PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTP

David Mundaye-mail: david.munday@uc.edu

Nick Heebe-mail: heebns@mail.uc.edu

Ephraim Gutmarke-mail: ephraim.gutmark@uc.edu

University of Cincinnati,Cincinnati, OH 45221

Junhui Liue-mail: jhliu@lcp.nrl.navy.mil

K. Kailasanathe-mail: kailas@lcp.nrl.navy.mil

Naval Research Laboratory,Washington, DC 20375

Supersonic Jet Noise ReductionTechnologies for Gas TurbineEnginesThis paper presents observations and simulations of the impact of several technologieson modifying the flow-field and acoustic emissions from supersonic jets from nozzlestypical of those used on military aircraft. The flow-field is measured experimentally byshadowgraph and particle image velocimetry. The acoustics are characterized by near-and far-field microphone measurements. The flow- and near-field pressures are simulatedby a monotonically integrated large eddy simulation. Use of unstructured grids allowsaccurate modeling of the nozzle geometry. The emphasis of the work is on “off-design”or nonideally expanded flow conditions. The technologies applied to these nozzles includechevrons, fluidic injection, and fluidically enhanced chevrons. The fluidic injection geom-etry and the fluidic enhancement geometry follow the approach found successful forsubsonic jets by employing jets pitched 60 deg into the flow, impinging on the shear layerjust past the tips of the chevrons or in the same axial position when injection is withoutchevrons. �DOI: 10.1115/1.4002914�

1 IntroductionThere is a growing need to significantly reduce the noise gen-

erated by high-performance, supersonic military aircraft. Thenoise generated during take-off and landing on aircraft carriers hasdirect impact on shipboard health and safety issues. Noise com-plaints are increasing, as communities move closer to militarybases or when there are changes due to base closures or realign-ment. Furthermore, U.S. and international noise regulations andpolicies will have an impact on operations and training unlesseffective steps are taken to reduce the noise.

1.1 Technical Background. There is a significant amount ofliterature dealing with noise reduction in civilian, subsonic air-craft. Some of the techniques found effective in that regime couldpossibly be applied for noise reduction in supersonic military jets.Many of these techniques use flow modifiers such as mechanicalchevrons to enhance the mixing of the jet with the surroundings toreduce the jet noise. A distinct difference between current civilianaircraft engines and military aircraft engines is that military en-gines tend to have low bypass ratios and high velocities and theirnoise tends to be dominated by jet noise, especially shock-associated noise. This is because during flight near the ground orby an aircraft carrier such as during take-off or landing, the ex-haust from the engines tends to be nonideally expanded. Hence,the current research is focused on these flow conditions.

Although there are still fundamental questions about the sourceand mechanisms of noise in supersonic jets, significant progresshas been made over the past few decades �1–10�. From previousstudies, it is known that the noise generated by an imperfectlyexpanded supersonic jet consists of discrete and high amplitudescreech tones, broadband shock-associated noise �BBSN� compo-nents, and mixing noise. The first two sources are related to theshock waves that are present in high-speed jet flow. While themixing noise dominates in the downstream direction, the shock-associated noise elevates the overall noise level in the upstreamdirection. The screech tones are thought to arise due to a feedbackloop involving the large scale flow structures shed from the nozzle

lip, their interactions with the shock cell structure producingacoustic waves, and the upwind propagating acoustic waves in-ducing more large scale structures at the nozzle lip.

Large eddy simulations �LES� have gained prominence as a toolto investigate the flow-field and noise from supersonic jets�11–23�. Typically, the near-field is computed using LES, and thefar-field noise is estimated using the information from LES alongwith Lighthill’s acoustic analogy �24�, or the Ffowcs Williams–Hawkins method �25� or the Kirchhof method.

1.2 Combined Experimental/Numerical Approach. Thetechnical approach employed is to design and fabricate nozzlesrepresentative of those used in military engines, conduct experi-ments and acquire data, compare this information to validate nu-merical simulations, and then use both experiments and simula-tions together to understand the sources of the jet noise andinvestigate specific methods to reduce it.

For this application, experimental measurement and numericalsimulation are particularly complementary. Measurement of pres-sure fluctuations within the jet itself is impossible by any nonin-trusive technique. Since the introduction of any probe will intro-duce new shocks into the jet and since shocks contributesubstantially to noise generation, the introduction of pressureprobes is quite a substantial intrusion. One can employ measure-ment techniques to learn a great deal about the acoustical behavioroutside the jet, but what can be measured inside the jet is limited.Within the jet, capturing the shock structures with numericalsimulations requires very high grid densities. The numerical costscales with the number of computational cells, so this makes itprohibitively expensive to cover a large domain. The presentproject, therefore, employs simulations to learn about the noisegeneration within the jet but limits in the domain simulated.Within the jet, we can make good nonintrusive measurements ofvelocity with particle image velocimetry �PIV�, allowing for wellfounded validation so one can have great confidence in the com-bined approach. A number of cases are run experimentally, andselected cases are examined in greater detail with validated LES.

Laboratory experiments are being conducted at the Universityof Cincinnati in Cincinnati, OH and numerical simulations at theNaval Research Laboratory in Washington, DC. Design help andoverall guidance to ensure relevance to field implementation inthe future is being provided by the General Electric Global Re-search �Niskayuna, NY�.

Contributed by the International Gas Turbine Institute �IGTI� of ASME for pub-lication in the JOURNAL OF ENGINEERING FOR GAS TURBINES AND POWER. Manuscriptreceived May 12, 2010; final manuscript received October 6, 2010; published onlinexxxxx-xxxxx-xxxxx. Editor: Dilip R. Ballal.

1

23456789

10

1112131415161718192021222324252627282930313233343536

APC:#1

373839404142434445

464748495051525354555657585960616263646566676869707172737475767778

APC:#2

APC:#3

Journal of Engineering for Gas Turbines and Power MAY 2011, Vol. 133 / 1-1Copyright © 2011 by ASME

PROOF COPY [GTP-10-1329] 005103GTP

PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTP

1.3 Scope of Research. This paper reports the progress todate on a project aimed at characterizing the source of the noisefrom military aircraft jets and investigating several of the promis-ing techniques to reduce it. The baseline nozzle and part of thechevron results were reported at last year’s ASME Turbo Expo�26�. This paper completes the chevron examination and explorestwo additional techniques, namely, fluidic injection in whichblowing from microjets replaces chevrons, and fluidically en-hanced chevrons in which microjets blow in addition to chevrons.

2 Experimental ApproachModels representing practical military aircraft engine exhaust

nozzles have been constructed for several design Mach numbers,Md. Md is a function only of area ratio. These nozzles have aconic contraction section, a sharp throat, and a conic expansionsection. These conical convergent-divergent �C-D� nozzles are areasonable approximation to the faceted nozzles on military air-craft since there is no significant circumferential flow within thenozzle. Unless otherwise noted, all cases presented in this paperhave a design Mach number of Md=1.50. The Md=1.5 nozzle isshown in Fig. 1. For this nozzle, the inflow diameter Di is3.124 in., the throat diameter Dt is 2.640 in. �67.06 mm�, andnozzle exit diameter De, or simply D, is 2.868 in. �72.85 mm�.The nozzle lip is very thin, with a thickness of 0.020 in. �0.61mm�. This particular nozzle geometry is chosen to be representa-tive of practical engine nozzles.

2.1 Flow Control Arrangements. A chevron cap has beenmachined to fit over the outside of the nozzle such that the innercontour of the chevron matches the inner contour of the nozzle atthe exit. Twelve chevrons extend 0.6 in. �15.24 mm� from thenozzle trailing edge with a half angle of 35 deg. The inner andouter curves of the chevrons are circular arcs, with the inner arcbeing tangent to the nozzle inner surface. There is a small gapbetween the chevrons approximating the gap that would be left bythe opening of the seals of the actual variable-geometry nozzle.The chevron geometry is shown in Fig. 2.

For the fluidic injection cases with and without chevrons, atoroidal plenum is fitted around the model with 12 0.125 in. �3.18mm� i.d. tubes extending to inject one microjet just past the tip ofeach chevron at an angle of 60 deg to the main jet. This arrange-ment follows Alkislar et al. �27�. The jets are maintained at apressure ratio of 3 for the experiments and 4 for the LES. A crosssection view of the fluidically enhanced chevron geometry is inFig. 3. The fluidic injection geometry is the same but with thechevron cap removed.

2.2 Instrumentation. The high-gradient features in the floware visualized by the shadowgraph technique. An Oriel 66056 arclamp is used for illumination. A pair of 12 in. �305 mm� parabolic

first-surface mirrors with a 72 in. �1.829 m� focal length are em-ployed to collimate the light before the model and then to refocusthe beam after. The image is captured with a LaVision imagerintense cross-correlation charge-coupled device �CCD� camerawith 1376�1040 pixel resolution and 12-bit intensity resolution.This gives a spatial resolution on the order of 0.01 in. �0.25 mm�or 0.004 throat diameters. A 28–300 mm zoom lens is mounted tothe camera, which allows optimization of the field of view. Theaperture is left completely open, and the exposure is controlled bymounting neutral density filters. Averaging 100 images eliminatesthe turbulence and gives a clear view of the shock and Prandtl–Meyer waves.

Detailed flow-field mapping is performed by PIV. The PIV sys-tem is built by LaVision, and the entire PIV suite �laser and cam-eras� is mounted on a traverse, which allows the system to betranslated undisturbed to any streamwise location, allowing manyfields of view to be measured without the loss of time in changingsetups and without the uncertainties, which come from repeatedadjustment of components. The flow is seeded with olive oil drop-lets with diameters on the order of 1 �m. A 500 mJ New WaveResearch neodymium doped yttrium aluminum garnet �Nd:YAG�double-pulse laser is passed through sheet-forming optics to illu-minate the seed, and the images are captured by a pair of LaVisionCCD cameras with 1376�1040 pixel resolution and 12-bit inten-sity resolution.

The model is mounted in the University of Cincinnati AcousticTest Facility �UC-ATF�, which is a 24�25�11 ft3 �7.3�7.6�3.4 m� test chamber that has been acoustically treated to beanechoic down to 500 Hz. Eight quarter-inch microphones arearrayed along an arc at angles ranging from �=35 deg to 150deg. � is measured from the upstream direction. The microphonesare placed at 135 in. �3.43 m� or 47 exit diameters from thenozzle exit. The facility was described in detail by Callender et al.

Fig. 1 The nozzle geometry for the baseline nozzle, with Md=1.50

Fig. 2 Geometry of the chevron cap, which fits over the nozzle

Fig. 3 Fluidic injection geometry „cross section view…,0.125 in. i.d. tube

798081828384858687

88

8990919293949596979899

100101102103

104105106107108109110111112113114115116117118119120121122

123124125

APC:#4

126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158

APC:#5

APC:#6

APC:#7APC:#8

APC:#9

APC:#10

1-2 / Vol. 133, MAY 2011 Transactions of the ASME

PROOF COPY [GTP-10-1329] 005103GTP

PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTP

�28�. The microphone signals are filtered above 100 kHz and re-corded at 200 kHz for 10 s. The flow variables are monitored at 4Hz in order to establish and hold the desired operating conditionof the model. Variables recorded include the ambient conditions inthe chamber, as well as the flow stagnation temperature andpressure.

In order to map the near-field acoustics, 32 microphones aremounted on a linear rake that is mounted on a large three-axistraverse. The rake is angled to keep the microphones out of theflow and is then traversed radially outward. The resulting grid ofpressure fluctuations can then be analyzed on an over-all soundpressure level �OASPL� basis or for particular frequencies perti-nent to the different mechanisms and can be mapped to look at thesource location and the direction of propagation.

3 Computational ApproachThe unsteady three-dimensional inviscid compressible flow

equations are solved with a finite element flow code FEFLO usingunstructured tetrahedral grids �29�. This code is capable of accu-rately representing complicated geometries such as the nozzle ge-ometries used in this work. No explicit subgrid scale model isused, and the modeling of subgrid scales is implicitly provided bythe embedded flux limiter. The present simulations are in theframework of monotonically integrated large eddy simulations�MILES� �30�. The finite element �FEM� version of flux-correctedtransport �FCT� algorithm �FEM-FCT� is used for the spatial dis-cretization, and a second order Taylor–Galerkin scheme is usedfor the time integration. FCT is ideal for simulating the shockcontaining flows because it is high-order, conservative, monotone,positivity-preserving �31�, and has previously been used to simu-late supersonic jet noise �32�.

3.1 Computational Domain and Boundary Conditions.The nozzle geometry used in the simulations is shown in Fig. 1.The computational domain is outlined in Fig. 4. Fine grids areclustered around both the nozzle and the jet wake. The area withthe fine grid is divided into two regions. The inner region is themost refined area, which covers the core of the jet flow to capturethe energy-containing turbulence scales. This region has a cell sizeof 0.0345D �D is the exit diameter�, and its length extends to 24Ddownstream in the axial direction. The radius of this region is1.4D near the nozzle exit, and it gradually increases to 1.9D at theend of the region. Since the nozzle lip is very thin, the cell sizenear the nozzle exit is further reduced to accommodate the lipthickness. However, only one element is used around the lip toavoid a time-step size that is too small. The cell size inside thenozzle geometry increases from 0.025 in. �0.635 mm� to 0.15 in.�3.81 mm� as it approaches the inflow boundary. Since the propa-gation of sound waves in the frequency range of interest allows alittle coarser cell size, the cell size in the outer region increases to0.065D. This region extends to 3D in the radial direction, 5D inthe upstream direction, and 24D downstream. Very coarse cell

sizes are used in the far-field and near outflow boundaries to avoidwave reflections from these boundaries. The overall domain sizeis 15D in the radial direction, 17D in the upstream direction of thenozzle exit, and 47D downstream. The mesh has roughly 11�106 grid points and 65�106 tetrahedral elements.

Characteristic boundary conditions are applied to both the far-field and outflow boundaries �with pressure relaxation�. The totalpressure is kept constant at the inlet of the nozzle. A slip boundarycondition is used for all solid surfaces.

Once a mature flow-field is established, the time-step size iskept constant and the data are collected at small time intervals.Since the number of grid points is large, it is not realistic to saveall the data at all points with a small time interval. Instead, thedata at points on a Cartesian mesh are saved after every 20 timesteps. These points essentially serve as nonintrusive numericalprobes in the flow-field. The distances between the neighboringprobes are 0.2D in both the axial and radial directions. More de-tails on the methodology and verification and validation simula-tions have already been presented elsewhere �33�. This paper ex-tends the investigation to include noise reduction techniques,namely, mechanical chevrons and fluidic injection.

4 Results and DiscussionThe flow through the nozzle is characterized by the fully ex-

panded Mach number, Mj, which is a function only of the nozzlepressure ratio �NPR� between the nozzle’s total pressure and theambient static pressure. The nozzle is at its design condition whenMj=Md. The off-design operation of these nozzles is of particularinterest, so overexpanded conditions Mj=1.22, 1.36, and 1.47 andunderexpanded conditions Mj=1.56, 1.64, and 1.71 will be exam-ined. All cases are unheated. All conditions are tested on the base-line, chevron, fluidic, and fluidically enhanced chevron configura-tions. Selected cases are presented below. The emphasis of thispaper is on the acoustic emissions.

4.1 Dynamics of Jets From Practical Military Nozzles.Mach contours from LES are shown in Fig. 5. This figure illus-trates the features typical of jets from practical military nozzles.The sharp throat generates a shock wave inside the nozzle. For therange of design Mach numbers explored in this project �Md=1.3−1.65�, this throat wave always forms a Mach disk, which sheds aslip line that can be observed leaving the nozzle. The diameter ofthe internal Mach disk and the slip line are dependent on Mj.Depending on the nozzle length and the Md, the onward wavefrom the Mach disk may reflect inside the nozzle or may emergefrom the nozzle and reflect from the shear layer outside the nozzleas it does for the Md=1.5 nozzle shown in Fig. 5. The reflectedwave sets up a shock diamond independent of the shock diamondsshed from the nozzle lip, producing a double-diamond pattern,which coalesces into a single diamond pattern after two or three

Fig. 4 The computational domain and boundary conditions

Fig. 5 Mach number contours of baseline nozzle, Mj=1.56159160161162163164165166167168169170171172

173

174175176177178179180181182183184185186187188

189190191192193194195196197198199200201202203204205206207208

209210211212213214215216217218219220221222223224225226227228229

230

231232233234235236237238239240241

242243244245246247248249250251252253254255256

Journal of Engineering for Gas Turbines and Power MAY 2011, Vol. 133 / 1-3

PROOF COPY [GTP-10-1329] 005103GTP

PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTP

reflections.Since the expansion section of these nozzles is conical, the flow

leaving the nozzles cannot be a parallel flow. This means that evenif the nozzle is operated with a perfect pressure match across theexiting shear layer, there will still be turning of the flow. This willproduce a shock wave even at the design condition. The details ofthe baseline flow have been more fully discussed elsewhere �34�.Validation of LES for this configuration has been presented inRef. �33�.

4.2 Flow Modification Due to Chevrons. The application ofchevrons produces a large convolution of the shear layer. Chev-rons force negative radial velocity behind the lobes and inducepositive radial velocity between the lobes. This convolution per-sists several diameters downstream of the exit. Figure 6 shows thecharacter of the convolution. The potential core of the jet is re-duced in diameter, and the sonic surface is also convoluted andmade radially smaller. The radially smaller supersonic region re-duces the shock cell spacing, which, in turn, shifts the peakfrequencies.

The shift in shock cell size can be readily observed in the av-eraged shadowgraph images in Fig. 7. Each pair has the baselinenozzle on the upper half and the chevron nozzle on the lower half.The shift in the shock cell size is most apparent in the lowervalues of Mj where the additional shock structures introduced bythe chevrons are less pronounced. At the higher values of Mj, itbecomes more difficult to discriminate the comparable lines in theimage.

In examining the baseline images �the upper half of each com-posite image�, it can be seen that the angle of the outer edge of thejet changes. For the most overexpanded case, the jet contracts sothe edge turns radially inward. This implies a radially inwardcomponent of velocity at the edge, and so the local velocity vectorat the edge of the jet points inward. For increasing values of Mj asmoving through perfect expansion and into underexpansion, onecan see the edge of the jet moving outward, implying an outwardpointing velocity vector at the edge of the jet, but as the velocityvector changes direction, the angle of the chevrons remains un-changed. This implies a change in the effective penetration of thechevrons. At the lowest values of Mj, the chevrons barely pen-etrate into the flow, and the flow is relatively undisturbed by thechevrons. As Mj increases, they penetrate more and more deeplyinto the undisturbed flow so their influence becomes morepronounced.

The shadowgraph technique is line integrating, so what is seenon the chevron images is the superposition of the flow between

chevrons, the flow from the chevron tips, and all points in be-tween. Two bright lines can be seen emerging from the edge of thenozzle beginning with the case of Mj=1.47. One inclines outwardand the other inclines inward. These are indications of the edge ofthe jet as it flows between the chevrons and from the chevron tip,respectively. These become more and more separated as Mj in-creases because the effective penetration grows and the chevronsproduce greater and greater convolution of the jet shear layer.

LES data can be plotted in arbitrary planes within the flow,avoiding the complication of line integration. The chevron tip andvalley planes are plotted separately in Fig. 8. Like in Fig. 7, theimages are composites with baseline on the top and chevrons onthe bottom. The lower half of the upper image is a plane bisectingthe valley between two chevrons, while the lower half of the

Fig. 6 PIV of velocity modification due to chevrons: Mj=1.46,z /D=2. The black circle has a diameter equal to the nozzle exit.The missing data regions are due to hardware reflections con-taminating the PIV images of the seeded flow.

Fig. 7 Shadowgraph images with and without chevrons

257258259260261262263264265

266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301

302303304305306307308309310311312313314315

1-4 / Vol. 133, MAY 2011 Transactions of the ASME

PROOF COPY [GTP-10-1329] 005103GTP

PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTP

lower image is a plane through a chevron tip. Notice in the chev-ron valley image that the outer edge of the shear layer is visiblealong the same line as the outer line on the shadowgraph and thatthe bright inner line on the shadowgraph matches the shear layeredge seen in the tip plane from LES.

The influence of fluidic injection is similar to chevrons, produc-ing convolution of the shear layer as chevrons do. Figure 9 showsa comparison of the LES of both cases. In both cases, pronouncedconvolution and substantial streamwise vorticity is induced.

4.3 Acoustic Influence of Chevrons. The far-field acousticcharacter of the baseline nozzle is similar to a smoothly contouredlaboratory nozzle except for the presence of shock-associatednoise components being present at the design condition. The ob-server angle � is defined as the angle from the upstream axis ofthe jet. The spectra from �=35 deg are shown in Fig. 10�a�. Thisis the direction in which screech and BBSN propagate moststrongly. Chevrons completely eliminate screech for the overex-panded cases and significantly reduce it for the perfectly expandedand underexpanded cases. This is due to chevrons breaking thehelical mode of the jet’s oscillation. The reduction in the chev-ron’s effectiveness at lower values of Mj is due to the reducedeffective penetration of chevrons as the flow turns increasinglyinward, while the angle of the chevrons remains the same. Thereis a similar trend with respect to the BBSN. The amplitudes of theBBSN are reduced, and the peak frequency is shifted to highervalues by the shortening of the shock cells. Chevrons do increasethe noise at higher frequencies and lower values of Mj, as ob-served in the forward quadrant.

The far-field spectra from the aft quadrant, at �=150 deg, areshown in Fig. 10�b�. This is in the direction in which mixing noiseis the dominant component of jet noise. Here one can see thatchevrons reduce mixing noise as they do for subsonic jets. Thebroadband reduction in mixing noise is offset at higher frequen-cies by the noise produced near chevrons by the small-scale tur-bulence they generate. This effect has been previously observed insubsonic jets with chevrons, and it is believed that the mechanismis the same here. Both effects are greatest at the largest Mj and arenegligible at the lowest value of Mj. This is due to both the re-duction in effective penetration at low Mj and the reduction inshear velocity.

4.4 Acoustic Influence of Fluidic Injection. Chevrons haveshown their value as a jet noise reduction tool in a number ofapplications, but they have a distinct disadvantage in that theyimpose a small thrust loss, which must be carried throughout theflight, even in parts of the flight when noise production is ofreduced importance. An alternative approach is to induce stream-wise vortices into the shear layer by fluidic injection or microjetblowing. The blowing has a cost in terms of bleed air, but it hasthe advantage that it can be turned on for noise sensitive portionsof the flight and turned off the rest of the time, reducing theperformance impact on the overall flight.

Figure 11 shows the far-field spectra for a configuration withblowing into the shear layer at the same location the chevron tipswould be. In this case, the fluidic injection jets are pitched into themain jet at an angle of 60 deg. The mass flow in this case is thesame for all values of Mj and ranges from 1.1% of the main jetmass flow for Mj=1.71 up to 2.1% for Mj=1.22. Qualitatively, thesame kinds of influence as with chevrons are shown. Screechpeaks are reduced, BBSN peaks are reduced and shifted to higherfrequencies, and mixing noise is reduced. The opposite trend,however, is apparent in terms of the influence of Mj. For thefluidic injection, the greatest effect is at the low values of Mjwhere the mass-flow ratio is the highest, and there is a reducedeffect as Mj increases and the mass-flow ratio reduces.

4.5 Acoustics of Fluidic Enhancement of Chevrons. Thetrends in terms of sensitivity to Mj are complementary betweenchevrons and fluidic injection. This suggests that combining thetwo approaches might be useful. Figure 12 shows spectra for flu-idically enhanced chevrons in which the blowing is applied at 60deg to the main jet at the tips of the chevrons. The mass flows arethe same as in the fluidic injection case. Combining the two ap-proaches does show some benefit in reducing the screech at thelower values of Mj where chevrons are least effective, but for themidvalues near the design condition, blowing increases the BBSN

Fig. 8 Density cross-cuts from LES: The upper image com-pares the nozzle without chevrons „top… with a plane throughthe valley between chevrons „bottom…. The lower image com-pares the nozzle without chevrons „top… with a plane throughthe chevron tips. Mj=1.56 for both images.

Fig. 9 LES of chevrons and fluidic injection, Mj=1.56, x/D=1.0. Injection tube pressure ratio was 4. Velocities are normal-ized by Uj. The top half is fluidic injection; the bottom half ischevrons.

316317318319320321322323324

325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355

356357358359360361362363364365366367368369370371372373374375376377378379

380381382383384385386387388389

Journal of Engineering for Gas Turbines and Power MAY 2011, Vol. 133 / 1-5

PROOF COPY [GTP-10-1329] 005103GTP

PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTPand produces a small increase in noise at high frequencies. At aft

angles shown in Fig. 12�b�, one can see that for the lower valuesof Mj, there is suppression of the screech tone and small broad-band reduction at lower frequencies. For the midvalues of Mj,there is some increased noise when blowing is applied to chev-rons.

5 Discussion and Conclusions

Three noise reduction technologies have been examined experi-mentally and numerically as they have been applied to overex-panded, perfectly expanded, and underexpanded supersonic jetsfrom convergent-divergent nozzles representative of those on high

Fig. 10 „a… Far-field spectra of baseline and chevron configurations for an observer at�=35 deg in the upstream quadrant. „b… Far-field spectra of baseline and chevron con-figurations for an observer at �=150 deg in the downstream quadrant.

390391392393394395

396

397398399400

1-6 / Vol. 133, MAY 2011 Transactions of the ASME

PROOF COPY [GTP-10-1329] 005103GTP

PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTPperformance military aircraft.

Chevrons have been shown to reduce or eliminate screech andreduce broadband shock-associated noise and mixing noise. Mix-ing noise and BBSN are known to be the most strongly generatednear the end of the potential core and the end of the shock train,respectively. For the case of Mj=1.56, this happens in the neigh-borhood of ten diameters. LES computation of pressure spectra is

shown in Fig. 13. This plot shows that at x/D�10.8, near the peaksource region chevrons produce a significant reduction in near-field pressure fluctuations below 9 kHz. This includes the frequen-cies at which BBSN peaks are observed in the far-field.

Figure 14 shows the instantaneous contours of pressure fromLES Mj=1.56. The effect of chevrons is quite striking. Beyond

Fig. 11 „a… Far-field spectra of baseline and fluidic injection configurations for an ob-server at �=35 deg in the upstream quadrant. „b… Far-field spectra of baseline and flu-idic injection configurations for an observer at �=150 deg in the downstream quadrant.

401402403404405406407

408409410411412413

Journal of Engineering for Gas Turbines and Power MAY 2011, Vol. 133 / 1-7

PROOF COPY [GTP-10-1329] 005103GTP

PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTPx /D=6 in the region where BBSN production is strongest the

shocks are becoming weak, the pressure fluctuations are domi-nated by the convecting large-scale structures. The chevrons sub-stantially reduce the size of these structures. Since the BBSN isproduced by the interaction between the large scale structures andthe shocks, reducing the size of the structures will reduce BBSNproduction. Chevrons also shift the shock-associated noise peaks

to higher frequency by reducing the diameter of the supersonicregion and, by this, the length of the shock cells.

Chevrons increase noise at higher frequencies due to an intensesmall-scale turbulence generation near the nozzle. For the appli-cation of chevrons to this baseline nozzle, the lobe of increasednoise near the nozzle has previously been shown to be centered atx /D=2 �35�. Figure 13 shows the near-field spectra in this region.

Fig. 12 „a… Far-field spectra of chevron and fluidically enhanced chevron configurationsfor an observer at �=35 deg in the upstream quadrant. „b… Far-field spectra of chevronand fluidically enhanced chevron configurations for an observer at �=150 deg in thedownstream quadrant.

414415416417418419420

421422423424425426427

1-8 / Vol. 133, MAY 2011 Transactions of the ASME

PROOF COPY [GTP-10-1329] 005103GTP

PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTP

This shows that chevrons increase near-field fluctuations above 10kHz, verifying that the high-frequency penalty imposed by chev-rons comes from the near-nozzle region.

Fluidic injection has been found to reduce screech, broadbandshock-associated noise, and mixing noise. It also shifts the shock-associated noise peaks to higher frequency and generates in-creased high-frequency noise as chevrons do. Examination of Fig.

13 reveals that fluidic injection produces the same reduction nearx /D=10 at midfrequencies and the same increase in high frequen-cies near the nozzle as chevrons. In fact, the spectra for chevronsand for fluidic injection are strikingly similar. The instantaneouspressure contours shown in Fig. 14 are also very similar betweenchevrons in the upper plot and fluidic injection in the lower plot.Both noise reduction techniques reduce the size of the large-scalestructures beyond x /D=6 and so both reduce BBSN by the samemechanism. The principal difference between chevrons and fluidicinjection is that for constant injection mass flow, the effectivenessof fluidic injection increases with decreasing values of Mj, whilefor chevrons, the trend goes the other way.

The application of blowing at the tips of the chevrons in aneffort to enhance chevrons shows modest additional noise reduc-tion in the form of screech suppression but shows some noiseincrease for a nozzle operating near the design condition.

AcknowledgmentThis work has been sponsored by the Strategic Environmental

Research and Development Program �SERDP� and the NRL 6.1Computational Physics Task Area. The authors also gratefullythank Prof. Rainald Lohner from the George Mason University forhis significant help with the computational code. Finally the au-thors would like to thank Dr. Steve Martens from GE GlobalResearch for his guidance in model and case selection.

NomenclatureC-D � convergent-divergent

D � exit diameterDe � exit diameterDi � inflow diameterDt � throat diameter

Md � design Mach numberMj � fully expanded Mach number

NPR � nozzle pressure ratio� � angle from the upstream jet axis

References�1� Tam, C. K. W., and Tanna, H. K., 1982, “Shock-Associated Noise of Super-

sonic Jets From Convergent-Divergent Nozzles,” J. Sound Vib., 81�3�, pp.337–358.

�2� Norum, T. D., and Seiner, J. M., 1982, “Broadband Shock Noise From Super-sonic Jets,” AIAA J., 20�1�, pp. 68–73.

�3� Seiner, J. M., 1984, “Advances in High Speed Jet Aeroacoustics,” AIAA PaperNo. 84-2275.

�4� Tam, C. K. W., 1991, “Jet Noise Generated by Large Scale Coherent Motion,”Aeroacoustics of Flight Vehicles: Theory and Practice, Vol. 1: Noise Sources,H. H. Hubbard, ed., NASA Langley Research Center, Hampton, VA, pp. 311–390.

�5� Tam, C. K. W., 1995, “Supersonic Jet Noise,” Annu. Rev. Fluid Mech., 27, pp.17–43.

�6� Krothapalli, A., Soderman, P. T., Allen, C. S., Hayes, J. A., and Jaeger, S. M.,1997, “Flight Effects on the Far-Field Noise of a Heated Supersonic Jet,”AIAA J., 35�6�, pp. 952–957.

�7� Raman, G., 1998, “Advances in Understanding Supersonic Jet Screech: Re-view and Perspective,” Prog. Aerosp. Sci., 34, pp. 45–106.

�8� Umeda, Y., and Ishii, R., 2001, “On the Sound Sources of Screech TonesRadiated From Choked Circular Jets,” J. Acoust. Soc. Am., 110�4�, pp. 1845–1858.

�9� Callender, B., Gutmark, E., and Martens, S., 2004, “Far-Field Acoustic Inves-tigation Into Chevron Nozzle Mechanisms and Trends,” AIAA J., 43�1�, pp.87–95.

�10� Panda, J., 2005, “Identification of Noise Sources in High Speed Jets via Cor-relation Measurements—A Review,” AIAA Paper No. 2005-2844.

�11� DeBonis, J. R., and Scott, J. N., 2001, “A Large-Eddy Simulation of a Turbu-lent Compressible Round Jet,” AIAA Paper No. 2001-2254.

�12� Shur, M. L., Spalart, P. R., Strelets, M., and Travin, A. K., 2003, “Towards thePrediction of Noise From Jet Engines,” Int. J. Heat Fluid Flow, 24, pp. 551–561.

�13� Uzun, A., Lyrintzis, A. S., and Blaisdell, G. B., 2004, “Coupling of IntegralAcoustics Methods With LES for Jet Noise Prediction,” Int. J. Aeroacoust.,3�4�, pp. 297–346.

�14� Kenzakowski, D. C., Kannepalli, C., and Brinckman, K. W., 2004, “Compu-tational Studies Supporting Concepts for Supersonic Jet Noise Reduction,”AIAA Paper No. 2004-0518.

Fig. 13 Near-field spectra from LES for baseline, chevrons,and fluidic injection. Mj=1.56. Upper figure x/D=10.8, r /D=2.2.Lower figure x/D=1, r /D=1.

Fig. 14 Instantaneous pressure from LES for Mj=1.56. Upperfigure compares chevrons to the baseline nozzle. Lower figurecompares fluidic injection to baseline.

428429430431432433434

435436437438439440441442443444445446447448449450

451

452453454455456457458

459460461462463464465466467468

469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506

Journal of Engineering for Gas Turbines and Power MAY 2011, Vol. 133 / 1-9

PROOF COPY [GTP-10-1329] 005103GTP

PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTP

�15� Bodony, D., and Lele, S., 2006, “Review of the Current Status of Jet NoisePredictions Using Large Eddy Simulations,” AIAA Paper No. pp. 2006-486.

�16� Shur, M. L., Spalart, P. R., Strelets, M. K., and Garbaruk, A. V., 2006, “FurtherSteps in LES-Based Noise Prediction for Complex Jets,” AIAA Paper No.2006-485.

�17� Berland, J., Bogey, C., and Bailly, C., 2006, “Large Eddy Simulation ofScreech Tone Generation in a Planar Underexpanded Jet,” AIAA Paper No.2006-2496.

�18� Viswanathan, K., Shur, M. L., Spalart, P. R., and Strelets, M. K., 2006, “Com-putation of the Flow and Noise of Round and Beveled Nozzles,” AIAA PaperNo. 2006-2445.

�19� Loh, C. Y., and Hultgren, L. S., 2006, “Jet Screech Noise Computation,”AIAA J., 44�5�, pp. 992–998.

�20� DeBonis, J. R., 2007, “Progress Toward Large-Eddy Simulations for Predic-tion of Realistic Nozzle Systems,” J. Propul. Power, 23�5�, pp. 971–980.

�21� Lo, S. C., Blaisdell, G. A., and Lyrintzis, A. S., 2008, “Numerical Simulationof Supersonic Jet Flows and Their Noise,” AIAA Paper No. 2008-2970.

�22� Eastwood, S., Tucker, P., and Xia, H., 2008, “Comparison of LES to LDA andPIV Measurements of a Small Scale High Speed Coflowing Jet,” AIAA PaperNo. 2008-2981.

�23� Liu, J., Kailasanath, K., Ramamurti, R., Munday, D., Gutmark, E., and Lohner,R., 2009, “Large-Eddy Simulations of a Supersonic Jet and Its Near-FieldAcoustic Properties,” AIAA J., 47, pp. 1849–1864.

�24� Lighthill, M. J., 1952, “On the Sound Generated Aerodynamically. I. GeneralTheory,” Proc. R. Soc. London, Ser. A, 211, pp. 564–587.

�25� Ffowcs Williams, J. E., and Hawkings, D. L., 1969, “Sound Generated byTurbulence and Surfaces in Unsteady Motion,” Philos. Trans. R. Soc. London,Ser. A, 264�1151�, pp. 321–342.

�26� Kailasanath, K., Liu, J., Gutmark, E., Munday, D., and Martens, S., 2009,“Impact of Mechanical Chevrons on Supersonic Jet Flow and Noise,” ASMEPaper No. GT2009-59307.

�27� Alkislar, M. B., Krothapalli, A., and Butler, G. W., 2007, “The Effect ofStreamwise Vortices on the Aeroacoustics of a Mach 0.9 Jet,” J. Fluid Mech.,578, pp. 139–169.

�28� Callender, B., Gutmark, E., and Dimicco, R., 2002, “The Design and Valida-tion of a Coaxial Nozzle Acoustic Test Facility,” AIAA Paper No. 2002-369.

�29� Löhner, R., Morgan, K., Vahdati, M., Boris, J. P., and Book, D. L., 1988,“FEM-FCT: Combining Unstructured Grids With High Resolution,” Commun.Appl. Numer. Methods, 4, pp. 717–729.

�30� Grinstein, F. F., and Fureby, C., 2005, “On Monotonically Integrated LargeEddy Simulation of Turbulent Flows Based on FCT Algorithms,” Flux-Corrected Transport: Principles, Algorithms, and Applications, D. Kuzmin, R.Lohner, and S. Turek, eds., Springer, New York, pp. 79–104.

�31� Boris, J. P., and Book, D. L., 1973, “Flux-Corrected Transport I. SHASTA, aFluid Transport Algorithm That Works,” J. Comput. Phys., 11, pp. 38–69.

�32� Kolbe, R., Kailasanath, K., Young, T. R., Boris, J. P., and Landsberg, A., 1996,“Numerical Simulations of Flow Modification of Supersonic RectangularJets,” AIAA J., 34�5�, pp. 902–908.

�33� Liu, J., Kailasanath, K., Ramamurti, R., Munday, D., Gutmark, E., and Lohner,R., 2009, “Large-Eddy Simulation of a Supersonic Jet and Its Near-FieldAcoustic Properties—Methodology and Validation,” 47th AIAA AerospaceSciences Meeting, Orlando, FL, AIAA Paper No. 2009-0500.

�34� Munday, D., Gutmark, E., Liu, J., and Kailasanath, K., 2009, “Flow Structureof Supersonic Jets From Conical C-D Nozzles,” 39th AIAA Fluid DynamicsConference, San Antonio, TX, Jun. 24, AIAA Paper No. 2009-4005.

�35� Munday, D., Heeb, N., Gutmark, E., Liu, J., and Kailasanath, K., 2009,“Acoustic Effect of Chevrons on Jets Exiting Conical C-D Nozzles,” 15thAIAA/CEAS Aeroacoustic Conference, Miami, FL, May 11, AIAA Paper No.2009-3128.

507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537

538539540541542543544545546547548549550551552553554555556557558559560561562563564565

APC:#11

1-10 / Vol. 133, MAY 2011 Transactions of the ASME

PROOF COPY [GTP-10-1329] 005103GTP

PROOF COPY [GTP-10-1329] 005103GTP

PROO

F COPY [G

TP-10-1329] 005103GTP

AUTHOR QUERIES — 005103GTP

#1 Au: All references must be cited in text and innumerical order. Please check our insertionof Ref. 1 cite here.

#2 Au: Please check our insertions in thissentence.

#3 Au: Please provide the location of “GeneralElectric Global Research.”

#4 Au: Please supply metric conversion forinstances where “in.” is used.

#5 Au: Please spell out CCD.

#6 Au: Please check our insertion of “mm” in thissentence.

#7 Au: Please check grammar changes and seeif intended meaning is preserved.

#8 Au: Please provide the location of “LaVision.”#9 Au: Please provide the location of “New

Wave Research.”#10 Au: Please provide metric conversion for

instances where “ft” is used.#11 Au: Please verify the page range in Ref. 32.

NOT FOR PRINT! FOR REVIEW BY AUTHOR NOT FOR PRINT!

PROOF COPY [GTP-10-1329] 005103GTP