AEDC TR-94-6 AD-A284 057LU l111111111 tElHll 1I!

Hypersonic Wind Tunnel Test Techniques

R. K. Matthews and R. W. RhudyCaispan Corporation/AEDC Operations

August 1994

Final Report for Period July 1992 - May 1993

DTICS ELECTE

SEP 02 1994 Dv G

FApprvedfor puW blimees; disriulon is ulmtd

___ QUALITY INSPECTED 5

0o) ARNOLD ENGINEERING DEVELOPMENT CENTERARNOLD AIR FORCE BASE, TENNESSEE0)

AIR FORCE MATERIEL COMMANDUNITED STATES AIR FORCE"94 9 01 234

NOInczs

When U. S. Government drawings, specifications, or other data are used for any purposeother than a defite related Government procurement operation, the Govermneit therebyincurs no responsibility nor any obligation whatsoever, and the fact that the Governmentmay have formulated, furnished, or in any way supplied the said drawings, specifications,or other data, is not to be regarded by implication or otherwise, or in any manner licensingthe holder or any other person or corporation, or conveying any rights or permission tomanufacture, use, or sell any patented invention that may in any way be related thereto.

Qualified users may obtain copies of this report from the Defense Technical InformationCenter.

References to named commercial products in this report are not to be considered in anyseine as an endorsement of the product by the United States Air Force or the Government.

This report has been reviewed by the Office of Public Affairs (PA) and is releasable tothe National Technical Information Service (NTIS). At NTIS, it will be available to thegeneral public, including foreign nations.

APPROVAL STATEMENT

This report has been reviewed and approved.

DENNIS N. HUPRICH, Major, USAFSpace and Missile Systems Test Division

Approved for publication:

FOR THE COMMANDER

CONRAD M. RITCHEY, Lt Col, USAFSpace and Missile Systems Test Division

Public reporting burden for this collection o1 information is es~timated tO average 1 hour per resoonse, including the time for reviewing instrfucions, fsearching eml$stilg data sources.

gathering and maint aining the data needed. and Ompl~eting and reviewing the collection of information Send comments regjarding this burden estimate Or any other aspe.ct: of this

collection of infor mation, including suggestions for reducing this burden. 10 Washington Headquarters Services. Directorate for Information Op~eratiOns and Reports. 121IS Jefferson

Davis, Hhhwa. Suite 1204t Arlin ton. VA 22202-430•r and to the Office of Man Kement and Iud• et Papermorki Reduction Prolect 10704-0188r Washin tOn.OC 20503

1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVEREDAuc ust 1994 Final -- July 1992 -May 1993

4. TITLE AND SUBTITLE 5. FUNDING NUMBERS

Hypersonic Wind Tunnel Test Techniques JN - 0979

6. AUTHOR(S)

Matthews, R. K. and Rhudy, R. W., Calspan Corporation/AEDCOperations ____________

7 PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATIONREPORT NUMBER

Arnold Engineering Development Center/DOF A D T-4Air Force Materiel Command AD -R9-Arnold Air Force Base, TN 37389-4000

9. SPONSORING/MONITORING AGENCY NAMES(S) AND ADDRESS(ES) 10. SPONSORING/'MONITORING

AGENCY REPORT NUMBER

Arnold Engineering Development Center/DOFAir Force Materiel CommandArnold Air Force Base, TN 37389-4000

11. SUPPLEMENTARY NOTES

Available in Defense Technical Information Center (DTIC).

1 2a. DISTRIBUTION/AVAILABILITY STATEMENT 1 :•b. DISTRIBUTION CODE

Approved for public release; distribution is unlimited.

13. ABSTRACT (Maxim urn 00 words)This report describes the procedures used in the continuous flow hypersonic tunnels ofthe AEDC for static stability, pressure, heat transfer, materials/structures, boundary-layertransition, and electromagnetic wave testing. Particular emphasis is placed on heat-transfer techniques because of the importance of defining the thermal environment ofhypersonic vehicles. An overview of the materials/structures test methodology used in thedevelopment of hypersonic vehicle components is presented. Unfortunately, themethodology to predict transition has eluded the aerodynamicist for over three decades,and there are still many unanswered questions. This report briefly touches on the manyparameters that affect transition and provides numerous references for those who areinterested in specializing in this topic. The methodology of using trip spheres is discussed,and illustrative data are presented. Electromagnetic wave testing represents a relativelynew test technique that involves the union of several d isciplines: aerotherrnodynamics,electromagnetics, materials/structures, and advanced diagnostics. The essence of this newtechnique deals with the transmission and possible distortion of electromagnetic waves(RF or ;R) as they pass through the bow shock, flow field, and electromagnetic (EM)window of a missile flying at hypersonic speeds. _________

14. SUBJECT TERMS 1 S, NUMBER OF PAGESelectromagnetic waves, missile seeker system, hypersonic vehicles, 59boundary layer, boresight error, radomes 16. PRICE CODE

17. SECURITY CLASSIFICATION' 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT

OF REPORT OF THIS PAGE OF ABSTRACTSA EASRP TU NCLASSIFIED UNCLASSIFIED UNCLASSIFIED SM SRPR

COMPUTER GENEaur[D Standard Form 298 (Rev, 2-89)Prescribed by ANSi Std Z39- I8298-102

AEDC-FR-94-6

FOREWORD

The hypersonic regime is the most severe of all flight regimes, and consequently demands smart utili-zation of ground testing and evaluation, flight testing, ano computation/simulation methodologies. Becauseof this challenge, von Karman Institute (VKI) asked the Arnold Engineering Development Center (AEDC)to develop a comprehensive course to define the "Methodology of Hypersonic Testing." Seven Americanscientists and engineers, representing AEDC and the University of Tennessee Space Institute (UTSI), for-mulated this course from their background of over a century of combined experience in hypersonic testing.

The objective of the course v i -nt a comprehensive overview of the methods used in hyper-sonic testing and evaluation, and Ah o,,! he principles behind those test techniques. Topics coveredinclude an introduction to hypersonic aer-Aynamics with descriptions of chemical and gas-dynamic phe-nomena associated with hypersonic fligN., -at,-ories and application of various hypersonic ground testfacilities; characterization of facility flow fields; measurement techniques (both intrusive and non-intru-sive); hypersonic propulsion test principles and facilities; computational techniques and their integrationinto test programs; ground-test-to-flight data correlation methods; and test program planning. The LectureSeries begins at the introductory level and progressively increases in depth, culminating in a focus on spe-cial test and evaluation issues in hypersonics such as boundary-layer transitie-i, shock interactions, electro-magnetic wave testing, and propulsion integration test techniqaes.

To obtain a complete set of notes from this course write to:

Lecture Series Secretaryvon Karman InstituteCharrissie de Waterloo, 72B-16409 Rhode-Saint-Genese (Belgium)

The information contained in this report is a subset of the work described above.

Accesion For

NTIS CRA&IDTIC TABUnannounced 5Justification ........................

By .....

Distribution I

Availability Codes

i Avail and I orDist I Special

S... .. .,.,•,= • nmmm lmm m~l~lllI ill I I ln mll =1

AEDC-TR-4-6

CONTENTS

bag

Test and Evaluation M ethods ........................................................... 5

M aterials/Structures Testing ............................................................ 29

Boundary-Layer Transition ............................................................ 41

Electromagnetic W ave Test ............................................................ 51

S. .. . .. . . . .ra m .i i i i i3

AEDC-TR-94-6

TEST AND EVALUATION METHODS

byR. K. MATTHEWS and R. W. RHUDY

Senior Staff EngineersCalspan Corporation/AEDC OperationsArnold Engineering Development Center

ABSTRACT E Heat gage output, mv

Test facility selection is generally the first step FA Axial force

toward planning a wind tunnel test, either through FN Normal forceavailability, simulation, or test technique require- Fy Side Forcements. The operating characteristics of the chosenfacility will immediately restrict the choice of test h Heat-transfer coefficient

techniques because of the tunnel operating mode. k Material conductivitySince the time response of the measurement technique

must be compatible with the run time of the facility. M Mach numbersome techniques are immediately ruled out for certain MI. 2. or x Balance moment

facilities. The aerodynamic and aerothermal test tech-niques used in conventional hypersonic tunnels are P Pressuregenerally similar throughout the world, varying only q Dynamics pressure

in the sophistication of the facilitys' instrumentationand data acquisition systems. This section describes q Heat flux

the procedures used in the continuous flow hyper- Re Reynolds number

sonic tunnels of the AEDC for static stability,pressure, and heat transfer testing. Particular t Time

emphasis will be on heat transfer techniques because T Temperature

of the importance of defining the thermal environ-ment of hypersonic vehicles and because the static S Reference area

stability and pressure techniques are very similar to S/R Surface distance-to-nose radius ratio

those used in subsonic and supersonic facilities.W Model weight

NOMENCLATURE x, y, z Coordinates of model CG in balance axissystem

Ao Intercept of q versus Tw for heat gage

data a Angle of attack

At Slope of h versus Tw for heat gage data / Angle of sideslip

b Thin-skin wall thickness Q Density

c Specific heat AT Temperature difference

C, Static stability coefficients: e.g. lift CL, 0 Roll angle

drag CD, pitching moment Cm Subscripts

Cp Pressure coefficient F Flight

CSF Heat gage calibration factor i Initial (time = 0)

C (tn) Coax gage calibration factor calculated at L Model lengthtime tn

pc Phase change

5

AEDC-T4"

T Tunnel dictate many of the pretest requirements/activities,r Recovery such as model scale, test techniques, etc. The

following sections cover the relatively standard static

w Wall stability test model, as shown in Fig. 2.

0 Free stream

0 Stilling chamber conditions

INTRODUCTION .= :

One of the primary considerations in k.

evaluating hypersonic facilities is the test CALCUTE: FORCE - • q,, ,techniques (or test methods) available in a given NOTES: 10 (ITqS

facility. A thorough understanding of test N - FEHICLF - FU610techniques is very important in planning a test T - WIND TUNNEL _ .program to address a precisely defined testobjective. This section describes the aero- PAT ~8"1dynamic and aerothermal test techniques thatare in general use. The aerodynamic metho-

dology includes force balance measurements FCOTEEand surface pressure measurements. Flow-field CALCU(Ar: T- q-_and other aerodynamic techniques are discussed W Nin later sections. Aerothermal methodology WIND TUNNEL MOELincludes thermal mapping, discrete gage Figure 1. Wind tunnel/flight simulation.

techniques and gage calibration. In general,each specific test technique is described in terms NOE- i, ý, Ad i AE ALWAYS IN

of (a) principle of operation, (b) apparatus, (c) TlHE IRIAKE AXIS SYSTEM,

data reduction and d) illustrative data. Lem FOIAID MOMENT NEVER IN THE MODEL AXIS.REFEIENE( POINT

STATIC STABILITY TESTING MODEL EC-v AXIS

I RALAE AXISStatic stability tests in a hypersonic wind tunnel -

are conducted in much the same manner and for thesame reasons as in lower-speedwind tunnels. Typicaltests are conducted to; verify that the performance :AXIS

of a particular design is accurate (parametric studies); JOO(L XIS 4Wverify theoretical codes (CFD); and/or prove that LEFT SIDE VIEW REAU VIEWproposed modifications to existing flight hardwarewill, in fact, improve the performance. In general, Figure 2. Model weight (tare) and center-of-gravity

compared to low speed facilities the tunnels and the locations referenced to balance axis.test models are smaller and the test cxwironmentmuch more severe. These smaller models and the The procedures described are for tests inextreme environment (i.e., low static pressure and

extremely high total pressure and temperature lead conventional, relatively long duration wind tunnels.to special requirements not encountered in lower Impulse type, short-duration testing requires other

speed wind tunnels. It is much cheaper and safer to special procedures to compensate for such things as

build small-scale vehicles and test in the wind tunnel the inertia forces from model vibrations.

("Test before flight'7) than to build the real thing andhave it fail ("Build it and see if it works"). The Apparatus

classical wind tunnel to flight correlation parametersare shown in Fig. I. Once the need for a static in general, static stability data in a hypersonic

stability wind tunnel test is determined and the test wind tunnel are obtained by use of a strain gage

objectives clearly established, the choice of what balance, usually mounted internal to the test model.

facility to use can be made. This choice naturally will Special circumstances may dictate deviations such as

6

AEDC-TR-944

extremely small models which require that the OUTPUTS OF GAESbalance be mounted external to the model with the NO. I A

use of a windshield, but these are so diverse that they A NJ -1 F

need to be addressed on a case-by-case basis and will - Ih - I' -not be covered here. Also, at times, requirements N - NJ OR 12 N " NO. I

arise for force/moment measurements to be made oncontrol surfaces or other individua' parts of a vehicle;

however, the techniques used and fabrication/cali-bration requirements are nearly identical to those des-cribed below for the "main" balance.

OUTPUITS aGMESStrain gage balances are constructed by machin- NO. I AND NO. 2 z, --

ing a thin section (called a flexure) in the balance and M F, + f2

bonding a strain gage to the surface. As loads are F - F, o F? NO. U. Iapplied, either by calibration or through the test it - F, (11 - 11)article, the elongation/compression of the flexure andthereby the strain gage causes an electromotive force b. Force balance(EMF) to be produced by the gage. This EMF can Figure 3. Force type and moment type balances.then be amplified and electronically processed inconjunction with the balance SW M AN YAWNG WNcalibration to calculate theapplied load. Machining severalflexures into a single balance atthe proper locations and in theproper plane with respect to the A

balance centerline creates the TOP VXIEL WOtC

capability to resolve all sixcomponents, i.e., normal, side

and axial force and pitching, oO-MAL FOE AU PITMWNG MOMNIE Nyawing, and rolling moments. AttltttIThe flexures and gages can bearranged to measure twomoments which are then resolved

into a force and a moment (e.g. / I--K PONIT

FN and Mx) or two forces which . L 0R WATER JAE ATITACHIENT

can be summed to obtain the SIE VIEWtotal force and, by using the Figure 4. Moment type internal balance.

balance geometry, resolved to amoment. These two types of balances and theirmeasurement resolutions are illustrated in Fig. 3. The Balance Calibration

hypersonic wind tunnels at AEDC normally use

moment-type balances. A schematic showing the Once a balance has been designed and fabricated,

normal arrangement of the flexures and gages for this it must be calibrated to determine the electrical output

type balance is given in Fig. 4. As stated earlier, (EMF) from the strain gages for a given applied load.

speciality balances measure from I to 6 components There are probably as many different techniques of

either force-type or moment-type or a combination; calibrating a balance as there are test facilities. For

however, the principle of operation, i.e., flexures and these notes, the procedures used to calibrate the

strain gages, is the same. Balances are designed and balances used in the AEDC hypersonic tunnels willfabricated for specific maximum loads. Obviously, be covered. Other techniques are very similar, and

the load range of a particular balance is determined the end result differs only by the desired precisionby the size of the flexures and the overall strength of the test data.

of the balance.

7

=.~~_ -i -ua Innu n miln l l l / • m •

AEDC-TR-9"

Balances are designed so that the flexures are the calibration loading is applied to the balance. The

aligned to measure a single component, i.e., strictly measured balance outputs are reduced to forces andFN, FA or Fy. However, it is impossible to machine moments using the previously calculated balance

the balance and install the strain gages so that they constants in a "Balance Loading Program" (BLP)

are perfectly aligned. Therefore, because of this and compared to the known applied loads. Since the

misalignment a pure normal force will produce some maximum applied loads are chosen equal to theoutput from the side force and other gages; expected maximums during the test, this comparisonconversely a pure side force will produce some output gives an insight into the test data precision. If thefrom the normal force and other gages; etc. These comparison 6f applied loads to calculated loads is

secondary outputs are called interactions, and must outside established precision bounds for thebe accotnted for with the calibration, particular balance, the balance is recalibrated.

The calibration of a balance consists of applying As shown in Fig. 6. a large number of balancesloads and combinations of loads (i.e., pure FN, FN are available for use in the AEDC hypersonic wind

+ FA, FN + Fy, etc.) in increments up to the rated tunnels. These balances vary in design and cover a

capacity of the balance and at several locations along wide range of load-carrying capabilities. For the high-the entire length of the balance. The electrical outputs est quality data, a balance should be selected that hasfrom the strain gages are measured, amplified, and maximum rated load capability close to the maximum

loaded into a very complex computer program. This expected test loads. Also shown at the top of Fig.program takes into account such things, in addition 6 is the water jacket which covers the balances when

to the applied load and its location, as the weight of used in the hypersonic wind tunnels. Since the output

the calibration equipment, balance deflections, bal- of the balance is highly temperature sensitive, it must

ance roll angle, etc. After the entire set of calibration be kept cool (near room temperature) during dataloads has been applied, the program calculates the taking. This can be accomplished by either gettingbalance constants to be used to resolve the forces and the data very quickly or by use of the water jacket

moments sensed by the b-dance during a wind tunnel which shields the balance from both radiative and

test. These constants, along with the other conductive heating.information shown in Fig. 5 are loaded into the testdata reduction computer. A balance calibration is Model Fabricationquite complicated and time consuming, typically

requiring 4 to 5 days; however, a single calibration The choice of the facility to be used to satisfy thewill normally be used for several tests. As a check, test objective will dictate the size of the model that

however, prior to each test a simplified version of can be tested. The model scale must be large enough

to maintain the fidelity of the full-scale configura-

ATTITUDEIAELA, AERO AND FLOW

CONSTANTS FLOW PARAMETERS COEF CONFIGURATION PARAMETERSAE RLO DYN A MI C (O ff qL a IA .

MDLCO P.ETC ET( •

00 ONSTANTfS REDUCTIONET EC

+

COMPARISONS TOTHEORY AND/OR PAEVIOUS DATA

TUNNEL CONDITIONS

TUNNEL FLOW BIALANCE MEASUREMIENTS

Figure 5. Static stability test overview.

8

AEDC-TR-94-6

_ __ disadvantages. In the point-pause method, data areSwobtained by positioning the model at a discrete angle

of attack and angle of sideslip, usually waiting forthe base pressure to stabilize, and recording the mea-surements electronically into the data reduction com-

puter. This is repeated for each model attitude

desired, usually in 2- or 3-deg increments over the

- ,•AERODYNAMIK .f

" GENERAL PURPOSE -COMPONENT BALANKES (24 AVAILALE)o NOMINAL BALANCE SIZES: LENGTH, IN. S.1 TO 13.8 (13-35 am) AANC REFERIENCE POINT

UDIAETER, IN. 0.6 ro 1.1 (QS-s3 cm) MONEL CENuTR OF GRAVITY ((I)6N • FIAWE RANIGIES, Uh: 20 TO 1.500SIM FOR CERANGES. LI: 2010 IO O STAINLESS STEEL MOOEL - AS LIGNT AS POSSI

.AXI L FORCE RANGES, LA: 4t0 300 * IDEALY M L €. 6. IN G M OF BALANC E lM GA O)" SPECIAL PURPOSE BALANCES (12 AVAJIAIL)*ELYML(..UCNE FRWIE(QIUSRS

* BALAICE AlSO LOCATED AS CLOSE AS POSSIILE TO THE• TYS AVAILAILE: M MASS ADDTION (3 TO 6 COMPONENT) AEOYNMC CENTER OF PRSSURE

• MAGNUS FORCE AND MOME1NTS e DATA PMECISiON DEPENDENT ON ACCURATE MEASURENTS OF:* REFERENCE AREAS

Figure 6. AEDC tunnels A/B/C balances. ( RKFERIENE LENGTHR TO BLA* €.6. LOCATION WlITl RESPECTO* MODEL COMPONENT WEIGHTS

tion; otherwise, the wind Figure 7. Model-balance arrangement.tunnel data will notaccurately predict the flight ADVANTAGES DISADVANTAGESperformance of the vehicle.

I. GREATER DATA PRECISION I. 11UCN StOWER THAN CONTINUOUS SWEEP-MORE COSTLYModel weight must be keptlow so that it is only a small 2. BAS PRESSURE MEASUREMENTS AT EVERY 21. ONE 1 DATA POINT PER RUN IN IMPULSE TUNNEL

percentage of the balance MODEL ATTITUDE

full-scale capability. 3. LESS LIKELY TO HAUE ANOMAL NYSTMEIS 3. MM DSTORTION DUE TO NON-UNIFORUM HEATING

However, because of the EFFECTS

high-temperature environ- 4. USE PITCH AND RO MIECHIANISM TO SET 4. DATA NOT NECESSARILY AT SANME MOEL ALTITUDE

ment, models for hypersonic PREDETERMINED a AND 0 FOR EACH CONFIGURATION

force testing are usually S. SIME DATA REDUCTION

fabricated from stainless Figure 8. Point-pause data taking technique.steel. After the model iscompleted, the reference areas, lengths, weight, and entire range of angle of attack and/or 5-deg incre-c.g. location must be accurately measured for input ments over the sideslip range. The advantages andto the data reduction program. These model dimens- disadvantages of this technriue, are shown in Fig. 8.ions and weights must be obtained for everyconfiguration to be tested. The test data pre- In the continuous sweep method, high-speed datacision/imprecision is a direct function oi how are taken as the model is pitched, rolled, or yawedaccurately these measurements are obtained and how through an angle range. Angle change rates varythey replicate the full-scale vehicle contours. The from a few degrees a second to tens-of-degrees permodel is then assembled with the balance, water second. The continuous data are "curve-fit" by ajacket, and balance sting, and the relative location computer routine, and finite data points are tabu-of the model c.g. with respect to the balance center lated at the desired angles. The data reduction pro-is determined (see Fig. 7). It is now ready to be gram takes into account the same items as in theinstalled in the wind tunnel. point-pause method, i.e., balance/sting deflections,

model weight, etc. The advantages and disadvantagesTesting Methods of the continuous sweep data method are shown in

Fig. 9.Wind tunnel force data are generally obtained in

one of two methods, either point-pause or continuous During both the point-pause and the continuoussweep. Each of these methods has it advantages and sweep tests, data are usually repeated for a pitch

9

AEDC-TR-D-

AIAN•WS DSADVANIA6$S quality final drag data

t. LiSS TUNNE riMIa. t1. wLEuSAA DATA require measurementswhich will allow calcula-

20lBTINi ENTHIR10 RATNA 011211KI ON RUN IN 2. US PRESWKlt INt.AWKNENrIS At START ,/AO (NO ONLYa2 " TINiI1i tions of the mass flow

through the duct over thefull range of test conditions

and pressure measurements4. MUCN LESS M00uL D6TOHTI0U Olt T0 t DATA SIUttON RIO~S M~lt (0iF1T1I STOW1 A[ TIME from which velocity can beDIFIEKHIIAL HATINGf

calculated. These measure-

Figure 9. Continuous sweep data taking technique. ments can be obtained byseparate test runs usingmass meters and a set of

series, but with the model rolled 180 deg. By com- well-designed pressure probes, or during the standardparing these data to the zero roll data, an evaluationof te efecs o nonnifrm M2,varitios, low point-pause force test runs using pressure probes andof the effects of nonuniform (M a. variations, flow svrlsm lfi g as m to s

angularity, etc.) tunnel flow can be made andcorrections programmed into the data reduction.Great care must be taken to interpret these data * I MEASUM A iFK AND TM MORIENTS (UNW IUE UNW

correctly and to not make erroneous adjustments to MMEINT AN HDT W K NOET)tWhW AIN DATA R 0NM VERY SWiR TO SAMN SlUIE

the data. WHEN USE- WI MA IMAIK10 1MU B EXEtCIHOD TO NWYNT IT VIM 31 tLEAS 00 NOT NTEIKEM WITI WAE IMAINC

Additional Test Methods 511RU-TMJNIIGJM SUDTIUSNINElUB* TMTAL A11t-OH(E MiUST BE AAII IONU UCT NM

The preceding section addressed the standard six- 0 NET IHUM CALCUIATED 1 MASING MASS RFO AW MO V10-

component force tests with base pressures. Very OTUM KE

often, additional measurements (Fig. 10) are required TM TJO SYSTEM (US)* UE IFO STORE SEPAMiin order to fulfill the test objective. Quite often, • 6g uNODUsJE1IOs

vehicles have control surfaces for which the loads at

various deflections are required in order to ensure Figure 10. Special test requirements.

the structural integrity of the component, its attach-ment, and the control system. These loads are mea- The AEDC Captive Trajectory System (CTS) is

sured with a small balance ("fin balance"), usually shown in Fig. 1i. CTS tests usually have a parentthree components, mounted internally to the model. vehicle mounted on a six-component balance sup-

Great care must be taken during wind tunnel installa- ported on the tunnel standard pitch mechanism andtion of these models to ensure that the electrical leads an additional model of an "external store," (missile

to the fin balances, which must "jumper" the main or bomb) mounted on a balance connected to thebalance, do not restrict the main balance deflections CTS. The CTS is a mechanism which can produceand therefore cause erroneous readings. These six degrees of freedom independent of the mainbalances are fabricated and instrumented very similar model support. The attitude and position of the CTSto the main balance and require the same type cali- model relative to the parent vehicle can either be

bration and data reduction. They normally measure preprogrammed points in the CTS computer (grida normal force, hinge moment, and the root bending mode) or determined by the forces and momentsmoment. sensed by the CTS balance (trajectory mode). This

is a highly complex type test and requires a largeAnother fairly common and yet non-standard amount of pretest set-up. However, once it is

model for hypersonic wind tunnel test requires the operational, a large amount of data can be obtainedsimulation of engines which require ram air. To truly in a relatively short amount of wind tunnel time. Thesimulate the flight vehicle, the wind tunnel model AEDC/CTS is the only such system known in themust be built with a "flow through" duct. The world to operate at hypersonic speeds. The spaceinternal part of these "engine simulators" cannot be shuttle solid rocket booster separation was tested atfabricated to truly reproduce the drag/thrust of the M = 4 in Tunnel A, and separation of the externalfull-scale vehicle; therefore, corrections to the mea- tank (E/T) from the orbiter was tested in Tunnel Bsured wind tunnel drag must be made. The data for at Mach 8.these corrections is provided by one of several means,depending on the desired precision. The highest

10

AEDC-TR-94-6

temperature instrumentation, and other inputs areamplified, converted to digital form, and fed to thedata reduction computer. These data are combined

(is SUPPORT with the previously input balance calibration and. j o, XYZ VALAE wind tunnel calibration data. The tunnel calibrationCOMPUTER PROGRA MMEDAOM CONTRIOLLED data are used to calculate the test conditions. Model

positioning readouts, are combined with theS• force/moment data from the balance to calculate the

aerodynamic coefficients and model attitude. Thesedata can thea be used to resolve the coefficients into

7any axis system (body, wind, pitch, etc.) desired.

MAIN TUNNEL MOIL Total drag is made up of skin friction, base drag, andSUPPO•T VARIIi! wave drag. Since the skir friction is generally a small

5011. DIAMETER part of the total drag, no adjustments are made forany difference between the wind tunnel value andflight value. This is not, however, the case with base

a. Tunnel B installation sketch axial force. Because of sting effects in the windtunnel, the base pressure may be very different fromthe flight value. Therefore, an adjustment to the axialforce measured by the balance (CAT)T is made. Asnoted in Fig. 12, the wind tunnel base axial force(CAB)T, as calculated from the measured basepressure is subtracted from the (CAT)-r to obtain theforebody axial force (CAF) which is the same forboth wind tunnel and flight. The flight total axialforce (CAT)F is then derived by aLdng a predictedflight base drag (CAn)F to the CAF. Otheradjustments may be applied to the data to com-pensate for such things as wind tunnel flowangularity, model/balance misalignments, etc. In thespecial case of a model which has a simulated engine

b. Test of shuttle booster rocket separation in duct, an adjustment to the drag data, similar to thetunnel A. base drag adjustment, must also be made. There are

Figure II. AEDC captive trajectory system. several methods used to calculate the internal dragof the wind tunnel duci with varying degrees of ac-

Data Reduction curacy. The most precise method is to measure themass flow through the duct using a pre-calibrated

The data reduction program for point-pause type mass flow-meter in addition to measurements (totaldata is quite complicated, and yet much simpler than and static pressures) from which velocity can bethat for the continuous sweep type. At each desired calculated, and then to adjust the total axial forcemodel attitude (data point), the electrical outputs by the momentum loss of the flow through the duct.from the balance, model attitude sensors, basepressure transducers, wind tunnel pressure and The data reduction

program for continuoussweep data is the sameas for point-pause,

NOTE: ASSUMES & dT - (Cf), except for the manner inwhich the data points to

be tabulated are gene-- TOTAL AXIAL FORCE COEF rated. As explained

t oSE AXIAL FORCE COEF C) = (€•7), - (Ci0)• (CA,), - CaF + (C4), earlier, instead of data

CA, -SKIN FRICTION C., (FROM MEASURED BASE PRESSURE) (PREDICED) at finite model pitch/C4 - FORtEBODY AXIAL FOR(CE OEF WIND TUNNEL ( )I FLIGHT ( )' yaw positions, a conti-

nuous stream of data is

Figure 12. Axial force accounting. generated over a pitch,

11

AEDC-TR-94-6

yaw, or roll sweep. These data are then fit with a to the measured values provide a quick-look indica-computer-generated polynomial for each variable, tion of the data quality and, in the case of aand then the finite data points are generated at the parametric study helps determine the particulardesired model attitudes. These finite points are then configuration that will provide the best flight results.operated on by the computer in the same way as for The comparisons can save a large amount of windthe point-pause data. If adjustments are to be made tunnel time and thereby sizably reduce the ovet allfor base pressure, internal drag, or other items which cost of the program. The total data reduction flowrequire pressure measurement, separate test runs is illustrated in Fig. 13, and typical stability data aremust be made in the point-pause mode to allow time shown in Fig. 14.for the pressure instrumentation to respoudaccurately. These data can also be curve fit and fed SURFACE PRESSURE TESTINGinto the same adjustment routines as in the point-pause mode. Surface pressure tests in hypersonic wind tunnels

are primarily conducted in association with surfaceIn addition to the calculations, adjustments, heat-transfer and/or flow-field probing measure-

and/or corrections discussed above, the data ments to provide inputs to or to validate a CFD code.reduction computer is quite often preprogrammed Of current high interest are pressure tests definingwith theoretical predictions (CFD) and/or previously the internal and external aerodynamics of scramjetmeasured values of the wind tunnel model aero- propulsion systems.dynamic performance. Comparisons of these values

MEASURED TEST DATA DATADATA CONS .T ADJUSTMENTS OUTPUT

OUTPUT

UMODEL AMTTTU.POSITION CONMD ON

'USEm IIALANE IN IoNTERNAL -NMIOF ATTITUDE

PUSS~~~DAT CDMPAIN CNTIUOU W

DATUTNNELEL

CONDITIONS

Figure 13. Data reduction.

. • 3.00D.6 -0.01 2.00

e0.4 - 0 1-002

0.2 -4.03- 0

0-0.04 -. 001-S 0 5 10 1s 20 25 -S 0 S 10 Is 2 25 -5 1 5 10 is 20 2S

AlNGLE Of ATTACK A OF ATTiCK ANGLE OF ATTACK

Figure 14. Example of on-line stability data.

12

AEDC-TR-94-6

Pressure Transducers Since it is important to obtain data as quickly aspossible because oi the test environment and/or toStatic pressure levels on test models are generally reduce costs, the transducer should be located close

very low in hypersonic flow. However, in cases of to the point of measurement on the model to reduce

shock wave interaction and/or impingement, they the pressure stabilization time.

can be orders of magnitude higher than the free-

stream static pressure. For this reason, great care Miniaturization has allowed large numbers ofmust be used in choosing the type of instrumentation individual transducers to be connected together intoto be used to obtain the best possible precision, and what is called an ESP (Electronically Scannedalso to protect the measurement device from Pressure) unit. These units can usually be housedoverload. In general, model surface pressures in within the test model or mounted very close withinhypersonic flow are measured with a pressure the mounting hardware. They must, however, betransducer similar to that shown schematicaly in Fig. cooled (usually wvith a water jacket) because the15a. This differential transducer senses the difference transducers are highly temperature sensitive. Thein pressure on the measuring side from that on the advantages of these modules, in addition to theirreference side. As seen by the schematic, a small extremely small size and small volume of the overalldifference in pressure will cause a deflection of the system, is that they require only one reference line,thin diaphragm, resulting in an electrical output from one calibration line, and one set of electrical leadsthe attached strain gage. The deflection and, for all of the transducers instead of individual lines.therefore, the output and maximum allowable The major disadvantage is that all surface pressurespressure differential, is a function of how rigid (thick) must be near the same value because all of the trans-the diaphragm is made. These transducers are ducers within a unit have the same maximum pressuremanufactured in pressure ,atings from a few rating and, as stated, use a common referencehundredths of a psid to several thousand psid. pressure. They are, however, usually protected

against a large overload. A typical unit capable ofsu1muUM measuring up to 32 model pressures and manu-

factured by Pressure Systems Incorporated is shown

MEASUREMENT in Fig. 15b.

MODEL SURFACE SIDE REFERENCEREFERENCE - Transducer CalibrationPRIESSURE - SIDE PRiESSURE-

L _Pressure transducers must be calibrated, as in the

ELECTRICAL LEADS case of the force balances, to determine the electricalFROM STRAIN GAGE output of the strain gage as a function of applied

a. Pressure transducer schematic. pressure differential across the diaphragm. Unlike thebalance calibrations, however, the pressure trans-ducers are usually calibrated at least once a dayduring use. These calibrations take only a fewminutes because there is only one component and,therefore, no interactions. The calibration can be

accomplished in one of several ways, depending onthe type and magnitude of the rated pressure of thetransducer. The most common method used fortransducers rated up to atmospheric pressure is toreduce the pressure on the reference side of the

diaphragm in increments by applying a known(6.3 cm) ( .4.3cm) pressure less than atmosphere. The magnitude of the

2.5 3 [-0. 2 1.7 applied pressure is measured with accurate instru-

__ • mentation (secondary standard) traceable to a

(2.S cm) meao (c ry nr. ) a l 000000000primary standard. A secondary standard is a field1.0 R'REFERENCE measurement device traceable to a laboratory pri-

EASUEENT TU - -mary standard which is, in turn, traceable to theSTUBES 32 CALIBRATE TUBE National Institute of Standards and Technology.

b. Electronically scanned pressure unit (ESP) Since most transducers are not linear through zeroFigure 15. Pressure measurement device, pressure differential, if values of the test pressure to

13

AEDC-TR-94-6

be measured are expected to be both above and below with stainless steel tubing attached by one or morethe value of the reference pressure to be used during of the methods shown in Fig. 16. When the confi-the test, the applied calibration pressure should also guration is such that the backside or inside of thecover both cases. Quite often, several individual model is accessible, the type installation shown at thetransducers being used for a test are connected to a left is the most desirable. In this type installation.

common reference, as is the case for the ESP units. a small hole, usually less than 0.050 in. (1.3 mm)

In these cases, all transducers on a common reference diam, is drilled through the model wall, a counter-

can be calibrated at the same time. bore the size of the O.D. of the tube is drilled partway through from the backside, and the stainless steel

If a transducer has a sealed vacuum on the tube is solderqd in place. Either of the two right-hand

reference side of the diaphragm (an "absolute" installations can be used when there is not enough

pressure transducer), or the maximum rated pressure room or access to solder the tubes on the backside.

is greater than atmospheric pressure, or it is an ESP Care must be used in installation and/or handling

unit of the type shown in Fig. 15, it must be calibrated this type model, or leaks can develop around the tube

from the measurement side of the diaphragm. This or the tapered plug, resulting in measurement errors.

is accomplished by sealing off the tube going to the Any of the three installations should be checked for

model surface and applying the reference pressure leaks by applying a vacuum to the surface tap, sealing

through a "T" as shown in Fig. 16. In the case of it off, and observing the test instrumentation over

the ESP unit, this valve is internal to the unit; a few minutes to obtain a leak rate. The entire qystem

however, for other types of systems, the valving and must be clean and free from foreign material (such

calibration tubing must be added to the system and as oil), or the outgassing may appear to be a leak.The number of pressure taps on a model can be limi-can become quite complicated to ensure there are no

leaks anywhere in the system. The calibration ted by available room for internal ESP units or forthe tubes routed through the mounting hardware. As

pressure is applied and measured in the same manner the cas of the fre models, herscale otas te rferece idecaliraton.in the case of the force models, the scale of the

as the reference side calibration.pressure models and the fabrication process must be

MODEL SUUtKE such as to maintain the,,_,:'",•/ f~•1Pl STANOAU0 fidelity of the full-scale

"1 --- SEE SKETCHES uW T--RANSDUCER ROMA configuration. The "as-built"REIFERENCI/AMIAATE location of the surface.. " SSU SURrLY pressure taps is critical to

---- _•. i.-ALTERNATE -M1ASUIRMENT obtaining data which willVA L'V EISE IAwIoATIIN SI.M accurately predict the flightMEASUREMENT SIO ST1-1ONO ANAO pressure loadings or whichCAUIIIATION SYSTEM IM can be used to validate

theoretical computer codes.

Testing Methods

SWARED OR SWEAT SOKID t[OREDAND MACHINED AFTERWARD SWAGA* PLUG Testing of surface pressureAND

MACHINED AFTEIVWARO distribution models in hyper-

ACC SUIE INACESSILE IACiASSD sonic wind tunnels can be verNtime consuming and therefore

Figure 16. Pressure measurement and calibration schematic. more expensive than staticstability testing. In the

intermittent tunnels, usually data for only one modelModel Fabrication attitude per run can be obtained because of the time

it takes for the pressure to stabilize. For the sameAs in the case of the static stability model, the reason, even in the continuous flow tunnels, it takes

choice of test facility dictates the maximum size longer to get the pressure data over a pitch or yawmodel. Weight is not a critical item in the pressure polar than even the "point-pause" type force data.models; therefore, reinforcing and other fabrication Even when the model and pressure systems aretechniques can be used to reduce the thermal distor- optimized to reduce the required stabilization time,tion during testing. Pressure models for hypersonic it still requires up to minutes per model attitude.testing are usually fabricated from stainless steel, Because of the expense in time/money, pressure tests

14

AEDC-TR-946

are usually designed for data points in wider incre- AEROTHERMAL METHODOLOGY

ments than the normal force test; i.e., where data for

a force test may obtained for - 15 < n < + 15 in Fundamentals and Simulation Parameters

2-deg increments, the pressure test would probably The requirements for developing hypersonic flight

be 5-deg increments, vehicles place increasing demands on ground test

Data Reduction capabilities. Of particular concern is the requirementto demonstrate that flight components such asleading edges, cowl lips, and structural panels will

Test data are combined with previously input survive the aerothermal flight environment. Specific

calibration data (transducers and wind tunnel), and components as shown in Fig. 18, can experience heat-

the data output parameters i.e., P, P/PC., Cp etc. ing rates ranging from 200 to 2,000 Btu/ft2-sec and

are calculated. A data reduction technique called the surface temperatures from 1100-1940 0C (2,000 topressure prediction routine can be used to reduce the 3,500°F). Ground test of flight components have pre-

amount of time required to obtain the pressure data viously been performed at test facilities like those at

for a given model attitude. To apply this technique, NASA and AEDC. This section presents an overview

the output of the pressure transducer is recorded in of the materials/structures test methodology and the

uniform time increments (- I/sec) for a period of test techniques used in the development of hypersonic

time (usually 30 sec) after the model has been set at vehicle components.the desired attitude. The results of these samplings arecurve fit within the computer, and the FATNER WINGITAILresults are extrapolated to the JOINTS AND SEALSN '(PROITUBERAN(ES) LEADING EDGESasymptotic value of the actual pres- REWIEQOUIPMENT / -IOUNItY LAYER TRANSITIONsure. The technique, illustrated in Fig. WINDOWS ___ / __

17, takes into account such things as loxpressure tube geometry, gas tempera- NOSE CONTROL

ture, viscosity, system geometry, etc. COOLING SURFACES

As shown by the illustration, the actual HOT STRUCTURE '/ U10LIM ENGINEINUATO 111OIWING PANELS COOLING

equilibrium value of the pressure was AND FUELpredicted very accurately by using the

input obtained during only 30 sec when Figure 18. Typical aerothermal structures/materials issues.

it would have taken well over a minute

for the system to stabilize to the final value. The wind A review of some fundamental heat-transfer

tunnel test time/costs can be greatly reduced with concepts is presented in Fig. 19. The typical textbooks

very little sacrifice in accuracy by applying this discuss the convective heating to a wall and relates

technique. the heat flux, q, to a temperature difference, Tg-Tw.

For aerodynamic heating the heat flux is also

NOWUA ran- rMy proportional to the temperature gradient at the wall,

1.i1 = t.0.04+0W TaIM, - s.St m II and the heat-transfer coefficient, h, is used to relate0.4482 0.06S - smot the heat flux, 4, and the temperature driving

09 0DImO FINA 1 msIU potential, Tr-Tw, where Tr is the recovery.392 0.055 + 4s i.t" i &1re 11.0as10 nsa) temperature. The experimentalist often uses the

44 45 *.111IN gtS1.1n Mo (00 o .•8 t.e (01111 prA) facility total temperature, To, in place of the more

. . .0 X. 0* - .15 06 P.1167 rI elusive recovery temperature Tr.

S0.2413 0.035 I%Ir 'A common approach used to solve aerothermal

0.1724 0.02S •issues is based on combining analysis with experi-0.11240.025mentation. It is imperative that analytical techniques

0.1034 0.015 , n i , r--- be used to plan the test and to analyze the final data.0 10 20 30 40 sO o 0 The two fundamental steps in the development pro-

TIME, SEC cess are:

Figure 17. Typical pressure stabilization curve. 1. defining the flight thermal environment

2. demonstrating hardware survivability

15

AEDC-TR-B44

is then used to extrapolate the resultsCONVECTION AEROHEATING to higher Mach numbers incorporating

TIS, k is real-gas and viscous effects as re-quired. This procedure is illustrated in

T Fig. 21, which includes the primary test

facilities used to obtain heat-transfer

Yly- 0 data in the U.S. The AEDC TunnelsWALL (W) I B and C are the national workhorseWAL 4!1YT. WALL (W) facilities in this category, and it has

q - 0(1 - TQ" q- 4 been estimated that 75 percent ofk'lp I - - existing hypersonic data defining

LET k- NTh',-W:] thermal environments have been

"WHERE: .l - HEATING RATE, STU"FT•E 5C obtained in these tunnels.

h - HEAT TRWNSFER COEFFICIENT, The test techniques availableITU/M SEC OR measure aero-heating are listed in I

T, - RECOVERY TEMPERATURE, °1 22, along with a reference wh)(V, T,) illustrates the use of the technique.

T. - WAlL TEMPERATURE, -1 Thermal mapping techniques providea comprehensive look at the entire

Figure 19. Basic aeroheating concepts. model and are often used to identify

PHASE I - DEFINING THEtML ENItOW TS (STEP 1) the location of high heating rates (e.g.,PHASE 1 . sEoNkG THERMALtiNn).NMHowev(STEthI* SCALE MODELS IN WINS TUNNELS * NAI RWSFEI TEST TECHNIQUE shock interaction). However, the

uncertainty of thermal mapping datais of the order of t 15 percent,whereas the discrete measurement

Pt"tS., C OE %EMaIIo1 =: RAPW. TO FUMG techniques can produce ± 6-percentSHEATING IPU ) data which are more reasonable forSTHERML ENINMENT code validation tests. Additional

details on heat-transfer measurementPHASE 2 - DENONSTRATE HARDWARE SURWMABILITY (STEPS 2, 3, 4) techniques may be found in Refs. 4, 7,

"MATERIAL TEST STRUCTURAL CO I FLIT HARDWARE and 10, and a brief overview of eachTEST DENO TEST technique is presented below.

(SAMLES) (CMONENTS) ((COMPONENTS)

SDUPLICATE LOCAL T ARTICLES Phase Change Paint Technique

ENVIPIUMENT tzfTT(a.*. q LOCAL - q FUENT) The Phase Change Paint technique

of measuring the heat transfer to aFigure 20. Methodology for aerothermal structures/materials model surface was developed by Jones

development. and Hunt.II This technique assumes

that the model wall temperature re-These two fundamental phases are illustrated in

Fig. 20. In defining the thermal environment, the STEP (1)versatility of analytical tools is combined with the DEFINING THERMAL ENVIRONMENTSexperimentally measured heating distributions. These (I.E., WHAT HEATING RATES/ITEMPERATURES ARE ENCOUNTERED IN FLIGHT?)

data, obtained on scaled models in simulated flow APPROACH:environments, are used to verify the accuracy of the ANALYSIS EXPERIMENTSanalytic tools. The important simulation parameters * ESTIMATE HEATING DISTRIBUTION * SCALE MODEL TESTS/SIMULATED

o ENGINEERING CODES MACH AND REYNOLDS NUMBERare Mach number and Reynolds number. However, 0 CGO e AEDC TUNNELS I & Cit may not be necessary to match the Mach number * EXTRAPOLATE DATA TO FLIGHT 9 NSW( TUNNEL 9

because of the "Mach number independence princi- * THERMAL RESPONSE CALC. * CALSPAN SHOCK TUNNELS* AMES 3.S FT TUNNEL

pIe." A commonly used procedure to define the o Lot( TUNNELSthermal environment is to use the data obtained atMach 8 or 10 to substantiate a code at precisely the Figure 21. Development process, step I.same conditions as the experimental data. The code

16

AEDC-TR-94-4

ANVAITAOEI INSIANTAUS TlF.

* PNASE-CmAMGE PAMN1 VIM ILLUSTRATION O NOT ISPOT MUST REAItY PANT. DATA PRESEIIATN 2NI6N SATIAL RESOLUTION CAMN BE WNFUI

* I S0 liS CAMERA (DUPUIERIENERTme PKlTS Ai 0OlO SAITAL REISUTION 3oMA, ANODNMUO US9W

* TiEID06APNK PNOSPINR 0UPFLEI M , GM SPATIAL MOEL PEPARATID1 AI NATA 4.111E1S1,1ON PIESENTATIN

"* 11N-Sm. NIGH N1ANlY MATA, MIE SPACING OXPHIIM MREL FAD, EU"I SEFFT•"* WAX GAGE EASY TO INSTALL, ONTOUWU, KOURIM M LO OUTPUT, TST IESon 6"* SiOIMMT.4ELTER 6AE IG OUTPUT, SUMi i TOfuML, FAN AN WJLIATUN TIE REUiM 7

VERY WRINLE"* GWEAN WAS (III TEMP, LO TEMP) YEWAl OF EIPERIIEC. FAST RESPOINSE GAGE ATTIIOI RATI, NOT IUM I"* INI-FU DENSE SPACING, FAST RESS, CAN U RELATIVELY FFIWLT iNTALLATKII, 9

USED ON SMAIL RAN MATERIIAL ]M(36

Figure 22. Test techniques available for measurement of heat transfer.

PRINCIPLE OF OPERATION

SURFA(E TEMPERATURE MEASUREMENT OF SEMI-INFINITE SL

T,

~ith~~ GaTIME

DATA REDUCTION EQUATIONS

TV - T, I

T,-Ti (B)2ACKTE- iUr#K

WHERE 0 - e (2)

Figure 23. Phase change paint technique.

sponse is similar to that of a semi-infinite slabsubjected to an instantaneous and constant heat--transfer coefficient (see Fig. 23). The surface walltemperature rise for a semi-infinite slab is given bythe equations shown in this figure.

A specific value of the wall temperature (Tpc) isindicated by a phase change paint (Tempilaq® ).These paints change from an opaque solid to atransparent liquid at a specified phase changetemperature (Tp). For known values of Ti, Tpc, t,and th hea-trnsfr coffiien (h)canI~e Figure 24. Typical examples of phase-changeand QV-eck, the heat-transfer coefficient (h) can be pitpoorps

calculated as a function of the time required for the paint photographs.phase change to occur by using mechanism at the desired test attitude, and the model

h ck (3) initial temperature (Ti) is measured. The model isA then injected into the airstream for approximately 25

sec, and during this time the model surface tempera-where P comes from the solution of Eq. (1) since the ture rise produces isotherm melt lines. The pro-left-hand side is known. gression of the melt lines is photographed with

70-mm sequenced cameras operating at one or twoPrior to each run, the model is cleaned and cooled frames per second. Typical examples of phase change

with alcohol and then spray-painted with Tempi- paint photographs obtained during a run are pre-laq® . The model is installed on the model injection sented in Fig. 24, and an example of phase change

17

AEDC-TR4-96

SIM DATA Infrared Sca"ningOPE 111111-SO MRla PAN, SIO MR FUSLG VENT,,O, FLAVAI PA1, PWUhAE 11101 111. PILOT WITA Thermal mapping techni-

Va - I NOiS WIL- q ques used in wind tunnel testII W •• applications generally involve

5.6 _ the use of heat-sensitive model

0.1 surface coatings. The majordrawback to these methodshas been the time required to

0 NkIGN.l:~ AIS~obtain quantitative data from

"0.01 - %( SMED DOW TO t SALE. photographic test results.00

0 oWith an infrared (IR)0 0 0 o o o scanning camera system, heat-

0.00 _____________________ transfer coefficient data in the0 0.1 0.2 0.3 0.4 0.5 form of tabulations, plots,

IlL and surface maps are pro-

Figure 25. Leeward centerline heat-transfer distributions at duced within minutes of test

ReL = 8.6 x 106, a = 30 deg. run completion.

A typical installation of the AGA Thermovision 680paint data compared to thin-skin data is presented scanning camera for an aerodynamic heating test isin Fig. 25. This figure also illustrates a common sketched in Fig. 26. The camera is positioned outside

technique used to extrapolate wind tunnel heating of the wind tunnel environment. The infrared

distributions to flight. The wind tunnel data are radiation emitted by the test model within the field

normalized by the Fay-Riddell stagnation point of view of the camera is collected by the system optics

heating on a 1-ft-radius sphere scaled down to the and focused on the camera detector. The signal gen-

model scale. To obtain heating rates in flight, the erated by the detector is proportional to the detected

distribution is multiplied by the Fay-Riddell heating infrared radiation. Two rotating prisms form an

on a I-ft nose radius sphere "flown along the flight optical scanner which controls the position of the

trajectory." The basic assumption is that the distri- instantaneous field of view (IFOV) of the camera.

bution at M = 8 is unchanged from tunnel to flight. _ rd._- L_.

A complete description of the phase change paint FLOW =.... , Oi1EBtechnique as applied to a particular test situation is iE

presented in Ref. 2. MON aOPIA %.. TESTR U M P 2 SC IU E TEST M DATA DATA

MODEL - SYM DIGITIZER11 CAMERA

WmaIND MAWAM I ~~~TUNNEL ALTPOS

DATA

NOE Figure 27. Infrared system schematic.

O.M (1.3.) The complete infrared system in use at AEDC is

schematically illustrated in Fig. 27. The system is

VIEW ILOIE NUIAUA1I K TMIL M composed of the test model, the AGA 680 camera,and a data system to collect, digitize, and convert the

Figure 26. Sketch of typical IR camera installation, camera signal to the desired data output. For a typical

18

AEDC-TR-94-6

test, there are several modes of data output that can area in the object plane. Since the detector in the

be selected to fulfill the test requirements. One type AGA 680 camera is circular in scope, a "spot

of data output is a tabulated output of model surface diameter" is viewed in the object (test) plane (see Fig.

temperature or heat-transfer coefficient for each 29a).desired position within the total field of view. Acapability of presenting the temperature map of the The ability of the IR system to track a tempera-

model surface in the form of a color plot is used, and ture profile across a "worst case" step-heating gradi-

a sample is presented in Fig. 28. Other forms of data ent will be discussed to aid in the understanding of

presentation consist of both 70-mm photographs and the data. Assume that the IR camera is scanning

16-mm movies of the color video monitor. along the centerline of a target thatis composed of a plate at a uni-

form temperature, Tc, that has acircular protuberance at anelevated uniform temperature of

MO Validat I ns• TH as shown in Fig. 29b. For aTis Esystem with perfect optics and

electronics, the ability to track astep increase in temperature is afunction of only the IFOV whichdetermines the scanning spotdiameter. If the camera had aninfinitesimal IFOV, the tempera-ture profile would be trackedexactly as shown in Fig. 29b. This

an 3st 3g 41o 46s sis s64 1s Go,£ ,, rs, os ,9 ,s, ,,, less 112i, tii. case is of academic interest only,

' 273 2a 378 *419 467 S6, SI US 6 14 1'3 409 ese go? 95s i, toss isz but represents what would be re-

quired to track a step change in

Figure 28. Illustration of IR thermal mapping data. temperature exactly.

To aid in the understanding and interpretation of SPO _Mthe data obtained with the IR scanning system - 01"

requires the following: (1) a definition of some basic Dr-YN -7Tterminology associated with the IR system and the UNE .--- -

data, and (2) a review of basic IR system operational SCAN O.4

characteristics and limitations. T_-. I

The total field of view of the camera at the windtunnel centerline is a function of the lens selected for a. IR target

the camera and the distance from the camera focalpoint to the centerline. The desired total field of view

is determined based on the size of the test model to TN

be viewed and the spatial resolution (to be discussedlater) that is desired. The total field of view is arectangle as shown in Fig. 29a. One complete scan

of this total field of view is defined as a frame ofdata. Each frame of data is composed of a matrix

of 70 line scans, each containing 110 points, for thetotal of 7,700 discrete measurements. Each line/point

*combination identifies the location of the centerline b. IFOV = infinitesimal

of the IFOV of the camera as it scans the total field Figure 29. IR scanning of a step change in

of view. temperature.

The IFOV of the camera is specified as the angle Next, we will assume a practical spot diameter of

in milliradians subtended by the projected detector 0.16 in. (4 mm) based on an 8-deg lens viewing the

19

AEDC-TR-9-6

target shown in Fig. 29 at a distance of 76 in. (1.9 As a result of the limitation discussed above, the

mm) (typical). The nominal temperature profiles for IR scanning technique of obtaining heat-transfer datavarious target diameters are presented in Fig. 30. This has not been used in recent years at AEDC. How-shows that the temperature profile is not tracked ever, a more complete description of the infrared

exactly. This is true because at each of the IR system, the data reduction, testing techniques, and

measurement points, the radiation is integrated over presentation of sample test results can be found in

the spot area. When the diameter of the target is Ref. 3.

equal to the diameter of the scanning spot, the correcttemperature would be recorded at only one scan Thermographic Phosphor Pala(

position. For a target diameter smaller than the spot

diameter, the IR system would never record the true Thermographic phosphors paints are sensitive to

temperature of the protuberance. temperature and, as the model surface temperatureincreases, their luminosity decreases. The luminosity

When the IR scan changes from viewing the of these paints is typically excited by UV or laser light

background temperature to viewing a heated target, and, as the temperature patterns develop, they are

such as the edges of a flat target, the temperature in photographed (see Fig. 31) or recorded with a video

a region along the leading edge is invalid because of camera. As recently as 1990, K. W. Nutt t 0 presented

these same integrated effects. This region of invalid a comprehensive discussion of the thermographic

data is indicated on the color plot in Fig. 28. phosphor technique. The reader is referred to thosenotes for this specific thermal

1-14G lENS mapping technique. Thermographic

W OISTAIKI - 7A IN. (193 an phosphor paints have been usedSCANNING ACROSS A CIRCULAR both in wind tunnel testing and in

1.0 -TEMPERATURE CHNGE I turbine engine tests. In the wind

0.8 - f tunnel, these paints are applied to0 ,~ models to measure surface tempera-

ture patterns which can be used to3 ~ .infer heating rates. In turbine

DDITER engine tests, the phosphor paints0.4 .,,s -o ,have been applied to turbine blades

0.2 -- .062S an-DIAMETER to infer temperature. Temperaturemeasurement with thermographic

0 0 0 phosphors is practical from about0.8 0.6 0.4 0.2 0 0.2 0.4 0.6 0.8 IN. 100 to 900°F (38-482 0C).2.0 I.5 1.0 0.5 0 0.5 1.0 1. 2.0 cm

DISTANCE FROMN SPOT CENTERLINE, IN/CM Discrete Measurement Techniques

Figure 30. Determination of spatial resolution.

Thin-Skin

The thin-skin technique (see Fig.

32) has been used for many years"and remains one of the most

DEVELOPED 1910ADVANTAGES a HIGH SENSITIVITY THERMAL PATTERNS DISPiAYEU BY accurate and reliable methods

a REVERSILE PHOSPHORS DURING A TEST RUN available.5 Thermocouple installa-S COLETE COVEGRAG IN tion is a key concern in this techni-SINGLE PHOTOGRAPH

OISADVANTAGES o DATA REDUCTION IS COML.EX que, and experience indicates that* LIMITED APPLICATION the method illustrated in Fig. 33

produces quality data. The reduc-ARK tion of thin-skin temperature dataCONTRIIBUTIONS 9 FO WIND TUNNEL USIET DATA REDUCTION TECHNIOUES COLOR CODED THERMIU PATTERNS to coefficient form normally in-

INCLUDING USE OF (OlOR COOING (SW PHOTOGRAPH Of COLOR PICTUEi) volves only the calorimeter heatbalance for the thin skin as shown

Figure 31. Thermal mapping (thermographic phosphors). in Equations 4 and 5.

20

AEDC-TR.94-O

S i nSince h, Q, b, and c are constants, then the(n = iorcd in skin derivatives in Eq. (7) must also be a constant. Hence,

qb = QI dTw/dt (4) 1he term

h Qbc dT,/dt (5) rT iTr Tw Tr - Tw n -TrTw 8)

Thermal radiation and heat conduction effects on is linear with time. This linearity assumes the validitythe thin-skin element are neglected in the data reduc- of Eq. (7) which applies for convective heating only.tion, and the skin temperature response is assumedto be due to convective heating only. It can be shown thus ifvthe dataso a nonlinty ec otherthan convective heating are present. In most cases,that for constant Tr, the following relationship is the nonlinearity will be caused by conduction effects.true: Machine plots of data from each thermocouple pro-

vide the opportunity for quick visual examination ofd ( Tr T dT/dt ( test data with the objective of evaluating conductiondt Tr T Tr - Tw effects. Once areas and/or time frames during which

significant conduction effects were present are

Substituting Eq. (6) in Eq. (5) and rearranging terms identified, the remaining valid data are used toyields: calculate the heat transfer coefficient via Eq. (7).

a0 0h d n Tr1Ti7) 0.4 0.0

Qbc dt TrTw 0.0.3 00.0

PRINCIPLE OF OPEIATION "o__ 0.2 ,0.04CALOIUAETRY (HEAT IN - HEAT STOtO IN SIlN) M AINO Em £

i.e. CDOUCTION - 0ItADIATION - 0 0. C 0.A

DATA IEIDUCTION EGUATIONS

a, -db-u_ SEC 0 Ia h - 1YtW 5K Of0

e1 Q M MATERIAL DENSITY TIME, SETb - RIAN TKNFiES T Figure 34. Representative thin-skin data.c - MATEtlA ,•t•IFK HEAT

Figure 32. Thin-skin technique. Representative data which demonstrate the above

procedure are shown in Fig. 34. Thermocouples werechosen which represent cases of no conduction, con-duction away, and conduction into the thermocouple

vicinity. Note that steady heating was not establisheduntil about 1.75 sec. This was caused by the processof injecting the model into the flow. The injectionwas actually completed at 2.10 sec. The nose thermo-

couple output (circles) is observed to be linear toabout 3 sec, and then it begins to fall away from theinitial slope. This indicates heat being lost from thisarea, as would be expected in a nose or leading-edgesituation. Around the nose (triangles), the data arelinear only up to about 2.60 sec, and then the slopebegins to increase, indicating heat flow into this area.This is most likely some of the heat being lost by thenose region. Finally, a thermocouple further back on

Figure 33. Typical thin-skin thermocouple the body (diamonds) shows no conduction effects forinstallation, the entire run. Obviously, longitudinal temperature

gradients are small in this area, and the observed

21

AEDC-TR-94-6

heating is solely due to convection. The magnitude construction, this gage has a very large temperatureof conduction effects can be estimated directly from operating range (- 320 to 1,0000 F and above) asthese plots since the change in slope with time is compared to other measurement techniques.directly proportional to the conduction error. Oncethese preliminary evaluations have been made and 0.12Sappropriate data reduction times selected which avoid (NO•i Ti.. wll i _ Tthe injection transient and conduction effects, the I,.MAIN "second step of the data reduction process can (ONSTANTA JCKETproceed. Linear least-squares curve fits of the selected EPOMYdata are obtained, and evaluations of the heat-transfer coefficients (Eq. (7)1 are made. Anillustration of thin-skin data quality is presented in lop" WItESFig. 35. Figure 36. Coax gage construction.

1.0 PIIMENTAL - THEOEKALOWMPARISON OF HNU TWSMFI The data reduction equation used to obtain the

DISTEITION ON A EMINIERE heat flux from the measured temperature of a coax

0.5 Egage is as follows:

______ v y y, (9)2C (tn) '

•/Tw - T

_

2~~ C t,

J40n - tN,+ j

where q surface heat flux at time tn

SYMIOL Ai 0.1 O 10 Tw. = surface temperature at time tj

= ~0HF1 - FAY-ItDELL TNIfuy The term C (tn) is essentially the scale factor for

0"050 0.4 0. ! 1.1 .6 2.0 each coax gage with the thermal propertiesM temperature dependence included.

Figure 35. Illustration of thin-skin data quality. The semi-infinite solid assumption used in the

development of the data reduction techniqueA complete discussion of the thin skin test described above is valid for only a limited time since

technique as applied to a particular test situation the gage has a finite depth. Another factor which can(including presentation and analysis of the resultant limit the accuracy and/or run time is lateraldata) is contained in Ref. 5. conduction effects. The combined thermophysical

properties of Chromel and constantan are quiteCorapilicag s wsimilar to stainless steel. Thus, if mounted in a steel

For applications where surface contour is critical model, the conduction errors usually become

or where measurement at severe wall temperature negligible for run times less than about five seconds.

conditions is required, the coaxial thermocouple gage The extremely fast response of these gages (50(or coax gage) is often used. The coax gage (Fig. 36) psec to a step input) makes possible their applicationis simply a surface thermocouple which is comprised to continuously moving models or movingof an insulated Chromel wire fixed concentrically components of models.within a constantan jacket. The thermocouple junc-tion is formed at the sensing surface by blending the The sketch in Fig. 37 illustrates the installationtwo materials together with a file. This filing process of coax gages in the body flap of the orbiter modelis also used to contour the gage surface to exactly which was 'ised to evaluate the continuous-sweepmatch the model surface. Because of its simple concept. The results presented in Fig. 38 are typicalconstruction, the coax gage can be made very small; of data obtained during testing of the orbiter bodygage diameters of 0.165 cm (0.065 in.) and 0.318 cm flap and illustrate the excellent agreement obtained(0.125 in.) are in common use. Also, because of its

22

AEDC-TR.-94-

between sweep measurements and discrete (fixedangle) measurements. These data are from the same .-

gage operated in both a discrete (fixed point) and 0 DATA o0Al INIMD VIA sWE EP MOsweep mode. Such results indicate that reliable heat- a 1 DA OTAINED VIA STANDARD MODE )

transfer data can be obtained if certain precautions - 4 (ON FAP ANGLE PER TUNNEL INJECTION) , 0

are followed. In particular, the data acquisition rate, x 12 00the model sweep rate, and the method of data filter- 1 0

ing (or averaging) used to reduce data scatter must I Lo UNCETAINTYbe coordinated to meet test data quality require- . 6 40ments. Also, care must be taken to avoid violating 4 o

the assumption of a model wall thermal response 2 2

identical to that of a semi-infinite solid. This implies 0 5 0 I 2that sweep data should be obtained only during 1 0 F 10 is 20

moderate (less than 5 sec) aeroheating exposures. For F

further details regarding test results and analysis ofthe advantages and restrictions of this technique, see

Ref. 6. standard modes.

Figure 38. Concluded.ANGLE OF ATTACK

ID! FLAP self-generating output signal directly proportional toCOAX incident heat flux, continuous service temperature of

I•NSTALLED 7000F, and semicontourability.

f 11.1 N. The principle of operation of the Schmidt-Boelter

+24 gage is based on axial heat conduction, and involvesFLAP HINGE measuring the temperature difference, AT, between

ANODIZED ALUMINUM WAFERFigure 37. Illustration of flap sweep and gage H 0.002-IN'-DIAM (0.05 mm)

*installation. CONSTANTAN WilE

Schmidt-Boelter Gage -T - 1, - Tc

The Schmidt-Boelter gage (Fig. 39) alleviates _.many disadvantages found in other gages. Schmidt- CONSTANTAN WIE ELECTROPLATED

Boelter gages have seen considerably wider usage in IN - TOP SURFACETEMPERATRI

recent years for heat-transfer measurements in - IT=TOgI SURFACEcontinuous-flow wind tunnels and flight test TEMPERTURE

applications.7 This is primarily due to the'attractiveoperating characteristics common to this type sensor.These include excellent durability, good sensitivity,

INSERT HETSNH- INTO FLP -HEAT SINK

TUNNEL SWEEP25

~20D

5

-* a. Schmidt-boelter gage concept

-n0 1 TIE2 S( 3 4 Figure 39. Construction details of Schmidt-Boelter

TIME SEC gage.

a. Body flap sweep profile

Figure 38. Continuous sweep coax-gage data.

23

AEDC-T 4

"21 fi 1.u04Nu €OSTAlITAN Representative data from Schmidt-Boehaer gagesWME EUCTEIWED WIN CMillON MKAL OF COIL (IN1Y t0IlE installed in a flat plate model are shown in Fig. 40.

S2.2 n. 1tinl v- 13S MAI.W09M0 ala Wra M. klI m, I v2.6 I. RAi. 14M M R

•-~W MO -. 0.3 K6 ra i

If Is 2-0.O1. AXIAL DISTAN2E FIN 6 L EM, X

L AL*U.I I

SHEAT UIK Figure 40. Axial heat flux distribution on a flatplate model.

O-VIlAPPING OF

aT MEAUREMIT EPOXY \\MI. AND

WEASIUMH

b. Section drawing of 3/16-in.-diam Schmidt-Boelter

Figure 39. Concluded.

UM VIEW Of ENSING FOWLtwo parallel planes on the top and bottom of a slab /-ENMl FOIL

or wafer which is backed by a heat sink, as shownin Fig. 39. This temperature difference is generallymeasured with a differential thermocouple. The hot WMI lM-junction temperature, TH, is on the top surface of €O MI TUMI"--the slab, and the cold junction temperature, TC, is N ILtMon the bottom surface. The material and thickness (Ol, NUT Simof the slab can vary widely; the heat sink is usuallya material with a high thermal conductivity such as l to*.AUIMt* Ialuminum, copper, etc. Excellent sensitivity is fIEINWIEachieved by using a series thermocouple (thermopile)arrangement to detect the temperature difference ALL WANS=I II INOCSbetween the top and bottom surfaces of the slab. The Figure 41. High-sensitivity gardon gage.thermopile is constructed by winding 20 to 40 turnsof small (0.002-in.-diam) constantan thermocouple- Gardon Gagegrade wire around the anodized aluminum wafer.One-half of the constantan coil is electroplated with The Gardon gage differs from other types of heatcopper, creating a multi-element copper-constantan gages because its principle of operation is based ondifferential thermocouple. The steady-state output lateral heat conduction.signal of the transducer is proportional to the incidentheat flux at the surface (). There are two types of Gardon gages. The

standard Gardon gage output is the result of a

Experimental calibrations of Schmidt-Boelter junction formed by a thin copper wire connected togages are performed using a radiant heat source as the center of a thin constantan sensing disk on thedescribed in Ref. 12. These experimental procedures gage face. The other type of heat gage (thermopileenable a calibration scale factor (CSF) to be obtained Gardon gage) derives its output from overlappingfor each gage and heating rates are calculated as antimony and bismuth deposits, forming a thermo-follows: couple which senses the temperature gradient on the

back side of the sensing foil (see Fig. 41). This

= (CSF) (AE) (10) arrangement greatly increases the output of these

24

AEOC-TR-94-6

gagcs as compared to the standard Gardon gage, and Heating distributions on the wing leading edge offor this reason it is more commonly used. a 0.025-scale space shuttle orbiter were measured

during a recent test. Approximately 6 in. of the wingWith q and T, obtained from gage leading edge of the model was replaced by aninstru-

measurements, the heat-transfer coefficient can be mented insert as shown in Fig. 42. The insert wascalculated as previously stated. fabricated as illustrated in Fig. 43. MACOR! 9658

41 (a glass ceramic material produced by Corning) wash = ( ()1) selected as a substrate because of its machinability

(Tr - Tw) and good thermal insulating properties. The

MACOR substrate was machined to an external con-Unfortunately, the recovery temperature is nottour corresponding to the wing leading edge with a

alwas kown nd ts eterinaioncan ecoe a channel in the back side to accommodate routing ofsignificant factor in utilizing test data, particularly channel iead side to smate roles offorelectrical lead wires. Fifty-four small holes weretheforsituations wehere T as b d00 Howevelop or, drilled normal to the surface of the test article intothe following technique has been developed for

determination of Tr for these cases. Utilizing the the lead wire channel on two angular rays to provide27 thin-film sensors 0.76 cm (0.30 in.) long with

continuouswegage: o oaspacings of 0.25 cm (0.10 in.) between them. Chromel

pins were cemented into the holes, with the topc of

*l = hTr - hTw (12) the pins protruding slightly from the sensing surfaceof the test article and the back of the pins extending

Equation (12) has the form of a straight line when well into the lead wire channel. The sensing surface

h and Tr are assumed constant (which is valid for this of the test article was then polished smooth (- 12

application) x 10-6 cm roughness) so the Chromel pins wereflush with the sensing surface. Thin-film resistance

S= A + A, Tw (13) thermometers of Hanovia Liquid BrightO platinumwere painted across the 0.76 cm (0.30 in.) space

Also note that h = - A, and setting 4 = 0 leads to between the pins. A silver preparation was then

the relationship for Tr, painted over the Chrome[ pins and in the areaimmediately surrounding the pins. The entire

Tr = -AO/AI (14) assembly was then successively fired and cooled

A description of the Gardon gage techniquedescribed above as applied to a specific test situationis included in Ref. 8.

Thin Film

Standard methods for obtaining convectiveheating data (i.e., heat gages, thin-skin thermo- ". .

couples, etc.) on wind tunnel models are restricted

to model regions with relatively flat surfaces and/or a. View of exposed surface of instrumented insert.small surface heating gradients. These limitationshave precluded obtaining accurate heating rate mea-

surements in the most crit,.al areas of a vehicle, suchas the wing and fin leading edges, nosetips, inter-ference flow regions, etc. By applying small thin-filmresistance thermometers to a contoured ceramic sur-

face (Fig. 42), a technique has been developed 9 formaking measurements in these regions, and a philo- b. View of backside of instrumented insert.sophy for data reduction was derived which allows figure 42. Space shuttle orbiter instrumented wing

application of this technique to models tested in leading edge.continuous-flow wind tunnel facilities.

25

AEOC-TR~4

| i i ! ;i;s;

ALL IMENSIONS IN INCHES 0 'fI-4tt 1 +4t-~

mrTrON VIEW10NOE tRA•a I U-KION

c. Sketch of instrumented insert installation in orbiter CF. t EXPOhU E TIME)

wing i0 a MFigure 42. Space shuttle orbiter instrumented wing

leading edge. S0. p I I

0 1 2 3EXPOSURE TIME, SEC

MAWINAILE Figure 44. Representative thin-film data.0.3 I CERAIC SUISIEOATE1.76 -(MAWI)

FLW V__ B

ROW lo SIM INSTRUMENTED WINGLEEUWINE WIWAING

SIMC INTERACIONPILADmIm ai0l l .1FILM (TV.) PlE I

,..OUTW EDGE Of low soCSINTERATION HEATING PEAK

0 000Figure 43. Typical thin-film installation. 0o 000000 0r 0.01 0O 0O

through several cycles to put a permanent set in the -

resistance vs. temperature characteristics of the thinfilms. After the firing, the area immediatelysurrounding the pins was polished to a smooth finish.

Copper lead wires were soldered to the pins on the O.0nI I I

back side of the test article. 0.5 0.6 0.7 0.8 0.9 1.0VBn

Each film provides a measurement of the surface

temperature response of the wing leading edge during Figure 45. Typical thin-film data application.

exposure to winid tunnel conditions. A sample of the Gag Calibration

temperature transient results is presented in Fig. 44.

The heat-transfer rate is calculated from the tempera- Experimental calibrations of heat flux gages are

ture transient curve via semi-infinite solid response performed using a radiant heat source. These experi-

considerations, then a correction is applied to account mental procedures enable a calibration scale factorfor the effect of the small local radius of the surface. (CSF) to be obtained for each gage, allowing heating

The correction technique is based on finite-element rates to be calculated. AEDC has recently developed

modeling of the heat conduction in the wing leading a high-heat flux gage calibrator (Fig. 46). Through

edge and was verified by a series of experiments on efficient collection of the radiant energy from a high-

cylindrical leading-edge pieces. Typical results from intensity xenon arc lamp, a lab calibration apparatusthe space shuttle orbiter wing leading-edge tests are (Fig. 46a) delivers up to 1,500 w/cm2-sec. Emphasis

illustrated in Fig. 45. has been placed on uniformity (Fig. 46b), and

repeatability. Continued development is underway to

26

AEDC-TR-9P-6

calibrate a gage at elevated temperatures to account comprehensive pretest planning, are essential for afor gage material property variations with successful test program. Designers and system

temperature. engineers are strongly encouraged to contact test

GAGE HOLDER engineers at the facilities for the most recentGAGE developments and improved test capabilities.

FAST RESPOSE -SHUTTER .REFERENCES

SOLENOID OPERATEDLENS - MEDIUM SPEED

SHUTTER I. Bernstein L. (1975) "Force measurements in0= •short duration hypersonic facilities." AGARD-

ograph 214.

MANUALLY OPERATEDWAFER SHUTTER 2. Matthews, R. K. and Gilley, G. E. "Reduction

-XENON LAMPHOUSE of Photographic Heat-Transfer Rate Data atAEDC." AEDC-TR-73-90 (AD-762928), June1973.

FLOOR 7 7 7 3. Boylan, D. E., Carver, D. B., Stallings, D. W.,

LAII FLOOR and Trimmer, L. L. "Measurement and Mapp-a. Apparatus ing of Aerodynamic Heating Using a Remote

Infrared Scanning Camera in Continuous Flow

256 Wind Tunnels." AIAA 10th AerodynamicDATA TAKEN AT EXIT PLANE OF OPTICAl. INTEGRATOR Testing Conference, San Diego, CA, April

192 19-21, 1978.

12 .4. Trimmer, L. L., Matthews, R. K., and

S644 Buchanan, T. D. "Measurement of Aero-0 dynamic Heat Rates at the von Karman Gas

Dynamics Facility." International Congress on-64 •Instrumentation in Aerospace Simulation Faci-

-1 --- lities, September 1973.

5. Carter, L. D. and Kaul, C. E. "Heat Transfer

., 0 _t -Tests on the Rockwell International Space&.510 Shuttle Orbiter With and Without Simulated

b. Heat flux uniformity Protuberances." AEDC-TR-75-20 (AD-

Figure 46. Capability for high heat-flux calibration. A012876), July 1975.

6. Wannenwetsch, G. D. and Martindale, W. R."Roughness and Wall Temperature Effects onBoundary Layer Transition on a 0.0175-Scale

The complexities of hypersonic vehicles challenge Space Shuttle Orbiter Model Tested at Mach

the experimentalist to develop new facilities and Number 8." AEDC-TR-77-19 (AD-A038895),innovative test techniques. This section briefly April 1977.

discussed the large variety of ground testingtechniques that can be used in conjunction with 7. Matthews, R. K., Nutt, K. W., Wannenwetsch,

analytical tools to develop hypersonic flight vehicles G.D., Kidd, C. T., and Boudreau, A. H. "Dev-and demonstrate their survivability and performance. elopments in Aerothermal Test Techniques atThe variety of test techniques available to the the AEDC Supersonic/Hypersonic Wind

experimentalist continues to grow, and the impor- Tunnels." Vol. 103, AIAA Progress intance of his selection is compounded by rising test Astronautics and Aeronautics, 1986.

costs. There are many factors that influence theselection of a test technique, but clearly the number 8. Matthews, R. K. "A Summary Report on Store

one consideration must be the test objective. A Heating Technology." AEDC-TR-78-46 (AD-

precisely defined test objective, coupled with A059415), September 1978.

27

AEDC-TR444

9. Wannenwetsch, G. D., Ticatch, L. A., Kidd, II. Jones, R. A. and Hunt, J. L. "Use of Fusible

C. T., and Arterbury, R. L. "Measurement of Temperature Indicators for Obtaining Quanti-

Wing-Leading-Edge Heating Rates on Wind tative Aerodynamic Heat-Transfer Data."

Tunnel Models Using the Thin-Film Techni- NASA TR-R-230, February 1966.

que." AIAA Paper 85-0972, June 1985.12. Kidd, C. T. "Determination of the Uncertainty

10. Nutt, K. W. "Thermographic Phosphor of Experimental Heat-Flux Calibrations."Technique." von Karman Institute for Fluid AEDC-TR-83-13 (AD-A131918), August 1983.Dynamics, Lecture Series 1990-05, June 1990.

28

AEDCTR-4)4

MATERIALS/STRUCTURES TESTING

byR. K. MATTHEWS

Senior Staff EngineerCalspan Corporation/AEDC OperationsArnold Engineering Development Center

ABSTRACT p Pressure

The requirements for developing hypersonic flight P0 Total pressure

vehicles place increasing demands on ground test q Heating ratecapabilities. Of particular concern is the requirement To Total temperatureto demonstrate that flight components such as lead- T. Wall temperatureing edges, cowl tips, and structural panels will survivethe aerothermal flight environment. Specific corn- VP Particle velocityponents can experience very high heating rates and Q Material density

surface temperatures from 1,100-1,9500C (2,000 to 7 Shear3,5000 F). In the U.S., ground test of flight compo-nents have previously been performed at test facilities • Micronslike those at NASA and AEDC. This section presents Subscripts

an overview of the materials/structures test methodo- L Local conditionslogy used in the development of hypersonic vehiclecomponents. INTRODUCTION

NOMENCLATURE In the development of hypersonic vehicles, struc-tural survivability is of fundamental concern, speci-

c Specific heat fically for components that comprise the external

dp Particle diameter, microns surface of the vehicle (see Fig. I). For the purposesk Conductivity of these notes the term "structural components" in-

cludes the following:LDV Laser doppler velocimeter * fuselage, wing, or tail panels up to 6 ft (1:8

M Mach number m) long and 4 ft (1.2 m) wide• protuberances, gaps, joints, seals, and the

Nd Particle number density surrounding structure

N I L CM I nKUAI~l5111•II NllI~l tA I

Figure 1. Development of vehicle structural integrity.

29

-;~~WONM OF TO IMMM RUMi~lmn ~i

AED•TR4h-

* nosecaps PHASE 2 - DEMONSTRATE HARDWARE SURAIIUTY (STEIPS , 3, 4)

* wing leading edge sections* cowl lip sections MATERIL TEST STRUCTURACONTEPT FUWI HARDWARETEST KNOl TEST* control surfaces (SAMPLES) COMPOiENTS) (PNENTS)I

* fins* inlets DUPIJCATE LOWAL TEST tATKiES

* radomes p.s. q LO(w. , FIT)* antenna windows

STEPS(2) SELF(T MATERIALStructures larger than those listed above (3) TES SEULEC M ER

would be too big for a high-enthalpy air- (4) PEO FLIGHT HARDWARE VERIFICATION TESTS

flow facility and typically are tested in astructures test facility using radiant heat-

ing. A structures test facility (no-flow) is - ESTIMATE FUIEHT ENVIRONMENT * FLIGHT (OMPONENT (01 LARGE* EN61 COES SIZE REPICA•) EXPOSD TOalso best suited to perform life-cycle a SULTED FLIGHT ENOSMED NT

testing. • THERMAL RESPONSE COOES IN TEMS OF TEIIPEATURE, HEAr(TIEMPIERATURE) RATE, SHIAM, PRESSURIE MOI LOAD0 STRUCTURAL LowADS AEK - APTU

In contrast, airflow facilities can STRAIN e AEKM AEOTHERAL TUNNELSprovide a better simulation of flight and * LaC 5-FT TUNNEL

* AMES 3.5-fl TUNNELa more comprehensive evaluation of S ARC FACILITIES SEEpotential failure modes. An obvious • STEP I IN PREEEDIN SECTIN * AES, K. MDNW, it TABLE I

example is the case where the hot airflowleaks into the substructure and exposes Figure 2. Methodology for aerothermal structures/materials

low-temperature elements to very high development.temperatures. In addition, an airflowfacility is needed to test components forwhich the primary failure mode is related

to aerodynamic heating, shear, acoustics,or vibration.

AEROTHERMAL METHODOLOGYMATERIAL SURVIVABILIYIS DETERMINED IN OUPLIATEO

A common approach used to solve FLIGHT ENVIRONMENTaerothermal issues is based on combininganalysis with experimentation. It is im-perative that analytical techniques be used TEST INSTALLAION

to plan the test and to analyze the finaldata. The two fundamental phases in the MATERIALS SUIIVAlILITY

IS A FUNCTION Ofdevelopment process are:

T, WALL TEMPERATIUE & ACOUSTK ENVIRONMENT1. defining the flight thermal en- -r SHEAR SPALLINGIEROSION

at THERMAL SHOCK PARTCLE IMPACTvironment q HEATING RATE PRESSURE

2. demonstrating hardware sur-vivability Figure 3. Materials testing in wind tunnels.

[the first phase of this processwas discussed in the preceding characterization of materials requires many hours of

section] laboratory experiments. Despite these efforts, it is

The steps and approach in accomplishing this often difficult to predict material failure modes in

second phase are illustrated in Fig. 2. The general a flight environment. Material survivability (Fig. 3)

approach is to use analysis tools to design "smart can be a function of many variables. Wind tunnels

tests" that simulate the flight environment, and then and arc-heated facilities often use the wedge testing

to expose flight components to this environment in techniques' to produce local flow environments that

ground test facilities. Material selection is perhaps simulate flight. The primary test results are test article

the most difficult and the most important step. appearance after the run (i.e., survivability). An

Determining thermo-physical properties and illustration of the wedge test technique is shown

30

AEDC-TR-04-

of interest. The data can be developedSIMULATED RENT EININMIINT TES ARICLE into an algorithm that characterizes the

MtH.SK INSTRUMENTATIOM: T1111000pij test article.

W, LUt STRAIN GW •Component survivability tests basi

- Tcally answer the question "Will com-

ponent X survive environment Y?" Thedisadvantage of this type test is that no

'5' quantitative data are produced for

AIDE THERMAL MDOITIONING .0&"6ther" flight conditions.LEADING E06E TO 1'"PATENl" FlI6T NIIAAUENVIRINIENTTEST ARIE (IN, MAY K REWIRE) Component thermal response tests

Figure 4. The wedge technique for aerothermal struc- consist of a fully instrumented test article

tures/materials testing. exposed to a wide variety of conditionsso that a relatively sophisticated math

model can be developed. The math model can be usedin Fig. 4. The simulated flight environment may be to predict internal component temperatures for a

priomarilyccharacterizedmbynheatinmprateuresfwallprimarily characterized by heating rate (€]), wall variety of flight profiles. The disadvantage of thistemperature (Tw), local surface pressure (p), and technique is that it is only valid for the specificslear (7). To produce this environment and to component that is used to generate the database. If

provide changes in the para-

meters, it may be necessary to ADVANTAGES DISADVANTAGES

pitch the wedge (e.g., - 5 to 25 . satEEN TEST RELATIVELY SIMILE, MNY TEST ONLY RELATIVE RAKIUNG •EEN

deg); of course, run time (or ex- ARTLES RUN IN SNORT TIM TEST ARTIES

posure time) is also a key test . CNWMMTEATION ALLOWS EVALUATION AT SELKTED REDUIRES TESTING OVER A

variable. FIHT CONDITIONS LRANW OF SEVERAL PARAMETERS,i. Tw, SHMEA

The test article may range * WOM SURVIVABILITY PROVIDES '11' OR 'WO' ANWEKR NO OUANTITATIVE DATA FOR

from a simple insulated material FOR SIMULATED FLIGHT CONDITION 'OTHER' FLIGHT (ONDITIONS

panel to a complex structure with * COIMONEIT TNHERAL RESPONSE CAN POV'IDE TEMPERATURES OIR A DATA ONLY VALID FORVARIETI OF SIMULATED FLIGHT SPECIIK COMPONENT TESTED

LH 2 cooling passages. In CONDITIONaddition, it may be necessary to ---

provide structural loading on the Figure 5. Types of aerothermal structures/materials test techniques.

test article using hydraulic actuators mounted under the component design changes after the test, it maythe wedge. If the test article incorporates backside be necessary to rerun the entire test for the new

cooling, it is important to simulate both the design. Examples of these test techniques are

aeroheating (4 n) and the heat removed by the discussed below.

coolant (4,ut). Therefore, the mass flow, heat

capacity, temperature, and viscosity of the coolant Illustrative Examplesmust be duplicated in the ground test.

A materials screening test was conducted on high-In general, materials/structures testing can be temperature polymeric and Boron-Aluminum

grouped into the four categories listed in Fig. 5. composite material specimens in AEDC Hypersonic

Screening tests make relative comparisons among Wind Tunnel (C), which is a continuous-flow, Mach

many candidate test articles in a constant (or 10 wind tunnel. Test specimens were attached to a

repeatable) test environment. It is desirable to wedge model adapted to serve as a specimen holder.

fabricate several identical test articles to investigate The use of a large wedge as a holder for material

the repeatability of failure modes. samples creates the two test regions illustrated in Fig.

6. Region I is limited in height by the distance of the

Characterization tests are the inverse of screening bow shock above the wedge boundary layer at thetests in that one test article design is exposed to a aft end of the wedge. This distance is about 10.9 cm

variety of test conditions. For example, the test (4.3 in.) for the 105.4 cm (41.5-in.)-long wedge usedvariables may be temperature and load, and each is in this test. Region I provided relatively uniform flow

held constant whiie the other is varied over the range for testing the curved surface samples that were

31

AEDC-TR4-4

wit. sicI ~ q 1l~xggg -2e

LARK! SAMPtES , 7. 3

TEST ShMftlS 4. 5

Figure 6. Sketch illustrating wedge flow-field

nomenclature. .,,

exposed to direct impact of the local flow (see Fig.7). Much larger samples can be tested in Region il.(In fact, samples up to 63.5 cm (25 in.) long have TEST SAMES 6, 7, 8.9

been used.) In Region 11, duplication of local shearstress, rl, is probably the most relevant parameter -1for the given total temperature level, To, whereas in 14Region 1, duplication of total pressure, PoL, totaltemperature, ToL, and Mach number is of primary UT SOMS 1interest. The desired flow conditions on the materialsample were produced by adjusting the wedge angle. Figure 8. Photographs of curved-surface specimensThe oblique shock wave generated by the wedge was during run.used to reduce the local Mach number on the wedgesurface to the desired supersonic level (M - 4). The from the leading edge. The other types of materialstunnel stagnation conditions were adjusted to were relatively good thermal insulators; as a result,produce the desired local pressure and temperature the surface temperature rise was very rapid. The datalevels. The fact that the Tunnel C clean air heater from this type test provide the necessary informationcan provide temperatures up to 9820C (l,800°F) is to select the best material for survival in the simulatedthe primary reason why this facility can be used for flight environment.this type of testing. The test specimens consisted of51 material samples: 39 curved configurations of The second example of materials testing is theseveral radii and thicknesses, and 12 flat panels. A characterization of recession in a high-enthalpysketch illustrating how the samples were installed on facility. Thermal protection of hypervelocity flightthe wedge is presented in Fig. 7. vehicles is often accomplished by using materials that

ablate during exposure to the severe environment.Ct•DED-fME •.The outer surface ablates away while the innerATEIL .UPLE$S surface remains relatively cool. The ideal material

would allow no heat to penetrate to the insidesurface, be very lightweight, and ablate very slowly.

ATOU PLATE The ablation rate, or "recession rate," is a criticalANDOH FLAT SAMPLES parameter in the selection of materials that can

PHENOUC SPACER survive exposure to the high-enthalpy flowenvironment of hypersonic flight vehicles.

Figure 7. Sketch of material support techniques. The Recession Rate Monitor (Fig. 9) operates on

Typical photographs showing the effect of the a triangulation principle (Fig. 10) where a focused

simulated Mach 4 environment on the curved panels laser beam is projected at a given angle onto a re-

are presented in Fig. 8. All but one of the samples ceding surface. When viewed from another angle, thewere glowing red-hot in the relatively short exposure laser spot shifts in the field of view in an amount

times of 12 to 20 sec. This sample was the best con- proportional to the surface recession. The measured

ductor; as a result, heat was conducted away

32

AEC-TAR-94-

produce profiles, the laser beam is rapidly scannedacross the surface, providing recession data from

many points (or a continuum of points) along thescanned line (Fig. 11).

S~LOIB WW The Recession Rate Monitor has been used in

ROM AEDC's HEAT-H I facility, 2 which uses a segmented

type arc heater. High-pressure air is heated by means11N111 aof an sustained arc plasma. The heated air is

UM SWaccelerated through a water-cooled nozzle to producea free-jet test environment.Sacelranged f hrom g 1. to12atmerndcold walzlheat flouxe

A representative test article installed in this facilityis shown in Fig. 12. Axial centerline surface pressuresranged from 1.5 to 12 atm0 and cold wall heat flux

Sranged from about 500 to 2,700 w/cm2 sec or

Figure 9. Recession rate monitor system. (Btu/ft2-sec). The test articles were carbon compositepanels 5 x 10 x 0.9 cm (2- x 4- x 0.36-in.) thick.

LtEo OBSERV For each test run, five test articles were mounted onthe radial inject struts of a rotating sector. Duringa run, the test articles were sequentially subjected tothe free-jet flow for about 3 sec each.

Recession data for one of the test articles areshown in Fig. 13. This figure can be viewed as a cen-

- 1CESS!N_&_ ------- -terline cross-section of the test article (greatlyDISPLACED SURFACE exaggerated thickness) with the upstream end to theFigure 10. Recession rate monitor operating left and the exposed surface up. The "pretest

principale surface" line identifies the location of the pretest

Di •the solid line curves in the figure represent the surface

profile obtained by the Recession Rate Monitor atthe corresponding times. The surface profile definedby the circles was obtained from a posttest

:S measurement of the surface made with a contact-typemeasurement device. The lowest of the family of solidline curves in the figure represents the profile

• •acquired by the Recession Rate Monitor just beforeORIGINAL POSITION DISPLACED POSITION the test article was rotated out of the flow.

Comparison of this curve with the post-testFigure 11. Profile measurement technique. measurements indicates an agreement of

USTl unU within ±0.013 cm (0.005 in.). This excellentCUM x 4M xM T agreement demonstrates the quality of mea-

sm surements obtainable with the RecessionRate Monitor3 and demonstrates the char-acterization of the recession rate of anablation test sample.

The third category of material/structuretesting is "Component Survivability," andan example of this test technique is presented

in Figs. 14 and 15. In the specific exampleFigure 12. Test article installation, the thickness of the shuttle tank insulation

material was the primary area of concern. Thisshift, multiplied by the calibration constant, yields material is a low-density Spary On Foam Insulationan accurate recession measurement at that point. To (SOFI) whose material properties (Q, c, k) are known.

33

AEDC-T-04-

The use of analysis tools to design a test

is illustrated in Fig. 14. An "aerothermal

ON.-Ij Ill " lNEWMa$S MollaM 10 rq TM INmg[nM response code" combines the material pro-MIOI0t P14 perties, the flight trajectory, and other inputs

0.4 with a heat conduction model of the testS1XI0M11E UK, 9[{ MIE•ST UWAIXO.

0.gs . , article to provide a prediction of surface

0.3 1.13 temperature versus time. A similar code11.1 6 17$51 . flCTIN MUL combines the results of the wedge calibration

data (i.e., i versus WA*) and the facility3.2 2.34 flow conditions to produce the test article

Z.I0 surface temperature predictions during the.1 3.1111 wind tunnel test. In this manner, the wedge

"TA FRM TEST IN HEAT.HI, WINEV MCL angle can be adjusted until the temperatures

agree as shown in the figure. For the test

0 ........ article to reach the predicted temperature

0151*K MM MML 11RFAG, IN. and temperature gradients, run times ofmany minutes may be required. Conse-

Figure 13. Characterization of material recession rate. quently, impulse facilities cannot be used forthis type of testing.

The primary data from this type of testare the appearance of the test sample afterexposure to the simulated flight environ-

Mrea i Ltt ment. As shown in Fig. 15 the SOFi didMIMOU / MW survive the test; however, it did experience

Smmi some recession, as was expected.ill

The fourth type of material/structuretest technique (thermal response) is far more

, complex than those previously described,,_______ ,___ and only a few of this type have been

i. SK conducted. Therefore, a similar discussionwill not be presented, however, Ref. 4 is an

Figure 14. Flight trajectory thermal response simulation. excellent example of this category for thosewho are interested.

WEATHER/EROSION TESTING

The requirement for all-weather flightcapability emphasizes the need for testingtechniques that evaluate particle impacteffects on vehicles operating at supersonicthrough hypersonic speeds in particle-laden environments. In particular, theeffects of raindrops or other particles onradomes, heatshield materials, andantenna window materials need to beevaluated over a wide range of flight

Figure 15. Shuttle external tank insulation material test. * WA - Wedge Angle

34

AEDC-TR-84-

speeds. The flight speed regimes and the strong function of the radome material, vehiclecorresponding test facilities are divided into the three speed, number of impacts, and the particle mass.classes listed below Historically, weather erosion tests of material

samples (coupons) and of full-size radomes have beenImpact Velocities, ft/sec Facility conducted at sled tracks equipped with rain

(km/sec) simulation apparatus. In the U.S. the Holloman sled

1,000-3,000 (0.3-0.9) Wind Tunnels/Sleds track5 is perhaps the best known of these facilities.

3,000-6,000 (0.9-1.8) Arc-Facilities In recent years, a new long-duration, multiple impact

6,000-20,000 (1.8-0.6) Aeroballistic Ranges capability has been developed 6 utilizing Tunnel C at

Each of these regimes is briefly discussed below. AEDC. The basic principles of this new capabilityare illustrated in Fig. 17. The initialdevelopment of a precipitation en-vironment simulation in a wind tunnel

°. considered the injection of ice particles

or water droplets into the tunnel air-

flow. Analytical studies indicated this

approach to be impractical for waterdroplets, since high shear forces would

SINGLE IMPACT MULTIPLE IMPACT cause disintegration of the droplets.4i 'The technique to inject ice particles

was derived from an injection method

proposed by Allen. 7 The ice particles

are suspended in a liquid-nitrogen"carrier and injected into the tunnel

* • . .flow upstream of the supersonic

-0.25 IN. ,i nozzle. As the flow field expandsthrough the nozzle, the carrier flashes

Figure 16. Missile radome in flight encountering rain/ice to a gaseous state, which causes veryenvironment, little effect on the flow constituents. During lab-

oratory checkout of the injection system, 500-IL iceFREE JET NOZZLE particles were discharged at rates of up to 5 x I05

CHAM1II particles per sec for durations up to 60 sec.

PARTICLE INam The Mach 4 Aerothermal Wind Tunnel (C)

COLO SCAEE used in the development of the precipitation en-tRADOME vironment is a closed-circuit, high-temperature,

supersonic free-jet wind tunnel with an axisym-

INSULATED PRESSURE-.. metric contoured nozzle and a 0.63 m (25-in.)-VESSEL .diam exit. The tunnel operates continuously over

FLOW REMIS a range of pressures from 15 psia to a maximumICE - 0.03 LgiMSE(LN 0.30 L•M/SEC LN(ICE _A of 180 psia (l-12 bar) and at maximum stagnationRUh DUR'ATION -60 SEC SLURRY "'temperatures up to i,900'R (1,050°K). The test

Gil FOR MIXING AND unit utilizes a model injection system which allowsPRESSUIRIZATION the test article to be withdrawn from the test

Figure 17. Aerodynamic particle erosion test section while the free-jet tunnel re'mains intechnique. operation. The test specimen is mounted on a sting

support mechanism in an installation tank directlyImpact Velocity 1,000-3000 ft/sec (0.3-0.9 m/sec) underneath the tunnel test section and is then injected

into the particulate flow. Data are obtained with aThe erosion caused by a single water drop pulsed holography system, the LDV, and the photo-

impacting a radome is shown on the left side of Fig. graphic systems (Fig. 18) in addition to the standard16. The right side of the figure shows the erosion that tunnel system during the exposure process. At acan result from multiple impacts. It is clear that flight selected time, the test article is retracted into the tank,through weather can cause significant damage; how- the particle injection system control valve is closed,ever, it must be emphasized that the damage is a

35

AEDC-TR444

one-dimensional particle acceleration code. Figure 20shows a demonstration test radome located on flowcenterline approximately 6 in. downstream of the

laser diagnostics systems during a precipitationenvironment exposure. This full-scale flight hardwarcwas exposed to the precipitation environment for 20sec and, as can bc seen approximately half of theRTV material was eroded away while the other side(candidate material) survived the multiple impactenvironment. A point that should be noted is the useof ice particles to represent water droplets in theprecipitation environment. Although ice particles

Figure 18. Wind tunnel test section area. make up a major portion of the flight environmentof interest, a portion of the environment exists where

a simulation of rain droplet impacts would be re-and the system depressurized in preparation for the quired. AEDC has conducted a brief study tonext exposure. The laser diagnostic systems provide determine the correlation between ice and waterparticulate cloud characteristics during the run. Ice impacts at about 762 m/sec (2,500 ft/sec). For theparticles of a size and concentration representative limited data obtained, there was negligible differenceof a precipitation environment are injected into the between the ice and water impact craters; however,airstream and accelerated by aerodynamic drag to this conclusion cannot be generalized for all materialssupersonic velocities. Characteristics of a particulate of interest, and each situation needs to be evaluated.

cloud before its impact with the test articleare determined with laser diagnostic [MIT ROMsystems. A vehicle operating in a flightregime below 50,000 ft (15 kin) mayencounter a precipitation environment of100- to 3 ,000-p (0.1- to 3.0-mm) iceparticles or water droplets falling at ratesup to 2.6 in./hr (66 mm/hr).

Figure 19 illustrates a typical hologramobtained during an exposure run showingthe two images of a particle from thedouble-exposure technique. Thesephotographs quantify both the size andgeometry of the particles, as well as theparticle velocity. The velocity data fromthese holograms agree to within 10 percent Figure 20. Radome following 20-sec exposure to multiple

of pretest predictions obtained from a impacts.

Impact Velocity 3,000-6,000 ft/sec

The AEDC HEAT-HI Test Unit has the capa-bility of graphite particle injection and accelerationfor combined ablation/erosion testing. This capa-bility is achieved by injecting various size graphiteparticles (60-400 pm) into a chamber just upstreamof the nozzle and allowing them to drag-accelerateto velocities between 3,000 and 6,000 ft/sec (0.9-1.8km/sec). A schematic of the erosion testing techniqueused in HEAT-HI is given in Fig. 21. Dust particleflow rates from 5 to 60 gm/sec can be provided, anddevelopment work is underway to reduce flow rates

Figure 19. Double-pulsed hologram of 500p particle. below the 5 gm/sec rate. The cylindrical nozzle

36

AEOC-TR-94-.

extensions shown in Fig. 21 are added tothe HEAT-HI contoured nozzles toprovide a longer transit time for theparticles to accelerate. A Mach number HEAT SHttLD SAPLE

3.5 nozzle has been used extensively fordust erosion tests, and a Mach number WEDG MODEL HOLDER FLOW2.2 nozzle is also available. Calculatedvelocities for a range of particle diametersof interest are shown in Fig. 21 with the30-in.-Iong dust accelerator installed.Laser diagnostic systems similar to thosepreviously mentioned are available to a. Pre-test photograph.characterize the particle cloud. Repre-sentative sample sizes are 1.9 cm (0.75 in.) [1SILSAPwide, 5.1 cm (2 in.) long, and dovetailfitted into a surrounding silica phenolicheatshield material, which in turn wasdovetailed into the stainless steel holder.Wedge model/holder hardware and a typ- WEDGE MODEL HOLDERical test installation photo is shown in Fig.22. The primary "data" from this typetest are obtained by close inspection of thesample before and after exposure to theerosion environment (see Fig. 22). b. Posttest photograph.

Figure 22. Arc heater erosion test mounted on twenty-five-deggte wedge sample.

.material samples through a specified erosive

environment at a given range pressure and modelNOZLE THkUT GAS VELOCTY NO EXIT velocity. The erosive environment is specified in terms

10,000 of particle type, particle size, and field concentration.S3,0 Glass spheres (simulating dustfield environments),-, 'I OIL PARTICL cirrus ice particles and dendritic snow flakes have

}VntOMY been routinely used. The range pressure and modelM 4,000 velocity are independently controlled to duplicate a

2,000 point of interest on a reentry trajectory.

0 1 The typical snowfield generator installation for

DISIMIE FROM Homn PLAE, IN. the free-flight mode of testing is shown in Fig. 23.These snowfield generators are typically mounted

Figure 21. Dust erosion testing in the arc-heated 0.61 m (2 ft) above the centerline at selected intervalsfacilities, along the range axis. The plates are vertical, but

oriented at 45 deg to the longitudinal axis of the range

Impact velocities 6,000-20,000 ft/sec to accommodate side-mounted cameras which photo-graph the falling snow. The snowflakes are dislodged

For these extremely high velocities, the primary from the copper plate by a mechanical shock devicefacility with this capability is the AEDC (solenoid plunger) at the desired time, and produceHypervelocity Range/Track G. This ballistic range a free-falling erosion field which can be varied in fieldhas the capability to test in the normal free flight length and concentration. A similar approach forballistic range mode or to utilizes a track system to generating the snow particles is used in the trackguide a model through the test environment, mode of testing.

The basic ablation/erosion test technique in The generated snowflakes that are formed whenTrack G is to launch the nosetip or heat shield the crystals are dislodged are shown in Fig. 24.

37

AEDC-TP4-

mm.K SO F "-I' I" FT O Snowfield concentration is con-/NIIGmIAum trolled by regulating crystal growth

time and filtering the falling snow-. flakes through perforated plates.

I . "Since the falling snow stratifies with-I"larger flakes near the head of the

IMtUUIETI1OM CWUhIN slH UIA01l FN fall and smaller ones near the tail,SUWAtLEA Dilm WAOS- JLIOuIED NO A IEUIUTIITI the impacted flake size can be con-

laUmCH jl,]---- 3 PUTE SNW trolled within the limits of model= 99hV- IST I SUIPtRi01 dispersion by appropriate timing of

"_L " ,model arrival. This model-snow-

,LO o4 field encounter at 12,000 ft/sec was--- •--- 1- obtained by utilizing a high speedI / _N 14 multi-frame camera with a framing

WAMI_ OF LUSER STATIS40KlTANS FlR UM ENTAIg, Ftr rate on the order of a million frames

-• per second.Figure 23. Snowfield generator installation in AEDC Range G.

A water droplet impact sequenceonto a track guided model shows

f4% (Fig. 25) that the water dropletremains intact in passing through

the bow shock. But it is also im-portant to recognize the extremely

"short time-frame of this event (i.e.,0.6 pec). It is interesting to note

A, the quantity of debris and addi-tional shock waves being placed into

the flow field by the impact debrisplume.

SUMMARY

A summary of the systems, testarticles, and type of test needed for

Figure 24. Model-snowfield encounter at 12,000 ft/sec. the development of structural/material components for hypersonicvehicles is presented in Fig. 26. It is

important to point out that thisI summary is limited to the topics

briefly discussed in this section, andthat this subject is much broader in

that only "flow facilities" wereT T- 0 P- 1.2 1m discussed. There are, of course,

S I- numerous non-flow facility tests

that are an essential aspect ofstructure/material development.

I- I.I 0,a- 2.4 14 SK

MOEL VELOWfY: 2,460 M/rnUNEPRESSURE: 3S4 toi1.2-m-Am WATER ORIOLT$

Figure 25. Model/water two droplet impact sequence.

38

AEDC;-TI:t-04-.6

S•IIIS IESI MIMES TEST TYMS

• HTPEISONK VENKLES •MOOELS Of PAJITS • KIIEENIN5(• , HiSV) 01F (OOifM41UTION • {• |IIZATIOli

• tACTiCAL MISSILES • STAGNATION RlfdOli • COMPOOIINT •IIWIV401LITY• liE-ENTITY VEHIILES • NOSIS • THEIUL4L lilSPOliSi• DICO• • LEII•ING EDI',• • WlJtHlilJllllO•N]li

• INLET LiPS * Eli T• $10ii• |M ruES * WINOOW •1•

• NON-ST/•iATION li[Gl• * IOItl• l llllOI• WINDWARD PANELS • IL40DWAliE DflJ4)OtSTUTION• LEESIDt PANELS • WAKE (• EIIIZATION• PIIOTUIIIIANCES• IM (ELECTIIOIL45NETK)

Figure 26. Summary of hypersonic ae•othermal tests for the development

of structural/material components.

Table I. Aerothermal test facilities.

NOZZLE STAC41ATION STAGNATIOII FLOW I• (lI.OWNEli FACILITY NAME I•KN SIZE, IN. PIIES•II[, ENTHAL•, UT[. MW

flUMlilli (2.$4 cm/'m.) kT¥ lTU/Lllli I.liM/SI(

NA..•L-JOHNSON SPACE CENTEli AIIMSEAP TP-I 3-12 I0 Io 40 • 10HOUSTON IX77051

• P TP-2 3-6 S 26,000 S

NAU-LANGLIY IIES£JLIKH 5 AIW TUNMfl. 2.0, 2.2 6, 9 DRAM 0,02-0.4 20Q-3,000 SCENTEli, HAMPTON VA23365

S€OMDUSTION.NIAIIED "l.S 13.2• x 13 26 3-12 3OO-SSO 16-40 NIA

SCIIAMJET •STFACILIWS.. ...... • lAW TUNNI• -- 4.2, 4.S, S.O, 0-,-12/I-61 22 b• 0.•-1.3• - 600-SiSO0 20

S.7AIIC-HEATED • 4.7, 6.0 II Idlk, II • 40 1,500 14 20TEST FAClLI•

I • HIGH T• UH• 5.0-7.3 IO • 112 750 NIATUNNEL

ARNOLD ENGINEERING HYPEIISONi( WIND TUNNEL 4, 10, li 2S ANO 50 OL4M ]36 4S0 SO SO

DEVELOPMENT CENTER, TUNNEL, {AliNOLD AFI, TU • ,i•O-ii.Y--- TFTOgi2-- •UPTO IM 1,2M-:Ii•10" 2-1• 4237389-S000

NIGH ENTI4AI.PI" I.il-3.1 11-4.0 IN. ULMt ]00 2,000 2-10 42401.ATiON TEST UNIT(NEAT) allNIGH ENTHALPY 1.8 to 3.$ 75 ro 3 IN. • ---- 120 -- SO0-IO00 O.S-8 • ,10

40LATION TEST UNITS.. .. .... (H• T) NI

AEIOOVNAMH:S AND z.o, 2.2, 2.ss, n-so OIAM 20 ' NiXPROPULSION TEST UNIT 2.72, 3.5, 4.1

S. . .. (•"tu)• G 1-22 (•L) 2.5 I O-S• 0,• NIA NIA

39

• . ,• • i i m nHimi ui • in i i ii ni imi•lU •

AEDC-TR-444

Table 1. Concluded.

OWNER FAIIYME ~ SIZE, STAGNATION STAGNATION FLOW PWROWNR FCILTY AMEIN. PRESSURE, ENTHALPY , RAFE,

NUMBER I. PSIA stun"R LRMJSK m

MARTIN MARIETTA CORP CONTROLLED FLOW I.S, 2.4, 3.2 3.5-26 DIAN 30-2W0 200-220 N/AP. 0. BOX S531 (CONFLOW)ORLANDO FL 32855 COMBIUSTION FACILITY

MCDONNII.I DOUGLAS CORP LARGE CORf ARC 0.9-24.0 DIAM 0.1-100 1,000- 10P. 0. SOX 516 TUNNEL (LCAT) 20,0010IT Louis #4O 63166 FACILITY

HIGH IMPAC 1.7-3.0 0.45-2.0 DIAN 10-250 1,000- 0.01-3.3 12PRESSURE (Hip) #'S00

NASA-AMES RESEARCH CENTER HIGH ENTHALPY ENTRY 1.7 4 DIAN 10,000 INMOFFEITT FIELD CA FACILITY_94035_

2 X 9-11N. TURIUMLINT 3.5 2 x 9 20 4,0W 20FLOW FACILITY

GIANT PLANET FACILITY 1.1 2.IS DUAM 10 40,000- 0.1-0.5 IS____ ___ ___ ______100,000

20 MW PANEL TEST 5.5 4 x It To 4W0- 0.1-2.0 20FACILITY SEMAIELLIPTIC 1IS0W60OMW INTERACTIOIN 5.5, 7.5 1 x32 a36 DIN 3,000- - 0HEATING FAILITY ___20,000

tRlANSITIONAL ROW I.7 AND 2.5 2.3-7.0 DIAN ID 3,000 20-40 100

FACILITYFACILIY 14,003.5 FT NYPEWOMK 5, 7, I0D 3.5 DIAN 133 1 1N/A

REFERENCES 5. Staff, Holloman Air Force Base, "The HighSpeed Test Track Facilities and Capabilities."

1. Matthews. R. K. and Stallings, D. W. 6585th Test Group Test Track Division, Hol-"Materials Testing in the VKF Continuous loman Air Force Base, NM, April 1989.Flow Wind Tunnels." AIAA 9th AerodynamicTesting Conference, Arlington, TX, June 7-9, 6. Beitel, G. R. and Matthews, R. K. "A New1976. Rain/ice Test Capability to Simulate Flight."

AIAA-86-9824, 3rd Flight Testing Conference,2. Horn, D. D. and Smith, R. T. "AEDC High- Las Vegas, NV, April 2-4, 1986.

Enthalpy Ablation Test (HEAT) Facility Des- 7.AlnR.M "A etoofceerigIe

AEDCptRion, 1 Juvlypet an198 irai.. Particles to Hypersonic Velocities in the Labor-AEDC-R-8110, uly 981.atory. " Boeing Report Number D 180-18691 -1,

3. Sherrouse, Peter and Carver, Dwayne. March 1975, Addendum October 1975."Demonstrated Real-Time Recession Measure-ments of Flat Materials During Testing in High-Enthalpy Flows." AIAA-92-0765, 30thAerospace Sciences Meeting & Exhibit, Reno,NV, January 6-9, 1992.

4. Wannenwetsch, G. D., Matthews, R. K. andHowanick, J. "Test Results and 1--redictionTechniques Used in the M505A3E2 Fuze Auto-Ignition Investigation." AIAA-81 -1041, AIAA16th Thermophysics Conference, Palo Alto,CA, June 23-25, 1991.

40

AEDC-TR-94-

BOUNDARY-LAYER TRANSITION

byR. K. MATTHF\VS

Senior Staff LagineerCalspan Corporation/AEDC OperationsArnold Engineering Development Center

ABSTRACT Re,, Transition Reynolds number based on axialdistance x

In the design of hypersonic vehick-: is extremelyimportant to predict the boundary-layer transition ReCF Cross flow Reynolds numberReynolds number. The boundary-layer state (laminaror turbulent) that approaches the vehicle control T Local temperaturesurfaces can significantly affect the control surfaceeffectiveness. In addition, the heating rates (k) U Velocityassociated with turbulent boundary layers are oftenten times higher than those of laminar boundary v Cross flow velocity component

layers. Unfortunately, the methodology to predicttransition has eluded the aerodynamicist for over x Axial distance

three decades, and there are still many unansweredquestions. This section briefly touches on the many Xtbeg Axial distance to beginning of transition

parameters that affect transition and provides numer-ous references for those who are interested in special- Xtend Axial distance to end of transition

izing in this topic. It should be emphasized thatduring wind tunnel testing it is very important toknow the boundary-layer state. Typically, heat-transfer distributions can provide this information; tS Viscosityhowever, it is often necessary to artificially trip theflow to induce a turbulent boundary layer. The Q Densitymethodology of using trip spheres is discussed, andillustrative data are presented. Wall shear stress

SubscriptsNOMENCLATURE OD Free stream

h Heat-transfer coefficient e Edge condition

href Reference heat-transfer coefficient Superscript

k Roughness (or trip) height - Unsteady.

L Length of model or vehicle FUNDAMENTALS

M Mach number

p Pressure Boundary-layer transition is a very complex andspecialized subject, and to be truly knowledgeable in

4w Wall heating rate this speciality requires many, many years of dedicated

Re Reynolds number work. In the U.S. two individuals who have done thisare Morkovin14 and K. F. Stetson 5-14 . Much of the

Rek. tr Transition Reynolds number based on material is this section has been taken directly fromroughness height k their work.

41

AEDC-TR-94-8

Figure I introduces the usual concept of boun-dary-layer transition as consisting of a regionbeginning at the termination of pure laminar flow("transition onset" at X1 , beg) and ending at the

beginning of turbulent flow ("transition completion" -.at Xt,end). The sketches below the figure show thequalitative variation of heat flux or surface shear inthe laminar, transitional, and turbulent flow regions.The relatively sharp change in value of either of these a. Laminarparameters at the beginning of transition offers a

method of detecting transition onset. The mostcommon method in use today for detecting transitiononset in either ground or flight testing is some formof heat flux or temperature measurement.

INVISCID FLOW REGION

U VSCOSFLOW REGO

b. Transitional

I X I I I• .LAMINAR FLOW TRNII sONAL TUI ruLENut".

I REGION I FLOW REGION I FLOW REMI.=.

w 1 4.,

Figure 1. Boundary layer concepts, terminology, and c. Turbulentmeasurement. Figure 3. Photograhs obtained on a sharp 9 deg cone

at M. = 8 in AEDC tunnel B.

- - .. ... - . -< turbulence, etc., that cause transition to bypass theI-7-. ... "linear processes. The photographs of Fig. 3 show

laminar (3a), transitional (3b), and turbulent (3c),RANDOM boundary layers on a sharp 9-deg cone at Mach 8 in

DISTURBANCE (GENERALLY NOT KNOWN FOR AEDC's Tunnel B. The laminar boundary layer isSPECTRUM GROUND O1 FLIGHT TEST) characterized by a white line which has a "roping"

SLOW LINEAR- NON.LINEAR IIO appearance in the transitional regime and eventually0-, AMPLIFICATION RESPONSE L completely disappears in the fully turbulent regime.

LAMINARLAYER "BYPASSES" A Specialist's View of Boundary-Layer Transi-(CROSSFLOW, ROUGHNESS, ETx.) tion, presented in Fig. 4, represents Morkovin's

Figure 2. Current understanding of boundary layer 19844 update of an earlier chart which describes the

transition. laminar boundary layer as either a linear or nonlinearoperator in its response to a random disturbance

Figure 2 is a simplified schematic representation spectrum. It depicts a summary of the transition

of Reshotko's 1 5 concept of transition of the laminar process based on several decades of research inboundary layer to a turbulent state as the "non-linear theoretical and experimental studies of transitionresponse of a very complicated oscillator (the laminar phenomena. A careful review of this figure and the

boundary layer) to a random forcing function whose 1969 treatise by Morkovin should help the transitionspectrum is assumed to be of infinitesimal amplitude student to develop an appreciation for the complexity

compared to the appropriate laminar flow quan- of the transition process with its many competingtities." Also shown are the "bypasses" discussed by "operation modifiers" and the race betweenMorkovin, 3 such as roughness, free-stream instability modes.

42

AEDC-TR-B4-8

D.C. DISTURBANCES hu 60111T SWEEPIACK INSTAIILITIES

AC. INPUT - DISTURBANCES • , of d ds-IM-Shw Inat NOW of dowobwIrt-m wid5owý ofamt" spots

hohi*uec 1 30 Few*u

0 I 1 . RAcEL I.am-- J •II!.i Mi OOESJ

SLOW LINEAR AMPLIFICATION OPERATION MODIFIERSBYPASSES: ofd W pds: - MEAN I.L PIIOPKTIES

IdNN ASI/ ce 2D : DIRECT 30 : INDIIECT30 NONLINEAR MqsVut ofSPACE-TIME dktusmbs pust .) 30 rmalmSS

DISTURBUNCES of "e d e TA, '.CUMULATIVE EFFECTS a NdiM of z

dudu ,- WM (Mi•o eWg d yew

rlest n 1* of todl 3D m-b*od STRONG I>>C lo•u V60 0I

WWX301 NTAUlES t

2), N: NAMi AND (RSS-FLOWslegs ,»>>0: /hA )

sam"'Im n fI• tetnroeghm .

-..m IMF" ... ..m,""e-.1. a w "NW

SEWOPED TURBULENT K i - -.......

Figure 4. A specialist's view of boundary-layer transition.

TURBULENT BOUNDARY LAYER Figure 5 illustrates the disturbance sources presentin wind tunnels, as identified by Kovasznay. 16 These

\\\RD , ATED) disturbance modes were discussed by Morkovin, 1 .2

-P u who speculated on their origins in a supersonic wind"//o T tunnel. The vorticity and entropy fluctuations were

traceable to conditions in the settling chamber, while

MACH NUMBER sound disturbances could originate in the stilling

RANGE TYPE DISTURBANCE EFFECT ON TRANSITION chamber and from the test section boundaries. The

VORTCITY (V) -USUALLY turbulent boundary layer on the tunnel wall was iden-(TURBULENCE) DOMINANT tified by Morkovin as a potential source for radiated

Mý - 1.5 TO 3 NOISE (•) - MAY BE DOMINANT noise. Later research by Pate 17 has shown this to beENTROPY (T) -NEGLIGIBLE the dominant factor in transition on wind tunnel(TEMP. SPOTS) models.

M. - 3 TO IS RADIATED NOISE (I) - USUALLYICOLD FLOW) DOMINANT The comparison of a large collection of wind

M., > I0 ENTROPY * (T) -? tunnel and flight transition datai8 on sharp conesARC TUNNELS over a wide range of Mach number and ReynoldsSHOCK TUNNELS RADIATED NOISE (I) - MAY BE DOMINANTMHD TUNNELS number is presented in Fig. 6. The flight data are

generally higher than the wind tunnel data correlationFigure 5. Flow disturbances in supersonic and line, presumably due to large disturbance levels in

hypersonic tunnels, the wind tunnels. It is noteworthy that the flight data

exhibit a unit Reynolds number dependence.

43

AEDC-TR410

101 FLIGHT DATA: 17 PONIS Another comparison of wind tunnel and flight-" transition data from the shuttle is presented in Fig.

MEAN Kiltl X 10- 8. These transition data obtained during reentry

S 3. l, (increasing Reynold number) were significantlyIOo_' influenced by disturbances caused by "surface

V.. .4 -- o roughness" in the form of the shuttle tiles. Transition

.Reynolds number based on tile height, RCK,. 1 shows

SWIND TUNNEL DATA CORRELATION: that once transition occurred, it progressed quickly568 POINTS (NOT SHOWN) toward the shuttle nose for a relatively small change

0 2 4 6 8 10 12 14 16 in Reynolds number. On the other hand the wind

A, tunnel results (shaded area) predicted a more gradual

Figure 6. Comparison of wind tunnel and flight forward progress of transition. This example is

transition data on sharp cones. interpreted as an example of "effective tripping" bythe roughness with high amplification rates that

As previously mentioned, boundary-layer transi- diminish differences in ground and flight results.

tion is often detected from heat flux or temperature COMMENTS ON TRANSITION PREDICTIONmeasurements, and a typical example is presented in MMEHO DOLOGN

Fig. 7. This figure shows shuttle temperature mea- METHODOLOGYsurements during reentry of flight STS-2. Just after1,000 sec, there was a significant rise in the flight i tw od be pnictiag toaputiforth ag

temperature data which is typically presumed to be reliable method for predicting transition such thatcaused by the increased heating associated with tur- those people who have to make estimates of transi-bulent flow. Note, that the predicted time of transi- tion Reynolds numbers would know exactly how to

tion was at about 800 sec which corresponds to a sig- proceed. Unfortunately, the problem is not thatnificantly lower value of transition Reynolds number. simple. The best that can be done is to provide some

This is just another example of the inability to predict general guidelines and encouragement to make thethe correct flight transition Reynolds number. most of the data that are currently available. Thus,the following are some comments on the methodo-

2,500 TRANSITION logy for hypersonic transition predictions (see Fig. 9).

-. 2,000 -IkI 1,250\NA 1,000 STEP I: OBTAIN BAIGROUND INFORMATION

- 750 STEP 2: OBTAIN FLOW FIELD DATA

z F1IGHT DAT SooJ) 0 STEP 3: CONSIDER DOMINANT MECHANISMSSoo PREDICTION 250 STEP 4: OBTAIN A BASE-LINE TRANSITION REYNOLDS NUMBER

0 STEP 5: BIAS THE TRANSITION RESULTS0 400 300 1.200 1,600

TIME FROM ENTRY INTERFACE, SEC STEP 6: ESTIMATE UPPER AND LOWER BOUNDS

Figure 7. Control surface temperature comparison, STEP 7: A MIRACLE OCCURS AND YOU INOW WHAT THE TRANSITION REYNOLDS

STS-2. NUMBER SHOULD BE

Figure 9. Transition prediction methodology.19

1,000 RANGE FOR ROUGH MODELS (W) TUNNEL-RELATED DISTURBANCES STEP 1: OBTAIN BACK-', V", T, GROUND

N8OS WATCH OUT FOR. I MAIN BYPASSES: Sometimes un-

FLIGHT CONDITIONS DISTURBANCES FUSELAGE expected phenomena such aso STS- I o STS-3 DUE TO MISALIGNED freestream turbulence or wall"TILES (VEHILE GENERATED cooling effects can greatly reduce

10 1 1 DISTURBANCES) the expected transition Reynolds0 0.2 0.4 O/ 0.6 t.1 numbers.05

XILFigure 8. Comparison of wind tunnel and flight transition for space shuttle.

44

AEDC-TR-94-8

CONFIGURATION DIFFERENCES: Be aware dary-layer thickness is used as the length dimension,of the influence of configuration differences on tran- for a value of ReCF = 200 it could be expected thatsition. The cone versus flat plate issue illustrates the crossflow instabilities would dominate and cause

problem. It had generally been assumed that one transition. 23 ,26

should obtain higher transition Reynolds numbers oncones than on flat plates, at least between Mach ReCF= max

numbers 3 and 8. Experiments were performed in the A

NASA/Langley Research Center Mach 3.5 quiettunnel to investigate this issue.2 °,21 The ratios of LAIGEG OTMN~O:I hconne-to-flaest-pate transitionReynolds numberas wee beginning of the leading edge of a swept wing is incone-to-flat-plate transition Reynolds num be rs were c n a t w t oi u f c e g ,af sl g ra w nfound to vary from about 0.8 for low-noise free- contact with a solid surface (e.g., a fuselage or a windstream conditions to about 1.2 for higher noise tunnel wall), the turbulence which is present in theconditions. boundary layer of the adjoining surface will con-

taminate the leading edge boundary of the swept

PROBLEMS OF WIND TUNNEL DATA: Con- wing. Such turbulence contamination has a signi-

ventional wind tunnels have large disturbances in ficant effect on the state of the leading edge boundary

their free-stream environments; thus, wind tunnel layer and can dominate the transition process on the

transition Reynolds numbers are generally lower than wing. 27

corresponding flight transition Reynolds numbers.GORTLER INSTABILITIES/ADVERSE

LENGTH OF THE TRANSITION REGION: It PRESSURE GRADIENTS: There is insufficient data

has been customary to assume that the length of the available to establish a general criterion to determine

transition region is the same as the length of the when adverse pressure gradient effects and Gortler

laminar region. 22 instabilities will dominate and produce an earlytransition.

STEP 2: OBTAIN FLOW-FIELD DATALinear stability theory (the eN method Refs. 28

Uncertainty in flow-field calculations directly and 29) has been used to predict the effects of Gortler

influences the uncertainty in the transition estimates. instabilities on transition of boundary layers on wind

Calculations of the boundary-layer properties are a tunnel nozzle walls.

very important part of the transition problem. 23

Close attention should be given to the flow-field STEP 4: OBTAIN A BASELINE TRANSITION

properties. REYNOLDS NUMBER

STEP 3: CONSIDER DOMINANT MECHANISMS Baseline transition Reynolds number can beobtained from a collection of transition data (such

NOSETIP: Because nosetip transition Reynolds as Fig. 6) or by correlation techniques. Remember

numbers can be influenced by roughness they can be that flight data such as contained in Fig. 6 already

two orders of magnitude less than frustum transition contains effects such as small nosetip bluntness, small

length Reynolds numbers, it is necessary to consider angles-of-attack, and some wall temperature vari-

nosetip transition independently from the rest of the ations.

configuration.24

STEP 5: BIAS THE TRANSITION RESULTS

EARLY FRUSTUM: Early frustum is defined asthe region just downstream of the nosetip, extending If there are several parameters which are expected

for several nose radii. The transition experiments to have a significant effect on transition, but no single

reported in Ref. II clearly identified the early cone effect is clearly dominant, then some accounting forfrustum as a region with its own transition criteria, the individual effects should be made. If a baseline

transition Reynolds number is obtained, make adjust-

CROSSFLOWS: The laminar boundary-layer ments to this number to account for the other param-

profile in a three-dimensional, viscous flow has a eter effects which are expected to be significant. Mosttwisted profile that can be resolved into tangential of the parametric trends come from wind tunnel data(u) and crossflow (v) velocity components. The cross- since wind tunnel experiments can be betterflow component of the velocity is used for the com- controlled than flight experiments.

putation of crossflow Reynolds number. If the boun-

45

AEDC-TR-4-

STEP 6: ESTIMATE UPPER AND LOWER COAMIED MODEL AND TRIP BOW WAVES

BOUNDS TMRL • WWAVE

All transition estimates will have an uncertainty

associated with them, and obviously it is important __ _ _ MoDELto consider this uncertainty when predicting transi-

tion location. b. Correctly sized tripFigure II. Concluded.

STEP 7: A MIRACLE OCCURS AND YOUKNOW WHAT THE TRANSITION flow over a model. In this case artificial tripping of

REYNOLDS NUMBER SHOULD BE the boundary layer is required. The fundamental

principles of boundary-layer tripping are illustratedB gnDArYLAYEro , TiPS 1in Fig. 11. Experience (e.g.,Refs. 30 and 31) has

shown that a spherical trip with a height on the orderof the boundary-layer thickness, 5 will produce a

The preceding discussion showed that wind tunnel turbulent boundary layer, yet will not cause excessive

transition data could not be depended upon to pre-

dict transition in flight. However, a typical vehicle downstream flow disturbances. A phase-change paint

will experience either a laminar or turbulent boun- photograph of the turbulent flow produced on the

dary layer for a significant part of its mission profile, nose of a shuttle model by a single trip sphere is

and wind tunnel tests can be designed to produce data shown in Fig. 12a.

on models with predominantly laminar or turbulent •.Iboundary layers. It is very important to know the -boundary-layer state and surface heat transfer mea-surements are routinely used for this purpose (e.g.,Ref. 26).

"* (AIT DEPENOD ^ GROUND TST1 FAILITIES MOKB ICTING FLIGHT TRANSITIONREYN~tOS NUMM(

"* BUT WE (AN OBTAIN MODEL DATA WITH LAMINAR, TRANSITIONAL OR TURBULENTBOUNDARY-LAYER STATE

" VERY IMPORTANT TO KNOW BOUNDAY.LAYER STATE OF TEST

"* TYPICALLY WE WANT LAMINAR Of TURBULENT NOT TRANITIONALa. Single sphere

"* BUT WIND TUNNEL le. IS TOO LOW TO PRODUE( A TURBULENTBOUNDARY LAYER

.. MUST USE BOUNDARY-LAYER TRIPPING TECHNIQUES

Figure 10. An engineering approach to transition.

Laminar flow over a model can usually beobtained by running at the low end of the operatingstagnation pressure range of the facility. On the other

hand, in many wind tunnels operation at the high endof the operating envelope may not provide a -

Reynolds number high enough to produce turbulent

DEFLECTED BOW WAVE b. Multiple spheres, single row

REECTED TRIP ROW WAVE -•Figure 12. Illustration of boundary layer trip

TRIP RCOMPRESSION WAVE effectiveness on nose of an orbitorik model.

Fig. 12b is a similar photograph, but with a singleMO"E THICKENSYED row of trips spaced about three diameters apart. In

general, three rows of staggered trip spheres, as illu-

a. Oversized trip strated in Fig. 13, is the most desirable trip configura-Figure 11. Fundamentals of trip sizing. tion. Note that the spheres are welded on to a thin

46

AEDC-TR-94-6

TRIP SPHERES

: • II. >hotograph of trip ring sith three rows IFigure 14. Photograph of numerous noses requiring

ot staqigcred ,phcres. trip rings for typical force test.

SYM Ret Just as predicting flightTUNNEL B. M= transition is a complex art, so is the

66x I0t proper sizing of boundary-layerftAG BOUNDARY LAYER TRIPS t F 1 i

LAMINAR-LOCAL SIMILARITY r trips. Figure 16 is an outline of a20 --- TURBULENT-SPALDING CHI 1*O computer code used to size trips.

h 7o "h-N on,- h INPUTh_ FREESTREAM CONDITIONS

S. AXISYMMITRIC - - MODEL GEOMETRYOf -AXISYMMETRI( LOCAL CONDITIONS

DI 0 oTRIP LOCATION

TWO DIMENSIONAL oWO2 0NNM KAX, MAX TRIP SIZE.00? UtO

0 0.2 04 06 08 t0 0 02 04 06 08 I0x/L x/LLOCAL STREAM PROPERTIES

0= I DEG 20 DEG IxlO, NATURAL TRANSITION i

Figure 15. Illustration of boundary layer trip application. LOCATIONI

GENERALIZED BOUNDARY LAYERPRO FILE I

detachable ring which is easy to install and remove STREAM PROPERTIES AT TOP Of

from a large number of similar configurations (see TRIP ELEMENT

Fig. 14). A common procedure is to select a trip F

sphere size (diam) and to build three detachable rings, fIND RITICALuMBTRIP REYNOLDS

one with the nominal trip diameter and the other two Ifwith a smaller and larger size, respectively. XT. TRANSITION LOCATION j

Typical data illustrating the use of boundary-layer

trips on a space shuttle model are shown in Fig. 15. YE]

Note that the flagged symbols correspond to the

"tripped data" and, in general, agree with the NOturbulent boundary layet predictions. The unflaggedhigh Reynolds number data at (y C 20 deg are YES

representative of a transitional boundary layer while

the tripped data moved transition all the way to themodel nose. Figure 16. Trip sizing computer code.

47

AEDC-T4"4

SUMMARY 8. Stetson, K. F., Thompson, E. R., Donaldson,J. C., and Siler, L. G. "Laminar Boundary

The primary conclusions of this section are: Layer Stability Experiments on a Cone at Mach8, Part 3: Sharp Cone at Angle of Attack."

" Boundary-layer transition prediction for AIAA Paper No. 85-0492, January 1985.flight has been studied by specialists formany years but it remains elusive 9. Stetson, K. F., Thompson, E. R., Donaldson,

J. C., and Siler, L. G. "Laminar Boundary" Boundary-layer transition observed in wind Layer Stability Experiments on a Cone at Mach

tunnels is often upstream of that observed 8, Part 4: On Unit Reynolds Number andin flight because of "noise" in the tunnel Environmental Effects." AIAA Paper No.

" Boundary-layer trips have been used for 86-1087, May 1986.

many years to produce turbulent flow in10. Stetson, K. F., Thompson, E. R., Donaldson,

spheres most effective) J. C., and Siler, L. G. "On Hypersonic Transi-

tion Testing and Prediction." AIAA Paper No.

REFERENCES 88-2007, May 1988.

11. Stetson, K. F. "Nosetip Bluntness Effects on1. Morkovin, M. V., "On Transition ExperimentsCoeFutmBndrLarTasionn

at Mderte uperoni Speds. Jorna ofCone Frustum Boundary Layer Transition inat Moderate Supersonic Speeds." Journal of HyrsncFo.AAAPprN.8-63

the Aeronautical Sciences, Vol. 24, No. 7, July Hypersonic Flow." AIAA Paper No. 83-1763,

1957, pp. 480-486. July 1983.

2. Morkovin, M. V. "On Supersonic Wind 12. Stetson, K. F. "On Nonlinear Aspects of

Tunnels with Low Free-Stream Disturbances." Hypersonic Boundary-Layer Stability." AIAA

J)urnal of Applied Mechanics, Tran. ASME, Journal, Vol. 26, No. 7, pp. 883-885, July 1988.

Vol. 26, Series E, June 1959,pp.319-324. 13. Stetson, K. F. "Boundary Layer Transition on

3. Morkovin, M. V. "Critical Evaluation of Blunt Bodies with- Highly Cooled Boundary

Transition From Laminar to Turbulent Shear Layers." J. A. S., Vol. 27, pp. 81-91, February

Layers with Emphasis on Hypersonically 1960. (Also, IAS Report No. 59-36, January

Traveling Bodies." AFFDL TR-68-149, March 1959).

1969. 14. Stetson, K. F. "Mach 6 Experiments of Transi-

tion on a Cone at Angle of Attack." Journal4. Morkovin, M. V., AIAA Lecture Series on

Boundary Layer Transition, 1984. of Spacecraft and Rockets, Vol. 19, No. 5, pp.397-403, Sept-Oct 1982.

5. Stetson, K. F. and Kimmel, R. L. "On Hyper- 1 Reshotko, Eli. "Stability and Transition, Howsonic Boundary-Layer Stability," AIAA Paper Much Do We Know?" Paper presented atNo. 92-0737, January 1992. Tenth U. S. National Congress of Applied

6. Stetson, K. F., Thompson, E. R., Donaldson, Mechanics, The University of Texas at Austin,

J. C., and Siler, L. G. "Laminar Boundary Austin, Texas, June 1986.

Layer Stability Experiments on a Cone at MachAIAAPape No. 16. Kovasznay, Leslie, S. G3. "Turbulence in8, Part 1: Sharp Cone. Supersonic Flow." Journal of the Aeronautical83-1761, July 1983. Sciences, Vol. 20, No. 10, October 1953, pp.

7. Stetson, K. F., Thompson, E. R., Donaldson, 657-674.

J3. C., and Suer, L. G. "Laminar Boundary 17. Pate, S. M. "Effects of Wind Tunnel Dis-Layer Stability Experiments on a Cone at Mach turbances on Boundary-Layer Transition with8, Part 2: Blunt Cone." AIAA Paper No. Emphasis on Radiated Noise: A Review."84-0006, January 1984. AIAA Paper No. 80-0431, March 1980.

48

AEDC-TR-94-6

18. Goodrich, W. D., Derry, S. M., and Bertin, J. 26. Matthews, R. K. "An Experimental Investiga-J. "Shuttle Orbiter Boundary-Layer Transi- tion of the Effects of Geometry and Attitudetion: A Comparison of Flight and Wind Tunnel on the Windward Heating of Conical ModelsData." AIAA Paper No. 83-0485, January at Mach 8." AEDC-TR-73-69 (AD-909-655L),

1983. April 1973.

19. Stetson, K. F. "Hypersonic Boundary-Layer 27. Creel, T. R., Jr., Beckwith, I. E., and Chen,Transition." The Second Joint Europe/US F. J. "Transition on Swept Leading Edges atShort Course in Hypersonics, GAMNI-SMAI, Mach 3.5." Journal Aircraft, Vol. 24, No. 10,January 1989. December 1987.

20. Beckwith, 1. E. "Development of a High 28. Chen, F. J., Malik, M. R., and Beckwith, 1.Reynolds Number Quiet Tunnel for Transition E. "Instabilities and Transition in the WallResearch." AIAA Journal, Vol. 13, No. 3, pp. Boundary Layers of Low Disturbance Super-

300-306, March 1975. sonic Nozzles." AIAA Paper No. 85-1573,1985.

21. Chen, F.-J., Malik, M. R., and Beckwith, I.E. "Comparison of Boundary Layer Transition 29. Beckwith, 1. E., Chen, F. J., and Malik, M.

on a Cone and Flat Plate at Mach 3.5." AIAA R., "Design and Fabrication Requirements forPaper No. 88-0411, January 1988. Low-Noise Supersonic/Hypersoqic Wind

Tunnels." AIAA Paper No. 88-0143, January

22. Pate, S. R. "Dominance of Radiated Aero- 1988.dynamic Noise on Boundary Layer Transitionin Supersonic-Hypersonic Wind Tunnels, 30. Matthews, R. K., Buchanan, T. D., Martin-Theory and Application." AEDC-TR-77-107 dale, W. R., ARO, Inc. and Warmbrod, J. D.

(AD-A052-621), March 1978. MSFC. "Experimental and Theoretical Aero-dynamic Heating and Flow Field Analysis of

23. Neumann, R. D., and Patterson, J. L. "Results a Space Shuttle Orbiter." NASA TM X-2508,of an Industry Representative Study of Code February 1972.to Code Validation of Axisymmetric Con-figurations at Hypervelocity Flight Condi- 31. levalts, J. 0., and Matthews, R. K. "Aero-tions." AIAA Paper No. 88-2691, June 1988. thermal Evaluation of High Temperature

Structural Materials for Use in High-Per-

24. Batt, R. G., and Legner, H. H. "A Review of formance Missile Design." AEDC-TR-79-38Roughness Induced Nosetip Transition." (AD-A080-023), January 1980.

AIAA Paper No. 81-1223, June 1981.

25. Pate, S. R. and Adams, J. C. "HypersonicSimulation for Lifting Body TransitionStudies." Proceedings of the Boundary LayerTransition Workshop, November 1971, Aero-space Report No. TOR-0172(S2816-16)-5,December 1971.

49

AEDC-TR-04-

ELECTROMAGNETIC WAVE TEST

byR. K. MATTHEWSSenior Staff Engineer

and

S. A. StepanekPrincipal Engineer

Calspan Corporation/AEDC OperationsArnold Engineering Development Center

APSTRACT recently, standard experimental methods providedmeasurements only for radomes at ambient

Electromagnetic wave testing represents a temperature. As missile velocities increase causingrelatively new test technique that involves the union aeroheating to be important, it is increasingly

of several disciplines: aerothermodynamics, electro- important that BSE measurements be made on

magnetics, materials/structures, and advanced dia- radomes under realistic flight conditions, especiallygnostics. The essence of this new technique deals with in terms of radome temperature levels andthe transmission and possible distortion of electro- distributions. In one such effort, Weckesser et al.,I

magnetic waves (RF or IR) as they pass through the investigated the effect of temperature on BSE forbow shock, flow field, and electromagnetic (EM) several ceramic radomes using a solar furnace as awindow of a missile flying at hypersonic speeds. radiant heat source for the radomes. OtherVariations in gas density along the optical path can investigators have used various nonaerodynamic heat

cause significant distortion of the electromagnetic sources, including electric blankets and radiationwaves ind, therefore the missile seeker system may lamps. One of the drawbacks with these methods is

not effectively track the target. Two specific test that the aerothermodynamic environment is not

techniques are described. The first example deals with closely simulated so that, although elevated radomethe combining of a wind tunnel and an RF range temperatures are achieved, correct distributions andwhile the second example discusses the complexities levels are not. In addition, the lack of simulation of

of evaluating IR seeker system performance. flight pressures and aerodynamic shear, which

contribute to the ablation/recession process, makesINTRODUCTION these techniques even less desirable for evaluating

Tactical missiles currently being designed or ablative type radomes.

developed will fly much faster than existing systems.The radomes on these missiles will experience high (ROSS.PtlAE ERIOR GIN WtU

aerodynamic heating during flight resulting in ele- Fvated radome temperatures which, in turn produce INPUNE ERROR

changes in the dielectric properties of the radome and -TENNcontribute to changes in boresight error (BSE) (illus- -

trated in Fig. I). BSE is a function of the radomematerial, thickness, temperature, and receiver operat- .,' I0RESGlT ERan (USE): DIFFENCE KBTEN

ing frequency. Therefore, characterization of radome APPARENT LINE OF SOT TO THE TARGET

EM transmission performance in terms of boresight lTtRST ANE D M THE RL ,SA UNE OF ERAT UTIA $S 1 - ftKDOE MATER, %WA, TEKAURE

error over the range of possible seeker-antenna SIULATING TARGET SoGAL FROEN(, ETC.)

gimbal angles is essential in the development of aradome. Figure 1. Radome cross-section sketch showing

boresight error.Because of the complex manner in which the

various radome and seeker-system design parameters Ri TEST TECHNIQUEinteract, it is very difficult to use predictive methodsto evaluate the boresight error characteristics of a The AEDC, working jointly with General Dy-

system. For that reason, radome electrical per- namics/Electronics (GD/E), has developed a unique

formance is determined experimentally. Until test technique which permits boresight error mea-

51

AEDC-TR-94-6

surements to be made on radomes exposed to AMUR l U•I I NMIconditions closely approximating flight. The UM= Wtechnique was developed by using the Aero- Ftmlthermal Wind Tunnel (C) in combination with -Ma radar range facility designed specifically for this S KTO UAW ispurpose. This concept takes advantage of NINEexisting wind tunnel capabilities and data systems IW Tan

and provides extensive coverage of the radometemperature and surface recession. FD i RK

The test program as described below provided an a. Facility sketch

example of the application of the boresight error testtechnique which consisted of two distinct test phases. M

Phase 1: Experimental Determination of the *Radome Thermal Response to the f

Wind Tunnel Environment.

Phase I: Validation of the Radome Rangeand Boresight Error Measurementson Aerodynamically HeatedRadomes.

The Aerothermal Wind Tunnel (C) used for these b. Tunnel after BSE test modificationtests is a continuous flow. 0.64 m (25-in.)-diam Figure 2. Radome range/aeroheating facility atvariable density, Mach 4 free-jet wind tunnel. This AEDC.unique tunnel (Fig. 2a) was developed specifically formaterials testing and provides "clean" aerodynamic aI

flow at temperatures, pressures, and shear loadssimilar to those experienced in actual flight. "Clean"refers to both the repeatable quality and uniformityof the flow as well as the fact that air is used as the "test gas, and only clean air constituents are present(i.e., no water vapor, dirt, dust or products of ':....,! - ",

combustion). The tunnel is equipped with a modelinjection system which translates the test article"into" and "out of" the flow field in a few seconds.However, since the airflow is continuous, the testarticle can remain in the flow for a time periodrequired for the measurement (e.g., 10 to 140 secduring the current test). A complete description ofTunnel C can be found in Ref. 2.

In order to provide a facility with a boresighterror measurement capability, a radar range wasinstalled directly below the tunnel nozzle as shownin Fig. 2b. The range consisted of an existingcylindrical pressure vessel which was matched to theinstallation tank door opening of the wind tunnel.A photograph of the inside of the range, taken fromthe transmitter looking towards the radome andreceiver, is shown in Fig. 3. Figure 3. View of AEDC boresight error range

looking toward radome from trans-External surface temperatures of the radome were mitter location.

monitored with an infrared data system. 3 With this

52

AEDC-TR-94-6

system, the surface temperature of the entire radome In Fig. 5, results obtained for one of the radomeswas mapped at various times during wind tunnel tested during the BSE phase are shown to examineexposure and during the BSE measurement process, the difference between BSE data for the ambientFigure 4 shows a sample surface temperature map (before injection into tunnel flow) and elevated (firstproduced from the IR system output. Surface sweep after retraction from tunnel flow) radome tem-temperatures determined from IR measurements peratures. Notice that the peak values of BSE arewere also plotted with interface and backwall thermo- different for the ambient and heated data and that

couple data as shown in Fig. 4c to aid in determining the gimbal angle at which the peak occurs is alsoradome wall temperature gradients. different. Therefore, these data indicate that the

elevated temperature and/or recession which occursHKIN~r 0.44 m (17.5 IN.), during exposure to the tunnel flow produced signi-

ficant changes in the BSE characteristics of thisradome.

5111k FOLRATINAN TEMPRATUREo NORIOAL

* VERTICAL ELEVATE (FRUST SWEEP* NORMDUAL M

SHADOWGRAPILUNI PIU11WNDWOULNE

a. Shadowgraph flow field 8&

EXPOSURE TIME - 120 sa

0

•00

o

0 4 U 12ain....,,.'.. wh GIMAL ANIG, DEGitin hi S i .S- Sii ii i.aa.iiiil "i a Ci iS S

.-- Figure 5. Comparision of ambient and heatedb. IR system mapping of surface temperature radome boresight error data.

response (top view)

1100 AERO-OPTICS TEST TECHNIQUE1o000SURFACE TEMPERATURE Understanding the interaction between optical900 FRLOMliR_00,.,OM It - RETUCT waves and aerodynamic flow fields is also critical in

S . -the development of optical guidance and tracking sys-700 BEGIN SE tems. Accurately discerning the location of radiat-0 MEASUREMENT

600 - ao ing targets is difficult from images that are blurred,

So 0 o jittered, and attenuatod by an unsteady aerodynamic4 0 INTERFACE TEMPERATURE flow field. Refraction of optical waves propagating400? NJE through shock waves and density gradients must be

300 I known for in-flight compensation. The distortion of

200 -A(IW&LL TEMPERATURE target radiation due to hypersonic aerodynamics for

t00 0 1 a cooled-window flight scenario is depicted in Fig.0 40 80 120 160 200 240 6. Fluctuating effects due to turbulence, shock/

TIME. s boundary layer interactions, flow separations,

c. IR and thermocouple temperature history. mixing/shear layer flows, etc. must also be evaluated.Figure 4. Sample data obtained from thermal phase.

53

AEDC-TR-94-6

currently unreliable. Furthermore, existingS~til(M Oground test facilities only simulate parts of the

-DISTOVION severe conditions encountered in high-speedAIS...•M0 ow MNIK flight. Therefore, the approach required for

M.I- N 1AVER modern aero-optic investigations is one where-a- .code-validalion experiments must be carefully

IMAT SOME • ""performed in available test facilities, and thenCOOLINGTt•--- LAM validated codes must be used to predict the aero-

optic effects at flight conditions. As shown inwINOW o TTIO- Fig. 7, the application of aero-optics technology

includes validating complex CFD and opticalSVwave propagation models using both aerodyn-

amic and optical measurements from carefullydesigned code validation experiments. Thisapproach is particularly difficult given the

Figure 6. Aero-cptical distortion scenario, current state-of-the-an of turbulence models andflow diagnostics where few validated techniquesexist.

Additionally, flight system effects, such asmechanical vibrations and thermal-mechanical The basic approach developed to obtain experi-window stress, can render guidance and tracking mental optical measurements in a large-scale ground

sensors useless. ' ýnsing effects from stress-induced test facility at AEDC is illustrated in Fig. 8. Thewindow distortions (exacerbated by aerodynamic objective of this .echnique is to isolate and characte-heating and pressure loads), and radiation from rize the flow-induced (aero-optic) distortion. An IRheated windows and hot external gases can preclude beam (representing the target radiation) originatingtarget discrimination. Realistic issues such as these outside the tunnel is introduced through a flat platecomplicate the specification, design and testing of and directed toward the test article. After passingoptical systems. through the test article flow, the then-distorted bean'

is directed out of the tunnel to the receiving optics.In evaluating aero-optic effects, designers struggle The primary test requirement is to provide an undis-

with theoretical predictions and experimental torted beam to the test article, such that as the beammethodologies because both are unvalidated and passes through the local test article flow, thus the only

FVLO IWJ.g DEWU EVLP onto M"I MODELS MumIUN MEASURIEMENT SYSTEM

MU I MIM0RVNI FLOW P"PW~oN

Il~miNl

WAW IMMI•T I

_IV R[ONM E-AIDATIO XEIET

OEIG FU• SYTE M

Figure 7. Applying aero-optics methodology to flight system design.

54

CFOlKllIi

AEDC-TR-94.4

Figure 10 shows the physicalinstallation inside the wind tunnel,

looking downstream. Optical distor-tion measured in this technique-valida-

Stion configuration represents the com-PLATEUIVING bined facility effects (including a

OPTICSM second flat plate and optical system

imperfections) and serves as an end to-end evaluation of the ability of thefacility to successfully provide anundistorted beam for fundamentalAERO-OPTICAL

CAI[IIIITION WEDGE aero-optics measurements.

Figure 8. AEDC aero-optics measurement system basic scheme. The beam passes through the local

flat plate boundary layer and bow

shock. The plate is pitched windwardslightly to ensure a well established but weak attached

optical distortion measured is caused by the test shock. Flat plate surface pressure and heat tr~.asferarticle flow. Only in this manner can the optical measurements are required to ensure that the flowdistorton be directly related to the flow for code over the source window is not disturbed by leadingvalidation. edge effects or wraparound flow from under the

plate. Heat transfer measurements are also used toCareful attention must be placed on minimizing locate boundary layer transition, which ideally should

the effects of facility vibration on the optical mea- occur well downstream of the window to allowsurements by using isolation damped optical tables passing the beam through a benign laminar boundaryand an evacuated, rubber-bellows-mounted beam layer.

path.

An overhead view of the physical installation of W 0mNaUr

the aero-optical measurement system is shown in U UW - nsFig. 9. AEDC's system is a fully integrated optical duiam Pun s-ss 0Mutmeasurement system providing imaging, image jitter,

and holographic measurements. Diffraction-limited . jperformance of the beam delivery system requiredthe use of highest-quality optics. Custom built suMu S1

electronics controlled and varied the duration of thebeam pulse so that the effects of camera-integrationtime could be determined.

Figure 10. View looking downstream at aero-opticwedges and supporting equipment.

The types of optical test measurements are

depicted in Fig. I1. Image spot data, used primarilyto quantify geometric degradation of the opticalwaves, are obtainable from intensity measurements

of a focused image spot. From imaging measure-I L ments, four parameters are commonly determined:boresight error, image blur, image jitter and Strehlratio. (These same parameters are determined theo-

retically by passing a plane wave, representing theFigure 9. Aero-optics equipment installed for test in target radiation, through the calculated aerodynamic

AEDC tunnel C, M = 8. field.) Boresight error refers to the difference betweenthe position inferred from the detected image and the

55

AEOC-TR-946

mean flow conditions, and the application of•OESIGI1T ERROR assumed turbulence models to define the fluctuating

MAGNITUDE AND DIRECTION flow properties, to the calculations of the optical"OF ST Sbeam distortion. An example of the utility of nume-

rous and complimentary measurement techniques to

. SHAPE CHANGE - •' c2 o verify various theoretical calculation steps is depicted(t) (t + '1) (1 + 2at) in Fig. 12.

$ 2

* JITTER 4 The use of advanced, non-intrusive flow diag-

• RANDOM MOTIO I mj<• .'l " nostic techniques is considered an enabling andnecessary feature of aero-optic testing in that experi-

* STREHL RATIO (I l mental characterization of the key fluctuating stateR S 09 / I properties (e. g., density) is the critical link betweenS/ /.,•,•\/•" (I + 2&t)

L t •.,c.X/N -t + lit) the aerodynamics and the optical effects. The- tif promise of advanced techniques to provide simulta-

DISTANCE neous measurement of more than one flow property

Figure ii. Aero-optical image-spot characteristics. may be invaluable to successful

PITOT modeling of turbulent mixing of1T PROBE A(OUSTI( dissimilar, reacting gases.

LIF, OPTICALPITOT MFA PRONE LDY LIF HOT FILM MEASUREMENTS To illustrate window/frame-

I induced distortion experiments in

P.p, ,, u, M -4 u, P, I -' . i etc tUi R ground test facilities, Fig. 13 includesBj I some imaging results from AEDC's

JITTE Tunnel C facility validation test. On

the left side, without installation of theMEAN AE1O FLUCTUATIONS AE10 OPTICAL DISTOuTIhN windows in the flat plate test articles,

the imaging data show a diffraction-

AERO-OPTIKS COMPUTATIONAL DEVELOPMENT limited spot, illustrating the successfuldevelopment of a near-perfect optical

Figure 12. Use of aero measurement techniques for code beaelivertem. However, whenvalidtion.beam delivery system. However, when

validation. the windows were installed, serious

actual beam (target) location. Image blur is the degradation of the beam resulted (center image). This

spreading of an imaged spot relative to its original was determined to be caused by mounting stressesspredin of n iagedspo reativ toits rignal in the window frame assembly. Finally, with the wind

shape, which results from variations in the refractive tn the window s atnarly with the

media, and its effect is seen as a degradation in image resulting image shows a different pattern of dis-

clarity. Image jitter is the random unsteady motion tortion. Unforeseen difficulties are common inof the image about a reference point, which results performing aero-optic tests, and later modificationsin the appearance of the image to flicker or blur, to the window frame were designed and validated in

depending on the frequency of the randomness and lab experiments.the sampling of the detection system. The Strehl ratiorefers to the diminished total energy of the beam after SUMMARYit passes through the aerodynamic field.

In summary, performing aero-optic investigationsThe validation of aero-optic prediction codes is a relatively new technology in appliJ aero-

requires experimental measurement of both the aero- dynamics, and from the experiences at AEDC,dynamic flow field and the resulting optical dis- several key observations can be made.tortion. Optical measurements alone are not suffi-cient to validate the fundamental aspects of the code I. In performing code validation experiments,

that relate the optical wavc distortion to the aero- both aerodynamic and optical measurementsdynamic flow. Given the complexity of the aero- are required - and most important is the

dynamic environment, as many theoretical steps in characterization of the turbulence, since thethe code as possible should be independently vali- weakest part of existing codes is the tur-

dated experimentally: from the calculation of local bulence model.

56

AEDC-TR-44

a. No windows b. Windows, no flow c. Windows and flow(mechanical stress) (thermal and mech-

anical stress)Figure 13. Image spots showing window-stress effects.

2. Meticulous attention must be given to the Error." Proceedings of the Fifteenthfacility-induced error sources, which can in- Symposium on Electromagnetic Windows,validate measurements made with even the Georgia Institute of Technology, June 18-20,most perfect optics 1980.

3. Any system designed and constructed must 2. Strike, W. T. "Calibration and Performance

be painstakingly tested and validated prior of the AEDC/VKF Tunnel C, Mach Num-

to its use. The analysis of measurement ber 4, Aerothermal Wind Tunnel." AEDC-

uncertainty of aero-optic testing is an TR-82-6 (AD-A 116279), June 1982.essentially-untouched frontier. 3. Boylan, D. W., Carver, D. B., Stallings, D.

4. Investigating fundamental wave-flow W., and Trimmer, L. L. "Measurement and

interactions requires a diversity of Mapping of Aerodynamic Heating Using aRemote Infrared Scanning Camera in Con-interrelated disciplines, including aero-

dynamics, electro-optics, spectroscopic tinuous Wind Tunnels." AIAA Paper No.

physics, material/structures, ground testing 78-779, April 1978.

and CFD. 4. Crosswy, F. L. and Havener, A. G. "Aero-Optics Measurement System for the AEDC

5. Planning and preparing for aero-optics Aero-optics Test Facility." AEDC-

testing requires careful consideration of test TR-90-20, Arnold Engineering Development

simulation issues, measurement require- Center, February 1991.

ments, facility and instrumentation error

sources, etc. 5. Havener, A. G. and Stepanek, S. A. "Aero-

optics Testing Capabilities & AEDC."6. Additional information regarding Aero- AIAA Paper 92-0760, 30th Aerospace

Optics Wave Testing can be found in Science Meeting, Reno, NV, January 6-9,References 4 through 6. 1992.

REFERENCES 6. Dixon, P. W., "Image Degradation Causedby Thermal Gradients in Windows." AIAA

1. Weckesser, L. B. et al. "Radome Aero- Paper 92-075, 30th Aerospace Sciencedynamic Heating Effects on Boresight Meeting, Reno, NV, January 6-9, 1992.

57