ICASPAPERNO.18

TECHNOLOGICAL ADVANCES INAIRFRAME-PROPULSION INTEGRATION

byDemetrius Zonars, Chief Scientist

Air Force Flight Dynamics LaboratoryWright-Patterson Air Force Base, Ohio USA

TheEighthCongressOfthe

InternationalcounciloftheAeronauticalsciences

INTERNATIONAAL CONGRESCENTRUM RAI-AMSTERDAM, THE NETHERLANDSAUGUST 28 TO SEPTEMBER 2, 1972

Price:3. Dfl.

á

TECHNOLOGICALADVANCES IN AIRFRAME-PROPULSION INTEGRATION

D. Zonars, Chief ScientistAir.Force Flight Dynamics Laboratory

Air Force Systems CommandWright-Patterson Air Force Base, Ohio

Abstract

F-111A aircraft inlet system and TF-30 enginecompatibility is reviewed based on an assessment oftime averaged and instantaneous distortion parame-ters. In addition, recent advances in research oninlet configurations associated with steady-stateand dynamic distortions are presented. A completerandom data acquisition, editing and processingmethod is described for accomplishing data analysisas an inlet flow diagnostic tool. Finally, recentafterbody and nozzle research results, which improvethe technology base for understanding airframe-nozzle interactions, are reviewed. A basic aircraftconfiguration incorporating a common forebody, wing,inlet system and a twin engine installationwasutilized during high Reynolds number wind tunneltests to determine the relative merits of a widespectrum of afterbody-nozzle geometrical variations.

I. Introduction

Over the past twenty years, both military andcivilian flight vehicle sophisticationhas drasti-cally increased as a result of the ever prolifer-ating demands for improved performance. Thisincrease in aerospace system sophistication hasbeen possible through new and rapidly emergingtechnologies including the advent of new designtechniques and facility testing methods.

The continuing degree of interest displayed inairframe-propulsionintegration is not surprisingsince this technical domain involves criticallyimportant influences which strongly impact vehicleperformance. More importantly, this area has beenunder significant scrutiny due to the difficultiesexperienced in the operation of tactical type air-craft. The most severe problem has been enginesurge or compressor stall due to steady-state anddynamic distortion emanating principally from theinlet. On the other hand, acknowledgment of theexcessive drag associated with inlet and nozzleinstallationshas been of a subtle nature with lowlevel attention given to this subject until recently.In both cases, variable geometry systems have beenemployed to accommodate the wide range of mass flowand pressure ratio characteristics required formatching airframe and propulsion systems.

The transonic flight regime still appears asthe most difficult portion of the speed range inwhich limited analysis methods can be applied topredict airframe-nozzle interactions. By necessity,one must turn to experimental means in order todevelop a viable and correct determination of thephenomena involved. The AF Flight Dynamics Labora-tory has undertaken a number of programs to inves-tigate the effects of air flows about fuselage andfuselage-wing configurations throughout the subsonic,transonic and supersonic speed regimes. A portionof this work is reported herein.

II. F-111 Inlet Flight Experiences

Advanced tactical aircraft are required to per-form a number of missions which demand a high de-gree of airframe propulsion integration includinglow flow distortion over a much larger range ofoperating conditions (Mach number, altitude, angleof attack, engine mass flow) than previous super-sonic tactical aircraft systems. Requirements formaneuvering flight in a low drag configurationnecessarily implies high angle of attack flightattitudes from subsonic to supersonic speeds inexcess of Mach number 2.0.

It has been, therefore, quite natural to utilizean inlet design for the F-111A which takes advantageof the flight vehicle fuselage and wing to reducethe effects of angle of attack and angle of yawduring maneuvering flight. The F-111A inlet shownin Figure 1 is an external compression 88 degreesegment of an axisymmetric inlet which is integratedwith the airframe fuselage-wing root intersection.Locating the inlet in the wing-fuselage flow fieldalso provides precompression for the inlet flow insupersonic flight which means that the inlet cap-ture area is reduced from that required at freestream conditions. Further, a significant vehicleweight savings is realized by integrating thesupporting structure of the inlet and relativelyshort duct with the vehicle structure.

The spike system of the inlet translates foreand aft and the second cone angle varies with flightMach number and angle of attack to vary the inletthroat area. Each of the inlets is matched to aPratt and Whitney TF-30-P-3 afterburning turbofanengine. The modulated afterburner improves thetactical ability of the F-111A by providing a vari-able thrust output in afterburner mode upon demandby the throttle. Therefore, in addition to beingclosely integrated with the airframe, the F-111Ainlet system is closely integrated with the engineto accommodate variations in airflow demand duringengine transient operation.

During prototype flight tests of the F-111A, itbecame apparent that the desired flight envelopewas restricted. Maneuverability of the aircraft athigh subsonic speeds and supersonic speeds was beinglimited by a rapid buildup of steady and dynamicinlet flow distortion resulting in engine compressorstall. This incompatibility of the inlet and enginein the F-111A aircraft was the impetus for a compre-hensive evaluation of flight test and wind tunneldata to identify the causes of the compressor stallsand define modifications to the inlet system toreduce the incidence of compressor stalls in bothsteady state and maneuvering flight.

In order to identify problem areas and suggestmodifications to improve the inlet system and itscompatibility with the engine, the inlet of a proto-type airplane was equipped with diagnostic totaland static pressure instrumentation in the inletand engine. In addition, the engine compressor

FIGURE1. AFT VIEWOF TYPICALF-111AINLET

face was equipped with 40 total pressure probes incentroids of equal areas to map the total pressurepattern entering the fan and low presAure compress-or of the engine. There were eight rakes with fiveprobes per rake.

Initially, a theoretical study was undertakento determine the anticipated steady state distor-tion at the compressor face during supersonicflight. The purpose of this study was to compare

uniformly expanded into a circumferentialprofile atthe compressor face. A plane of symmetry wasassumed half way between the 88 degree sector of theinlet system. The total pressures resulting fromsuch calculations were plotted and compared withthose of the 1/6 scale wind tunnel and flight case.Figure 2 shows such a comparison. The wind tunneland flight data showed a remarkably similar profile,however the comparison with theory is understandablydifferent due to the absence of viscous effects.

such information with similar 1/6 scale wind tunneland actual flight test data and thus provide in-sight as to trouble areas resulting from specifictheoretical-experimentaldifferences.The condi-

tions of Mach number 2.2 and angle of attack of5.5 degrees with an initial inlet cone angle of12.5 degrees and second cone angle of 24 degreeswas chosen.The theoretical approach first con-

sisted of estimating the Mach number aft of theconical wave system resulting from the fuselage-wing glove intersection.This yielded a Machnumber of 2.08 wherein the flow field was assumedto be uniform to the inlet.An exact Taylor-

Maccoll solution was developed for the initial cone

More importantly,clearly identifiablelower portion ofdifferent thantheory. It wastion should beand duct flow relatingairflow. Valuespresented as derivedfollows:

5

i=1

the main differencelow energy

the compressorthat predictedtherefore clear

undertaken toto this

of steady statefrom the

ttavmin

stems from aarea on the inboard,face which is quite

by the inviscidthat an investiga-

survey the oncomingportion of the entiredistortion are also

expression K.DA as

-

Ptav-P

i= 1ci

angle with subsequent use of the method of char-acteristics to generate the flow field about thesecond cone. A normal shock was then assumed atthe entrance to the cowl lip. The resulting iso-bars for this 88 degree inlet segment were then

2

THEORY 1/6 SCALEWINDTUNNEL FLIGHT

tr. LEFTHANDENGINELOOKINGAFTPARAMETER

AIR FLOW,LBSISEC.-11-14-Hi-11-AVE.RECOVERYPRESSURERATIO

DISTORTIONFACTOR,KDA

THEORY WINDTUNNEL FLIGHT

168 168 16290.8 85 87

100 383 428

FIGURE2. COMPARISONOF F-1114COMPRESSORFACETOTALPRESSURESFOR MACHNUMBER2.2 and AN ANGLEOF ATTACKOF 5.5 DEGREES

whereC = ratio of compressor inlet radius to ring

radius

= number of ring

8-= largest continuous arc of the ring overwhich the total pressure is below the ringaverage pressure

P = ring average total pressuretav

Pt= ring minimum total pressuremin

Here again, the low theoretical KDA value is due tothe absence of the low, inboard impact pressuresand the inviscid assumptions.

In the analysis of the flight test data takenwith this instrumentation, several approaches wereemployed. First, compressor face total pressuremaps were compared, which showed the changes inflow distortion as a compressor stall condition wasapproached. Although this analysis indicated a lowtotal pressure region on the inboard side, theresults were inconclusive and so time variations ofdata from other sets of instrumentation further up-stream in the duct were examined for many stallsequences in order to identify problem areas in theinlet flow field as they developed. From the timesequence plots, selected data for a particular timecut were used to define duct static pressure dis-tributions or boundary layer total pressure pro-files. The static pressure distributions were usedto locate shock wave positions, indicate boundarylayer bleed effectiveness, estimate stream veloci-ties, and indicate regions of separated flow.Total pressure profiles were used to define regionsof low energy flow ahead of and in the inlet, and

to indicate regions of separated flow. In aparallel study coordinated with this quasi-steadydata evaluation, the dynamic pressure fluctuationsindicated by traces from the flight telemetry andmagnetic tape output of the individual probes werebeing carefully analyzed. Under certain flightconditions, the traces indicated extreme "turbu-lence" at the compressor face. This was known tocause a loss in engine stability in other 9nOnesas reported by Gabriel, Wallner and Lubic01) andwas felt to be a contributory factor in the stallproblems of the F-111A. Although a complete corre-lation between quasi-steady flow distortion anddynamic pressure fluctuations was not undertaken,it was realized that there was a cause and effectrelationship between these two types of distortionand the approach was to address the cause ofunsteady and non-uniform flow in the inlet andattempt to eliminate it. A corresponding reductionof the severity of the dynamic-pressure fluctuations(dynamic distortion) would be expected, but it wasimportant to know the relationships between steadyand dynamic flow to the limits of the instrumenta-tion signal available.

III. F-111 Inlet Pressure Fluctuation Effects

The effects of transient disturbances, or morespecifically, the fluctuating nature of the measuredtotal pressures at the compressor face were con-sidered to have a strong influence in the stallproperties of the engine. This influence andcorresponding effect were considered to be abovethe acceptable steady-state distortion which couldbe accommodated by the engine. The flight regimein which this phenomenon commenced was found to beat low supersonic speeds, with increasing disturb-ance intensity as a function of increasing Machnumber. These disturbances exhibited a wide range

3

of amplitude-frequency content showing both slowand rapid transients. The slaw transients couldpossibly be compensated for by inlet and enginecontrols. For such law frequency disturbances,engine performance is basically similar to steady-state operation since normally, the outlet pres-sures will follow the inlet flow variations inmagnitude and phase such that the overall compress-or pressure ratio will remain the same. However,the majority of actual transient disturbances andtotal pressure fluctuationswere found to be sig-nificantly faster than any of the aforementionedcontrol capabilities. Under these circumstances,specific outlet pressures lag the inlet pressurevariations in both amplitude and phase. Conse-quently, the pressure ratio across the compressorcan differ considerably from the steady-statevalueon the operating line, and conditions can bereached wherein compressor stall margin reductionand even stall will be experienced. Data fromreference (1) has shown this to be the case for asimply induced sinusoidal pressure variation inputto a compressor.

It is important to note here that the originalcompressor face total pressure instrumentationonthe prototype test aircraft was never intended forthe accurate measurement and analysis of transientdisturbances. Hence, an effort to correlate tran-sient disturbances with the lawer frequency averagevalues of the measured total pressures to the com-pressor face required special data reductionmethods. Total pressure readout from flightmagnetic tapes at conditions appropriate to enginestall were first identified and then processedthrough narrow band pass filters by the FieldMeasurements Group of the Air Force Flight DynamicsLaboratory. Figure 3 shows two typical frequencyspectrums obtained from the filtering process ofthe flight test data wherein 85.55 inch longpressure carrying lines were provided between thepressure probes and the transducers. At first

FIGURE3. TYPICALF-111ACOMPRESSORFACEPRESSUREAMPLITUDE-FREQUENCYSPECTRUM

glance, the higher amplitude data would appear tooccur at the lower frequencies; however, the utili-zation of 85.55 inch lines (tubulation) for steadystate pressure measurements suggested the possi-bilities of transient signal attenuation to thetransducer due to classical acoustic type dissipa-tion. In order to correct for this tubulationeffect, an experimental program was undertaken bythe Aero-Acoustics Branch of the Air Force FlightDynamics Laboratory to apply corrections to the

measured pressure variations for conditions justprior to engine stall. Theoretical predictionswere provided by Air Force Aero Propulsion Labora-tory personnel. Figure 4 shows the nature of theamplitude corrections as a function of frequencywhen examined for an average pressure of 14.7 psiand varying temperature. The experimentaldatataken at room temperature showed excellent agree-ment with theory.

10 100 1000

FREQUENCY CYCLES/SECOND

FIGURE4. F-111AOSCILLATORYPRESSURETUBULATIONCHARACTERISTICS

Many supersonic flight conditions associatedwith engine stall were examined with specificemphasis on the high frequency aspects of the tran-sient disturbances. Multiplexing of the instrumen-tation, as is normally accomplished on test air-craft for measuring steady-state parameters, wasfound to have a strong influence on the high fre-quency transient data. In an effort to isolatethese effects, several flights were performedinvolving a minimum of multiplexing with a 15 inchline replacing one of the 85.55 inch lines forreduced tubulation effects. A comparison ofpressure transients in a 85.55 inch line withmultiplexing versus a 15 inch length tubing withminimal multiplexing is shown in Figure 5.

FIGURE5. COMPARISONOF F-111ATUBULATION

Above approximately 250 cps it can be seen that thehigh frequency transients recorded with the 85.55inch line were due to multiplexing, and not presentexcept for some disturbances in the 525-660 cpsrange. It was, therefore, decided to utilize the

COMPRESSORFACEPRESSusPROSES4

3

RMS % OFTRANSDUCER 2FULL SCALE

AIRCRAFTW.16MYLIJ a • 4.8°

1

03 10 100 1000

FREQUENCY CYCLES / SECOND

100

10

OSCIL. PRESS.CORRECTIONFACTOR

0 DPERIWMAL/,

If/1

/

/ °°

IXACTTHEORY,TEMP• CIA EXACT THEORY,TEMP• 2002T

DUCT THEORY,TEMP

5.55" TUBELENGTH

AIRCRAFTF. 16. Mo• L Rita .4.80, 86" TUBE, MULTIPLE5D

AIRCRAFTI. II, Mo. 2.18. 5.5.7°, 15" TUBE. SINGLECHANREL

10 100 1000

FREQUENCY,••• CYC LES/ SECOND

0.2OSCI L.PRESS

PSI RMS0.1

4

0 - 250 cps frequency range for data analysis whentransient data was subject to multiplexing, and0 - 1000 cps for the data with minimal multiplexing.

A specific comparison of steady-state compress-or face recovery pressures with corresponding tran-sient disturbance values obtained from flight isshown in Figure 6. The oscillatory or transientpressure data is based on 0 - 250 cps as discussedabove. Figure 6 shows, generally, that the high,

angle of attack on the dynamic characteristics ofthis probe at a constant Mo = 0.77. Although manydiscrete frequencies were identified from the spec-trum analyzer output, specific frequencies of 130,230, and 525 cps appeared to persist with relative-ly high amplitude for this test condition, whichwas at a military power engine setting. The ampli-tudes of these particular frequencies appeared tobe fairly constant up to moderate angle of attackwith a tendency to converge and further increase in

INBOARD

TOP.war

11/1• ••- 1 ••ON ....MI• • ==.., "JIMM•

.1M• • • /-1• ••• •/".. •W 0, •

• • •• •• • • •• ••• •• ••••'. • •• • •,

1••• •• ••• •,. INT •-W.•.C.III I•I n. IMO/. :. ••

11••••-.7 ..,/• • • ••Ir-f• .4 ..1••

Alb,. • /// n V.l••/SAM• • -•.".• • • •••,--.., • _

.MMK.... • • •21 11/

agegoot..••• .mow

"WM a •• • •• I •

'MOB ,WAWSKRAMWMWM MOM

A,MwWW m

s,MMS•B*A.MMo,W-VM.MWM1

IMMIMI

MMMMP1 ••

IN MEW A N N .0 Oa= ME, IMM1 A0 MEM •=, MMI •••0 IN. = MI I=MI AVnr•0 1• • MI INIM••• I • Eh. A•11• ••••••••••••

Ma

1 IMI•• ••• •1 ••• • .•• NEM 1••

-AKINXIM••• 0 iv Anealfe• =

I• • ••••MV.V.WM •=1, Ztr•i••••••••.•_ ___ _____ m..,••••• ••• WEN W.., • NM INN ,;, ,i1..'„,!..:,....ii.......... W."n V 1.., UNE 1•M PM. d•-•.: i • era004.N-,fial••••••••• C , ARM MMEMJ•W •V.• IIMa =11••• COM OS Ai d1.100... -.MEW 1. Y •••••"••

.

=MK •.1.1.11•11a1 •151...le•a. +NW. •db • MI• at=1.•••••• AKKRK.•••••••••11. ..,.....000C4, • MW •• MMIEW

II •••.-^••••=•••••••rn.*•••.•••.• 0 0 • I7M.I.

• •••• ••• • •• .,.... --.•• ••••••,•171•••• . • • • • /.

I. NEM IMMLIM••• ••,sain......Y.,OL,••‘

1.1•• •• ••••• • 1 M. Ai ••••••*•••• n by MN •

•1•1 11••••••••• •• ••• •• • •• •••••••• ••. ••• • ••• ••••• • • •••• • . ••••••• • ..•••••• •• •••._ • •• •••••• I.•••• • .•1_.••••-,....... _-••••••••••

STEADY- STATECOMPRESSORFACE RMS OSCILWORY PRESSURES--RECOVERYPRESSURES % OF AVG. TOTALPRESSURE

AIRCRAFT16; Mo= 1.98' 0(-4.8°, 0.2 SECONDSBEFOREENGINESTALL

FIGURE6. F-111A COMPRESSORFACE STEADY-STATE AND OSCILLATORY PRESSURES

steady-state recovery pressures corresponded toareas of reduced transient or oscillatory pressures,whereas low recovery pressures related to regionsof higher transient values. The region in thelower left-hand corner of the oscillatory pressuremap was of particular interest. This area of high-est transient disturbance values correspondeddirectly to the lower left-hand portion of theinlet which was most susceptible to boundary layeringestion. In addition, the steady-state analysisfrom flight demonstrated the upward spreading oflow total pressures from the bottom of the sharpcowl lip with increasing angle of attack. The dataof Figure 6 would indicate that, in addition tobeing of a very low recovery nature, this portionof the flow possessed a high degree of flow unstead-iness sufficient to cause engine stall.

The transient disturbance analysis for all 40compressor face pressure measurements for a particu-lar stall condition would have required a prohibi-ted expenditure of manhours and it was, therefore,decided to use an available single pitot tube of 15inch tubulation to examine the fluctuating natureof the duct flow. Figure 7 shows the effect of

amplitude at higher angles of attack until enginecompressor stall was experienced. Also shown onFigure 7 are the effects of first zone afterburneroperation for cruise angle of attack. The

1.4

1.2

1.0

RMS %OF AVG. 0.8RECOVERY

0.6

0.4

0.2

00 25 50 75 100

ANGLE OF ATTACK •••% OF STALL ANGLE

FIGURE7. F-111A COMPRESSORFACE PRESSUREFLUCTUATIONSVS. ANGLE OF ATTACK

C)PP033

FIRST ZONE

AFTERBURNER

M0.037

STILL

0 230 CPSMILITARY RATED POWER

525 CPS130 CPS

•

5

amplitude associated with 130 cps waa found tochange a small amount, however, there was a sub-stantial amplitude increase in the 230 and 525 cpsfrequencies.

The condition of aircraft acceleration forcruise angle of attack at maximum afterburner powerwas examined with results as presented in Figure 8.

20 CM 0•

0 IslEt 1 1 1 llllllllll

0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4

MACH NUMBER

FIGURE8. F-111ACOMPRESSORFACEPRESSUREFLUCTUATIONSVS.MACHNUMBER

Here again, the influence of afterburner operationis shown in the amplitudes of the 240 and 525 cpsfrequencies for transonic flight conditions. How-ever, amplitudes at these frequencies decreasedwith increasing Mach number and correspondingdecreases in corrected air flow up to approximatelyMach number 2. Beyond Mach number 2, there was adramatic increase in all three amplitudes up toMach number 2.2 where engine stall was experienced.

From the quasi-steady and transient disturbanceclatastudied during 1967, it was clear that theengine compressor stall characteristicswerestrongly influenced by inlet pressure pulsations athigh frequencies and this effect must be consideredin conjunction with the "steady-state"distortion.

IV. Detailed Studies of F-111A Inlet and Engine

Air Flow Fluctuation Effects

Subsequent to the initial study of the F-111Ainlet engine incompatibilityeffort discussed 42_6,Section III, a limited number of investigations' 'have been carried out in order to shed light onthis important problem area. A typical example ofone of the more significant and recent programs wasreported by Plourde.and Brimelow(7). Recently,Burcham and Hughes(8) have modified and utilizedthe Pratt and Whitney KDA distortion factor forpredicting surge. The engine compressor face wassub-divided into 5 equal areas through concentriccircles or rings. Probes were placed on ringswhich were maintained at a constant radii from thecompressor centerline. The modified KDA distortionparameter was defined as follows:

-PtPtmax ci

( 2)1=1 av

Kimc

x100

i=1

whereC = ratio of compressor inlet radius to ring

radius

= number of ring

Er= largest continuous arc of the ring overwhich the total pressure is below the ringaverage pressure

In this specific effort a flight test F-111A air-craft was utilized to determine the dynamic natureof inlet pressure fluctuations related to engineoperational stability. Derived steady-state flowdistortion patterns as developed from low responsepressure instrumentationwere compared with boththe KDA and KDM distortion parameters calculatedfrom high response Instrw2cntation. A typical com-parison is shown in Figure 9 for the flight case ofMach number 1.6 at an altitude of 45,000 feet withoff-design inlet spike position. Here it is clear-ly seen that the low response data technique func-tioning at a sampling rate of 50 cuts per seconddid not yield information indicative of the com-pressor stall. On the other hand, utilizing thehigher response data technique and calculatingeither the KDA or Kim distortion parameter at 400

FIGURE9. COMPARISONOF HIGHAND LOW RESPONSEKDAWITHHIGH RESPONSEKm. F-111AFLIGHTCONDITIONS;MACW NUMBER= 1.6,ALTITUDE= 45,000FEETAND OFF-DESIGNSPIKEPOSITION

1400

1200

KDA 1000(high 800

response,400sps) 600

400

200

1200—

Kw 100°—

ighresponse,400sps)

4

200'

Surgeinitiation—\ SurgeI /

—

KDA 10011—(low

response,50sps) 600

1 1

.2 .3

Time, sec

1 /I /

I I I.5 .6

1.0

0.8

RMS % 0.6OF AVG.RECOVERY

0.4

0.2

FULLAFTERBUFNERcc• 5.8

STALL

Pt= ring maximum total pressuremax

Pt= ring average total pressureav

P = ring minimum total pressuretmin

6

samples per second did yield a substantial peakapproximately 15 milliseconds prior to surge.Figure 10 shows a time history of the probe dataand distortion factor for Mach number 2.17 and analtitude of 44,000 feet. Probes A and B showincreases in pressure as the stall condition isapproached whereas probes AI and B1 are decreasingand hence result in a maximum distortion value.It is interesting to note that the instantaneouspressure recovery map shows a larger high pressurearea along with a more intense low pressure area.

p1av- 0.83 P1av• "3

Figure 11 shows the surge characteristics fortransonic flight at Mach number 0.9 and 30,000 feetaltitude. This particular stall occurred as aresult of the off-design conditions of the inletcone and is generally recognized as a "drift" typeof surge. This is demonstrated by the fact thatpeak values of the distortion factor occurred sever-al times during the time period examined.

The modified dist9Won parameter as developedby Burcham and Hughes") was found to be approxi-mately 80 percent effective in identifying surgewhen dynamic conditions prevailed within approxi-mately 90 percent of the maximum steady state dis-tortion Riau!. Needless to say, additional infor-mation 0'IU and more exacting methods must bedeveloped to predict engine instability due todynamic inlet conditions.

ptav- 0.87 ptav• 0.87

PressurerecoveryInstantaneousatsurgeinitiation

— Surgeinitiatio_ _ _ _ _ _ity

.1

FIGURE11. COMPARISONOF AVERAGERECOVERYANDINSTANTANEOUSRECOVERYMAPS WITHKDA. F-1:1:FLIGHTCONDITIONS;MACH NUMBER= 0.9,ALTITUDE=30,000FEETAND OFF-DESIGNSPIKEPOSITION

V. Advanced Inlet Configuration Studies

More recently the Air Force Flight DynamicsLaboratory has undertaken a number of programs 'toinvestigate flows about fuselage and wing-fuselagecombinations throughout the subsonic, transonic andsupersonic speed regimes. The objectives of theseprograms are to develop a clear understanding ofinlet-airframe interactions and, more importantly,to attain an experimental data bank and corres-ponding analytical approach for assessing the dynam-ic phenomena associated with engines and inlets.The Laboratory has initiated prcject Tailor-Mate inorder to examine the effects of configuration vari-ations on flow field dynamics and related effectsto the engine system. Figures 12 and 13 show atypical 1/3 scale wind tunnel model along withvarious aircraft configurations studied. Configur-ations A-1 and A-2 are examples of side mountedtype inlets whereas A-3 is an example of a fuselageshielded inlet, and wing shielded inlets are shownby configurations B-3 and B-4. One quarter scalefuselage models were constructed for wind tunneltesting purposes with appropriate fuselage staticpressure distributions, boundary layer measurementsand more importantly, the dynamic nature of theflow fields at the proposed inlet stations. Inaddition to the flow field measurements made in thearea of the entrance to the inlet, two side mountedand two shielded external compression inlets weretailored for the flow fields defined by the fore-body as shown in Figure 14. The detailed instru-mentation for such a duct system is shown in Figure15. Instrumentationwas utilized to document theinlet performance and included static pressurerakes near the cowl lip, in the diffuser and at the

20.5

Pressure, 19.5psia 18.5

17.520

Pressure, 19psia 18

17

1400

1200

KDA 1000

800

Surgeinifiafion SurgeA

.2 •Timt, sec

FIGURE10. COMPARISONOF AVERAGERECOVERYANDINSTANTANEOUSRECOVERYMAPS WITHKDA. F-111AFLIGHTCONDITIONS:MACH NUMBER= 2.17,ALTITUDE=44,000FEET

Pressure recoverymap Pressure recoverymapTimeaveraged Instantaneousat surge initiation

PressurerecoveryTimeaveraged

1400

1200

1000

KDA goo

600

40°0.2 .3Time, sec

.4

7

FIGURE12. TYPICAL1/3SCALETAILOR-MATEWINDTUNNELMODEL

FIGURE13. REPRESENTATIVECONFIGURATIONSFOR FOREBODYFLOWFIELDTESTS

simulated compressor face. It is important to notethat the compressor Face instrumentationcontainedhigh response type transducers to identify thefluctuation nature of the inlet flow. Figure 16shows the results of wind tunnel tests for the fourconfigurationsmentioned. These tests were per-formed at Mach number 2.2 with varying angle ofattack. Figure 16 shows both wing shielded inletsystems experienced lower distortion as indicatedby the simple distortion index along with low"turbulence" as a function of angle of attack. Asmight be expected the side mounted type of inlets

experienced higher distortion with correspondinglyhigher indices of "turbulence."

The importance of the flow field generated bythe forebody of theffuselage has been pointed outby Surber and Stava°-1) and Zonars(12). An exampleof such sensitivity is shown in Figure 17 whereinthe side mounted 2-dimensional inlet on body A-1was examined in conjunction with body A-2. Thisfigure shows the vastly different characteristicof distortion vs "turbulence." Surprisinglyenough, the small change in contour of the A-2

8

STATION 2 STATION 2

Iii

POROUS-PLATEBLEEDFIGURE14. NOMENCLATUREAND COMPARISONOF FOUR (4)INLETDESIGNS

STATICPRESSURELOCATIONSSTEADYSTATE

HIGHRESPONSEI(40 PROBES*,•

Itow fAirewSTATION2

A-1B-4

o A-2a B-3

DISTORTION

FIGURE15. TYPICALINLETINSTRUMENTATION

APTRMSPT

ay

0-10 0 10 2U 30

ANGLE OF ATTAC;( •-•-•DEGREES

TURBULENCE

.04

.02

0

-10 0 10 20 .30

ANGLE OF ATTACK DEGREES

(P

- P

MAX TMIN)

PTay

FIGURE16. COMPARISONOF STEADY- STATEAND DYNAMICDISTORTIONFOR VARIOUSCONFIGURATIONSAT MACH NUMBER2.2

9

1.a

. 60-FLOWDISTORTIONPARAMETER-Dx. 4

. 2

00 01 02 03 04 05 06

TURWLENCE,-.APT ///PTav

FIGURE 17. INFLUENCE OF FOREBODYCONTOURONENGINE STABILITY AT MACHNUMBER2.2

fuselage was found to have a substantially bettercharacteristic than A-1. Thia is undoubtedly dueto a lower local outwash and hence a reduced tend-ency toward flow separation on the inboard side ofthe inlet. In the event the designer is confinedto the A-1 inlet configurationand cannot readjustthe body contour as shown by the A-2 characteris-tics, he must then look for other means by whichhe can suppress both the steady state and "turbu-lent" distortion. A longer inlet duct has a sur-prisingly favorable characteristicas denoted inFigure 18. There is a considerable reduction inboth distortion parameters which puts the opera-tional mode of the inlet well within the stablebounds of engine operation.

1.0

.8

FLOW .6DISTORTIONPARAMETER-DX .4

.2

00 .01 .02 .03 .04 .05 .06

TURBULENCE;..APT/

TRMS av

FIGURE 18. EFFECT OF LONGERDUCT ON SUPPRESSINGSTEADY STATE AND "TURBULENT" DISTORTION

VI. Dynamic Data Handling Technique

Among the problems that exist in handlingdynamic data are the tremendous quantities of ana-log data tapes generated during inlet study anddevelopment programs and that past efforts toanalyze dynamic data has depended heavily upon whathas been seen relative to the behavior of the steadystate or average component of compressor face totalpressure. As a consequence, only about one percentof the data is actually examined since considerable

digitization is required to review one case. Morespecifically, the area of interest centers on theanalysis of the dynamic or fluctuating componentof total pressure measured at the compressor faceplane of a trisonic type inlet. The pressures aremeasured by means of fast response instrumentationlocated in rakes radiating from the hub. Thisdata is recorded on analog tape and represents thebeginning of our problem.

The solution to this problem has been todevelop an analog editing system for screening andediting inlet dynamic data based on the use ofengine distortion parameters. As a result, largequantities of tape can be screened and those partsof the data identified which would have adverseeffects on airframe-propulsioncompatibility.

A typical Air Force Flight Dynamics Laboratoryinlet program consists of 5000 data points whereinone Mach number, angle of attack and yaw, andcapture area ratio comprises a single data point.At least 200 feet of tape are used for each datapoint and as a result, it can be seen that about100reels of tape are required for a program of thismagnitude. For a more extensive inlet developmentprogram, such as associated with advanced flightvehicles, as many as 500 tapes are required. Inany event, the data of interest is contained ononly about one percent of the tape which is notnecessarily the same one percent of tape mentionedpreviously. The principle question that arises ishow does one expeditiouslyand economically locatethe data of interest?

In the development of the analog editing system,certain goals were established. First, it wasdesirable to utilize parameters involving all thecompressor face steady-state and dynamic data whichhad a direct relationship to engine stability.Second, a scheme was desired that would identifyhigh levels of dynamic flow activity on the tapesand where this event occurred. Third, a fastresponse capability was a requirement in order toaccount for model scale. For example, if a particu-lar engine is sensitive to pressure fluctuations upto 200 cycles/second,and the inlet wind tunnelmodel is one-tenth scale, then valid data out to2000 cycles/second is required from the model.Model scaling characteristic9have been hypothe-sized by Sherman and Motycka"13). Fourth, adesirous capability was to use more than one para-meter in the screening process to determine whichwas most meaningful and acceptable and hence avoidtape re-runs. Fifth, the system should be flexibleto permit digitization of data and possess a dataplayback capability at the recorded speed.

Among the parameters selected for data screen-

ing was the Pratt & Whitney engine distortion param-etersthe basisshowndescribeparticular

KA =

KRA13

KA, KRADandor—experimental

below, whichthe levelcompressor

Ko + bK.RAD

1

KA which

in partof distortion

face

have been formulateddata. The expressionsrelates to reference(12),

associatedpattern.

-x

Di

(Ptav -Ptav,j

on

with a

-xQav

i=1i=1

A-1 INLETWITHA-2 FOREBODY

ALWAYS.--\I STABLE1

10

whereb = constant depending on engine design and

entrance Mach number

x = weighting factor depending on-distortionsensitivity

K describes the influences associated with a cir-cumferential distortion pattern while KRAD describesthe pattern variation associated with r-iglaldis-tortion. When a combined pattern exists, which istypically the case, Ke and KRAD are added togetherin a weighted manner to form KA. In addition tothe Pratt and Whitney parameters, a set of GeneralElectric engine distortion parameters have beenprogrammed. These expressions are used to identifyhigh levels of dynamic activity in the air flowprocess. These data are subsequently subjected tofurther analysis which in turn aids in determiningthe necessary modifications required to alleviatethe compatibility problem.

The dynamic data screening device or system wasdeveloped jointly between the Air Force FlightDynamics Laboratory and the Aeronautical SystemsDivision Computer Center using a hybrid computer.This program was initiated in January 1970 bySedlock and Marous (14) with the acquisition of a72 channel multiplex discriminator system, a 14track direct playback tape transport, tape searchunit peak detectors, and a 48 channel data filter-ing system. The complete system shown in Figure 19became operational in July 1971. The systemdescribed above is,qpilar to OA developed byCrites and Heckart-"), Critesivi, Lynch andSlade(17) except for the added flexibility due to ahybrid computer capability.

The current status of the Air Force FlightDynamics Laboratory system is that both GeneralElectric and Pratt and Whitney engine distortionparameters have been programmed on the computer andup to five parameters can be tracked simultaneouslywith an order of priority established for eachparameter. The primary requirement of the systemis to identify dynamic peaks and the time of

occurrence. The resolution of the tape search unitpermits identificationof the peak value within onemillisecond. Center frequencies used in the dis-criminators have been selected for greatest compat-ibility with those being used by USAF and contract-or facilities. The dynamic data can be filteredfrom 125 to 9000 cycles/second in six discreteincrements in order to account for model scale andfiltering of any unwanted high frequency informa-tion such as probe resonance. Both the enginedistortion parameters and pressure data can bedigitized at various sampling rates.

Our past and current efforts have includedreview of the compatibility points in the B-1 Inter-face Control Document and the Arnold EngineeringDevelopment Center 1/10 scale inlet test data. Inaddition to continued support of the B-1 program,data from the RA-5C wind tunnel-flight test corre-lation and Tailor-Mate programs will be reviewed.

McDonnell-Douglas personnel(16) have developeda screening system for use during the F-15 inletdevelopment program. In examining this capabilityto review dynamic data based on conventional means,it was estimated that six manyears and one milliondollars were required to review one percent of a250 data point program which represented some fourmillion pressure distributions. The developmentof an analog editing system reduced this task tosix weeks with less than 1000 feet of tape to beexamined. Our own experience has shown that onereel of tape containing 30 data points can beexamined in approximately one hour. To accomplishthe same task with a digital computer would require15 hours to digitize the data and approximately20 hours of computer time to process the informa-tion. This estimate is based on 200 samples perdata point. The development described above repre-sents a major step in handling the extremely largeamount of data associated with an inlet developmentprogram.

A description of how the engine distortion areimplemented on the analog/digital computer toaccomplish the goal of editing and screening the

HYBRID COMPUTER

ANALOG DIGITAL

'

LINEPRINTER

MAGTAPE

FM MULTIPLEXDISCRIMINATOR

SYSTEM

TAPEDECK

FILTERS

PEAKDETECTORS

TAPESEARCHUNIT

FIGURE 19. DYNAMIC DATA SCREENING AND EDITING SYSTEM

11

data will now be addressed. This can be accomplishedby relating the expression for Ke to a particularconfiguration of total pressure probes at the com-pressor face. In this particular case, consider aconfiguration that consists of 48 probes with sixrings and eight rakes. The implementationof Equa-tion (3) on the analog computer is quite simple.The steady-state and gain adjusted fluctuating pres-sure are summed together to form the total componentof pressure. Each pressure is multiplied by its re-spective sine® and cosine®, summed around eachring, and then squared. These two terms are addedtogether and then the square root is taken of thissummation. Finally, this value is multiplied bythe value of the leading term to attain Ke for onering. This process is repeated for each ring andthen the individual ring Ke are summed together toform the total Ke. While the value of this expres-sion is being calculated, a similar process is oc-curring simultaneously for the other parameters.

The editing process is accomplished by consid-ering the time history of the parameter Ke, Ke canbe generated as a continuous function since an ana-log computer is a continuous type of machine. Theoperator has the ability to set a threshold levelfor each of the parameters such that only informa-tion occurring about that level will be examined.The engineer must know when a peak in Ke has beenexperienced and the time of this occurrence. In ad-dition, he is interested in the value of Ke when itexceeds a given threshold level and when it returnsto a lower value. Special peak detector networksare utilized to accomplish these objectives. Thesepeak detectors track an increasing signal to thepeak level and maintain that level until it is re-set. The peak detectors are normally reset whenthe value of the parameter drops below the thresholdlevel in order that successive peaks can be detectedeven though such peaks may be of a lower value thana preceding peak. In addition, the peak detectorscan be used as a signal generator that signifies apeak has been detected. Judgement must be made asto identifying both threshold crossings and peaksor peak values alone. When a threshold crossingoccurs, an interrupt signal is generated, and theinformation is transferred from the analog to thedigital computer. No on-line manipulation of thisinformation is permitted in order to transfer thedata as quickly as possible. The current responsetime from signal interrupt thru information trans-fer is 300 microseconds. An important feature ofthe program is the identificationof which para-meters triggered the interrupt signal. Wheneverthe interrupt signal occurs, the peak value of theparameter is stored as well as the value of theother parameters at this particular instant. Thenext output from the editor is the time when thepeak was detected. The time resoltuion is to with-in one millisecond.

Many electronic components make up the editingsystem. A 14 track tape transport is used to play-back the dynamic data through the discriminatorsystem which de-multiplexes the individual signals.Each pressure signal is filtered before it is sentto the analog computer. Coupled vith the tape deckand hybrid computer is the tape search unit. Thesearch unit allows one to find a particular time-pressure history on the tape while it also servesas the time reference frame for the hybrid computerThe peak detectors, mentioned earlier, are coupledto the analog computer. The information stored inthe digital computer can be printed out on the

line printer or stored on magnetic tape.

VII. Advanced Nozzle Configuration Studies

The Air Force Flight Dynamics Laboratory hasundertaken a number of programs(16) to investigatethe effects of airframe-nozzle interactions through-out the subsonic, transonic and supersonic speed re-gimes. The objective of these programs is to de-velop a clear understanding of nozzle-airframe in-teractions and hence improve the technology base.A basic aircraft configuration incorporatinga com-mon forebody, wing, inlet system, and a twin engineinstallation shown in Figure 20, was utilized duringhigh Reynolds number (0.5 - 6 x 106 per foot) windtunnel tests to determine the relative merits of a

FIGURE20. BASICTWIN-ENGINEWINDTUNNELMODEL;ALL DIMENSIONSIN INCHES

wide spectrum of aft-body nozzle geometrical varia-tions. This figure shows a schematic of the modeland the support system required to conduct tests inthe 16 foot AEDC Propulsion Transonic and SupersonicWind Tunnels. Cross-sectionalarea distributions,including the influence of the support strut sys-tem and the effect of nozzle spacing, is shown inFigure 21.

FIGURE21. WINDTUNNELMODELAND SUPPORTSYSTEMCROSS-SECTIONALAREA CHARACTERISTICSThe same type of area distributions, involving the

113 189 133.192

38° 31._

-1.341

17.940

27 4

31 AV

77.6

aff// //

ILLNNNELROMlo°

NOZZLESPACING RATIO

WIDEI NTERNEDIAifNARROW

430

CROSS SEVIONALAREA - I6/`

300

161THOUTSTRUT

\CONVERGENT -DIVERGENT

100

20 M M 80 100 120

MODELFUSELAGESTATION( F.S. I - IN.

12

FIGURE22. WINDTUNNELMODELCROSS-SECTIONALAREACHARACTERISTICS

which were connected or disconnected at fuselagestation 133.182 inches for non-afterburningcondi-tions are shown in Figure 23. Maximum afterburning

nozzle contours are shown in Figure 24 and weresimilarly attached to the same fuselage location.The large scale wind tunnel model shown in Figure

COWERGENT FLAP CONVERGENT-DIVERGENT

ALTERNATE

CONVERGENT-DIVERGENT

COW ERGENT-IRI SUNSHROUDEDPLUG

FIGURE24. AFTERBURNINGNOZZLECONFIGURATIONSTESTED

25 was configured in such a manner as to employ asystem of balances which measured both the aft-bodyboattail drag and the nozzle boattail drag. In ad-dition, a thrust balance was incorporated to measurethe internal thrust and nozzle boattail drag. It isimportant to note that both the horizontal and ver-tical tail were not included in the drag measure-ments aforementioned.

The nozzle boattail drag characteristics forthe non-afterburning condition is shown in Figure26 for Mach number 0.9. The drag coefficients dis-played in this figure are referenced to the cross-sectional area located at fuselage station 133.182inches. The convergent-divergentand convergentiris nozzles, which are basically slender in nature,

EMPENNAGEON

300CROSSSECTIONALAREA -IN2 203

effect of wing, tail and the limits of nozzle set-ting, are shown in Figure 22. The various nozzles,

WITH WING

AIR NOZZLE

DRYPOWER

40 60 80 100 120

MODELFUSELAGESTATIONI F.S. I - i N.

140 160

COWERGENTFLAP

ALTERNATECONVERGENT- DIVERGENT

COWERGENT- DIVERGENT

CONVERGENTIRIS UNSHROUDEDPLUG

FIGURE23. SYSTEMOF NON-AFTERBURNINGNOZZLESCONFIGURATIONSTESTED

FIGURE25. PHOTOGRAPHOF MODELINSTALLEDIN THE 16 FOOT AEDC TRANSONICPROPULSIONWINDTUNNEL

13

SYMBOL

4

0

0

0

NOZZLE

COWERGENTFLAP

COWERGENT-DIVERGENT

ALTERNATECOWERGENT-DIVERGENT

COWERGENT IRIS

UNSHROUDEDPLUG

0 2 4 6 8 10 12 14NOZZLEPRESSURERATIO - Prn/ PeD

NOZZLES

COWERGENTFLAP

COWERGENT-DIVERGENT

ALTERNATECOWERGENT-DIVERGENT

COWERGENTIRI S

UNSHROUDEDPLUG

O.04

NOZZLEBOATTAILDRAGCOEFFICI ENT-C 0

(IBM

04

408

O.12

O.08

0.04

NOZZLE BOATTAI LDRAG COEFFICIENT-CD 0

BM

-0.04

-O.08

8=5,

2 3 4 5 6

NOZZLE PRESSURERATIO - Pn PI co

O.12

0.08

FIGURE26. NON—AFTERBURNINGNOZZLEBOATTAILDRAGCHARACTERISTICS;MACH NUMBER= 0.9

displayed the lowest drag characteristics.In fact,these two nozzles, along with the convergent flapnozzle, experienced pressurizationof the external

aft-facing surfaces and, hence, were subjected to athrust rather than a drag force. The alternateconvergent-divergentnozzle experienced the highestdrag due to flow separation caused by the rearwardfacing step between the nozzle and fuselagefairing. The unshrouded plug nozzle also ex-perienced separated flow, which basically preventedpressurization of the external surfaces with cor-respondingly lower relative drag.

The nozzle boattail drag based on maximum after-burner nozzle position for Mach number 0.9 isshown in Figure 27. The convergent-flapand con-vergent-iris nozzles show good pressure recovery

SYMBOL NOZZLES

COWERGENTFLAP

COWERGENT-DIVERGENT

ALTERNATECOWERGENT- DIVERGENT

0 COWERGENTIRIS

a UNSHROUDEDPLUG

FIGURE27. AFTERBURNINGNOZZLEBOATTAILDRAGCHARACTERISTICS;MACH NUMBER= 0.9

characteristics, particularly with increased nozzlepressure ratio. The remaining three nozzles yielded

higher drag characteristics. The same nozzle boat-tail drag characteristics associated with the maxi-

mum afterburner position for Mach number 1.2 isshown in Figure 28. Here it can clearly be seen that

the convergent-flap and convergent-irisnozzles in-volve a substantial drag penalty due to the largeprojected frontal area;however, the drag coefficient

of these two nozzles diminishes with increased pres-sure ratio as might be expected with a Mach number

FIGURE28. AFTER—BURNINGNOZZLEBOATTAILDRAGCHARACTERISTICS;MACHNUMBER= 1.2

1.2 flight requirement.The convergent-divergentandulstsrouded-plugnozzles have lower drag charac-teristics than the two aforementionednozzles pri-marily due to the reduced aft-facing surface areas.The alternate convergent-divergeatnozzle displayedthe best characteristicswith very little influenceof nozzle exhaust pluming effects.

The aft-body boattail drag coefficient, basedupon the cross-sectionalarea at fuselage station113.188 associated with the non-afterburningnoz-zle and afterburning nozzle positions for Machnumber 0.9, is shown in Figures 29 and 30. The

0 1 2 3 4 5 6NOZZLEPRESSURERATIO - P.i.n/Pco

FIGURE29. AFT—BODYBOATTAILDRAG CHARACTERISTICS;NON—AFTERBURNINGNOZZLES,MACH NUMBER= 0.9

O.12AFTERBODY BOATTAI L

DRAG COEFFICIENT-CD 0.08

BM

O.04

0 1 2 3 4 5 6

NOZZLE PRESSURERATIO - P PTn/

FIGURE30. AFT—BODYBOATTAILDRAG CHARACTERISTICS;AFTERBURNINGNOZZLES,MACHNUMBER= 0.9

0.04NOZZLEBOATTAILDRAGCOEFFICIENT-CD 0

BTN

2 3 4 5NOZZLEPRESSURERATIO - PIn

0.1

AFTERBODY BOATTAI LDRAG COEFFIC I ENT-

0.CD

BTA

o.

SYMBOL NOZZLE

COWERGENTFLAP

CONVERGENT-DIVERGENT

a ALTERNATECONVERGENT-DIVERGENT

CONVERGENTIR IS

0 UNSHROUDEDPLUG

0.12

0.08

SYMBOL NOZZLE

CONVERGENTFLAP

ALTERNATECOWERGENT-DIVERGENT

0 CONVERGENTIRISa UNSHROUDEDPLUG

14

same aerodynamic coefficient for Mach number 1.6with maximum afterburner nozzle conditions is shownin Figure 31. In all three cases, the afterbody

0.16SYMBOL NOZZLE

CONVERGENT-DIVERGENTCONVERGENT-IRIS

UNSHROUDEDPLUG

0 0 2 4 6 8 10 12

NOZZLEPRESSURERATIO - P rTr, co

FIGURE31. AFT-BODYBOATTAILDRAGCHARACTERISTICS;AFTERBURNINGNOZZLES,MACHNUMBER= 1.6

boattail drag was of a positive nature with smalldifferences caused by downstream changes in thenozzle boattail geometry.

The total drag coefficient, i.e., the summa-tion of aft-body boattail drag and nozzle boattaildrag, is shown in Figure 32. Here we see that the

The nozzle test characteristics and resultspresented above represent a complex interactionbetween external and internal flows. The issue isfurther clouded by viscous considerations. Theaccurate simulation of all important parameters isvery difficult, even to the point where some param-eters are not fully defined or identified. A re-search and development effort is required to definethe proper parameters and their associated impor-tance.

NOZZLE PRESSURERATIO

FIGURE33. AFT-BODYDRAG CHARACTERISTICSAS AFUNCTIONOF REYNOLDSNUMBER;MACH NUMBER= 0.9

0.015

0 COI

MI BODY 0.53DRAG

COEFFICIENT0.

0.001

0

451

-0 001

UNIT REYNOLDSNUMBER

106

1.5 106

L4. 146

; 2';

DRAG METRIC

0.12

0.08AFTERBODYBOATTAILDRAGCOEFFICIENT-CD 0.04

BM

REFERENCES

0.6 0.8 1.0 1.2 1.4 L6 1.8 2.0

FREESTREAMMACH NUMBER - Mco

FIGURE32. TOTALAFT-ENDDRAG CHARACTERISTICS

total aft-end drag variations based upon aircraftwing area is principally due to aft-body boattaildrag characteristics previously described. Also,one can observe that the total aft-end drag is ofa positive nature mainly due to the aft-body boat-tail section. Variations of this drag are associ-ated with the nozzle boattail drag characteristics.

Reynolds number effects on the total aft-bodydrag are shown in Figure 33. These results, whichshow increasing aft-body drag as a function of in-creasing unit Reynolds number, was unexpected sinceskin friction effects and separated flow conditionsnormally improve such circumstances with lowerresulting drag. Although not presented in thispaper, the Reynolds number effects on the nozzleboattail drag were of an opposite nature.

Gabriel, D.S., Wallner, L.E., and Lubick, R.J.,"Some Effects of Transients in Inlet Pressures andTemperatures on a Turbojet Engine," Institute ofthe Aeronautical Sciences Twenty-Fifth Annual Meet-ing, (January 1957).

Martin, A.W. and Kostin, L.C., "PropulsionSystem Dynamic Test Results," North American Avia-tion, Inc., Report No. NA-67-386, (April 1967).

Winslow, L.J., et. al., "Inlet Distortion In-vestigation, Upstream Engine Influence and ScreenSimulation," The Boeing Company, Technical ReportAFAPL-TR-68-140, (January 1969).

Wasserbauer, J.F. and Willoh, R.G., "Experimen-tal and Analytical Investigation of the DynamicResponse of a Supersonic Mixed Compression Inlet,"AIAA Preprint 68-651 (1968).

Martin, A. and Kostin, L., "Dynamic Distortionat the Exit of a Subsonic Diffuser of a Mixed Com-pression Inlet," North American Rockwell, ReportTFD-69-588 (1969).

Oates, G.C., Sherman, D.A., Motycka, D.L.,"Experimental Study of Inlet-Generated PressureFluctuations," PWA Report 3682 (1969).

Plourde, G.A. and Brimelow, B., "PressureFluctuations Cause Compressor Instability," AirForce Airframe/PropulsionCompatibility Symposium,(June 1969).

0.010

0.038

0.006

0.034

0.002

SYMBOL NOZZLE

CI CONVERGENTFLAP

0 0 COFNERGENT-DIVERGENT

ALTERNATECONVERGENT-DIVERGENT

CONVERGENTIRIS

Oa UNSHROUDEDPLUGTOTALDRAGCOEFFICIENT-

CDTw

15

Burcham, F.W., Jr. and Hughes, D.L., "Analysisof In-Flight Pressure Fluctuations Leading toEngine Compressor Surge in an F-111A Airplane forMach Number to 2.17," AIAA Preprint 70-264 (1970).

Calogeras, J.E., Burstadt, P.L. and Coltrin, R.E., "Instantaneous and Dynamic Analysis of Super-sonic Inlet-Engine Compatibility,"AIAA Preprint71-667 (1971).

Jansen, W., "Compressor Sensitivity to Tran-sient and Distorted Transient Flows," AIAA Preprint71-670 (1971).

Surber, L.E. and Stava, D.J., "SupersonicInlet Performance and Distortion During ManeuveringFlight," AGARD Propulsion and Energetics PanelSpecialists' Meeting on "Inlets and Nozzles forAerospace Engines," AGARD Conference ProceedingsCP-91-71, (1971).

Zonars, D., "Problems on Inlets and Nozzles,"7th Congress of the InternationalCouncil of theAeronautical Sciences, (August 1970).

Sherman, D.A., and Motycka, D.L.,"ExperimentalEvaluation of a Hypothesis for Scaling Inlet Tur-bulence Data," AIAA Preprint 71-669 (1971).

Sedlock, D. and Marous, J., private communi-cation, (1972).

Crites, R.C. and Heckart, M.V., "Application ofRandom Data Techniques to Aircraft Inlet Diagnos-tics," AIAA Preprint 70-597 (1970).

Crites, R.C., "The Philosophy of AnalogTechniques to the Analysis and High Speed Screeningof Dynamic Data, "AIAA Preprint 70-595 (1970).

Lynch, F.R. and Slade, C.J., "Data Acquisitionand Automated Editing Techniques for Engine-InletTests," AIAA Preprint 70-596 (1970).

Air Force Flight Dynamics Laboratory,"Experimental and Analytical Determination ofIntegrated Airframe-Nozzle Performance," InterimReport (1972).