National Aeronautics and Space AdministrationLangley Research Center • Hampton, Virginia 23681-0001

NASA Technical Paper 3475

Analysis of Fluctuating Static PressureMeasurements in the National TransonicFacilityWilliam B. IgoeLangley Research Center • Hampton, Virginia

March 1996

Printed copies available from the following:

NASA Center for AeroSpace Information National Technical Information Service (NTIS)800 Elkridge Landing Road 5285 Port Royal RoadLinthicum Heights, MD 21090-2934 Springfield, VA 22161-2171(301) 621-0390 (703) 487-4650

Available electronically at the following URL address: http://techreports.larc.nasa.gov/ltrs/ltrs.html

iii

Contents

Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2. Dynamic Measurements in NTF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32. Test Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.1. Benefits of Cryogenic Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.2. Description of NTF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.3. Instrumentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3.1. Pressure Transducer Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3.2. Pressure Transducer Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.3.2.1. Static calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3.2.2. Dynamic check-calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3.2.3. Resonant frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.3.3. Pressure Transducer Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.3.4. Signal Conditioners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.3.5. Dynamic Data Acquisition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.3.6. NTF Steady-State Data System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.4. Data Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.4.1. Free-Stream Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.4.2. Fluctuating Static Pressure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82.4.3. Statistical Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.4.4. Data Repeatability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3. Test Conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3.1. Static Pressure Gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3.2. Air Mode Tests. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3.3. Nitrogen Mode Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104. Discussion of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

4.1. Effect of Hot Wall and Cold Wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

4.2. Effect of Fixing Boundary Layer Transition on 10.6° Cone . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

4.3. Comparison of Air and Gaseous Nitrogen Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

4.4. Effect of Fan Drive Power Variation at Constant Mach Number and Reynolds Number . . . . . . 14

4.5. Variation of Fluctuating Pressure Coefficient With Reynolds Number . . . . . . . . . . . . . . . . . . . 144.5.1. Effect of Constant Stagnation Pressure, Stagnation Temperature, or Fan Drive Power . . . . 144.5.2. Effect of Reynolds Number in Air. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

4.6. Variation of Fluctuating Pressure Coefficient With Mach Number . . . . . . . . . . . . . . . . . . . . . . 144.6.1. Effect of Mach Number at Constant Reynolds Number and Fan Drive Power. . . . . . . . . . . 144.6.2. Nitrogen Mode Performance Envelope Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154.6.3. Air Mode Performance Envelope Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

4.6.3.1. Effect of test section slot covers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164.6.3.2. Effect of downstream choke . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .174.6.3.3. Effect of fan speed or inlet guide vane variation for velocity change . . . . . . . . . . . . . . . 174.6.3.4. Comparison with other wind tunnels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

iv

4.7. Effect of Test Section Geometry Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

4.8. Effect of Liquid Nitrogen Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

4.9. Fluctuating Pressure Coefficient in Settling Chamber. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

4.10. Fluctuating Pressure Coefficient in High-Speed Diffuser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

4.11. Fluctuating Pressure Coefficient in Plenum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21

4.12. Convection Velocities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Appendix A—Detailed Description of NTF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Appendix B—Longitudinal Static Pressure and Mach Number Gradients . . . . . . . . . . . . . . . . . . . . . . 43

Appendix C—Accuracy of Test Section Pressure Gradients by Least-Squares Method. . . . . . . . . . . . 45

Appendix D—Preliminary Test Results With Steady-State Calibration ProbeInstalled in Test Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Appendix E—Estimation of Edge Tone Frequency in Free-Shear Layer . . . . . . . . . . . . . . . . . . . . . . . 51

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

v

Symbols

Amax model maximum cross-sectional area, ft2

An polynomial fit area, ft2

Ax model cross-sectional area, ft2

AT test section throat cross-sectional area, ft2 (66.77 ft2 for NTF)

a slope of least-squares straight line

b intercept of least-squares straight line

Cx longitudinal buoyancy-force coefficient,

c speed of sound, fps

wing mean aerodynamic chord, ft

ct stagnation speed of sound, fps

d1, d2 cavity depth of outer and inner chambers, respectively, of dual Helmholtz resonator,in., (fig. A14)

e base of Naperian logarithms (2.71828. . .)

Fx longitudinal buoyancy force, lb

f frequency, Hz

fp calculated fundamental frequency of slotted wall-plenum resonance, Hz

fs characteristic slot frequency, Hz

g gravitational acceleration, ft/sec2

h test section height at throat, ft (8.202 ft for NTF)

i i th data point

J highest order exponent in polynomial fit to area distribution

j upper limit of summation for least-squares straight-line fit

k1 model blockage area ratio,

k2 wing area to test section area ratio,

k3 model length to test section height ratio

L lip-to-wedge gap distance for edge tone geometry, in.

l length of model, ft

M test section free-stream Mach number,

m stage number for edge tone frequencies

N amplification factor; logarithmic exponent of disturbance growth ratio

n exponent in polynomial fit to area distribution

OA1, OA2 open area ratio of perforated liner for outer and inner chambers, respectively, of dualHelmholtz resonator (fig. A14)

P static pressure measurement, psi (fig. C1)

Pi static pressure measurement atith point, psi (fig. C1)

p static pressure, psi

pi true static pressure atith point, psi

Fx

qS------

c

Amax

AT------------

SAT-------

lh---

Vc----

vi

pmax maximum amplitude range of pressure sensor, psi

pt free-stream stagnation pressure, psi

root-mean-square value of fluctuating component of static pressure (with the meansubtracted), psi

( /q)2 power spectral density function of the fluctuating static pressure coefficient, per Hz

q test section free-stream dynamic pressure, psf

qsc dynamic pressure in settling chamber, psf

R Reynolds number

Reynolds number based on wing mean aerodynamic chord

S model wing area; other reference area, ft2

St Strouhal number

Taw adiabatic wall temperature,°FTt free-stream stagnation temperature,°FTw wall temperature,°FT∞ free-stream static temperature,°Fuc streamwise convection velocity, fps (appendix E)

uc(x) measured overall streamwise convection velocity, fps

V free-stream velocity, fps

v volume, ft3

x longitudinal distance (positive downstream), ft

αrf reentry flap angle (positive away from flow), deg

γ lag factor in edge tone feedback emittance

∆ increment or change in a quantity or variable

∆a error in least-squares straight-line slopea

δ root sum squares ofδi, psi

δi deviation between measured pressurePi and least-squares straight-line fit atith point,psi (fig. C1)

δmsw model support wall angle (positive away from flow), deg

εi error in measured pressurePi, psi

θtsw test section wall divergence angle (positive away from flow), deg

φ phase angle, deg

Abbreviations:

AEDC Arnold Engineering and Development Center

BPF blade passage frequency

dia. diameter

DRA Defence Research Agency (formerly RAE)

ESP electronically scanned pressure

FM frequency modulation

freq. frequency

p

p

Rc

vii

GN2 gaseous nitrogen

ID inner diameter

IGV inlet guide vane

IRIG Inter-Range Instrumentation Group

LaRC Langley Research Center

LHS left-hand side

LN2 liquid nitrogen

max. maximum

min. minimum

NASA National Aeronautics and Space Administration

NBS National Bureau of Standards

NIST National Institute of Standards and Technology (formerly NBS)

NLR—HST National Aerospace Laboratory—High-Speed Tunnel

NTF National Transonic Facility

OD outer diameter

RAE Royal Aerospace Establishment

re with respect to

RHS right-hand side

rms root mean square

rss root sum square

TPT Transonic Pressure Tunnel

Summary

Dynamic measurements of fluctuating static pressurelevels were taken with flush-mounted, high-frequencyresponse pressure transducers at 11 locations in the cir-cuit of the National Transonic Facility (NTF) across thecomplete operating range of this wind tunnel. Measure-ments were taken at test-section Mach numbers from 0.1to 1.2, at pressures from 1 to 8.6 atm, and at temperaturesfrom ambient to−250°F, which resulted in dynamic flowdisturbance measurements at the highest Reynolds num-bers available in a transonic ground test facility. Testswere also made by independent variation of the Machnumber, the Reynolds number, or the fan drive powerwhile the other two parameters were held constant,which for the first time resulted in a distinct separation ofthe effects of these three important parameters.

This report contains a description of the NTF andemphasizes flow quality features; details of instrumentcalibration; results of measurements with the test sectionslots covered and the downstream choke; effects of liquidnitrogen injection and gaseous nitrogen venting; compar-isons between air and nitrogen modes of operation; isola-tion of the effects of Mach number, Reynolds number,and fan drive power; and identification of the sources ofsignificant flow disturbances. The results indicate thatprimary sources of flow disturbance in the NTF may beedge tones generated by test section sidewall reentryflaps and the venting of nitrogen gas from the return legof the tunnel circuit between turns 3 and 4 in the cryo-genic mode of operation. The tests to isolate the effectsof Mach number, Reynolds number, and fan drive powerindicate that Mach number effects predominate. A com-parison with other transonic wind tunnels shows that theNTF has low levels of test section fluctuating static pres-sure, especially in the high-subsonic Mach number rangeof 0.7 to 0.9.

1. Introduction

A wind tunnel is primarily a means of creating aflow over a model or body to determine the influence ofone on the other. With the exception of specialized windtunnels for the study of specific fluid dynamic problems,the principal objective of the wind tunnel is to study theflow about configurations while duplicating full-scale,free-flight conditions to the fullest extent possible. Windtunnels such as the National Transonic Facility (NTF)fulfill the full-scale condition by achieving full-scaleReynolds numbers. In general, the free-flight condition isaddressed in a variety of ways. The interference createdby the test section walls is alleviated by wall ventilationsuch as slots or perforations; adaptive walls are some-times used in an attempt to remove the interference morecompletely. The interferences created by model supports

are minimized by support of the models from the rear bystings or blades; the interference can be removed morecompletely by the use of magnetic suspension. To date,the uniformity and steadiness of the flow has not beencompletely resolved. Few wind tunnels, if any (espe-cially transonic wind tunnels), can approach therelatively quiescent conditions of free air. Therefore,examination of the disturbance levels of wind tunnels isnecessary to assess their capability to perform diverseresearch roles.

1.1. Background

The influence of flow disturbances such as velocityfluctuation and noise on aerodynamic phenomena haslong been widely recognized; in recent years, thedynamic flow quality of wind tunnels has received closeattention. Lately, efforts to develop natural laminar flow,laminar flow control airfoils, and other aerodynamicsurfaces for use on commercial aircraft have increasedthe interest in the magnitude and the frequency character-istics of flow disturbances in wind tunnels.

In the early 1900’s, an example of the effect ofdynamic flow quality on wind tunnel measurements byPrandtl (1914) involved the discrepancy in sphere dragdata measured under comparable test conditions in thewind tunnels of Prandtl and Eiffel. Here of course, thediscrepancy was resolved after recognition that highervelocity turbulence levels in the Eiffel wind tunnel hadcaused transition of the sphere boundary layer fartherupstream, which resulted in the sphere having moreresistance to flow separation and lower drag. Animportant byproduct of discovering the cause of thesphere drag discrepancy was that wind tunnel investiga-tors should proceed with caution when applying theresults of measurements taken in turbulent wind tunnelairstreams to aircraft in nominally quiescent free air.

As used by Mabey (1976), the term wind tunnelunsteadiness is a general one which refers to fluctuationsin velocity, pressure, and temperature. Timme (1973)distinguished between acoustic (i.e., noise) disturbances,which show wave forms with a phase velocity corre-sponding to the speed of sound and turbulence, whichhas stochastic fluctuations with a phase velocity that issome fraction of the flow velocity. However, Timmenoted that the difference between the two may not alwaysbe distinct. Figure 1, adapted from Mabey (1971), showsmany of the sources of flow unsteadiness identified byMabey in transonic wind tunnels.

Although flow disturbances may have an effect onall measurements taken in wind tunnels, the effects aremore pronounced in certain types of aerodynamicresearch. As indicated by Timme (1973) and Mabey(1976), those research areas that can be most affected by

2

wind tunnel unsteadiness include boundary layer transi-tion from laminar to turbulent flow, turbulent boundarylayer development, shock-wave–boundary-layer interac-tion, separated flows and wakes, flow reattachment, inletand control surface buzz, buffeting, and flutter.

In wind tunnels, fluctuations in velocity have beengenerally considered to have the greatest influence ondynamic flow quality. Prandtl (1914) proposed that theabrupt change with Reynolds number in the drag coeffi-cient of a sphere be used to indicate a measure of thevelocity fluctuations. However, as discussed by Drydenand Abbott (1948), spheres were not reliable indicatorsfor low-turbulence wind tunnels at turbulence levelsbelow approximately 0.5 percent. Also, because of theeffects of compressibility, Robinson (1937) showed thatspheres were not suitable for high-speed wind tunnels atMach numbers greater than≈0.35. Because of theselimitations, the sphere drag test became less useful as aturbulence indicator.

As described by Dryden and Abbott (1948), the hot-wire anemometer became the standard instrument for themeasurement of velocity fluctuations. Kovásznay (1950,1953) developed the application of hot-wire measure-ment techniques for supersonic flow. Spangenberg(1955) showed that typical hot-wire sensitivities were afunction of both Mach number and Knudsen number.Morkovin (1956) improved and extended Kovásznay’stechniques, but the application for high-subsonic com-pressible flow and transonic flow remained in question.Horstman and Rose (1977) and Rose and McDaid (1976)applied the supersonic flow hot-wire techniques totransonic flow with the assumption that the hot-wiresensitivities for velocity and for density changes wereequal. However, Stainback, Johnson, and Basnett (1983)and Jones (1991) have shown that the velocity anddensity sensitivities for hot-wires are not equal. A three-wire technique of Stainback, Johnson, and Basnett and ofJones was developed to separate the effects of velocity,density, and temperature changes but has not yetreceived universal acceptance. The velocity and densityfluctuations measured with this technique appear unusu-ally large, and the reasons for these results have not beendetermined.

As shown by Meyers and Wilkinson (1982), laservelocimeters have been used to measure velocity fluctua-tion and have indicated reasonable comparisons with hot-wire technique results for the streamwise component ofvelocity fluctuation. However, the high-speed burstcounter technique of Meyers and Wilkinson has a high-amplitude threshold of detection of approximately0.5 percent. Meyers and Clemmons (1987) indicated thatthis threshold of detection can be lowered to approxi-mately 0.2 percent by using a frequency domain signal

processor, but even this amplitude threshold is notsufficiently sensitive for low-turbulence wind tunnelmeasurements.

The length of the laminar boundary layer run beforethe transition to turbulence occurs is considered a sensi-tive measure of flow quality. In a review of the influenceof flow disturbances, Michel (1988) has indicated that,although the boundary layer transition process is prima-rily sensitive to velocity fluctuations, aerodynamic soundcan control the transition process below a minimumthreshold value of approximately 0.2 percent. Dougherty(1980) has used a cone having an included angle ofrevolution of 10° to evaluate the relationship of boundarylayer transition sensitivity and wind tunnel flow qualityin a large number of wind tunnels. The Arnold Engineer-ing and Development Center (AEDC) 10° transition conewas instrumented with a traversing surface probe todetect the location of transition and with surface-mounted microphones or pressure transducers to measurethe fluctuating static pressure on the cone surface. Thiscone has been used to determine the dynamic flow qual-ity in 23 wind tunnels in both this country and Europe. Inaddition to wind tunnel tests, Dougherty and Fisher(1980) describe flight tests of the cone on the nose of anF-15 fighter airplane. With so much test history in somany research environments, the cone has attained thestatus of a calibration standard. The 10° cone testsprovided a distinct opportunity to obtain dynamic flowquality information in a large number of different testfacilities with identical hardware and instrumentation,which assured comparability of all measurements.Unfortunately, the AEDC 10° cone is not compatiblewith a cryogenic environment either in materials or inoperational capability; thus, it could not be tested in thenitrogen mode of the NTF.

The initial analysis by Dougherty (1980) of thevariation of the transition Reynolds number with themeasured fluctuating static pressure coefficient of theAEDC 10° cone showed apparent correlation betweenthe fluctuating pressure coefficient and the transitionReynolds number based on either the beginning or theend of transition. However, the transition data showconsiderable scatter when data from one facility arecompared with that from another. Spangler and Wells(1968) have further shown that the effectiveness of soundin promoting transition is highly dependent on thefrequency of the excitation. The correlation between theroot-mean-square (rms) noise level and the location oftransition is difficult because both the frequencyspectrum of the noise and the receptivity of the boundarylayer are involved. The apparent correlation byDougherty of the data from the AEDC 10° cone has beenchallenged by an analysis of the data by Murthy andSteinle (1985, 1986) in which they attempted to account

3

for the effects of Mach number in the correlation. Theyconcluded that there was no acceptable data correlationfor the AEDC 10° cone between the transition Reynoldsnumber and the fluctuating pressure coefficient, at leastat low-noise levels.

The sensitivity of an initially laminar boundary layerto flow disturbances which cause transition to turbulencebased on a comparison of the boundary layer transitionlocations calculated by theeN method of the linearcompressible stability theory with those locationsactually measured in a wind tunnel has been used byElsenaar (1990) as a measure of dynamic flow quality.The comparison yielded representativeN factors whichare characteristic of the flow for specific configurations.On a two-dimensional laminar flow airfoil in theNational Aerospace Laboratory—High-Speed Tunnel(NLR—HST), Elsenaar obtainedN factors of 6 to 12,depending on Mach number and Reynolds number;high-N factors close to free-flight values were associatedwith good dynamic flow quality.

Both Timme (1973) and Mabey (1976) havedescribed the difficulties of measuring velocity fluctua-tions at transonic speeds and have indicated that flowunsteadiness in transonic wind tunnels is usuallyestimated from measurement of the fluctuating staticpressure. Elsenaar (1990) has also indicated that themeasurement of the fluctuating static pressure in the testsection provides a first indication of the dynamic flowquality in a wind tunnel.

By comparing the static pressure fluctuationmeasured on a body of revolution on the tunnel center-line with that on a sidewall, Mabey (1971) concludedthat the pressure fluctuations were almost the same andthat the pressure fluctuation field was approximatelyone-dimensional. Dolling and Dussauge (1989) also indi-cated that fluctuating pressures measured on a wall aredependent on the surrounding flow and reflect the salientfeatures of that flow; the wall measurements have theadditional advantage of being essentially nonintrusive.

Siddon (1969) has cautioned that measurements inthe free stream of fluctuating static pressures with probescan result in significant error. The interaction betweenthe probe and the fluctuating flow can result in errors thatare either positive or negative, which depends onwhether the scale of a typical eddy size in the flow issmaller or larger, respectively, than the dimensions of theprobe. Eckelmann (1990) has also indicated that pressureprobe measurements within turbulent flows are notreliable because velocity fluctuations in the flow producerandom pressure fluctuations on the probe, which wouldnot occur if the probe were not present.

In addition to the fluctuations occurring in the freestream, wall pressure measurements include contribu-tions from disturbance levels generated within the turbu-lent boundary layer itself. Within the turbulent boundarylayer, interactions of the turbulence with the mean shearand of the turbulence with itself occur. However, Mabey(1971) has indicated that the latter contribution to wallpressure measurements represents only a small correc-tion at high frequency and is frequently approximated asa constant as was shown by Lowson (1968).

Because of the difficulties associated with othermeans of determining wind tunnel dynamic flow qualityand the advantages associated with the wall measurementof fluctuating static pressures as cited previously, thewall pressure approach was adopted for the preliminaryassessment of dynamic flow quality in the NTF.

1.2. Dynamic Measurements in NTF

Because the NTF is a wind tunnel which operates athigh Reynolds numbers, verification of dynamic flowquality was considered essential to gain confidence thatthe NTF would be suitable for laminar flow research andfor dynamic aeroelastic research such as flutter andbuffet testing. The cryogenic feature of this wind tunnel,which contributes so importantly to its high Reynoldsnumber capability, also introduces additional factorswhich affect the flow quality and complicate its measure-ment. To obtain these measurements, high-frequencyresponse pressure transducers were installed flush withthe surface at 11 locations in the NTF circuit. Dynamicmeasurements were made of the fluctuating staticpressure levels across the complete operating range ofthe NTF at test section Mach numbersM from 0.1 to 1.2,at pressuresp from 1 to 8.6 atm at temperaturesTt fromambient to−250°F, and at a maximum unit Reynoldsnumber of approximately 146× 106 ft−1. These combina-tions of test conditions resulted in data at the highestReynolds numbers available in a transonic ground testfacility.

The capability to test across a wide range of temper-atures permitted measurements of fluctuating staticpressures to be made at variable Mach number whilekeeping Reynolds number and fan drive power constantby appropriate variation of temperature and pressure.Similarly, additional tests were made at variableReynolds number while keeping Mach number and fandrive power constant, and at variable fan drive powerwhile keeping Mach number and Reynolds numberconstant, which for the first time resulted in a distinctseparation of the effects of these important parameters.

The importance of the fan drive system as a sourceof wind tunnel noise has been described in several papersincluding Williams (1977), Michel and Froebel (1988),

4

and Chiu and Lauchle (1989). As indicated by Williams,the wind tunnel fan drive system represents one of theprimary sources of background noise in the test section.According to Chiu and Lauchle, the fan aerodynamicnoise has a broadband spectrum, which sometimesincludes a series of discrete frequency peaks associatedwith the fan blade passage frequency and its harmonics.The major broadband noise sources include blade vortexshedding from the blade trailing edge, blade-to-bladevortex interaction, flow separation from the blade uppersurface, and random fluctuating blade forces caused bothby the blade boundary layer and by blade interaction withinflow turbulence. The discrete frequency peaks are pri-marily due to inflow distortion where the blades interactwith wakes from upstream obstructions. This unsteadyinteraction causes fluctuating pressure fields, which in acompressible medium radiate as dipole sound sources.

When comparing broadband dipole fan noise undersimilar operating conditions, Williams (1977) indicatedthat the fan overall sound power is approximatelyproportional to the cube of the fan tip speed times the fanaerodynamic shaft power times one minus the fanaerodynamic efficiency. To isolate the effect of blade tipspeed from that of fan drive power, the NTF wasoperated across the same wind tunnel speed and fan drivepower ranges and either fan speed change or inlet guidevane angle change was used to load the fan and changewind tunnel speed.

2. Test Apparatus

Before beginning a description of the NTF, a briefreview of the cryogenic concept may be worthwhile. His-torically, the use of modest cooling to increase Reynoldsnumber was first proposed by Margoulis (1921), and thepotential benefits of further temperature reduction werelater pointed out by Smelt (1945). Many years were topass by before the cryogenic concept would be success-fully demonstrated at Langley Research Center (LaRC)by Kilgore (1974). (Also, see Goodyer and Kilgore(1972).)

2.1. Benefits of Cryogenic Concept

The benefits of the cryogenic approach can best beillustrated in figure 2, which are taken from Kilgore,Adcock, and Ray (1974) and shown for a Mach numberof 1. In figure 2(a), the variation of the gas propertieswith temperature is shown with reference to the proper-ties at 120°F. As shown in figure 2, the density increases,and both the viscosity and speed of sound decrease withdecreasing temperature. The Reynolds number dependson density in the numerator and viscosity in the denomi-nator, and the variation of Reynolds number with tem-perature is shown in figure 2(b), again in relative terms.

In the extreme, an increase of sixfold or more inReynolds number can be obtained, although the tempera-ture is rarely taken this low in actual testing. The diffi-culty arises in approaching the gas condensationboundary too closely. A more realistic factor for theincrease of Reynolds number with temperature reductionwould be four to five. Because the speed of sound isreduced with reduced temperature, the velocity toachieve a given Mach number is also reduced so therequired fan drive power is reduced. All this is obtainedwith no change in dynamic pressure. This demonstrateswhy the cryogenic concept is an attractive way to obtainhigh Reynolds numbers. Any further increase ofReynolds numbers would normally be obtained byincreasing the stagnation pressure with the accompany-ing increase in dynamic pressure and fan drive power.

Some additional benefits of the cryogenic conceptwere pointed out by Kilgore, Adcock, and Ray (1974)and are shown in figures 3–5. These figures are concep-tual and do not represent actual performance of the NTF.Figure 3 shows that, at a constant Mach number,dynamic pressure and hence model loads and deflectionscan be held constant while the Reynolds number isvaried; conversely, Reynolds number can be heldconstant while dynamic pressure is varied for pureaeroelastic studies. Figure 4 shows that at a constantReynolds number, Mach number can be varied whiledynamic pressure is held constant, or dynamic pressurecan be varied while Mach number is held constant.Figure 5 shows that, at a constant dynamic pressure,either Mach number or Reynolds number can be variedindependently. Thus, this feature of the cryogenicconcept permits pure Mach number, pure Reynolds num-ber, or pure aeroelastic studies to be made while the otherparameters are held constant. Later in this paper, someadditional benefits of the cryogenic concept for dynamicflow quality testing will be presented, which show thatthe effects of fan drive power can be separated fromthose of Mach number and Reynolds number.

2.2. Description of NTF

The NTF characteristics have been amply describedas has the wind tunnel evolved during planning, design,construction, and initial operation. Howell andMcKinney (1977), Igoe (1980), and Bruce (1985) aretypical sources of information on the NTF and cite manyother references. A brief description of the NTF ispresented here; a more detailed description, whichincludes those components that influence the dynamicflow quality, is presented in appendix A. These descrip-tions, which are essentially a review of the previouslymentioned sources, draw material freely from them andtheir other cited references.

5

In most respects, the NTF is a rather conventionalwind tunnel with only a few unconventional features.The circuit lines and overall dimensions of the windtunnel are shown in figure 6. The tunnel circuit isapproximately 200 ft long and 48.6 ft wide betweencenterlines, which results in an internal circuit length ofapproximately 497 ft and enclosed volume of approxi-mately 230000 ft3. It was constructed on the site of thedeactivated 4-Foot Supersonic Pressure Tunnel, andincorporated the induction drive motors as well as someof the other equipment from that tunnel. The plane of thetunnel circuit is tilted approximately 9° with the center-line of the fan at a lower elevation than the centerline ofthe test section. The fan and test section centerlines lie inhorizontal planes, and the walls of the test section are ori-ented horizontally and vertically. The reason for the tiltwas to accommodate the fan driveshaft centerlinepositioning with respect to the existing induction drivemotors and to minimize extensive below-grade excava-tion requirements in the test section-plenum region.

The wind tunnel has two basic modes of operation:one at near-ambient temperatures with air as the test gasand the other at cryogenic temperatures with nitrogen asthe test gas. In the air mode of operation, cooling isaccomplished by a conventional water cooled heatexchanger inside the tunnel circuit. For cryogenicoperation, cooling is accomplished by spraying liquidnitrogen directly into the tunnel circuit. The minimumMach number is 0.1, the maximum is 1.2, and themaximum unit Reynolds number is about 146× 106 ft−1

at a Mach number of about 1.0.

2.3. Instrumentation

2.3.1. Pressure Transducer Characteristics

The fluctuating pressures were measured with minia-ture electrical pressure transducers. The transducercasing was 0.092 in. in diameter by 0.5 in. long with adifferential pressure range of±10 psi. They had an aniso-tropically etched silicon diaphragm 0.05 in. in diameterwith an active four-arm piezoresistive bridge diffusedinto the diaphragm. The diaphragm was recessed belowthe surface of the transducer under a 0.03-in-diameterorifice with a dead volume of 0.000015 in3. The speci-fied resonant frequency of the diaphragm was 130 kHz;the sensitivity to acceleration was 0.00015 psi/g. Tem-perature compensation for bridge resistance change withtemperature is accomplished with integral hybridelectrical circuitry. The transducers were temperature-compensated across a temperature range of approxi-mately−280°F to 150°F. The excitation voltage for boththe data measurement and the bench calibration was 10 Vand was continuously monitored during measurements.

2.3.2. Pressure Transducer Calibration

The primary calibration of the pressure transducerswas done in a laboratory environment with calibrationstandards traceable to the National Institute of Standardsand Technology (NIST), formerly the National Bureau ofStandards (NBS). Calibrations were performed bothstatically and dynamically at an excitation voltage of10 V.

2.3.2.1. Static calibration.The pressure transducerswere calibrated statically over their full pressure range of±10 psi at 2 psi pressure increments at temperatures from135°F to −280°F to span the expected NTF operatingtemperature range of 120°F to −250°F. The individualcalibration data at each temperature were fitted with aleast-squares straight line. The slopes of the straight-linefits for one of the pressure transducers (64RY), whichwas used in the NTF test section sidewall at station 13 onthe right-hand side, are shown in figure 7(a). The dashedline through the data points is a least-squares straight lineversus temperature. The variation of calibration sensitiv-ity with temperature was fitted this way for all thepressure transducers used in this investigation. The staticcalibration sensitivities were used to reduce the data fromthe dynamic pressure transducers because they wereconsidered the most reliable.

2.3.2.2. Dynamic check-calibration.The pressuretransducers were dynamically check-calibrated with amicrophone calibrator. The output of the calibrator wasverified with a 1/2-in. microphone as a standard. Thepressure transducers were check-calibrated at a constantfrequency of 1 kHz with input amplitudes varying from0.001 to 0.1 psi and at a constant input amplitude of0.05 psi at frequencies ranging from 50 Hz to 2 kHz.These dynamic calibrations were only performed at roomtemperature. The results of the constant frequency andconstant amplitude dynamic calibrations for transducer64RY are shown in figures 7(b) and 7(c). The variationin sensitivity of all pressure transducers with eitheramplitude or frequency variation was within the ±1/2-dBband normally expected from this type of calibration.The ordinate scale ranges in figures 7(b) and 7(c) (34to 38 mV/psi and 33 to 37 mV/psi, respectively) corre-spond approximately to a ±1/2-dB range. In addition tothe laboratory calibration, all transducer outputs wereverified in position in the wind tunnel with a portablecalibrator at a signal amplitude of 0.0015 psi and afrequency of 2 kHz.

2.3.2.3. Resonant frequency.Because of the highresonant frequency (130 kHz) of the pressure transducersand the frequency limitation of the spectral analyzerwhich was used (100 kHz at the maximum digitalsampling rate of 256 kHz in the single-channel mode of

6

operation), the resonant peaks of these transducers do notshow up in any of the power spectra plots. However, fre-quency counting on an oscilloscope trace yielded a fre-quency of approximately 137 kHz, which because of itsproximity to the resonant frequency of 130 kHz specifiedfor the transducers, is probably the actual resonantfrequency.

2.3.3. Pressure Transducer Installation

Before the pressure transducers were installed in thetunnel circuit, they were first mounted in brass instru-ment plugs as shown in figure 8. Brass was chosen as thematerial for the plugs because it is easy to machine,braze, or solder, and its thermal coefficient of expansionis similar to that of the aluminum structure of the windtunnel to which they were attached. This method ofinstallation was utilized for all the transducers except forthe two that will be described later. With guidance fromthe results of Coe (1969) and the recommendations ofHanly (1975), the pressure transducers were mountedeither flush or slightly below (0.001 in.) the surface ofthe brass plug to minimize transducer-generated flowdisturbances. A static pressure orifice and a copper-constantan thermocouple were included in the instrumentplug. The static pressure orifice was connected to thereference pressure side of the pressure transducerthrough a 100-ft coil of 0.04-in-inside-diameter flexiblepressure tubing to provide damping of the referencepressure. A short length (≈1/2 in.) of 0.01-in-inside-diameter stainless steel tubing connected the flexiblepressure tubing and the reference side of the pressuretransducer. The thermocouple was used to monitor thetemperature of the pressure transducer environment. Thepressure transducers were potted in place in the plugswith an instrument-grade silicone rubber compound.Prior to cryogenic operation, the reference pressure tubeswere purged with dry nitrogen gas.

Eight instrumented brass plugs were installed in theNTF circuit flush with the local surface, at the tunnelstation mid-height, and with the thermocouple orienteddownstream. The locations are shown in figure 9 and areas follows: one on the left-hand side (LHS) when lookingupstream at station 6.5 and another opposite to it on theright-hand side (RHS), three closely spaced (2.25 in.apart streamwise) on the RHS at station 13, one on theRHS at station 16, one on the LHS in the high-speeddiffuser at station 68, and one on the RHS just down-stream of turn 1 and adjacent to the liquid nitrogen injec-tors. An additional instrumented brass plug was installedin the plenum on the RHS at station 0 near the plenumwall and at the same height as the test section top wall.

Because of space limitations, the pressure transducerinstalled in the settling chamber was first potted into a

drilled 1/4”-20 stainless steel bolt and then installed flushon the RHS of the settling chamber wall at station−52,1 ft downstream of the last screen and approximately 30°up the wall from bottom center. The reference pressurefor the transducer was supplied from an adjacent staticpressure orifice and again damped through a 100-ft coilof flexible pressure tubing. The temperature environmentfor this transducer was assumed to be equal to the windtunnel stagnation temperature. A conventional pitot-static pressure probe was installed downstream of theheat exchanger to measure the local flow conditions inthe settling chamber.

During this investigation, the test section was nomi-nally empty; that is, there was no test model installed inthe test section. To cover up the blunt centerbody on themodel support arc-sector strut, a conical fairing withan included angle of 10.6° was installed as shown in fig-ure 10. Most of the rearward section of the fairing wasmade of fiberglass-reinforced plastic. The tip section ofthe fairing was made of stainless steel and is shown infigures 11 and 12. The apex of the cone was blunted witha 0.06-in-radius tip. The machined surface of the conefront section was estimated to have a 32-µin. finish.

A flush-mounted dynamic pressure transducer wasinstalled on the cone surface 10.4 in. downstream of theblunted cone tip. Because the transducer had a flat0.092-in-diameter face and was mounted in a conicalsurface with a radius of 1.03 in. at that point, the so-called flush mounting was not really flush. The philoso-phy, which was followed in mounting the pressure trans-ducer in the cone, was to avoid any forward-facingsurfaces or other protrusions of the transducer. Althoughthe installation was as clean as could be under thecircumstances, it still represented some discontinuity inthe cone surface. When installed in the tunnel, the conetip was on the test section centerline at station 16.74. Thetransducer was on the RHS and faced the test sectionRHS wall at station 17.60. A copper-constantan thermo-couple was installed in the cone on the side opposite ofthe pressure transducer. The reference pressure for thetransducer was the plenum static pressure and wassupplied to the transducer through a 100-ft coil of flexi-ble pressure tubing. The plenum static pressure is nearlythe same as the free-stream static pressure in the testsection across the entire operating range of the windtunnel.

2.3.4. Signal Conditioners

The signal conditioners used for the dynamicpressure transducers were all silicon, solid-state, feed-back amplifiers with gain settings of 0 to 60 dB (equiva-lent to a linear scale of 1 to 1000). The filter settingsranged from 10 Hz to 100 kHz (wideband setting) with a

7

12-dB/octave Bessel filter characteristic. The inputimpedance was 100 MΩ.

Before the signal conditioners were installed for thisinvestigation, they were carefully matched so that signalpairs that were to be analyzed with respect to phaseangle, and cross-correlation were connected to amplifierswith similar phase shift characteristics. For example, thesignal conditioners for the two adjacent pressure trans-ducers at station 13, for which phase angle and cross-correlation information are presented, had a phase shiftwithin 1° of each other across the entire range of gainsettings of the signal conditioners.

The amplifiers were operated in the manual gain-setting mode but the gain settings were acquiredautomatically on the NTF steady-state data acquisitionsystem. During the entire investigation, all of the filterswere set at the wideband setting which exceeded theupper limit of the frequency response of the FM data taperecorder.

2.3.5. Dynamic Data Acquisition

All of the dynamic pressure transducer data signalswere recorded on magnetic tape with a 28-track FM taperecorder. The tape recorder was operated in the wideband1 mode so that, at the tape recording speed of 60 ips, theresultant bandwidth was 40 kHz. All data signals werecontinuously monitored on-line with oscilloscopedisplays, and the signal conditioner amplifier gainsettings were manually adjusted to keep the recordedsignal amplitude as high as possible without exceeding±1 V peak to peak. A time code generator signal wassynchronized with the NTF steady-state data acquisitionsystem clock and was also recorded on the FM tape.

Some selected data signals were also simultaneouslyacquired on-line with a four-channel spectral analyzer.The analyzer digitized the input signal, performed a fastFourier transform on the digital data, and computedpower spectra or other frequency domain or time domainstatistical quantities, which were then stored on acomputer disk. The dynamic data acquisition instrumen-tation in the NTF control room and a simplified wiringblock diagram of the system are shown in figures 13and 14, respectively.

2.3.6. NTF Steady-State Data System

The current configuration of the NTF steady-statedata system utilizes four 16-bit, serial processor, digitalcomputers each with 2 MB of memory and a 167-MBhard disk drive. The computers are linked together andshare four magnetic tape drives. Each computer supportsspecific operations in the NTF: one computer is dedi-cated to research data acquisition and processing, another

supports data management and communication, a thirdserves as a process monitor for operation of the NTF, anda fourth is dedicated to wind tunnel control. Descriptionsof the data acquisition system are given by Fuller (1981),Boyles (1986), and Foster and Adcock (1987).

Because the NTF operates at high pressures in the airmode and at high pressures and low temperatures in thenitrogen mode, the test gas can depart significantly fromperfect gas behavior. Adcock (1976), Adcock andJohnson (1980), and Hall and Adcock (1981) haveshown that imperfect gas effects can be adequatelyaccounted for with Beattie-Bridgman-type equations ofstate solved iteratively for the appropriate gas flowparameters. The most serious departures from perfect gasbehavior in the NTF occur at low temperatures as the gascondensation boundary is approached.

All of the steady-state gas flow parameters for thisinvestigation were computed with imperfect gas effectstaken into account as indicated by Foster and Adcock(1987). As noted previously, the amplifier gain settingsfrom the signal conditioners for the dynamic pressuretransducers were acquired on the steady-state data acqui-sition system. In addition, the outputs of the thermocou-ples and the monitor signals of the excitation voltages ofthe dynamic pressure transducers were also acquired onthat system along with all the usual steady-state flowparameter measurements for the wind tunnel.

2.4. Data Accuracy

The uncertainty expected in the fluctuating staticpressure data will be considered in three categories. Thefirst category includes the free-stream parameters andother wind tunnel-related data. The second categoryincludes the actual measurement of the fluctuating staticpressure. The third category includes the statisticalreliability of the spectral data derived from statisticalanalysis of the fluctuating static pressure measurements.

2.4.1. Free-Stream Parameters

On the basis of information presented on the NTFinstrumentation by Kern, Knight, and Zasimowich(1986), on the NTF data acquisition system by Foster andAdcock (1987), and on the NTF static wind tunnel cali-bration (private communication from M. Susan Williamsand Jerry B. Adcock of the NTF staff), the estimateduncertainties for the free-stream parameters for thisinvestigation are as follows:

M .......................................................................±0.002

R, ft−1........................................................... ±0.2× 106

q, percent value.....................................................±0.1

V, fps ........................................................................±2

8

pt, percent value................................................ ±0.025

Tt, °F.....................................................................±0.1

Fan rotational speed, rpm ........................................±2

qsc, percent value..................................................... ±5

2.4.2. Fluctuating Static Pressure

The static calibration of the pressure transducersyielded least-squares straight-line fits (sensitivities) witha maximum deviation of less than±1 percent of full scaleand, generally, less than±0.5 percent, which indicatesgood linearity and very little hysteresis. The variations insensitivity with temperature were fitted with least-squares straight lines with a maximum deviation of lessthan±1 percent for all the pressure transducers.

The dynamic check-calibrations show basically thatno significant anomalies existed in the dynamic perfor-mance of the pressure transducers, at least over the rangecovered by the dynamic check-calibrations. The varia-tions in performance which were indicated by thedynamic check-calibrations are primarily a characteristicof the dynamic calibrator system which was used and arenot a characteristic of the pressure transducers (privatecommunication from John J. Chapman of the LaRCInstrument Research Division staff).

Dolling and Dussauge (1989) list several sources oferrors in fluctuating static pressure measurements withwall-mounted pressure transducers. For transducers withdiaphragm sensors mounted in cavities beneath orifices,diaphragm and cavity resonances exist that are to beavoided. For the pressure transducers used in this investi-gation, the diaphragm resonance was about 137 kHz andthe cavity Helmholtz resonance was estimated to beapproximately 75 kHz under no-flow conditions. Both ofthese frequencies are well above the 20 kHz upper cutofffrequency used for the analysis of the rms fluctuatingstatic pressure coefficient data.

For turbulent boundary layers, Dolling andDussauge (1989) indicated that a typical frequency forenergy containing eddies is of the order of the velocity atthe outer edge of the boundary layer divided by the thick-ness of the boundary layer. They stated that a safe uppercutoff frequency would be on the order of five times thistypical frequency for energy containing eddies. Thiscriterion was developed for velocity fluctuations butwas also assumed to apply to pressure fluctuations. Thetest section sidewall turbulent boundary layer at station13 in the NTF has been found by measurement to varyfrom approximately 2.5 to 4 in. in thickness; the thick-ness depends on Reynolds number and Mach number(private communication from Jerry B. Adcock of theNTF staff). For an average wall boundary layer thickness

of 3 in., the upper frequency cutoff criterion for the flowconditions of this investigation would be approximately20 kHz.

The effect of orifice size on the spatial resolution ofthe measurement is another source of error discussed byDolling and Dussauge (1989). Disturbances whose scalesare small compared with the orifice diameter tend to beaveraged out, and the spectrum is therefore under-estimated at the higher frequencies. By using the previ-ously noted upper frequency cutoff criterion, theseauthors concluded that an orifice diameter less than 0.04times the boundary layer thickness should be adequate.By applying this criterion to an average 3-in-thickboundary layer, the orifice diameter would be about 0.12in., which is large when compared with the actual diame-ter of 0.03 in. of the NTF transducers.

By using empirically determined relationships forthe one-dimensional longitudinal and lateral cross-spectral density for the wall turbulent boundary layer,Corcos (1963, 1967) developed a correction procedurefor the power spectral density as a function of a non-dimensionally (similarity) reduced frequency. The cor-rection gives the ratio of the measured to actual powerspectral density at a given reduced frequency caused bythe averaging effect of the finite size of the transducerorifice. However, both Willmarth and Roos (1965) andSchewe (1983) indicate that the Corcos correction maynot be adequate at high frequencies.

Schewe (1983) measured the wall fluctuating staticpressures beneath turbulent boundary layers withpressure transducers of different sizes, and the resultsshowed increased spatial resolution as a function of thereduced diameter of the pressure transducer diaphragmexpressed in terms of wall coordinates (i.e., essentiallythe Reynolds number of the diaphragm diameter basedon the friction velocity). Schewe concluded that a dia-phragm diameter of approximately 20 wall coordinateswas adequate to resolve the pressure structures in a tur-bulent boundary layer. For the NTF, Schewe’s criterionfor transducer diameter would impose an unusually strin-gent requirement for the pressure transducers. For theNTF sidewall installation, the orifice diameter in wallcoordinates was estimated to vary from approximately 50to greater than 6000. Only the settling chamber pressuretransducer with an estimated orifice diameter in wallcoordinates of approximately 13 could even approachSchewe’s criterion and then only at the lowest free-stream Reynolds number. The conclusion was that norealistically sized orifice diameter for the NTF pressuretransducers could have satisfied Schewe’s criterion at theNTF test conditions. For the NTF data, no correctionshave been made to the spectra for the turbulent boundary

9

layer fluctuations or for the size of the transducer relativeto the boundary layer.

The data channels for the test section RHS sidewallstation 13 and the 10.6° cone were two of the four chan-nels analyzed on-line with the spectral analyzer. Themean-square data and the 20-kHz bandwidth spectra forthese two channels on the spectral analyzer had a maxi-mum dynamic range of 76 dB. All other data wereanalyzed off-line and were limited to the maximumsignal-to-noise ratio of 47 dB characteristic of the play-back performance of the 28-track FM tape recorder.

2.4.3. Statistical Reliability

Bendat and Piersol (1980) list several factors affect-ing errors in the statistical analysis of random data.Among these factors are measurement transducers,signal conditioners, magnetic tape recorders, analog-to-digital conversion, preanalysis data conditioning, station-arity (i.e., ergodicity), finite sample length, random error,and bias (i.e., systematic) error. The latter few factors arethose of importance in statistical analysis errors. For theNTF data, stationarity is obtained by holding all of thetest conditions which are subject to control as nearlyconstant as possible during the time interval of the datasample. A long sample length on the order of 30 sec wasused for most of the data analyses; however, even longersamples on the order of several minutes were sometimesused when advantageous to do so (e.g., for some of thecross-correlation analyses).

With certain simplifying assumptions, Bendat andPiersol (1980) indicated that the normalized randomerror for a power spectral density estimate is inverselyproportional to the square root of the number of distinctaverages used in the computation. For most analyses,100 averages were used, which resulted in a normalizedrandom error on the order of 10 percent for the NTFpower spectral density data and on the order of 5 percentfor the rms data. A Hanning window was used for all thespectral analyses. For the type of random data analyzedherein, the normalized bias error is expected to be smallcompared with the random error and, because of beingfrequency specific, cannot be stated in general terms.

2.4.4. Data Repeatability

During the nitrogen mode of testing, a nominal testcondition (except for the wall-to-gas temperature ratio) atM = 0.8 andR= 40× 106 ft−1 was repeated to give a totalof five data points. The repeatability of these test condi-tions and the rms fluctuating pressure coefficient on thetest section RHS sidewall station 13 was as follows:

M ....................................................................... ±0.001

R, ft−1 ...........................................................±0.5× 106

q, percent value.....................................................±0.3

V, fps........................................................................±3

pt, percent value....................................................±0.1

Tt, °F......................................................................±1.5

/q, percent value....................................................±2

3. Test Conditions

With the exception of some preliminary tests whichwere performed during the steady-state calibration with acenterline calibration probe in the test section, thedynamic investigation was performed with the testsection empty; as noted previously, a 10.6° cone fairingcovered the model support strut centerbody. The NTFsteady-state calibration was done only with the testsection slots open, and the plenum static pressure was thecalibration reference pressure. With two exceptions,dynamic measurements were taken with the test sectiongeometry variables of wall divergence, reentry flapangle, and model support wall angle at the settings devel-oped during the NTF steady-state calibration to obtainminimum longitudinal static pressure gradients. The twoexceptions occurred when the effects of variation in walldivergence, flap angle, and model support wall anglewere being investigated and when the slots were covered.The slot covers changed the test section static pressuregradient and also rendered the plenum pressure unusableas a reference. For the latter case with the slots covered,special steps were necessary not only to obtain a satisfac-tory reference pressure but also to ascertain the effects ofthe slot covers on the static pressure gradient.

3.1. Static Pressure Gradient

The longitudinal static pressure gradients in the testsection are of interest because gradients tend to promoteinteraction between the turbulent fields of fluctuations invelocity, pressure, and temperature and can distort thepower spectra at high frequencies. For the dynamicmeasurements to be representative of aerodynamicresearch conditions, these measurements must be takenat the same flow conditions as are encountered during theaerodynamic research.

Generally, for aerodynamic research purposes, assmall a gradient as possible is desirable as a means ofmore closely duplicating free-air conditions. Whenlongitudinal static pressure gradients are encountered inaerodynamic research (e.g., force test models), the gradi-ents are usually accounted for by introduction of longitu-dinal buoyancy corrections to the force data. Because theNTF is capable of variable wall divergence, the longitu-dinal buoyancy in the NTF can be hypothetically reducedto an extremely small amount but is seldom ever actually

p

10

accomplished. One of the objectives in the NTF steady-state calibration was to determine the sensitivity of thelongitudinal static pressure gradient to wall divergenceangle. In a solid wall test section, the sensitivity is highand only small wall angle changes are needed to cancelgradients. However, in a ventilated wall test section, theflow is in intimate communication with the plenum,which is a uniform pressure reservoir. As a consequence,the longitudinal gradients are naturally very small but arealso less sensitive to a change of the wall angle.

If consideration is restricted to a linear change ofstatic pressure with longitudinal distance and if somefurther simplifying assumptions are invoked, somegeneral results on buoyancy effects are obtained as indi-cated in appendix B. These results are based on the max-imum recommended size limits for models in transonictesting as indicated by Baals and Stokes (1971) andMonti (1971). The results are shown in figure 15 as a testsection longitudinal static pressure gradient that wouldcause one count (defined as 0.0001) of buoyancy-induced incremental drag coefficient on a large model.Other gradients are obtained by linearly scaling up ordown according to the chosen allowable level of buoy-ancy drag. Figure 16 shows the same gradient expressedin terms of Mach number.

For the steady-state calibration, an extensive distri-bution of 25 static pressure orifices was available in thecenterline calibration probe. The individual pressures foreach orifice were determined with an electronicallyscanned pressure (ESP) unit with a maximum pressurerange of±2.5 psi and a specified error of no more than±0.15 percent of full range. Appendix C shows that aleast-squares straight-line fit to the longitudinal staticpressure variation determined with the ESP instrumenta-tion has similar accuracy to the lower dashed line shownin figure 15 for the gradients determined with thecenterline calibration probe. To determine the static pres-sure gradients with the slots covered, static pressureorifices in the walls were used. The same ESP unit wasused, but only 13 orifices were available, which resultedin the somewhat degraded accuracy shown in figure 15by the upper dashed line. The same gradient error infor-mation is shown is figure 16 in terms of Mach number.

3.2. Air Mode Tests

The NTF performance envelope for air mode opera-tion (private communication from Jerry B. Adcock of theNTF staff) is shown in figure 17. The boundaries areformed on the bottom by the minimum operatingpressure of about 1 atm, on the LHS by the minimumMach number of about 0.1, on the RHS by the maximumfan speed, on the top left by the maximum pressure limitof the shell of 130 psi, and on the top right by the cooling

capacity of the water-cooled heat exchanger. Thesymbols shown in this figure indicate the test conditionsat which dynamic data were obtained. These conditionsinclude Mach number variations along the minimumpressure boundary, along the maximum performanceboundary, and at a constantR = 6× 106 ft−1 as well asReynolds number variations at a constant Mach numberof 0.5.

At a constantR= 6 × 106 ft−1, the effect of changingMach number by adjustment of the inlet guide vanes witha constant fan rotational speed of 550 rpm was investi-gated atM = 0.6 to 1.0. A comparison was made at thesame Reynolds number, but the Mach number waschanged by varying the fan rotational speed and keepingthe inlet guide vanes fixed at 0°. In addition,R= 6 × 106 ft−1 andM = 0.8, the effects of variable testsection wall geometry (i.e., test section wall divergencefrom −0.3° (converged) to 0.3° (diverged), reentry flapdeflection from−1.5° (toward flow) to 2° (away fromflow), and test section wall to model support wall stepheight from 4.0 to 6.2 in. (0.08 to 0.13 as a fraction oftest section half-height)) were investigated. Some of theNTF test section geometry variables are shown infigure A10.

The effect of slot covers was investigated along theminimum pressure boundaryM = 0.2 to 0.9. The slotcovers are shown in place in figure 18. The effect of adownstream choke, as shown in figures 18 and A10, wasinvestigated both with slots open and slots covered atM = 0.8. On the 10.6° cone, the effect of free transitionwas investigated atM = 0.1 to 0.8.

At M = 0.5, the effect of Reynolds number variationfrom R = 2.9× 106 to 19× 106 ft−1 by variation in totalpressure was investigated with boundary layer transitionboth fixed and free on the 10.6° cone. For all other tests,the transition on the cone was fixed by a transition stripnear the tip. The transition strip consisted of No. 80 gritsparsely distributed in a 0.1-in-wide band 2 in. down-stream of the tip. The choice of grit size and location wasguided by the criteria given by Braslow and Knox (1958)and by Braslow, Hicks, and Harris (1966).

3.3. Nitrogen Mode Tests

Data point coverage within the performance enve-lope of the NTF in the nitrogen mode of operation isshown in figure 19. The symbols indicate the test condi-tions at which dynamic measurements were obtained. Asshown at the bottom of the figure, the Mach number wasvaried fromM = 0.2 to 1.05 atR= 6 × 106 ft−1. This testwas done in warm nitrogen to correspond to the similartest in the air mode and provide a direct comparisonbetween results in air and in nitrogen. The test points inthe nitrogen mode were taken at the same temperature as

11

those in the air mode, and the pressure was adjusted togive the same Reynolds number for each Mach number.In retrospect, a better procedure might have been toadjust the temperature to give the same stagnation speedof sound as in air and then adjust the pressure to matchthe Reynolds number. This will be examined furtherwhen the results are discussed. (See section 4.3.)

The other test points shown in figure 19 were chosento cover the operating envelope as completely as possibleunder the circumstances of limited resources of liquidnitrogen. At stagnation pressurept = 43.2 psi and stagna-tion temperatureTt = −250°F, the Mach number was var-ied from M = 0.2 to 1.0. A single point was taken atM = 0.8 andpt = 80 psi at the sameTt = −250°F. Thepoints at high Reynolds numbers were taken at near-maximum pressure andTt = −250°F. A single point wastaken atM = 1.2,pt = 20 psi, andTt = −158°F.

As mentioned previously, the special cryogenicfeatures of the NTF permitted isolation of effects such asvariations in Mach number, Reynolds number, and fandrive power. In a conventional wind tunnel, the operatingtemperature is usually fixed within relatively narrowlimits, and Reynolds number changes are obtained bypressure changes with an accompanying change in fandrive power. Generally, Mach number variations are alsoaccompanied by a change in fan drive power. However,in a cryogenic wind tunnel such as the NTF, Mach num-ber can be varied while holding Reynolds number andfan drive power constant by appropriate variation of thepressure and temperature. Similarly, Reynolds numbercan be varied while holding Mach number and fan drivepower constant, or to vary fan drive power while holdingMach number and Reynolds number constant. Theselatter two variations are indicated in figure 20, whichshows an operating envelope (private communicationfrom Jerry B. Adcock of the NTF staff) forM = 0.8. TheReynolds number variation at constant power is shownalong a constant power line of 30 MW. The fan drivepower variation is shown forR= 40× 106 ft−1. TheMach number variation at constant Reynolds number andfan drive power can be visualized as occurring normal tothe page and going through successive points at differentMach numbers, all atR= 40 × 106 ft−1 and a fan drivepower of 30 MW. In addition, figure 20 shows a varia-tion of Reynolds number by pressure variation whiletemperature is held constant atTt = −230°F and bytemperature variation while pressure is held constant atpt = 43.2 psi. Note that all of these variations passthrough a common point atM = 0.8, R= 40 × 106 ft−1,fan drive power of 30 MW,pt = 43.2 psi, andTt =−230°F. This point was repeated each time but was usu-ally approached from different conditions so that the walltemperatures were not in thermal equilibrium with thegas temperature, which permitted a limited study of the

effect of hot wall or cold wall on the measured wall pres-sure fluctuations.

The effects of hot wall and cold wall were studiedfurther during the initial cooldown of the NTF from near-ambient temperature to cryogenic temperatures. One ofthe design requirements for the structure of the NTF wasthat it be able to withstand the thermal effects of a rapidcooldown or warmup of 80°F. This rapid change oftemperature was done atM = 0.8 andpt = 25 psi, whichproduced fairly large differences in wall temperaturewhen compared with the adiabatic wall temperature. Forthese hot wall and cold wall tests, the adiabatic wall tem-perature was calculated by ignoring imperfect gas effectsand assuming a recovery factor equal to the cube root ofthe Prandtl number.

During the steady-state calibration of the NTF, somepreliminary dynamic measurements with a centerline cal-ibration probe in the test section were attempted. If therequirements for both the steady-state and the dynamicmeasurements could have been satisfied at the sametime, an efficient use of tunnel test time and of liquidnitrogen resources would have resulted. These prelimi-nary measurements are presented and discussed inappendix D.

4. Discussion of Results

The dynamic data are presented in the form of adynamic pressure coefficient /q, where is the root-mean-square (rms) value of the fluctuating static pressurereadings with the mean subtracted. The fluctuating pres-sure coefficients were computed for all outputs of thedynamic pressure transducers from the dynamic pressurein the test section except for the transducer in the settlingchamber where the local flow conditions were measuredwith a pitot-static probe. For the pressure coefficients ofthis transducer, the dynamic pressure in the settlingchamber was used.

The fluctuating pressure components were recordedin analog form on FM magnetic tape and were playedback into a spectral analyzer, four channels at a time. Theanalyzer had an upper frequency limit of 20 kHz perchannel as a result of the maximum digital sampling rateof 51.2 kHz per channel when four channels are analyzedsimultaneously. The mean-square values were obtainedby integration of the power spectra from 0 to 20 kHz.The power spectra presented in figure 21 show someof the consequences of terminating the integration ofthe spectra at 20 kHz. One of the power spectra(fig. 21(a)) is for the transducer in the test section side-wall at station 13. The other (fig. 21(b)) is for the trans-ducer in the 10.6° cone. Both power spectra are fromambient temperature air mode tests atM = 0.801 andR = 38 × 106 ft−1, which corresponds to the minimum

p p

12

pressure boundary in air. The numbers at the top of thegrid in figure 21 are the rms fluctuating pressure coeffi-cients corresponding to the integrated mean-square val-ues when the integration is terminated at that frequency.There is a 1- to 1.5-percent reduction in the coefficient asthe integration range is shortened from 40 kHz to20 kHz. The dB levels which are included in figure 21for reference are with respect to the standard 20µPa.

As indicated by Mabey (1971), disturbances propa-gating upstream from the extraction region (where thetest section flow which has entered the plenum throughthe slots is returned to the mainstream) and the high-speed diffuser are major sources of high levels of fluctu-ating static pressure in slotted transonic test sections athigh-subsonic speeds. Data presented by Mabey showthat the disturbance levels are a strong function of thelongitudinal location of the measurement; the levels arehighest near the downstream end of the test section anddiminish sharply toward the upstream end. A similarvariation occurred in the NTF test section as shown infigure 22 where RHS sidewall data at stations 6.5, 13,and 16 are shown with data from the 10.6° cone at station17.6 for aM = 0.8. Mabey’s results were obtained fromthe reduction in magnitude of a particular spectral peak.However, the results in figure 22 are for overall rmsmagnitudes and therefore show a less pronounced varia-tion. Because of the variation of disturbance level withlocation in the test section and because station 13 corre-sponds with the center of the calibrated region of the testsection and is the center of pitch rotation for modelstested at angle of attack, data for this station are used torepresent the test section disturbance levels for the NTF.

4.1. Effect of Hot Wall and Cold Wall

Whenever the test gas temperature is changed, thewind tunnel structure thermally lags the gas temperature;the greater and more rapid the temperature change, thegreater the lag. When the NTF is cooled down fromambient temperature to cryogenic temperatures, the cool-ing process can take 4 to 5 hr to avoid large temperaturedifferences in the structure and the thermal strains whichaccompany them. During this cooling process, the windtunnel flows are just high enough to promote satisfactoryheat transfer without consumption of too much liquidnitrogen in the process. However, when gas temperaturechanges are made at research conditions, the liquid nitro-gen flow rates can be much greater, and any delays instabilizing test conditions can be very costly in terms ofnitrogen consumption. During the dynamic investigation,the concern was whether differences between walltemperature and gas temperature would have a signifi-cant effect on the measured fluctuating pressures.Because temperature differences affect the wall shearstress and the thickness and stability of the boundary

layer, the question was to what extent the fluctuatingpressures would be similarly affected.

Fluctuating pressure data obtained on the test sectionRHS sidewall at station 13 during the NTF initialcooldown for this investigation are shown by the squaresymbols in figure 23. There is a tendency for the coldwall data to have greater fluctuations than the hot walldata but the differences are slight except for the point onthe extreme right in the figure. The temperaturedifferences obtained during the cooldown were greaterthan those encountered during the normal research testconditions. The data point shown in figure 20 forM = 0.8,R = 40 × 106 ft−1, pt = 43.2 psi, andTt = −230°Fwas repeated several times, which resulted in the pres-sure fluctuation data shown plotted with the circles infigure 23. The temperature differences encountered forthese data are more typical of what occurred during thedynamic investigation. Across this more limited range,the effects appear quite small and indicate that wall tem-perature differences can be ignored in the dynamic data.

4.2. Effect of Fixing Boundary Layer Transitionon 10.6° Cone

The fluctuating pressure coefficient measured on the10.6° cone with fixed and free boundary layer transitionis shown in figure 24(a). These data were taken along theminimum Reynolds number boundary in the air mode ofoperation at ambient temperature andpt = 15 psi. Asdescribed earlier, the transition strip consisted of No. 80grit sparsely distributed in a 0.1-in-wide band 2 in. down-stream of the tip of the cone. With free transition at lowMach numbers fromM = 0.1 to 0.4, the cone apparentlyhad a laminar boundary layer extending past the locationof the cone pressure transducer at 10.4 in. from the tip.Boundary layer transition from laminar to turbulent flowwas detected in time history traces of the 10.6°-conepressure transducer signal on an oscilloscope by observa-tion of the occurrence of intermittent pressure spikes. AtM = 0.5, the boundary layer was transitional at the pres-sure transducer and continued to be so up toM = 0.7, thepoint at which the boundary layer was fully turbulent anddeveloped trends similar to the results for fixed transi-tion. For the free-transition case, very low levels (as lowas 0.001) of fluctuating pressure coefficient weremeasured beneath the laminar boundary layer and veryhigh levels (as high as 0.023) were measured beneath thetransitional boundary layer.

Further effects of fixing boundary layer transition onthe 10.6° cone are shown in figure 24(b) forM = 0.5 andR ≈ 3 × 106 to 20× 106 ft−1. For R> 6 × 106 ft−1, theresults for free transition are very close to those for fixedtransition, which indicate that the boundary layer is fullyturbulent in this range. However, atR = 3 × 106 ft−1, the

13

results for free transition are influenced by a transitionalboundary layer. The lack of agreement for the repeatpoints here and in figure 24(a) is indicative of how sensi-tive the transitional boundary layer is to minor variationsin test conditions. To avoid this sensitivity and such wideand abrupt variations in transducer response as shown infigure 24, the dynamic investigation was performedmostly with the boundary layer transition fixed on the10.6° cone.

Measurements beneath a laminar boundary layerwould ordinarily be preferable because they would beuncontaminated by the higher pressure fluctuation levelsassociated with a turbulent boundary layer and wouldthereby more closely represent the fluctuation levelsoccurring in the free stream. However, such measure-ments were not possible across most of the operatingrange of the NTF because of the minimum physical sizeof the pressure transducers and the high unit Reynoldsnumber of the wind tunnel flow.

4.3. Comparison of Air and Gaseous NitrogenResults

Because air is roughly 78-percent nitrogen and bothgases behave as diatomic perfect gases at standard condi-tions, measurements in the two media could reasonablybe expected to compare well. Results for air and gaseousnitrogen, as measured on the test section RHS sidewall atstation 13, are shown in figure 25 forR= 6 × 106 ft−1.Power spectra (0 to 20 kHz) for the rms data in figure 25are shown in figure 26. The differences between thepower spectra of the air and nitrogen mode tests areprimarily broadband in nature with the exception of thepower spectra forM = 0.2 and 0.7 in figures 26(a)and 26(f), respectively. AtM = 0.2, the power spectrumfor nitrogen shows a peak at about 3.2 kHz. As discussedlater in section 4.9, this peak is thought to be due to anacoustic standing wave associated with the heatexchanger in the settling chamber. AtM = 0.7, both theair and the nitrogen power spectra show a peak at about850 Hz. Reduced bandwidth (0 to 2 kHz) power spectrafor this Mach number are shown in figure 27. Theimproved frequency resolution in this bandwidth showsthat the peak in air is at 840 Hz and in nitrogen at 855 Hz.These frequencies appear proportional to velocity andboth have approximately the same reduced frequency,which suggests that they are possibly aerodynamic inorigin.

Although the Mach numbers and Reynolds numberswere the same for the air and nitrogen data, the velocitieswere not the same. This mismatch in velocity was a con-sequence of the way the test conditions were reproduced.As mentioned previously, the Mach numbers and the

stagnation temperatures were matched, and the stagna-tion pressures were adjusted to match the Reynolds num-bers. Because of the difference in gas constants, whichare nominally 1716 ft2/sec2-°R for air and 1775 ft2/sec2-°R for nitrogen, the velocities are approximatelymismatched by the square root of the ratio of the two gasconstants. If the stagnation speed of sound had beenmatched instead of the temperatures, the velocities wouldthen have been matched. The importance of matchingvelocity lies in the fact that the frequencies of aero-dynamic disturbances such as vortex shedding or edgetones are proportional to velocity; to reproduce theseaerodynamic disturbances faithfully, the velocity shouldbe matched as well as Mach number and Reynoldsnumber. The large difference in amplitude between thetwo peaks in figure 27 raises the possibility that theaerodynamic disturbance may be coupling with anotherdisturbance that is sensitive to resonance conditions andmay be sharply tuned.

Power spectra for the settling chamber, plenum,high-speed diffuser, and liquid nitrogen injection stationare shown in figure 28 at ambient temperature for air andnitrogen atM = 0.7. Stations in the settling chamberupstream of the test section, in the high-speed diffuser,and at the liquid nitrogen injectors downstream of the testsection do not show peaks in the 850 Hz frequencyrange, which indicate that the source of this disturbanceis apparently localized in the vicinity of the test sectionand the plenum. The disturbance is present in the testsection at the same Mach number with the slots coveredas shown by the power spectrum in figure 29, which indi-cates that the disturbance is not directly connected withthe slots or the extraction region where the flow enteringthe plenum through the slots is reintroduced into themainstream. However, there may be an indirect connec-tion with the extraction region because various mechani-cal gaps exist even with the slots covered, and manypossible sources (e.g., edge tones) remain and cannot beeliminated from consideration.

The phase angle and coherence between adjacentpressure transducers spaced 2.25 in. apart streamwise onthe test section RHS sidewall at station 13 are shown infigures 30(a) and 30(b), respectively, as a function offrequency in the range from 0 to 2 kHz atM = 0.7 forboth air and nitrogen. The phase angle is shown for thedownstream transducer signal with respect to theupstream transducer signal; a positive phase shift indi-cates that the downstream signal is leading the upstreamsignal, and therefore, the disturbance is propagating inthe upstream direction. Further comment on the source ofthe 850 Hz disturbance will be reserved until the Machnumber effects at constant Reynolds number and fandrive power are discussed in section 4.6.1.

14

4.4. Effect of Fan Drive Power Variation atConstant Mach Number and Reynolds Number

Mach number, Reynolds number, and fan drivepower are three of the most influential factors affectingthe disturbance level in wind tunnels. The test matrixshown in figure 20 illustrates the way in which eitherReynolds number or fan drive power can be varied whilethe other two parameters are held constant. Although notshown in figure 20, the same can be done for Mach num-ber. The significance of this test technique is that it sepa-rates the effects of the three variables, something notpossible before the advent of the cryogenic wind tunnel.

As indicated in figure 20, the fan drive power wasvaried from approximately 24 MW to 53 MW forM = 0.8 andR= 40× 106 ft−1. The results of this test areshown in figure 31 for the test section RHS sidewall atstation 13. The variation of fluctuating pressure coeffi-cient is mostly flat with a slight tendency to rise withincreased power. The results indicate that, for these testconditions, the disturbance level as measured by the fluc-tuating static pressure coefficient on the test section side-wall is relatively insensitive to variations in fan drivepower. The variable power data in figure 31 were takenwith the fan drive system at a constant synchronousspeed of 360 rpm. The effect of blade tip speed will beexamined in section 4.6.3.3.

4.5. Variation of Fluctuating Pressure CoefficientWith Reynolds Number

4.5.1. Effect of Constant Stagnation Pressure,Stagnation Temperature, or Fan Drive Power

The matrix of test points shown in figure 20 includesReynolds number variations along three paths: constantpressure, constant temperature, and constant fan drivepower. The results of these three variations are shown infigure 32 for the test section RHS sidewall at station 13for M = 0.8. ForR> 40× 106 ft−1, the disturbance levelsare all about the same with a coefficient value of approx-imately 0.0095. At lower Reynolds numbers, the highlevels or low levels of disturbance are associated with thepresence or absence of discrete frequency peaks in therespective power spectra. The three data points atR= 40× 106 ft−1 are all essentially repeat test data at thesame values ofpt = 43.2 psi, Tt = −230°F, fan drivepower of 30 MW, andM = 0.8. As mentioned in thediscussion of hot wall and cold wall effects, the distur-bance levels could be affected by the different wall tem-peratures which occur as the data points were approachedfrom the prior run warmer or colder temperature level.However, the differences in disturbance level are slight.

The variation of disturbance level with Reynoldsnumber at constant fan drive power shown in figure 32 is

of particular significance because, as has already beenindicated, this type of data has not been previously avail-able. A principal result of this test series is that the varia-tion of disturbance level with Reynolds number atconstant fan drive power forM = 0.8 is relatively flatfrom R= 20× 106 to 50× 106 ft−1. Note that peakdisturbance levels generally occur aroundM = 0.8 (e.g.,fig. 25), and that these peak disturbance levels (fig. 32)appear relatively insensitive to Reynolds numbervariation.

4.5.2. Effect of Reynolds Number in Air

The variation of the fluctuating pressure coefficientwith Reynolds number on the test section RHS sidewallat station 13 is shown in figure 33 forM = 0.5 in air atambient temperature. The Reynolds number range ofR ≈ 3 × 106 to 20× 106 ft−1 was obtained by the variationof stagnation pressure frompt = 15 to 105 psi. As aconsequence, the fan drive power varied from 7.7 to41.4 MW. The disturbance level on the sidewalldecreased monotonically with increasing Reynolds num-ber in this range of test variables.

The disturbance level measured on the 10.6° conewith fixed transition at these same test conditions hasalready been shown in figure 24(b). Except at the lowestReynolds number, the trend of the disturbance level isupward with increasing Reynolds number. This, ofcourse, is opposite to what was observed previously onthe test section sidewall. However, the boundary layersin these two instances are noticeably very different; theboundary layer on the cone is undoubtedly very thincompared with that on the sidewall. On the cone, thedistance from the origin of the boundary layer to thelocation of the orifice is less than 1 ft. For the test sectionsidewall, the virtual origin for the boundary layer proba-bly lies somewhere in the upstream section of thecontraction as far as 50 ft from the transducer. Therefore,the Reynolds numbers based on the turbulent boundarylayer lengths of the wall and cone differ by approxi-mately 50 to 1 and may not be comparable at all witheach other.

4.6. Variation of Fluctuating Pressure CoefficientWith Mach Number

4.6.1. Effect of Mach Number at ConstantReynolds Number and Fan Drive Power

As mentioned in the description of the tests, appro-priate variation of pressure and temperature will result ina variation of Mach number if Reynolds number and fandrive power are held constant. The significance of thistest technique is that it permits an isolation of the effectsof Mach number from those of Reynolds number and fan

15

drive power, which have already been noted as beingthree of the most influential factors in wind tunnel distur-bances. The fluctuating pressure coefficients measuredon the test section RHS sidewall at station 13 are shownin figure 34 for constantR= 40× 106 ft−1 and fan drivepower of 30 MW forM = 0.6 to 1.0. The disturbancelevel variation with pure Mach number variation is simi-lar to that shown previously in figure 25 and confirmsthat, for these and all other test section results, the distur-bance level typically peaks at high-subsonic Mach num-bers near 0.8 and falls off as a Mach number of 1.0 isapproached. The falloffM > 0.8 may be at least partiallydue to a choking effect that prevents downstream distur-bances from propagating upstream into the test section.At near-sonic speeds, all results tend to converge to alower level of the fluctuating pressure coefficient ofapproximately 0.0055.

The behavior of the fluctuating pressure coefficientat M = 0.7 to 0.8 in figure 34 is similar to what wasshown in figure 25 in the air-nitrogen comparison. Thereasons are again found in the power spectra that areshown in figure 35. Frequency peaks in the 0.8- to1.0-kHz range occur in almost all data sets, particularlyat 860 Hz forM = 0.694, at 900 Hz forM = 0.742, andsomewhat less prominently at 960 Hz forM = 0.793 infigures 35(b), 35(c), and 35(d), respectively. For theseMach numbers, the phase shift between adjacent pressuretransducers at station 13, which is shown in figure 36(a)along with the coherence in figure 36(b), again indicatedan upstream propagation of disturbances at thesefrequencies. Figure 35(g) forM = 0.992 also shows adisturbance at 960 Hz, which indicates that the distur-bances at these frequencies were not being choked offand therefore probably did not originate downstream ofthe test section. The disturbances did not appear in thepower spectra for the settling chamber, which supportedthe indications that the source is probably localized in thetest section or plenum. The most likely area of origin isthe extraction region at the downstream end of the testsection.

The reduced frequencies shown in figure 35 are notconstant. The lack of constancy may be associated withtest conditions which required that constant Reynoldsnumber and constant fan drive power be achieved simul-taneously. Both the temperature and the pressure had tobe varied across a fairly wide range, the pressure from 71to 34 psi and the temperature from−174°F to −250°F.The temperature changes cause thermally inducedchanges in the dimensions of the test section; the pres-sure changes can contribute to dimensional changes aswell. Thus, if the extraction region of the test section isinvolved in the disturbances, then the dimensionalchanges can be responsible for changes in frequency,especially if the disturbances are associated with edge

tone effects. As in the case of the air-nitrogen compari-son and in the absence of more information, the probablesource of these disturbances is speculative.

An estimation of possible edge tone frequenciesassociated with the geometry of the sidewall reentry flapsat the downstream end of the test section near station 20is presented in appendix E. While not conclusive, thefrequency estimation can be viewed as supportive of thepossibility that the disturbance peaks in the power spec-tra of figures 27 and 35 are caused by edge tones.

Within the range of the three variables consideredhere (i.e., Mach number, Reynolds number, and fan drivepower), Mach number has the greatest effect on the fluc-tuating pressure coefficients. This result should not beinterpreted as indicating that Reynolds number and fandrive power are unimportant in affecting disturbancelevels but rather that the coefficient formed by dividingthe rms fluctuating pressure by the dynamic pressureserves to collapse some of these effects, particularly inthe case of fan drive power. This collapsing influence ofthe dynamic pressure also occurs on the Mach numbereffects as would become apparent if the coefficient wereformed by dividing by the static pressure instead of thedynamic pressure as is sometimes done.

4.6.2. Nitrogen Mode Performance EnvelopeResults

The fluctuating pressure coefficients measured onthe test section RHS sidewall at station 13 for the testpoints in figure 19 for the nitrogen mode of operation areshown in figure 37. The data for the maximum Reynoldsnumber boundary and for a constantpt = 43.2 psi wereobtained at a constantTt = −250°F. The data in figure 25for R= 6 × 106 ft−1 at ambient temperatures are includedhere for comparison. A single data point obtained atM = 1.2 andR = 14.3× 106 ft−1 is also included.

To show frequency content, power spectra for thehigh Reynolds number data in figure 37 are presented infigures 38 and 39. For those Mach numbers where com-parable data exist, the power spectra for the maximumReynolds number boundary andpt = 43.2 psi are verysimilar. Figure 38 shows power spectra from 0 to20 kHz; figure 39 shows the same data across a reducedbandwidth of 0 to 2 kHz. ForM = 0.2 and 0.4, thefrequency peak at about 2 kHz is thought to result froman acoustic standing wave associated with the heatexchanger in the settling chamber and will be discussedlater in section 4.9. AtM = 0.6 and 0.7, the frequencypeaks at about 800 Hz are thought to be associated withedge tones originating at the sidewall reentry flaps.

The previously shown insensitivity of the distur-bance coefficient levels to fan drive power and Reynolds

16

number at aM = 0.8 is reflected in the results shown infigure 37. The data for the constantpt = 43.2 psi and forthe maximum Reynolds number boundary show closeagreement at aM = 0.8 despite a fan drive power incre-ment from 26.7 to 76.6 MW and a Reynolds numberincrement from 46.0× 106 to 132.4× 106 ft−1. The close-ness of agreement across the rest of the Mach numberrange prompts speculation that the demonstrated insensi-tivity may not be limited to justM = 0.8 but may occurmore widely. Note that the peak fluctuating pressurecoefficient /q = 0.0953 at M = 0.8 and R= 132.4×106 ft−1 corresponds to a peak sound pressure level of161.4 dB re 20µPa.

One result of the apparent insensitivity of the NTFflow disturbance level to Reynolds number may be thepossible absence of what is referred to by Elsenaar,Binion, and Stanewsky (1988) as a pseudo-Reynoldsnumber effect. (Also, refer to the discussion by Bobbitt(1981) on unit Reynolds number effects.) This effect isattributed to the variation of the wind tunnel disturbancelevel with wind tunnel Reynolds number. As alreadynoted, the wind tunnel disturbance level can alter thelocation of the boundary layer transition and cause falseresults if the disturbance level varies when the windtunnel Reynolds number is varied. Elsenaar, Binion, andStanewsky indicate that this pseudo-Reynolds numbereffect occurs most readily if the location of the boundarylayer transition is not fixed; however, it can also occurwhen the boundary layer transition is fixed. The apparentinsensitivity of the flow disturbance level to Reynoldsnumber in the NTF does not completely ensure thatpseudo-Reynolds number effects will not occur in thiswind tunnel but is clearly a favorable indicator.

4.6.3. Air Mode Performance Envelope Results

The fluctuating pressure coefficients have been mea-sured as a function of Mach number in the air mode atthe Reynolds number ranges indicated in the perfor-mance envelope in figure 17. The results for the mini-mum and maximum Reynolds number boundaries arepresented in figure 40 for the test section RHS sidewallat station 13. There is a tendency for more separation ofthe data with Reynolds number in the air mode than wasobserved in figure 37 for the nitrogen mode, and theoverall level near the peak atM = 0.8 is lower. Further, incontrast with the nitrogen mode results, the maximumReynolds number boundary results for the air mode areeverywhere lower than those for the minimum Reynoldsnumber boundary. Generally, the Mach number effectsare quite similar to what has been observed previously.

4.6.3.1. Effect of test section slot covers.Ventilatedwall test sections tend to be much noisier than compara-ble solid wall test sections. Overall, slotted wall test

sections tend to be quieter than perforated wall testsections. In perforated wall test sections, the primaryadditional noise source tends to be edge tones associatedwith the perforation holes. In slotted wall test sections,the primary additional noise sources are the free-shearlayers in the slots and the extraction region of the testsection where plenum flow reenters the mainstream. Bycovering the slots, both of these additional noise sourcesare eliminated. The slot covers that were used are shownin place in the test section in figure 18.

With the slots covered and the choke off, the testsection wall divergence angle was set at 0.1° on the topand bottom walls; the sidewalls remained parallel. Thesewall settings resulted in a slight positive static pressuregradient. Quantitatively in terms of Mach number,the gradients varied fromdM/d(x/h) = −0.00010 to−0.00595, or in terms of equivalent buoyancy-induceddrag coefficient increments in figure 16, from less than ahalf count (0.00005) to somewhat more than four counts(0.0004) of negative buoyancy drag coefficient in theMach number range of 0.2 to 0.9. For reference, thenormal operating conditions for the NTF with slots openresult in less than one count of buoyancy-induced dragcoefficient across the entire operating range.

The fluctuating pressure coefficients measured onthe test section RHS sidewall at station 13 with the slotscovered are shown in figure 41. The data were takenfrom M = 0.2 to 0.9 along the minimum Reynolds num-ber boundary in air. Data taken along the same boundarywith the slots open are also shown for comparison. Thereduction in disturbance level with slots covered occursonly at the high-subsonic Mach numbers (M > 0.6). Thischaracteristic may be connected with incompletewave reflection at the slotted wall-plenum interface,which allows test section disturbances to pass throughinto the plenum and become dissipated atM < 0.618. Asnoted previously in the discussion of the air-nitrogencomparison, the power spectrum with the slots covered(fig. 29) shows a disturbance peak at 850 Hz. The pres-ence of these disturbances with the slots covered elimi-nates any direct connection between these disturbancepeaks and the slot flow or the reentry process in theextraction region. This result is supportive of the proba-bility that edge tone effects associated with the sidewallreentry flaps are responsible for the large peaks occurringin the power spectra in figures 27 and 35.

Power spectra (0 to 20 kHz) with the slots open andthe slots covered are shown in figure 42 at the tunnelconditions plotted in figure 41. ForM ≤ 0.6, the powerspectral densities with the slots open are slightly higherthan with the slots covered at low frequencies but arelower at high frequencies. ForM > 0.6, the power spec-tral densities at low frequencies for the slots open are

p

17

significantly higher than with the slots covered but arestill lower at the high frequencies. The most significantdifference in power spectra apparently occurs fromM = 0.7 to 0.9 at the low frequencies from 0 to≈1 kHzwhere the power spectral densities with slots open aremuch higher than with slots covered.

4.6.3.2. Effect of downstream choke.A major con-tributor of broadband noise at low frequencies is thenoise propagating upstream from the diffuser and modelsupport sections into the test section. To investigate theeffect of a downstream choke, the variable geometry fea-tures of the NTF test section were used to create a mini-mum flow area at the downstream end of the test section.As shown in figure A10, the minimum flow area waslocated at the hinge line of the top and bottom wallreentry flaps, which created a two-wall choke atstation 25. The area was sized to choke the flow at thislocation when the test sectionM ≈ 0.8. Although the testsection geometry is capable of being fully variable whilethe tunnel is running, the test section wall angle, themodel support wall angle, and the reentry flap angleswere all preset before tunnel start-up and were not variedduring the choke runs. The choke geometry was set bothfor slots-open and slots-covered conditions. Because ofdifferences in test section wall boundary layer growthwith the slots open and covered, the preset wall geometrywas not identical for the two conditions. The test sectionwall divergence angle was set to accommodate the calcu-lated boundary layer growth for the closed wall configu-ration to minimize the longitudinal static pressuregradient, and the reentry flap angles were set to blendwith the test section wall. The wall geometry settings forthe different runs are summarized in table I. The wallgeometry at the downstream end of the test section isshown in figure 43. The settings are pictured for theslots-open condition. The photograph was taken whenthe wall geometry settings were rehearsed prior to theactual dynamic investigation. The sting configurationshown in the photograph was for a model test which wasin preparation at the time and was removed for thedynamic investigation.

The effect of the downstream choke is shown in fig-ure 41 with the flagged solid symbols. In operation with

the choke in place, the wind tunnel speed was increaseduntil further increases in fan drive power did not result inany further increase in wind tunnel speed as shown infigure 44. The relatively small decrease in disturbancelevel with the choke deployed (on the order of 0.001 incoefficient) may be an indication that disturbances origi-nating downstream of the test section do not contributegreatly to the disturbance level in the test section. Thepower spectra for the configuration with the slots open,both choked and unchoked, are shown in figure 45 forM = 0.8. The reduction in disturbance levels due to thechoke occurs mainly at the low frequencies between 0and 5 kHz. From the data shown in figures 41 and 45, theuse of a two-wall downstream choke to reduce flow dis-turbance levels in the NTF test section resulted in onlymarginal improvements. However, a different choke con-figuration might have been more effective.

4.6.3.3. Effect of fan speed or inlet guide vane vari-ation for velocity change.As described in appendix A,the NTF tunnel has two relatively independent means ofchanging tunnel speed. In normal wind tunnel operation,when only the power of the induction motors is requiredand the synchronous motor is not energized, the windtunnel speed can be changed by either fan speed or inletguide vane (IGV) angle variation; the method dependsupon circumstances. The preferred mode of operation isto select a fixed fan speed that can be maintained whilewind tunnel speed is varied over the desired range usingIGV variation. This is especially true when the windtunnel is operated automatically under computer control.Wind tunnel speed changes can be made much morerapidly by using IGV variation than by using fan speedvariation. When the additional power of the synchronousmotor is required and the fan speed is fixed at synchro-nous speed, then IGV variation must be used for windtunnel speed changes.

A brief test was made to determine if the test sectiondisturbance level would be affected by operation in eitherone or the other of the wind tunnel speed-changingmodes. AtR= 6 × 106 ft−1 in the air mode, the wind tun-nel speed was changed fromM = 0.2 to 1.0 with fanspeed variation from 160 to 595 rpm with the IGV fixedat 0° (neutral position). At the same test conditions, the

Table I. NTF Test Section Wall Geometry Variables

Test section configuration Wall angle, deg, at—Slots Choke Test section Model support section Reentry flap angle, deg Mach number range

Covered On 0.1 −4.23 0.87 0.2 to 0.8 Covered Off 0.1 −3.79 −0.1 0.2 to 0.9 Open On 0 −4.23 1.86 0.2 to 0.8 Open Off 0 −1.76 0 0.2 to 1.05

18

wind tunnel speed was also varied fromM = 0.6 to 1.0with IGV variation from 25° (fan unloaded) to−20° (fanloaded) at a fan speed of 550 rpm. The results are shownin figure 46 for the test section RHS sidewall atstation 13. The close agreement between the two sets ofresults would indicate that fluctuating static pressurelevel is not dependent upon the method of wind tunnelspeed changes or the combinations of IGV settings andfan speed settings used to set a particular Mach number.

The results of this IGV versus fan speed investiga-tion also provide some information on the noise charac-teristics of the NTF fan system. As noted previously, fansound power is usually considered proportional to the fanpower times the cube of the blade tip speed. Although theinflow velocities at the fan were not measured, theyshould be essentially a function of Mach number for thetest conditions in figure 46 and be fairly similar for boththe variable speed and the variable IGV data points.Because the blade tip speed is obtained by a vector reso-lution of the fan rotational speed and the inflow velocity,its variation over the Mach number range is different forthe variable speed and the variable IGV data points. Fordata points in figure 46 atM < 0.9, the blade tip speedwould be higher for the variable IGV data compared withthe variable speed data, and the opposite is true forM > 0.9. However, the data for the disturbance levels donot show a similar tendency. From the results infigures 31 and 46, the disturbance level in the NTF testsection appears to be insensitive to variations in eitherthe blade tip speed or the shaft power of the NTF fandrive system.

4.6.3.4. Comparison with other wind tunnels.Thefluctuating pressure coefficients measured on the NTFtest section RHS sidewall at station 13 for the minimumReynolds number boundary in air (atmospheric stagna-tion pressure and ambient temperature) and plotted infigure 40 are replotted in figure 47. Data from Jones(1991) for the Langley 8-Foot Transonic Pressure Tunnel(8-Foot TPT) at similar test conditions are shown forcomparison. The Langley 8-Foot TPT data are from apressure transducer located on the test section LHS side-wall at a station corresponding to the location of testmodels. The LHS sidewall of the Langley 8-Foot TPT,which is downstream of the inside corner of turn 4 issimilarly positioned to the RHS sidewall of the NTF.

No test model was in the Langley 8-Foot TPT at thetime of Jones’ (1991) measurements, but a nose conesupporting five probes was mounted on the centerlinemodel support system of the wind tunnel. This modelsupport system regularly utilizes a pair of guy wiresdownstream of the model location to provide lateralrestraint for the sting support system. A frequency spikecaused by vortex shedding from these guy wires was

identified by Jones in the spectra of the Langley 8-FootTPT fluctuating pressure data. Because guy wires arenormally present for conventional model testing in theLangley 8-Foot TPT, their influence is a normal part ofthe flow disturbance measurements in that wind tunnel.Jones did not indicate to what extent the guy wire inter-ference may have affected the overall level of themeasurements.

The data for the Langley 8-Foot TPT sidewall showthe same characteristics as observed for the NTF data,that of peaking at high-subsonic Mach numbers near 0.8and falling off steeply as sonic speeds are approached.Both wind tunnels show similar disturbance levels on theorder of 0.6 percent at low-supersonic speeds. At thepeak nearM = 0.8, the level is approximately 1.5 percentfor the Langley 8-Foot TPT and 0.8 percent for the NTF.

For reference, the fluctuating pressure coefficientdata measured in the NLR—HST by Ross and Rohne(1973) are also shown in figure 47. The NLR—HST datawere measured on the AEDC 10° cone which isdescribed by Dougherty (1980). The HST data appear torepresent a maximum envelope of disturbance level forthe test conditions in that wind tunnel. These data maynot be directly comparable with the Langley 8-Foot TPTor NTF data because of the differences in the methods ofmeasurement. However, all three wind tunnels appear tohave relatively quiet flows. The peak level for theNLR—HST is approximately 1 percent. On the basis ofthe data in figure 47, the NTF has low levels of test sec-tion fluctuating static pressure as measured on the testsection sidewall, especially in the high-subsonic Machnumber range from 0.7 to 0.9.

As mentioned previously, wall pressure fluctuationsmeasured beneath a turbulent boundary layer are influ-enced by disturbance levels generated within the turbu-lent boundary layer itself. There is an interaction of theturbulence with the mean shear and an interaction of theturbulence with itself. These disturbance levels representa floor or minimum level that can be measured on a wall.Lowson (1968) has derived the empirical expression forestimating this minimum level for attached equilibriumturbulent boundary layers of

(1)

which is also shown in figure 47. Most of the wind tunneldata are above this line.

4.7. Effect of Test Section Geometry Variables

As indicated in figure A10, the NTF test sectiongeometry variables consist of variable top and bottomtest section wall divergence angles, variable reentry flap

pq--- 0.006

1 0.14M2

+----------------------------=

19

angles, and variable top and bottom model support sec-tion wall angles. The role of the test section wall diver-gence in controlling longitudinal static pressure gradientshas already been mentioned. The reentry flaps can simi-larly control pressure gradients near the downstream endof the test section. The model support section wall anglevariation was used in the downstream choke test to forma minimum flow area at the location of the reentry flaphinge line. All three variables affect the wind tunnelpower consumption. The results of the steady-state cali-bration (private communication from M. Susan Williamsand Jerry B. Adcock of the NTF staff) were used to selectsettings of these geometry variables for normal operationof the wind tunnel, and these settings were used for thedynamic measurements as well.

The effect that the test section geometry variableshave on the disturbance level in the test section wasinvestigated briefly by varying each setting through asmall range while the other two were held fixed. Theresults are shown in figure 48 for the test section RHSsidewall at station 13 forM = 0.8 andR= 6 × 106 ft−1.With the exception of the test section wall divergenceangle of 0.3°, all the effects are slight. Broadband (0 to20 kHz) power spectra for the rms data in figure 48 areshown in figure 49. For the data point at 0.3° wall diver-gence in figure 48(a), the power spectrum in figure 49(a)does not show any frequency spikes, only a small broad-band increase in disturbance level in the frequency rangefrom ≈100 Hz to≈2 kHz. Figure 48 shows that the lowestlevels of disturbance are obtained at test section wallangles from parallel to slightly converged, reentry flapangles away from the flow, and model support wallangles toward the flow.

4.8. Effect of Liquid Nitrogen Injection

A process capable of spraying as much as 1000 lb/sec of volatile liquid in a confined space has the potentialof having a significant influence on the test section dis-turbance levels. Tests in the air mode and in the nitrogenmode provided an opportunity to compare the distur-bance levels at the liquid nitrogen station both with andwithout injection but at otherwise substantially the sametest conditions. This comparison is shown in figure 50 atR= 6 × 106 ft−1. The pressure coefficient data are plottedas a function of the test section Mach number. Becausethe liquid nitrogen injection station is downstream of thetest section, the choke effect at the test section does nottend to reduce the disturbance levels as test section sonicspeeds are approached, and the disturbance levelscontinue to rise as the Mach number is increased. Powerspectra for the rms data in figure 50 are shown infigure 51.

At M > 0.6, the disturbance levels are greater innitrogen than in air. A comparison of the power spectrafor these conditions (figs. 51(e)–51(l)) showed that theincrease was primarily broadband with no apparent par-ticular frequency selectivity. The frequency peaks thatwere so prominent in the power spectra of the test sectionpressure transducers at these test conditions (figs. 26and 27) were not evident in the power spectra at theliquid nitrogen injection station.

To gain further insight into the effect of liquid nitro-gen injection, the output of the dynamic instrumentationwas continuously recorded as the injection process wasabruptly turned off. The initial test conditions for thecutoff test wereM = 0.8,R= 12.6× 106 ft−1, pt = 20 psi,and Tt = −160°F. A playback of the continuouslyrecorded data is shown in figure 52 for the settling cham-ber, the test section RHS sidewall at station 13, the high-speed diffuser, and the liquid nitrogen injection station.The initiation and completion times for the cutoff ofinjection are shown on the upper grid line. The totalcutoff time from initiation to completion took about12 sec. The time for a disturbance to propagate byconvection completely around the tunnel circuit at thistest condition has been estimated to be slightly less than7 sec. The effects of the nitrogen cutoff are so impercep-tible in figure 52, and the moment of cutoff is nearlyimpossible to detect from the transducer signals. As willbe discussed in section 4.9, the settling chamber distur-bance level may be influenced by the gaseous nitrogenexhaust which is automatically controlled by the windtunnel control process to maintain stagnation pressurewhen the nitrogen injection is stopped.

During the cutoff procedure, the wind tunnel controlsystem maintained the Mach number and the stagnationpressure. The stagnation temperature increased rapidlyand the test was terminated after a temperature increaseof 25°F. Because the Mach number was being held con-stant, the velocity increased with the temperature.Because of the rapid increase in temperature and velocityfollowing the nitrogen cutoff, the test conditions were nolonger completely constant, and the statistical analysismethods used herein were no longer strictly appropriate.However, because the disturbance amplitude did notshow drastic changes as seen from the time history tracesin figure 52, a short relatively stationary time sample ofabout 10 sec before and after cutoff was analyzed forpower spectral content and rms level. The power spectraare shown in figure 53 for the same four wind tunnel sta-tions whose time traces are shown in figure 52. The rmslevels listed on the power spectra indicate that the set-tling chamber disturbance level decreased slightly whenthe liquid nitrogen injection was stopped and either

20

remained the same or increased slightly for the otherthree stations. Sound pressure levels in dB re 20µPa arealso shown in figure 53.

The apparent lack of influence of the liquid nitrogeninjection process on the level of flow disturbancesdetected in the test section may be associated with thepresence of suspended droplets in the liquid nitrogenspray. Such a droplet suspension could be inhibiting theupstream propagation of broadband fan noise similar tothe attenuation of sound propagation in atmospheric fogand partially offsetting any direct noise created by theinjection process.

4.9. Fluctuating Pressure Coefficient in SettlingChamber

The fluctuating pressure coefficients for the settlingchamber at the test conditions along the maximumReynolds number boundary in the nitrogen mode areshown in figure 54. In this and subsequent figures show-ing the settling chamber disturbance levels, the fluctuat-ing pressure coefficients are formed by using thedynamic pressures in the settling chamber and are plottedas a function of the Mach number or Reynolds number inthe test section. Although the coefficient levels arehigher in the settling chamber, a significant resemblanceexists between figure 54 and figure 37, which showed thedisturbance level in the test section for the same test con-dition. For the settling chamber, a sharp drop-off of thefluctuating pressure coefficient occurs from approxi-mately 0.275 atM = 0.8, to approximately 0.125 as sonicspeeds are approached in the test section. The similarcharacteristics shown in figures 37 and 54 suggest thatthe disturbance levels in the settling chamber are stronglyaffected by the levels in the test section, which indicatethat these disturbances may originate in the test sectionor further downstream and propagate upstream from thetest section into the settling chamber. This supposition isfurther supported by the effect of the downstream choke,which is shown for the minimum Reynolds numberboundary in air in figures 55(a) and 55(b) with the testsection slots open and covered, respectively. When thedownstream second minimum cross section is activelychoking flow, the fluctuating levels drop to the samelevel as when sonic speeds are approached, which mirrorthe results of the test section. (See fig. 41.) However,there is a significant difference between the levels for themaximum Reynolds number boundary (fig. 54) and theminimum Reynolds number boundary (fig. 55(a)), whichraises the question of whether other influences (e.g., pos-sibly fan noise and other disturbances caused by thewide-angle diffuser, the heat exchanger, and the screens)are present in the settling chamber as well and bias theresults from minimum to maximum Reynolds number.

Some possible sources of disturbance will be exam-ined next to see if they could be responsible for the largedifferences between the rms data in figures 54 and 55.The settling chamber power spectra for the maximumand minimum Reynolds number boundary atM = 0.8 areshown in figure 56. A comparison of the power spectrashows that at frequencies above≈2 kHz, the higher levelsof disturbance for the maximum Reynolds numberboundary are primarily broadband. However, the contri-bution at these frequencies to the overall power is slight.The major differences between the maximum and mini-mum Reynolds number boundary power spectra are inthree broad peaks with most of their power concentratedbelow ≈1.2 kHz. The lowest of the three broad peakscontains a small peak at the blade passage frequency of150 Hz and another small peak at 110 Hz, which is prob-ably associated with vortex shedding from the tubes ofthe heat exchanger. The small peak at 35 Hz is probablyassociated with vortex shedding from some of the heatexchanger support structure, which consists of verticalplates on either side of the individual tube bundles in theheat exchanger. The minor peak at 2165 Hz is probablyassociated with vortex shedding from the screen wires.The tentative frequency identifications just referred toand that follow are based on the assumed Strouhalnumber for the vortex-shedding characteristics of eachcomponent.

At the maximum Reynolds number boundary testcondition, the fan was operated at its synchronous speedof 360 rpm. The 25 blades of the fan produce distur-bances at the fundamental blade passage frequency(BPF) of 150 Hz and harmonics at integral multiples ofthat frequency. Except for the fundamental tone at150 Hz, none of the blade passage harmonic frequenciesare apparent in the spectrum for the maximum Reynoldsnumber boundary. At the minimum Reynolds numberboundary test condition, the fan was operated at its maxi-mum speed of 600 rpm. The fundamental BPF of 250 Hzis barely evident among other minor peaks in that fre-quency range. The first, second, and third harmonics ofthe BPF at 500 Hz, 750 Hz, and 1000 Hz are evident buthigher harmonics are not.

For the rest of the low Reynolds number boundarypower spectrum, the only peaks which can be tentativelyidentified are again the following vortex-sheddingfrequencies: at 50 Hz from the support structure of theheat exchanger, at 210 Hz from the heat exchanger tubes,and at 3420 Hz from the screen wires. These tentativelyidentified sources for the minimum and maximumReynolds number boundaries represent only a smallfraction of the total power in the respective spectra. If thesource identifications are correct, these sources cannotby themselves be responsible for the differences in thedisturbance levels, and some other as yet undetermined

21

source must be a factor. Note that disturbances from thewide-angle diffuser have not been eliminated fromconsideration, but these disturbances would likely bebroadband in nature and not cause the discrete peakswhich have been observed. Another possibility is theventing region in the crossleg between turns 3 and 4. Forthe minimum Reynolds number boundary in air, the ventvalves are normally closed. However, the valves are openfor the maximum Reynolds number boundary in nitrogenand vent a mass flow at a rate equal to the liquid nitrogeninjection rate.

The tests of air versus nitrogen provide an opportu-nity to compare the settling chamber disturbance levelsboth with and without venting at otherwise similar testconditions. The fluctuating pressure coefficients in thesettling chamber for these two tests are shown infigure 57 for a test sectionR= 6 × 106 ft−1. Power spec-tra at selected Mach numbers for the test points infigure 57 are shown in figure 58. Because the test condi-tions for the air and the nitrogen mode tests are nearlyidentical, the differences between the tests should bedirectly attributable to the nitrogen injection process andthe accompanying venting. The data in figure 50 showedthat the injection process did not have much effect at theliquid nitrogen injection station. The injection processcan be assumed to have even less effect at the settlingchamber, so the differences that are observed can beattributed primarily to the venting.

The fluctuating pressure coefficient data in figure 57show an increase in the disturbance level across theMach number range for the nitrogen test. AtM ≥ 0.6, thepower spectra in figure 58 show both narrowband andbroadband increases. AtM = 0.2 in the nitrogen mode(fig. 58(a)), the peak at 3.2 kHz is thought to be due to anacoustic standing wave associated with the heatexchanger in the settling chamber. This tentative identifi-cation is based on the observation that the frequency didnot vary with velocity changes but did vary approxi-mately with the square root of the absolute temperature,and the lateral spacing of the heat exchanger tubes wasabout right for a standing wave of this frequency. AtM = 0.2 and 0.4, the air mode test data show major fre-quency peaks at 14.8 kHz and 15.3 kHz, respectively,which are not present in the nitrogen mode test data. Noinformation is currently available on the noise character-istics of the vent region on which to base any furthercomment. This area of the wind tunnel circuit and theventing process require further study.

Figure 59 shows comparative data taken at twodifferent test conditions in the nitrogen mode: one atconstantR= 40× 106 ft−1 and constant fan drive powerof 30 MW and the other at constantpt = 43.2 psi and con-stantTt = −250°F. The fluctuating pressure coefficients

in the settling chamber are relatively insensitive to thedifferences in the test conditions in this intermediateReynolds number range. Data taken across a broad rangeof Reynolds numbers atM = 0.8 are shown in figure 60for three test conditions: constant drivepower of 30 MW,constant stagnation temperature of−230°F, and constantstagnation pressure of 43.2 psi. These data are also rela-tively insensitive to the difference in the three test condi-tions but show an increasing level of disturbance withincreasing Reynolds number. The effect of the variationof fan drive power is shown in figure 61 forM = 0.8 andR= 40× 106 ft−1. The disturbance levels are relativelyinsensitive to changes in fan drive power with only aslight tendency to decrease with increasing fan drivepower. Figures 59–61 show that the settling chamberresponses are similar to those observed in the test sectionwhere disturbance levels are a strong function of Machnumber and relatively insensitive to fan drive power.Although still slight, the effects of Reynolds numberappear more distinct in the settling chamber.

4.10. Fluctuating Pressure Coefficient inHigh-Speed Diffuser

The pressure transducer in the high-speed diffuserwas installed on the LHS at midheight, about halfwaydownstream at station 68. The fluctuating pressure coef-ficient was formed with the dynamic pressure in the testsection, and the coefficients were plotted as a function oftest section Mach number. Data for the comparison testsin air and in nitrogen at ambient temperatures are shownin figure 62 forR= 6 × 106 ft−1. Just as with the liquidnitrogen injection station data, the disturbance levelscontinue to rise as sonic speeds in the test section areapproached and exceeded because the measuring stationis downstream of the test section; there is no tendency forthe choke effect at the test section to reduce the distur-bance levels as test section sonic speeds are approached.In fact, the tendency is for the disturbance levels to risemore steeply forM > 0.8 probably as a result of theformation of unsteady shocks between the end of the testsection and the beginning of the diffuser with theirattendant increase in noise levels.

4.11. Fluctuating Pressure Coefficient in Plenum

The pressure transducer in the plenum was installedon the RHS at station 0, near the plenum wall, and at thesame height as the test section top wall. The fluctuatingpressure coefficients calculated with the test sectiondynamic pressure are shown as a function of the testsection Mach number in figure 63. The test conditionsrepresented in figure 63 are the comparison tests for airand nitrogen atR= 6 × 106 ft−1 and the minimum andmaximum Reynolds number boundaries, respectively.The disturbance levels in the plenum for these test

22

conditions are very low and increase from about 0.001 to0.002 with an increase in Mach number from 0.2 to 1.05.There is very little difference between the air and thenitrogen mode test results atR = 6 × 106 ft−1 or betweenthe minimum and maximum Reynolds numberboundaries.

Power spectra for the minimum Reynolds numberboundary data forM = 0.2 to 0.642 are shown infigure 64 for a frequency bandwidth from 0 to 1 kHz. AtM = 0.2 (fig. 64(a)), line frequency interference peaks atmultiples of 60 Hz are quite evident. The somewhatelevated disturbance level shown for the rms fluctuatingpressure coefficient at this Mach number in figure 63(b)may partially result from the line interference present inthe measurement. Note that this is an exceptional condi-tion where an already quiet signal from the plenum ismeasured at the lowest operating condition of the tunnel.At most other conditions, the data signal is of sufficientmagnitude that the small line frequency interference isnot an appreciable part of the measurement. AtM = 0.3(fig. 64(b)), the line frequency interference is still evidentbut not nearly so intrusive as at the lower Mach number.At the higher Mach numbers, the line frequency interfer-ence is still identifiable in some places but is not a signif-icant part of the measurement.

All of the power spectra in figure 64 show that manydisturbance peaks affect the plenum. Few, if any, of thepeaks can be positively identified as to source except, ofcourse, the line frequency interference. Another possibleexception is the fan BPF, which is known fairly accu-rately. The BPF is noted on each of the power spectra. Acareful examination of the power spectra shows peaks ator very near this frequency. For instance, atM = 0.6 to0.642 (figs. 64(e)–64(g)), the BPF is 227 Hz. There is amodest peak in the power spectra at 221 Hz. The fre-quency resolution in the power spectra is approximately1 Hz so this peak is not close enough in frequency to beidentified as the BPF. However, the peak at 221 Hz isdistorted on the high-frequency side as though a secondpeak might be there as well. A higher resolution powerspectrum in figure 65 forM = 0.6, which corresponds tothe power spectrum of figure 64(e), shows a separatepeak of about 225 Hz, which probably corresponds to thefan blade passage frequency peak.

Mabey (1976) has indicated that the slots themselvescan be sources of regular disturbances with characteristicslot frequencies having Strouhal numbers in the range of0.03 to 0.04 based on slot width. If a slot Strouhalnumber of 0.035 is assumed, then the characteristic slotfrequenciesfs for the NTF slots are as indicated on thepower spectra in figure 64. Positive correlation of thesefrequencies with frequency peaks in the power spectra

was not possible and evidence of their existence in thepower spectra is not strong.

Multiple acoustic resonances can exist in the ple-num, but these disturbances have not been considered inthe present analysis. However, wind tunnel models sub-ject to unsteady load are, therefore, sources of excitationand can be affected by one particular kind of resonanceassociated with the ventilated test section walls and theplenum. For this kind of resonance, the model is thesource of the excitation either through forced oscillationor aeroelastic vibration response such as flutter or buffet.In this test section plenum resonance, transverse wavesfrom the test section are partially transmitted through thetest section plenum boundary, travel outward into theplenum, and are reflected from the outer plenum wallsback to the ventilated wall interface. If the returningwave is in phase with the outgoing wave, resonance willoccur.

This kind of resonance problem has been studied byMabey (1978), Barger (1981), and Mokry (1984) whodeveloped analytical solutions; more recently, Lee andBaik (1991) extended a finite element numerical solutionto include the slotted wall boundary conditions. For reso-nance frequency estimates in the NTF, the closed formresults of Barger were applied because that analysis usedaccurate boundary conditions for the slotted walls. Thecalculated fundamental frequencies for the test sectionplenum resonancefp are indicated in the power spectra infigures 64(a)–64(e) up toM = 0.6. ForM > 0.618, thesolution changes character, the reflection from the slottedboundary is complete, and the influence of the plenum isgreatly diminished.

The presence of the test section plenum resonancepeaks cannot be confirmed in the power spectra infigure 64. In the absence of discrete excitation as mightbe provided by an oscillating or vibrating test model, thisparticular form of resonance is not considered to be aproblem in the NTF. Apparently, flow unsteadiness byitself is not sufficient to excite this resonance to detect-able levels.

From the power spectra in figure 64, the excitationassociated with fan blade passage frequencies, character-istic slot frequencies, and slotted wall plenum resonancesdo not contribute significantly to the level of disturbancemeasured in the plenum.

4.12. Convection Velocities

In a shear layer, disturbance patterns are transportedwith the stream at some fraction of the free-stream veloc-ity, which depends on the scale of the disturbance and itslocation within the shear layer. The magnitude of thetransport velocity generally depends on the frequency of

23

the disturbance and on the separation distance overwhich it is measured.

In the NTF, the overall streamwise convectionvelocity uc(x) was measured in the test section. Threepressure transducers were spaced 2.25 in. apart stream-wise on the test section RHS sidewall at station 13. Crosscorrelations of unfiltered signals from an adjacent pair ofthese transducers were used to obtain the transit timebetween the two transducers for disturbances convectingdownstream. A sample cross-correlation plot is shownin figure 66 forM = 0.998 andR= 6.1× 106 ft−1. Thismeasurement was made in the nitrogen mode at ambienttemperatures. For this test point, an overall streamwiseconvection velocity ratiouc(x)/V = 0.746 was computed.Data for this and other Mach numbers from 0.2 to 1.05are shown in figure 67 for comparison with similar datameasured in air. The agreement between the two sets ofdata is sufficient to indicate that there are no significantdifferences in convection velocities in the two test gases.Although the convection velocity ratio is relatively con-stant across the Mach number range at≈0.76, there is aslight tendency for the level to rise with increasing Machnumber.

A similar comparison is shown in figure 68 for slotsopen compared with slots covered. These data weremeasured in the air mode along the minimum Reynoldsnumber boundary. Again, comparative data indicate nosignificant difference between slots open and slotscovered. The overall convection velocity ratio of 0.74 atminimum Reynolds numbers is slightly lower than forR= 6 × 106 ft−1 with a slight tendency for the level torise with increasing Mach number. Because Reynoldsnumber increases with increasing Mach number alongthe minimum Reynolds number boundary, the apparentincrease of convection velocity ratio with Mach numbermay really be a Reynolds number effect.

The effects of hot wall versus cold wall on the con-vection velocity are shown in figure 69 forM = 0.8.There is a tendency for the convection velocities to behigher for the hot wall condition than for the cold wallcondition; overall, the convection velocity ratio is≈0.77.The Reynolds numbers are not constant for these dataand range from≈7.5× 106 ft−1 for the cold wall data to≈10 × 106 ft−1 for the hot wall data. The tendency in theconvection velocity data has been for the level toincrease slightly as the Reynolds number is increased,which might partially account for the apparent effect ofthe hot wall in figure 69.

The effect of Reynolds number mentioned previ-ously is somewhat more apparent in the data in figure 70where the convection velocity ratios for the maximum

Reynolds number boundary are shown as a function ofMach number. The data for the minimum Reynolds num-ber boundary (air mode with slots open) are repeatedfrom figure 68 for comparison. The convection velocityratios for intermediate Reynolds numbers atpt = 43.2 psiare also shown. There is a positive increment of about0.05 in convection velocity ratio from the minimum tothe maximum Reynolds number boundary with most ofthe increment occurring between the low and intermedi-ate Reynolds numbers.

5. Conclusions

Dynamic measurements of the fluctuating staticpressure levels have been made at 11 locations in the cir-cuit of the NTF across the complete operating range,which resulted in flow disturbance measurements at thehighest Reynolds numbers available in a transonicground test facility. Tests were made with independentvariation of Mach number, Reynolds number, and fandrive power; for each test, two of the three parameterswere kept constant, which for the first time allowed adistinct separation of the effect of these important param-eters. Tests were also made with independent windtunnel speed variation by either fan speed or inlet guidevane angle variation. An analysis of these dynamicflow disturbance measurements has led to the followingconclusions:

1. The results of tests to isolate the effects of Machnumber, Reynolds number, and fan drive power onflow disturbance levels indicate that Mach numbereffects predominate. The flow disturbance levelsappear relatively insensitive to Reynolds number andfan drive power variations.

2. One of the primary sources of noise in the NTFappears to be flow surface gaps associated with thesidewall reentry flaps at the downstream end of thetest section. The gaps appear capable of producingedge tones at some flow conditions.

3. The downstream second minimum flow area formedon the top and bottom walls to choke the flow atsubsonic test section Mach numbers produces onlymarginal reduction in the flow disturbance levels inthe test section at a chokeM = 0.8.

4. The effects on flow disturbance level of intentionaldifferences in temperature between the wall bound-aries and the test gas are small with a tendency forthe cold wall to have a slightly higher disturbancelevel than the hot wall. Smaller incidental tempera-ture differences during the tests show an almost neg-ligible effect on the data.

24

5. Although in general agreement, a comparisonbetween results in air and in nitrogen show differ-ences that are thought to be caused at least in part bya mismatch in velocity between the two tests.

6. The liquid nitrogen injection process does not con-tribute significantly to the level of flow disturbancein the test section.

7. Flow disturbance levels in the settling chamberappear to be adversely affected by the gas-ventingprocess, which occurs during cryogenic operationswith nitrogen.

8. Tests with the slots covered show reductions in thetest section sidewall static pressure fluctuation levelsonly atM > 0.6.

9. Test section sidewall static pressure fluctuationlevels are relatively independent of the test sectiongeometry settings of wall divergence, reentry flapangle, and model support wall angle.

10. Wind tunnel speed changes can be obtained witheither fan speed changes or inlet guide vane anglechanges with no significant difference in test sectionflow disturbance levels.

11. Fan blade passage frequency peaks are not a signifi-cant contribution to the flow disturbance levels in thewind tunnel at most operating conditions. Thesefrequency peaks are only apparent in the powerspectra at low tunnel operating conditions (i.e., low-Mach number, low-stagnation pressure, and low fandrive power) when the background level of distur-bance is sufficiently low. The disturbances associ-ated with the characteristic slot frequencies and slot-ted wall-plenum resonances are also insignificant.

12. Overall streamwise convection velocity ratiosuc(x)/V measured with the use of unfiltered pressuretransducer signals on the test section sidewall are rel-atively unaffected by the change in test gas from airto nitrogen, by open or covered test section slots, orby differences in wall-to-gas temperature. The con-vection velocity ratios increase slightly with increas-ing Reynolds number.

13. A comparison with other transonic wind tunnelsshows that the NTF has low levels of test sectionfluctuating static pressure especially in the high-subsonic Mach number range from 0.7 to 0.9.

Some additional comments on the measured databeyond those specifically enumerated are included here.

From comments by Mabey (1991), the weak sensi-tivity of the measured flow disturbances to Reynolds

number variation in the NTF may indicate that the flowdisturbances being measured could be due primarily toaerodynamic noise sources that are controlled by turbu-lent eddy viscosity effects. Such effects are typicallydependent on velocity and eddy sizes and independent ofReynolds number. Mabey’s theory specifically considersthe noise that emanates from the extraction region at thedownstream end of the test section where flows, whichhave entered the plenum through the slots, reenter themainstream. The results in this report do not specificallyconfirm Mabey’s theory, but the comparison at high-sub-sonic Mach numbers between slots open and slots cov-ered (fig. 41) is consistent with it.

The lower flow disturbance levels in the test sectionat M < 0.6 with the slots open compared with slotscovered may be connected with the partial transmissionof waves at the slotted wall-plenum interface, whichallows test section disturbances to pass through into theplenum atM < 0.618. Waves that enter the plenum maybe subject to dissipation effects through multiple reflec-tions from the thermal insulation surface on the interiorof the pressure shell or may become trapped by plenumstructural elements. AtM > 0.618, the reflection ofwaves from the slotted boundary is complete, and theattenuating influence of the plenum could be greatlydiminished.

Although considered separately in the list of conclu-sions, the effect of fan drive power variation and thecomparison of fan speed versus inlet guide vane angle forchanging wind tunnel speed, when taken together, indi-cate that the flow disturbance levels in the NTF testsection are insensitive to either the fan blade tip speed orthe fan shaft power. The relationship, if any, between thisapparent insensitivity and the extensive noise attenuationtreatment of the fan nacelle nose cone and tail cone wasnot determined.

The apparent lack of influence of the liquid nitrogeninjection process on the level of flow disturbancedetected in the test section may be associated with thepresence of suspended droplets in the liquid nitrogenspray. Such a droplet suspension could be inhibiting theupstream propagation of broadband fan noise similar tothe attenuation of sound propagation in atmospheric fog,which would partially offset any direct noise created bythe injection process.

NASA Langley Research CenterHampton, VA 23681-0001October 23, 1995

25

Appendix A

Detailed Description of NTF

The National Transonic Facility (NTF) is a single-return pressurized transonic cryogenic wind tunnel with aslotted square test section and can operate at Mach num-bers up to 1.2, pressures up to 130 psi, and temperaturesdown to−320°F. Specific components of the wind tunnelare described in this appendix.

Thermal Insulation

Thermal insulation for the wind tunnel (shownshaded in fig. 6) is internal rather than external to thepressure shell. The internal insulation shields the pres-sure shell from large temperature changes as the windtunnel temperature is varied during the cryogenic modeof operation. Because the pressure shell with its largethermal inertia is not directly subjected to changes in gastemperature, liquid nitrogen consumption is reduced andthermal cycling of the pressure shell is avoided.

The insulation is a closed-cell, high-density, rigidfoam of modified polyurethane material varying in thick-ness from≈6.0 to 7.5 in. and is attached to the inside ofthe pressure shell as shown in figure A1. Its excellentfire-retardant property is an important feature for anymaterial used in a wind tunnel such as the NTF, whichcan be pressurized to 130 psi in air. As shown in figure 6,the insulation is completely isolated from the flow by aninternal aerodynamic liner in the high-speed section ofthe wind tunnel from the beginning of the wide-anglediffuser downstream to the end of the high-speeddiffuser. Except for the fan shroud region, the insulationin the remainder of the wind tunnel circuit is separatedfrom the flow stream by a relatively thin liner as shownin figure A1. For economy of fabrication, the liner platesshown in figure A2 are flat aluminum panels installed ina 24-sided polygon cross-sectional shape in the wind tun-nel essentially from turn 1 to turn 4.

Principal Components

The principal components of the NTF circuit areshown in figure A3. As previously mentioned, the cryo-genic mode of operation uses nitrogen as the test gaswith cooling accomplished by the injection of liquidnitrogen directly into the flow stream. The liquid nitro-gen injection nozzles are located upstream of the fannacelle. Adcock (1977) has shown that liquid nitrogeninjection upstream of the fan results in lower powerrequirements and lower liquid nitrogen flow rates whencompared with downstream injection. This location mayalso be more favorable for complete evaporation of theinjected liquid, attenuation of upstream moving fan noise

by liquid nitrogen spray droplets, and reduction in theinjection noise levels which reaches the test section.

The aerodynamic design of the NTF has beenstrongly influenced by the need for economy of opera-tion. The cryogenic concept permits the achievement ofhigh Reynolds numbers at relatively low energy con-sumption levels compared with other high Reynoldsnumber ground test facility concepts. However, even forthe cryogenic wind tunnel, the overall consumption ofenergy is high and must be carefully managed. The prin-cipal energy requirement of the NTF in the cryogenicmode of operation is the energy consumed to produce theliquid nitrogen used for cooling. To minimize the cost ofnitrogen required to pressurize the tunnel and to reducethe cost of the pressure shell, the internal volume of theNTF circuit was designed to be as small as practical.Within this limitation, the settling chamber was madewith as great a length as the economics of the pressureshell and the internal fill volume would permit.

The corners of the NTF circuit are mitered to form90° turns. The turning vanes in these corners have anarithmetically progressive spacing, which was intro-duced initially by Dimmock (1950) for gas turbineresearch and used in other wind tunnels such as the LaRC0.3-Meter Transonic Cryogenic Tunnel and the DefenceResearch Agency (DRA) (formerly Royal AerospaceEstablishment (RAE)) 5 Metre Pressurized Low SpeedWind Tunnel.

Wide-Angle Diffuser

The wide-angle diffuser shown in figure A4 islocated immediately upstream of the settling chamber.The use of a wide-angle diffuser at this location mini-mized the diameter of the return duct of the tunnel circuitbut still permitted a large settling chamber and a high-contraction ratio with its attendant benefits on windtunnel performance and flow quality. The wall curvaturein the wide-angle diffuser was designed in the mannerdescribed by Küchemann and Weber (1953) for a nearlyconstant static pressure along the walls in the streamwisedirection. This desired pressure gradient with its reducedtendency for boundary layer separation is obtained byproper curvature of the walls. The centrifugal forceacting on the flow as it follows the curved wall contour isbalanced by the stream pressure gradient as the flow isslowed by the increased area of the wide-angle diffuser.At the downstream end, the flow must be returnedtoward the axial direction. In the NTF, the turning isaccomplished by the finned tube heat exchanger, whichalso contributes to the downstream pressure loss requiredto prevent flow separation from the diffuser walls. Theheat exchanger cooling water tubes are elliptical in crosssection and are oriented vertically. The plate-like fins

26

attached to the tubes are oriented horizontally. The aero-dynamic and gravity loads on the heat exchanger aresupported by a truss with radial and annular elementslocated near the downstream end of the wide-anglediffuser. The annular elements have been shaped toconform to the flow field curvature in the streamwisedirection. The wide-angle diffuser has an exit-to-inletarea ratio of 2.08:1, a length-to-inlet diameter ratio of0.465:1, and an exit wall angle of about 61°.

Turbulence Damping Screens

The exit of the wide-angle diffuser is followed by asettling chamber approximately 19 ft long. There arefour turbulence damping screens in the settling chamberthat are spaced 2 ft apart with the last screenapproximately 5 ft from the beginning of the contraction.The four screens are identical; the square-mesh wirecloth is woven of 0.032-in-diameter wires at a spacingof 6 wires/in. with a resulting porosity of approximately0.65. Because of the large diameter (35.7 ft) of the set-tling chamber and the high pressure (130 psi) at whichthe NTF can operate, the limiting factor in the selectionof the screens was the stress in the screen wires underload. The 18-ft-wide rolls of screen cloth are joinedtogether at their edges by butt-welding the individualwires thereby producing aerodynamically clean seams.The screens are installed preslacked to allow them todeflect about 2 ft downstream under maximum load as ameans of reducing the screen wire stress.

Contraction Section

The contraction section has an area ratio 14.95 to 1and was designed to produce uniform flow at the throatunder choke conditions (i.e., to have an essentiallystraight sonic line). The prescribed area distribution forthe contraction was calculated by a streamline curvaturemethod developed by Barger (1973) for axially symmet-rical flow with the use of the exact equations for an invis-cid compressible flow.

The NTF contraction consists of the three subsec-tions shown in figure A5. The first subsection is axiallysymmetrical with the prescribed area distributionmatched exactly. The second subsection is a transition ofthe cross-sectional shape from round to flat-sided withradial corner fillets of progressively shorter radius andarc length. Here, the prescribed area distribution ismatched only approximately. The third subsection is acontinuation of the essentially square cross section withcorner fillets; the corner fillet shape changes fromcircular arc to flat approximately 9 ft upstream of the testsection.

The length of the contraction is approximately 48 ft.An upstream 39-ft-long section of the contraction is a

movable structure that can be detached from the rest ofthe contraction and moved upstream within the pressureshell to permit deployment of one of two isolation valvesthat seal the test section and plenum from the rest of thecircuit. The isolation valves permit access to the test sec-tion without depressurizing the entire tunnel circuit.

Test Section

A plan view of the NTF test section is shown in fig-ure A6. The design of the NTF test section closelyresembles that of the Langley 8-Foot Transonic PressureTunnel, especially in the flow reentry region at the down-stream end of the test section. The cross section is nomi-nally 8.2 by 8.2 ft square with flat corner fillets at 45°,which results in a test section throat cross-sectional areaof 66.77 ft2. There are six longitudinal slots in each ofthe horizontal (top and bottom) walls and provision fortwo slots in each of the vertical (side) walls.

The NTF slot shape for the top and bottom walls isshown in figure A7. In the test section length normallyoccupied by test models, the slots are of constant width(i.e., open area ratio of 6 percent). The upstream lengthof the slots is contoured to obtain a uniform Machnumber distribution at a Mach number of 1.2. Thecontour shape was designed with a modified method ofcharacteristics developed by Ramaswamy and Cornette(1980). The side wall slots are currently blanked offwhich provides a solid side wall. Because the currentconfiguration of the test section is only slotted on the topand bottom walls, the overall wind tunnel open area ratiois then 3 percent.

The length of the slotted region is approximatelythree test section widths. The side walls are parallel toeach other, but the top and bottom walls have flexures atthe upstream end which permit variation in wall anglefrom approximately 0.5° converged to 1.0° diverged. Asshown in the sketches of figures A8 and A9, the side-walls have provision for three large windows for flowfield observations; smaller ports are located in the top,bottom, and side walls for lighting and viewing.

Remotely adjustable reentry flaps equal to20-percent slot length are located at the downstream endsof the slots. The angle of these flaps can be varied to con-trol test section flow gradients and to minimize powerconsumption. The range of flap angle adjustment is from4° toward the flow to 15° away from the flow on the topand bottom walls and from 0° to 15° away from the flowon the side walls.

Model Support Section

The model support section, which is located immedi-ately downstream of the test section, is rectangular in

27

cross section with corner fillets that are tapered in thestreamwise direction. The side walls adjacent to themodel support strut are indented to relieve strut block-age. The top and bottom walls in this section are attachedto flexures at their downstream ends. The angle of incli-nation of these walls can be varied from 0° (walls paral-lel to tunnel centerline) to approximately 4.5° inward(leading edge toward flow). The vertical height offsetbetween the top and bottom test section walls and themodel support section walls is independently variablefrom an offset near zero to approximately 8 in. This ver-tical height offset can be varied as a function of windtunnel flow conditions to accommodate varying reentryflow requirements and, thus, to minimize wind tunnelpower consumption. Generally, test models are sting sup-ported from a circular arc strut as shown in figure A9.The elevation sketch in figure A10 shows the NTF testsection geometry variables including the geometry forthe downstream choke.

High-Speed Diffuser

The high-speed diffuser shown in figure A11, islocated immediately downstream of the model supportsection. It consists of two sections: a three-stage transi-tion section and a conical section. The three stages of thetransition section approximate the area distribution of acone with a half-angle of approximately 2.6°, the sameangle as the actual conical section. The transition cross-sectional shape progresses from a rectangular sectionwith flat corner fillets to a fully round section in threestages of nearly equal length. The flat corner fillets arefaired out within the first stage of the transition. Exceptfor these fillets, the shapes in the transition sectionconsist of flat panels joined at the corners of the crosssection by quarter-round conical sections. The diffuser,which includes the model support section, has an overallarea ratio of 2.92 to 1. As is the contraction, the high-speed diffuser is also a movable structure. It can bedetached from the model support section and moveddownstream within the pressure shell to permit deploy-ment of the downstream isolation valve. (See fig. A9.)

Fan Drive System

The fan is located 29 ft downstream of turn 2. Asshown in figure A12, the upstream fan nacelle fairing isbent through that corner. The single-stage fan has 25fixed-pitch blades fabricated of fiberglass-reinforcedplastic, and the fan load is changed by either the angle ofthe 24 variable inlet guide vanes (IGV) upstream ofthe fan or rotational speed. There are 26 fixed statorsdownstream.

Flat acoustic panels are located on the fan nacelleand the adjacent wind tunnel walls at the nose and tail

cones of the nacelle. These panels are shown infigures A13(a) and A13(b) for the nose and tail cone ofthe nacelle, respectively. These panels are intended toattenuate the fan noise propagating upstream and down-stream from the fan. As described by Lassiter (1981), thedesign uses a dual Helmholtz resonator concept and pro-vides approximately 13-dB reduction of fan noise at thetest section. The geometry of the two-layer perforatedsheet honeycomb lining that forms the dual Helmholtzresonator acoustic panels is shown in figure A14.

The fan is powered by two variable-speed, wound-rotor induction motors and a synchronous motor withmaximum power ratings of 66 000 hp (49.2 MW) and60 000 hp (44.8 MW), respectively. As mentioned previ-ously, the induction motors were salvaged from the4-Foot Supersonic Pressure Tunnel that was originallylocated on the NTF site. As shown in the upper part offigure A15, the introduction motors are coupled to thefan driveshaft through a two-speed gear box with gearratios (motor-to-fan speed) of 835:360 in low gear and835:600 in high gear. The purpose of the two-speed gearbox is to provide a better match of the available motortorque with the required fan torque at different operatingtemperatures. The synchronous motor is in line with thefan driveshaft and rotates at fan speed at all times.

The maximum shaft power available from the drivemotor combination as a function of fan rotational speedis shown in the lower part of figure A15 for both thehigh- and low-gear ratios. The synchronous motor isoperated at the fan shaft speed corresponding to the max-imum speed of the induction motors in the low-gearratio, and is brought up to synchronous speed by theinduction motors. The rotational speed of the inductionmotors is controllable within±0.25 percent over theentire range with a modified Kraemer drive controlsystem.

Under high-power conditions, when both the induc-tion and synchronous motors are required, the fan isrotated at the synchronous motor speed of 360 rpm(6 Hz), and wind tunnel speed control is accomplished byvariation of the inlet guide vane angles. At low-powerconditions when only the induction motor power isrequired, tunnel wind speed can be varied either by inletguide vane angle variation or motor rotational speedvariation.

Because cooling in the air mode of operation isaccomplished with the water-cooled heat exchanger, themaximum usable power is limited by the design capacityof the heat exchanger cooling towers; this limit isapproximately 55 000 hp (41 MW).

28

Exhaust System

When the NTF is operating in the cryogenic mode,liquid nitrogen is continuously introduced into the circuitto maintain temperature. To maintain pressure andconstant mass flow, an equal amount of gaseous nitrogenmust be removed from the circuit. As seen in figure A3,the exhaust ports are in the crossleg between turns 3

and 4; a sketch of the exhaust muffler and vent stack isshown in figure A16. The maximum flow rate of liquidnitrogen into the circuit is on the order of 9000 gpm orapproximately 1000 lb/sec. The muffler and vent stackare sized to exhaust an equal mass flow rate of gaseousnitrogen. The exhaust system is also used to vent thetunnel pressure in either the air mode or nitrogen mode ofoperation.

29

Figure A1. Thermal insulation and liner in NTF.

Aluminum linerEpoxy-glass pultrusion

rib and cap strip

Tab and link system

Three layers:insulation,adhesive,

and fiberglass

Shell

Pultrusion

Links

Liner Flow

Shell

30

Figure A2. Return leg of the NTF circuit with aluminum liner plate flow surface.

31

Figure A3. Principal components of NTF circuit.

Turning vanesScreen

support section Downstreamnacelle

Downstreamshroud

Upstream and fancontainment

shroud Upstreamnacelle

LN2injectors

High-speeddiffuser

Model supportHeat

exchanger

Contractionsection

Test sectionwindows

Truss support forheat exchanger

GN2exhaust

ports Acousticpanels

Wide-angle diffuser

Fixed contraction,test, and

model support section

32

Figure A4. NTF wide-angle diffuser.

Figure A5. NTF contraction section.

Truss support forheat exchanger

33

Figure A6. Plan view of NTF test section. All dimensions are in feet.

25

20

13

Flow

Test section,8.2 by 8.2

Reentry flaps

Plenum, 27 dia.

34

Figure A7. NTF slot shape for top and bottom walls.

Figure A8. NTF test section.

.2 .4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

x/h

0

.02

.04

.06

.08

.10

Slot-openarea ratio

35

Figure A9. Elevation of NTF test section.

Plenum

High-speeddiffuser

Contraction(retracted)

Isolation valve(closed)

Isolation valve(stored)

Test section

36

Figure A10. Elevation of NTF test section geometric variables (not to scale).

Tunnelcenterline

Model centerof pitch rotation

(station 13)

Choke off

Choke on

Choke station

–5 0 5 10 15 20 25 30 35

Tunnel station, ft

Bottom test section wall(slots open)Flex plateContraction

Plenum

Reentryflap Flap

hinge lineModel support

section wall Flex plate

Flow

37

Figure A11. NTF high-speed diffuser.

38

Figure A12. NTF fan section assembly.

Outlet stators

Fan

Inlet guidevanes

Shaft mainradial bearing

Acousticliners

Guide vaneactuator ring

Flow

Acoustic liners

39

(a) Nacelle nose cone.

(b) Nacelle tail cone.

Figure A13. NTF nacelle fairings with artist’s conception of flat acoustic liner panels.

40

Figure A14. NTF fan nacelle acoustic panel dual Helmholtz resonator concept.

Nacellecone

Nose

Tail

Open area — Cavity depths —

5.5

5.5

4.5

4.5

1.3

2.5

4.6

7.5

OA1, percent OA2, percent d1, in. d2, in.

d1

d2

OA1

OA2

41

Figure A15. NTF fan drive system power 10-min. rating.

100

80

60

40

20

100 200 300 400 500 600

Fan Two-speedgear box

Existing variable-speedinduction motors,

49.2 MW

New synchronousmotor, 44.8 MW

Synchronousmotor

Inductionmotors

Low-gear ratio835:360 High-gear ratio

835:600

0Fan speed, rpm

Motor shaftpower, MW

42

Figure A16. NTF exhaust muffler and vent stack.

Mixing plane

Secondary air

Fan air

Acousticenclosure

Discharge vessel

MufflerStack

43

Appendix B

Longitudinal Static Pressure and MachNumber Gradients

The longitudinal buoyancy force induced on a bodyby a static pressure gradient is

(B1)

which, for a linear change in static pressure with longitu-dinal distance, simplifies to

(B2)

where the volume is

(B3)

A somewhat idealized cross-sectional area distribu-tion of a representative transport aircraft model is shownin figure B1. A polynomial fit to the area distribution canbe expressed as

(B4)

For simplicity, the polynomial in equation (B4) is limitedto its second order as shown in figure B1

(B5)

Substitution of equation (B5) into equation (B3) resultsin the volume for the second-order polynomial fit

(B6)

After substitution of this result into equation (B2), thelongitudinal force is obtained from

(B7)

If a coefficient form is introduced into equation (B7),then

(B8)

and the various terms are nondimensionalized as

(B9)

The following geometrical ratios, which relate themodel size to the test section dimensions, can then berecognized:

1.

2.

3.

Then

(B10)

To avoid significant test section wall interferences,the recommended maximum limits for these ratios aregiven as follows:

No firm criteria exist for what constitutes an accept-able pressure gradient. In reality, the pressure gradient ispractically never exactly zero; even when it is quitesmall, higher order effects because of nonlinearity mayassume some importance. One possible approach is toexamine the magnitude of a gradient, which would causean increment in drag coefficient of±0.0001 (i.e.,±1 dragcount in the terminology of wind tunnel experimental-ists). The nondimensionalized pressure gradient isobtained from equation (B10) as

(B11)

and substitution of the previous values ofand results in

(B12)

If perfect gas relations are assumed,q/pt is solely a func-tion of Mach number, and the nondimensional staticpressure gradient can be plotted as shown in figure 15where the allowable gradients under present assumptionsare those that fall below the line. Perfect gas relations areadequate in this instance because the results are to beapplied to air mode tests with the test section slots cov-ered, the stagnation pressure near atmospheric, and thestagnation temperature near ambient.

Test section gradients are often expressed in terms ofMach number instead of static pressure. With theassumption of perfect gas relations and the ratio of

∆Fxpd

dx------ Axdx

0

l

∫=

∆Fxpd

dx------ V=

V Axdx0

l

∫=

Ax

Amax------------ An

xl--

n

n 0=

J

∑=

Ax

Amax------------ 4

xl-- 1 x

l--–

=

v23--- lAmax=

∆Fx23--- dp

dx------ lAmax=

∆Cx

∆Fx

qS----------

23--- dp

dx------ l

qS------ Amax= =

∆Cx23--- 1

q/pt----------

dp/pt

dx/h------------- l /h

S/AT------------

Amax

AT------------=

k1

Amax

AT------------ model blockage ratio,=

k2S

AT------- wing area to test section area ratio,=

k3lh--- model length to test section height ratio,=

∆Cx23--- 1

q/pt----------

dp/pt

dx/h-------------

k1k3

k2-----------=

k1 0.005 (Baals and Stokes (1971))≤

k2 0.05 (Monti (1971))≤

k3 0.6 (Baals and Stokes (1971))≤

dp/pt

dx/h-------------

32--- ∆Cx

qpt-----

k2

k1k3-----------=

∆Cx k1 k2,, ,k3

dp/pt

dx/h------------- 0.0025

qpt-----=

44

specific heats of 1.4, equation (B12) can be rewritten interms of Mach number as

(B13)

and is shown plotted in figure 16. The question of accu-racy of the measurement of the local static pressures thatmake up the pressure gradient and the effect of this accu-racy on the computed magnitude of that gradient isaddressed in appendix C.

dMdx/h----------- 0.00125M 1 0.2M

2+( )=

Figure B1. Cross-sectional area distribution for typical transport model.

1.0

.5

x/l1.0.50

2d-orderpolynomial

Typicaltransport

Ax/Amax

45

Appendix C

Accuracy of Test Section Pressure Gradientsby Least-Squares Method

The accuracy of individual pressure measurementsrequired for least-squares slopes of a given accuracy willbe evaluated here. First, the usual equations for a least-squares straight-line fit will be shown. In figure C1, thedeviation between theith data point and the least-squaresstraight line is

(C1)

The sum of the squares of these deviations is to beminimized by

(C2)

where the values of the slopea and the interceptb are tobe chosen to obtain the minimization

(C3)

(C4)

The summation is now understood to be oni from 1 toj.

(C5)

(C6)

Equations (C5) and (C6) are solved simultaneously foraandb

(C7)

(C8)

Equations (C7) and (C8) are the standard equations todetermine the slope and intercept for a least-squaresstraight-line fit to the data pointsPi. A simplification isconveniently introduced here such that

which would be true, for instance, if the distribution ofthe pointsxi were symmetrical aboutx = 0. In the presentcase, the simplification is justified because the origin for

the xi coordinates is at the midpoint of the distributionand the value of the intercept is to be used as the Machnumber at the midpoint. With this simplification equa-tions (C7) and (C8) become

(C9)

(C10)

In equation (C9), note that the influence of a particulardata pointPi depends on its positionxi in the distributionand on the total number of points.

Now, errorsεi are introduced in the data pointsPisuch that

(C11)

then

(C12)

where the first term on the right may be associated withthe true value ofa and the second term with the error. Aroot-sum-square (rss) version of this latter term is

(C13)

For the National Transonic Facility (NTF) steady-statecalibration, the error in pressure measurement was

then

(C14)

Also, for the distribution of static pressure orifices in theNTF centerline calibration tube, 25 orifices were spaced3 in. apart in a length of 6 ft centered on test section sta-tion 13, so

and

(C15)

δi Pi axi b+( )–=

δ2 δi2

i 1=

j

∑ Pi axi b+( )–[ ]2

i 1=

j

∑ min.= = =

∂δ2

∂a-------- 2Σxi Pi axi b+( )–[ ]– 0= =

∂δ2

∂b-------- 2Σ Pi axi b+( )–[ ]– 0= =

ΣxiPi aΣxi2

bΣxi–– 0=

ΣPi aΣxi jb–– 0=

ajΣxiPi ΣxiΣPi–

jΣxi2 Σxi( )2

–----------------------------------------=

bΣxiΣxiPi Σxi

2ΣPi–

Σxi( )2jΣxi

2–

-----------------------------------------------=

Σxi 0=

aΣxiPi

Σxi2

-------------=

b1j---ΣPi=

Pi pi εi+=

a ∆a+Σxi pi εi+( )

Σxi2

-----------------------------Σxi pi

Σxi2

-------------Σxiεi

Σxi2

------------+= =

∆arss

Σ xiεi( )21/2

Σxi2

--------------------------------=

εi 0.003pmax±=

∆arss

0.003pmax

Σxi2( )

1/2-------------------------=

1h--- Σxi

2

1/2

1.099=

dp/pt

dx/h-------------

error

0.0031.099-------------

pmax

pt------------=

46

The electronically scanned pressure (ESP) unit used forthe steady-state calibration had a maximum pressurerange of±2.5 psi; if a worst case (lowest) value of totalpressure is taken as 15 psi, then the static pressure gradi-ent error term becomes

(C16)

The gradient error can be expressed in terms of Machnumber again with the assumption of perfect gas rela-tions and the ratio of specific heats of 1.4 to get

(C17)

During the dynamic investigation with the slotscovered, the longitudinal static pressure gradient wasdetermined by a distribution of 13 test section wall staticpressure orifices in a length of 7.5 ft centered on test sec-tion station 13. In this case

and because the same ESP unit for pressure measurementwas used, the pressure gradient error term equivalent toequation (C16) becomes

(C18)

The Mach number gradient error term then increases to

(C19)

This indicates that the accuracy of the determinationof test section longitudinal gradients for the configura-tion with the slots covered was slightly degradedcompared with that for the steady-state calibration. Equa-tions (C16) and (C18) and equations (C17) and (C19) areincluded in figures 15 and 16, respectively, for reference.

dp/pt

dx/h-------------

error

0.000455=

dMdx/h-----------

error

0.0003251 0.2M

2+

9/2

M-------------------------------------=

1h--- Σxi

2

1/2

0.892=

dp/pt

dx/h-------------

error

0.00056=

dMdx/h-----------

error

0.000401 0.2M

2+

9/2

M-------------------------------------=

Figure C1. Least-squares straight-line fit to pointsP1, ...,Pn.

Least-squaresstraight line

P

xi

piδi

X

47

Appendix D

Preliminary Test Results With Steady-StateCalibration Probe Installed in Test Section

As mentioned in section 3.3 on nitrogen mode testsprior to the start of the present dynamic investigation andwhile the steady-state calibration investigation was beingplanned, simultaneous testing for both the steady-stateand the dynamic characteristics of the tunnel was consid-ered possible as an efficient use of the tunnel test timeand of liquid nitrogen resources if the requirements forboth sets of measurements could be satisfied concur-rently. For steady-state calibration purposes, a long slen-der survey probe containing several hundred staticpressure orifices was installed on the tunnel centerline;the probe stretched the entire length of the test sectionand extended well upstream into the contraction where itwas secured with support cables as shown in figure D1.At the downstream end, the survey probe was mounted inthe arc-sector centerbody, and the junction was aerody-namically faired with the same fiberglass-reinforcedplastic conical fairing as was used for this investigation.

Fluctuating static pressure data were taken at someof the test conditions covered in the steady-state calibra-tion. The test points for the power spectra shown infigure D2 are for test conditions near the maximumReynolds number boundary in the nitrogen mode(105 psi <pt < 125 psi andTt = −250°F). The data are forthe test section RHS sidewall at station 13. A large peakexists in the unfiltered power spectra and ranges in fre-quency from≈1.5 kHz atM = 0.2 to≈5.2 kHz atM = 1.0.

To help identify the source of this peak, observe that thepeak frequency changes with Mach number and thereforewith velocity, which indicates that the source is likelyaerodynamic. Further, observe that the source is notchoked off atM = 1.0, which indicates that the source isnot downstream of the test section. With these readilyobvious clues, the probe upstream support cables wouldbe a reasonable suspect. To substantiate this suspicion,the vortex-shedding frequency of the cables was esti-mated by assuming a Strouhal number of 0.2 and isshown as a function of test section Mach number in fig-ure D3 for test conditions corresponding to the ambienttemperature, low-pressure air mode tests. The frequen-cies of the peaks in the power spectra at these same testconditions are also shown for comparison, which indi-cate a reasonable probability that support cable vortexshedding is the source.

To determine the magnitude of the added distur-bance, a notch or band-reject filter consisting of a combi-nation of low-pass and high-pass filters was used to filterthe suspect peak. Filtered results were obtained for all ofthe power spectra and are shown by the dashed lines infigure D2. The integrated rms fluctuating pressure coeffi-cients for these power spectra are shown in figure D4.The large reduction due to filtering (as much as 25 per-cent at M = 0.5 and 0.6) indicates the importance ofobtaining dynamic data with all extraneous interferencesremoved. These preliminary results led to the decision toobtain dynamic data under dedicated conditions with thetest section empty. The actual empty test section distur-bance levels were even lower than those indicated by thefiltered data.

48

L-90-3430

Figure D1. Centerline calibration probe with support cables in NTF test section.

RHS sidewallreentry flaps

Slots

Support cables

RHS sidewallreentry flaps

Slots

Support cables

49

Figure D2. Power spectra obtained with centerline calibration probe in test section; test section RHS sidewall at station 13; 105 psi <pt < 125 psi;Tt = −250°F.

UnfilteredFiltered

Arbitraryscale

0 20Frequency, kHz

0 20Frequency, kHz

0 20Frequency, kHz

0 20Frequency, kHz

Arbitraryscale

M = 0.2 M = 0.5

M = 0.6

M = 0.7

M = 0.8

M = 0.9

M = 1.0M = 0.3

50

Figure D3. Calculated and measured vortex-shedding frequencies of centerline probe cables.pt = 17 psi;Tt = 120°F; airmode.

Figure D4. Effect of filtering centerline calibration probe support cable vortex-shedding frequency on fluctuating pres-sure coefficient.

Calculated, St = 0.2

Measured

9

8

7

6

5

4

3

2.2 .3 .4 .5 .6 .7 .8 .9 1.0 1.1 1.2

M

f, kHz

.1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.1 1.2M

Test section RHS sidewall station 13

.005

.006

.007

.008

.009

.010

.011

.012

pq

Unfiltered

Filtered~

51

Appendix E

Estimation of Edge Tone Frequency inFree-Shear Layer

The subject of edge tones has been extensively stud-ied as indicated, for example, by the survey paper on jet-wedge edge tones by Karamcheti et al. (1969) and by theintroductory paper on free-shear layer edge tones byHussain and Zaman (1978). The purpose here is not toanalyze the edge tone generation characteristics of free-shear layers but rather to investigate the plausibility thatedge tones occur in the National Transonic Facility(NTF) test section. Attention is concentrated on aerody-namic gaps that exist at the downstream end of the testsection wall surface. These gaps are on the test sectionsidewall immediately upstream of the leading edges ofthe sidewall reentry flaps and are visible in figure 43.

A sketch of the gap profile is shown in figure E1.The upstream lip is considered to act as the point of sepa-ration of the wall boundary layer to form a free-shearlayer between the lip and the downstream wedge, whichis formed by the leading edge of the sidewall reentryflap. As shown in the sketch, a vortex is considered to beperiodically shed from the lip, to travel downstream inthe free-shear layer, and to impact on the wedge, whichcreates an edge tone; an acoustic feedback signal isassumed to travel through the quiescent region on theplenum side of the free-shear layer and sustain the tone.

This tone phenomenon appears similar to what isencountered in open cavities. An expression for the fre-quency of periodic disturbances found with open cavitieswas given by Rossiter (1964) as

(E1)

wherem is the stage number for the periodic disturbance,andγ is a factor for the lag between the interaction of thevortex with the downstream edge and the emittanceof the associated acoustic feedback disturbance. Equa-tion (E1) can be rewritten as

(E2)

The denominator in equation (E2) can be recognized asthe sum of the time required for the shed vortex to travelthe length of the gapL downstream at a convection

velocity uc in the free-shear layer and the time for anacoustic disturbance emitted by the interaction of thevortex with the downstream edge to travel back to the lipat an acoustic velocityc through the quiescent region.

The acoustic velocity in equation (E2) has beenassumed to be equal to the free-stream speed of sound.Heller, Holmes, and Covert (1971) have shown thatbetter agreement between predicted and measuredfrequencies for cavities is obtained if the stagnationspeed of sound is used. This modified form of theRossiter (1964) equation is then written as

(E3)

or in reduced frequency form as

(E4)

The value for the convection velocity ratiouc/V used byRossiter was 0.57. The lag factorγ has been shown byRossiter to be a function of the length-to-depth ratio ofthe cavity and diminishes nearly linearly from a value of0.54 at a length-to-depth ratio of 8:1 to a value of 0.25at a length-to-depth ratio of 4:1. The gap profile of fig-ure E1 is considered to represent an open cavity with alength-to-depth ratio approaching zero because the ple-num wall is on the order of 9.0 ft from the test sectionsidewall.

Reduced frequencies calculated from equation (E4)for the test conditions of figures 27 and 35 are shown intable E1 and figure E2. For these calculations, the con-vection velocity ratiouc/V was assumed as 0.6 (which isappropriate for a free-shear layer) the lag factorγ wastaken as 0, and the velocityV was taken as the free-stream velocity in the test section.

The measured frequency results (lower at the lowReynolds numbers and higher at the high Reynolds num-bers) are listed in table E1 and shown in figure E2 tobracket the calculated values. The agreement is consid-ered good enough to support the possibility that the mea-sured frequency peaks are caused by free-shear layeredge tones, which are generated at the gap upstream ofthe sidewall reentry flaps under certain flow conditionsand are detectable in the test section.

fVL---- m γ–

V/uc M+-----------------------=

fm γ–

L/uc( ) L/c( )+-----------------------------------=

fm γ–

L/uc( ) L/ct( )+-------------------------------------=

fLV------ m γ–

V/uc( ) V/ct( )+--------------------------------------=

52

aQuantities in parentheses are possible second-stage frequencies.

Table E1. Free-Shear Layer Edge Tone Frequency Data

fL/V—

Gas type M R, ft−1 V, fps ct, fps Tt, °F f, measured, Hz measured calculated

Air 0.695 6.0× 106 764 1151 91.4 840 0.36 0.43

Nitrogen .695 6.0 777 1172 92.2 855 .36 .43

Nitrogen .598 39.9 481 834 −174.0 890 .60 .45

Nitrogen .694 39.9 519 784 −206.2 860 .54 .43

Nitrogen .742 39.8 538 764 −218.6 900 .54 .42

Nitrogen .793 39.8 558 747 −228.7 960 .56 .41

Nitrogen .839 39.8 576 734 −236.2 (1410)a (.80)a (.82)a

Nitrogen .892 39.9 601 725 −241.4 960 .52 .40

Nitrogen .992 40.1 644 710 −250.2 960 .48 .39

53

Figure E1. Gap profile between NTF test section sidewall and reentry flap. All dimensions are in inches.

Flow

Test section sidewall

L = 3.9

Sidewall reentry flap

6.5

Lip 1/4 rad.

Plenum

1.6

54

Figure E2. Comparison of measurement-reduced frequencies with free-shear layer edge tone-reduced frequenciescalculated from the modified Rossiter equation.

1.0

.8

.6

.4

.2

0.6 .7 .8 .9 1.0

M

fLV

Modified Rossiter eq. (E4)

R = 40 × 106 ft–1

R = 6 × 106 ft–1

2d stage

1st stage

55

References

Adcock, Jerry B. 1977:Effect of LN2 Injection Station Location onthe Drive Fan Power and LN2 Requirements of a CryogenicWind Tunnel. NASA TM X-74036.

Adcock, Jerry B. 1976:Real-Gas Effects Associated With One-Dimensional Transonic Flow of Cryogenic Nitrogen. NASATN D-8274.

Adcock, Jerry B.; and Johnson, Charles B. 1980:A TheoreticalAnalysis of Simulated Transonic Boundary Layers inCryogenic-Nitrogen Wind Tunnels. NASA TP-1631.

Baals, Donald D.; and Stokes, George M. 1971: A Facility Con-cept for High Reynolds Number Testing at Transonic Speeds.Facilities and Techniques for Aerodynamic Testing at Tran-sonic Speeds and High Reynolds Number, AGARD CP No. 83,pp. 28-1–28-12.

Barger, Raymond L. 1973:Streamline Curvature Design Proce-dure for Subsonic and Transonic Ducts. NASA TN D-7368.

Barger, Raymond L. 1981:Theory for Predicting BoundaryImpedance and Resonance Frequencies of Slotted-Wall WindTunnels, Including Plenum Effects. NASA TP-1880.

Bendat, Julius S.; and Piersol, Allan G. 1980:EngineeringApplications of Correlation and Spectral Analysis. John Wiley& Sons, Inc.

Bobbitt, Percy J. 1981: Report of the Panel on Fluid Dynamics.High Reynolds Number Research—1980, L. Wayne McKinneyand Donald D. Baals, eds., NASA CP-2183, pp. 169–195.

Boyles, George B., Jr. 1986: Description and Operational Status ofthe National Transonic Facility Computer Complex. AIAA-86-0383.

Braslow, Albert L.; Hicks, Raymond M.; and Harris, Roy V., Jr.1966: Use of Grit-Type Boundary-Layer-Transition Trips onWind-Tunnel Models. NASA TN D-3579.

Braslow, Albert L.; and Knox, Eugene C. 1958:Simplified Methodfor Determination of Critical Height of Distributed RoughnessParticles for Boundary-Layer Transition at Mach NumbersFrom 0 to 5. NACA TN-4363.

Bruce, Walter E., Jr. 1985: The U.S. National Transonic Facility—II. Special Course on Cryogenic Technology for Wind TunnelTesting, AGARD-R-722, pp. 15-1–15-10.

Chiu, Wen-Shyang; and Lauchle, Gerald C. 1989: Subsonic AxialFlow Fan Noise and Inflow Velocity Disturbance.Inter-Noise 89—Engineering for Environmental Noise Control,Proceedings of the 1989 International Conference on NoiseControl Engineering, George C. Maling, Jr., ed., Volume 1,pp. 133–138.

Coe, Charles F. 1969: Surface-Pressure Fluctuations AssociatedWith Aerodynamic Noise.Basic Aerodynamic Noise Research,Ira R. Schwartz, ed., NASA SP-207, pp. 409–424.

Corcos, G. M. 1963: Resolution of Pressure in Turbulence.J. Acoust. Soc. America, vol. 35, no. 2, pp. 192–199.

Corcos, G. M. 1967: The Resolution of Turbulent Pressures at theWall of a Boundary Layer.J. Sound & Vib., vol. 6, no. I,pp. 59–70.

Dimmock, N. A. 1950: The Development of a Simply ConstructedCascade Corner for Circular Cross Section Ducts. National GasTurbine Establ. Memo. No. 78, British Ministry Supply.

Dolling, D. S.; and Dussauge, J. P. 1989: Fluctuating Wall-Pressure Measurements.A Survey of Measurements and Mea-suring Techniques in Rapidly Distorted Compressible Turbu-lent Boundary Layers, H. H. Fernholz, P. J. Finley, J. P.Dussauge, A. J. Smits, and Eli Reshotko, eds., AGARD-AG-315. (Available from DTIC as AD A211 107.)

Dougherty, N. Sam, Jr. 1980:Influence of Wind Tunnel Noise onthe Location of Boundary-Layer Transition on a Slender Coneat Mach Numbers From 0.2 to 5.5, Volume I—ExperimentalMethods and Summary of Results. AEDC-TR-78-44-VOL-1,U.S. Air Force. (Available from DTIC as AD A083165.)

Dougherty, N. S., Jr.; and Fisher, D. F. 1980: Boundary-LayerTransition on a 10-Deg Cone: Wind Tunnel/Flight Correlation.AIAA-80-0154.

Dryden, Hugh L.; and Abbott, Ira H. 1948:The Design of Low-Turbulence Wind Tunnels. NACA Rep. 940.

Eckelmann, Helmut 1990: A Review of Knowledge on PressureFluctuations. Near–Wall Turbulence—1988 Zoran Zari’cMemorial Conference, S. J. Kline and N. H. Afgan, eds., Hemi-sphere Publishing Corp., pp. 328–347.

Elsenaar, A. 1990: The Windtunnel as a Tool for Laminar FlowResearch.ICAS Proceedings—1990, pp. 174–185. (Availableas ICAS-90-6.1.1.)

Elsenaar, A.; Binion, T. W., Jr.; and Stanewsky, E. 1988:ReynoldsNumber Effects in Transonic Flow. AGARD-AG-303.

Foster, Jean M.; and Adcock, Jerry B. 1987:User’s Guide for theNational Transonic Facility Data System. NASA TM-100511.

Fuller, Dennis E. 1981:Guide for Users of the National TransonicFacility. NASA TM-83124.

Goodyer, Michael J.; and Kilgore, Robert A. 1972: The HighReynolds Number Cryogenic Wind Tunnel. AIAA PaperNo. 72-995.

Hall, Robert M.; and Adcock, Jerry B. 1981:Simulation of Ideal-Gas Flow by Nitrogen and Other Selected Gases at CryogenicTemperatures. NASA TP-1901.

Hanly, Richard D. 1975: Effects of Transducer Flushness on Fluc-tuating Surface Pressure Measurements. AIAA-75-534.

Heller, H. H.; Holmes, D. G.; and Covert, E. E. 1971: Flow-Induced Pressure Oscillations in Shallow Cavities.J. Sound &Vib., vol. 18, no. 4, pp. 545–553.

Horstman, C. C.; and Rose, W. C. 1977: Hot-Wire Anemometry inTransonic Flow.AIAA J., vol. 15, no. 3, pp. 395–401.

Howell, Robert R.; and McKinney, Linwood W. 1977: The U.S.2.5-Meter Cryogenic High Reynolds Number Tunnel.High

56

Reynolds Number Research, Donald D. Baals, ed., NASACP-2009, pp. 27–51.

Hussain, A. K. M. F.; and Zaman, K. B. M. Q. 1978: The FreeShear Layer Tone Phenomenon and Probe Interference.J. Fluid Mech., vol. 87, pt. 2, pp. 349–383.

Igoe, William B. 1980: Characteristics and Status of the U.S.National Transonic Facility. Cryogenic Wind Tunnels,AGARD-LS-111, pp. 17-1–17-11.

Jones, Gregory Stephen 1991: The Measurement of Wind TunnelFlow Quality at Transonic Speeds. Ph.D. Diss., Virginia Poly-tech. Inst. & State Univ.

Karamcheti, K.; Bauer, A. B.; Shields, W. L.; Stegen, G. R.; andWoolley, J. P. 1969: Some Features of an Edge-Tone FlowField. Basic Aerodynamic Noise Research, Ira R. Schwartz,ed., NASA SP-207, pp. 275–304.

Kern, F. A.; Knight, C. W.; and Zasimowich, R. F. 1986: NationalTransonic Facility Mach Number System.ISA Trans., vol. 25,no. 2.

Kilgore, Robert Ashworth 1974: The Cryogenic Wind Tunnel forHigh Reynolds Number Testing. Ph.D. Thesis, Univ. ofSouthampton. (Available as NASA TM X-70207.)

Kilgore, Robert A.; Adcock, Jerry B.; and Ray, Edward J. 1974:Flight Simulation Characteristics of the Langley High Rey-nolds Number Cryogenic Transonic Tunnel.J. Aircr., vol. 11,no. 10, pp. 593–600.

Kovásznay, Leslie S. G. 1950: The Hot-Wire Anemometer inSupersonic Flow.J. Aeronaut. Sci., vol. 17, no. 9, pp. 565–572,584.

Kovásznay, Leslie S. G. 1953: Turbulence in Supersonic Flow.J. Aeronaut. Sci., vol. 20, no. 10, pp. 657–674, 682.

Kuchemann, Dietrich; and Weber, Johanna 1953:Aerodynamics ofPropulsion, First ed. McGraw-Hill Book Co., Inc.

Lassiter, William S. 1981: Design Predictions for Noise Control inthe Cryogenic National Transonic Facility.Noise Control Eng.,pp. 76–84.

Lee, In; and Baik, Ki-Young 1991: Resonance Prediction forSlotted Circular Wind Tunnel Using Finite Element.AIAA J.,vol. 29, no. 12, pp. 2266–2269.

Lowson, M. V. 1968:Prediction of Boundary Layer PressureFluctuations. AFFDL-TR-67-167, U.S. Air Force. (Availablefrom DTIC as AD 832 715.)

Mabey, D. G. 1971:Flow Unsteadiness and Model Vibrationin Wind Tunnels at Subsonic and Transonic Speeds.CP No. 1155, British ARC.

Mabey, D. G. 1978:The Resonance Frequencies of VentilatedWind Tunnels. R. & M. No. 3841, British ARC.

Mabey, D. G. 1991:A Semi-Empirical Theory of the Noise in Slot-ted Tunnels Caused by Diffuser Suction. Tech. Rep. 91023,British RAE.

Mabey, D. G. 1976:Some Remarks on the Design of TransonicTunnels With Low Levels of Flow Unsteadiness. NASACR-2722.

Margoulis, W. 1921:A New Method of Testing Models in WindTunnels. NACA TN-52.

Meyers, James F.; and Clemmons, James I., Jr. 1987: FrequencyDomain Laser Velocimeter Signal Processor—A New SignalProcessing Scheme. NASA TP-2735.

Meyers, James F.; and Wilkinson, Stephen P. 1982: A Comparisonof Turbulence Intensity Measurements Using a Laser Veloci-meter and a Hot Wire in a Low Speed Jet Flow. Presented atInternational Symposium on Applications of Laser-DopplerAnemometry to Fluid Mechanics.

Michel, Roger 1988: Boundary Layer Development andTransition.Boundary Layer Simulation and Control in WindTunnels, AGARD-AR-224, pp. 217–249.

Michel, U.; and Froebel, E. 1988: Lower Limit for theVelocity Fluctuation Level in Wind Tunnels.Exp. Fluids,vol. 6, pp. 49–54.

Mokry, M. 1984: Prediction of Resonance Frequencies For Venti-lated Wall Wind Tunnels.Wind Tunnels and Testing Tech-niques, AGARD-CP-348, pp. 15-1–15-10.

Monti, R. 1971: Wall Corrections for Airplanes With Lift inTransonic Wind Tunnel Tests.Report of the AGARD Ad HocCommittee on Engine-Airplane Interference and Wall Correc-tions in Transonic Wind Tunnel Tests, AGARD-AR-36-71,pp. III-1–III-13. (Available from DTIC as AD 729 568.)

Morkovin, Mark V. 1956:Fluctuations and Hot-Wire Anemometryin Compressible Flows. AGARDograph 24.

Murthy, S. V.; and Steinle, F. W. 1985: On Boundary Layer Tran-sition in High-Subsonic and Transonic Flow Under the Influ-ence of Acoustic Disturbances and Free-Stream Turbulence.AIAA-85-0082.

Murthy, Sreedhara V.; and Steinle, Frank W. 1986: Effects ofCompressibility and Free-Stream Turbulence on BoundaryLayer Transition in High-Subsonic and Transonic Flows.ACollection of Technical Papers—AIAA 14th Aerodynamic Test-ing Conference, pp. 242–250. (Available as AIAA-86-0764.)

Prandtl, L. 1914: Der Luftwiderstand von Kugleln. Nachrichtender Königlichen Gesellschaft d. Wiss. zu Göttingen.Math.-Phys. K1., no. 2, pp. 177–190.

Ramaswamy, M. A.; and Cornette, E. S. 1980: Supersonic FlowDevelopment in Slotted Wind Tunnels.A Collection of Techni-cal Papers—AIAA 11th Aerodynamic Testing Conference,pp. 165–172. (Available as AIAA-80-0443.)

Robinson, Russell G. 1937: Sphere Tests in the N.A.C.A.8-Foot High-Speed Tunnel.J. Aeronaut. Sci., vol. 4, no. 5,pp. 199–201.

Rose, William C.; and McDaid, Edward P. 1976: Turbulence Mea-surement in Transonic Flow.Proceedings—AIAA 9th Aerody-namic Testing Conference, pp. 267–271.

57

Ross, R.; and Rohne, P. B. 1973: Noise Environment in the NLRTransonic Windtunnel HST. NLR TR-74128 U, Natl. Aerosp.Lab. (NLR).

Rossiter, J. E. 1964: Wind-Tunnel Experiments on the Flow OverRectangular Cavities at Subsonic and Transonic Speeds. R. &M. No. 3438, British ARC. (Supersedes RAETR 64037.)

Schewe, Günter 1983: On the Structure and Resolution ofWall-Pressure Fluctuations Associated With TurbulentBoundary-Layer Flow.J. Fluid Mech., vol. 134, pp. 311–328.

Siddon, Thomas E. 1969:On the Response of Pressure MeasuringInstrumentation in Unsteady Flow (An Investigation of ErrorsInduced by Probe-Flow Interaction). AFOSR 68-2466, U.S.Air Force. (Available from DTIC as AD 682 296.)

Smelt, R. 1945:Power Economy in High-Speed Wind Tunnels byChoice of Working Fluid and Temperature. Rep. No.Aero. 2081, British RAE.

Spangenberg, W. G. 1955:Heat-Loss Characteristics of Hot-WireAnemometers at Various Densities in Transonic and Super-sonic Flow. NACA TN-3381.

Spangler, J. G.; and Wells, C. S., Jr. 1968: Effects of FreestreamDisturbances on Boundary-Layer Transition.AIAA J., vol. 6,no. 3, pp. 543–545.

Stainback, P. Calvin; Johnson, Charles B.; and Basnett, ConstanceB. 1983: Preliminary Measurements of Velocity, Density andTotal Temperature Fluctuations in Compressible SubsonicFlow. AIAA-83-0384.

Timme, Adalbert 1973: Effects of Turbulence and Noise onWind-Tunnel Measurements at Transonic Speeds.FluidMotion Problems in Wind Tunnel Design, AGARD-R-602,pp. 5-1–5-12.

Williams, John 1977: Appendix 4—Aeroacoustic Requirementsfor Model Noise Experiments in Subsonic Windtunnels.A Fur-ther Review of Current Research Related to the Design andOperation of Large Windtunnels, AGARD-AR-105, pp. 61–91.

Willmarth, W. W.; and Roos, F. W. 1965: Resolution and Struc-ture of the Wall Pressure Field Beneath a Turbulent BoundaryLayer.J. Fluid Mech., vol. 22, pt. 1, pp. 81–94.

58

Figure 1. Sources of unsteadiness in transonic wind tunnels adapted from Mabey (1971).

Fan driveSlotted

test section

Heatexchanger

Screens

Flow

Plenum chamber

Return circuit

Boundarylayer noise

Small-scalehigh-frequencyturbulence fromscreens

Unsteadinessfrom plenumchamber flow

Unsteadinessfrom extractionregion

Edge tonesfrom reentryflaps

Noise at fan bladepassagefrequencyandharmonics

Aerodynamicnoise-acousticstanding wavesfrom heat exchanger

Aerodynamicnoise fromducts and turns

Unsteadinessfrom model supportarc-sectorand diffuser

59

(a) Gas properties. (b) Test conditions and fan drive power.

Figure 2. Variation of gas properties and wind tunnel conditions with temperature.M = 1.0; constantpt; and wind tunnelsize.

Figure 3. Operating envelope at constant Mach number.M = 1.0;

Ratio,values to

120°Fvalue

-460 -280 -100 80 260

Tt, °F

4

3

2

1

Density

Speed of sound

Viscosity

8

6

4

2

Reynolds number

Dynamicpressure

Power

Tt, °F

-460 -280 -100 80 260

Ratio,values to

120°Fvalue

8000

6000

4000

2000

Pure Reynolds

number studies

Pure aeroelasticstudies

Tt, min

Tt, max = 150°F pt, max = 130 psi

0 20 40 60 80 100 120 140 × 106

pt, min = 15 psi

Rc

q, psf

c 0.82 ft.=

60

Figure 4. Operating envelope at constant Reynolds number.

Figure 5. Operating envelope at constant dynamic pressure.

Pure Mach number studies

Pure aeroelastic studies

Tt, min

Mmax.

pt, max = 130 psi

M

8000

6000

4000

2000

0 .2 .4 .6 .8 1.0 1.2 1.4

q, psf

Rc 50 106× ; c 0.82 ft.= =

100

80

60

40

20

Pure Mach number studies

Pure Reynolds number studies

RcTt, min

Mmax

Tt, max = 150°F

pt, max =130 psi

M 0 .2 .4 .6 .8 1.0 1.2 1.4

× 106

q 2089 psf;c 0.82 ft.==

61

Figure 6. Plan view of NTF tunnel circuit. All linear dimensions are in feet.

Low-speed diffuser19.7-dia. fan

Turn 2

Turn 1

Turn 3

Turn 4

Cooling coil

Wide-angle diffuser

27-dia. plenum High-speed diffuser2.6° half-angle

16.8 dia.

200

48.6

25 dia.

Screens

Slotted test section8.2 by 8.2

35.7 dia.14.95:1 contraction

62

(a) Static calibration sensitivity versus temperature.

(b) Dynamic range at 1 kHz.

(c) Frequency response at 0.05-psi amplitude.

Figure 7. Pressure transducer 64RY static and dynamic calibration sensitivities.

34

33

32-300 -200 -100 0 100 200

Calibration temperature, °F

Sensitivity,mV/psi

38

36

34.001 .01 .1

Pressure amplitude, psi

Sensitivity,mV/psi

37

35

33

Frequency, Hz0 400 800 1200 1600 2000

Sensitivity,mV/psi

63

Figure 8. Pressure transducer instrumentation plug. All linear dimensions are in inches.

B

B

1.5 dia.

.75 dia.A

A

60° 60°

.040-dia. orifice hole

1.0

.25

Section A-A

.060-ODstainless steel tubing

24-gaugecopper-constantanthermocouple leads

.5

The pressure transducer,thermocouple, and orificeinstrumentation holes areon .375-in-dia. circle.

.102-dia. holefor pressure transducer

Section B-B

.062-dia. holefor thermocouple

Pressuretransducer

.01-IDreferencepressure tube

Electrical leads

64

Figure 9. Locations of dynamic pressure transducers installed in NTF circuit.

Settling chamber, station –52Plenum, station 0

Test section RHS sidewall, station 6.5Station 13 (3 spaced 2.25 in. apart)

Station 16

10.6° cone, station 17.6

Test section LHSsidewall, station 6.5

High-speed diffuser, station 68

Liquid nitrogeninjection station

Station 0

65

L-90-13688

Figure 10. NTF model support strut center body with 10.6° conical fairing.

66

L-90-13687

Figure 11. Tip of 10.6° conical fairing with dynamic pressure transducer installed.

Pressure transducer

67

Figure 12. Tip of 10.6° cone fairing. All linear dimensions are in inches except wind tunnel stations which are in feet.

5.10 2.5

.102-dia. hole forpressure transducer

10.4

5.3°

Station 16.7418.00

.06 rad.

3.00 dia.

Station 17.60

2 Transition stripgrit location

68

L-91-06638

Figure 13. Dynamic data acquisition instrumentation in NTF control room.

69

Figure 14. NTF dynamic data acquisition system wiring block diagram.

Pressuretransducer

Oscilloscope

(monitor)

28-trackFM taperecorder

rmsdigital

voltmeter(monitor)

Spectralanalyzer

Time codegenerator

IRIG A

Calibrationsignal

22-gauge, 2-conductortwisted pair; shieldgrounded at source

10-Vpowersupply

Voltagemonitordigitizer

Facility data acquisitionmainframe computer

Patchboard

Signalconditioner

Digitizedgain signal

70

Figure 15. Longitudinal static pressure gradient for 1 count (0.0001) of buoyancy-induced drag coefficient.

Figure 16. Longitudinal Mach number gradient for 1 count (0.0001) of buoyancy-induced drag coefficient.

.0010

.0008

.0006

.0004

.0002

eq. (B12)

eq. (C18)

eq. (C16)

Assumptions for eqs. (C16) and (C18):

Assumptions for eq. (B12):

0.5-percent blockage area ratio; 5-percent wing area ratio; 0.6 model length ratio

pt = 15 psi;

±0.0075 psi static pressure measurement error

1.21.0 .8.6.4.20M

dp/pt

dx/h

.0020

.0016

.0012

.0008

.0004

eq. (B13)

eq. (C19)

eq. (C17)

1.21.0 .8.6.4.20M

dMdx/h

Assumptions for eq. (B12):

0.5-percent blockage area ratio; 5-percent wing area ratio; 0.6 model length ratio

Assumptions for eqs. (C17) and (C19):pt = 15 psi;

±0.0075 psi static pressure measurement error

71

Figure 17. NTF performance envelope in air mode.Tt = 120°F.

130 psi

120 psi

105 psi

90 psi

75 psi

60 psi

45 psi

30 psi

15 psi

25

20

15

10

5

.1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.1 1.2 1.30

M

R, ft–1

× 106

Data points at test conditions

72

L-91-05320

Figure 18. NTF test section walls with slot covers in place.

RHS sidewallreentry flapsRHS sidewallreentry flaps

Choke locationChoke location

Slot coversSlot covers

Choke locationChoke location

73

Figure 19. NTF performance envelope in nitrogen mode.

Figure 20. NTF operating envelope forM = 0.8.

1.21.0.8.6.4.20

20

40

60

80

100

120

140

160

R, ft–1

M

Data points at test conditions

× 106

20 MW

30 MW

40 MW

50 MW

60 MW-270°F

-240°F-225°F

-260°F-200°F

-150

°F

-119

°F

-100

°F-5

0°F

+0°

F+

75°F

+15

0°F

130

120

110

100

90

80

70

60

50

40

30

20

10

0 20 40 60 80 100 120 140 160 × 106

R, ft–1

pt, psi

Saturationat M = 1.0

Saturationat M = 0.8

88 MW

80 MW 70 MW

Data points at test conditions

74

(a) Test section RHS sidewall station 13.

(b) 10.6° cone.

Figure 21. Power spectral density function versus frequency for fluctuating static pressure coefficient.M = 0.801;R= 3.8× 106 ft−1; pt = 14.92 psi;Tt = 121.1°F; air; dB re 20µPa.

2E-7

2E-8

2E-9

2E-10

2E-11

~p

=q

.007

36

.007

91

.008

18

.008

31

.008

39

.008

45

.008

47

.008

50

.008

51

.008

53

per Hzpq

~ 2

Frequency, kHz0 4 8 12 16 20 24 28 32 36 40

(141

.0)

(141

.6)

(141

.9)

(142

.0)

(142

.1)

(142

.2)

(142

.2)

(142

.2)

(142

.2)

(142

.2)

dB =

Frequency, kHz0 4 8

per Hz

12 16 20 24 28 32 36 40

2E-7

2E-8

2E-9

2E-10

2E-11

.010

1

.010

4

.010

7

.010

8

.010

8

.010

8

.010

9

.010

9

.010

9

.010

9~p

=q

pq

~ 2

(143

.7)

(144

.0)

(144

.2)

(144

.3)

(144

.3)

(144

.3)

(144

.4)

(144

.4)

(144

.4)

(144

.4)

dB =

75

Figure 22. Variation of fluctuating pressure coefficient with streamwise location in test section.M = 0.8; R = 3.8 ×106 ft−1; minimumR boundary; air mode.

Test section RHS sidewall

10.6° cone

Cone

.012

.010

.008

.006

.004

.002

0 2 4 6 8 10 12 14 16 18

Test section station, ft

pq

~

76

Figure 23. Effect of hot wall and cold wall on fluctuating pressure coefficients at test section RHS sidewall station 13.

-.6 -.4 -.2 0 .2 .4 .6 .8 1.0 1.2

Tw – TawT

t – T∞

0

.002

.004

.006

.008

.010

Cold wall Hot wall

pq

pt = 43.2 psi; T

t = –230°F; M = 0.8

pt = 25 psi; –49°F < T

t < 60°F; M = 0.8~

77

(a) pt = 15 psi; ambient temperature; minimum Reynolds number boundary; air mode.

(b) M = 0.5; ambient temperature; air mode.

Figure 24. Effect of fixing boundary layer transition on 10.6° cone.

.2 .4 .6 .8 1.0 1.2

.024

.020

.016

.012

.008

.004

0

Transition fixedTransition freeRepeat test

M

p~

q

2 4 10 20 × 106

.010

.008

.006

.004

.002

0

R, ft–1

Transition fixedTransition freeRepeat test

p~

q

78

Figure 25. Fluctuating pressure coefficient measured at test section RHS sidewall station 13.R= 6 × 106 ft−1; ambienttemperature; air and nitrogen modes.

.2 .4 .6 .8 1.0 1.2M

Nitrogen

Air

0

.002

.004

.006

.008

.010

pq

~

79

(a) M = 0.2.

Figure 26. Power spectra of fluctuating pressure coefficient at test section RHS sidewall station 13.R= 6 × 106 ft−1; ambient temperature;air and nitrogen modes.

Frequency, kHz

0 2010

5E-7

5E-8

5E-9

5E-10

5E-11

~ 2p

qper Hz

AirNitrogen

3.2

kHz

80

(b) M = 0.3.

Figure 26. Continued.

Frequency, kHz

0 2010

2E-7

2E-8

2E-9

2E-10

5E-11

~ 2p

qper Hz

AirNitrogenAirNitrogen

81

(c) M = 0.4.

Figure 26. Continued.

Frequency, kHz

0 2010

4E-7

4E-8

4E-9

4E-10

4E-11

~ 2p

qper Hz

AirNitrogen

82

(d) M = 0.5.

Figure 26. Continued.

Frequency, kHz

0 2010

2E-7

2E-8

2E-9

2E-10

2E-11

~ 2p

qper Hz

AirNitrogen

83

(e) M = 0.6.

Figure 26. Continued.

Frequency, kHz

0 2010

2E-7

2E-8

2E-9

2E-10

2E-11

~ 2p

qper Hz

AirNitrogen

84

(f) M = 0.7.

Figure 26. Continued.

Frequency, kHz

0 2010

4E-7

4E-8

4E-9

4E-10

4E-11

~ 2p

qper Hz

AirNitrogen

850

Hz

85

(g) M = 0.75.

Figure 26. Continued.

Frequency, kHz

0 2010

4E-7

4E-8

4E-9

4E-10

4E-11

~ 2p

qper Hz

AirNitrogen

86

(h) M = 0.8.

Figure 26. Continued.

Frequency, kHz

0 2010

4E-7

4E-8

4E-9

4E-10

4E-11

~ 2p

qper Hz

AirNitrogen

87

(i) M = 0.85.

Figure 26. Continued.

Frequency, kHz

0 2010

4E-7

4E-8

4E-9

4E-10

4E-11

~ 2p

qper Hz

AirNitrogen

88

(j) M = 0.9.

Figure 26. Continued.

Frequency, kHz

0 2010

1E-6

1E-7

1E-8

1E-9

1E-10

~ 2p

qper Hz

AirNitrogen

89

(k) M = 1.0.

Figure 26. Continued.

Frequency, kHz

0 2010

1E-7

1E-8

1E-9

1E-10

1E-11

~ 2p

qper Hz

AirNitrogen

90

(l) M = 1.05.

Figure 26. Concluded.

Frequency, kHz

0 2010

1E-7

1E-8

1E-9

1E-10

1E-11

~ 2p

qper Hz

AirNitrogen

91

Figure 27. Power spectra of fluctuating pressure coefficient at test section RHS sidewall station 13.M = 0.7;R= 6 × 106 ft−1; ambient temper-ature; air and nitrogen modes.

Frequency, kHz

0 2010

4E-6

4E-7

4E-8

4E-9

4E-10

~ 2p

qper Hz

Parameter Air Nitrogen

V, fps 764 777

Peak frequency, Hz 840 855

Reduced 9.02 9.03

frequency, fhV

AirNitrogen

92

Figure 28. Power spectra of fluctuating pressure coefficient in settling chamber, plenum, high-speed diffuser, and liquid nitrogen injection station.M = 0.7;R= 6 × 106 ft−1; ambient temperature; air and nitrogen modes.

Frequency, kHz0 2 4 6 8 10

2E-4

2E-8

~ 2p

qper Hz

Air Nitrogen

1E-8

1E-124E-7

4E-112E-7

2E-11

Settling chamber

Plenum

High-speed diffuser

Liquid nitrogen injection station

~ 2p

qper Hz

~ 2p

qper Hz

~ 2p

qper Hz

sc

93

Figure 29. Power spectra of fluctuating pressure coefficient at test section RHS sidewall station 13.M = 0.7; minimum Reynolds numberboundary; slots covered; air mode.

Frequency, kHz0 2010

4E-7

4E-8

4E-9

4E-10

4E-11

pq per Hz~ 2

94

(a) Phase angle.

Figure 30. Phase angle and coherence of disturbance signal between adjacent pressure transducers at test section RHS sidewall station 13.M = 0.7;R= 6 × 106 ft−1; ambient temperature; comparison of air and nitrogen modes.

Frequency, kHz0 21

270

180

90

0

-90

Air

Nitrogen

Phase angle φ, deg

95

(b) Coherence.

Figure 30. Concluded.

Frequency, kHz21

1.0Air

Nitrogen

Coherence .5

0

96

Figure 31. Effect of fan drive power variation on fluctuating pressure coefficient at test section RHS sidewall station 13.M = 0.8;R= 39.7× 106 ft−1.

Figure 32. Fluctuating pressure coefficient at test section RHS sidewall station 13.M = 0.8; nitrogen mode.

pq

~

.010

.008

.006

.004

.002

Power, MW0 10 20 30 40 50 60

Constant power = 30 MWConstant Tt = –230°FConstant pt = 43.2 psi

.012

.010

.008

.006

.004

.002

0 10 20 30 40 50 60 × 106

R, ft–1

pq

~

97

Figure 33. Fluctuating pressure coefficient at test section RHS sidewall station 13.M = 0.5; air mode; ambienttemperature.

Figure 34. Fluctuating pressure coefficient at test section RHS sidewall station 13.R= 40× 106 ft−1; fan drivepower = 30 MW.

pq

~

.008

.006

.004

.002

02 4 10 20 × 106

R, ft–1

M

p

q

0 .2 .4 .6 .8 1.0 1.2.002

.004

.006

.008

.010

~

98

(a) M = 0.598.

Figure 35. Power spectra of fluctuating pressure coefficient at test section RHS sidewall station 13.R= 40× 106 ft−1; fan drive power =30 MW.

Frequency, kHz0 42

1E-6

1E-7

1E-8

1E-9

1E-10

130

170

210

260

540 84

089

0

2490

~ 2pq per Hz

31

V = 481 fpsf (peak) = 890 Hzfh/V = 15.2

99

(b) M = 0.694.

Figure 35. Continued.

Frequency, kHz

0 42

1E-6

1E-7

1E-8

1E-9

1E-10

5010

5

220

270

390

860

1730

2490

150

~ 2p

qper Hz

V = 519 fpsf (peak) = 860 Hzfh/V = 13.6

31

100

(c) M = 0.742.

Figure 35. Continued.

Frequency, kHz

0 42

1E-6

1E-7

1E-8

1E-9

1E-10

5014

080 26

0

480

90018

0~ 2p

qper Hz

V = 538 fpsf (peak) = 900 Hzfh/V = 13.7

1 3

101

(d) M = 0.793.

Figure 35. Continued.

Frequency, kHz

0 42

2E-6

2E-7

2E-8

2E-9

2E-10

100

960

190

~ 2p

qper Hz

V = 558 fpsf (peak) = 960 Hzfh/V = 14.1

31

102

(e) M = 0.839.

Figure 35. Continued.

Frequency, kHz

0 42

2E-6

2E-7

2E-8

2E-9

2E-10

40

1410

8010

0 160

210

~ 2p

qper Hz

V = 576 fps

31

103

(f) M = 0.892.

Figure 35. Continued.

Frequency, kHz

0 42

2E-6

2E-7

2E-8

2E-9

2E-10

40

2610

8018

013

0

960

~ 2p

qper Hz

31

V = 601 fpsf (peak) = 960 Hzfh/V = 13.1

104

(g) M = 0.992.

Figure 35. Concluded.

Frequency, kHz

0 42

2E-6

2E-7

2E-8

2E-9

2E-10

60

200

960~ 2

p

qper Hz

31

V = 644 fpsf (peak) = 960 Hzfh/V = 12.2

105

(a) Phase angle.

Figure 36. Phase angle and coherence of disturbance signals between adjacent pressure transducers at test section RHS sidewall station 13.R= 40× 106 ft−1; fan drive power = 30 MW; nitrogen mode.

Frequency, kHz

0 1.0.5

270

180

90

0

-90

Phase angle

φ, deg

0.694.742.793

M

106

(b) Coherence.

Figure 36. Concluded.

Frequency, kHz

1.0.5

1.0

Coherence .5

0

0.694.742.793

M

107

Figure 37. Fluctuating pressure coefficient at test section RHS sidewall station 13. Nitrogen mode.

.010

.008

.006

.004

.002

0 .2 .4 .6 .8 1.0 1.2

p

q

~

M

R = 6 × 106 per ft–1; ambient temperature

pt = 43.2 psi; Tt = –250°FMax. R boundary; Tt = –250°Fpt = 20 psi; Tt = –158°F

108

Figure 38. Power spectra of fluctuating pressure coefficient at test section RHS sidewall station 13.Tt = −250°F; maximumR boundary andpt = 43.2 psi.

Max. R boundarypt = 43.2 psi

Arbitraryscale

0 20Frequency, kHz

0 20Frequency, kHz

0 20Frequency, kHz

0 20Frequency, kHz

Arbitraryscale

M = 0.2 M = 0.4

M = 0.8

M = 0.6

M = 0.9

M = 0.7

M = 1.0M = 0.75

109

(a) M = 0.2.

Figure 39. Power spectra of fluctuating pressure coefficient at test section RHS sidewall station 13.Tt = −250°F; maximumR boundary andpt = 43.2 psi; 0- to 2-kHz bandwidth.

2E-6

2E-8

2E-100 1 2

Frequency, kHz

pq

2 per Hz

~( )

Max. R boundary

pt = 43.2 psi

110

(b) M = 0.4.

Figure 39. Continued.

4E-6

4E-8

4E-100 1 2

Frequency, kHz

pq

2 per Hz

~( )

Max. R boundary

pt = 43.2 psi

111

(c) M = 0.6.

Figure 39. Continued.

2E-6

2E-8

2E-10

0 1 2

Frequency, kHz

pq

2 per Hz

~( )

pt = 43.2 psi

112

(d) M = 0.7.

Figure 39. Continued.

4E-6

4E-8

4E-10

0 1 2

pq

2 per Hz

~( )

pt = 43.2 psi

Frequency, kHz

113

(e) M = 0.75.

Figure 39. Continued.

4E-6

4E-8

4E-10

0 1 2

Frequency, kHz

pq

2 per Hz

~( )

Max. R boundary

114

(f) M = 0.8.

Figure 39. Continued.

4E-6

4E-8

4E-10

0 1 2

Frequency, kHz

pq

2 per Hz

~( )

Max. R boundary

pt = 43.2 psi

115

(g) M = 0.9.

Figure 39. Continued.

4E-6

4E-8

4E-10

0 1 2

Frequency, kHz

pq

2 per Hz

~( )

Max. R boundary

pt = 43.2 psi

116

(h) M = 1.0.

Figure 39. Concluded.

2E-6

2E-8

2E-10

0 1 2

Frequency, kHz

pq

2 per Hz

~( )

Max. R boundary

pt = 43.2 psi

117

Figure 40. Fluctuating pressure coefficient at test section RHS sidewall station 13. Air mode.

Figure 41. Effect of slot covers and downstream choke on fluctuating pressure coefficient at test section RHS sidewallstation 13. Minimum Reynolds number boundary; air mode.

.2 .4 .6 .8 1.0 1.2M

0

.002

.004

.006

.008

.010

pq

Min. R boundary; air mode

Max. R boundary; air mode

~

Slots open

Slots covered

Downstream choke on

.2 .4 .6 .8 1.0 1.2M

0

.002

.004

.006

.008

.010

pq

~

118

(a) M = 0.2.

Figure 42. Power spectra of fluctuating pressure coefficient at test section RHS sidewall station 13. Minimum Reynolds number boundary;air mode; slots open and covered.

2E-7

2E-9

2E-110 10 20

Frequency, kHz

pq

2 per Hz

~( )

Slots open

Slots covered

119

(b) M = 0.3.

Figure 42. Continued.

4E-7

4E-9

4E-11

0 10 20

Frequency, kHz

pq

2 per Hz

~( )

Slots open

Slots covered

120

(c) M = 0.4.

Figure 42. Continued.

2E-7

2E-9

2E-11

0 10 20

Frequency, kHz

pq

2 per Hz

~( )

Slots open

Slots covered

121

(d) M = 0.5.

Figure 42. Continued.

4E-7

4E-9

4E-11

0 10 20

Frequency, kHz

pq

2 per Hz

~( )

Slots open

Slots covered

122

(e) M = 0.6.

Figure 42. Continued.

4E-7

4E-9

4E-11

0 10 20

Frequency, kHz

pq

2 per Hz

~( )

Slots open

Slots covered

123

(f) M = 0.7.

Figure 42. Continued.

4E-7

4E-9

4E-11

0 10 20

Frequency, kHz

pq

2 per Hz

~( )

Slots open

Slots covered

124

(g) M = 0.8.

Figure 42. Continued.

2E-7

2E-9

2E-11

0 10 20

Frequency, kHz

pq

2 per Hz

~( )

Slots open

Slots covered

125

(h) M = 0.9.

Figure 42. Concluded.

2E-7

2E-9

2E-11

0 10 20

Frequency, kHz

pq

2 per Hz

~( )

Slots open

Slots covered

126

L-89-00663

Figure 43. NTF test section wall geometry variables for downstream choke with slots open.

RHS sidewallreentry flapsRHS sidewallreentry flaps

Bottom wallreentry flapsBottom wallreentry flaps

Downstreamchoke location

SlotsSlots

Downstreamchoke location

127

Figure 44. Effect of slot covers and downstream choke on NTF fan drive power. Minimum Reynolds number boundary; air mode.

.2 .4 .6 .8 1.0 1.2

Slots open

Slots covered

Downstream choke on

M

0

10

20

30

Fan drive power,MW

128

Figure 45. Power spectra of fluctuating pressure coefficient at test section RHS sidewall station 13. Slots open;M = 0.8;R= 4 × 106 ft−1; airmode.

Frequency, kHz0 4020

2E-7

2E-8

2E-9

2E-10

2E-11

Choke offChoke on

~ 2pq

per Hz

129

Figure 46. Effect of fan speed and inlet guide vane variation for velocity change on fluctuating pressure coefficient attest section RHS sidewall station 13.R= 6 × 106 ft−1; ambient temperature; air mode.

Figure 47. Comparison of NTF and 8-Foot TPT fluctuating pressure coefficient on test section sidewall. Atmosphericstagnation pressure; ambient temperature; air mode.

0 .2 .4 .6 .8 1.0M

Fan speed variation; IGV = 0°

Inlet guide vane variation; 550 rpm

.002

.004

.006

.008

.010

pq

~

.2 .4 .6 .8 1.0 1.2 1.4M

0

.004

.008

.012

.016

8-Foot TPNTFNLR—HST (10° cone)Lowson eq. (1)

Sidewallpt = 1 atm

pq

~

130

(a) αrf = 0°; δmsw= −1.76°. (b) θtsw = 0°; δmsw= −1.76°. (c) θtsw = 0°; αrf = 0°.

Figure 48. Effect of test section geometry variables on fluctuating pressure coefficient at test section RHS sidewall station 13. M = 0.8.;R= 6 × 106 ft−1;ambient temperature; air mode.

Toward flow

–2.4

Model support wall angle, deg

pq

.010

.008

.006

.004

.002

0

~

–2.0 –1.6 –1.2

Toward

flowAway from

flow

Reentry flap angle, deg

pq

.010

.008

.006

.004

.002

0

~

–2 0 2

pq

Towardflow

Away from flow

–.3

Wall divergence angle, deg

.010

.008

.006

.004

.002

0

~

0 .3

131

(a) Wall divergence angle;αrf = 0°; δmsw= −1.76°.

Figure 49. Effect of test section geometry variables on power spectra of fluctuating pressure coefficient at test section RHS sidewallstation 13.M = 0.8.;R= 6 × 106 ft−1; ambient temperature; air mode.

Frequency, kHz

0 2010

4E-7

4E-8

4E-9

4E-10

4E-11

~ 2p

qper Hz

–0.30 0.3

θtsw, deg

132

(b) Reentry flap angle;θtsw = 0°; δmsw= −1.76°.

Figure 49. Continued.

Frequency, kHz

0 2010

4E-7

4E-8

4E-9

4E-10

4E-11

~ 2p

qper Hz

αrf, deg

–1.50 2

133

(c) Model support wall angle;θtsw = 0°; αrf = 0°.

Figure 49. Concluded.

Frequency, kHz

0 2010

4E-7

4E-8

4E-9

4E-10

4E-11

~ 2p

qper Hz

δmsw, deg

–1.34–1.76–2.28

134

Figure 50. Fluctuating pressure coefficient at liquid nitrogen injection station.R= 6 × 106 ft−1; ambient temperature; air and nitrogen modes.

pq

~

.006

.004

.002

M

Air

Nitrogen

0 .2 .4 .6 .8 1.0 1.2

135

(a) M = 0.2.

Figure 51. Power spectra of fluctuating pressure coefficient at liquid nitrogen injection station.R= 6 × 106 ft−1; ambient temperature;air and nitrogen modes.

1E-6

1E-10

1E-140 10 20

Frequency, kHz

Air Nitrogen

pq per Hz

~ 2

136

(b) M = 0.3.

Figure 51. Continued.

1E-6

1E-10

1E-140 10 20

Air Nitrogen

pq per Hz

~ 2

Frequency, kHz

137

(c) M = 0.4.

Figure 51. Continued.

2E-6

2E-10

2E-140 10 20

Frequency, kHz

Air Nitrogen

pq per Hz

~ 2

138

(d) M = 0.5.

Figure 51. Continued.

1E-6

1E-10

1E-140 10 20

Frequency, kHz

Air Nitrogen

pq per Hz

~ 2

139

(e) M = 0.6.

Figure 51. Continued.

1E-6

1E-10

1E-140 10 20

Frequency, kHz

Air Nitrogen

pq per Hz

~ 2

140

(f) M = 0.7.

Figure 51. Continued.

2E-6

2E-10

2E-140 10 20

Frequency, kHz

Air Nitrogen

pq per Hz

~ 2

141

(g) M = 0.75.

Figure 51. Continued.

2E-6

2E-10

2E-140 10 20

Frequency, kHz

Air Nitrogen

pq per Hz

~ 2

142

(h) M = 0.8.

Figure 51. Continued.

1E-6

1E-10

1E-140 10 20

Frequency, kHz

Air Nitrogen

pq per Hz

~ 2

143

(i) M = 0.85.

Figure 51. Continued.

4E-6

4E-10

4E-140 10 20

Frequency, kHz

Air Nitrogen

pq per Hz

~ 2

144

(j) M = 0.9.

Figure 51. Continued.

4E-6

4E-10

4E-140 10 20

Frequency, kHz

Air Nitrogen

pq per Hz

~ 2

145

(k) M = 1.0.

Figure 51. Continued.

4E-6

4E-10

4E-140 10 20

Frequency, kHz

Air Nitrogen

pq per Hz

~ 2

146

(l) M = 1.05.

Figure 51. Concluded.

4E-6

4E-10

4E-140 10 20

Frequency, kHz

pq per Hz

~ 2

Air Nitrogen

147

Figure 52. Time history traces of fluctuating static pressure coefficient at four tunnel stations during cutoff of liquid nitro-gen injection.M = 0.8;pt = 20 psi; andTt = −160°F at initiation.

Time, sec0 8 16 24 32 40

.02

0

-.02

Liquid nitrogen injection cutoff

Initiated Completed

p

q

p

q

p

q

p

q

.4

0

-.4

.005

0

-.005

.01

0

-.01

Liquid nitrogen injection station

Settling chamber

Test section RHS sidewall station 13

High-speed diffuser

sc

148

(a) Liquid nitrogen injection station.

Figure 53. Power spectra of fluctuating pressure coefficient before and after cutoff of liquid nitrogen injection.M = 0.8; pt = 20 psi;Tt = −160°Fbefore cutoff; and dB re 20µPa.

Frequency, kHz

0 2010

4E-6

4E-10

4E-14

~ 2p

qper Hz

LN2 p/q dB

On 0.00330 136.4

- - - - Off .00364 137.2

~

149

(b) Settling chamber.

Figure 53. Continued.

Frequency, kHz

0 2010

2E-4

2E-6

2E-8

LN2 p/q dB

On 0.184 122.4

- - - - Off .179 122.2

sc~

pqsc per Hz

2~

150

(c) Test section RHS sidewall station 13.

Figure 53. Continued.

Frequency, kHz

0 2010

4E-7

4E-9

4E-11

~ 2p

qper Hz

LN2 p/q dB

On 0.0104 146.4

- - - - Off .0107 146.6

~

151

(d) High-speed diffuser.

Figure 53. Concluded.

Frequency, kHz

0 2010

1E-6

1E-10

1E-14

~ 2p

qper Hz

LN2 p/q dB

On 0.00150 129.6

- - - - Off .00151 129.6

~

152

Figure 54. Fluctuating pressure coefficient in settling chamber for maximum Reynolds number boundary in nitrogenmode.

.4

.3

.2

.1

0 .2 .4 .6 .8 1.0

M

~p

qsc

153

(a) Test section slots open.

(b) Test section slots covered.

Figure 55. Fluctuating pressure coefficient in settling chamber. Minimum Reynolds number boundary; air; choked andunchoked.

.2

.1pqsc

~

0 .2 .4 .6 .8 1.0 1.2M

Downstream choke

OffOn

.2

.1pqsc

~

0 .2 .4 .6 .8 1.0 1.2M

Downstream choke

OffOn

154

(a) Linear frequency scale.

Figure 56. Power spectra of fluctuating pressure coefficient in settling chamber.M = 0.8; minimum and maximum Reynolds numberboundaries.

Frequency, kHz

pqsc per Hz

0 42

2E-3

2E-4

2E-5

2E-6

2E-7

175

2

565

925

1140

2165

2330

3130 38

57

3420

132.4 x 106

3.8 x 106Nitrogen

Air

R, ft–1 Gas type

~

1 3

155

(b) Logarithmic frequency scale.

Figure 56. Concluded.

Frequency, Hz

5 400040

2E-3

2E-4

2E-5

2E-6

2E-7

35

50

400

110

175

150

200

565

140 15

5

90

175 22

023

528

5 505

655 92

5

1140

2330

3130 38

57

pqsc

per Hz

2

132.4 x 106

3.8 x 106Nitrogen

Air

R, ft–1 Gas type

~

BP

F

156

Figure 57. Fluctuating pressure coefficient in settling chamber.R= 6 × 106 ft−1; ambient temperature; air and nitrogen modes.

pqsc

~

Air

Nitrogen

.3

.2

.1

0 .2 .4 .6 .8 1.0 1.2M

157

(a) M = 0.2.

Figure 58. Power spectra of fluctuating pressure coefficient in settling chamber.R= 6 × 106 ft−1; ambient temperature; air and nitrogenmodes.

4E-4

4E-8

4E-5

4E-6

4E-7

pqsc

per Hz2

NitrogenAir

0 10 20Frequency, kHz

950

1000

1300

3250

5775

6350

7375

8050

9750

3750

3200

1415

0

1480

0

~

158

(b) M = 0.4.

Figure 58. Continued.

1E-4

1E-8

1E-5

1E-6

1E-7

pqsc

per Hz2

NitrogenAir

0 10 20Frequency, kHz

950

1000

1300

3225

5875

6300

1850

1465

015

275

2650

2550

1850

3350

200

150

~

159

(c) M = 0.6.

Figure 58. Continued.

4E-5

4E-9

4E-6

4E-7

4E-8

pqsc

per Hz2

NitrogenAir

0 10 20Frequency, kHz

950

1000

3450

7525

7550

1410

0

1590

0

3300

200

200

3400

3750

1420

0

~

160

(d) M = 0.7.

Figure 58. Continued.

2E-4

2E-8

2E-5

2E-6

2E-7

pqsc

per Hz2

NitrogenAir

0 10 20Frequency, kHz

600

1000

3600

6250

8100

1600

250

3350

4150

175

~

161

(e) M = 0.75.

Figure 58. Continued.

2E-4

2E-8

2E-5

2E-6

2E-7

pqsc

per Hz2

NitrogenAir

0 10 20Frequency, kHz

950

1200

3600

6350

7050

3250

200

4100

250

950

4250

6500

9600

1175

0

~

162

(f) M = 0.8.

Figure 58. Continued.

650

1700

6400

7800

8450

150

2200

4350

200

950

4250

6600

1050

0

1550

2250

2800

2E-4

2E-8

2E-5

2E-6

2E-7

pqsc

per Hz2

NitrogenAir

0 10 20Frequency, kHz

~

163

(g) M = 0.85.

Figure 58. Continued.

2E-4

2E-8

2E-5

2E-6

2E-7

pqsc

~per Hz

2

NitrogenAir

0 10 20Frequency, kHz

700

100

150

1000

4250

4650

4700

6600

6500

164

(h) M = 1.0.

Figure 58. Continued.

2E-4

2E-8

2E-5

2E-6

2E-7

pqsc

per Hz2

NitrogenAir

0 10 20Frequency, kHz

500

1200

2600

2800

5050

4550

4600

6575

500

~

165

(i) M = 1.05.

Figure 58. Concluded.

2E-4

2E-8

2E-5

2E-6

2E-7

pqsc

per Hz2

NitrogenAir

0 10 20Frequency, kHz

500

500

650

1250

2725 50

50

6750

6600

2800 45

5049

50

~

1175

166

Figure 59. Fluctuating pressure coefficient in settling chamber as function of test section Mach number in nitrogen mode.

R = 40 x 106 ft

–1; power = 30 MW

pt = 43.2 psi; Tt = –250°F

.3

.2

.1

0 .2 .4 .6 .8 1.0M

pqsc

~

167

Figure 60. Fluctuating pressure coefficient in settling chamber as function of test section Reynolds number atM = 0.8.

Figure 61. Effect of fan drive power on fluctuating pressure coefficient in settling chamber.M = 0.8;R= 40× 106 ft−1.

Power = 30 MWTt = –230°Fpt = 43.2 psi

.3

.2

.1

p qsc

~

0 10 20 30 40 50 60 70 × 106

R, ft–1

.3

.2

.1

0 10 20 30 40 50 60

Fan drive power, MW

p qsc

~

168

Figure 62. Fluctuating pressure coefficient in high-speed diffuser as function of test section Mach number.R= 6 ×106 ft−1; air and nitrogen modes; and ambient temperature.

.010

.008

.006

.004

.002

0 .2 .4 .6 .8 1.0 1.2

p

q

~

M

Air

Nitrogen

169

(a) R= 6 × 106 ft−1.

(b) Minimum and maximum Reynolds number boundary.

Figure 63. Fluctuating pressure coefficient in plenum.

p q

~

.004

.002

Air

Nitrogen

0 .2 .4 .6 .8 1.0 1.2M

0 .2 .4 .6 .8 1.0 1.2

Min. R boundary; air

Max. R boundary; nitrogen

.002

M

p q

~

170

(a) M = 0.2.

Figure 64. Power spectra of fluctuating pressure coefficient in plenum. Minimum Reynolds number boundary; air mode.

Frequency, kHz

0 1.0.5

1E-6

1E-8

1E-10

1E-12

1E-14

BPFfsfp

Hz

8199

1932039

5460

81 B

PF

120 18

019

1

240 25

7.5 30

0

360

420

464

480

540

600

660 72

0

780

840 90

0

960p

q per Hz2~

171

(b) M = 0.3.

Figure 64. Continued.

Frequency, kHz

0 1.0.5

2E-7

2E-8

2E-9

2E-10

2E-11

BPFfsfp

Hz

12114619411

2140

6086

100 12

014

717

7.5

190

221

257.

5

420

480

300

600

663 73

4

900

pq per Hz

2~

172

(c) M = 0.4.

Figure 64. Continued.

Frequency, kHz

0 1.0.5

2E-7

2E-8

2E-9

2E-10

2E-11

BPFfsfp

Hz

162193196

1126

20

60

196

100

119

160

177

189

221

257 46

2.5

660

pq per Hz

2~

173

(d) M = 0.5.

Figure 64. Continued.

Frequency, kHz

0 1.0.5

1E-7

1E-8

1E-9

1E-10

1E-11

Hz

194241216

1127

.520

60

195

100

120

180

176

590

659

714

39

87

222

215

240 27

1

779

BPFfsfp

pq per Hz

2~

174

(e) M = 0.6.

Figure 64. Continued.

Frequency, kHz

0 1.0.5

4E-8

4E-9

4E-10

4E-11

4E-12

Hz227288412

27.5

20

60 29210

012

0

181

176

659

714

87

221

494

225

395

731

BPFfsfp

pq per Hz

2~

175

(f) M = 0.618.

Figure 64. Continued.

Frequency, kHz

0 1.0.5

4E-8

4E-9

4E-10

4E-11

4E-12

BPFfs

Hz

227297

27.5

20 60

29410

012

0

176

660

717.

5

87

221

271

732.

5

11

900

39pq per Hz

2~

176

(g) M = 0.642.

Figure 64. Concluded.

Frequency, kHz

0 1.0.5

4E-8

4E-9

4E-10

4E-11

4E-12

BPFfs

Hz

227308

2920

310

100

120

177

659

721

221

292.

5

850

11

320

60pq per Hz

2~

177

Figure 65. High frequency resolution power spectrum of fluctuating pressure coefficient in plenum.M = 0.6; minimum Reynolds num-ber boundary; air mode; and BPF = 227 Hz.

1E-7

1E-9

1E-11

pq per Hz

2~

200 225 250

Frequency, Hz

178

Figure 66. Cross correlation of adjacent pressure transducer signals at test section RHS sidewall station 13.M = 0.998;R= 6.1× 106 ft−1; nitrogen mode; ambient temperature; and transducer separation distance 2.25 in.

1

–1

0

Crosscorrelation

–.0025 0 .0025Time, sec

Peak time = 0.0002246 sec

uc(x) = 835 fps

uc(x) = 0.746

V

179

Figure 67. Convection velocity ratio at test section RHS sidewall station 13.R= 6 × 106 ft−1; ambient temperature; airand nitrogen modes.

Figure 68. Convection velocity ratio at test section RHS sidewall station 13. Minimum Reynolds number boundary; airmode; ambient temperature; and slots open and covered.

0 .2 .4 .6 .8 1.0 1.2

M

.2

.4

.6

.8

1.0

uc(x)V

Nitrogen

Air

M

0 .2 .4 .6 .8 1.0 1.2.2

.4

.6

.8

1.0

Slots open

Slots covered

uc(x)V

180

Figure 69. Effect of hot and cold walls on convection velocity ratio at test section sidewall station 13.M = 0.8; nitrogenmode.

Figure 70. Effect of Reynolds number on convection velocity ratio at test section sidewall station 13.

-.4 -.2 0 .2 .8 1.0 1.2.4 .6

Tw – TawTt – T∞

0

.2

.4

.6

.8

1.0Cold wall Hot wall

uc(x)V

0 .2 .4 .6 .8 1.0 1.2.2

.4

.6

.8

1.0

M

Max. R boundary; nitrogen mode

Min. R boundary; air mode

Intermediate R; nitrogen mode; pt = 43.2 psi

pt = 20 psi; Tt = –158°F; nitrogen mode

uc(x)V

Form ApprovedOMB No. 0704-0188

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources,gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of thiscollection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 JeffersonDavis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503.

1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

4. TITLE AND SUBTITLE 5. FUNDING NUMBERS

6. AUTHOR(S)

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)

11. SUPPLEMENTARY NOTES

8. PERFORMING ORGANIZATIONREPORT NUMBER

10. SPONSORING/MONITORINGAGENCY REPORT NUMBER

12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

13. ABSTRACT (Maximum 200 words)

14. SUBJECT TERMS

17. SECURITY CLASSIFICATIONOF REPORT

18. SECURITY CLASSIFICATIONOF THIS PAGE

19. SECURITY CLASSIFICATIONOF ABSTRACT

20. LIMITATIONOF ABSTRACT

15. NUMBER OF PAGES

16. PRICE CODE

NSN 7540-01-280-5500 Standard Form 298 (Rev. 2-89)Prescribed by ANSI Std. Z39-18298-102

REPORT DOCUMENTATION PAGE

March 1996 Technical Paper

Analysis of Fluctuating Static Pressure Measurements in the NationalTransonic Facility WU 505-59-54-01

William B. Igoe

L-17371

NASA TP-3475

Dynamic measurements of fluctuating static pressure levels were taken with flush-mounted, high-frequencyresponse pressure transducers at 11 locations in the circuit of the National Transonic Facility (NTF) across thecomplete operating range of this wind tunnel. Measurements were taken at test-section Mach numbers from 0.1to 1.2, at pressures from 1 to 8.6 atm, and at temperatures from ambient to−250°F, which resulted in dynamic flowdisturbance measurements at the highest Reynolds numbers available in a transonic ground test facility. Tests werealso made by independent variation of the Mach number, the Reynolds number, or the fan drive power while theother two parameters were held constant, which for the first time resulted in a distinct separation of the effects ofthese three important parameters.

Transonic; Dynamic flow quality; Cryogenic; Pressure fluctuations; High Reynoldsnumber

187

A09

NASA Langley Research CenterHampton, VA 23681-0001

National Aeronautics and Space AdministrationWashington, DC 20546-0001

Unclassified–UnlimitedSubject Category 09Availability: NASA CASI (301) 621-0390

Unclassified Unclassified Unclassified