JOURNAL OF PROPULSION AND POWER

Vol 17 No 2 MarchndashApril 2001

Experimental Testing of a Hypersonic Inlet withRectangular-to-Elliptical Shape Transition

M K Smartcurren

NASA Langley Research Center Hampton Virginia 23681

Wind-tunnel testing of a hypersonic inlet with rectangular-to-elliptical shape transition has been conducted atMach 62 These tests were performed to validate the use of a recently developed design methodology for xed-geometry hypersonic inlets suitable for airframe integrated scramjets Results indicated that ow features withinthe inlet were similar to design and that the inlet typically captured 96 of the available air ow Typical mass- ow-weighted total pressure recoveries of 55 were obtained for compression ratios of 148 throughout the testprogram Assessment of the inlet starting characteristics indicated that the inlet self-started at Mach 62 despite thefact that it had an internal contraction ratio well above the Kantrowitz limit (Kantrowitz A and Donaldson CldquoPreliminary Investigation of Supersonic Diffusersrdquo NACA WR L-713 1945) These results demonstrate thathigh-performance xed-geometry inlets can be designed to combine a nearly rectangular capture with a smoothtransition to an elliptical throat

NomenclatureA C E = pressure instrumentation streamlinesG I KAFM = ow meter throat areaCD = discharge coef cient for the mass- ow meterCpNi = speci c heat of nickelH = enthalpy zero-base referenceM = Mach numbermc = mass capture percentagePm = mass ow ratePav = average pressure on instrumentation streamlinesPR = inlet compression ratio pe=p1

PT = inlet total pressure recovery pte=pt1

p = pressureq = dynamic pressurePq = heat transfer rateR = gas constant for air 287035 Jkg KRe = Reynolds numberT = temperatureTR = inlet temperature ratio Te=T1

t = timeu = x component of velocityv = volumex = axial length along modely = heightz = widthdeg = ratio of speci c heatsacuteKD = process ef ciency

1 iexcl PR1 iexcl deg =deg

1 iexcl Tt1=TtePR=PT 1 iexcl deg =deg

acuteKE = kinetic energy ef ciency

Tte

Tt1

micro1 iexcl

2=deg iexcl 1

M 21

copyPT 1 iexcl deg =deg iexcl 1

ordfpara

frac12 = density

Received 30 September 1999 revision received 28 February 2000 ac-cepted for publication 9 March 2000 Copyright cdeg 2000 by M K SmartPublished by the American Institute of Aeronautics and Astronautics Incwith permission

currenNRC Research Associate currently Aeronautical Engineer LockheedMartin Engineering and Sciences NASA Langley Research Center MS168 Hampton VA 23681 Senior Member AIAA

Subscripts

c = stagnation chambercap = inlet capturee = inlet exiti = inside cowlm = modelo = outside cowlt = totalt2 = pitotw = wall1 = inlet entrance

Introduction

T HE design of ef cient inlets for hypersonic vehicles utilizingairframe integrated scramjet modules is a subject of interest

in the high-speed propulsion community In these con gurationsthe vehicle bow shock performs the initial compression and thecapture shape for the inlet of each scramjet module is required tobe rectangular Other requirements are that inlets start before ram-jet takeover speeds are reached (Mach 3ndash4) that they operate overa large Mach number range and that they be ef cient during ve-hicle cruise For overall system simplicity there is also a strongdesire to have an intake with xed geometry and no requirement forboundary-layer bleed Another desirable feature of a hypersonic in-let for some scramjet applications is a transition from a rectangularcapture to anellipticalthroat The inletmay then be used in combina-tion with an elliptical combustor which is superior to a rectangularcombustor in terms of the structural weight required to withstand aspeci ed pressurethermal load and the wetted surface area neededto enclose a speci ed cross-sectionalareaThis type of con gurationalso reduces the undesirable effects of hypersonic corner ows

A number of three-dimensional curved inlets leading to circularor elliptical combustors were designed and tested in the 1960s1iexcl4

These xed-geometry inlets performed well during wind-tunneltests and self-started with internal contraction ratios considerablyabove the one-dimensional theoretical starting limit rst introducedby Kantrowitz and Donaldson5 The current study involves a furtherinvestigation of the advantages of three-dimensional curved inlets atask greatly simpli ed by the advent of modern computational toolsThese tools have enabled a design methodology to be devised forhypersonic inlets in which all internal surfaces of the inlet performa portion of the compression corners are removed with a minimumof disturbance to the ow and three-dimensional shock cancellationis utilized All of these features contributed to the design of a high-performance inlet con guration that includes a transition from a

276

SMART 277

nearly rectangular capture to an elliptical throat A detailed method-ology for the design of xed-geometry rectangular-to-ellipticalshape transition (REST) inlets was reported in Ref 6 The currentpaper describes the experimental testing of a Mach 60 REST inletin a hypersonic wind tunnel with a freestreamMach number of 62

Experimental Program

Wind-Tunnel and Test Conditions

The current experiments were conducted in the NASA LangleyResearch Center Arc Heated Scramjet Test Facility (AHSTF)7 Inthis facility an electric arc heats a portion of the air ow to approx-imately 4450 K and the remaining air ow is injected downstreamof the arc heater to produce the desired test conditions The AHSTFcan provide test ows with a total enthalpy corresponding to Mach8 ight In the current tests however actual ight total enthalpywas not required and the AHSTF was operated at low temperatureto maximize Reynolds number (120ndash156 pound 106m) and to meetmodel temperature restrictions The Mach 62 nozzle used for thetests had a 277 pound 277 cm exit and exhausted into a 122 m (4 ft)diam test chamber that was connected to a 305 m (100 ft) vacuumsphere

AHSTF Mach 62 nozzle calibrations are reported in Ref 8 fora range of wind-tunnel total conditions These indicated that thenozzle ow contains vortex pairs at the midpoint of each wall thatcontaminate the core ow The inlet was therefore positioned sothat its projected capture tube contained as high a proportion ofundisturbed core ow as possible Figure 1 shows the inlet capturestream tube superimposed on a contour plot of the nozzle exit Machnumber for a typical test condition indicating that the inlet capturedonly a small portion of the ow affected by the vortex pairs Twonominal test points were established for the experimental programTable 1 lists the facility stagnation chamber conditions for these testpoints and the equivalent one-dimensional properties of the owwithin the inlet capture stream tube These values were generatedbyscaling the nozzle calibrationdata8 at stagnation chamber conditionspc D 224 MPa (325 psia) and Hc D 0614 MJkg (264 Btulbm) tothe test points utilized in the current program

Experimental Model

A photograph of the fully assembled experimental modelmounted in the AHSTF test section is shown in Fig 2 The complete

Fig 1 Mach number contours at the Mach 62 nozzle exit with theREST inlet capture area shown

Table 1 Wind-tunnel test conditions

Property Test point 1 Test point 2

pc MPa (psia) 301 (437) 267 (390)Hc MJkg (Btulbm) 0702 (302) 0556 (239)M1 618 618p1 kPa (psia) 1393 (0202) 1241 (0180)T1 K (plusmnR) 811 (1459) 640 (1153)pt1 MPa (psia) 2672 (3875) 2366 (3431)Tt1 K (plusmnR) 690 (1242) 550 (991)Re1 pound 106 miexcl1 (ftiexcl1) 120 (366) 156 (477)q1 kPa (psf) 373 (779) 332 (694)Pm1 kgs (lbs) 0757 (1669) 0760 (1675)Capture area cm2 (in2) 1138 (1764) 1138 (1764)

Fig 2 Photograph of the fully assembled REST inlet model in theAHSTF test section

model consistedof the REST inlet a cruciformrake a dump-isolatortube a mass- ow meter and supporting structure Three views ofthe fully assembled model are shown in Fig 3 with key dimensionslabeled The complete model was approximately 175 cm (69 in) inlength and was suspended in the test section from two overhead I-beam sections The inlet had a total length of 946 cm (372 in) withcowl closure 508 cm (199 in) from the most forward point thethroat a further 286 cm (113 in) downstream of cowl closure anda 152 cm (6-in) long elliptical isolator downstream of the throatThe capture area of 1138 cm2 (176 in2 ) was 152 cm (60 in) wideand 110 cm (43 in) high at its plane of symmetry and had sharpleading edges The inlet was manufactured from pure nickel usingan electroforming technique The rst step in this process involvedthe machining of an aluminium mandrel to the internal shape andsurface nish required for the inlet This was followed by nickel de-position to approximately 35 mm thickness nal machining andthen removal of the mandrel by the use of an acid bath This overallprocess provided an economical means of manufacturing the three-dimensional curved internal shape to within the desired tolerance ofsect013 mm (sect0005 in)

The xed-geometry inlet was designed using the methodologyoutlined in Ref 6 This method combined a quasi-streamtraced in-viscid technique with a correction for three-dimensional boundary-layer growth to design an inlet with nearly rectangular capture and

278 SMART

Fig 3 Three views of the REST inlet model con guration

smooth transition to an elliptical throat Its highly notched cowlwas designed to allow for ow spillage when operating below thedesign Mach number The particular inlet shown in Figs 2 and 3was designed to be mounted beneath a vehicle cruising at Mach71 on a constant q D 50 kPa (1044 psf) trajectory In combina-tion with a 6 deg forebody compression the inlet was requiredto supply a scramjet combustor with ow at or above 50 kPa(05 atm) In this instance ow enters the inlet at M1 D 60 andthe required inlet compression ratio is PR D 135 Further assump-tions used in the design of the model were an entrance Reynoldsnumber of Re1 D 26 pound 106=m a wall-to-total temperature ratio ofTw =Tt1 D 05 and a transition to turbulent ow that began 125 cm(05 in) downstream of all leading edges The resulting inlet modelhad an overall contraction ratio of 474 and an internal contractionratio of 215

The cruciform rake mounted at the rear of the inlet containedboth pitot and static pressure probes for surveying the inlet exit ow An 89 cm (35 in) diameter dump-isolator tube was connectedto the rear of the rake Fueled engine operation was simulated byreducing the exit areaof the dump isolator via translationof a conicalplug Throughout the test program this device was used to 1) obtainan estimate of the inlet mass capture and 2) examine the self-startcapability and back-pressured performance of the inlet

Instrumentation and Data Acquisition

Model instrumentation consisted of surface pressure taps pitotand static pressure probes coaxial surface thermocouples and to-tal temperature probes Model pressures were measured using aPressureSystems IncModel 8400 electronicallyscanning pressuresystem Four pressure ranges were utilized in the tests 0ndash68 kPa(0ndash1 psia) 0ndash344 kPa (0ndash5 psia) 0ndash1021 kPa (0ndash15 psia) and0ndash6807 kPa (0ndash100 psia) The error associated with the use ofthese transducers was sect05 of full scale The type E coaxial sur-face thermocouples were manufactured by Medtherm Corporationand were electrically isolated from the model The error associatedwith the use of these thermocouples was sect05 of the temperaturemeasured above the 273 K reference The type K thermocouplesused in the total temperature probes were of beaded constructionand the error associated with the use of the total temperature probeswas sect075 of the temperature measured above the 273 K ref-erence Thermocouple output was converted to temperature by auniversal temperature reference system and processed by a 16-bitNEFF AD converter Facilitydata were also processedby the NEFFAD converter All data recorded using the Autonet Version 40 dataacquisition software on a personal computer Pressure data associ-ated with the model recorded at a rate of 10 Hz and all thermocoupleand facility data recorded at a rate of 50 Hz All error bars shownin the paper were calculated using standard uncertainty analysistechniques and the aforementioned instrumentation error ranges

Figure 4 shows a schematic of the nominal streamlines alongwhich surface pressure taps were distributed in the inlet Pressuretaps were concentrated on the right-hand side of the model (as the

Fig 4 Schematic of theREST inlet pressure instru-mentation streamlines

Fig 5 Photograph of the cruciform rake containing both pitot andstatic probes

model had a vertical plane of symmetry) and were placed on the sixstreamlines (labeled A C E G I and K in Fig 4) at 15 differ-ent axial stations along the inlet These instrumentation streamlineswere approximately equally spaced around the internal surface ofthe inlet and gave a good indication of the overall surface pressuredistribution Thermocouples were placed at three representative po-sitions on the inside surface of the model that is the front of theinlet downstream of the inlet throat and on the cowl just down-stream of the inlet crotch (see Fig 3) These were used to monitorthe temperature of the inlet and also to obtain an estimate of the heattransfer to the model

A photograph of the cruciformrake is shown inFig 5 It contained24 pitot probes with an outer diameter of 10 mm and 12 staticprobes with an outer diameter of 15 mm The static probes wereequally spacedalong the horizontal and verticalbranches of the rakewhereas the pitot probes were concentratednear the walls tomeasurethe viscous losses in the inlet more accurately The static probeswere designed for internal supersonic ow measurements using themethod of Pinckney9 A full pitot and static pressure survey of theinlet exit ow required two runs with the cruciform rake rotated180 deg between runs

SMART 279

Fig 6 Schematic of the ow meter and the back-pressured ow eld

Figure 6 shows a schematic of the device used to measure thecaptured inlet mass- ow rate These measurements were obtainedby back pressuring the inletdump isolator with the conical pluguntil a shock train was stabilized near the throat of the inlet Flowwas then subsonic in the dump isolator and choked at its exit In thecurrent apparatus (referred to in the remainder of the paper as the ow meter) four pitot probes and two total temperature probes weresituated slightly upstream of the conical plug to obtain values forthe pitot pressure and total temperature of the ow exiting the owmeter Assuming a thermally perfect gas and a sonic Mach numberat the throat of the ow meter the captured inlet mass- ow rate isthen given by

Pm cap D pt p=pt p

deg =RTt Tt =T CD AFM (1)

where p=pt deg and T=Tt are all functions of Tt and are calculatedusing the methods described in Ref 10 The discharge coef cient forthe mass- ow meter CD D 0978 was obtained from a previouslyperformed calibration The error associated with the mass- ow ratecalculations was estimated at sect17

Test Program

The test program was separated into two sections 1) determina-tion of the inlet starting characteristics and back-pressured perfor-mance and 2) determination of the inlet ef ciency capability andmass capture The starting characteristics of the REST inlet wereassessedby reducing the exit areaof the ow meter until the inlet un-started then removing the ow constriction Restarting of the inletafter such a mechanically imposed unstart is generally considered tobe de nitive proof that the inlet will self-start at similar conditionsThe maximum back pressure tolerated by the inlet before unstart isalso an important design parameterbecause it provides an indicationof the amount of pressure rise that the inlet will withstand duringdual-mode scramjet engine operation

To de ne the conditions of the air entering the combustor of ascramjet engine both an inlet ef ciency term like total pressurerecovery as well as an inlet capability term like compression ratiomust be supplied11 The mass- ow rate of the air captured by theinlet is also required The ef ciency capability and mass capture ofthe REST inlet were determined using the pitot and static pressuresurveys at the inlet exit plane the surface pressure distributionswithin the inlet and the ow meter measurements Run times of35 s allowed the effect of different wall temperature conditions tobe investigated but the model was never allowed to reach adiabaticwall conditions due to the temperature limit of the epoxy used toseal the surfacepressure taps (500 K) Test sectionReynolds numberand dynamic pressure were varied as indicated in Table 1

Results and DiscussionInlet Starting Characteristics

If the inlet did not start then no useful data could be obtainedfrom the tests Hence inlet start-ability both with the initial tunnelstartup and after mechanically induced unstarts was critical to thetest program As discussed in Ref 12 and other places the keyparameter for characterizing inlet starting is the Mach number at itsplane of closure Based on computational uid dynamics analysesthe mass- ow averaged Mach number at the cowl closure planein the current experiments was Mcap D 468 The Kantrowitz limit5

indicates a maximum internal contraction of 153 at Mach 468 The

REST inlet was designed with an internal contraction ratio of 215so it violated the generally used rule of thumb for inlet starting bya signi cant margin Despite this the REST inlet was observed tostart with the wind tunnel for all test conditions Figure 7 shows aschlieren image of a typical test Flow is from left to right and theexternal shock structure generated by the inlet can be seen at thebottom of the image The appearance of a cowl shock emanatingfrom the crotch of the inlet was found to be a good indicator thatthe inlet was started

Figure 8 shows an instantaneous schlieren image of the externalshock structure observed during a mechanically induced unstartThis ow eld exhibited high-amplitude uctuations with a strongshock oscillating well upstream of the inlet crotch Figure 9 shows

Fig 7 Schlieren image of the started ow eld

Fig 8 Instantaneous schlieren image of the highly unsteady unstarted ow eld

Fig 9 Plot of the ow meter throat area and the key surface pressuresvs time during a self-start test

280 SMART

Table 2 Comparison of actual and designtest conditions

Property Actual Design

M1 618 60Re pound 106 120156 (366477) 260 (792)

miexcl1 (ftiexcl1)Tw=Tt1 044ndash071 050

a time-dependent plot of ow meter throat area AFM and two key sur-face pressure taps during a typical unstart test These taps were sit-uated just downstream of the inlet crotch one on the inner surfaceof the inlet and one on the outer surface The behavior of these twopressure measurements gave a clear indication of whether the inletwas started or not During a self-start test AFM was gradually re-duced until unstart was observed This ow process is indicated inFig 9 by the crossing over of the pi and po pressures traces Aftera period in which the inlet remained unstarted an increase of AFM

allowed the inlet to restart at an area slightly larger than the area atwhich the unstart began The REST inlet tested in the current exper-iments exhibited the ability to self-start after mechanically inducedunstarts for all test conditions

Surface and Probe Based Pressure Measurements

The REST inlet tests were initially planned to be undertaken inthe NASA Langley Research Center 20 in Mach 6 tunnel but thetests were transferred to the AHSTF due to scheduling dif cultiesFor this reason the test Mach number and Reynolds number wereslightly different from design Table 2 shows a comparison of theactual test conditions with those used to design the REST inletWhereas the wall thermal conditions of the tests were similar todesign both the increasedMachnumber and the decreasedReynoldsnumber produced a higher inlet compression ratio than the designvalue of PR D 135

Figure 10 shows the normalized surface pressure distributionsalong instrumentation streamlines A C E G and I (see Fig 4) fortest point 1 with Tw =Tt1 D 0438 together with the design valuesIn the portion of the inlet upstream of the crotch the experimentaldata was almost identical to the design values Downstream of thecrotch experimental data showed some differences in local pres-sure levels but the overall trend of the experimental and designpressure distributions remained similar The local discrepancies ob-served downstream of the crotch are believed to be due to featuresof the shockboundary-layer interactions not accounted for in thedesign process (see Ref 6) The results shown in Fig 10 are typicalof the surface pressure distributions observed throughout the testprogram It was generally observed that the inlet surface pressuredistributions were somewhat sensitive to the wall thermal conditionbut exhibited minimal sensitivityto the changes in Reynolds numberbetween the two test points

Both the pitot and static pressure distributions at the exit of theinlet were measured using the cruciform rake In combination withthe mean total temperature of the ow at the inlet exit these mea-surements were used to calculatea complete set of properties at eachprobe location assuming the air behaved as a thermally perfect gasTo determine the total temperature of the inlet exit ow an estimateof the average heat transfer rate to the model Pqm was obtained using

Pqm DdTm

dtC pNi frac12Ni vm (2)

As noted earlier the inside surface temperature was measured atthree representative locations on the model (see Fig 3) As themodel was a thin-walled structure (wall thickness frac1435 mm) thetemperature gradient of the entire structure dTm =dt was assumedto be equal to the average of the three representative surface tem-perature gradients Employing the speci c heat of nickel CpNi thedensity of nickel frac12Ni and the volume of the model vm use of Eq (2)resulted in typical heat transfer rates to the model of between 06and 15 of the captured total enthalpy ux Subtracting the heat

Fig 10 Surface pressure comparison between design and experimentalong instrumentation streamlines

transfer rate from the captured total enthalpy ux gave the total en-thalpy ux of the ow exiting the inlet The methods described inRef 10 were used to calculate the total temperature correspondingto the exit ow enthalpy ux

Figures 11a and 11b show the normalized pitot and staticpressuredistributions obtained along the horizontal and vertical branches ofthe rake for test point 2 with Tw=Tt1 D 0631 Figure 11a indicatesthat both the pe and pt2e distributions are nearly symmetric aboutthe centerline of the inlet Interestingly pe was observed to be ap-proximately constant across the horizontal branch whereas pt2e

showed plateaus near both walls smooth peaks approximately mid-way between the center and walls and a central dip The dip inpt2e at the center of the inlet was an unexpected feature of the ow especially given the relatively constant pe Probe interferenceand other possible problems associatedwith the central pitot probeswere investigated as causes of this phenomenon without result It is

SMART 281

a) Horizontal branch

b) Vertical branch

Fig 11 Pitot and static pressure probe measurements at the inlet exit

possible that the central dip in pt2e is a consequence of the nonuni-formities present in the inlet capture ow (Fig 1) because the in-let was designed with the assumption of uniform freestream owFigure 11b shows that pe remained constant on the lower two-thirdsof the vertical branch then jumped to a level approximately 50higher on the upper one-third The corresponding pt2e distributionpeaked toward the top of the inlet and exhibited a sharp dip in theviscous-dominated region near the top wall On the bottom portionof the vertical branch pt2e dropped from its peak at a more gradualrate due to increased shock losses near the cowl These distributionsare typical of those observed at the exit of the inlet throughout thetest program

Figures 12a and 12b show the Mach number and total pressuredistributions calculated from the rake measurements On the hor-izontal branch (Fig 12a) both Me and pte show a similar shapeto pt2e with the Mach number peaking at Me D 39 approximatelyhalfway between the center and the wall and dipping to Me D 35 atthe center In general the presence of the centraldip in pitot pressurehad a signi cant effect on all of the calculated inlet ow propertiesThe vertical branch distributions of Me and pte (Fig 12b) were alsosimilar to pt2e with the Mach number peaking at Me D 38 towardthe top of the inlet exit Note that the distribution of frac12eue (not shown)was also observed to show a similar trend to the distribution of pt2e

Inlet Ef ciency Capability and Mass Capture

Equivalent one-dimensional mass- ow-weighted properties havebeencalculatedat the inlet exit using the probe measurementsThesewere generatedby dividing the inlet exit area into four sections thenconstructing a series of elliptical-annular strips centered on each ofthe pitot probe positions on the horizontal and vertical branches ofthe rake The properties at the pitot probe positions were then as-sumed to correspond to the average values over the strip When inte-

Table 3 Equivalent one-dimensionalproperties at the inlet exit for test point 2a

Property Test point 2

pt MPa (psia) 1269 (1840)Tt K (plusmnR) 5436 (9784)Ht MJkg (Btulb) 0548 (2358)M 3264p kPa (psia) 1876 (2721)T K (plusmnR) 1813 (3263)frac12 kgm3 (lbft3) 0377 (235 pound 10iexcl2)u ms (fts) 8544 (2803)Pm kgs (lbs) 0732 (1615)Area cm2 (in2) 2437 (3778)

aTw=Tt1 D 0631

a) Horizontal branch

b) Vertical branch

Fig 12 Calculated Mach number and total pressure ratio at the inletexit

grated over the entire exit area of the inlet this averaging techniquewas expected to supply a best possible one-dimensional estimate ofthe inlet performance given the available data Table 3 includes acomplete list of the exit ow properties corresponding to the datashown in Figs 11 and 12

Numerous experiments were conducted at both test points withdifferent wall temperature conditions Figure 13 shows plots ofmass- ow-weighted equivalent one-dimensional ef ciency param-eters namely the total pressure recovery PT the kinetic energyef ciency acuteKE and the process ef ciency acuteKD vs wall-to-total tem-perature ratio Tw=Tt1 Note from Fig 13 that the ef ciency curvesfor both test points are almost identical where they overlap Thisindicated that the 30 difference in Reynolds number between testpoints 1 and 2 had a minimal effect on inlet ef ciency Also note

282 SMART

Fig 13 Plots of inlet ef ciency parameters vs wall-to-total tempera-ture ratio

Fig 14 Plots of inlet capability parameters vs wall-to-total tempera-ture ratio

from Fig 13 that Tw=Tt1 had a signi cant effect on PT For exam-ple increasing Tw=Tt1 from 044 to 071 reduced PT from 0565 to0517 This decrease in PT was due to the increasedboundary-layerthickness for the higher Tw =Tt1 Process ef ciency showed a lesspronounced decrease with increasing Tw=Tt1 with values in therange acuteKD D 0895ndash0877 for the tests In contrast to PT and acuteKDkinetic energy ef ciency showed a slight increase with increasingTw =Tt1 with values in the range acuteKE D 0963ndash0968 for the tests

The wall-to-total temperature ratio effect was also manifested inthe inlet capability as indicatedby parameterslike compression ratioPR temperature ratio TR and exit Mach number Me which are plot-ted vs Tw=Tt1 in Fig 14 The increased boundary-layer thicknessfor the higher Tw =Tt1 values raised the pressure and temperatureratio by increasing the aerodynamic contraction experienced by theinviscid core ow The same phenomenon also reduced the Machnumber at the inlet exit Increase of Tw=Tt1 from 044 to 071 pro-duced a 107 increase in PR to a maximum for the current tests ofPR D 1555 Both TR and Me showed a less pronounced dependenceon Tw=Tt1 with typical values of TR D 281 and Me D 329

Table 4 lists the ranges of the mass- ow-weighted equivalent one-dimensional values for PR TR Me PT acuteKE and acuteKD measured forthe REST inlet at the two test points The values listed in Table 4correspond to mass capture percentages of m c D 960 and 966 fortest points 1 and 2 respectively Both of these values are within theerror range of the corresponding mass- ow meter measurementswhich were mc D 962 sect 17 and 965 sect 17 for test points 1and 2 respectively

Table 4 Equivalent one-dimensional performanceparameter ranges for test points 1 and 2

Test point 1 Test point 2Property Tw=Tt1 D 044ndash059 Tw =Tt1 D 054ndash071

PR 1410ndash1518 1465ndash1555TR 272ndash284 279ndash290Me 335ndash328 332ndash322PT 0565ndash0544 0555ndash0512acuteKE 0963ndash0966 0965ndash0968acuteKD 0895ndash0887 0888ndash0877mc 962 sect 17 965 sect 17Pqm =Ht1 120ndash061 148ndash113

Fig 15 Surface pressure distributions at the maximum back-pressurecondition

Inlet Back Pressure Performance

Figure 15 shows a typical pressure distribution in the inlet fortest point 2 at maximum back pressure that is just before a me-chanically induced unstart This data supplies an indication of themaximum pressure the REST inlet could sustain in a combustor dur-ing dual-mode scramjet operation A plot of the average pressurealong the inlet without back pressure has been added to Fig 15 forreference Note that when fully back pressured the axial pressuredistributions along all streamlinesbecome similarThis is an indica-tion that ow downstream of the inlet crotch is fully separated witha shock train situated inside the enclosed portion of the inlet Basedon the data shown in Fig 15 the ratio of maximum back pressureto the pressure generated with no back pressure is 57 for this RESTinlet at Mach 62 ( note also that the maximum back pressure is ap-proximately 90 times the freestreampressure) Measured maximumback-pressure ratios varied between 53 and 57 throughout the testprogram

ConclusionsResults of Mach 62 wind-tunnel testing of a xed-geometry

hypersonic inlet were reported This three-dimensional quasi-streamtraced inlet had an overall contraction ratio of 474 an in-ternal contraction ratio of 215 and a Mach 60 design point Italso included a transition from a nearly rectangular capture to anelliptical throat The tests were conducted to validate a previouslydeveloped design methodology for xed-geometry inlet con gura-tions that include signi cant cross-sectional shape changes The in-let was tested at a Reynolds numbers of Re D 120 and 156 pound 106m(366 and 477 pound 106ft) and at wall-to-total temperature ratios rang-ing between 044 and 071

The inlet was observed to start with the wind tunnel at all timesduring the test program It also exhibited self-startingcapability af-ter mechanically induced inlet unstarts These results indicated thatthe Kantrowitz5 starting limit does not necessarily apply to all inlet

SMART 283

con gurations Surface pressure distributions within the inlet werefound to be similar to design However the overall compressionratio generated by the inlet was higher than design because the in-let was tested above the design Mach number and below the designReynolds number Equivalent one-dimensional performance param-eters were obtained using both pitot and static pressure probes atthe inlet exit On a mass- ow-weigthed basis capability parameterssuch as compression ratio and exit Mach number ranged between141 and 156 and 335 and 322 respectively Ef ciency param-eters such as total pressure recovery and kinetic energy ef ciencyranged between 0565 and 0512 and 0963 and 0968 respectivelyIn general increased values of wall-to-total temperature ratio led toincreased inlet capability but at a reduced total pressure recoveryInlet mass capture was deduced from measurements to be approx-imately 96 of the available air ow Heat transfer to the modelranged between 06 and 15 of the captured total enthalpy uxFinally back-pressurizationdata indicated that the maximum back-pressure ratio relative to unhindered operation varied between 53and 57 This test program demonstrated that the inlet shape tran-sition required to utilize elliptically shaped combustors in airframeintegrated scramjets can be accomplished without sacri cing inletperformance

AcknowledgmentsThis work was supported by the Hypersonic Airbreathing Propul-

sion Branch and the Hyper-X Program Of ce at NASA Langley Re-search Center The author would like to thank E Ruf and A Auslen-der for many useful discussions of the work and R Irby J Connoverand J Simmons for their efforts during the test program

References1Hartill W B ldquoAnalytical and Experimental Investigation of a Scram-

jet Inlet of Quadriform Shaperdquo US Air Force Rept AFAPL-TR-65-74Marquardt Corp Aug 1965

2Billig F S ldquoSupersonic Combustion Ramjet Missilerdquo Journal ofPropulsion and Power Vol 11 No 6 1995 pp 1139ndash1146

3Kutshenreuter P H ldquoHypersonic Inlet Tests inHelium and AirrdquoGeneralElectric Co Advanced Engine and Technology Departments CincinnatiOH June 1965

4Molder S and Romeskie J M ldquoModular Hypersonic Inlets with Con-ical Flowrdquo CP-30 AGARD 1968

5Kantrowitz A and Donaldson C ldquoPreliminary Investigation of Super-sonic Diffusersrdquo NACA WR L-713 1945

6Smart M K ldquoDesign of Three-Dimensional Hypersonic Inlets withRectangular-to-Elliptical Shape Transitionrdquo Journal of Propulsion andPower Vol 15 No 3 1999 pp 408ndash416

7Guy R W Rogers R C Puster R L Rock K E and Diskin G SldquoThe NASA Langley Scramjet Test Complexrdquo AIAA Paper 96-3243 July1996

8Thomas S R Voland R T and Guy R W ldquoTest Flow CalibrationStudy of the Langley Arc- Heated Scramjet Test Facilityrdquo AIAA Paper 87-2165 June 1987

9Pinckney S Z ldquoImproved Static Probe Designrdquo AIAA Journal Vol 12No 4 1974 p 562

10Witte D W and Tatum K E ldquoComputer Code for Determination ofThermally Perfect Gas Propertiesrdquo NASA TP-3447 Sept 1994

11Molder S McGregor R J and Paisley T W ldquoA Comparison of ThreeHypersonic Inletsrdquo Sec 6 ldquoInvestigations in the Fluid Mechanics of Scram-jet Inletsrdquo US Air Force Rept WL-TR-93-2091 Ryerson Polytechnic InstToronto 1993

12Van Wie D M Kwok F T and Walsh R F ldquoStarting Characteristicsof Supersonic Inletsrdquo AIAA Paper 96-2914 July 1996

SMART 277

nearly rectangular capture to an elliptical throat A detailed method-ology for the design of xed-geometry rectangular-to-ellipticalshape transition (REST) inlets was reported in Ref 6 The currentpaper describes the experimental testing of a Mach 60 REST inletin a hypersonic wind tunnel with a freestreamMach number of 62

Experimental Program

Wind-Tunnel and Test Conditions

The current experiments were conducted in the NASA LangleyResearch Center Arc Heated Scramjet Test Facility (AHSTF)7 Inthis facility an electric arc heats a portion of the air ow to approx-imately 4450 K and the remaining air ow is injected downstreamof the arc heater to produce the desired test conditions The AHSTFcan provide test ows with a total enthalpy corresponding to Mach8 ight In the current tests however actual ight total enthalpywas not required and the AHSTF was operated at low temperatureto maximize Reynolds number (120ndash156 pound 106m) and to meetmodel temperature restrictions The Mach 62 nozzle used for thetests had a 277 pound 277 cm exit and exhausted into a 122 m (4 ft)diam test chamber that was connected to a 305 m (100 ft) vacuumsphere

AHSTF Mach 62 nozzle calibrations are reported in Ref 8 fora range of wind-tunnel total conditions These indicated that thenozzle ow contains vortex pairs at the midpoint of each wall thatcontaminate the core ow The inlet was therefore positioned sothat its projected capture tube contained as high a proportion ofundisturbed core ow as possible Figure 1 shows the inlet capturestream tube superimposed on a contour plot of the nozzle exit Machnumber for a typical test condition indicating that the inlet capturedonly a small portion of the ow affected by the vortex pairs Twonominal test points were established for the experimental programTable 1 lists the facility stagnation chamber conditions for these testpoints and the equivalent one-dimensional properties of the owwithin the inlet capture stream tube These values were generatedbyscaling the nozzle calibrationdata8 at stagnation chamber conditionspc D 224 MPa (325 psia) and Hc D 0614 MJkg (264 Btulbm) tothe test points utilized in the current program

Experimental Model

A photograph of the fully assembled experimental modelmounted in the AHSTF test section is shown in Fig 2 The complete

Fig 1 Mach number contours at the Mach 62 nozzle exit with theREST inlet capture area shown

Table 1 Wind-tunnel test conditions

Property Test point 1 Test point 2

pc MPa (psia) 301 (437) 267 (390)Hc MJkg (Btulbm) 0702 (302) 0556 (239)M1 618 618p1 kPa (psia) 1393 (0202) 1241 (0180)T1 K (plusmnR) 811 (1459) 640 (1153)pt1 MPa (psia) 2672 (3875) 2366 (3431)Tt1 K (plusmnR) 690 (1242) 550 (991)Re1 pound 106 miexcl1 (ftiexcl1) 120 (366) 156 (477)q1 kPa (psf) 373 (779) 332 (694)Pm1 kgs (lbs) 0757 (1669) 0760 (1675)Capture area cm2 (in2) 1138 (1764) 1138 (1764)

Fig 2 Photograph of the fully assembled REST inlet model in theAHSTF test section

model consistedof the REST inlet a cruciformrake a dump-isolatortube a mass- ow meter and supporting structure Three views ofthe fully assembled model are shown in Fig 3 with key dimensionslabeled The complete model was approximately 175 cm (69 in) inlength and was suspended in the test section from two overhead I-beam sections The inlet had a total length of 946 cm (372 in) withcowl closure 508 cm (199 in) from the most forward point thethroat a further 286 cm (113 in) downstream of cowl closure anda 152 cm (6-in) long elliptical isolator downstream of the throatThe capture area of 1138 cm2 (176 in2 ) was 152 cm (60 in) wideand 110 cm (43 in) high at its plane of symmetry and had sharpleading edges The inlet was manufactured from pure nickel usingan electroforming technique The rst step in this process involvedthe machining of an aluminium mandrel to the internal shape andsurface nish required for the inlet This was followed by nickel de-position to approximately 35 mm thickness nal machining andthen removal of the mandrel by the use of an acid bath This overallprocess provided an economical means of manufacturing the three-dimensional curved internal shape to within the desired tolerance ofsect013 mm (sect0005 in)

The xed-geometry inlet was designed using the methodologyoutlined in Ref 6 This method combined a quasi-streamtraced in-viscid technique with a correction for three-dimensional boundary-layer growth to design an inlet with nearly rectangular capture and

278 SMART

Fig 3 Three views of the REST inlet model con guration

smooth transition to an elliptical throat Its highly notched cowlwas designed to allow for ow spillage when operating below thedesign Mach number The particular inlet shown in Figs 2 and 3was designed to be mounted beneath a vehicle cruising at Mach71 on a constant q D 50 kPa (1044 psf) trajectory In combina-tion with a 6 deg forebody compression the inlet was requiredto supply a scramjet combustor with ow at or above 50 kPa(05 atm) In this instance ow enters the inlet at M1 D 60 andthe required inlet compression ratio is PR D 135 Further assump-tions used in the design of the model were an entrance Reynoldsnumber of Re1 D 26 pound 106=m a wall-to-total temperature ratio ofTw =Tt1 D 05 and a transition to turbulent ow that began 125 cm(05 in) downstream of all leading edges The resulting inlet modelhad an overall contraction ratio of 474 and an internal contractionratio of 215

The cruciform rake mounted at the rear of the inlet containedboth pitot and static pressure probes for surveying the inlet exit ow An 89 cm (35 in) diameter dump-isolator tube was connectedto the rear of the rake Fueled engine operation was simulated byreducing the exit areaof the dump isolator via translationof a conicalplug Throughout the test program this device was used to 1) obtainan estimate of the inlet mass capture and 2) examine the self-startcapability and back-pressured performance of the inlet

Instrumentation and Data Acquisition

Model instrumentation consisted of surface pressure taps pitotand static pressure probes coaxial surface thermocouples and to-tal temperature probes Model pressures were measured using aPressureSystems IncModel 8400 electronicallyscanning pressuresystem Four pressure ranges were utilized in the tests 0ndash68 kPa(0ndash1 psia) 0ndash344 kPa (0ndash5 psia) 0ndash1021 kPa (0ndash15 psia) and0ndash6807 kPa (0ndash100 psia) The error associated with the use ofthese transducers was sect05 of full scale The type E coaxial sur-face thermocouples were manufactured by Medtherm Corporationand were electrically isolated from the model The error associatedwith the use of these thermocouples was sect05 of the temperaturemeasured above the 273 K reference The type K thermocouplesused in the total temperature probes were of beaded constructionand the error associated with the use of the total temperature probeswas sect075 of the temperature measured above the 273 K ref-erence Thermocouple output was converted to temperature by auniversal temperature reference system and processed by a 16-bitNEFF AD converter Facilitydata were also processedby the NEFFAD converter All data recorded using the Autonet Version 40 dataacquisition software on a personal computer Pressure data associ-ated with the model recorded at a rate of 10 Hz and all thermocoupleand facility data recorded at a rate of 50 Hz All error bars shownin the paper were calculated using standard uncertainty analysistechniques and the aforementioned instrumentation error ranges

Figure 4 shows a schematic of the nominal streamlines alongwhich surface pressure taps were distributed in the inlet Pressuretaps were concentrated on the right-hand side of the model (as the

Fig 4 Schematic of theREST inlet pressure instru-mentation streamlines

Fig 5 Photograph of the cruciform rake containing both pitot andstatic probes

model had a vertical plane of symmetry) and were placed on the sixstreamlines (labeled A C E G I and K in Fig 4) at 15 differ-ent axial stations along the inlet These instrumentation streamlineswere approximately equally spaced around the internal surface ofthe inlet and gave a good indication of the overall surface pressuredistribution Thermocouples were placed at three representative po-sitions on the inside surface of the model that is the front of theinlet downstream of the inlet throat and on the cowl just down-stream of the inlet crotch (see Fig 3) These were used to monitorthe temperature of the inlet and also to obtain an estimate of the heattransfer to the model

A photograph of the cruciformrake is shown inFig 5 It contained24 pitot probes with an outer diameter of 10 mm and 12 staticprobes with an outer diameter of 15 mm The static probes wereequally spacedalong the horizontal and verticalbranches of the rakewhereas the pitot probes were concentratednear the walls tomeasurethe viscous losses in the inlet more accurately The static probeswere designed for internal supersonic ow measurements using themethod of Pinckney9 A full pitot and static pressure survey of theinlet exit ow required two runs with the cruciform rake rotated180 deg between runs

SMART 279

Fig 6 Schematic of the ow meter and the back-pressured ow eld

Figure 6 shows a schematic of the device used to measure thecaptured inlet mass- ow rate These measurements were obtainedby back pressuring the inletdump isolator with the conical pluguntil a shock train was stabilized near the throat of the inlet Flowwas then subsonic in the dump isolator and choked at its exit In thecurrent apparatus (referred to in the remainder of the paper as the ow meter) four pitot probes and two total temperature probes weresituated slightly upstream of the conical plug to obtain values forthe pitot pressure and total temperature of the ow exiting the owmeter Assuming a thermally perfect gas and a sonic Mach numberat the throat of the ow meter the captured inlet mass- ow rate isthen given by

Pm cap D pt p=pt p

deg =RTt Tt =T CD AFM (1)

where p=pt deg and T=Tt are all functions of Tt and are calculatedusing the methods described in Ref 10 The discharge coef cient forthe mass- ow meter CD D 0978 was obtained from a previouslyperformed calibration The error associated with the mass- ow ratecalculations was estimated at sect17

Test Program

The test program was separated into two sections 1) determina-tion of the inlet starting characteristics and back-pressured perfor-mance and 2) determination of the inlet ef ciency capability andmass capture The starting characteristics of the REST inlet wereassessedby reducing the exit areaof the ow meter until the inlet un-started then removing the ow constriction Restarting of the inletafter such a mechanically imposed unstart is generally considered tobe de nitive proof that the inlet will self-start at similar conditionsThe maximum back pressure tolerated by the inlet before unstart isalso an important design parameterbecause it provides an indicationof the amount of pressure rise that the inlet will withstand duringdual-mode scramjet engine operation

To de ne the conditions of the air entering the combustor of ascramjet engine both an inlet ef ciency term like total pressurerecovery as well as an inlet capability term like compression ratiomust be supplied11 The mass- ow rate of the air captured by theinlet is also required The ef ciency capability and mass capture ofthe REST inlet were determined using the pitot and static pressuresurveys at the inlet exit plane the surface pressure distributionswithin the inlet and the ow meter measurements Run times of35 s allowed the effect of different wall temperature conditions tobe investigated but the model was never allowed to reach adiabaticwall conditions due to the temperature limit of the epoxy used toseal the surfacepressure taps (500 K) Test sectionReynolds numberand dynamic pressure were varied as indicated in Table 1

Results and DiscussionInlet Starting Characteristics

If the inlet did not start then no useful data could be obtainedfrom the tests Hence inlet start-ability both with the initial tunnelstartup and after mechanically induced unstarts was critical to thetest program As discussed in Ref 12 and other places the keyparameter for characterizing inlet starting is the Mach number at itsplane of closure Based on computational uid dynamics analysesthe mass- ow averaged Mach number at the cowl closure planein the current experiments was Mcap D 468 The Kantrowitz limit5

indicates a maximum internal contraction of 153 at Mach 468 The

REST inlet was designed with an internal contraction ratio of 215so it violated the generally used rule of thumb for inlet starting bya signi cant margin Despite this the REST inlet was observed tostart with the wind tunnel for all test conditions Figure 7 shows aschlieren image of a typical test Flow is from left to right and theexternal shock structure generated by the inlet can be seen at thebottom of the image The appearance of a cowl shock emanatingfrom the crotch of the inlet was found to be a good indicator thatthe inlet was started

Figure 8 shows an instantaneous schlieren image of the externalshock structure observed during a mechanically induced unstartThis ow eld exhibited high-amplitude uctuations with a strongshock oscillating well upstream of the inlet crotch Figure 9 shows

Fig 7 Schlieren image of the started ow eld

Fig 8 Instantaneous schlieren image of the highly unsteady unstarted ow eld

Fig 9 Plot of the ow meter throat area and the key surface pressuresvs time during a self-start test

280 SMART

Table 2 Comparison of actual and designtest conditions

Property Actual Design

M1 618 60Re pound 106 120156 (366477) 260 (792)

miexcl1 (ftiexcl1)Tw=Tt1 044ndash071 050

a time-dependent plot of ow meter throat area AFM and two key sur-face pressure taps during a typical unstart test These taps were sit-uated just downstream of the inlet crotch one on the inner surfaceof the inlet and one on the outer surface The behavior of these twopressure measurements gave a clear indication of whether the inletwas started or not During a self-start test AFM was gradually re-duced until unstart was observed This ow process is indicated inFig 9 by the crossing over of the pi and po pressures traces Aftera period in which the inlet remained unstarted an increase of AFM

allowed the inlet to restart at an area slightly larger than the area atwhich the unstart began The REST inlet tested in the current exper-iments exhibited the ability to self-start after mechanically inducedunstarts for all test conditions

Surface and Probe Based Pressure Measurements

The REST inlet tests were initially planned to be undertaken inthe NASA Langley Research Center 20 in Mach 6 tunnel but thetests were transferred to the AHSTF due to scheduling dif cultiesFor this reason the test Mach number and Reynolds number wereslightly different from design Table 2 shows a comparison of theactual test conditions with those used to design the REST inletWhereas the wall thermal conditions of the tests were similar todesign both the increasedMachnumber and the decreasedReynoldsnumber produced a higher inlet compression ratio than the designvalue of PR D 135

Figure 10 shows the normalized surface pressure distributionsalong instrumentation streamlines A C E G and I (see Fig 4) fortest point 1 with Tw =Tt1 D 0438 together with the design valuesIn the portion of the inlet upstream of the crotch the experimentaldata was almost identical to the design values Downstream of thecrotch experimental data showed some differences in local pres-sure levels but the overall trend of the experimental and designpressure distributions remained similar The local discrepancies ob-served downstream of the crotch are believed to be due to featuresof the shockboundary-layer interactions not accounted for in thedesign process (see Ref 6) The results shown in Fig 10 are typicalof the surface pressure distributions observed throughout the testprogram It was generally observed that the inlet surface pressuredistributions were somewhat sensitive to the wall thermal conditionbut exhibited minimal sensitivityto the changes in Reynolds numberbetween the two test points

Both the pitot and static pressure distributions at the exit of theinlet were measured using the cruciform rake In combination withthe mean total temperature of the ow at the inlet exit these mea-surements were used to calculatea complete set of properties at eachprobe location assuming the air behaved as a thermally perfect gasTo determine the total temperature of the inlet exit ow an estimateof the average heat transfer rate to the model Pqm was obtained using

Pqm DdTm

dtC pNi frac12Ni vm (2)

As noted earlier the inside surface temperature was measured atthree representative locations on the model (see Fig 3) As themodel was a thin-walled structure (wall thickness frac1435 mm) thetemperature gradient of the entire structure dTm =dt was assumedto be equal to the average of the three representative surface tem-perature gradients Employing the speci c heat of nickel CpNi thedensity of nickel frac12Ni and the volume of the model vm use of Eq (2)resulted in typical heat transfer rates to the model of between 06and 15 of the captured total enthalpy ux Subtracting the heat

Fig 10 Surface pressure comparison between design and experimentalong instrumentation streamlines

transfer rate from the captured total enthalpy ux gave the total en-thalpy ux of the ow exiting the inlet The methods described inRef 10 were used to calculate the total temperature correspondingto the exit ow enthalpy ux

Figures 11a and 11b show the normalized pitot and staticpressuredistributions obtained along the horizontal and vertical branches ofthe rake for test point 2 with Tw=Tt1 D 0631 Figure 11a indicatesthat both the pe and pt2e distributions are nearly symmetric aboutthe centerline of the inlet Interestingly pe was observed to be ap-proximately constant across the horizontal branch whereas pt2e

showed plateaus near both walls smooth peaks approximately mid-way between the center and walls and a central dip The dip inpt2e at the center of the inlet was an unexpected feature of the ow especially given the relatively constant pe Probe interferenceand other possible problems associatedwith the central pitot probeswere investigated as causes of this phenomenon without result It is

SMART 281

a) Horizontal branch

b) Vertical branch

Fig 11 Pitot and static pressure probe measurements at the inlet exit

possible that the central dip in pt2e is a consequence of the nonuni-formities present in the inlet capture ow (Fig 1) because the in-let was designed with the assumption of uniform freestream owFigure 11b shows that pe remained constant on the lower two-thirdsof the vertical branch then jumped to a level approximately 50higher on the upper one-third The corresponding pt2e distributionpeaked toward the top of the inlet and exhibited a sharp dip in theviscous-dominated region near the top wall On the bottom portionof the vertical branch pt2e dropped from its peak at a more gradualrate due to increased shock losses near the cowl These distributionsare typical of those observed at the exit of the inlet throughout thetest program

Figures 12a and 12b show the Mach number and total pressuredistributions calculated from the rake measurements On the hor-izontal branch (Fig 12a) both Me and pte show a similar shapeto pt2e with the Mach number peaking at Me D 39 approximatelyhalfway between the center and the wall and dipping to Me D 35 atthe center In general the presence of the centraldip in pitot pressurehad a signi cant effect on all of the calculated inlet ow propertiesThe vertical branch distributions of Me and pte (Fig 12b) were alsosimilar to pt2e with the Mach number peaking at Me D 38 towardthe top of the inlet exit Note that the distribution of frac12eue (not shown)was also observed to show a similar trend to the distribution of pt2e

Inlet Ef ciency Capability and Mass Capture

Equivalent one-dimensional mass- ow-weighted properties havebeencalculatedat the inlet exit using the probe measurementsThesewere generatedby dividing the inlet exit area into four sections thenconstructing a series of elliptical-annular strips centered on each ofthe pitot probe positions on the horizontal and vertical branches ofthe rake The properties at the pitot probe positions were then as-sumed to correspond to the average values over the strip When inte-

Table 3 Equivalent one-dimensionalproperties at the inlet exit for test point 2a

Property Test point 2

pt MPa (psia) 1269 (1840)Tt K (plusmnR) 5436 (9784)Ht MJkg (Btulb) 0548 (2358)M 3264p kPa (psia) 1876 (2721)T K (plusmnR) 1813 (3263)frac12 kgm3 (lbft3) 0377 (235 pound 10iexcl2)u ms (fts) 8544 (2803)Pm kgs (lbs) 0732 (1615)Area cm2 (in2) 2437 (3778)

aTw=Tt1 D 0631

a) Horizontal branch

b) Vertical branch

Fig 12 Calculated Mach number and total pressure ratio at the inletexit

grated over the entire exit area of the inlet this averaging techniquewas expected to supply a best possible one-dimensional estimate ofthe inlet performance given the available data Table 3 includes acomplete list of the exit ow properties corresponding to the datashown in Figs 11 and 12

Numerous experiments were conducted at both test points withdifferent wall temperature conditions Figure 13 shows plots ofmass- ow-weighted equivalent one-dimensional ef ciency param-eters namely the total pressure recovery PT the kinetic energyef ciency acuteKE and the process ef ciency acuteKD vs wall-to-total tem-perature ratio Tw=Tt1 Note from Fig 13 that the ef ciency curvesfor both test points are almost identical where they overlap Thisindicated that the 30 difference in Reynolds number between testpoints 1 and 2 had a minimal effect on inlet ef ciency Also note

282 SMART

Fig 13 Plots of inlet ef ciency parameters vs wall-to-total tempera-ture ratio

Fig 14 Plots of inlet capability parameters vs wall-to-total tempera-ture ratio

from Fig 13 that Tw=Tt1 had a signi cant effect on PT For exam-ple increasing Tw=Tt1 from 044 to 071 reduced PT from 0565 to0517 This decrease in PT was due to the increasedboundary-layerthickness for the higher Tw =Tt1 Process ef ciency showed a lesspronounced decrease with increasing Tw=Tt1 with values in therange acuteKD D 0895ndash0877 for the tests In contrast to PT and acuteKDkinetic energy ef ciency showed a slight increase with increasingTw =Tt1 with values in the range acuteKE D 0963ndash0968 for the tests

The wall-to-total temperature ratio effect was also manifested inthe inlet capability as indicatedby parameterslike compression ratioPR temperature ratio TR and exit Mach number Me which are plot-ted vs Tw=Tt1 in Fig 14 The increased boundary-layer thicknessfor the higher Tw =Tt1 values raised the pressure and temperatureratio by increasing the aerodynamic contraction experienced by theinviscid core ow The same phenomenon also reduced the Machnumber at the inlet exit Increase of Tw=Tt1 from 044 to 071 pro-duced a 107 increase in PR to a maximum for the current tests ofPR D 1555 Both TR and Me showed a less pronounced dependenceon Tw=Tt1 with typical values of TR D 281 and Me D 329

Table 4 lists the ranges of the mass- ow-weighted equivalent one-dimensional values for PR TR Me PT acuteKE and acuteKD measured forthe REST inlet at the two test points The values listed in Table 4correspond to mass capture percentages of m c D 960 and 966 fortest points 1 and 2 respectively Both of these values are within theerror range of the corresponding mass- ow meter measurementswhich were mc D 962 sect 17 and 965 sect 17 for test points 1and 2 respectively

Table 4 Equivalent one-dimensional performanceparameter ranges for test points 1 and 2

Test point 1 Test point 2Property Tw=Tt1 D 044ndash059 Tw =Tt1 D 054ndash071

PR 1410ndash1518 1465ndash1555TR 272ndash284 279ndash290Me 335ndash328 332ndash322PT 0565ndash0544 0555ndash0512acuteKE 0963ndash0966 0965ndash0968acuteKD 0895ndash0887 0888ndash0877mc 962 sect 17 965 sect 17Pqm =Ht1 120ndash061 148ndash113

Fig 15 Surface pressure distributions at the maximum back-pressurecondition

Inlet Back Pressure Performance

Figure 15 shows a typical pressure distribution in the inlet fortest point 2 at maximum back pressure that is just before a me-chanically induced unstart This data supplies an indication of themaximum pressure the REST inlet could sustain in a combustor dur-ing dual-mode scramjet operation A plot of the average pressurealong the inlet without back pressure has been added to Fig 15 forreference Note that when fully back pressured the axial pressuredistributions along all streamlinesbecome similarThis is an indica-tion that ow downstream of the inlet crotch is fully separated witha shock train situated inside the enclosed portion of the inlet Basedon the data shown in Fig 15 the ratio of maximum back pressureto the pressure generated with no back pressure is 57 for this RESTinlet at Mach 62 ( note also that the maximum back pressure is ap-proximately 90 times the freestreampressure) Measured maximumback-pressure ratios varied between 53 and 57 throughout the testprogram

ConclusionsResults of Mach 62 wind-tunnel testing of a xed-geometry

hypersonic inlet were reported This three-dimensional quasi-streamtraced inlet had an overall contraction ratio of 474 an in-ternal contraction ratio of 215 and a Mach 60 design point Italso included a transition from a nearly rectangular capture to anelliptical throat The tests were conducted to validate a previouslydeveloped design methodology for xed-geometry inlet con gura-tions that include signi cant cross-sectional shape changes The in-let was tested at a Reynolds numbers of Re D 120 and 156 pound 106m(366 and 477 pound 106ft) and at wall-to-total temperature ratios rang-ing between 044 and 071

The inlet was observed to start with the wind tunnel at all timesduring the test program It also exhibited self-startingcapability af-ter mechanically induced inlet unstarts These results indicated thatthe Kantrowitz5 starting limit does not necessarily apply to all inlet

SMART 283

con gurations Surface pressure distributions within the inlet werefound to be similar to design However the overall compressionratio generated by the inlet was higher than design because the in-let was tested above the design Mach number and below the designReynolds number Equivalent one-dimensional performance param-eters were obtained using both pitot and static pressure probes atthe inlet exit On a mass- ow-weigthed basis capability parameterssuch as compression ratio and exit Mach number ranged between141 and 156 and 335 and 322 respectively Ef ciency param-eters such as total pressure recovery and kinetic energy ef ciencyranged between 0565 and 0512 and 0963 and 0968 respectivelyIn general increased values of wall-to-total temperature ratio led toincreased inlet capability but at a reduced total pressure recoveryInlet mass capture was deduced from measurements to be approx-imately 96 of the available air ow Heat transfer to the modelranged between 06 and 15 of the captured total enthalpy uxFinally back-pressurizationdata indicated that the maximum back-pressure ratio relative to unhindered operation varied between 53and 57 This test program demonstrated that the inlet shape tran-sition required to utilize elliptically shaped combustors in airframeintegrated scramjets can be accomplished without sacri cing inletperformance

AcknowledgmentsThis work was supported by the Hypersonic Airbreathing Propul-

sion Branch and the Hyper-X Program Of ce at NASA Langley Re-search Center The author would like to thank E Ruf and A Auslen-der for many useful discussions of the work and R Irby J Connoverand J Simmons for their efforts during the test program

References1Hartill W B ldquoAnalytical and Experimental Investigation of a Scram-

jet Inlet of Quadriform Shaperdquo US Air Force Rept AFAPL-TR-65-74Marquardt Corp Aug 1965

2Billig F S ldquoSupersonic Combustion Ramjet Missilerdquo Journal ofPropulsion and Power Vol 11 No 6 1995 pp 1139ndash1146

3Kutshenreuter P H ldquoHypersonic Inlet Tests inHelium and AirrdquoGeneralElectric Co Advanced Engine and Technology Departments CincinnatiOH June 1965

4Molder S and Romeskie J M ldquoModular Hypersonic Inlets with Con-ical Flowrdquo CP-30 AGARD 1968

5Kantrowitz A and Donaldson C ldquoPreliminary Investigation of Super-sonic Diffusersrdquo NACA WR L-713 1945

6Smart M K ldquoDesign of Three-Dimensional Hypersonic Inlets withRectangular-to-Elliptical Shape Transitionrdquo Journal of Propulsion andPower Vol 15 No 3 1999 pp 408ndash416

7Guy R W Rogers R C Puster R L Rock K E and Diskin G SldquoThe NASA Langley Scramjet Test Complexrdquo AIAA Paper 96-3243 July1996

8Thomas S R Voland R T and Guy R W ldquoTest Flow CalibrationStudy of the Langley Arc- Heated Scramjet Test Facilityrdquo AIAA Paper 87-2165 June 1987

9Pinckney S Z ldquoImproved Static Probe Designrdquo AIAA Journal Vol 12No 4 1974 p 562

10Witte D W and Tatum K E ldquoComputer Code for Determination ofThermally Perfect Gas Propertiesrdquo NASA TP-3447 Sept 1994

11Molder S McGregor R J and Paisley T W ldquoA Comparison of ThreeHypersonic Inletsrdquo Sec 6 ldquoInvestigations in the Fluid Mechanics of Scram-jet Inletsrdquo US Air Force Rept WL-TR-93-2091 Ryerson Polytechnic InstToronto 1993

12Van Wie D M Kwok F T and Walsh R F ldquoStarting Characteristicsof Supersonic Inletsrdquo AIAA Paper 96-2914 July 1996

278 SMART

Fig 3 Three views of the REST inlet model con guration

smooth transition to an elliptical throat Its highly notched cowlwas designed to allow for ow spillage when operating below thedesign Mach number The particular inlet shown in Figs 2 and 3was designed to be mounted beneath a vehicle cruising at Mach71 on a constant q D 50 kPa (1044 psf) trajectory In combina-tion with a 6 deg forebody compression the inlet was requiredto supply a scramjet combustor with ow at or above 50 kPa(05 atm) In this instance ow enters the inlet at M1 D 60 andthe required inlet compression ratio is PR D 135 Further assump-tions used in the design of the model were an entrance Reynoldsnumber of Re1 D 26 pound 106=m a wall-to-total temperature ratio ofTw =Tt1 D 05 and a transition to turbulent ow that began 125 cm(05 in) downstream of all leading edges The resulting inlet modelhad an overall contraction ratio of 474 and an internal contractionratio of 215

The cruciform rake mounted at the rear of the inlet containedboth pitot and static pressure probes for surveying the inlet exit ow An 89 cm (35 in) diameter dump-isolator tube was connectedto the rear of the rake Fueled engine operation was simulated byreducing the exit areaof the dump isolator via translationof a conicalplug Throughout the test program this device was used to 1) obtainan estimate of the inlet mass capture and 2) examine the self-startcapability and back-pressured performance of the inlet

Instrumentation and Data Acquisition

Model instrumentation consisted of surface pressure taps pitotand static pressure probes coaxial surface thermocouples and to-tal temperature probes Model pressures were measured using aPressureSystems IncModel 8400 electronicallyscanning pressuresystem Four pressure ranges were utilized in the tests 0ndash68 kPa(0ndash1 psia) 0ndash344 kPa (0ndash5 psia) 0ndash1021 kPa (0ndash15 psia) and0ndash6807 kPa (0ndash100 psia) The error associated with the use ofthese transducers was sect05 of full scale The type E coaxial sur-face thermocouples were manufactured by Medtherm Corporationand were electrically isolated from the model The error associatedwith the use of these thermocouples was sect05 of the temperaturemeasured above the 273 K reference The type K thermocouplesused in the total temperature probes were of beaded constructionand the error associated with the use of the total temperature probeswas sect075 of the temperature measured above the 273 K ref-erence Thermocouple output was converted to temperature by auniversal temperature reference system and processed by a 16-bitNEFF AD converter Facilitydata were also processedby the NEFFAD converter All data recorded using the Autonet Version 40 dataacquisition software on a personal computer Pressure data associ-ated with the model recorded at a rate of 10 Hz and all thermocoupleand facility data recorded at a rate of 50 Hz All error bars shownin the paper were calculated using standard uncertainty analysistechniques and the aforementioned instrumentation error ranges

Figure 4 shows a schematic of the nominal streamlines alongwhich surface pressure taps were distributed in the inlet Pressuretaps were concentrated on the right-hand side of the model (as the

Fig 4 Schematic of theREST inlet pressure instru-mentation streamlines

Fig 5 Photograph of the cruciform rake containing both pitot andstatic probes

model had a vertical plane of symmetry) and were placed on the sixstreamlines (labeled A C E G I and K in Fig 4) at 15 differ-ent axial stations along the inlet These instrumentation streamlineswere approximately equally spaced around the internal surface ofthe inlet and gave a good indication of the overall surface pressuredistribution Thermocouples were placed at three representative po-sitions on the inside surface of the model that is the front of theinlet downstream of the inlet throat and on the cowl just down-stream of the inlet crotch (see Fig 3) These were used to monitorthe temperature of the inlet and also to obtain an estimate of the heattransfer to the model

A photograph of the cruciformrake is shown inFig 5 It contained24 pitot probes with an outer diameter of 10 mm and 12 staticprobes with an outer diameter of 15 mm The static probes wereequally spacedalong the horizontal and verticalbranches of the rakewhereas the pitot probes were concentratednear the walls tomeasurethe viscous losses in the inlet more accurately The static probeswere designed for internal supersonic ow measurements using themethod of Pinckney9 A full pitot and static pressure survey of theinlet exit ow required two runs with the cruciform rake rotated180 deg between runs

SMART 279

Fig 6 Schematic of the ow meter and the back-pressured ow eld

Figure 6 shows a schematic of the device used to measure thecaptured inlet mass- ow rate These measurements were obtainedby back pressuring the inletdump isolator with the conical pluguntil a shock train was stabilized near the throat of the inlet Flowwas then subsonic in the dump isolator and choked at its exit In thecurrent apparatus (referred to in the remainder of the paper as the ow meter) four pitot probes and two total temperature probes weresituated slightly upstream of the conical plug to obtain values forthe pitot pressure and total temperature of the ow exiting the owmeter Assuming a thermally perfect gas and a sonic Mach numberat the throat of the ow meter the captured inlet mass- ow rate isthen given by

Pm cap D pt p=pt p

deg =RTt Tt =T CD AFM (1)

where p=pt deg and T=Tt are all functions of Tt and are calculatedusing the methods described in Ref 10 The discharge coef cient forthe mass- ow meter CD D 0978 was obtained from a previouslyperformed calibration The error associated with the mass- ow ratecalculations was estimated at sect17

Test Program

The test program was separated into two sections 1) determina-tion of the inlet starting characteristics and back-pressured perfor-mance and 2) determination of the inlet ef ciency capability andmass capture The starting characteristics of the REST inlet wereassessedby reducing the exit areaof the ow meter until the inlet un-started then removing the ow constriction Restarting of the inletafter such a mechanically imposed unstart is generally considered tobe de nitive proof that the inlet will self-start at similar conditionsThe maximum back pressure tolerated by the inlet before unstart isalso an important design parameterbecause it provides an indicationof the amount of pressure rise that the inlet will withstand duringdual-mode scramjet engine operation

To de ne the conditions of the air entering the combustor of ascramjet engine both an inlet ef ciency term like total pressurerecovery as well as an inlet capability term like compression ratiomust be supplied11 The mass- ow rate of the air captured by theinlet is also required The ef ciency capability and mass capture ofthe REST inlet were determined using the pitot and static pressuresurveys at the inlet exit plane the surface pressure distributionswithin the inlet and the ow meter measurements Run times of35 s allowed the effect of different wall temperature conditions tobe investigated but the model was never allowed to reach adiabaticwall conditions due to the temperature limit of the epoxy used toseal the surfacepressure taps (500 K) Test sectionReynolds numberand dynamic pressure were varied as indicated in Table 1

Results and DiscussionInlet Starting Characteristics

If the inlet did not start then no useful data could be obtainedfrom the tests Hence inlet start-ability both with the initial tunnelstartup and after mechanically induced unstarts was critical to thetest program As discussed in Ref 12 and other places the keyparameter for characterizing inlet starting is the Mach number at itsplane of closure Based on computational uid dynamics analysesthe mass- ow averaged Mach number at the cowl closure planein the current experiments was Mcap D 468 The Kantrowitz limit5

indicates a maximum internal contraction of 153 at Mach 468 The

REST inlet was designed with an internal contraction ratio of 215so it violated the generally used rule of thumb for inlet starting bya signi cant margin Despite this the REST inlet was observed tostart with the wind tunnel for all test conditions Figure 7 shows aschlieren image of a typical test Flow is from left to right and theexternal shock structure generated by the inlet can be seen at thebottom of the image The appearance of a cowl shock emanatingfrom the crotch of the inlet was found to be a good indicator thatthe inlet was started

Figure 8 shows an instantaneous schlieren image of the externalshock structure observed during a mechanically induced unstartThis ow eld exhibited high-amplitude uctuations with a strongshock oscillating well upstream of the inlet crotch Figure 9 shows

Fig 7 Schlieren image of the started ow eld

Fig 8 Instantaneous schlieren image of the highly unsteady unstarted ow eld

Fig 9 Plot of the ow meter throat area and the key surface pressuresvs time during a self-start test

280 SMART

Table 2 Comparison of actual and designtest conditions

Property Actual Design

M1 618 60Re pound 106 120156 (366477) 260 (792)

miexcl1 (ftiexcl1)Tw=Tt1 044ndash071 050

a time-dependent plot of ow meter throat area AFM and two key sur-face pressure taps during a typical unstart test These taps were sit-uated just downstream of the inlet crotch one on the inner surfaceof the inlet and one on the outer surface The behavior of these twopressure measurements gave a clear indication of whether the inletwas started or not During a self-start test AFM was gradually re-duced until unstart was observed This ow process is indicated inFig 9 by the crossing over of the pi and po pressures traces Aftera period in which the inlet remained unstarted an increase of AFM

allowed the inlet to restart at an area slightly larger than the area atwhich the unstart began The REST inlet tested in the current exper-iments exhibited the ability to self-start after mechanically inducedunstarts for all test conditions

Surface and Probe Based Pressure Measurements

The REST inlet tests were initially planned to be undertaken inthe NASA Langley Research Center 20 in Mach 6 tunnel but thetests were transferred to the AHSTF due to scheduling dif cultiesFor this reason the test Mach number and Reynolds number wereslightly different from design Table 2 shows a comparison of theactual test conditions with those used to design the REST inletWhereas the wall thermal conditions of the tests were similar todesign both the increasedMachnumber and the decreasedReynoldsnumber produced a higher inlet compression ratio than the designvalue of PR D 135

Figure 10 shows the normalized surface pressure distributionsalong instrumentation streamlines A C E G and I (see Fig 4) fortest point 1 with Tw =Tt1 D 0438 together with the design valuesIn the portion of the inlet upstream of the crotch the experimentaldata was almost identical to the design values Downstream of thecrotch experimental data showed some differences in local pres-sure levels but the overall trend of the experimental and designpressure distributions remained similar The local discrepancies ob-served downstream of the crotch are believed to be due to featuresof the shockboundary-layer interactions not accounted for in thedesign process (see Ref 6) The results shown in Fig 10 are typicalof the surface pressure distributions observed throughout the testprogram It was generally observed that the inlet surface pressuredistributions were somewhat sensitive to the wall thermal conditionbut exhibited minimal sensitivityto the changes in Reynolds numberbetween the two test points

Both the pitot and static pressure distributions at the exit of theinlet were measured using the cruciform rake In combination withthe mean total temperature of the ow at the inlet exit these mea-surements were used to calculatea complete set of properties at eachprobe location assuming the air behaved as a thermally perfect gasTo determine the total temperature of the inlet exit ow an estimateof the average heat transfer rate to the model Pqm was obtained using

Pqm DdTm

dtC pNi frac12Ni vm (2)

As noted earlier the inside surface temperature was measured atthree representative locations on the model (see Fig 3) As themodel was a thin-walled structure (wall thickness frac1435 mm) thetemperature gradient of the entire structure dTm =dt was assumedto be equal to the average of the three representative surface tem-perature gradients Employing the speci c heat of nickel CpNi thedensity of nickel frac12Ni and the volume of the model vm use of Eq (2)resulted in typical heat transfer rates to the model of between 06and 15 of the captured total enthalpy ux Subtracting the heat

Fig 10 Surface pressure comparison between design and experimentalong instrumentation streamlines

transfer rate from the captured total enthalpy ux gave the total en-thalpy ux of the ow exiting the inlet The methods described inRef 10 were used to calculate the total temperature correspondingto the exit ow enthalpy ux

Figures 11a and 11b show the normalized pitot and staticpressuredistributions obtained along the horizontal and vertical branches ofthe rake for test point 2 with Tw=Tt1 D 0631 Figure 11a indicatesthat both the pe and pt2e distributions are nearly symmetric aboutthe centerline of the inlet Interestingly pe was observed to be ap-proximately constant across the horizontal branch whereas pt2e

showed plateaus near both walls smooth peaks approximately mid-way between the center and walls and a central dip The dip inpt2e at the center of the inlet was an unexpected feature of the ow especially given the relatively constant pe Probe interferenceand other possible problems associatedwith the central pitot probeswere investigated as causes of this phenomenon without result It is

SMART 281

a) Horizontal branch

b) Vertical branch

Fig 11 Pitot and static pressure probe measurements at the inlet exit

possible that the central dip in pt2e is a consequence of the nonuni-formities present in the inlet capture ow (Fig 1) because the in-let was designed with the assumption of uniform freestream owFigure 11b shows that pe remained constant on the lower two-thirdsof the vertical branch then jumped to a level approximately 50higher on the upper one-third The corresponding pt2e distributionpeaked toward the top of the inlet and exhibited a sharp dip in theviscous-dominated region near the top wall On the bottom portionof the vertical branch pt2e dropped from its peak at a more gradualrate due to increased shock losses near the cowl These distributionsare typical of those observed at the exit of the inlet throughout thetest program

Figures 12a and 12b show the Mach number and total pressuredistributions calculated from the rake measurements On the hor-izontal branch (Fig 12a) both Me and pte show a similar shapeto pt2e with the Mach number peaking at Me D 39 approximatelyhalfway between the center and the wall and dipping to Me D 35 atthe center In general the presence of the centraldip in pitot pressurehad a signi cant effect on all of the calculated inlet ow propertiesThe vertical branch distributions of Me and pte (Fig 12b) were alsosimilar to pt2e with the Mach number peaking at Me D 38 towardthe top of the inlet exit Note that the distribution of frac12eue (not shown)was also observed to show a similar trend to the distribution of pt2e

Inlet Ef ciency Capability and Mass Capture

Equivalent one-dimensional mass- ow-weighted properties havebeencalculatedat the inlet exit using the probe measurementsThesewere generatedby dividing the inlet exit area into four sections thenconstructing a series of elliptical-annular strips centered on each ofthe pitot probe positions on the horizontal and vertical branches ofthe rake The properties at the pitot probe positions were then as-sumed to correspond to the average values over the strip When inte-

Table 3 Equivalent one-dimensionalproperties at the inlet exit for test point 2a

Property Test point 2

pt MPa (psia) 1269 (1840)Tt K (plusmnR) 5436 (9784)Ht MJkg (Btulb) 0548 (2358)M 3264p kPa (psia) 1876 (2721)T K (plusmnR) 1813 (3263)frac12 kgm3 (lbft3) 0377 (235 pound 10iexcl2)u ms (fts) 8544 (2803)Pm kgs (lbs) 0732 (1615)Area cm2 (in2) 2437 (3778)

aTw=Tt1 D 0631

a) Horizontal branch

b) Vertical branch

Fig 12 Calculated Mach number and total pressure ratio at the inletexit

grated over the entire exit area of the inlet this averaging techniquewas expected to supply a best possible one-dimensional estimate ofthe inlet performance given the available data Table 3 includes acomplete list of the exit ow properties corresponding to the datashown in Figs 11 and 12

Numerous experiments were conducted at both test points withdifferent wall temperature conditions Figure 13 shows plots ofmass- ow-weighted equivalent one-dimensional ef ciency param-eters namely the total pressure recovery PT the kinetic energyef ciency acuteKE and the process ef ciency acuteKD vs wall-to-total tem-perature ratio Tw=Tt1 Note from Fig 13 that the ef ciency curvesfor both test points are almost identical where they overlap Thisindicated that the 30 difference in Reynolds number between testpoints 1 and 2 had a minimal effect on inlet ef ciency Also note

282 SMART

Fig 13 Plots of inlet ef ciency parameters vs wall-to-total tempera-ture ratio

Fig 14 Plots of inlet capability parameters vs wall-to-total tempera-ture ratio

from Fig 13 that Tw=Tt1 had a signi cant effect on PT For exam-ple increasing Tw=Tt1 from 044 to 071 reduced PT from 0565 to0517 This decrease in PT was due to the increasedboundary-layerthickness for the higher Tw =Tt1 Process ef ciency showed a lesspronounced decrease with increasing Tw=Tt1 with values in therange acuteKD D 0895ndash0877 for the tests In contrast to PT and acuteKDkinetic energy ef ciency showed a slight increase with increasingTw =Tt1 with values in the range acuteKE D 0963ndash0968 for the tests

The wall-to-total temperature ratio effect was also manifested inthe inlet capability as indicatedby parameterslike compression ratioPR temperature ratio TR and exit Mach number Me which are plot-ted vs Tw=Tt1 in Fig 14 The increased boundary-layer thicknessfor the higher Tw =Tt1 values raised the pressure and temperatureratio by increasing the aerodynamic contraction experienced by theinviscid core ow The same phenomenon also reduced the Machnumber at the inlet exit Increase of Tw=Tt1 from 044 to 071 pro-duced a 107 increase in PR to a maximum for the current tests ofPR D 1555 Both TR and Me showed a less pronounced dependenceon Tw=Tt1 with typical values of TR D 281 and Me D 329

Table 4 lists the ranges of the mass- ow-weighted equivalent one-dimensional values for PR TR Me PT acuteKE and acuteKD measured forthe REST inlet at the two test points The values listed in Table 4correspond to mass capture percentages of m c D 960 and 966 fortest points 1 and 2 respectively Both of these values are within theerror range of the corresponding mass- ow meter measurementswhich were mc D 962 sect 17 and 965 sect 17 for test points 1and 2 respectively

Table 4 Equivalent one-dimensional performanceparameter ranges for test points 1 and 2

Test point 1 Test point 2Property Tw=Tt1 D 044ndash059 Tw =Tt1 D 054ndash071

PR 1410ndash1518 1465ndash1555TR 272ndash284 279ndash290Me 335ndash328 332ndash322PT 0565ndash0544 0555ndash0512acuteKE 0963ndash0966 0965ndash0968acuteKD 0895ndash0887 0888ndash0877mc 962 sect 17 965 sect 17Pqm =Ht1 120ndash061 148ndash113

Fig 15 Surface pressure distributions at the maximum back-pressurecondition

Inlet Back Pressure Performance

Figure 15 shows a typical pressure distribution in the inlet fortest point 2 at maximum back pressure that is just before a me-chanically induced unstart This data supplies an indication of themaximum pressure the REST inlet could sustain in a combustor dur-ing dual-mode scramjet operation A plot of the average pressurealong the inlet without back pressure has been added to Fig 15 forreference Note that when fully back pressured the axial pressuredistributions along all streamlinesbecome similarThis is an indica-tion that ow downstream of the inlet crotch is fully separated witha shock train situated inside the enclosed portion of the inlet Basedon the data shown in Fig 15 the ratio of maximum back pressureto the pressure generated with no back pressure is 57 for this RESTinlet at Mach 62 ( note also that the maximum back pressure is ap-proximately 90 times the freestreampressure) Measured maximumback-pressure ratios varied between 53 and 57 throughout the testprogram

ConclusionsResults of Mach 62 wind-tunnel testing of a xed-geometry

hypersonic inlet were reported This three-dimensional quasi-streamtraced inlet had an overall contraction ratio of 474 an in-ternal contraction ratio of 215 and a Mach 60 design point Italso included a transition from a nearly rectangular capture to anelliptical throat The tests were conducted to validate a previouslydeveloped design methodology for xed-geometry inlet con gura-tions that include signi cant cross-sectional shape changes The in-let was tested at a Reynolds numbers of Re D 120 and 156 pound 106m(366 and 477 pound 106ft) and at wall-to-total temperature ratios rang-ing between 044 and 071

The inlet was observed to start with the wind tunnel at all timesduring the test program It also exhibited self-startingcapability af-ter mechanically induced inlet unstarts These results indicated thatthe Kantrowitz5 starting limit does not necessarily apply to all inlet

SMART 283

con gurations Surface pressure distributions within the inlet werefound to be similar to design However the overall compressionratio generated by the inlet was higher than design because the in-let was tested above the design Mach number and below the designReynolds number Equivalent one-dimensional performance param-eters were obtained using both pitot and static pressure probes atthe inlet exit On a mass- ow-weigthed basis capability parameterssuch as compression ratio and exit Mach number ranged between141 and 156 and 335 and 322 respectively Ef ciency param-eters such as total pressure recovery and kinetic energy ef ciencyranged between 0565 and 0512 and 0963 and 0968 respectivelyIn general increased values of wall-to-total temperature ratio led toincreased inlet capability but at a reduced total pressure recoveryInlet mass capture was deduced from measurements to be approx-imately 96 of the available air ow Heat transfer to the modelranged between 06 and 15 of the captured total enthalpy uxFinally back-pressurizationdata indicated that the maximum back-pressure ratio relative to unhindered operation varied between 53and 57 This test program demonstrated that the inlet shape tran-sition required to utilize elliptically shaped combustors in airframeintegrated scramjets can be accomplished without sacri cing inletperformance

AcknowledgmentsThis work was supported by the Hypersonic Airbreathing Propul-

sion Branch and the Hyper-X Program Of ce at NASA Langley Re-search Center The author would like to thank E Ruf and A Auslen-der for many useful discussions of the work and R Irby J Connoverand J Simmons for their efforts during the test program

References1Hartill W B ldquoAnalytical and Experimental Investigation of a Scram-

jet Inlet of Quadriform Shaperdquo US Air Force Rept AFAPL-TR-65-74Marquardt Corp Aug 1965

2Billig F S ldquoSupersonic Combustion Ramjet Missilerdquo Journal ofPropulsion and Power Vol 11 No 6 1995 pp 1139ndash1146

3Kutshenreuter P H ldquoHypersonic Inlet Tests inHelium and AirrdquoGeneralElectric Co Advanced Engine and Technology Departments CincinnatiOH June 1965

4Molder S and Romeskie J M ldquoModular Hypersonic Inlets with Con-ical Flowrdquo CP-30 AGARD 1968

5Kantrowitz A and Donaldson C ldquoPreliminary Investigation of Super-sonic Diffusersrdquo NACA WR L-713 1945

6Smart M K ldquoDesign of Three-Dimensional Hypersonic Inlets withRectangular-to-Elliptical Shape Transitionrdquo Journal of Propulsion andPower Vol 15 No 3 1999 pp 408ndash416

7Guy R W Rogers R C Puster R L Rock K E and Diskin G SldquoThe NASA Langley Scramjet Test Complexrdquo AIAA Paper 96-3243 July1996

8Thomas S R Voland R T and Guy R W ldquoTest Flow CalibrationStudy of the Langley Arc- Heated Scramjet Test Facilityrdquo AIAA Paper 87-2165 June 1987

9Pinckney S Z ldquoImproved Static Probe Designrdquo AIAA Journal Vol 12No 4 1974 p 562

10Witte D W and Tatum K E ldquoComputer Code for Determination ofThermally Perfect Gas Propertiesrdquo NASA TP-3447 Sept 1994

11Molder S McGregor R J and Paisley T W ldquoA Comparison of ThreeHypersonic Inletsrdquo Sec 6 ldquoInvestigations in the Fluid Mechanics of Scram-jet Inletsrdquo US Air Force Rept WL-TR-93-2091 Ryerson Polytechnic InstToronto 1993

12Van Wie D M Kwok F T and Walsh R F ldquoStarting Characteristicsof Supersonic Inletsrdquo AIAA Paper 96-2914 July 1996

SMART 279

Fig 6 Schematic of the ow meter and the back-pressured ow eld

Figure 6 shows a schematic of the device used to measure thecaptured inlet mass- ow rate These measurements were obtainedby back pressuring the inletdump isolator with the conical pluguntil a shock train was stabilized near the throat of the inlet Flowwas then subsonic in the dump isolator and choked at its exit In thecurrent apparatus (referred to in the remainder of the paper as the ow meter) four pitot probes and two total temperature probes weresituated slightly upstream of the conical plug to obtain values forthe pitot pressure and total temperature of the ow exiting the owmeter Assuming a thermally perfect gas and a sonic Mach numberat the throat of the ow meter the captured inlet mass- ow rate isthen given by

Pm cap D pt p=pt p

deg =RTt Tt =T CD AFM (1)

where p=pt deg and T=Tt are all functions of Tt and are calculatedusing the methods described in Ref 10 The discharge coef cient forthe mass- ow meter CD D 0978 was obtained from a previouslyperformed calibration The error associated with the mass- ow ratecalculations was estimated at sect17

Test Program

The test program was separated into two sections 1) determina-tion of the inlet starting characteristics and back-pressured perfor-mance and 2) determination of the inlet ef ciency capability andmass capture The starting characteristics of the REST inlet wereassessedby reducing the exit areaof the ow meter until the inlet un-started then removing the ow constriction Restarting of the inletafter such a mechanically imposed unstart is generally considered tobe de nitive proof that the inlet will self-start at similar conditionsThe maximum back pressure tolerated by the inlet before unstart isalso an important design parameterbecause it provides an indicationof the amount of pressure rise that the inlet will withstand duringdual-mode scramjet engine operation

To de ne the conditions of the air entering the combustor of ascramjet engine both an inlet ef ciency term like total pressurerecovery as well as an inlet capability term like compression ratiomust be supplied11 The mass- ow rate of the air captured by theinlet is also required The ef ciency capability and mass capture ofthe REST inlet were determined using the pitot and static pressuresurveys at the inlet exit plane the surface pressure distributionswithin the inlet and the ow meter measurements Run times of35 s allowed the effect of different wall temperature conditions tobe investigated but the model was never allowed to reach adiabaticwall conditions due to the temperature limit of the epoxy used toseal the surfacepressure taps (500 K) Test sectionReynolds numberand dynamic pressure were varied as indicated in Table 1

Results and DiscussionInlet Starting Characteristics

If the inlet did not start then no useful data could be obtainedfrom the tests Hence inlet start-ability both with the initial tunnelstartup and after mechanically induced unstarts was critical to thetest program As discussed in Ref 12 and other places the keyparameter for characterizing inlet starting is the Mach number at itsplane of closure Based on computational uid dynamics analysesthe mass- ow averaged Mach number at the cowl closure planein the current experiments was Mcap D 468 The Kantrowitz limit5

indicates a maximum internal contraction of 153 at Mach 468 The

REST inlet was designed with an internal contraction ratio of 215so it violated the generally used rule of thumb for inlet starting bya signi cant margin Despite this the REST inlet was observed tostart with the wind tunnel for all test conditions Figure 7 shows aschlieren image of a typical test Flow is from left to right and theexternal shock structure generated by the inlet can be seen at thebottom of the image The appearance of a cowl shock emanatingfrom the crotch of the inlet was found to be a good indicator thatthe inlet was started

Figure 8 shows an instantaneous schlieren image of the externalshock structure observed during a mechanically induced unstartThis ow eld exhibited high-amplitude uctuations with a strongshock oscillating well upstream of the inlet crotch Figure 9 shows

Fig 7 Schlieren image of the started ow eld

Fig 8 Instantaneous schlieren image of the highly unsteady unstarted ow eld

Fig 9 Plot of the ow meter throat area and the key surface pressuresvs time during a self-start test

280 SMART

Table 2 Comparison of actual and designtest conditions

Property Actual Design

M1 618 60Re pound 106 120156 (366477) 260 (792)

miexcl1 (ftiexcl1)Tw=Tt1 044ndash071 050

a time-dependent plot of ow meter throat area AFM and two key sur-face pressure taps during a typical unstart test These taps were sit-uated just downstream of the inlet crotch one on the inner surfaceof the inlet and one on the outer surface The behavior of these twopressure measurements gave a clear indication of whether the inletwas started or not During a self-start test AFM was gradually re-duced until unstart was observed This ow process is indicated inFig 9 by the crossing over of the pi and po pressures traces Aftera period in which the inlet remained unstarted an increase of AFM

allowed the inlet to restart at an area slightly larger than the area atwhich the unstart began The REST inlet tested in the current exper-iments exhibited the ability to self-start after mechanically inducedunstarts for all test conditions

Surface and Probe Based Pressure Measurements

The REST inlet tests were initially planned to be undertaken inthe NASA Langley Research Center 20 in Mach 6 tunnel but thetests were transferred to the AHSTF due to scheduling dif cultiesFor this reason the test Mach number and Reynolds number wereslightly different from design Table 2 shows a comparison of theactual test conditions with those used to design the REST inletWhereas the wall thermal conditions of the tests were similar todesign both the increasedMachnumber and the decreasedReynoldsnumber produced a higher inlet compression ratio than the designvalue of PR D 135

Figure 10 shows the normalized surface pressure distributionsalong instrumentation streamlines A C E G and I (see Fig 4) fortest point 1 with Tw =Tt1 D 0438 together with the design valuesIn the portion of the inlet upstream of the crotch the experimentaldata was almost identical to the design values Downstream of thecrotch experimental data showed some differences in local pres-sure levels but the overall trend of the experimental and designpressure distributions remained similar The local discrepancies ob-served downstream of the crotch are believed to be due to featuresof the shockboundary-layer interactions not accounted for in thedesign process (see Ref 6) The results shown in Fig 10 are typicalof the surface pressure distributions observed throughout the testprogram It was generally observed that the inlet surface pressuredistributions were somewhat sensitive to the wall thermal conditionbut exhibited minimal sensitivityto the changes in Reynolds numberbetween the two test points

Both the pitot and static pressure distributions at the exit of theinlet were measured using the cruciform rake In combination withthe mean total temperature of the ow at the inlet exit these mea-surements were used to calculatea complete set of properties at eachprobe location assuming the air behaved as a thermally perfect gasTo determine the total temperature of the inlet exit ow an estimateof the average heat transfer rate to the model Pqm was obtained using

Pqm DdTm

dtC pNi frac12Ni vm (2)

As noted earlier the inside surface temperature was measured atthree representative locations on the model (see Fig 3) As themodel was a thin-walled structure (wall thickness frac1435 mm) thetemperature gradient of the entire structure dTm =dt was assumedto be equal to the average of the three representative surface tem-perature gradients Employing the speci c heat of nickel CpNi thedensity of nickel frac12Ni and the volume of the model vm use of Eq (2)resulted in typical heat transfer rates to the model of between 06and 15 of the captured total enthalpy ux Subtracting the heat

Fig 10 Surface pressure comparison between design and experimentalong instrumentation streamlines

transfer rate from the captured total enthalpy ux gave the total en-thalpy ux of the ow exiting the inlet The methods described inRef 10 were used to calculate the total temperature correspondingto the exit ow enthalpy ux

Figures 11a and 11b show the normalized pitot and staticpressuredistributions obtained along the horizontal and vertical branches ofthe rake for test point 2 with Tw=Tt1 D 0631 Figure 11a indicatesthat both the pe and pt2e distributions are nearly symmetric aboutthe centerline of the inlet Interestingly pe was observed to be ap-proximately constant across the horizontal branch whereas pt2e

showed plateaus near both walls smooth peaks approximately mid-way between the center and walls and a central dip The dip inpt2e at the center of the inlet was an unexpected feature of the ow especially given the relatively constant pe Probe interferenceand other possible problems associatedwith the central pitot probeswere investigated as causes of this phenomenon without result It is

SMART 281

a) Horizontal branch

b) Vertical branch

Fig 11 Pitot and static pressure probe measurements at the inlet exit

possible that the central dip in pt2e is a consequence of the nonuni-formities present in the inlet capture ow (Fig 1) because the in-let was designed with the assumption of uniform freestream owFigure 11b shows that pe remained constant on the lower two-thirdsof the vertical branch then jumped to a level approximately 50higher on the upper one-third The corresponding pt2e distributionpeaked toward the top of the inlet and exhibited a sharp dip in theviscous-dominated region near the top wall On the bottom portionof the vertical branch pt2e dropped from its peak at a more gradualrate due to increased shock losses near the cowl These distributionsare typical of those observed at the exit of the inlet throughout thetest program

Figures 12a and 12b show the Mach number and total pressuredistributions calculated from the rake measurements On the hor-izontal branch (Fig 12a) both Me and pte show a similar shapeto pt2e with the Mach number peaking at Me D 39 approximatelyhalfway between the center and the wall and dipping to Me D 35 atthe center In general the presence of the centraldip in pitot pressurehad a signi cant effect on all of the calculated inlet ow propertiesThe vertical branch distributions of Me and pte (Fig 12b) were alsosimilar to pt2e with the Mach number peaking at Me D 38 towardthe top of the inlet exit Note that the distribution of frac12eue (not shown)was also observed to show a similar trend to the distribution of pt2e

Inlet Ef ciency Capability and Mass Capture

Equivalent one-dimensional mass- ow-weighted properties havebeencalculatedat the inlet exit using the probe measurementsThesewere generatedby dividing the inlet exit area into four sections thenconstructing a series of elliptical-annular strips centered on each ofthe pitot probe positions on the horizontal and vertical branches ofthe rake The properties at the pitot probe positions were then as-sumed to correspond to the average values over the strip When inte-

Table 3 Equivalent one-dimensionalproperties at the inlet exit for test point 2a

Property Test point 2

pt MPa (psia) 1269 (1840)Tt K (plusmnR) 5436 (9784)Ht MJkg (Btulb) 0548 (2358)M 3264p kPa (psia) 1876 (2721)T K (plusmnR) 1813 (3263)frac12 kgm3 (lbft3) 0377 (235 pound 10iexcl2)u ms (fts) 8544 (2803)Pm kgs (lbs) 0732 (1615)Area cm2 (in2) 2437 (3778)

aTw=Tt1 D 0631

a) Horizontal branch

b) Vertical branch

Fig 12 Calculated Mach number and total pressure ratio at the inletexit

grated over the entire exit area of the inlet this averaging techniquewas expected to supply a best possible one-dimensional estimate ofthe inlet performance given the available data Table 3 includes acomplete list of the exit ow properties corresponding to the datashown in Figs 11 and 12

Numerous experiments were conducted at both test points withdifferent wall temperature conditions Figure 13 shows plots ofmass- ow-weighted equivalent one-dimensional ef ciency param-eters namely the total pressure recovery PT the kinetic energyef ciency acuteKE and the process ef ciency acuteKD vs wall-to-total tem-perature ratio Tw=Tt1 Note from Fig 13 that the ef ciency curvesfor both test points are almost identical where they overlap Thisindicated that the 30 difference in Reynolds number between testpoints 1 and 2 had a minimal effect on inlet ef ciency Also note

282 SMART

Fig 13 Plots of inlet ef ciency parameters vs wall-to-total tempera-ture ratio

Fig 14 Plots of inlet capability parameters vs wall-to-total tempera-ture ratio

from Fig 13 that Tw=Tt1 had a signi cant effect on PT For exam-ple increasing Tw=Tt1 from 044 to 071 reduced PT from 0565 to0517 This decrease in PT was due to the increasedboundary-layerthickness for the higher Tw =Tt1 Process ef ciency showed a lesspronounced decrease with increasing Tw=Tt1 with values in therange acuteKD D 0895ndash0877 for the tests In contrast to PT and acuteKDkinetic energy ef ciency showed a slight increase with increasingTw =Tt1 with values in the range acuteKE D 0963ndash0968 for the tests

The wall-to-total temperature ratio effect was also manifested inthe inlet capability as indicatedby parameterslike compression ratioPR temperature ratio TR and exit Mach number Me which are plot-ted vs Tw=Tt1 in Fig 14 The increased boundary-layer thicknessfor the higher Tw =Tt1 values raised the pressure and temperatureratio by increasing the aerodynamic contraction experienced by theinviscid core ow The same phenomenon also reduced the Machnumber at the inlet exit Increase of Tw=Tt1 from 044 to 071 pro-duced a 107 increase in PR to a maximum for the current tests ofPR D 1555 Both TR and Me showed a less pronounced dependenceon Tw=Tt1 with typical values of TR D 281 and Me D 329

Table 4 lists the ranges of the mass- ow-weighted equivalent one-dimensional values for PR TR Me PT acuteKE and acuteKD measured forthe REST inlet at the two test points The values listed in Table 4correspond to mass capture percentages of m c D 960 and 966 fortest points 1 and 2 respectively Both of these values are within theerror range of the corresponding mass- ow meter measurementswhich were mc D 962 sect 17 and 965 sect 17 for test points 1and 2 respectively

Table 4 Equivalent one-dimensional performanceparameter ranges for test points 1 and 2

Test point 1 Test point 2Property Tw=Tt1 D 044ndash059 Tw =Tt1 D 054ndash071

PR 1410ndash1518 1465ndash1555TR 272ndash284 279ndash290Me 335ndash328 332ndash322PT 0565ndash0544 0555ndash0512acuteKE 0963ndash0966 0965ndash0968acuteKD 0895ndash0887 0888ndash0877mc 962 sect 17 965 sect 17Pqm =Ht1 120ndash061 148ndash113

Fig 15 Surface pressure distributions at the maximum back-pressurecondition

Inlet Back Pressure Performance

Figure 15 shows a typical pressure distribution in the inlet fortest point 2 at maximum back pressure that is just before a me-chanically induced unstart This data supplies an indication of themaximum pressure the REST inlet could sustain in a combustor dur-ing dual-mode scramjet operation A plot of the average pressurealong the inlet without back pressure has been added to Fig 15 forreference Note that when fully back pressured the axial pressuredistributions along all streamlinesbecome similarThis is an indica-tion that ow downstream of the inlet crotch is fully separated witha shock train situated inside the enclosed portion of the inlet Basedon the data shown in Fig 15 the ratio of maximum back pressureto the pressure generated with no back pressure is 57 for this RESTinlet at Mach 62 ( note also that the maximum back pressure is ap-proximately 90 times the freestreampressure) Measured maximumback-pressure ratios varied between 53 and 57 throughout the testprogram

ConclusionsResults of Mach 62 wind-tunnel testing of a xed-geometry

hypersonic inlet were reported This three-dimensional quasi-streamtraced inlet had an overall contraction ratio of 474 an in-ternal contraction ratio of 215 and a Mach 60 design point Italso included a transition from a nearly rectangular capture to anelliptical throat The tests were conducted to validate a previouslydeveloped design methodology for xed-geometry inlet con gura-tions that include signi cant cross-sectional shape changes The in-let was tested at a Reynolds numbers of Re D 120 and 156 pound 106m(366 and 477 pound 106ft) and at wall-to-total temperature ratios rang-ing between 044 and 071

The inlet was observed to start with the wind tunnel at all timesduring the test program It also exhibited self-startingcapability af-ter mechanically induced inlet unstarts These results indicated thatthe Kantrowitz5 starting limit does not necessarily apply to all inlet

SMART 283

con gurations Surface pressure distributions within the inlet werefound to be similar to design However the overall compressionratio generated by the inlet was higher than design because the in-let was tested above the design Mach number and below the designReynolds number Equivalent one-dimensional performance param-eters were obtained using both pitot and static pressure probes atthe inlet exit On a mass- ow-weigthed basis capability parameterssuch as compression ratio and exit Mach number ranged between141 and 156 and 335 and 322 respectively Ef ciency param-eters such as total pressure recovery and kinetic energy ef ciencyranged between 0565 and 0512 and 0963 and 0968 respectivelyIn general increased values of wall-to-total temperature ratio led toincreased inlet capability but at a reduced total pressure recoveryInlet mass capture was deduced from measurements to be approx-imately 96 of the available air ow Heat transfer to the modelranged between 06 and 15 of the captured total enthalpy uxFinally back-pressurizationdata indicated that the maximum back-pressure ratio relative to unhindered operation varied between 53and 57 This test program demonstrated that the inlet shape tran-sition required to utilize elliptically shaped combustors in airframeintegrated scramjets can be accomplished without sacri cing inletperformance

AcknowledgmentsThis work was supported by the Hypersonic Airbreathing Propul-

sion Branch and the Hyper-X Program Of ce at NASA Langley Re-search Center The author would like to thank E Ruf and A Auslen-der for many useful discussions of the work and R Irby J Connoverand J Simmons for their efforts during the test program

References1Hartill W B ldquoAnalytical and Experimental Investigation of a Scram-

jet Inlet of Quadriform Shaperdquo US Air Force Rept AFAPL-TR-65-74Marquardt Corp Aug 1965

2Billig F S ldquoSupersonic Combustion Ramjet Missilerdquo Journal ofPropulsion and Power Vol 11 No 6 1995 pp 1139ndash1146

3Kutshenreuter P H ldquoHypersonic Inlet Tests inHelium and AirrdquoGeneralElectric Co Advanced Engine and Technology Departments CincinnatiOH June 1965

4Molder S and Romeskie J M ldquoModular Hypersonic Inlets with Con-ical Flowrdquo CP-30 AGARD 1968

5Kantrowitz A and Donaldson C ldquoPreliminary Investigation of Super-sonic Diffusersrdquo NACA WR L-713 1945

6Smart M K ldquoDesign of Three-Dimensional Hypersonic Inlets withRectangular-to-Elliptical Shape Transitionrdquo Journal of Propulsion andPower Vol 15 No 3 1999 pp 408ndash416

7Guy R W Rogers R C Puster R L Rock K E and Diskin G SldquoThe NASA Langley Scramjet Test Complexrdquo AIAA Paper 96-3243 July1996

8Thomas S R Voland R T and Guy R W ldquoTest Flow CalibrationStudy of the Langley Arc- Heated Scramjet Test Facilityrdquo AIAA Paper 87-2165 June 1987

9Pinckney S Z ldquoImproved Static Probe Designrdquo AIAA Journal Vol 12No 4 1974 p 562

10Witte D W and Tatum K E ldquoComputer Code for Determination ofThermally Perfect Gas Propertiesrdquo NASA TP-3447 Sept 1994

11Molder S McGregor R J and Paisley T W ldquoA Comparison of ThreeHypersonic Inletsrdquo Sec 6 ldquoInvestigations in the Fluid Mechanics of Scram-jet Inletsrdquo US Air Force Rept WL-TR-93-2091 Ryerson Polytechnic InstToronto 1993

12Van Wie D M Kwok F T and Walsh R F ldquoStarting Characteristicsof Supersonic Inletsrdquo AIAA Paper 96-2914 July 1996

280 SMART

Table 2 Comparison of actual and designtest conditions

Property Actual Design

M1 618 60Re pound 106 120156 (366477) 260 (792)

miexcl1 (ftiexcl1)Tw=Tt1 044ndash071 050

a time-dependent plot of ow meter throat area AFM and two key sur-face pressure taps during a typical unstart test These taps were sit-uated just downstream of the inlet crotch one on the inner surfaceof the inlet and one on the outer surface The behavior of these twopressure measurements gave a clear indication of whether the inletwas started or not During a self-start test AFM was gradually re-duced until unstart was observed This ow process is indicated inFig 9 by the crossing over of the pi and po pressures traces Aftera period in which the inlet remained unstarted an increase of AFM

allowed the inlet to restart at an area slightly larger than the area atwhich the unstart began The REST inlet tested in the current exper-iments exhibited the ability to self-start after mechanically inducedunstarts for all test conditions

Surface and Probe Based Pressure Measurements

The REST inlet tests were initially planned to be undertaken inthe NASA Langley Research Center 20 in Mach 6 tunnel but thetests were transferred to the AHSTF due to scheduling dif cultiesFor this reason the test Mach number and Reynolds number wereslightly different from design Table 2 shows a comparison of theactual test conditions with those used to design the REST inletWhereas the wall thermal conditions of the tests were similar todesign both the increasedMachnumber and the decreasedReynoldsnumber produced a higher inlet compression ratio than the designvalue of PR D 135

Figure 10 shows the normalized surface pressure distributionsalong instrumentation streamlines A C E G and I (see Fig 4) fortest point 1 with Tw =Tt1 D 0438 together with the design valuesIn the portion of the inlet upstream of the crotch the experimentaldata was almost identical to the design values Downstream of thecrotch experimental data showed some differences in local pres-sure levels but the overall trend of the experimental and designpressure distributions remained similar The local discrepancies ob-served downstream of the crotch are believed to be due to featuresof the shockboundary-layer interactions not accounted for in thedesign process (see Ref 6) The results shown in Fig 10 are typicalof the surface pressure distributions observed throughout the testprogram It was generally observed that the inlet surface pressuredistributions were somewhat sensitive to the wall thermal conditionbut exhibited minimal sensitivityto the changes in Reynolds numberbetween the two test points

Both the pitot and static pressure distributions at the exit of theinlet were measured using the cruciform rake In combination withthe mean total temperature of the ow at the inlet exit these mea-surements were used to calculatea complete set of properties at eachprobe location assuming the air behaved as a thermally perfect gasTo determine the total temperature of the inlet exit ow an estimateof the average heat transfer rate to the model Pqm was obtained using

Pqm DdTm

dtC pNi frac12Ni vm (2)

As noted earlier the inside surface temperature was measured atthree representative locations on the model (see Fig 3) As themodel was a thin-walled structure (wall thickness frac1435 mm) thetemperature gradient of the entire structure dTm =dt was assumedto be equal to the average of the three representative surface tem-perature gradients Employing the speci c heat of nickel CpNi thedensity of nickel frac12Ni and the volume of the model vm use of Eq (2)resulted in typical heat transfer rates to the model of between 06and 15 of the captured total enthalpy ux Subtracting the heat

Fig 10 Surface pressure comparison between design and experimentalong instrumentation streamlines

transfer rate from the captured total enthalpy ux gave the total en-thalpy ux of the ow exiting the inlet The methods described inRef 10 were used to calculate the total temperature correspondingto the exit ow enthalpy ux

Figures 11a and 11b show the normalized pitot and staticpressuredistributions obtained along the horizontal and vertical branches ofthe rake for test point 2 with Tw=Tt1 D 0631 Figure 11a indicatesthat both the pe and pt2e distributions are nearly symmetric aboutthe centerline of the inlet Interestingly pe was observed to be ap-proximately constant across the horizontal branch whereas pt2e

showed plateaus near both walls smooth peaks approximately mid-way between the center and walls and a central dip The dip inpt2e at the center of the inlet was an unexpected feature of the ow especially given the relatively constant pe Probe interferenceand other possible problems associatedwith the central pitot probeswere investigated as causes of this phenomenon without result It is

SMART 281

a) Horizontal branch

b) Vertical branch

Fig 11 Pitot and static pressure probe measurements at the inlet exit

possible that the central dip in pt2e is a consequence of the nonuni-formities present in the inlet capture ow (Fig 1) because the in-let was designed with the assumption of uniform freestream owFigure 11b shows that pe remained constant on the lower two-thirdsof the vertical branch then jumped to a level approximately 50higher on the upper one-third The corresponding pt2e distributionpeaked toward the top of the inlet and exhibited a sharp dip in theviscous-dominated region near the top wall On the bottom portionof the vertical branch pt2e dropped from its peak at a more gradualrate due to increased shock losses near the cowl These distributionsare typical of those observed at the exit of the inlet throughout thetest program

Figures 12a and 12b show the Mach number and total pressuredistributions calculated from the rake measurements On the hor-izontal branch (Fig 12a) both Me and pte show a similar shapeto pt2e with the Mach number peaking at Me D 39 approximatelyhalfway between the center and the wall and dipping to Me D 35 atthe center In general the presence of the centraldip in pitot pressurehad a signi cant effect on all of the calculated inlet ow propertiesThe vertical branch distributions of Me and pte (Fig 12b) were alsosimilar to pt2e with the Mach number peaking at Me D 38 towardthe top of the inlet exit Note that the distribution of frac12eue (not shown)was also observed to show a similar trend to the distribution of pt2e

Inlet Ef ciency Capability and Mass Capture

Equivalent one-dimensional mass- ow-weighted properties havebeencalculatedat the inlet exit using the probe measurementsThesewere generatedby dividing the inlet exit area into four sections thenconstructing a series of elliptical-annular strips centered on each ofthe pitot probe positions on the horizontal and vertical branches ofthe rake The properties at the pitot probe positions were then as-sumed to correspond to the average values over the strip When inte-

Table 3 Equivalent one-dimensionalproperties at the inlet exit for test point 2a

Property Test point 2

pt MPa (psia) 1269 (1840)Tt K (plusmnR) 5436 (9784)Ht MJkg (Btulb) 0548 (2358)M 3264p kPa (psia) 1876 (2721)T K (plusmnR) 1813 (3263)frac12 kgm3 (lbft3) 0377 (235 pound 10iexcl2)u ms (fts) 8544 (2803)Pm kgs (lbs) 0732 (1615)Area cm2 (in2) 2437 (3778)

aTw=Tt1 D 0631

a) Horizontal branch

b) Vertical branch

Fig 12 Calculated Mach number and total pressure ratio at the inletexit

grated over the entire exit area of the inlet this averaging techniquewas expected to supply a best possible one-dimensional estimate ofthe inlet performance given the available data Table 3 includes acomplete list of the exit ow properties corresponding to the datashown in Figs 11 and 12

Numerous experiments were conducted at both test points withdifferent wall temperature conditions Figure 13 shows plots ofmass- ow-weighted equivalent one-dimensional ef ciency param-eters namely the total pressure recovery PT the kinetic energyef ciency acuteKE and the process ef ciency acuteKD vs wall-to-total tem-perature ratio Tw=Tt1 Note from Fig 13 that the ef ciency curvesfor both test points are almost identical where they overlap Thisindicated that the 30 difference in Reynolds number between testpoints 1 and 2 had a minimal effect on inlet ef ciency Also note

282 SMART

Fig 13 Plots of inlet ef ciency parameters vs wall-to-total tempera-ture ratio

Fig 14 Plots of inlet capability parameters vs wall-to-total tempera-ture ratio

from Fig 13 that Tw=Tt1 had a signi cant effect on PT For exam-ple increasing Tw=Tt1 from 044 to 071 reduced PT from 0565 to0517 This decrease in PT was due to the increasedboundary-layerthickness for the higher Tw =Tt1 Process ef ciency showed a lesspronounced decrease with increasing Tw=Tt1 with values in therange acuteKD D 0895ndash0877 for the tests In contrast to PT and acuteKDkinetic energy ef ciency showed a slight increase with increasingTw =Tt1 with values in the range acuteKE D 0963ndash0968 for the tests

The wall-to-total temperature ratio effect was also manifested inthe inlet capability as indicatedby parameterslike compression ratioPR temperature ratio TR and exit Mach number Me which are plot-ted vs Tw=Tt1 in Fig 14 The increased boundary-layer thicknessfor the higher Tw =Tt1 values raised the pressure and temperatureratio by increasing the aerodynamic contraction experienced by theinviscid core ow The same phenomenon also reduced the Machnumber at the inlet exit Increase of Tw=Tt1 from 044 to 071 pro-duced a 107 increase in PR to a maximum for the current tests ofPR D 1555 Both TR and Me showed a less pronounced dependenceon Tw=Tt1 with typical values of TR D 281 and Me D 329

Table 4 lists the ranges of the mass- ow-weighted equivalent one-dimensional values for PR TR Me PT acuteKE and acuteKD measured forthe REST inlet at the two test points The values listed in Table 4correspond to mass capture percentages of m c D 960 and 966 fortest points 1 and 2 respectively Both of these values are within theerror range of the corresponding mass- ow meter measurementswhich were mc D 962 sect 17 and 965 sect 17 for test points 1and 2 respectively

Table 4 Equivalent one-dimensional performanceparameter ranges for test points 1 and 2

Test point 1 Test point 2Property Tw=Tt1 D 044ndash059 Tw =Tt1 D 054ndash071

PR 1410ndash1518 1465ndash1555TR 272ndash284 279ndash290Me 335ndash328 332ndash322PT 0565ndash0544 0555ndash0512acuteKE 0963ndash0966 0965ndash0968acuteKD 0895ndash0887 0888ndash0877mc 962 sect 17 965 sect 17Pqm =Ht1 120ndash061 148ndash113

Fig 15 Surface pressure distributions at the maximum back-pressurecondition

Inlet Back Pressure Performance

Figure 15 shows a typical pressure distribution in the inlet fortest point 2 at maximum back pressure that is just before a me-chanically induced unstart This data supplies an indication of themaximum pressure the REST inlet could sustain in a combustor dur-ing dual-mode scramjet operation A plot of the average pressurealong the inlet without back pressure has been added to Fig 15 forreference Note that when fully back pressured the axial pressuredistributions along all streamlinesbecome similarThis is an indica-tion that ow downstream of the inlet crotch is fully separated witha shock train situated inside the enclosed portion of the inlet Basedon the data shown in Fig 15 the ratio of maximum back pressureto the pressure generated with no back pressure is 57 for this RESTinlet at Mach 62 ( note also that the maximum back pressure is ap-proximately 90 times the freestreampressure) Measured maximumback-pressure ratios varied between 53 and 57 throughout the testprogram

ConclusionsResults of Mach 62 wind-tunnel testing of a xed-geometry

hypersonic inlet were reported This three-dimensional quasi-streamtraced inlet had an overall contraction ratio of 474 an in-ternal contraction ratio of 215 and a Mach 60 design point Italso included a transition from a nearly rectangular capture to anelliptical throat The tests were conducted to validate a previouslydeveloped design methodology for xed-geometry inlet con gura-tions that include signi cant cross-sectional shape changes The in-let was tested at a Reynolds numbers of Re D 120 and 156 pound 106m(366 and 477 pound 106ft) and at wall-to-total temperature ratios rang-ing between 044 and 071

The inlet was observed to start with the wind tunnel at all timesduring the test program It also exhibited self-startingcapability af-ter mechanically induced inlet unstarts These results indicated thatthe Kantrowitz5 starting limit does not necessarily apply to all inlet

SMART 283

con gurations Surface pressure distributions within the inlet werefound to be similar to design However the overall compressionratio generated by the inlet was higher than design because the in-let was tested above the design Mach number and below the designReynolds number Equivalent one-dimensional performance param-eters were obtained using both pitot and static pressure probes atthe inlet exit On a mass- ow-weigthed basis capability parameterssuch as compression ratio and exit Mach number ranged between141 and 156 and 335 and 322 respectively Ef ciency param-eters such as total pressure recovery and kinetic energy ef ciencyranged between 0565 and 0512 and 0963 and 0968 respectivelyIn general increased values of wall-to-total temperature ratio led toincreased inlet capability but at a reduced total pressure recoveryInlet mass capture was deduced from measurements to be approx-imately 96 of the available air ow Heat transfer to the modelranged between 06 and 15 of the captured total enthalpy uxFinally back-pressurizationdata indicated that the maximum back-pressure ratio relative to unhindered operation varied between 53and 57 This test program demonstrated that the inlet shape tran-sition required to utilize elliptically shaped combustors in airframeintegrated scramjets can be accomplished without sacri cing inletperformance

AcknowledgmentsThis work was supported by the Hypersonic Airbreathing Propul-

sion Branch and the Hyper-X Program Of ce at NASA Langley Re-search Center The author would like to thank E Ruf and A Auslen-der for many useful discussions of the work and R Irby J Connoverand J Simmons for their efforts during the test program

References1Hartill W B ldquoAnalytical and Experimental Investigation of a Scram-

jet Inlet of Quadriform Shaperdquo US Air Force Rept AFAPL-TR-65-74Marquardt Corp Aug 1965

2Billig F S ldquoSupersonic Combustion Ramjet Missilerdquo Journal ofPropulsion and Power Vol 11 No 6 1995 pp 1139ndash1146

3Kutshenreuter P H ldquoHypersonic Inlet Tests inHelium and AirrdquoGeneralElectric Co Advanced Engine and Technology Departments CincinnatiOH June 1965

4Molder S and Romeskie J M ldquoModular Hypersonic Inlets with Con-ical Flowrdquo CP-30 AGARD 1968

5Kantrowitz A and Donaldson C ldquoPreliminary Investigation of Super-sonic Diffusersrdquo NACA WR L-713 1945

6Smart M K ldquoDesign of Three-Dimensional Hypersonic Inlets withRectangular-to-Elliptical Shape Transitionrdquo Journal of Propulsion andPower Vol 15 No 3 1999 pp 408ndash416

7Guy R W Rogers R C Puster R L Rock K E and Diskin G SldquoThe NASA Langley Scramjet Test Complexrdquo AIAA Paper 96-3243 July1996

8Thomas S R Voland R T and Guy R W ldquoTest Flow CalibrationStudy of the Langley Arc- Heated Scramjet Test Facilityrdquo AIAA Paper 87-2165 June 1987

9Pinckney S Z ldquoImproved Static Probe Designrdquo AIAA Journal Vol 12No 4 1974 p 562

10Witte D W and Tatum K E ldquoComputer Code for Determination ofThermally Perfect Gas Propertiesrdquo NASA TP-3447 Sept 1994

11Molder S McGregor R J and Paisley T W ldquoA Comparison of ThreeHypersonic Inletsrdquo Sec 6 ldquoInvestigations in the Fluid Mechanics of Scram-jet Inletsrdquo US Air Force Rept WL-TR-93-2091 Ryerson Polytechnic InstToronto 1993

12Van Wie D M Kwok F T and Walsh R F ldquoStarting Characteristicsof Supersonic Inletsrdquo AIAA Paper 96-2914 July 1996

SMART 281

a) Horizontal branch

b) Vertical branch

Fig 11 Pitot and static pressure probe measurements at the inlet exit

possible that the central dip in pt2e is a consequence of the nonuni-formities present in the inlet capture ow (Fig 1) because the in-let was designed with the assumption of uniform freestream owFigure 11b shows that pe remained constant on the lower two-thirdsof the vertical branch then jumped to a level approximately 50higher on the upper one-third The corresponding pt2e distributionpeaked toward the top of the inlet and exhibited a sharp dip in theviscous-dominated region near the top wall On the bottom portionof the vertical branch pt2e dropped from its peak at a more gradualrate due to increased shock losses near the cowl These distributionsare typical of those observed at the exit of the inlet throughout thetest program

Figures 12a and 12b show the Mach number and total pressuredistributions calculated from the rake measurements On the hor-izontal branch (Fig 12a) both Me and pte show a similar shapeto pt2e with the Mach number peaking at Me D 39 approximatelyhalfway between the center and the wall and dipping to Me D 35 atthe center In general the presence of the centraldip in pitot pressurehad a signi cant effect on all of the calculated inlet ow propertiesThe vertical branch distributions of Me and pte (Fig 12b) were alsosimilar to pt2e with the Mach number peaking at Me D 38 towardthe top of the inlet exit Note that the distribution of frac12eue (not shown)was also observed to show a similar trend to the distribution of pt2e

Inlet Ef ciency Capability and Mass Capture

Equivalent one-dimensional mass- ow-weighted properties havebeencalculatedat the inlet exit using the probe measurementsThesewere generatedby dividing the inlet exit area into four sections thenconstructing a series of elliptical-annular strips centered on each ofthe pitot probe positions on the horizontal and vertical branches ofthe rake The properties at the pitot probe positions were then as-sumed to correspond to the average values over the strip When inte-

Table 3 Equivalent one-dimensionalproperties at the inlet exit for test point 2a

Property Test point 2

pt MPa (psia) 1269 (1840)Tt K (plusmnR) 5436 (9784)Ht MJkg (Btulb) 0548 (2358)M 3264p kPa (psia) 1876 (2721)T K (plusmnR) 1813 (3263)frac12 kgm3 (lbft3) 0377 (235 pound 10iexcl2)u ms (fts) 8544 (2803)Pm kgs (lbs) 0732 (1615)Area cm2 (in2) 2437 (3778)

aTw=Tt1 D 0631

a) Horizontal branch

b) Vertical branch

Fig 12 Calculated Mach number and total pressure ratio at the inletexit

grated over the entire exit area of the inlet this averaging techniquewas expected to supply a best possible one-dimensional estimate ofthe inlet performance given the available data Table 3 includes acomplete list of the exit ow properties corresponding to the datashown in Figs 11 and 12

Numerous experiments were conducted at both test points withdifferent wall temperature conditions Figure 13 shows plots ofmass- ow-weighted equivalent one-dimensional ef ciency param-eters namely the total pressure recovery PT the kinetic energyef ciency acuteKE and the process ef ciency acuteKD vs wall-to-total tem-perature ratio Tw=Tt1 Note from Fig 13 that the ef ciency curvesfor both test points are almost identical where they overlap Thisindicated that the 30 difference in Reynolds number between testpoints 1 and 2 had a minimal effect on inlet ef ciency Also note

282 SMART

Fig 13 Plots of inlet ef ciency parameters vs wall-to-total tempera-ture ratio

Fig 14 Plots of inlet capability parameters vs wall-to-total tempera-ture ratio

from Fig 13 that Tw=Tt1 had a signi cant effect on PT For exam-ple increasing Tw=Tt1 from 044 to 071 reduced PT from 0565 to0517 This decrease in PT was due to the increasedboundary-layerthickness for the higher Tw =Tt1 Process ef ciency showed a lesspronounced decrease with increasing Tw=Tt1 with values in therange acuteKD D 0895ndash0877 for the tests In contrast to PT and acuteKDkinetic energy ef ciency showed a slight increase with increasingTw =Tt1 with values in the range acuteKE D 0963ndash0968 for the tests

The wall-to-total temperature ratio effect was also manifested inthe inlet capability as indicatedby parameterslike compression ratioPR temperature ratio TR and exit Mach number Me which are plot-ted vs Tw=Tt1 in Fig 14 The increased boundary-layer thicknessfor the higher Tw =Tt1 values raised the pressure and temperatureratio by increasing the aerodynamic contraction experienced by theinviscid core ow The same phenomenon also reduced the Machnumber at the inlet exit Increase of Tw=Tt1 from 044 to 071 pro-duced a 107 increase in PR to a maximum for the current tests ofPR D 1555 Both TR and Me showed a less pronounced dependenceon Tw=Tt1 with typical values of TR D 281 and Me D 329

Table 4 lists the ranges of the mass- ow-weighted equivalent one-dimensional values for PR TR Me PT acuteKE and acuteKD measured forthe REST inlet at the two test points The values listed in Table 4correspond to mass capture percentages of m c D 960 and 966 fortest points 1 and 2 respectively Both of these values are within theerror range of the corresponding mass- ow meter measurementswhich were mc D 962 sect 17 and 965 sect 17 for test points 1and 2 respectively

Table 4 Equivalent one-dimensional performanceparameter ranges for test points 1 and 2

Test point 1 Test point 2Property Tw=Tt1 D 044ndash059 Tw =Tt1 D 054ndash071

PR 1410ndash1518 1465ndash1555TR 272ndash284 279ndash290Me 335ndash328 332ndash322PT 0565ndash0544 0555ndash0512acuteKE 0963ndash0966 0965ndash0968acuteKD 0895ndash0887 0888ndash0877mc 962 sect 17 965 sect 17Pqm =Ht1 120ndash061 148ndash113

Fig 15 Surface pressure distributions at the maximum back-pressurecondition

Inlet Back Pressure Performance

Figure 15 shows a typical pressure distribution in the inlet fortest point 2 at maximum back pressure that is just before a me-chanically induced unstart This data supplies an indication of themaximum pressure the REST inlet could sustain in a combustor dur-ing dual-mode scramjet operation A plot of the average pressurealong the inlet without back pressure has been added to Fig 15 forreference Note that when fully back pressured the axial pressuredistributions along all streamlinesbecome similarThis is an indica-tion that ow downstream of the inlet crotch is fully separated witha shock train situated inside the enclosed portion of the inlet Basedon the data shown in Fig 15 the ratio of maximum back pressureto the pressure generated with no back pressure is 57 for this RESTinlet at Mach 62 ( note also that the maximum back pressure is ap-proximately 90 times the freestreampressure) Measured maximumback-pressure ratios varied between 53 and 57 throughout the testprogram

ConclusionsResults of Mach 62 wind-tunnel testing of a xed-geometry

hypersonic inlet were reported This three-dimensional quasi-streamtraced inlet had an overall contraction ratio of 474 an in-ternal contraction ratio of 215 and a Mach 60 design point Italso included a transition from a nearly rectangular capture to anelliptical throat The tests were conducted to validate a previouslydeveloped design methodology for xed-geometry inlet con gura-tions that include signi cant cross-sectional shape changes The in-let was tested at a Reynolds numbers of Re D 120 and 156 pound 106m(366 and 477 pound 106ft) and at wall-to-total temperature ratios rang-ing between 044 and 071

The inlet was observed to start with the wind tunnel at all timesduring the test program It also exhibited self-startingcapability af-ter mechanically induced inlet unstarts These results indicated thatthe Kantrowitz5 starting limit does not necessarily apply to all inlet

SMART 283

con gurations Surface pressure distributions within the inlet werefound to be similar to design However the overall compressionratio generated by the inlet was higher than design because the in-let was tested above the design Mach number and below the designReynolds number Equivalent one-dimensional performance param-eters were obtained using both pitot and static pressure probes atthe inlet exit On a mass- ow-weigthed basis capability parameterssuch as compression ratio and exit Mach number ranged between141 and 156 and 335 and 322 respectively Ef ciency param-eters such as total pressure recovery and kinetic energy ef ciencyranged between 0565 and 0512 and 0963 and 0968 respectivelyIn general increased values of wall-to-total temperature ratio led toincreased inlet capability but at a reduced total pressure recoveryInlet mass capture was deduced from measurements to be approx-imately 96 of the available air ow Heat transfer to the modelranged between 06 and 15 of the captured total enthalpy uxFinally back-pressurizationdata indicated that the maximum back-pressure ratio relative to unhindered operation varied between 53and 57 This test program demonstrated that the inlet shape tran-sition required to utilize elliptically shaped combustors in airframeintegrated scramjets can be accomplished without sacri cing inletperformance

AcknowledgmentsThis work was supported by the Hypersonic Airbreathing Propul-

sion Branch and the Hyper-X Program Of ce at NASA Langley Re-search Center The author would like to thank E Ruf and A Auslen-der for many useful discussions of the work and R Irby J Connoverand J Simmons for their efforts during the test program

References1Hartill W B ldquoAnalytical and Experimental Investigation of a Scram-

jet Inlet of Quadriform Shaperdquo US Air Force Rept AFAPL-TR-65-74Marquardt Corp Aug 1965

2Billig F S ldquoSupersonic Combustion Ramjet Missilerdquo Journal ofPropulsion and Power Vol 11 No 6 1995 pp 1139ndash1146

3Kutshenreuter P H ldquoHypersonic Inlet Tests inHelium and AirrdquoGeneralElectric Co Advanced Engine and Technology Departments CincinnatiOH June 1965

4Molder S and Romeskie J M ldquoModular Hypersonic Inlets with Con-ical Flowrdquo CP-30 AGARD 1968

5Kantrowitz A and Donaldson C ldquoPreliminary Investigation of Super-sonic Diffusersrdquo NACA WR L-713 1945

6Smart M K ldquoDesign of Three-Dimensional Hypersonic Inlets withRectangular-to-Elliptical Shape Transitionrdquo Journal of Propulsion andPower Vol 15 No 3 1999 pp 408ndash416

7Guy R W Rogers R C Puster R L Rock K E and Diskin G SldquoThe NASA Langley Scramjet Test Complexrdquo AIAA Paper 96-3243 July1996

8Thomas S R Voland R T and Guy R W ldquoTest Flow CalibrationStudy of the Langley Arc- Heated Scramjet Test Facilityrdquo AIAA Paper 87-2165 June 1987

9Pinckney S Z ldquoImproved Static Probe Designrdquo AIAA Journal Vol 12No 4 1974 p 562

10Witte D W and Tatum K E ldquoComputer Code for Determination ofThermally Perfect Gas Propertiesrdquo NASA TP-3447 Sept 1994

11Molder S McGregor R J and Paisley T W ldquoA Comparison of ThreeHypersonic Inletsrdquo Sec 6 ldquoInvestigations in the Fluid Mechanics of Scram-jet Inletsrdquo US Air Force Rept WL-TR-93-2091 Ryerson Polytechnic InstToronto 1993

12Van Wie D M Kwok F T and Walsh R F ldquoStarting Characteristicsof Supersonic Inletsrdquo AIAA Paper 96-2914 July 1996

282 SMART

Fig 13 Plots of inlet ef ciency parameters vs wall-to-total tempera-ture ratio

Fig 14 Plots of inlet capability parameters vs wall-to-total tempera-ture ratio

from Fig 13 that Tw=Tt1 had a signi cant effect on PT For exam-ple increasing Tw=Tt1 from 044 to 071 reduced PT from 0565 to0517 This decrease in PT was due to the increasedboundary-layerthickness for the higher Tw =Tt1 Process ef ciency showed a lesspronounced decrease with increasing Tw=Tt1 with values in therange acuteKD D 0895ndash0877 for the tests In contrast to PT and acuteKDkinetic energy ef ciency showed a slight increase with increasingTw =Tt1 with values in the range acuteKE D 0963ndash0968 for the tests

The wall-to-total temperature ratio effect was also manifested inthe inlet capability as indicatedby parameterslike compression ratioPR temperature ratio TR and exit Mach number Me which are plot-ted vs Tw=Tt1 in Fig 14 The increased boundary-layer thicknessfor the higher Tw =Tt1 values raised the pressure and temperatureratio by increasing the aerodynamic contraction experienced by theinviscid core ow The same phenomenon also reduced the Machnumber at the inlet exit Increase of Tw=Tt1 from 044 to 071 pro-duced a 107 increase in PR to a maximum for the current tests ofPR D 1555 Both TR and Me showed a less pronounced dependenceon Tw=Tt1 with typical values of TR D 281 and Me D 329

Table 4 lists the ranges of the mass- ow-weighted equivalent one-dimensional values for PR TR Me PT acuteKE and acuteKD measured forthe REST inlet at the two test points The values listed in Table 4correspond to mass capture percentages of m c D 960 and 966 fortest points 1 and 2 respectively Both of these values are within theerror range of the corresponding mass- ow meter measurementswhich were mc D 962 sect 17 and 965 sect 17 for test points 1and 2 respectively

Table 4 Equivalent one-dimensional performanceparameter ranges for test points 1 and 2

Test point 1 Test point 2Property Tw=Tt1 D 044ndash059 Tw =Tt1 D 054ndash071

PR 1410ndash1518 1465ndash1555TR 272ndash284 279ndash290Me 335ndash328 332ndash322PT 0565ndash0544 0555ndash0512acuteKE 0963ndash0966 0965ndash0968acuteKD 0895ndash0887 0888ndash0877mc 962 sect 17 965 sect 17Pqm =Ht1 120ndash061 148ndash113

Fig 15 Surface pressure distributions at the maximum back-pressurecondition

Inlet Back Pressure Performance

Figure 15 shows a typical pressure distribution in the inlet fortest point 2 at maximum back pressure that is just before a me-chanically induced unstart This data supplies an indication of themaximum pressure the REST inlet could sustain in a combustor dur-ing dual-mode scramjet operation A plot of the average pressurealong the inlet without back pressure has been added to Fig 15 forreference Note that when fully back pressured the axial pressuredistributions along all streamlinesbecome similarThis is an indica-tion that ow downstream of the inlet crotch is fully separated witha shock train situated inside the enclosed portion of the inlet Basedon the data shown in Fig 15 the ratio of maximum back pressureto the pressure generated with no back pressure is 57 for this RESTinlet at Mach 62 ( note also that the maximum back pressure is ap-proximately 90 times the freestreampressure) Measured maximumback-pressure ratios varied between 53 and 57 throughout the testprogram

ConclusionsResults of Mach 62 wind-tunnel testing of a xed-geometry

hypersonic inlet were reported This three-dimensional quasi-streamtraced inlet had an overall contraction ratio of 474 an in-ternal contraction ratio of 215 and a Mach 60 design point Italso included a transition from a nearly rectangular capture to anelliptical throat The tests were conducted to validate a previouslydeveloped design methodology for xed-geometry inlet con gura-tions that include signi cant cross-sectional shape changes The in-let was tested at a Reynolds numbers of Re D 120 and 156 pound 106m(366 and 477 pound 106ft) and at wall-to-total temperature ratios rang-ing between 044 and 071

The inlet was observed to start with the wind tunnel at all timesduring the test program It also exhibited self-startingcapability af-ter mechanically induced inlet unstarts These results indicated thatthe Kantrowitz5 starting limit does not necessarily apply to all inlet

SMART 283

con gurations Surface pressure distributions within the inlet werefound to be similar to design However the overall compressionratio generated by the inlet was higher than design because the in-let was tested above the design Mach number and below the designReynolds number Equivalent one-dimensional performance param-eters were obtained using both pitot and static pressure probes atthe inlet exit On a mass- ow-weigthed basis capability parameterssuch as compression ratio and exit Mach number ranged between141 and 156 and 335 and 322 respectively Ef ciency param-eters such as total pressure recovery and kinetic energy ef ciencyranged between 0565 and 0512 and 0963 and 0968 respectivelyIn general increased values of wall-to-total temperature ratio led toincreased inlet capability but at a reduced total pressure recoveryInlet mass capture was deduced from measurements to be approx-imately 96 of the available air ow Heat transfer to the modelranged between 06 and 15 of the captured total enthalpy uxFinally back-pressurizationdata indicated that the maximum back-pressure ratio relative to unhindered operation varied between 53and 57 This test program demonstrated that the inlet shape tran-sition required to utilize elliptically shaped combustors in airframeintegrated scramjets can be accomplished without sacri cing inletperformance

AcknowledgmentsThis work was supported by the Hypersonic Airbreathing Propul-

sion Branch and the Hyper-X Program Of ce at NASA Langley Re-search Center The author would like to thank E Ruf and A Auslen-der for many useful discussions of the work and R Irby J Connoverand J Simmons for their efforts during the test program

References1Hartill W B ldquoAnalytical and Experimental Investigation of a Scram-

jet Inlet of Quadriform Shaperdquo US Air Force Rept AFAPL-TR-65-74Marquardt Corp Aug 1965

2Billig F S ldquoSupersonic Combustion Ramjet Missilerdquo Journal ofPropulsion and Power Vol 11 No 6 1995 pp 1139ndash1146

3Kutshenreuter P H ldquoHypersonic Inlet Tests inHelium and AirrdquoGeneralElectric Co Advanced Engine and Technology Departments CincinnatiOH June 1965

4Molder S and Romeskie J M ldquoModular Hypersonic Inlets with Con-ical Flowrdquo CP-30 AGARD 1968

5Kantrowitz A and Donaldson C ldquoPreliminary Investigation of Super-sonic Diffusersrdquo NACA WR L-713 1945

6Smart M K ldquoDesign of Three-Dimensional Hypersonic Inlets withRectangular-to-Elliptical Shape Transitionrdquo Journal of Propulsion andPower Vol 15 No 3 1999 pp 408ndash416

7Guy R W Rogers R C Puster R L Rock K E and Diskin G SldquoThe NASA Langley Scramjet Test Complexrdquo AIAA Paper 96-3243 July1996

8Thomas S R Voland R T and Guy R W ldquoTest Flow CalibrationStudy of the Langley Arc- Heated Scramjet Test Facilityrdquo AIAA Paper 87-2165 June 1987

9Pinckney S Z ldquoImproved Static Probe Designrdquo AIAA Journal Vol 12No 4 1974 p 562

10Witte D W and Tatum K E ldquoComputer Code for Determination ofThermally Perfect Gas Propertiesrdquo NASA TP-3447 Sept 1994

11Molder S McGregor R J and Paisley T W ldquoA Comparison of ThreeHypersonic Inletsrdquo Sec 6 ldquoInvestigations in the Fluid Mechanics of Scram-jet Inletsrdquo US Air Force Rept WL-TR-93-2091 Ryerson Polytechnic InstToronto 1993

12Van Wie D M Kwok F T and Walsh R F ldquoStarting Characteristicsof Supersonic Inletsrdquo AIAA Paper 96-2914 July 1996

SMART 283

con gurations Surface pressure distributions within the inlet werefound to be similar to design However the overall compressionratio generated by the inlet was higher than design because the in-let was tested above the design Mach number and below the designReynolds number Equivalent one-dimensional performance param-eters were obtained using both pitot and static pressure probes atthe inlet exit On a mass- ow-weigthed basis capability parameterssuch as compression ratio and exit Mach number ranged between141 and 156 and 335 and 322 respectively Ef ciency param-eters such as total pressure recovery and kinetic energy ef ciencyranged between 0565 and 0512 and 0963 and 0968 respectivelyIn general increased values of wall-to-total temperature ratio led toincreased inlet capability but at a reduced total pressure recoveryInlet mass capture was deduced from measurements to be approx-imately 96 of the available air ow Heat transfer to the modelranged between 06 and 15 of the captured total enthalpy uxFinally back-pressurizationdata indicated that the maximum back-pressure ratio relative to unhindered operation varied between 53and 57 This test program demonstrated that the inlet shape tran-sition required to utilize elliptically shaped combustors in airframeintegrated scramjets can be accomplished without sacri cing inletperformance

AcknowledgmentsThis work was supported by the Hypersonic Airbreathing Propul-

sion Branch and the Hyper-X Program Of ce at NASA Langley Re-search Center The author would like to thank E Ruf and A Auslen-der for many useful discussions of the work and R Irby J Connoverand J Simmons for their efforts during the test program

References1Hartill W B ldquoAnalytical and Experimental Investigation of a Scram-

jet Inlet of Quadriform Shaperdquo US Air Force Rept AFAPL-TR-65-74Marquardt Corp Aug 1965

2Billig F S ldquoSupersonic Combustion Ramjet Missilerdquo Journal ofPropulsion and Power Vol 11 No 6 1995 pp 1139ndash1146

3Kutshenreuter P H ldquoHypersonic Inlet Tests inHelium and AirrdquoGeneralElectric Co Advanced Engine and Technology Departments CincinnatiOH June 1965

4Molder S and Romeskie J M ldquoModular Hypersonic Inlets with Con-ical Flowrdquo CP-30 AGARD 1968

5Kantrowitz A and Donaldson C ldquoPreliminary Investigation of Super-sonic Diffusersrdquo NACA WR L-713 1945

6Smart M K ldquoDesign of Three-Dimensional Hypersonic Inlets withRectangular-to-Elliptical Shape Transitionrdquo Journal of Propulsion andPower Vol 15 No 3 1999 pp 408ndash416

7Guy R W Rogers R C Puster R L Rock K E and Diskin G SldquoThe NASA Langley Scramjet Test Complexrdquo AIAA Paper 96-3243 July1996

8Thomas S R Voland R T and Guy R W ldquoTest Flow CalibrationStudy of the Langley Arc- Heated Scramjet Test Facilityrdquo AIAA Paper 87-2165 June 1987

9Pinckney S Z ldquoImproved Static Probe Designrdquo AIAA Journal Vol 12No 4 1974 p 562

10Witte D W and Tatum K E ldquoComputer Code for Determination ofThermally Perfect Gas Propertiesrdquo NASA TP-3447 Sept 1994

11Molder S McGregor R J and Paisley T W ldquoA Comparison of ThreeHypersonic Inletsrdquo Sec 6 ldquoInvestigations in the Fluid Mechanics of Scram-jet Inletsrdquo US Air Force Rept WL-TR-93-2091 Ryerson Polytechnic InstToronto 1993

12Van Wie D M Kwok F T and Walsh R F ldquoStarting Characteristicsof Supersonic Inletsrdquo AIAA Paper 96-2914 July 1996