Low-Speed Fan Noise Reduction
-
Upload
aryatushar -
Category
Documents
-
view
223 -
download
0
Transcript of Low-Speed Fan Noise Reduction
-
8/6/2019 Low-Speed Fan Noise Reduction
1/40
Daniel L. SutliffSEST, Inc., Middleburg Heights, Ohio
Daniel L. TweedtAP Solutions, Inc., Cleveland, Ohio
E. Brian Fite and Edmane EnviaGlenn Research Center, Cleveland, Ohio
Low-Speed Fan Noise ReductionWith Trailing Edge Blowing
NASA/TM2002-211559
May 2002
-
8/6/2019 Low-Speed Fan Noise Reduction
2/40
The NASA STI Program Office . . . in Profile
Since its founding, NASA has been dedicated tothe advancement of aeronautics and spacescience. The NASA Scientific and TechnicalInformation (STI) Program Office plays a key partin helping NASA maintain this important role.
The NASA STI Program Office is operated byLangley Research Center, the Lead Center forNASAs scientific and technical information. TheNASA STI Program Office provides access to theNASA STI Database, the largest collection ofaeronautical and space science STI in the world.The Program Office is also NASAs institutionalmechanism for disseminating the results of itsresearch and development activities. These resultsare published by NASA in the NASA STI ReportSeries, which includes the following report types:
TECHNICAL PUBLICATION. Reports ofcompleted research or a major significantphase of research that present the results ofNASA programs and include extensive dataor theoretical analysis. Includes compilationsof significant scientific and technical data andinformation deemed to be of continuingreference value. NASAs counterpart of peer-reviewed formal professional papers buthas less stringent limitations on manuscript
length and extent of graphic presentations.
TECHNICAL MEMORANDUM. Scientificand technical findings that are preliminary orof specialized interest, e.g., quick releasereports, working papers, and bibliographiesthat contain minimal annotation. Does notcontain extensive analysis.
CONTRACTOR REPORT. Scientific andtechnical findings by NASA-sponsoredcontractors and grantees.
CONFERENCE PUBLICATION. Collectedpapers from scientific and technicalconferences, symposia, seminars, or othermeetings sponsored or cosponsored byNASA.
SPECIAL PUBLICATION. Scientific,technical, or historical information fromNASA programs, projects, and missions,often concerned with subjects havingsubstantial public interest.
TECHNICAL TRANSLATION. English-language translations of foreign scientificand technical material pertinent to NASAsmission.
Specialized services that complement the STIProgram Offices diverse offerings includecreating custom thesauri, building customizeddata bases, organizing and publishing researchresults . . . even providing videos.
For more information about the NASA STIProgram Office, see the following:
Access the NASA STI Program Home Pageat http://www.sti.nasa.gov
E-mail your question via the Internet [email protected]
Fax your question to the NASA AccessHelp Desk at 3016210134
Telephone the NASA Access Help Desk at3016210390
Write to:NASA Access Help Desk
NASA Center for AeroSpace Information7121 Standard DriveHanover, MD 21076
-
8/6/2019 Low-Speed Fan Noise Reduction
3/40
Daniel L. SutliffSEST, Inc., Middleburg Heights, Ohio
Daniel L. TweedtAP Solutions, Inc., Cleveland, Ohio
E. Brian Fite and Edmane EnviaGlenn Research Center, Cleveland, Ohio
Low-Speed Fan Noise ReductionWith Trailing Edge Blowing
NASA/TM2002-211559
May 2002
National Aeronautics and
Space Administration
Glenn Research Center
Prepared for theEighth Aeroacoustics Conferencecosponsored by the American Institute of Aeronautics and Astronauticsand the Confederation of European Aerospace SocietiesBreckenridge, Colorado, June 1719, 2002
-
8/6/2019 Low-Speed Fan Noise Reduction
4/40
Acknowledgments
The CFD design was done as a part of a contract NAS300158, AP Solutions, Inc. Testing, analysis, and reportingwere under contract NAS300170, SEST, Inc. The contributions of Earl Anderson and Sergy Samorezov,
ZIN Technologies, Inc.; Richard Martin, Cleveland State University; Ken Weiland, Test Installation Division,Herb Lawrence, Manufacturing Engineering Division, Tony Shook, Mike Ernst, and Cameron Cunningham,Engineering Design and Analysis Division, and Dr. George Baaklini and Larry Heidelberg, Structures and
Acoustics Division, NASA Glenn were vital.
Available from
NASA Center for Aerospace Information7121 Standard DriveHanover, MD 21076
National Technical Information Service5285 Port Royal RoadSpringfield, VA 22100
Available electronically at http://gltrs.grc.nasa.gov/GLTRS
http://gltrs.grc.nasa.gov/GLTRShttp://gltrs.grc.nasa.gov/GLTRS -
8/6/2019 Low-Speed Fan Noise Reduction
5/40
LOW-SPEED FAN NOISE REDUCTION WITH TRAILING EDGE BLOWING
Daniel L. Sutliff*
SEST, Inc.Middleburg Heights, Ohio 44130
Daniel L. Tweedt
AP Solutions, Inc.
Cleveland, Ohio 44135
E. Brian Fite
and Edmane Envia
National Aeronautics and Space AdministrationGlenn Research CenterCleveland, Ohio 44135
*Senior Aeroacoustics Researcher, Senior Member AIAA
Aerospace Engineer
Aerospace Engineer, Senior Member AIAA
Abstract
An experimental proof-of-concept test was conductedto demonstrate reduction of rotor-stator interactionnoise through rotor-trailing edge blowing. The velocitydeficit from the viscous wake of the rotor blades wasreduced by injecting air into the wake from a trailingedge slot. Composite hollow rotor blades with internalflow passages were designed based on analytical codesmodeling the internal flow. The hollow blade withinterior guide vanes creates flow channels throughwhich externally supplied air flows from the root of theblade to the trailing edge. The impact of the rotor wake-stator interaction on the acoustics was also predictedanalytically.
The Active Noise Control Fan, located at the NASAGlenn Research Center, was used as the proof-of-
concept test bed. In-duct mode and farfield directivityacoustic data were acquired at blowing rates (defined asmass supplied to trailing edge blowing system dividedby fan mass flow) ranging from 0.5 to 2.0%. The firstthree blade passing frequency harmonics at fanrotational speeds of 1700 to 1900 rpm were analyzed.The acoustic tone power levels (PWL) in the inlet andexhaust were reduced 11.5 and 0.1, 7.2 and 11.4, 11.8and 19.4 PWL dB, respectively. The farfield tonepower levels at the first three harmonics were reduced5.4, 10.6, and 12.4 dB PWL. At selected conditions,two-component hotwire and stator vane unsteadysurface pressures were acquired. These measurementsillustrate the physics behind the noise reduction.
Acronyms and Symbols
AAPL Aero-Acoustic Propulsion LaboratoryANCF Active Noise Control FanBPF blade passing frequencyC chord lengthHP horse powerins insertsL/D length-to -diameter ratiom circumferential mode ordern radial mode orderMint integrated fan mass flowopt optimumPWL power levelr radial positionr/s rotor stator interactionR duct radiusRnom nominal radius to farfield mics
rpmc revolutions-per-minute, correctedSPL sound pressure levelTERB Trailing Edge Rotor BlowingU upwash velocityV mean velocity
mean flow angle
stator vane angle hub-to-tip-ratio
Introduction
The velocity deficit due to the viscous wakes of therotor blades is a prime component of rotor-statorinteraction noise.
1The periodic wake disturbance
interacts with the stator causing unsteady surfacepressures on the stator vane that in turn couple to the
NASA/TM2002-211559 1
-
8/6/2019 Low-Speed Fan Noise Reduction
6/40
duct acoustic modes. The strength of the deficitcorrelates to the acoustic levels. It has beendemonstrated analytically that reducing the harmoniccontent of the wake will have a substantial effect onreducing the tone component of the fan noise.
One method to reduce the velocity deficit is to fill the
wakes by injecting air into the wakes from a slot in thetrailing edge. Prior experiments using rotor trailingedge blowing in a blow-down facility
2and inlet guide
vane trailing edge blowing3
have shown that filling thewake through trailing edge blowing reduces theharmonic content of the wake that is responsible forinteraction tones.
Composite hollow rotor blades were designed for thisexperiment with interior passages consisting of aplenum and guide vanes to create flow channelsthrough which air is channeled from the root of theblade to the trailing edge. The air for this experiment is
supplied by a separate external 3-lobed rotary positivedisplacement blower. Analytical codes were used todetermine the optimum shape of the internal passagesand predict the injected wake characteristics along theblade span. The analytical codes also modeled themixing of the injected flow with the free stream. Thismixed wake profile was then used as input to ananalytical noise prediction code to determine theoptimum design-blowing rate.
The trailing edge slot created a thick or blunt trailingedge that with no blowing, was unsuitable for baselinenoise measurements due to vortex shedding. Therefore,a set of inserts that created a sharp trailing edge was
installed to more closely model a realistic rotor blade.Although this extended the chord approximately 0.5 in.(nominal chord, 5 in.) this effect was ignored and therotor blades with inserts were defined as the baselinerotor for comparison.
Blowing rates (defined as mass flow injected at trailingedge divided by fan mass flow) of 0.5 to 2.0% at fanrotational speeds of 1500 to 1900 rpm were tested. In-duct acoustic mode, two-component hotwire velocity,stator vane surface unsteady pressure, farfielddirectivity acoustic data, and fan flow performance datawere acquired. The optimum blowing rate for reducingthe tone noise was found to be between 1.6 and 1.8%.
In addition, a low blowing rate of about 0.5 to 0.6%occurs due to the centrifugal force from the rotation.For this paper, this is defined as self-blowing.
Experimental Apparatus
ANCF Test Bed
A proof-of-concept test was performed on the NASAGlenn 48 in. Active Noise Control Fan
4(ANCF). It is
located in the Aero-Acoustic Propulsion Laboratory(AAPL) shown in figure 1a. The ANCF is a ducted fanused to test noise reduction concepts (figure 1b). Thefour foot diameter fan produces a tip speed of~425 ft/sec resulting in a Blade Passing Frequency(BPF) of approximately 500 Hz. A 16-bladed rotor incombination with a variable stator vane count and
spacing produces the desired rotor-stator interactionmodal content. For the Trailing Edge Rotor Blowing(TERB) test, 14 stator vanes at one-chord spacing wereused. This combination results in a single rotor-statorinteraction mode each at 1BPF and 2BPF, two modes at3BPF.
Trailing Edge Blowing Rotor
The ANCF facility was chosen for this experimentbecause the relatively low speed allows for a relativelysimple design. Sixteen composite hollow rotor bladeswere installed in the ANCF for this experiment. Aphotograph of the installed blades is shown in figure 1c.
The final blade construction is rather complex.Figure 2a shows a model of the assembled blade withthe pressure side skin removed to illustrate the flowpassages. Figure 2b shows an exploded diagram of theblade components. Each component is fabricatedseparately. The base is axisymmetric to allow for fanstagger changes and is fabricated from aluminum. Theinternal flow channels are created by an internalsintered part and the airfoil skins. The forward and aftflow channel boundaries are contained in a singlecomponent fabricated using laser-sintering techniques.Blade skins are made of graphite/epoxy laminates. Finalassembly is completed through use of a cast mold that
locates and holds the components while adhesive iscured to consolidate the components. The internalgeometry is critical in delivering the air to the trailingedge with minimal losses. Care was taken to assure bestpossible surface finishes on all wetted areas. Inaddition, the base of each fan blade was matched to itsmating supply channel in the hub. The hub contained animpeller device that accepted flow from the centraldrive shaft, turned the flow radial, and delivered it tothe fan blade with the proper rotational velocity. A lugon the base of each blade was matched to the top ofeach impeller channel to fix the blade-setting angle.Introducing the supply air through the facility drive
shaft allowed the injection air to be introduced into theANCF rig without affecting the existing flow path andmeasurement envelopes leading to a cleaner researchassessment of the technology capability.
Installation on ANCF
The injection of mass flow through the rotor required adelivery system. The rotor shaft was the obvious designchoice. An 8 in. diameter supply pipe and hose led froma 3-lobed rotary positive displacement blower to the aft
NASA/TM2002-211559 2
-
8/6/2019 Low-Speed Fan Noise Reduction
7/40
base support of the rig. The single pipe was split intofour flexible 4 in. diameter hoses which were furtherreduced to 3 in. before each enters the aft centerbody ofthe ANCF and finally the aft bearing manifold shown infigure 3a. Four ports in the aft bearing manifold acceptinjection air from the 3 in. flexible hoses and supply itto a circumferential plenum that surrounds the drive
shaft. From the plenum, air enters the shaft through fourhelical slots that include angled sides, 15, to match the
incoming flow angle at the design point speed andinjection air Mach number through the slot. Thepassage in the shaft is 5 in. diameter, reducing to 3.6 in.at the entrance to the impeller at the right in figure 3a.The impeller (a sintered part for low cost) turns theflow from axial to radial (and spinning at the rotorspeed) and divides the flow among 16 rectangularpassages of approximately 0.5 in. thick by 0.75 in. widethat supply air to the base of each rotor blade(figure 3b).
Analytical Prediction Methodology
Flow
The ANCF/TERB rotor was designed using a modifiedversion of the NASA developed compressor designprogram
5in conjunction with a three-dimensional
viscous computational fluid dynamics (CFD) code forturbomachinery, RVC3D.
68Through an iterative design
process, several key aerodynamic parameters needed bythe design code were obtained and adjusted based onthe CFD simulation results. In particular, the span wisedistributions of blade row total-pressure loss and exitflow deviation angle (turning) were determined from
the CFD solutions. A two-dimensional viscous CFDcode, DVC2D,9
was used to a limited extent, forexample, to simulate the flow field in the axisymmetricinlet upstream of the rotor, providing inlet boundarycondition data for the rotor computational domain.
Since it was desirable to use the existing stator with theTERB rotor, an RVC3D simulation was performed forthe stator using flow conditions obtained from the rotorsimulation. These results indicated that the existingstator would work well with the TERB rotor.
Simulations of the TERB ANCF rotor were performedusing the RVC3D code augmented with a one-
dimensional flow model for the TERB flowcharacteristics. The one-dimensional model providedspan wise distributions of total-pressure, total-temperature, and flow direction for the TERB jet, basedon flow conditions specified at the rotor center linewhere the TERB supply flow enters the rotor and wasassumed to have known conditions. The model includedthe effects of rotation (centrifugal pumping, work) on theTERB air flowing through the hollow rotor disk andblades, as well as the total-pressure losses associated
with those internal channel flows. Estimates of the total-pressure losses were obtained largely from DVC2D andRVC3D simulation results for portions of the internalflow passages and guide vane array. The resultingsimulated (external) rotor flow field includes the TERB jet emitting from a narrow trailing-edge slot of varyingwidth and extending over most of the blade span.
All rotor CFD simulations, with and without TERB,implemented the Baldwin-Lomax turbulence model
10
for the effects of boundary layer and wake turbulence.In addition to providing valuable assessments of theANCF rotor performance, the simulations also provideddownstream flow field wake predictions suitable foracoustic analyses described in the next section.
It was known prior to CFD simulations that to fill theviscous wake momentum deficit the injection velocitymust be higher than the relative flow velocity(figure 4a). That is since the injection slot height must
necessarily be less the wake thickness. The design massflow distribution was required to weighted to the tipnecessitating an increasing slot thickness with span asshown in figure 4b.
Aeroacoustic Analysis of the Design of the TrailingEdge Blowing Fan
To estimate the noise benefits of the proposed designfor rotor trailing edge blowing, the aeroacousticperformance of the ANCF with and without trailingedge blowing was analyzed using the V072
11code prior
to the test. For a given rotor gust input (i.e., fan wakes),the V072 code computes the three-dimensional acousticresponse of the stator vanes to an incident gust on a
harmonic basis. The code utilizes simplifieddescriptions of the rotor and stator geometry andaerodynamics to provide estimates of the acoustic modelevels produced by the interaction of rotor wakes withthe stator vanes. The accuracy of the predictions can beimproved by utilizing measured or CFD-based three-dimensional descriptions of the rotor wake.
12CFD-
based wakes generated as part of the blown rotordesign process were used to provide the necessary gustinput to the V072 code for the results presented herein.
Mode levels produced by several different blowingrates were computed. The predictions were carried out
for the first three harmonics of the blade passingfrequency tone with 0.0 (i.e., no-blowing), 1.9, 2.0, and2.1% blowing rates. Based on these results, it wasthought that the 2.0% blowing along part of the spanwould offer the optimal combination of aero andacoustic benefits for rotor trailing edge blowing. Inassessing these theoretical benefits, no considerationwas given to the potential broadband noise impact ofthe rotor trailing edge blowing (e.g., vortex sheddingfrom a blunt trailing edge). Neither was the self-noise
NASA/TM2002-211559 3
-
8/6/2019 Low-Speed Fan Noise Reduction
8/40
-
8/6/2019 Low-Speed Fan Noise Reduction
9/40
The self-blowing case brought air into the passagesthrough a duct whose inlet was 10 feet below the ductcenterline. From operational experience, there areknown to be significant vertical temperature gradientsinside the AAPL facility. The iteration process indicatesthat the temperature in the wake is indeed cooler, to amaximum of ~5 F cooler near the wake centerline.
For the blowing case, it is expected that there is atemperature rise in the wake. The iterative methodindicated that the peak rise is ~4.5 F near thecenterline. In addition, the variation in the wake resultsin an iteratively converged velocity that has asignificantly different characteristic than the velocityprofile from either overheat ratio. The presumed actualvelocity profile is overblown, a characteristic notindicated from the unadjusted profiles. In addition, bothcases have nearly identical (~11.5 F) bulk temperatureincreases that are probably due to the error intemperature that arise from the different locations of the
hotwire and the temperature measurement device; and asystematic error in the temperature correction due to thelarge difference between the calibration and theexperiment.
These adjustments produce physically reasonableresults, though unconfirmed by independentmeasurement. However, it is noted that the analyticalprofile solution (see Comparison to Analytical Resultssection) for the optimum-blowing case is predicted tobe slightly overblown. The adjusted profile matches thisprofile better. The constant temperature correctionderived in the inviscid portion of the velocity profileindicates the correction is valid. In addition, the
adjustments are similar across a wide variety ofconditions. Therefore, all hotwire data presented hereinare presented with this iteratively adjusted velocity andtemperature profile.
Surface Pressures
Unsteady stator vane surface pressures were alsoacquired only at 1800 rpm for the baseline rotor andoptimum blowing rate. The suction and pressure side ofa single stator vane were each instrumented with30 microphones as detailed in figure 8. Themicrophones were flush mounted on the surfaces anddistributed along three span locations (r/R = 0.49, 0.74,
and 0.91) and a radial line at 20% chord.
The time histories were acquired synchronous to theshaft rotation at 256 samples-per-revolution for500 revolutions. A frequency domain averaged FFTwith an ensemble length of five revolutions wasobtained from the time histories. The harmonics of theblade passing frequency up to the Nyquist frequencywere obtained from the spectra, with the first threeharmonics being of the most interest. The tonal
component of the unsteady surface pressure has beendemonstrated
14to be directly related to the acoustic
levels.
Rotating Rake
The rotating rake instrumentation system provides a
complete map of the duct modal signature at 1BPF,2BPF, and 3BPF for either the inlet or exhaust duct.The circumferential modes arise from a Dopplerinduced frequency shift due to the unique and discreterotation rate of each m-order. Radial modes (n) arecomputed from a least squares data fit of the radialpressure profile using hardwall Bessel functions as thebasis functions.
15Rotating rake data were acquired for
the entire fan speed range and blowing rates.
The modal data from the rotating rake will be presentedin 3-D format. The base plane axes are m- and n-order,and the vertical value axis in the PWL in the (m,n)mode. The mode power level is the sum of all cut-on
rotor-stator interaction modes. Along the wall of the m-order axis the sum of all the radials provides the powerin that circumferential mode. The sum of all providesthe PWL in the harmonic presented. The typical 3-Dchart provides information as to the dominant modespresent, usually those due to the rotor-stator interaction.Of secondary interest, are all other modes that may bedue to inflow distortions (often called extraneousmodes). Thus, a table for each 3-D chart will bepresented to indicate the total power in the harmonic,the total power in just the rotor-stator interactionmode(s), and the power in the extraneous modes.The Tyler-Sofrin rotor-stator modes
16expected with
16 blades and 14 stators with their cut-on rotational
speeds are presented on figure 10.
Farfield
Farfield acoustic data were also acquired over the entirerange blowing rates and fan speeds. Twenty-eightmicrophones were distributed along an arc ofapproximately 40 ft. radius with 5 increments.Figure 10 provides the farfield microphone locations.Data were synchronously sampled at 256 sample-per-revolution and were obtained by frequency domain fiverevolution ensembles. Tonal Sound Pressure Level(SPL) directivity was obtained at each blade passingharmonic. The SPL directivity was integrated over the
directivity angle assuming constant SPL over theazimuthal angle to obtain the tone PWL.
Experimental Results
The rotating rake acoustic data was acquired first todetermine the optimum-blowing rate. The optimumblowing rate was determined by the minimum of thesummation of the first three tone PWLs. However, thedata is presented from the noise generating mechanismto the ultimate metric, the farfield directivity.
NASA/TM2002-211559 5
-
8/6/2019 Low-Speed Fan Noise Reduction
10/40
The design goal was to reduce the rotor-statorinteraction mode at the fan blade passing frequencies.This was determined by comparing the levels for thefirst three harmonics for a given blowing rate andcomparing to the baseline. Unless otherwise mentionedthe results are for full span blowing.
Rotor Wake
The circumferentially averaged mean values forvelocity and flow angle as a function of radial span arepresented in figure 11. These data are taken one rotorchord behind the rotor in the plane of the stator leadingedge, although the stators are not present. Three basicconditions are presented: (a) rotors with trailing edgeinserts installed, (b) the self-blowing case (0.6%blowing rate), (c) optimum-blowing of 1.8%. Thebaseline rotor velocity profile is reasonably uniform.The velocity profile with self-blowing also appears tobe more uneven compared to that with inserts. This isprobably due to the centripetal forces that create the
self-blowing result in an un-even flow from the trailingedge, as well as a possibility of slight flow circulationbetween the passages. The self-blowing rate appears toreduce the velocity at the inner portion of the span. Thevelocity profile for the optimum-blowing rate is alsosimilarly reduced at the inner span. Integrating themean velocity profile along the radial direction for thecase with inserts and comparing to the blowing casesindicates a slight decrease in overall mass flow (0.8%)with self blowing; and a increase in mass flow (1.4%)with optimum blowing. The overall mass flow has beenincreased with optimum blowing, approximately theamount that has been injected.
The change in profile may be due to the blunt trailingedge causing vortex shedding or flow separation fromthe blade, which has been noted to reduce the meanflow.
It is likely that the blowing prevents this
undesirable flow from forming. This is especially truenear the tip. However, by design, less mass flow isdirected to the inner span. It is possible that in the innerspan, vortex shedding/flow separation is occurring,resulting in the lower mean velocity. There is lesspoint-to-point variation along the radial profile with theapplication optimum blowing due to the positive massflow not allowing the circulation between passages. Theangle profiles are similar for all three cases; with
perhaps a slight decrease (~1 to 2) in the turning angleas blowing is increased.
Hotwire measurements were taken behind the rotor when the trailing
edge was fully taped, creating a blunt trailing edge. Thesemeasurements showed a 5 to 10 fps drop in the velocity profile acrossthe span. The trailing edge taped rotor was judged unsuitable for anexperimental baseline, but may have provided insights into the resultswith blowing.
The passage averaged circumferential velocity, flowangle, and upwash velocity as a function of radialposition are presented in figure 12 as contour plots.Selected radial profiles from the circumferentialpassage are shown in figure 13. The self-blowing caseactually increases the velocity deficit. This is probablydue to the thickness of trailing edge (compared to the
sharp trailing edge) creating a thick wake that the lowblowing rate does not fill. The optimum-blowing rateactually over fills the wake, or over-blows, at radialstations from about 50% to the near the tip. The hubseparation is greater with the thicker trailing edge,which is not remedied by increased blowing. The wakeangle deviation is affected by blowing. The flow anglewith self-blowing is somewhat less than with theinserts. The flow angle resulting when optimumblowing is applied is considerably less. The flowdeviation reverses direction when the wake isoverfilled. The result is that the upwash as calculatedfrom Eq. (1) is modestly reduced over most of the
radial span when self-blowing is applied, andconsiderably reduced with optimum blowing.
Stator Vane Surface Pressures
The unsteady stator vane surface pressures for the firstthree harmonics at the 20% chord line are presented infigure 14 for the suction side and figure 15 for thepressure side. The unsteady pressure at the 20% chordline has been shown to be the major contributor to tonenoise for this fan and indicative of the overall levels.
13
The vane surface SPL for the case with the inserts andwith self-blowing are approximately the same for allthree harmonics. The surface SPLs with optimumblowing applied are significantly lower, especially near
the tip. This is the case for both the suction and pressuresides of the stator vane. The optimum-blowing caseshows an extreme minimum and a phase reversal nearthe 50% span. This location corresponds to thetransition between under- and over-blowing indicatedby the hotwire.
Acoustic Duct Modes
Figures 16, 17, and 18 show the modal decompositionfor the inlet and exhaust for the first three fanharmonics. At BPF, in the inlet, with optimum-blowinga reduction in m=2 of 11.5 dB occurs. In the exhaust adecrease of 5.0 dB is noted with self-blowing, but an
increase of 0.1 dB results when optimum-blowing isapplied.
The second harmonic rotor-stator mode (m=4, with tworadials) reductions are 7.2 dB (inlet) and dB 11.4(exhaust). It also becomes apparent that the non-rotorstator modes are reduced 1.5 dB (inlet) and 6.1 dB(exhaust). The overall harmonic PWL is reduced 6.4 dBand 10.4 dB.
NASA/TM2002-211559 6
-
8/6/2019 Low-Speed Fan Noise Reduction
11/40
The third harmonic rotor-stator interaction modes arereduced 11.8 and 19.4 dB with the application ofoptimum blowing. For the most part all modes arereduced to the measurement noise floor. The overallPWL reductions in the 3rd harmonic are 13.6 dB and18.5 dB in the inlet and exhaust ducts, respectively.
The significance is that a reduction in the rotor-statormode has a nearly 1-to-1 dB impact on the harmonicPWL. To the extent the extraneous modes are due tothe blade-to-blade rotor wake differences interactingwith the stator rather than distortions in the inflow fieldreacting with the rotor this is an expected, but usefulresult.
The effect of varying the blowing rate from self-blowing (0.6%) to 2.0% is presented in figure 19. Thelack of effectiveness at BPF maybe due to thatharmonic being a result of the strong tip flow, which isnot modified with the application of blowing. A second
possibility is the BPF levels are from the interaction ofthe rotor potential field variation rather than the wakedeficit interaction with the stator vane. At the secondand third harmonics the clear minimum at 1.8% isnoted. Also, note that this fan is dominated by the (m,0)modes. Blowing reduces all radial modes reasonablyuniformly.
The effect of blowing along only part of the rotor spanwas investigated by taping the trailing edge except for20% span from the tip, with the results shown infigure 20. It is seen that the minimum occurs at a lowerblowing rate of about 1.1%. (It is known that the designfull-span flow rate is heavily tip weighted.) The
reductions are approximately the same or a few dB lesscompared to full span blowing at 1.8%. This is becausethe ANCF is dominated by the (m,0) modes and fillingthe wake at 20% tip couples very well to the tip-dominated (m,0) modes. The reduction in the overall m-order is primarily due to reduction in the (m,0) mode.Unlike the full span blowing case, the higher modes aremostly unaffected by tip blowing. This indicates thatcarefully selected blowing at only spans that couple todominant acoustics may result in lower blowing rates toaccomplish similar reductions. However, the tapedsection of the rotor was effectively a blunt trailing edgethat created enormous broadband noise,
**which will be
briefly described in the farfield results. If part-span
**Note: farfield measurement were taken with two other rotor trailing
edge conditions: (i) the rotor trailing edge completely taped, and(ii) the duct inflow to the blowing system completely blockedresulting in no net mass flow through the blade. These configurationsresulted in modest changes in the tones but tremendous (~20 dB)increases in the farfield broadband SPL at certain frequencies aswould be expected
**from what is effectively a blunt trailing edge. It
is only mentioned here to indicate that caution must be exercisedwhen designing the rotor blade to anticipate blowing failure or part-span blowing conditions.
blowing is to be useful, resulting in lower blowingrates, very careful design to condition will be required.
The effect of blowing at other fan speeds resulted insimilar mode level versus blowing rate profiles.Figure 21 compares interaction mode PWLs from thebaseline case to the levels at the optimum blowing, as
determine for each speed and harmonic. The maximumreduction generally occurs at 1.8%. A few cases, mostlyBPF in the exhaust, better reductions are obtained atlower blowing rates.
Farfield Directivity
The farfield directivities for the first three harmonicsfrom 0.6 to 2.0% are presented in figure 21. For eachblowing rate, the tonal directivity is plotted along withthe tonal directivity with rotor trailing edge inserts forcomparison. The integrated tone PWL is noted on theplot. The farfield results confirm the in-duct modemeasurements. For example, note the directivity for the
first harmonic at optimum blowing (1.8%): the inletlobe is significantly lowered by 7.5 dB SPL at the peakof the lobe, but an increase of 1.2 dB SPL at the exhaustlobe peak. This matches the in-duct results very well.However, the farfield directivity indicates that 1.6%blowing produced greater reduction than 1.8% for allthree harmonics. This is partially a result of uncertaintyin the blowing mass flow measurement. The uncertainty
in the blowing ratio is estimated to be about 0.05%.An additional factor is that the presence of the exhaustrake may slightly back-pressure the fan effecting themeasurements taken with the rotating rake. It may alsobe that finer blowing rate increments would identify aminimum that both sets of measurements agree upon.
Comparative Tonal Summary
The reductions in the tone harmonics calculated ormeasured from the different measurement methods arepresented in figure 23. First, the average upwash acrossthe radial span as measured by the hotwire is calculatedfrom the FFT of the passage-averaged upwash. This isexpected to correlate to the duct mode PWLs asoutlined in references 2 and 11. The optimum-blowingharmonic upwash is referenced to that calculatedbehind the rotor with inserts. A reduction in theharmonic upwash of 8 to 10 dB is calculated. Next, theaverage SPL for all vane microphones for the first three
harmonics is computed. This metric has been shown inreference 13, for this fan only, to correlative very wellto the in-duct mode PWLs. The measured in-duct andfarfield PWLs are also presented. Overall, the trendsand comparisons for each measurement method agreeand compliment each other very well.
NASA/TM2002-211559 7
-
8/6/2019 Low-Speed Fan Noise Reduction
12/40
Comparison to Analytical Results
The other goal of this project was to validate the flowand acoustic codes used as design tools. Comparisonsof the experimental to the predicted data are presentedin this section. The comparisons between theexperimental data and CFD results are not at exactly
matching conditions primarily because the hotwire datawere acquired at 1 rotor chord (5 in.) behind the rotor,while the CFD results are 4.82 in. downstream of therotor stacking axis, which corresponds to the CFD gridexit. That is, the CFD results are 2.7 in. upstream of thehotwire. Furthermore, for the baseline comparisons, theexperimental results are from a sharp trailing edge,while that CFD results are from a rotor with a thickertrailing edge, though with out modeling the vortexshedding. Finally, it was decided to compare optimum-blowing, to optimum-blowing as defined by theseparate methodologies, i.e., 1.8% experiment versus2.0% CFD.
PerformanceThe CFD results predicted the baseline fan wouldproduce a mass flow of 131.7 lb
m /sec at 2000 rpmc at
maximum absorbed horsepower. Experimentally(interpolating figure 6), the fan produced 131.5 lb
m/sec
at 1910 rpmc at 100% horsepower. It is speculated fromin-situ measurements that blade setting angle asinstalled was slightly higher than designed.
Flow
The comparison of mean wake profile behind the rotoris shown in figure 24 for the baseline. The turning angleagreement is excellent, with the code over predicting
the angle by less than 1 to 2 degrees. The mean velocitycomparison shows agreement within 5 fps for thebaseline rotor, about 10 fps with blowing applied. Morenotable is the character of the profile. The code predictsa higher velocity at the hub. The experimental profileshowed a uniform profile with no blowing, andnoticeable unloading at the hub with blowing applied.
Figure 25 presents the wake profiles at the selectedradial positions for the baseline rotor. In general, thecode overpredicts the velocity deficits and thedeviation angle. This is probably because themeasurements correspond to a location furtherdownstream, but also because the Baldwin-Lomax CFD
may not have enough turbulent mixing17 in the wake.The phase of the CFD results was adjusted to accountfor convection by the mean swirl.
The comparisons in the wake profiles with blowingapplied are shown in figure 26. The code predicted theslight over-blowing near tip-ward and slight under-blowing hub-ward. The code shows less deviation thanwas measured experimentally. This may be a result ofuneven radial distribution in the experiment.
Acoustic
The predicted harmonic tone levels from V072 usingthe CFD results are compared to those measured by therotating rake in figure 27. The absolute levels for thebaseline rotor and with blowing applied as well as thereductions obtained with blowing are presented. In
general, the reductions due to blowing are estimatedaccurately, except for exhaust BPF. The analyticalresults predicted substantial reduction that was notmeasured experimentally in the exhaust.
Conclusions
The rotor blades of a low-speed fan were designed toreduce the rotor-stator interaction noise through the useof rotor trailing edge blowing. Composite hollow rotorblades were designed with internal passages to deliverthe injected flow at the design pressure and flow rate tofill the wake momentum deficit. CFD and analyticalcodes were developed and used as tools to optimize thedesign.
Types of data acquired were: (i) two-componenthotwire behind the rotor, (ii) unsteady surface pressureson a stator vane, (iii) acoustic duct modes, and(iv) farfield directivity. These data were analyzed fortonal character.
The rotating rake tonal analysis indicated that theviscous wake is essentially filled at a blowing rate of1.8% of the fan mass flow rate. The optimum-blowingrate as defined by the minimum acoustic levels wasbetween 1.6 and 1.8%. Blowing had modest effects onthe BPF tone in the exhaust. Blowing significantlyreduced all rotor-stator interaction modes and otherextraneous modes at the second and third harmonics.Acoustic tone power levels in the inlet and exhaustwere reduced 11.5 and 0.1, 7.2 and 11.4, 11.8 and19.4 PWL dB, respectively, at the first three harmonicsof the Blade Passing Frequency. The farfield directivityconfirmed the reductions obtained. The reductionsobtained in the farfield were 5.4 (1BPF), 10.6 (2BPF),and 12.4 (3BPF) dB tone PWL.
Reduction in the fan tone levels by filling the rotorviscous wake through trailing edge blowing has beendemonstrated to achieve substantial tone reduction at1.6 to 1.8% of the fan mass flow rate. Indirect methods
indicate that broadband reduction of rotor-statorinteraction noise may result.
The design codes used in this work were validatedas reliable tools for predicting the behavior of trailingedge blowing for a low speed fan. Simulations ofthe TERB ANCF rotor using the RVC3D code,augmented with a one-dimensional flow model forthe TERB flow characteristics, predicted theexperimentally values very well. Using these results
NASA/TM2002-211559 8
-
8/6/2019 Low-Speed Fan Noise Reduction
13/40
-
8/6/2019 Low-Speed Fan Noise Reduction
14/40
-
8/6/2019 Low-Speed Fan Noise Reduction
15/40
(a) ANCF Hollow Fan Blade with pressure side skin removed
(b) Exploded view showing blade components.
Figure 2. Details of Composite Trailing Edge Rotor Blade
NASA/TM2002-211559 11
-
8/6/2019 Low-Speed Fan Noise Reduction
16/40
-
8/6/2019 Low-Speed Fan Noise Reduction
17/40
9
12
15
18
21
24
0 0.02 0.04 0.06 0.08 0.1 0.12
0 0.01 0.02 0.03
Rad
ius
(in)
unit mass flow (lbm/sec / in)
trailing edge slot thickness (in)
9
12
15
18
21
24
200 250 300 350 400 450
Radius
(in)
Velocity - (fps )
(a) Injection velocity as a function of span
(b) Slot thickness and unit mass flow as a function of span
Figure 4. Trailing edge design parameters
NASA/TM2002-211559 13
-
8/6/2019 Low-Speed Fan Noise Reduction
18/40
Stator Vanes(14 @ 1 rotor chord)
Rotor16 Blades
InletRotating
Rake = 0.0
ExhaustRotating
Rake = 0.5
InflowControlDevice
rotor-stator planehub-to-tip ratio = 0.375
Fan Diameter = 48"
ExhaustL/D ~1
InstrumentedStator Vane
Hotwire Measurement Planeat leading edge of stator vanes,
centered between 2 vanes.(same location with or without stators)
InletL/D ~1
PITOT Staticmeasurementplane @ exit
Figure 5. Schematic of ANCF showing measurement locations
NASA/TM2002-211559 14
-
8/6/2019 Low-Speed Fan Noise Reduction
19/40
-
8/6/2019 Low-Speed Fan Noise Reduction
20/40
-
8/6/2019 Low-Speed Fan Noise Reduction
21/40
-
8/6/2019 Low-Speed Fan Noise Reduction
22/40
INLET 1018 887 1547 833 1304 1696 1072 1568
EXHAUST 894 863 1473 829 1247 1709 1071 1533
MODE--> (2,0) (4,0) (4,1) (6,0) (6,1) (6,2) (-8,0) (-8,1)(m,n)
1BPF 2BPF 3BPF
index/angle/radius
1 2.4 46.7
2 7.5 45.1
3 13.1 43.54 18.8 43.0
5 24.2 42.1
6 29.8 40.8
7 35.8 39.9
14 82.4 37.3
13 76.0 37.2
12 67.5 37.5
8 43.2 39.1
9 49.6 39.210 54.8 38.8
15 88.6 37.2
17 102.0 38.7
18 108.4 39.0
16 95.2 38.1
20 121.0 41.219 114.6 39.6
27 160.3 47.6
28 169.2 45.1
22 134.0 42.4
23 138.2 43.2
24 144.2 44.1
25 149.7 45.2
26 155.0 46.3
index/angle/radius
11 61.6 37.6
21 126.8 41.3
Rnom = 40'
angle measuredfrom inlet axis
radius in feet
Figure 9. Modal characteristics
Figure 10. Farfield arena microphone locations
NASA/TM2002-211559 18
-
8/6/2019 Low-Speed Fan Noise Reduction
23/40
-
8/6/2019 Low-Speed Fan Noise Reduction
24/40
-
8/6/2019 Low-Speed Fan Noise Reduction
25/40
-
8/6/2019 Low-Speed Fan Noise Reduction
26/40
-
8/6/2019 Low-Speed Fan Noise Reduction
27/40
-
8/6/2019 Low-Speed Fan Noise Reduction
28/40
-
8/6/2019 Low-Speed Fan Noise Reduction
29/40
-
8/6/2019 Low-Speed Fan Noise Reduction
30/40
-
8/6/2019 Low-Speed Fan Noise Reduction
31/40
-
8/6/2019 Low-Speed Fan Noise Reduction
32/40
-
8/6/2019 Low-Speed Fan Noise Reduction
33/40
70
80
90
100
110
120
70
80
90
100
110
120
PWL
(
dB)
70
80
90
100
110
120
70
80
90
100
110
120
PWL
(dB)
70
80
90
100
110
120
70
80
90
100
110
120
PWL
(dB)
(c)RPMc = 1700
INLET EXHAUST INLET EXHAUST INLET EXHAUST
1xBPF 2xBPF 3xBPF
(b)RPMc = 1800
(a)RPMc = 1900
Inserts Optimim
1.8%
1.6%
1.6% 1.4%
1.8% 1.8% 1.8% 1.8%
1.6% 1.8% 1.8%
0.6%
1.6% 1.8% 1.4% 1.8% 1.6%1.4%
INLET EXHAUST INLET EXHAUST INLET EXHAUST
1xBPF 2xBPF 3xBPF
INLET EXHAUST INLET EXHAUST INLET EXHAUST1xBPF 2xBPF 3xBPF
Figure 21. Maximum reductions obtained by blowing, several RPM
NASA/TM2002-211559 29
-
8/6/2019 Low-Speed Fan Noise Reduction
34/40
-
8/6/2019 Low-Speed Fan Noise Reduction
35/40
1BPF 2BPF 3PBF
0
2
4
6
8
10
12
14
16
18
20
Reduction
inIndicatedMetric(
dB)
HARMONIC
20log10(Upwash
opt/Upwash
ins)
Average SPL of all Vane mics (suction side)
Average SPL of all Vane mics (pressure side)
In-duct PWL (inlet)
In-duct PWL (exhaust)
Integrated Farfield PWL
Figure 23. Comparison of reduction obtained with optimum blowing using various measurement methods
NASA/TM2002-211559 31
-
8/6/2019 Low-Speed Fan Noise Reduction
36/40
9
12
15
18
21
24
10 15 20 25 30 35
145 155 165 175 185 195
Radius
(in
)
Angle (deg)
Velocity (fps)
angle (experimental)
angle (CFD)
velocity (experimental)
velocity (CFD)
(a) Trailing Edge Inserts
(b) Optimum Blowing
9
12
15
18
21
24
10 15 20 25 30 35
145 155 165 175 185 195
Radius
(in)
Angle (deg)
Velocity (fps)
NOTE:
Experiment measurement plane was
5.0" from rotor trailing edge.
CFD computational plane was
2.2" from rotor trailing edge.
Figure 24. Comparison of predicted and experimental flow parameters downstream of the fan
NASA/TM2002-211559 32
-
8/6/2019 Low-Speed Fan Noise Reduction
37/40
-
8/6/2019 Low-Speed Fan Noise Reduction
38/40
150
160
170
180
V
el
(fps)
160
170
180
190
Vel
(fps)
160
170
180
190
Vel
(f
ps)
160
170
180
190
Vel
(fps)
160
170
180
190
Vel
(fps)
155
165
175
185
0 2 4 6 810121416182022
Vel
(fps)
10
15
20
25
A
ng
(deg)
10
15
20
25
Ang
(deg)
10
15
20
25
Ang
(deg)
10
15
20
25
Ang
(deg)
15
20
25
30
Ang
(deg)
25
30
35
40
Ang
(d
eg)
23.65
22
19
16
13
10
Radius
(inches)
ROTATION
Hotwire measurements
(1.8% blowing rate)
CFD Results
(2.0% blowing rate)
Figure 26. Comparison of measured and predicted fan wake profiles with optimum blowing
NASA/TM2002-211559 34
-
8/6/2019 Low-Speed Fan Noise Reduction
39/40
BPF 2BPF 3BPF
85
90
95
100
105
110
115
PWL
(dB)
BPF 2BPF 3BPF
0
2
4
6
8
10
12
14
16
18
20
del-P
WL
(dB)
(a) Inlet
(b) Exhaust
BPF 2BPF 3BPF
85
90
95
100
105
110
115
PWL
(dB)
BPF 2BPF 3BPF
0
2
4
6
8
10
12
14
1618
20
del-PWL
(dB)
PREDICTION
ABSOLUTE LEVELS
baseline
optimum
blowing
EXPERIMENT
OPTIMUM BLOWING
REDUCTIONS
PREDICTIONEXPERIMENT
Figure 27. Comparison of measured and predicted tone power levels
NASA/TM2002-211559 35
-
8/6/2019 Low-Speed Fan Noise Reduction
40/40
This publication is available from the NASA Center for AeroSpace Information, 3016210390.
REPORT DOCUMENTATION PAGE
2. REPORT DATE
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.
Form Approved
OMB No. 0704-0188
12b. DISTRIBUTION CODE
8. PERFORMING ORGANIZATION
REPORT NUMBER
5. FUNDING NUMBERS
3. REPORT TYPE AND DATES COVERED
4. TITLE AND SUBTITLE
6. AUTHOR(S)
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
11. SUPPLEMENTARY NOTES
12a. DISTRIBUTION/AVAILABILITY STATEMENT
13. ABSTRACT (Maximum 200 words)
14. SUBJECT TERMS
16. PRICE CODE
15. NUMBER OF PAGES
Technical Memorandum
National Aeronautics and Space Administration
John H. Glenn Research Center at Lewis Field
Cleveland, Ohio 44135 3191
1. AGENCY USE ONLY (Leave blank)
10. SPONSORING/MONITORING
AGENCY REPORT NUMBER
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)
National Aeronautics and Space AdministrationWashington, DC 20546 0001
Available electronically athttp://gltrs.grc.nasa.gov/GLTRS
May 2002
NASA TM2002-211559
E13339
WU781301100
41
Low-Speed Fan Noise Reduction With Trailing Edge Blowing
Daniel L. Sutliff, Daniel L. Tweedt, E. Brian Fite, and Edmane Envia
Acoustics
Unclassified - Unlimited
Subject Categories: 07 and 71 Distribution: Nonstandard
Prepared for the Eighth Aeroacoustics Conference cosponsored by the American Institute of Aeronautics and Astronau-
tics and the Confederation of European Aerospace Societies, Breckenridge, Colorado, June 1719, 2002. Daniel L.
Sutliff, SEST, Inc., Middleburg Heights, Ohio; Daniel L. Tweedt, AP Solutions, Inc., Cleveland, Ohio; and E. Brian Fite
and Edmane Envia, NASA Glenn Research Center. Responsible person, Daniel L. Sutliff, organization code 5940,
2164336290.
An experimental proof-of-concept test was conducted to demonstrate reduction of rotor-stator interaction noise through
rotor-trailing edge blowing. The velocity deficit from the viscous wake of the rotor blades was reduced by injecting air
into the wake from a trailing edge slot. Composite hollow rotor blades with internal flow passages were designed based on
analytical codes modeling the internal flow. The hollow blade with interior guide vanes creates flow channels through
which externally supplied air flows from the root of the blade to the trailing edge. The impact of the rotor wake-stator
interaction on the acoustics was also predicted analytically. The Active Noise Control Fan, located at the NASA Glenn
Research Center, was used as the proof-of-concept test bed. In-duct mode and farfield directivity acoustic data were
acquired at blowing rates (defined as mass supplied to trailing edge blowing system divided by fan mass flow) ranging
from 0.5 to 2.0 percent. The first three blade passing frequency harmonics at fan rotational speeds of 1700 to 1900 rpm
were analyzed. The acoustic tone power levels (PWL) in the inlet and exhaust were reduced 11.5 and 0.1, 7.2 and 11.4,11.8 and 19.4 PWL dB, respectively. The farfield tone power levels at the first three harmonics were reduced 5.4, 10.6,
and 12.4 dB PWL. At selected conditions, two-component hotwire and stator vane unsteady surface pressures were
acquired. These measurements illustrate the physics behind the noise reduction.
http://gltrs.grc.nasa.gov/GLTRShttp://gltrs.grc.nasa.gov/GLTRShttp://gltrs.grc.nasa.gov/GLTRS