the nozzle towards the flaps, as represented by the square symbols, increases the noise below the wing because of the higher impingement velocities. Simi- larly, moving the nozzle up to the position represented by the triangular symbols increases the noise even more because of the impingement on the wing leading edge and increased scrubbing area.

Noise Sources The externally-blown-flap noise sources are now discussed in more de- tail (fig. XII-15). It is apparent that the externally blown flap with the en- gine under the wing can generate more noise than the engine alone. The en- gine alone generates internal noise and jet mixing noise. The internal noise comes from the engine machinery, such as the fan. Impingement of the en- gine exhaust jet upon the surfaces of the wing and flaps generates additional noise: scrubbing noise, leading-edge noise, and trailing-edge noise.

It is difficultto separately investigate all these noise sources. These sources, therefore, are discussed in terms of the results of a simple ex- periment (fig. XII-16). The jet from a 2-inch nozzle was directed at three vastly different surface shapes, in turn. These surfaces were a large flat board, a wing with flaps, and finally a slotless wing that followed the inner boundary of the wing with flaps. At a given nozzle exhaust velocity the peak impingement velocity would be the same for each surface. The impingement area and impingement angle were also nearly the same so far as noise gen- eration was concerned. The only major difference was the surface shape.

The results of the experiment are summarized in table XII- 1. The TABLE XII-l. - EFFECT OF SURFACE SHAPE ON NOISE Surface Source of noise Power law Jet Scrubbing Trailing Leading mixing edge edge Nozzle alone X V8 ] Large flat board X X V.8 Slotless wing X X X V 6 to Vi8 Wing with flaps X X X X nozzle alone has only jet mixing noise because there is no internal noise in this experiment. The flat board has the nozzle jet mixing noise and, in addition, there is scrubbing noise, which is associated with air flowing along a large surface. The flatboard is so large that air leaving the edges of the board is of very low velocity. Therefore the board will have no trailing- edge noise. Further, with no leading edges it, of course, has no leading- edge noise. The slotless wing also has jet mixing noise and scrubbing noise.

In addition, trailing-edge noise now becomes important, because the exhaust velocity off this small surface is high. Again there are no leading edges exposed to the flow. The wing with flaps has small surfaces and flap leading edges, therefore it has all the noise sources.

The noise below the "wing, " for the three surfaces, is shown in fig- ure XII-17. This figure illustrates the relative noise level from each noise source. The noise level is plotted as a function of nozzle exhaust velocity; however, it could also have been plotted as a function of the peak impinge- ment velocity for the surfaces. As indicated previously, the noise below the aircraft follows the sixth power of velocity for the small surfaces that make up the wing with flaps. However, the board, with its large surface, follows a much higher power law. These power laws are followed by these two surfaces at all angles measured from the inlet. The slotless wing fol- lows power laws somewhere in between these two limits, depending on the angle. Below the aircraft the slotless wing is quieter than the wing with flaps at low velocity, because it has no leading edges. For all three surfaces the noise from the nozzle jet is small compared to the noise generated by jet impingement.

The noise spectra for the three surfaces are compared at two velocities in figure XH-18. The normalized power spectral density in this figure is plotted as a function of a Strouhal number. The spectra are normalized to take out the variations in the overall noise level which are caused by varia- tions of surface shape and velocity. Only spectral variations can show up on this type of plot. The spectral shapes for all three surfaces are very simi- lar, in spite of the different noise characteristics of the surfaces.

Noise Levels for STOL Aircraft Estimates based on the 1/2-scale model data have been made of the flap noise during takeoff of a four- engine 170 000- pound externally- blown- flap airplane having 94 000 pounds total thrust (fig. XII- 19). The maximum 500- foot-sideline perceived noise level during a 15 ° climb is given as a function of core exhaust velocity. The fan exhaust velocity was assumed to be ap- proximately 0.8 of the core velocity. An airplane having engines with a by- pass ratio of 6 and a fan pressure ratio of 1.5 would have a perceived noise level of 114 PNdB at the 500-foot sideline. This is an unacceptably high noise level for commercial STOL airplanes. Fortunately, the estimates

showthat there is a very large decreasein noise as the engineexhaustve-

locities are decreased. An airplane havingengineswith a very high bypass

ratio of 15 wouldhavea maximumsideline noise of only 95 PNdB. For core

velocities above 700 feet per second, some form of flap noise suppression will be required in order to get the flap noise down to acceptable levels.

Noise Suppression A possible means of noise suppression is to reduce the flap impingement velocity. For the conventional fan jet nozzle, the potential core flow extends back almost to the flap assembly (fig. XII-20). In this case the impinge- ment velocity is nearly equal to the nozzle exhaust velocity.

To reduce the impingement velocity a multitube mixer nozzle can be used. The mixer nozzle (fig. XII-21) takes advantage of the fact that the potential core length is proportional to the diameter of the tube. The short potential core causes good mixing up to the point where the individual jets coalesce into a single larger jet.

Calculations show that in order to obtain the needed flap noise reduction the nozzle should be designed so that the impingement velocity is approxi- mately equal to one-half the nozzle exhaust velocity.

Instead of round tubes, a multilobed, or daisy, nozzle can be used to get the same type of effect.

Flap noise suppression tests (fig. XII-22) were run with the Lewis 1/2- scale model by replacing the bypass nozzle assembly with a seven-lobed mixer nozzle placed at the core exhaust position. Similar tests were run on the FRC externally-blown-flap test facility by replacing the conical nozzle with a daisy mixer nozzle (fig. XII-23).

Results of the Lewis tests are shown in figure XH-24 for a flap angle of 30°-60 °, which is a landing flap setting. The nozzle exhaust velocity is 773 feet per second. The mixer nozzle overall sound pressure level data are shown by the inner curve. For comparison, data taken using a round con- vergent nozzle having the same throat area are shown by the outer curve.

At all angles the mixer nozzle was quieter than the convergent nozzle. For example, there is a 6-decibel suppression at 70 ° from the inlet. The 70 ° angle gives the maximum flyover noise during landing. This is just about the amount of noise suppression that would be predicted from the reduction in impingement velocity when the change in exhaust width is accounted for.

Tests at the I0°-20 ° flap angle used during takeoff showed that the mix- er nozzle did not reduce the overall sound pressure level. Actually, the perceived noise level was slightly higher. This is because the higher fre- quency characteristic of the mixer nozzle itself plays a much more impor- tant role since the flap interaction noise is smaller at low flap angles. The increase in perceived noise level caused by the increase in high-frequency noise was found to cancel out all benefits of velocity reduction.

The daisy mixer nozzle tests at the Flight Research Center gave very similar results for both landing and takeoff flap angle settings.

Means of reducing mixer nozzle high-frequency noise are presently being studied. One method might be to use a thin layer of low-velocity sec- ondary airflow surrounding each lobe. Another means might be to use an acoustically lined ejector shroud around the nozzle.

In addition to the use of mixer nozzles, other methods of flap noise sup- pression, which try to reduce the flap noise at the source, are being exam- ined and tests are being run. They include compliant and porous flap sur- face treatment and the use of air injection next to the surface near the flap trailing edge.

Another approach to obtaining acceptable flap noise levels is to develop engines with very high bypass ratios and low exhaust velocities (see fig. XII-19). As an example, consider the 170 000-pound airplane equipped with engines having a bypass ratio of 15.5. The fan design pressure ratio is I. 25; the area ratio is 7.25. The takeoff thrust at I00 knots is 23 500 pounds force. The exhaust velocities for takeoff and approach are us follows: Takeoff Approach Exhaust velocity, ft/sec Core Fan

The fan duct inlet is treated to give 9.2- PNdB suppression at the 500-foot

sideline. The exhaust duct is treated to give 22.5-PNdB suppression. The fan exhaust duct requires more acoustic treatment because the blade passage frequency tones are louder in the rearward dirqction and because the 120 ° maximum lobe occurs at nearly the same angle as the flap noise maximum during takeoff.

The perceived noise level contours for a four-engine STOL airplane equipped with these high-bypass-ratio engines are shown in figure XII-25.

The contours were estimated from flap noise data and from estimates of the suppressed fan noise spectra. The inner boundary of each contour is deter- mined solely by the flap noise. The outer boundary was obtained by adding the suppressed fan noise spectra to the flap noise spectra at each point on the ground. The height above the runway is given by the upper curves for approach and takeoff. The noise "footprint" shows that this would be a very quiet STOL airplane, considering that powered lift is being employed by ex- ternal lower-surface flap blowing. The 95-PNdB contour for example ex- tends out to approximately the 500-foot-sideline distance and closes within 4000 feet on takeoff.

UPPER-SURFACE BLOWING Another means of obtaining lift augmentation from an externally-blown- flap system is to place the engine above the wing and make provision for the exhaust flow to attach to the upper surface of the wing and flaps. Such an arrangement might look like the artist's conception shown in figure XII-26.

For this aircraft all the primary and bypass flows from the engine are mixed and exhausted from a slot or D-nozzle over the wing. The flow attaches to the wing and is turned dox_nward over the flap. From a noise standpoint this engine-wing arrangement could be very favorable because the wing flap sys- tem can shield the community from some of the noise. The data that are presented for upper-surface blowing configurations are preliminary small- scale data taken at the Lewis Research Center. Additional noise tests are being run at the Langley Research Center. The aerody_mmic performance of these config_rations is presently being evaluated.

There are several ways by which lift augmentation can be achieved by means of exhaust flow attachment. Three possible methods are shown in fig- ure XII-27. In the first arrangement, shown at the top of the figure, the engine is mounted above the wing and the exhaust jet is turned downward to the wing by a deflector. Another way is to cant the engine or its exhaust nozzle so that the flow is directed downward to the wing. The flow is not turned downward during cruise for either of these methods. The last method shown uses a slot or D-nozzle on top of the wing rather than a circular jet.

The jet flow attaches to the wing upper surface immediately upon exhausting from the nozzle.

The noise characteristics below an aircraft employing the upper-surface blowing arrangements described in the preceding paragraph depend on which noise-producing mechanisms dominate. For purpose of discussion let us examine the various noise sources associated with the deflector configura- tion (fig. XII-28). The usual engine-alone noises - the internal noise from the engine turbomachinery and the jet mixing noise - are still there. The presence of the deflector in the jet exhaust flow also generates noise.

Scrubbing noise also arises because the deflected exhaust flow strikes the upper surfaces of the wing-flap system. Finally, there is the noise associ- ated with the exhaust flow over the trailing edge of the flap. In this sketch it is assumed that the gaps between the wing and slots are closed. If these were opened, leading-edge noise and the leakage of noise through the slots would have to be considered.

The small-scale deflector configuration model that was tested is shown in figure XII-29. This model has a 2-inch-diameter circular nozzle and de- flector on top of a 13-inch-chord slotless wing. The wing has a 60 ° trailing flap.

The noise radiation pattern in the flyover plane for this model is plotted in figure XH-30 (open symbols) for an exhaust velocity of 585 feet per sec- ond. These conditions are typical of landing. For comparison, the equiva- lent engine-under-the-wing configuration is also given (solid symbols). Sig- nificantly lower overall noise levels occur below the engine-above-the-wing configuration.

It is also important to examine the noise spectra below the model air- craft using a deflector at these same test conditions (fig. XII-31) in order to obtain an idea of the magnitude of the noise sources. The three curves

shownin this figure represent the spectra for the nozzle alone, the nozzle

with the deflector, andthe nozzle with deflector abovea slotless wing.

Whenthe deflector is addedto the nozzle, the noise level increases over the

nozzle-alonecaseat all frequenciesandparticularly at the higher frequen-

cies. Whenthe wing is added, a significant changein the spectrum occurs.

The high frequenciesare effectively shielded by the wing surface; however,

considerablelow-frequencytrailing-edge noise is generatedas the flow ex-

haustsat the trailing edge. The perceivednoise level is primarily influ-

encedby the high-frequencypart of the spectrum whenthesemodel data are

scaledup to a full-sized aircraft. The low-frequency noise must also be

consideredbecauseof its effect on aircraft andcommunitystructures.

Upper-surface-blowing noise data for the cantednozzle configuration is

shownin figure XII-32. The nozzle is canteddownward,with the nozzle ex-

haustdirected toward the slotless wing in sucha waythat the flow attaches.

The noise spectrum below this model wing sectionis plotted as solid square

symbols for conditionstypical of takeoff. For takeoff the exhaustvelocity

from the circular nozzle is 750feet per second, andthe trailing flap has

been set at 20° . The spectrum for the deflector arrangementat the same

conditionsis also plotted on this figure as opencircular symbols. The spec-

tra for eachof these attached-flowcasesis aboutthe same. As a further

comparison the nozzle was blown over the slotless wing with no attachment.

This spectrum is shown by the diamond symbols. The two attached-flow cases have additional low-frequency trailing-edge noise. For the conditions noted in this figure the spectra for all the wing cases come together at high frequency, regardless of the degree of attachment. Of interest also is that the noise below the model wing is less than the noise from the nozzle alone at high frequency, because of shielding.

Noise measurements (fig. XII-33) have also been obtained on the upper- surface flow arrangement consisting of a slot nozzle placed immediately ad- jacent to the wing upper surface. Noise spectra below the model aircraft are shown here for slot nozzles having slot-width-to-height ratios of 5:1 and 10:1. The exhaust areas of the rectangular nozzles are equivalent to a cir- cular nozzle of 2-inch diameter. The nozzles exhausted over slotless wings having 60 ° trailing flaps and flap-length-to-slot-height ratios of 17, 28, and 58.

All the spectra have significant low-frequency trailing-edge noise. In- creases in the flap-length-to-slot-height ratio generally result in lower noise levels at high frequency because of improved shielding. The quietest arrangement was the 10:l slot with a 58:1 flap-length-to-slot-height ratio.

This arrangement would actually be more typical of the conventional jet flap, where air is supplied from a fan stage through internal wing ducts. The other two configurations shown might possibly be employed with external upper-surface blowing. As a matter of interest, the noise spectrum below the model of the 5:1 slot nozzle is nearly the same as the previously shown circular nozzle with deflector.

Noise radiation patterns, for the same test conditions, are shown in fig- ure XII-34 for the above-the-wing arrangement using the 5:1 and 10:l slot nozzles. Also shown for comparison is the noise radiation pattern for the equivalent engine under the wing. It should be noted that the upper-surface slot-flap arrangements produce significantly less noise than the engine- under-the-wing configuration.

There is some miscellaneous information that can be mentioned about the upper-surface blowing configurations. All the upper-surface blowing data shown were for slotless wing-flap combinations. Open slots will raise the noise level below the aircraft a few decibels. If the circular nozzle of the previous examples is replaced by a mixer nozzle in either the canted- nozzle or deflector configurations, such an arrangement is somewhat quieter.

The noise level below the aircraft also proved to be sensitive to the nozzle location and orientation for all the upper-surface blowing arrangements.

In addition to the shielding of aerodynamic noise, the upper-surface blowing arrangements should also provide shielding of internal engine noise; such as the noise from the compressor, fan, and turbine. In other words, with the engine above the wing, the wing and flaps can shield the community below from some of the internal noise that passes through the nozzle. The results presented in figure XH-35 show the amount of internal noise reduc- tion, or shielding, in the sideline and flyover planes. The data are for the model wing section using a 2-inch nozzle with deflector above a slotless wing with a 60° trailing flap. For this data, internal machinery noise was simu- lated by placing an orifice upstream of the exhaust nozzle. Noise reductions of 4 to 10 decibels were observed directly below the wing. This result is probably conservative. It suggests that less exhaust duct treatment would

be required to reduce the internal enginenoise belowSTOLaircraft with the

engineslocated over the wing.

From strictly a noise standpoint, the small-model tests indicate that the

engine-over-the-wing externally- blown-flap configuration may haveimpor-

tant advantages ovel the engine-below-the-wingconfiguration. However, the

embodimentof this upper-surface blowing conceptinto a practical airplane

requires considerationof other factors, suchas the low- andhigh-speed

aerodynamicperformanceof engine-over-the-wing airplanes. Also the in-

ternal performance of slot and other noncircular exhaustnozzles must be

considered. Althoughthe work completedto datedoesnot indicate serious

problems in theseareas, considerableinvestigation andrefinement will be

required to define a completeupper-surface blowing configuration.

SUMMARY This progress report on externally-blown-flap noise research can be summarized by the following remarks: With lower-surface blowing, the sources of the flap noise are beginning to be understood and the noise scaling laws have been established. Further, progress has been made on suppress- ing the flap interaction noise at the large flap deflections used during landing.

More work is needed to solve the flap noise problem at the low flap deflec- tions used during takeoff. Recent small-scale noise tests of configurations using external upper- surface blowing indicate that engine- over- the- wing confi_urations may be promising; however, large-scale noise tests are needed to determine the noise scaling laws and to better assess the potential of this externally- blown- flap configuration.

EXTERNALLY-BLOWN-FLAP STOL AIRPLANE Figure XII-1 ONE-HALF SCALE EXTERNALLY BLOWN FLAP MODEL Figure XII-2 EXTERNALLY BLOWN FLAP TEST FACILITY Figure XII-3 RADIATION PATTERNS AT 50 FT EXHAUSTVELOCITY: CORE, 76.5FT/SEC; FAN, 582 FT/SEC o 300-600 a 10o-20 o zx 0° (RETRACTED} 270o a PYLON-NOZZLE ALONE OVERALLSOUND PRESSURE dB LEVE 0 ° 1'10/ 100/ 9.0_ _ "_go'--100' flO 1800 90o Figure XII-4 I/5-OCTAVE SPECTRA AT 50 FT EXHAUST VELOCITY: CORE, 765 FTISEC; FAN, .582 FT/SEC 100- 850 FROMINLET SOUND _o CONFIGURATION PRESSURELEVE_

i

dB 300-60o FLAPS 8O -_ 10°-200 FLAPS z_ O 0 (RETRACTED) _BYPASS NOZZLE ALONE I I I 4O 125 400 1250 4000 12.500 C_;- (. 30q(_ FREQUENCY, Hz Figure XII-5 I/2-SCALE EXTERNALLY BLOWN FLAP MODEL 13-1N. DIAMETERNOZZLE C-71 - 2039 CS-(J3083 Figure XII-6 1/13-SCALE EXTERNALLY BLOWN FLAP MODEL 2-IN. DIAMETER NOZZLE r: 7 J - 6_7 Figure XII-7 cs-_ _oso COMPARISON OF LARGE-AND SMALL-MODEL DATA AT 50 FT FLAP ANGLE, 30°-60 ° • 1/2-SCALE DATA o 1/13-SCALE DATA PLUS 2.3dB + OVERALLSOUND,' ....

PRESSURE_VEL 130 120/110 100b--:_'Nl00110$ r120_D0 0 o 1800 90o CS-63145 Figure Xll-8 COMPARISON OF LARGE-AND SMALL-MODEL SPECTRA AT 50 FT FLAPANGLE, 30o-60°;NOZZLEPRESSURE RATIO 1.7; 85 o FROMINLET 110 - olll • I/2-SCALE MODEL ° m _ O SCALED UP FROM 105- gll _O IlI3-SCALE MODEL SOUND

100-- PRESSURE LEVEL, @

Di• dB o 9O 125 400 1250 4000 12 500 4O FREQUENCY, Hz cs-63o9z Figure XII-9 AT 50 FT EFFECT OF VELOCITY ON FLAP NOISE LEVEL 80 ° FROM INLET 130-- - ....030o-60° FLAPS 120- Jd_<> i0o_20o FLA pS OVERALL 110-- J_ 0°(RETRACTED) SOUND PRESSURE _Z__/_[_ NOZZLEALONE LEVEL, dB I //'& _ SOLID II2-SCALE DATA 90 _/ 17" OPEN II13-SCALE DATA I I V8/_ PLUS2.3dB ivy, , I, I1, 400 6011 800 1000 1201) NOZZLE EXHAUSTVELOCITY, FTISEC cs-63o97 FigureXII-1O COMPARISON OF IMPINGEMENT VELOCITY PROFILES 80(] -- 60 o FLAP o EDWARD6200-RPM DATA • LEWIS 1/2-SCALE DATA; PRESSURERATIO, 1.4 ,,-EDWARDS LEWl VELOCITY, 40C 22 IN.-' _ FT/SE C FLAP CHORDLENGTH,1..S PROFILE WIDTH, 1.6 SCALEFACTORS:

I I

16 8 O 8 16 24 RADIAL DISTANCE,IN.

CS-63143 Figure XII-ll COMPARISON OF LARGE-SCALE MODEL EXTERNALLY-BLOWN-FLAP SPECTRA AT IOO FT 400 FROhi INLET; 600 TRAILING FLAP; IMPINGEMENTVELOCITY, 77.5 FT/SEC O EDWARD 6200-RPM DATA 110F [] LEWIS PRES SURE-RATIO-1.4 / (i)0 DATA SCALEDTO EDWARDS 1001-- ° 0 ,_ E]_ CONDITIONS SOUND ]0 OFIooUC}OI_ PRESSURE 0(0_1:1] 11100 E_]I_{Z] dB LEVEL [][i] L_C_{] 7O I I I I I I I :50 100 200 500 1000 2000 _]0010000 FREQUENCY, Hz CS-63286 Figure XII-12 EFFECT OF NOZZLE DIAMETER ON NOISE AT IO FT IMPINGEMENT VELOCITY, 700 FTISEC; 30o-60 o FLAP; 80 ° FROM INLET D / OVERALL T SOUND PRESSURE 10 dB dB / LEVEL, l / I I I I ._ 2 3 NOZZLE DIAM, IN. cs-631oo Figure XII-13 EFFECT OF MOVING NOZZLE WITH RESPECT TO WING EXHAUSTVELOCITY, 965 FT/SEC; NOZZLE DIAM, 2.06 IN.; 30°-60° FLAP /x / / o° - _._ _ - 1_ \ NORMAL _' b POSITION 9t0 o OVERALL SOUND PRESSURE LEVEL dB I 0 :30 60 90 120 150 180 210 ANGLE FROM INLET, DEG cs-G3o95 Figure XIl-14 EXTERNALLY-BLOWN-FLAP NOISE SOURCES / NOISE /-LEADING EDGE

,NTERNA, /_

/ NOISE':' i__.._ IF SCRUBBING ,--_.______ ,, REFLECTION -j "_.j_j_" !: i / / TRAILING EDGE NOISE':' J '_ ':'FLAPSOURCES cs 63o98 Figure Xll-15 SURFACE-SHAPE NOISE TESTS _,/- LARGE FLAT WING WITH FLAPS--, \ \>'::_'/ BOARD '_SLO_,_ss w,.o _':_ \_,, VELOCITY,V: -/ ' ' ' : _ ' : 1 _ , " _ J IMPINGEMENT VELOCITY,V i -/ "_ Figure XIl-16 cs-63o89 EFFECT OF SURFACE SHAPE ON NOISE AT IO FT NOZZLE DIAMETER,2.06 IN. ; IMPINGEMENTANGLE, 600; 80o FROM INLET OVERALL ---o--- WING WITH FLAPS SOUND _" PRESSURE --8-- BOARD ilOo, .

LEVEL, / ----- NOZZLE ALONE dB z_ /s 90-- vS/

/1" I I I

500 600 800 i000 1200 NOZZLEEXHAUST VELOCITY, FTISEC CS-(_3142 Figure XII-II EFFECT OF SURFACE SHAPE ON NOISE SPECTRA EXHAUSTVELOCITY, 10- FTISEC SURFACE 920 600 • [] WING WITH FLAPS NORMALIZED & SLOTLESSWING POWER © FLAT BOARD SPECTRAL -10 DEN SflY, dB -20

-" %

[] 1 I I -30 • 1 1 10 • Ol D STROUHALNO. : f CS-63094 Figure Xll-18 ESTIMATE OF MAXIMUM 5OO-FT-SIDELINE FLAP NOISE DURING TAKEOFF 170,000 LB GROSSWEIGHTAIRPLANE;94,000 LBTHRUST(4 ENGINES); 15 0 CLIMB ANGLE BYPASS RATIO, 6 1 115 -- FAN PRES SURERATIO, 1.5,t'-"_ 110- PERCEIVED 105- NOISE LEVEL PNdB

15 _ _L_.._:_ '- VFAN -08

]. 251" _.:_ VCORE 9O 600 700 800 900 1000 1100 COREEXHAUSTVELOCITY, FT/SEC cs-F_Jz_ Figure XII-19 EXHAUST VELOCITY DECAY CONVENTIONALFAN-JET NOZZLE

I I

Vj Vi _=Vj Figure XII-20 C5- _ 3088 EXHAUST VELOCITY DECAY MUI_TITUBE MIXER NOZZLE ___1%£:_. _,::;'_'_ _'% _'_" : :'C _2_::'_ _ _ 4_,-_:2

}

vi V_ J Figure XII-2] CS-63087 I/2-SCALE MODEL WITH MIXER NOZZLE !

/ C-71-3984 CS- 63082 Figure Xli-22 DAISY MIXER NOZZLE TEST C$-63Z92 OF RADIATION PATTERNS AT 50 FT COMPARISON EXHAUSTVELOCITY,773 FT/SEC; FLAPANGLE, 300-60 ° NOZZLETYPE 270° O MIXER RALLSOUND '_:_ _'_PRES SURE LEVEl.,

oo ;'i'_ _',_ 18oo

1 • 100 _llOl 120 900 cs-63141 Figure XII-24 PERCEIVED NOISE LEVEL CONTOURS 170 000-LB-GROSS-WEIGHT STOL AIR PLANE HEIGHT, TAKEOFF APPROACH FT -1.50 - 1000 - 500 PNdB SIDELINE DISTANCE, FT I 1 I I I t I 1 I 400O .5_)0 -I000 0 I000 2000 3000 -4000 -3000 -2000 DISTANCE FROM CENTER Of RUNWAY, FT CS-63288 FigureXII-25 EXTERNALLY BLOWN FLAP AIRPLANE UPPER SURFACE BLOWING CS-63139 Figure XII-26 EXTERNALLY BLOWN FLAP UPPER SURFACE BLOWING CONFIGURATIONS _,_!_ ....

":V,_,_?: ". DEFLECTOR CANTED NOZZLE Figure Xll-2/ NOISE SOURCES FOR DEFLECTOR CONFIGURATION ,,- DEFLECTOR NOISE INTERNAL NOISE -_ I- SCRUBBING / NOISE _JET MIXING NOISE _TRAILING EDGE NOISE CS-63090 Figure XII-28 EXTERNALLY BLOWN FLAP MODEL WITH UPPER SURFACE BLOWING 2-IN. DIAMETER NOZZLEWITH DEFLECTOR CS-63290 Figure XII-29 COMPARISON OF NOISE PATTERNS FOR EXTERNALLY BLOWN FLAPS AT I0 FT EXHAUSTVELOCITY, 585 FTISEC; NOZZLE DIAMETER, 2.06 IN.; 60°TRAILING FLAP ENGINEABOVEWING 270° DEFLECTOR WITH N_ ENGINEBELOW Wl VERALL SOUND / f_ PRESSURELEVEL, t800 0 o _ EXHAUST CS-63144 Figure XlI-30 NOISE SPECTRA FOR DEFLECTOR CONFIGURATION WITH ROUND NOZZLE AT IO FT EXHAUST VELOCITY, 585 FT/SEC; NOZZLE DIAMETER, 2 IN.; SLOTLESS WING CHORD,13 IN. ; 60o FLAP ANGLE; 80 o FROM INLET 80 DEFLECTOR & NOZZLE SOUND | ^_,.( __,oo_,,.,_-'a,_.A_ WING & DEFLECTOR PRESSURE

7or--Ar_"__p '-_ _-__, NOZZLE

LEVEL, dB 60_ _ _ I --_'NOZZLEALONE 50/I I II I I III I 200 400 i000 4000 1000020000 FREQUENCY, Hz cs-63140 Figure XII-31 COMPARISON OF NOISE SPECTRA AT IO FT ROUND NOZZLE DIAM, 2 IN.; EXHAUST VELOCITY, 750 FT/SEC; SLOTLESS WINGWITH 20 ° FLAP_120 ° FROMENGINE INLET ----- NOZZLE ALONE _---,- 100 -0- NO ATrACHMENT-'7_..T" --B-CANTED NOZZLE _._....

_ __FL_ECTOR SOUND PRESSURE LEVEL dB

- <_/,'

I I I I I I 400 1000 2000 4000 I0000 20000 100 200 FREQUENCY, Hz CS-63287 Figure XII-32 EFFECT OF FLAP LENGTH AND SLOT NOZZLE ASPECT RATIO EXHAUSTVELOCITY,590 FT/SEC; SLOTAREA, 3.4 IN. 2; 60o TRAILINGFLAP; MICROPHONEAT 10 FT; 800 FROM INLET 9O L w 80 • 17 5 SOUND o 28 10 PRESSURE o 58 10 LEVEL, 70- ,_h "o, 60- 50i,,I I I I I I I I00 200 400 1000 20004000 I0 000 20 000 FREQUENCY, Hz cs-63_s8 Figure XII-33 COMPARISON OF NOISE PATTERNS FOR EXTERNALLY BLOWN FLAP MODELS AT IO FT EXHAUSTVELOCITY,590FTISEC, NOZZLEAREA, 3.4 IN. 2, TRAILINGFLAP, 60 o • ENGINEBELOW WING ENGINEABOVEWING A 5:ISLOTWITH Uh" 17 ----_, o 10:.1 SLOTWITH Uh = 28 OVERALL SOUND_ \\ PRESSURE z___x 900 CS-63Z89 Figure XII-!WI SHIELDING OF INTERNAL NOISE BY SLOTLESS WING DEFLECTOR CONFIGURATION WING CHORD, ]3 IN.; 60o FLAP ANGLE; NOZZLE PRESSURERATIO, 1.25 20- _ ,_ 0o __o i0 90°_ FLYOVER SHIELDING, ASPI., , __?_!.:,_-_k _, _ SIDELINE clB -5 3O 60 90 120 150 180 210 ANGLE FROM INLET, DEG cs-G3o99 Figure XII-35 XlII. DESIGN INTEGRATION AND NOISE STUDY FOR A LARGE STOL AUGMENTOR WING TRANSPORT Jack V. O'Keefe* The major noise components of an augmentor wing STOL airplane en- gine are forward arc radiated inlet noise, rear arc radiated jet noise, and augmentor noise, which is directed immediately under the aircraft on a fly- over (fig. XIII-1). The beam patterns of these components are separated sufficiently by directivity that they are nonadditive for peak perceived noise level predictions. The required reductions are plotted in the figure.

Primary jet velocity is the governing parameter for jet noise levels through engine cycle selection. A two-stream engine installation with all the fan air being routed to the wing to provide an 80/20 percent thrust split is shown in figure XIII-2. Internal flow turning duct wall acoustic lining eliminates aft arc fan noise. Inlet noise reduction using a sonic inlet is the subject of the paper by F. Klujber.

The objective of this NASA-Ames - Boeing augmentor program is to de- velop through analysis, design, experimental static testing, wind-tunnel testing, and design integration studies an augmentor wing jet flap configura- tion for a jet STOL transport aircraft having maximum propulsion and aero- dynamic performance with minimum noise generation. The program has three basic elements: (1) static testing of a scale wing section to demon- strate augmentor performance and noise characteristics, (2) two- dimensional wind-tunnel testing to determine flight-speed effects on per- formance, and (3) system design and evaluation that optimizes the complete system and ensures that the design is compatible with the requirements for a large STOL transport having a 500-foot sideline noise of 95 perceived noise decibels (PNdB) or less.

All performance and acoustic tests were performed at the Boeing North Field Mechanical Laboratories, Seattle, Washington. The laboratories have a facility designed for large-scale combined acoustic and thrust performance *The Boeing Company.

test programs. The augmentor thrust is measured with a six-component, platform balance bridged with high-pressure air; the noise can be measured in a 180 ° arc in an acoustic arena (fig. XIII-3). The thrust stand accurately measures model forces using either hot (300 ° F) or ambient-temperature air. Nozzle flow rates are determined with precision using ASME venturi flowmeters calibrated against a Boeing standard nozzle. An acoustically treated muffler plenum, located on the balance platform upstream of the test nozzle plenum, prevents any noise generated by the air supply lines and con- trol valves from reaching the test nozzles. A flap system is shown installed on the test stand in figure XIII-4.

Initial static testing included slot nozzles with aspect ratios between 50 and 400, a convergent-divergent nozzle, and various multielement nozzles with lobe and tube shapes. A representative array of nozzles is shown in figure XIII-5. Subsequent static testing included improved nozzles in full augmentor systems. The results presented in this paper showhow the noise and performance objectives are met with the augmentor.

The static testing included multirow tube and lobe nozzles of array area ratios varying from 4 to 8. Array area ratio is defined as the ratio of total array area to nozzle exit area. Several configurations of augmentor geom- etries with internal design variations were investigated with a range of acoustically tuned linings. Combinations of flap and shroud lengths were tested for their thrust augmentation and noise characteristics. The best configurations for performance and noise used high array area ratio multi- element nozzles in augmentors with symmetrical internal contours.

The achievement of the large noise suppression required to meet the 95-PNdB (perceived noise decibels) noise level depends on a series of care- fully integrated design steps. These are shown starting with a high-aspect- ratio unaugmented slot nozzle and progressing through an augmentor with acoustically tuned lining.

A peak level of 116 PNdB would be measured on the 500-foot sideline at takeoff of a jet STOL aircraft employing a high-aspect-ratio slot nozzle for bypass thrust and a nozzle pressure ratio of 2.6. A suppression of 21 PNdB, relative to such a slot nozzle, is required to achieve the goal of this task.

This suppression is equivalent to a reduction of annoyance by a factor of 4 to 5. The noy-weighted spectrum (fig. XIII-6) of the slot nozzle identifies the

problem area as the mid- and high-frequency range of the spectrum. The

annoyance levels in the 2- to 4-kilohertz bands (to which the human ear is most sensitive) are particulary prominent. While a large array area ratio (AAR) multirow lobe nozzle alone alleviates the midfrequency band level problem (fig. XIII-6), the effect in the critical high-frequency bands is only enough to deliver a suppression of 8 PNdB for the nozzle alone.

Adding an unlined augmentor to the lobe nozzle reduces the noise level in high-frequency bands, but it increases noise in the low-frequency bands and provides a net noise suppression of 10 PNdB (fig. XIII-7). Additional reduc- tions of noise in higher frequency bands are achieved by screech suppression and by using acoustically tuned lining in the augmentor.

The aerodynamically induced jet screech produced by the multirow lobe nozzle is eliminated by extending one side of each lobe six lobe widths (fig.

xm-8).

The basic nozzle suppression with unlined augmentors increases to 12 PNdB, and the tuned lining suppression (relative to unlined augmentor) increases from 3.5 PNdB without screech shielding to 7PNdB when the screech shields are used, resulting in a 19-PNdB suppression relative to the slot nozzle. Matching the core depths of the tuned lining to the frequency distribution along the jet axis and also installing a baffle at the lower secondary air gap results in a total suppression of 21 PNdB (fig. XIII-9).

In summary, the noise suppression available from one of the best con- figurations is illustrated in figure XIII-10. The noise objective of 95 PNdB at the 500-foot sideline is met by this configuration and the static thrust augmentation level (total thrust vector; flaps on/flaps off) is above 1.42 at takeoff flap setting.

An important element in obtaining the 21 PNdB noise suppression is the proper design and application of tuned lining for the angmentor surfaces.

This was accomplished by testing a matrix of seven linings. The results confirm the design procedure. The best lining was in the eye of the matrix and achieved a 7-PNdB suppression relative to the unlined augmentor. The sound pressure level spectrum for this is shown in figure XIII-11.

Although the lining designs are successful, the initial suppression was below the 6 PNdB predicted. The effectiveness of the acoustic lining was initially masked by nozzle screech. Consequently, an investigation was made to develop an effective screech suppressor for the primary nozzles.

The screech shield was devised and applied to the AAR = 6, 172-lobe pri- mary nozzle. The same tuned lining reduced the noise by 7 PNdB relative to the unlined augmentor.

The spectrum and beam patterns of the best lined augmentor tested are shown in figures XIII-12 and XIH-13. The primary nozzle has screech shields; the augmentor has mixed single-layer lining and a lower gap baf- fle. The sound pressure level spectrum is based on the same data as the noy-weighted spectrum of figure XIII-10, and it is compared with a basic slot nozzle. The suppressed spectrum is fiat with no pure tones and varies less than 10 decibels through the frequency range. The maximum perceived noise level is reduced by 21 PNdB relative to the slot nozzle. The beam pattern of the suppressor is highly directional, providing additional advan- tages with respect to the time duration effects for this noise component.

High static thrust augmentation is developed with multirow lobe nozzles of large array height operating with a relatively short angmentor system.

The multielement nozzles in augmentors demonstrate high static thrust aug- mentation through 45 ° of internal flow turning without the need for special boundary-layer control slots on the flap or other devices for energizing the boundary layer.

The relation among the achievable augmentation ratios, nozzle array area ratios, and nozzle ventilation is illustrated in figure XIH-14. As the nozzles increase in length, in array area ratio, and in the number of ele- ments (and nozzle perimeter), the augmentation ratio increases from 1.2 to 1.48.

Since thrust augmentation does not include nozzle internal loss effects, these must be identified to evaluate total airplane performance. The small penalties for nozzles with large numbers of elements are predictable. For example, the 172-lobe nozzle has a velocity coefficient of 0.95 at a design nozzle pressure ratio of 2.6. Suppressor nozzles tested correlate well as a function of hydraulic diameter.

Compromises were made in order to satisfy airplane installation re- quirements. One of the most important factors in augmentor wing airplane designs is the trade-off in augmentor system length. Although increasing the nozzle length significantly improves augmentation (fig. XIII-14), no change in noise suppression was measured. Increasing the flap length, how- ever, provides both acoustic and augmentation improvement (fig. XIII-15).

Within the range of lengths considered, it is important to use as long an aug- mentor as possible.

Figure XIII-16 is a review of the major acoustic predictions, data, and configuration milestones of the program. The objective of future tasks is to improve the noise suppression by 5 PNdB and the thrust augmentation ratio by five counts. The acoustic improvement lies primarily in the area of the application of multielement acoustic linings, while the thrust improvement lies in the area of better mixing primary nozzles, improved nozzle ventila- tion, and refinements in internal augmentor contours.

The projected peak 500-foot sideline noise of the 1978 augmentor wing airplane is 90 PNdB. The airplane takeoff noise "footprint" generated by the 1978 augmentor at maximum STOL takeoff weight is given in figure XIII-17. The 90-PNdB closure point directly under the flight path is located at 7200 feet from brake release. The total takeoff noise area encompassed by the 90-PNdB isocontour is approximately 100 acres. This type of foot- print should be acceptable for a large majority of STOL ports, especially since the noise spectrum will not contain sharp, pure tones. A 90-PNdB short-duration transportation noise level will blend with many community ambient noise levels and result in little or no annoyance.

AUGMENTOR WING STOL AIRPLANE FLYOVER NOISE (PREDICTED)

oo, ' ' / _ _]'_ ' ' 118°°

/ / ,

//

ANGLE TO INLEI Figure XIII-I AUGMENTOR WING PROPULSION SYSTEM FAlsE-" 1 _"__/P'__ AUGMENTOR

I _ " _ _ _ Noise

[ _ '_JET & CORE NOISE Figure XIII-2 CS-63128 MICROPHONE ARRAY NOISE FACILITY Figure XIII-3 NOZZLE FLAP SYSTEM ON NOISE TEST FACILITY \ \ \ Figure Xlll-4 TYPICAL NOZZLE TEST SPECIMENS CS-6._46Z Figure XIII-5 TAKEOFF PERCEIVED NOISE LEVELS WITHOUT AUGMENTOR NOYWEIGHTED SPECTRA TESTED CONFIGURATIONS 0 SLO1 NOZZLE 0 NULTIRO'I LOBE NOZZLE

'I--

IOP_B

_/g cm

I I lO00 10,000 FREQUENCY F_ure XIll-6

TAKEOFF PERCEIVED NOISE LEVELS WITH AUGMENTOR

TESTED CONFIGURATIONS NOY WEIGHTED SPECTRA 0 SLOT NOZZLE [] MULTIROW LOBENOZZLE <> MULTIROW LOBE NOZZLE IN AUGMENTOR MULTIROW LOBENOZZLE |0 PNIIB IN LINED AUGMENTOR I le_P " 10,000 FREQUENCY Figure XIII-7 NOZZLE SCREECH SUPPRESSORS (SCREECH SHIELDS) SHIELD NOZZLE EXIT AREA -----/ Figure XIII-8 TAKEOFF PERCEIVED NOISE LEVEL WITH LINED AUGMENTOR, SCREECH SHIELD AND BAFFLE TESTED CONFIGURATIONS NOY IEIGHTEO SPECTRA r 0 SLOT NOZZLE / 0 NULTIR(YI/LOBE NOZZLE / a MULTIR011 LOBE NOZZLE / F IN AUGMENTOR / A MULTIRON LOBE NOZZLE L IN LINEO AUGMENTOR /

[

I_ UULTIROW LOBE NOZZLE IITH / ) SCREECH SHIELD IN LINED AUGMENTOR l / NITH LOIER AIR GAP BAFFLE I / FREQUENCY Figure XIII-9 TAKEOFF PERCEIVED NOISE LEVELS 0 SLOT NOZZLE 0 IULTIR011 LOBE NOZZLE iZ0 0 MULTIRON LOBE NOZZLE IN AUGMENTOR A MULTIROI LOBE NOZZLE IN LINED AUGNENTOR I% MULT1ROI LOBE NOZZLE WITH I10 SCREECH SHIELD IN LINEO AUGMENTOB WiTH LOWER AIR GAP BAFFLE NOY WEIGHTED SPECTRA 500-FT SIDELINE PRESENT PERCEIVED ,',-OBJECTIVE NOISE LEVEL 4 FUTURE PNdB OBJECTIVE 8O NPR Z.6 FOUR ENGINES T |0 PNdl 20,000 LB/ENGINE (Fn) T T 300"F AIRSPEED lO0 KN i ].00 II)O0 lO.O0_ FREQUENCY FigureXIII-IO COMPARISON OF NOISE SPECTRA UNLINEDVS LINEDAUCCAENTOR, AS MEASURED MODEL SCALE(I/2 LOBENOZZLEWITH SCREECHSHIELDS, MIXED LINING I00 T NPR 2.6 i o i r 0 D0 I 50 FT POLAR UNLINED_ !

9O 172LOB ,E NOZZ_AA_ THIRD-OCTAVE-BAND SPL, dB IRE: 85

__/-_

200PICOBAR) \s / \ \ \ I 1 7(] ------.--..L___-_L_ __ I I cs-_, _,is; 102 2 5 103 2 5 104 2 5 105 FREQUENCY, HZ Figure XIII-II 21-PNdB SUPPRESSOR SPECTRA AND BEAM PATTERNS 2700 130 260°Z'_ ,n^ 320-FT PNL, PNdB 240°_--_i0

2mOT-'-- _/_

20o0 V"-Ik"".,_X/_ \_._i io

o SLOT NozzLE :_ok"_)</_\ 12o

n 172-LOBE, AAR 6, NOZZLE I_LO_//V_J ] 1 j_,._.,.,_"

73,>L o

o EXHAUST 1600_ l I _,li_.^_P NPR : 2.6 Tl- : 300 P AYI(:--" ,ou lm_AolO00

A:,: 770 SQ IN. 5f: 350 ...... OROU,ND-_ "_ 140o i_o120° "'°

CS-_3484 Ill/l/llllllllllll II I Figure Xlii-12 21-PNdB SUPPRESSOR SPECTRA AND BEAM PATTERNS THIRD-OCTAVE- BAND SPL, dB iRE: O. 0002 MICROBAR)

I ""

70" " I i I i I i J , , CS-63483 63 125 250 .500 103 2 4 8 FREQUENCY, Hz Figure XIII-13 PRIMARY NOZZLE GEOMETRY EFFECTS AUG,'VENTOR GEOMETRY B AUGMENTOR GEOMETRY A N%R Z O0 2.6 " 150 172 LOBE NOZZLE NPR - 2.6 140 TUBE NOZZLE AAR-8 ,-_ 1.5 _ - 1._..

/WAR - 4 - x _ 1.4 THRUST 1.3 -- 1.27 AUGMENTATION, ].2 1.1 i.20 C5-63486 I.O Figure XIII-14 AUGMENTOR LENGTH EFFECTS 1.50 1.40 PNdB// 1.30 15 // _// THRUST _LPNL SUPPRESSION, AUGMENTATION.

PNdB, RE: BARE SLOT NOZZLE |0 IIit// 1.20 1.10 0 1._ i i i i i 0 20 40 60 80 CS-63485 AUGMENTOR LENGTH/NOZZLE EQUIVALENT SLOT HEIGHT. L_h E Figure XIII-15 AUGMENTOR WING NOISE LEVELS 150-PASSENGER STOL FOUR ENGINES -- -- PROGRAN OBJECTIVE (TASK I) ZO, O00 LB/ENGINE (Fn ) x2g FLIGHT CONDITIONS IlO0 KN) -- ESTIMATED (10/70) TEST DATA • SLOT NOZZLE, WITH LOWER LIP EXTENSION ll{ 0 70-LOBE, AAR m 4 _O-FT • 70,-LODE, AAN == 4 IN.

SIDELINE AUGIENTOR. SOLID PNL, SURFACES, OPEN GAP PNdB • 70-LOSE., AAR m 4 IN.

AUGNENTOR, BULK LINING, / ___o_ A OPEN GAP .L'_ ....

0 172-LOBE, AAR = $ IN.

IN AUGNENTOR, TUNED LINING d. 17Z-LONE, AAR 8 IN.

AUGNENTOR, TUNED LINING I II B), OPEN GAP O 172 LOQES, AAR= 6, 1 I J SCREECH SHIELDS IN AUGMENTO_, Z.0 Z.S 3.0 MIXED TUNED LINING, LON'ER NOZZLE PRESSURE RATIO SECONDARY GAP BAFFLE Figure XIII-16 TAKEOFF FOOTPRINTS FOR TWO-STREAM AUGMENTOR WING AIRPLANE 3 x 103 Z ALTITUDE, FT I I I I 0 2 O,STAN E RON BRAKE RELEASE, 4 6 8 10 12 14 lb x 103 0 I__FT SIDE'_INE 100 ENd8 ' / ' ' / ], ---- - LATERAL 80 PNdB (AREA _ 500 ACRES)* DISTANCE, FT 2 t 90 PNdB (AREA :100 ACRES)* 3 x 10 3 FOUR STF 395D (BN-I) ENGINES AT 18,b00 LB THRUST (SLS) AVERAGE GRADIENT 0. 149 AVERAGE AIRSPEED 91 KTAS +AREA MEASURED FROM BRAKE RELEASE Figure XIII-17 XIV. SONIC INLET DEVELOPMENT FOR TURBOFAN ENGINES Frank Klujber* STOL propulsion installation requires substantial inlet noise reduction to satisfy anticipated airplane noise requirements. Depending on the engine and airplane configuration chosen, a 15 to 30 perceived noise decibel (PNdB) inlet noise suppression is required to match other reduced noise source levels.

As indicated in figure XIV-1, the augmentor wing airplane requires an inlet noise suppression of 25 to 30 PNdB. For such high levels of inlet noise suppression, sonic or high Mach number inlets become attractive can- didates. In addition to the augmentor wing propulsion system, a broad range of applications of sonic or high Mach number inlets is foreseen for future propulsion systems.

Suppression of engine inlet noise by use of a sonic throat has been under consideration for a number of years. A limited amount of testing has been conducted by NASA, the Boeing Company, and others. However, there has been a need for a systematic investigation of fundamental noise suppression characteristics, evaluation of the several candidate configurations, and a study of the design and operational characteristics required for practical flight installations. Such a program was undertaken in January 1972 by the Boeing Company under contract to the Lewis Research Center.

The basic acoustic principle of sonic inlet operation is that sound waves are attenuated while traveling upstream in a high Mach number flow stream.

The limit case is achieved when a complete sonic plane is generated. In this case, theoretically, no sound waves can travel past the sonic plane (fig.

XIV-2). The aerodynamic principle of operation is that the flow is acceler- ated to produce a high velocity throat and then is diffused to provide accep- table flow velocities to the engine.

There are two basic airplane operating conditions under which low noise * The Boeing Company.

(i.e., choked inlet mode) is desired for community noise reduction - takeoff and landing approach. Since the engine power settings for these two condi- tions are different, the area must be variable for maximum noise reduction.

Based on the flow area requirements for the critical operating condi- tions, one can establish the basic requirements for the design of a sonic inlet: (1) Provide throat area reduction in the inlet to produce sonic or near sonic flow conditions.

(2) Provide a diffuser beyond the throat to reduce flow velocity to the engine inlet.

(3) Provide area variability for takeoff, approach, and cruise.

In order to arrive at an optimum choice for an operationally and eco- nomically viable inlet, the merits of each design must be determined by the amount of noise reduction balanced against other technical requirements.

The following are the most important of these requirements: (1) Minimization of inlet flow recovery losses (2) Minimization of flow distortion (3) Minimization of external drag (4) Minimization of weight (5) Mechanical feasibility An added requirement for the best sonic inlet choice would be to satisfy these performance requirements under actual operating conditions of angle of attack and crosswind.

Many sonic inlet configurations could satisfy the previous design re- quirements. These inlets fall in two basic categories, namely, single pas- sage and multiple passage types. Examples of single passage inlets are the following: (1) Translating centerbody (2) Expanding centerbody (3) Contracting cowl wall Multiple passage inlet types include the following: (1) Movable concentric rings (2) Movable radial vanes, etc.

Boeing is currently conducting a program under an NASA Lewis contract for sonic inlet configuration selection and design technology development

(fig. XIV-3). The first phase of the current program consists of designing

and testing different inlet models in order to select the most promising con- cepts for further development (fig. XIV-3). After these initial screening tests two concepts will be further refined and optimized.

The test program is being conducted in an anechoic chamber (fig.

XIV-4) where inlet noise is separated from other fan noise sources and background noise is minimized. This arrangement allows evaluation of the actual noise reduction potential of sonic inlets. Detailed state-of-the-art aerodynamic design and instrumentation also allowthe comparison of differ- ent inlets in terms of their noise reduction potential and aerodynamic per- formance (i.e., pressure recovery and flow distortion).

A baseline inlet (fig. XIV-5) has been tested in the program to study the near field noise attenuation characteristics of sonic inlets. This inlet was designed with adequate length to ensure good aerodynamic performance and boundary layer separation free operation. Instrumentation included a line of wall static pressures, boundary layer rakes, and inlet wall kulite probes for noise measurement (fig. XIV-5). Total pressure rakes were traversed at the exit plane of the inlet to obtain recovery and flow distortion data. This instrumentation is typical for all test models.

The baseline inlet has been probed in the near field with a specially con- structed probe for continuous measurement of noise and static pressure in- side the inlet (fig. XIV-6). The probe is mounted on an X-Y traverse mech- anism to enable continuous recording in both axial and radial planes.

Typical Mach number and blade passage frequency noise contours are shown based on the axial traverses at various radial positions (fig. XIV-7).

The data show that the noise level at the fan face is greatly dependent on radial position. The hub noise levels are the lowest, and the highest values are measured at the tip. In the throat region rapid noise reduction takes place when the Mach number exceeds 0.7 with the minimum noise measured at the throat. Radial noise gradients disappear at the throat.

Specific attention was centered on the attenuation of shock waves and multiple pure tones by sonic inlets at supersonic fan tip speeds. Noise mea- surements were taken near the fan face, downstream of the throat, and in the throat. The spectrum comparisons for these measuring points indicate that all pure tones were effectively attenuated by the sonic inlet (fig. XIV-8).

Another point of interest is the possible noise generated by the sonic

plane itself. Therefore, measurements were also taken ahead of the shock plane (fig. XIV-9). The comparisons indicate that no significant noise in- crease was experienced because of noise generation by the sonic plane for the test conditions investigated.

Testing to date has been completed on several inlet models. Single passage inlet test results show (fig. XIV-10) that a 28 to 30 PNdB inlet noise reduction at approach engine power setting can be achieved by sonic inlets with 0.97 inlet recovery within acceptable inlet distortion limits. The test data show that sonic inlets within realistic length limits (L/Y) = 1) can be de- signed to achieve high noise reduction and acceptable aerodynamic perform- ance on a static performance basis. Takeoff configuration tests have shown similarly good results for approach on the static tests (fig. XIV-11). The measured single-passage-fan face total-pressure distortion levels are shown for a range of Mach number conditions (fig. XIV-12). With the exception of the last test condition shown, no flow separation is indicated by these mea- surements. A further investigation of the flow separation phenomenon at a high Mach number and low recovery operation is planned to extend the range of operating envelope of these inlets.

One multipassage sonic inlet has been tested to date. This inlet was a radial-vane-type sonic inlet with 36 retractable radial vanes forming the sonic plane. Acoustic and aerodynamic data from this test are shown in comparison to the L/D = 1 centerbody inlet on figure XIV-13. The multi- passage inlet produced lower noise reduction with significantly lower pres- sure recovery.

In spite of the poorer static performance of the multipassage inlet, in- terest will be continued in these configurations until wind tunnel evaluation of the different concepts can be made. Multipassage inlets are believed to be less sensitive to angle of attack and crosswind conditions.

Spectral comparison of the noise for the baseline and sonic inlet indi- cates that some dependence of attenuation effectiveness exists as a function of frequency (fig. XIV-14). Pure tones of high frequency and high frequency broadband noise are more effectively attenuated than the tones of low fre- quency noise. This phenomenon seems to be Mach number dependent, indi- cating that flow velocity effects may interact with the direction of noise 3O8 propagation. Thus, plane waves (low frequency noise) traveling in the axial direction will propagate from the inlet at high flow Mach numbers, while traversely propagating modes (high frequency) get reflected at lower inlet flow velocities. Further analytical work is required to fully understand and quantify these relations.

Noise directivity measurements taken in the acoustic chamber with ade- quate inlet sidewall insulation and other noise sources minimized (fig.

XIV-15) indicate that the noise is effectively reduced at all angles in the for- ward arc. This result helps to clarify some questions with regard to side- line (90 ° ) effectiveness of sonic inlets.

In summary, the static program has shown to date that very large noise reductions can be achieved by the sonic inlet concept with realistic inlet length and good aerodynamic performance. It is also shown that different inlet concepts produce substantially different results. It should be empha- sized that the conclusions drawn here are based on static performance eval- uation necessitated by the current status of the program. Therefore, ade- quate caution must be exercised until distortion and angle of attack effects are fully determined on the performance of these inlets.

3O9 TYPICAL SONIC INLET APPLICATION AUGMENTOR - WING QUIET NACELLE 12.0 _ ,UNSUPPRESSEOINLE T NOISELEVEL _-" APPROACH MODE CRUISE MODE 500 FT 105 SIDELINE NOISE - PN/8 SONIC INLET (4 ENGINES) NOISEREDUCTION 06JECTIVES (25-310 A PNdS) ENGINE PRIMARY NOISE AUGNENTOR NOISE BY CYCLE BUSHED _ SU PPRESSORTARGET Figure )GV-!

SONIC INLET OPERATIONAL REQUIREMENTS AND DIES IGN CONSIDERATIONS SONIC PLANE __OIFFUSER / !

AND CONTROL MAIN DESIGN CONSIDERATIONS : • MINIMUM LOSs MINIMUM FLOW 01STORTION • LIGHT WEIGHT • MECHANICAL FEASIBILITY AN0 SIMPLICITY Figure XIV-2 NASA/BOEING SCALE MODEL SONIC INLET PROGRAM FINAL MODEL TEST SCREENING TESTS SINGLE

':st,?

MULTi- __-- " ......

Figure XIV-3 12-iNCH RIG MICROPHONE ARRAY Figure XIV-4 SONIC INLET NOISE MECHANISMS TEST INSTRUI/£NTATION 0 STATIC PRESSURE PORT • HIGH FREQUENCY RESPONSE TRANSDUCER BOUNDARy LAYER FIXED RAKE ROW OF 4 KULITES ROW OF ].8WALL _ FLUSH MOUNTED STATICS AT 355/ AT _._

" J

! _-m'TH ROA T . TRAVERSE _ "" "'" ,_¢--,d'_ TOTAL // [ / \ PRESSURE __ ___ _)_ TRAVERSE RADIAL PLAME COVERED BY KULITE & PS TRAVERS PROBE Figure XIV-5 ]2-INCH FAN RIG SYSTEM i Figure XIV-6 EFFECT OF RADIAL AND AXIAL LOCATION ON FAN TONELEVEL THROAT 1.0

RAO,O..,.CH

INCH STREAJJ BACH BACH NO. = .7_ NUMBER AVERAGE THROAT RADIUS 4 INCH

s_I /2'NCH

FAN TONE SPL ~ dE

//CERTERL,NE

. _ ; ' , -3 _ _ 6 0 uPSTREAM DOWNSTREAM AXIAL DISTANCE FROM THROAT PLANE (iNCH) Figure XlV-7 NOISE SPECTRUM COMPARISON BEHIND A SONIC INLET BLADE PASSAGE d FREQUENCY • --MULTIPLE PURE TONES l / / /_l/_ THROAT WALL MN == 1.06 ® SOUND PRESSURE LEVEL ~ dB 120 • I ZO I0 FREQUENCY ~ KHz CS-63507 Figure XIV-8 NOISE SPECTRUM COMPARISON AHEAD OF A SONIC INLET 130[ r THROAT WALL i N sl.06 PRE_JRE (_ LEVEL SOUND ]20 / - dO 110 I

t

,L I I I 0 5 10 15 FREQUENCY - KHz Figure XIV-9 SOMC IM.Lrl " PERFORMANCE - APPROACH PNL REDUCTION A PNdB _0 AT 500 FT S.L.

4°fj

O CONTRACTING COWL WALL L./O • _.0 D CENTERBOD¥ L/D • 1.3 INLET RECOVERY .q_ X CENTERBOOY L/D • l.O .% 1 I .lo I .5 • 3 .B .9 1.0 NONMALIZEO THROAT MACH NO.

Figure XlV-lO SONIC INLET PERFORMANCE - TAKEOFF ~ A P_B AT 500 FT S.L ZO A CONTRACTING COWL NALL L."D • 1.0 41, CENTERBOOY L/D • 13 INLET RECOVERY .t .% .10 .05 PTAv G (PT MAX - PTMIN ) f I - DISTORTION '_ L | i I ' .5 • 3 .R .9 1.0 NORMALIZED THROAT MACH NO.

Figure XIV-11 SONIC INLET RADIAL PRESSURE PROFILE 5.5 5.0 4_ RAOIUS - INCHES 3.5 MACHTHROAT RECAvG .5_ .9% O 3_ .994 4.0 (3 ._1 ._ .M9 .990 3.0 C, V ._9 .964 ._5 ._0 O 2..5 .U5 ,966 # i HUB 2.0 .06 .qO ,_ ._ ._ ._ 1.00 TRAVERSE PROBE RECOVERY CEMTERBOOY APPROACH CONFIGURATION INLET L/I) • |.3 Figure _IV-12 PERFORMANCE COMPARISON OF SINGE AND MULTIPASSAGE SONIC INLETS - APPROACH PNL REDUCTION - _ PNd8 ZO

AT 500 FT S.L. '° f

0 ,, L_,._....._-aL-__;; I x CENTER80OV L 0 • 1.0 INLET RECOVERY RADIAL VANES L/D= L0 .% .OS (PTNA x - PT MIN ) PT AVG _STORTION i , .5 .6 .7 .8 .9 1.0 NORMALIZED THROAT MACH NO.

Figure XIV-13 NOISE SPECTRUM COMPARISON la] APPROACH /---BASELINE L'3 O.B. SPL AT 50 ° FROM INLET CENTERLIHE ~ dB

fCEHTER ®Y O.,C,.LET

6O _U H . 1.0 4O , A , I . , i I _ l 2L5 1.0 1.6 2 5 4.0 63 10 16 40 FREQUENCY KHz (hi TAKEOFF 8ASELINE I00 L_ O.B. SPL AT 50 ° FROM INLET CENTERLINE ~ dB ODY SONIC INLET N N • ._ i i '. 4.'0 I , i i 4'0 1.0 1.6 2 5 63 10 16 Z5 FREQUENCY KHz Figure XlV-14 FAN TONE DIRECTIVITY COMPARISON ta) APPROACH • BASELINE • CONTRACTING COWL WAL L L/D • 2.0 O.B. FAN TONE SPL O CENTERBOOY L/D • 1.3 ~ dB 80 _ mN ~ .9 x CENTERBODY L/D = 1,0 bO U N ~ .9 NN ~ 1.0 40 • i i l i i I i , 0 lO 20 30 40 50 50 70 80 gO DtRECTIVITY: DEGREE FRON INLET CENTERLINE (b} TAKEOFF • BASELINE 1/3 (3.8. FAN TONE SPL A CONTRACTING COWL WALL L/D • 1.0 a...

. dB 0 CENTERBODY L/D = 1.3 ld N ~ LO M N 1.0 ; 1'o ;o ;o,o 5'o ,'o ;o ,o_ DIRECTIVIT Y: DEGREE FROM INLET CENT ER L IN E Figure XIV-15 E-6978 _ 2

Overview

Key points

Document

CASE

AIRCRAFT ENGINE

NOISE REDUCTION

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

AIRCRAFT ENGINE

NOISE REDUCTION

like this were used on many airliners. More complicated nozzles (fig. I-2)

in the joint DOT-NASA Civil Aviation Research and Development Study re-

Machnumber, fan pressure ratios of 1.9 to 2.1 wouldbe of interest (cross-

It would appearthat evenhigher bypassratios may be desirable for

lower fuel consumption;however, the enginesbecomevery large. This is

illustrated in figure II-2; the lower-pressure-ratio fan enginesare asso-

ciated with lower fan jet velocities and, as shownin the sketch, are very

large. At the higher pressure ratios, substantially higher fan jet velocities

occur but the fans are relatively smaller and of lower bypassratios. The

higher-pressure-ratio fans may require two or more stages. The low-

pressure-ratio fans can be single stage, but they require a large nacelle and

installation drag becomesa large fraction of the thrust. The large fans

must be driven by the relatively small-diameter turbine. This may require

increasing the number of turbine stagesto achievethe necessarypower from

the small core flow. This increasedengineweight must be consideredas a

The fan must also be suitable to meet all the required engineoperating

conditions. This can be illustrated by referring to a fan performance map

such as shownin figure II-3. Flow curves are given for three rotational

Abovethe cruise operatingpoint, the fan encountersstall andbecomesun-

stable. Similar pressure ratio flow curves are indicated at the speedre-

quired for takeoff andat a lower speed. The fan must producethe desired

pressure andflow alongthe cruise andtakeoff operatingconditions The

positions of the takeoff andcruise operatingpoints dependon the cruise

Machnumber andthe enginecycle. Theseoperatingconditionsshouldbe in

the region of high fan efficiency. They must also be sufficiently removed

from the stall line to allow engineacceleration andtransient operation. The

margin betweenthe operatinglines andthe fan stall must be sufficiently

large to take into accountinlet flow distortions, as shownschematicallyin

figure II-4. Crosswindsandinlet flow distortion causethe stall line to be

movedtoward the operatinglines. There are a number of additional reasons

for building or selecting the fan design to havea high stall margin. These

includethe eventualperformancedegradationwhich results from blade for-

eign object damageandprovision of a thrust growth potential. Enginethrust

growth requires an increase in either fan flow or pressure ratio. Fans de-

agatesthrough the inlet andfan dischargeducts, andpropagatesto the

groundfrom both inlet anddischargeduct planes. Additional important en-

gine noise sourcesare indicatedandinclude let mixing noise from fan and

primary streams, turbine noise, and core compressor noise radiated for-

During an airplane flyover, noisesfrom theseenginesourcespeakat

different instants andcontribute to a complicatedhistory of the flyover noise

event. Sinceall enginenoise sourcesare also presentduring static ground

noise tests, special meansmust be usedto separatefan noise so that its

characteristics canbe studied. Someof the special test rigs usedin fan

noise programs will be described. Throughtheir use, supplemented by

Fan noise radiates in a complicateddirectivity pattern from the inlet

anddischargeducts. At eachangular position in the far field the noise is

convenientlydescribedby meansof its spectrum, or frequencydistribution

of acoustic energy. Figure II-9 gives two illustrative spectra of forward-

radiated fan noise. In the top half of this figure, which is typical of

subsonic-blade-tip-speedsingle-stage fan operation, a broad baseof con-

This noise is generatedby random pressure fluctuations in the fan and

sounds,generally, like jet mixing noise. Prominent features of the spec-

trum are the two sharppeaksor "spikes" at fan bladepassagefrequency

andits harmonic. Thesediscrete tones correspondto fan whineandare

At supersonicfan bladetip speeds,noise of a different character ap-

pears. It is exhibitedin figure II-9 as a closely spacedseries of discrete

frequency spikes separatedby shaft rotation frequencyandhas a distinctive

sounddescribedvariously as deep,rich, raspy, or buzz saw. Multiple-pure-

tone (MPT) noise, combination-tonenoise, andbuzz saware the usual de-

scriptive terms. In brief, it results from the rotating pattern of shock

wavesfrom the fan bladeswhentheir tip speedis supersonic. Becauseof

the nonlinear propagationbehavior of shockwaves, slight variations in the

shock structure from bladeto blade, causedinitially by small manufacturing

deviations, are amplified with distance from the fan into a pattern having

markedirregularities. Sincethis irregular pattern turns with the rotor, its

frequencyis that of shaft rotation speedandall multiples thereof rather

thanbeing restricted to blade passagefrequencies, which wouldresult from

These noise spectral features maybe incorporated in a subjective

rating of the noise exposureby meansof units suchas PNdBin a manner

describedpreviously. But a better understandingof fan noise details for

purposesof noise reduction can only be reachedif the processesthat pro-