Mixing Process in a Lobed Jet Flohuhui/paper/journal/2002-aiaa-SPIV.pdfreduce takeoff jet noise and speci” c fuel consumption.1;2Morere-cently, lobed mixers/nozzles have also emerged

AIAA JOURNAL

Vol 40 No 7 July 2002

Mixing Process in a Lobed Jet Flow

Hui Hucurren Tetsuo Sagadagger Toshio KobayashiDagger and Nobuyuki Taniguchisect

University of Tokyo Tokyo 153-8505 Japan

A high-resolution stereoscopic particle image velocimetry (PIV) system was used to conduct three-dimensionalmeasurement of an air jet exhausted from a lobed nozzle The characteristics of the mixing process in the lobedjet ow are revealed clearly and quantitatively from the stereoscopic PIV measurement results The instantaneousvelocity and streamwise vorticity distributions revealed that the large-scale streamwise vortices generated bythe lobed nozzle break into smaller but not weaker streamwise vortices as they travel downstream This is theproposed reason why a lobed nozzle would enhance both large-scale mixing and small-scale mixing reported byother researchers The overall effect of the lobed nozzle on the mixing process was evaluated by analyzing theensemble-averaged streamwise vorticity distributions The strength of the ensemble-averaged streamwise vorticeswas found to decay very rapidlyover the rst two diameters ( rst six lobeheights) then to decay at a more moderaterate farther downstream The averaged turbulent kinetic energy pro le also indicated that most of the intensivemixing between the core jet and ambient ow occurred within the rst two diameters These results indicate thatthe maximum effectiveness region for the lobed nozzle to enhance mixing is about the rst two diameters of thelobed nozzle ( rst six lobe heights)

NomenclatureD = diameter of the lobed nozzle 40 mmH = lobe height 15 mmK = turbulent kinetic energy 1=2U 2

0 u 0 2 C v0 2 C w0 2NK z = averaged-turbulentkinetic enegy

Rfrac12W x y zK x y z dx dyR

frac12W x y z dx dy

Re = Reynolds numberU V W = velocity three componentsU0 = mean velocity of the air jet at the inlet of the

lobed nozzle 200 msu 0 v 0 w0 = rms values of velocity uctuationsx y z = coordinatesmicroin = inward penetration angle of the

lobed nozzle 22 degmicroout = outward penetration angle of the

lobed snozzle 14 degsup1 = viscosity of airfrac12 = density of air$z = streamwise vorticity D=U0v=x iexcl u=y

Introduction

L OBED mixersnozzles which are essentially splitter plateswith corrugatedtrailing edges are uid mechanic devicesused

to augment mixing in a variety of applications Such devices have

Received 27 February 2001 revision received 7 November 2001acceptedfor publication 18 December 2001 Copyright cdeg 2002 by the authors Pub-lished by the American Institute of Aeronautics and Astronautics Inc withpermission Copies of this paper may be made for personal or internal useon condition that the copier pay the $1000 per-copy fee to the CopyrightClearance Center Inc 222 Rosewood Drive Danvers MA 01923 includethe code 0001-145202 $1000 in correspondence with the CCC

currenJapan Society for the Promotion of Science Research Fellow Instituteof Industrial Science 4-6-1 Komaba Meguro-Ku currently PostdoctoralResearch Associate Turbulent Mixing and Unsteady Aerodynamics Lab-oratory A22 Research Complex Engineering Michigan State UniversityEast Lansing MI 48824 huhuiegrmsuedu Member AIAA

daggerResearch Associate Institute of Industrial Science 4-6-1 KomabaMeguro-Ku

DaggerProfessor Institute of Industrial Science 4-6-1 Komaba Meguro-KusectAssociate Professor Institute of Industrial Science 4-6-1 Komaba

Meguro-Ku

been applied widely in turbofan engine exhausts and ejectors toreduce takeoff jet noise and speci c fuel consumption12 More re-cently lobed mixersnozzles have also emerged as an attractive ap-proach for enhancing mixing between fuel and air in combustionchambers to improve the ef ciency of combustionand to reduce theformation of pollutants3

Besides the continuousefforts to optimize the geometry of lobedmixersnozzles for better mixing enhancement performance andwiden the applications of lobed mixersnozzles extensive studiesabout the mechanism by which the lobed mixersnozzles substan-tially enhancemixing have also been conductedin past yearsBasedon pressure temperature and velocity measurements of the ow- eld downstreamof a lobed nozzlePaterson4 revealedthe existenceof large-scalestreamwise vortices in lobed mixing ows inducedbythe special geometry of lobed nozzles The large-scale streamwisevortices were suggested to be responsible for the enhanced mixingWerle et al5 and Eckerle et al6 found that the streamwise vorticesin lobed mixing ows follow a three-step process by which thestreamwise vortices form intensify and then break down and sug-gested that the high turbulenceresulting from the vortex breakdownimproved the overall mixing process Elliott et al7 suggested thatthere are threeprimarycontributorsto the mixing processes in lobedmixing ows The rst is the spanwise vortices which occur in anyfree shear layers due to the KelvinndashHelmholtz instability The sec-ond is the increased interfacial contact area due to the convolutedtrailing edge of the lobed mixer The last element is the stream-wise vortices produced by the special geometry of the lobed mixerBased on pulsed laser sheet ow visualizationwith smoke and three-dimensional velocity measurements with a hot- lm anemometerMcCormick and Bennett8 suggested that it is the interaction of thespanwise KelvinndashHelmholtz vortices with the streamwise vorticesthat produce the high levels of mixing

Although the existence of unsteady vortices and turbulent struc-tures in lobed mixing ows has been revealed in those previousstudies by qualitative ow visualization the quantitative instanta-neous whole- eld velocity and vorticity distributionsin lobed mix-ing ows have never been obtained until recent work of the presentauthors9 In Ref 9 both planar laser induced uorescence (PLIF)and conventionaltwo-dimensionalparticle image velocimetry(PIV)techniques were used to study lobed jet mixing ows Based on thedirectly perceived PLIF ow visualization images and quantitativePIV velocity vorticity and turbulence intensity distributions theevolutionand interactionof variousvorticaland turbulentstructuresin the lobed jet ows were discussed

The conventionaltwo-dimensionalPIV system used in the earlierwork of the authors9 is only capable of obtaining two components

1339

1340 HU ET AL

of velocity vectors in the planes of illuminating laser sheets Theout-of-planevelocity component is lost whereas the in-plane com-ponents may be affected by an unrecoverable error due to perspec-tive transformation10 For highly three-dimensional ows such aslobed jet mixing ows conventionalPIV measurement results maynot be able to reveal their three-dimensional features successfullyA high-resolution stereoscopic PIV system which can provide allthree components of velocity vectors in a measurement plane si-multaneously is used in the present study to measure an air jet owexhausted from a lobed nozzleBased on the stereoscopicPIV mea-surement results some characteristics of the mixing process in thelobed jet ow are discussed

Experimental Setup and Stereoscopic PIV SystemFigure 1a shows the lobed nozzle used in the present study It has

six lobe structuresat the nozzle trailingedgeThe width of each lobeis 6 mm and the lobe heightis 15 mm (H D 15 mm) The inward andoutward penetration angles of the lobed structures are microin D 22 degand microout D 14 deg respectivelyThe diameter of the lobed nozzle is40 mm (D D 40 mm)

Figure 1b shows the air jet experimental rig used in the presentstudy A centrifugal compressor was used to supply air jet ows Acylindricalplenumchamberwith honeycombstructureswas used tosettle theair ow Througha convergentconnection(convergentratio501) the air ow is exhausted from the lobed nozzle The velocityrange of the air jet out of the convergent connection (at the inlet ofthe testnozzle) couldbevariedfrom5 to 35ms A meanspeedof theair jet ow at the inletof the lobednozzleof U0 D 200 ms was usedwhich corresponds to a Reynolds number Re D 5517 pound 105 (basedon the nozzlediameter D D 40mm) The air jet owwas seededwith1 raquo 5-sup1m di-2-ethylhexyl-sebact (DEHS) droplets generated by aseeding generatorThe DEHS droplets out of the seeding generator

a) Lobed nozzle

b) Experimental rig

Fig 1 Lobed nozzle and air jet experimental rig

were divided into two streams one is used to seed the core jet owand the other for ambient air seeding

Figure2 shows the schematicof thestereoscopicPIV systemusedThe lobed jet ow was illuminated by a double-pulsed NdYAGlaser set (NewWave 50-mJpulse) with the laser sheet thickness ofabout 20 mm The double-pulsedNdYAG laser set can supply thepulsed laser (pulsed illumination duration 6 ns) at a frequency of15Hz Two high-resolution(1000pound 1000)cross-correlationcharge-coupled device (CCD) cameras (TSI PIVCAM10-30) were used toperform stereoscopicPIV image recording The two CCD cameraswere arranged in an angular displacement con guration to get alarge overlapped view With the installation of tilt-axis mounts thelenses and camera bodies were adjusted to satisfy the Scheimp ugcondition (see Ref 11) In the present study the distance betweenthe illuminating laser sheet and image recording planes of the CCDcameras is about650 mm and the anglebetween the view axis of thetwo cameras is about 50 deg For such an arrangement the size ofthe stereoscopicPIV measurement window is about 80 pound 80 mm2

The CCD cameras and double-pulsed NdYAG lasers were con-nected to a workstation (host computer) via a synchronizer (TSILaserPulse synchronizer) which controlled the timing of the lasersheet illumination and the CCD camera data acquisition In thepresent study the time interval between the two pulsed illumina-tions is 30 sup1s

A general in situ calibration procedure was conducted to obtainthemappingfunctionsbetweenthe imageplanesandobjectplanes12

A target plate (100 pound 100 mm) with 100-sup1m-diam dots spaced atintervals of 25 mm was used for the in situ calibration The frontsurface of the target plate was aligned with the center of the lasersheet and then calibration images were captured at three locationsacross the depthof the laser sheetsThe space intervalbetween theselocations is 05 mm for the present study

HU ET AL 1341

The mapping function used was taken to be a multidimensionalpolynomial which is fourth order for the directions (X and Y di-rections) parallel to the laser sheet plane and second order for thedirection (Z direction) normal to the laser sheet plane The coef- cients of the multidimensional polynomial were determined fromthe calibration images by using a least-square method

The two-dimensional particle image displacements in each im-ageplanewas calculatedseparatelyby usinga hierarchicalrecursive

Fig 2 Stereoscopic PIV system

a) Instantaneous velocity eld

b) Ensemble-averaged velocity eld

c) Instantaneous streamwise vorticity distribution

d) Ensemble-averged streamwise vorticity distribution

Fig 3 Stereoscopic PIV measurement results in the ZD = 025 (ZH = 067) cross plane

PIV (HR-PIV) software developed in-houseThe HR-PIV softwareis based on HR processes of a conventional spatial correlation op-eration with offsetting of the displacement estimated by the formeriterationstep and hierarchicalreductionof the interrogationwindowsize and search distance in the next iteration step13 Compared withconventional cross-correlation-based PIV image processing meth-ods the HR-PIV method has advantagesin spuriousvector suppres-sion and spatial resolution improvement of the PIV result Finally

1342 HU ET AL

when the mapping functions obtained by the in situ calibration andthe two-dimensionaldisplacementsin the two imageplanesareusedall three componentsof the velocityvectors in the illuminating lasersheet plane were reconstructed

Because 32 pound 32 pixel interrogation windows with 50 overlapwere used for the nal step of the recursive correlation operationthe spatial resolution of the present stereoscopicPIV measurementis expected to be about 25 pound 25pound 20 mm3 To evaluate the mea-surement accuracy of the present stereoscopic PIV system a laserdoppler velocimetry (LDV) system was used to conduct simulta-neous measurement of the lobed jet ow The stereoscopic PIVmeasurement results were compared with the LDV results It wasfound that the velocitydifferencesbetween the measurementresultsof the stereoscopicPIV system and the LDV data are less than 20at the comparedpointsTherefore the accuracy level of the velocityvectors obtained by the stereoscopic PIV system is expected to beabout 20 and that of the rms of the velocity uctuations u0 v0and w0 and turbulent kinetics energy K are about 50 An adap-tive scheme14 was used in the present study to calculate streamwisevorticity $z distributions from the velocity elds obtained by thestereoscopic PIV measurement The accuracy level of the stream-wise vorticity data in the present study is expected to be within100 Further details about the in situ calibration reconstructionproceduresof the stereoscopicPIV systemand thecomparisonof thestereoscopicPIV and LDV measurementsmay be found in Ref 15

Experimental Results and DiscussionFigures 3ndash5 shows the stereoscopic PIV measurement results in

three typical cross planes of the lobed jet ow which include typi-cal instantaneousvelocity elds simultaneous streamwise vorticity

a) Instantaneous velocity eld

b) Ensemble-averaged velocity eld

c) Instantaneous streamwise vorticity distribution

d) Ensemble-averged streamwise vorticity distribution

Fig 4 Stereoscopic PIV measurement results in the ZD = 15 (ZH = 40) cross plane

distributions ensemble-averaged velocity and streamwise vortic-ity elds The ensemble-averagedvelocity and streamwise vorticity elds given in Figs 3ndash5 were calculated based on 250 frames ofinstantaneousstereoscopicPIV measurement results

In the Z=D D 025 (Z=H D 067) cross plane (almost at the exitof the lobed nozzle) the high-speed core jet ow was found tohave the same geometry as the lobed nozzle The signature of thelobed nozzle in the form of a six-lobe structure can be seen clearlyfrom both the instantaneousand ensemble-averagedvelocity elds(Fig 3) The existenceof very strongsecondarystreams in the lobedjet ow is revealedvery clearly in the velocityvectorplotsThe corejet ow ejects radiallyoutward in the lobe peaks and ambient owsinject inward in the lobe troughs Both the ejection of the core jet ow and the injection of the ambient ows generally are followingthe outward and inward contours of the lobed nozzle which resultsin the generationof six pairs of counterrotatingstreamwise vorticesin the lobed jet ow The maximum radial ejection velocity of thecore jet ow in the ensemble-averagedvelocity eld is found to beabout 50 ms which is almost equal to the value of U0 cent sinmicroout Alarge high-speedregion can also be seen clearly from the ensemble-averaged velocity eld which represents the high-speed core jet ow in the center of the lobed nozzle

The existence of the six pairs of large-scale streamwise vor-tices due to the special geometry of the lobed nozzle can be seenmore clearly and quantitatively from the streamwise vorticity dis-tributions shown in Figs 3c and 3d The size of these large-scalestreamwise vortices is found to be on the order of the lobe heightCompared with those in the instantaneousstreamwise vorticity eld(Fig 3c) the contours of the large-scale streamwise vortices in theensemble-averagedstreamwise vorticity eld (Fig 3d) were found

HU ET AL 1343

a) Instantaneous velocity eld

b) Ensemble-averaged velocity eld

c) Instantaneous streamwise vorticity distribution

d) Ensemble-averged streamwise vorticity distribution

Fig 5 Stereoscopic PIV measurement results in the ZD = 30 (ZH = 80) cross plane

to be much smoother However they have almost the same distri-bution pattern and magnitude as their instantaneous counterpartsThe similarities between the instantaneousand ensemble-averagedstreamwise vortices might suggest that the generation of large-scale streamwise vortices at the exit of the lobed nozzle to be quitesteady

The lobed jet ow was found to become much more turbulentin the Z=D D 15 Z=H D 60 cross plane compared to that at theexit of the lobed nozzle Instead of generally following the inwardand outward contours of the lobed trailing edge at the exit of thelobed nozzle the secondary streams revealed in the instantaneousvelocity eld (Fig 4a) become much more random and unsteadyThe signature of the lobed nozzle in the form of a six-lobe structureis nearly indistinguishable from the instantaneous velocity eldAlthough the high-speed region in the center of the lobed nozzlestill can be discerned from the ensemble-averaged velocity eldthe size of the high-speed region was found to decrease substan-tially due to intensive mixing between the core jet ow and ambient ow The ensemble-averaged velocity vector plot also shows thatthe ensemble-averagedsecondarystreams in the lobed jet ow havebecome much weaker The maximum radial ejection velocity of thecore jet ow has decreased to about 20 ms which is only about40 of that at the exit of the lobed nozzle

The six pairs of large-scale streamwise vortices generated by thelobed nozzle could no longer be identi ed from the instantaneousstreamwise vorticity distribution in the Z=D D 15 Z=H D 40cross plane shown in Fig 4c Instead of large-scale streamwisevortices many smaller streamwise vortices were found in the in-stantaneousstreamwise vorticity eld This indicates that the large-

scale streamwise vortices observed at the exit of the lobed nozzlehavebrokendown into many smaller streamwise vorticesHowevernote that the maximum vorticity value of the instantaneous small-scale streamwise vortices is found to be almost at the same level asthose at the exit of the lobed nozzle

From the ensemble-averagedstreamwise vorticity distribution inthis cross plane (Fig 4d) note that the strength of the large-scaleensemble-averaged streamwise vortices in the lobed jet ow hasdissipated The maximum value of the ensemble-averagedstream-wise vorticity is only about one- fth of that at the exit of the lobednozzle Because the present lobed jet ow is a freejet the centers ofthe large-scale streamwise vortices were found to have spread out-ward as they travel downstream which is also revealed very clearlyin the ensemble-averagedvelocity vector plot given in Fig 4b Thesame phenomena were also found in the LDV measurement resultsof Belovich and Samimy16

As the downstream distance increases to Z=D D 30 Z=H D80 the lobed jet mixing ow became so turbulent that the sig-nature of the lobed nozzle can no longer be identi ed from theinstantaneous velocity eld (Fig 5a) The ow eld is completely lled with many small-scale vortices and turbulent structures Theensemble-averagedvelocity eld in this cross plane shows that thedistinct high-speed region in the center of the lobed jet ow hasdissipated so seriously that isovelocity contours of the high-speedcore jet ow have become small concentric circles The ensemble-averagedsecondarystreams in this crossplane becomeso weak (themaximum secondary stream velocity of less than 08 ms) that theycan not be identi ed easily from the ensemble-averaged velocityvector plot

1344 HU ET AL

From the instantaneous streamwise vorticity distribution in theZ=D D 30 Z=H D 80 crossplane given in Fig 5c note that thereare so many small-scale streamwise vortices in the lobed jet mixing ow that they almost lled the measurementwindow However themaximum vorticityvalueof these instantaneoussmall-scale stream-wise vortices was found to be still at the same level of those in theupstream cross planes Because of serious dissipation caused bythe intensive mixing between the core jet ow and ambient owalmost no apparent streamwise vortices can be identi ed from theensemble-averagedstreamwise vorticty distribution(Fig 5d) in theZ=D D 30 Z=H D 60 cross plane

Based on the stereoscopic PIV measurement results at 12 crossplanes the ensemble-averaged three-dimensional ow eld in thenear downstream region of the lobed jet ow was reconstructedFigure 6a shows the three-dimensionalensemble-averagedvelocityvectors The velocity isosurfaces of the reconstructed three-dimensional ow eld are given in Fig 6b The velocitymagnitudesof these isosurfaces are 40 80 120 and 160 ms respectivelyNote that the high-speed core jet ow has the same geometry asthe lobed nozzle that is six-lobe structure at the exit of the lobednozzle Because of the stirring effect of the large-scale streamwisevortices generated by the lobed nozzle the six-lobe structure of thecore jet ow was rounded up rapidly At Z=D D 30 Z=H D 80downstreamthe isosurfaceswere found to becomeconcentriccylin-ders which are very similar to those in a circular jet ow

a) Velocity vectors

b) Velocity isosurfaces

Fig 6 Reconstructed three-dimensional ow eld

The precedingstereoscopicPIV measurementresultshave shownthat the size of the instantaneous streamwise vortices in the lobedjet ow decreases as the downstream distance increases that is thelarge-scalestreamwisevorticesgeneratedby the lobednozzle brokeinto smaller vortices as they traveldownstreamHowever the maxi-mum vorticityvaluesof the smaller vorticeswere found to be almostat the same level as their parent streamwise vortices These resultssuggested that the dissipationof the large-scalestreamwise vorticesgenerated by the corrugated trailing edge of the lobed nozzle didnot happen abruptly but rather appeared to be a gradual processThe large streamwise vortices were found to break down into manysmaller but not weaker streamwise vortices as the downstream dis-tance increasesThus besides the large-scalemixing enhancementmixing at a ner scale could also be achieved in the lobed jet mix-ing ow The result therefore agrees well with those obtained byBelovich et al17

The lobed nozzle might be thought to act as a uid stirrer in thelobed jet ow to mix the core jet ow with ambient ow The stirringeffect of the lobed nozzle on the mixing processes can be evaluatedfrom the evolution of the ensemble-averaged streamwise vorticesFigure 7 shows the decay pro le of the ensemble-averagedstream-wise vorticity in the lobed jet ow Note that over the rst twodiameters downstream of the lobed nozzle ( rst six lobe heights)the ensemble-averaged streamwise vorticity decayed very rapidlyand then the decay rate slowed farther downstream This may in-dicate that the stirring effect of the lobed nozzle to enhance themixing between the core jet ow and ambient ow is concen-trated primarily in the rst two diameters ( rst six lobe heights) Inthe cross plane of Z=D D 30 Z=H D 80 the ensemble-averagedstreamwise vorticity dissipated so seriously that the strength of theensemble-averagedstreamwise vortices became only about 1

10 th ofthat at the exit of the lobed nozzle This means that the stirring ef-fect of the lobed nozzle has almost disappeared at this downstreamlocation

The characteristicsof themixingprocessbetweenthecore jet owand ambient ow in the lobed jet ow may be revealedmore quanti-tatively from the averaged-turbulentkinetic energy pro le given inFig 8 In the present paper the averaged-turbulentkinetic energy

Fig 7 Decay of the ensemble-averaged streamwise vorticity

Fig 8 Averaged turbulent kinetic energy pro le

HU ET AL 1345

of the lobed jet ow at each measured cross plane was calculatedby using the following equation

NK z DR

frac12W x y zK x y z dx dyRfrac12W x y z dx dy

It was found that the averaged-turbulentkinetic energy grew veryrapidly within the rst diameter of the lobed nozzle and then de-creased to a more moderate rate farther downstreamThe averaged-turbulent kinetic energy pro le reached its peak value at aboutZ=D D 20 (Z=H D 533) and then began to decrease This maybe explained as follows Within the rst one diameter of the testnozzle due to the stirring effect of the large-scale streamwise vor-tices generatedby the lobed nozzle revealedin the given streamwisevorticity distributions the core jet ow and ambient ow mixed in-tensively In the downstream region of Z=D gt 10 (Z=H gt 267)the large-scale streamwise vortices generated by the lobed nozzlewere found to break down into smaller vortices that is the stirringeffect of the large-scale streamwise vortices weakened Thus thegrowth rate of the averaged turbulent kinetic energy was found todecrease and to reach its peak at about Z=D D 20 (Z=H D 533)The ensemble-averaged streamwise vorticity has become so weakthat it almost cannot affect the mixing process in the farther down-stream region Because most of the turbulent kinetic energy in thelobed jet mixing ow dissipated in the intensive mixing upstreamthe averaged turbulent kinetic energy begins to decrease It mayalso indicate that most of the enhancementof mixing caused by thelobe nozzle occurs within the rst two diameters This result is alsofound to agree with the ndings of McCormick and Bennett8 andGlauser et al18 in a planar lobed mixing ow who suggested thatthe maximum effectiveness region for a lobed mixer is about the rst six lobe heights

Because large-scale streamwise vortices generated by thelobed nozzle were dissipated in the farther downstream region(Z=D gt 30 Z=H gt 80) they no longer can enhance the mixingprocess in the lobed jet ow The ensemble-averagedvelocity dis-tributionsand velocity isosurfacesshown in Figs 5 and 6 reveal thatthe isovelocity contours in the lobed jet ow in the farther down-stream region (Z=D gt 30 Z=H gt 80) become concentric circleswhichare similar to thosein a circularjet ow Thereforethe mixingbetween the core jet ow and the ambient ow in the farther down-stream region (Z=D gt 30 Z=D gt 80) is expected to occur by thesame gradient-typemechanism as that in a conventionalcircular jet ow

SummaryA high-resolutionstereoscopicPIV system was used to measure

an air jet ow exhausted from a lobed nozzle The evolution of thelarge-scale streamwise vortices in the lobed jet ow which origi-nated from the strong secondary streams due to the geometry of thelobed trailing edge was revealed instantaneouslyand quantitativelyfrom the stereoscopicPIV measurement results The instantaneousvelocity elds reveal that the core jet ow expands radially out-ward in the lobe peaks and ambient ow injects inward in the lobetroughs at the exit of the lobed nozzle which results in the gen-eration of large-scale counter-rotating streamwise vortices in thelobed jet ow The large-scale streamwise vortices break down intosmaller but not weaker vortices as they travel downstream This isproposedas the reason that a lobednozzlewould enhanceboth large-scale mixing and small-scale mixing reported by other researchersThe overall effect of the lobed nozzle on the mixing process inthe lobed jet ow was evaluated by analyzing the evolution of theensemble-averagedstreamwise vorticity distributionsThe strengthof the ensemble-averagedstreamwise vortices was found to decayvery rapidly in the rst two diameters of the lobed nozzle then turnto a moderate decay rate farther downstreamThis may indicate thatthe stirring effect of the lobed nozzle to enhance mixing between

the core jet ow and ambient ow is concentrated primarily in the rst two diameters of the lobed nozzle ( rst six lobe heights) Theaveraged turbulent kinetic energy pro le also indicated that mostof the intensive mixing between the core jet ow and ambient owoccurred over the rst two diameters of the lobed jet ow ( rst sixlobe heights) The isovelocity contours of the lobed jet ow in thefarther downstreamregions (Z=D gt 30 Z=H gt 120) are found tobe concentriccircleswhich arequite similar as those in a circular jet ow These results indicate that the mixing enhancement caused bythe special geometry of the lobed nozzle concentrates primarily inthe rst two diametersdownstreamof the lobednozzle ( rst six lobeheights) The mixing between the core jet ow and ambient ow inthe fartherdownstreamregion (Z=D gt 30 Z=H gt 120) will occurby the same gradient-typemechanism as that for a circular jet ow

References1Kuchar A P and Chamberlin R ldquoScale Model Performance Test In-

vestigation of Exhaust System Mixers for an Energy Ef cient Engine (E3 )rdquoAIAA Paper 80-0229 1980

2Presz W M Jr ReynoldsG and McCormick D ldquoThrust Augmenta-tion Using MixerndashEjectorndashDiffuser Systemsrdquo AIAA Paper 94-0020 1994

3Smith LL Majamak A J Lam I T Delabroy O Karagozian A RMarble F E and Smith O I ldquoMixing Enhancement in a Lobed InjectorrdquoPhysics of Fluids Vol 9 No 3 1997 pp 667ndash678

4Paterson R W ldquoTurbofan Forced Mixer Nozzle Internal Flow eldrdquoNASA CR-3492 1982

5Werle M J Paterson R W and Presz W M Jr ldquoFlow Structure ina Periodic Axial Vortex Arrayrdquo AIAA Paper 87-6l0 1987

6Eckerle W A Sheibani H and Awad J ldquoExperimental Measurementof the Vortex Development Downstream of a Lobed Forced Mixerrdquo Journalof Engineering for Gas Turbine and Power Vol 114 Jan 1992 pp 63ndash71

7Elliott J K Manning T A Qiu Y J Greitzer C S Tan C Sand Tillman T G ldquoComputational and Experimental Studies of Flow inMulti-Lobed Forced Mixersrdquo AIAA Paper 92-3568 1992

8McCormick D C and Bennett J C Jr ldquoVortical and TurbulentStruc-ture of a Lobed Mixer Free Shear Layerrdquo AIAA Journal Vol 32 No 91994 pp 1852ndash1859

9Hu H Saga T Kobayashi T and Taniguchi N ldquoResearch on theVortical and Turbulent Structures in the Lobed Jet Flow by Using LIF andPIVrdquo Measurement Science and Technology Vol 11 No 6 2000 pp 698ndash

71110Prasad A K andAdrianR J ldquoStereoscopicParticle Image Velocime-

try Applied to Fluid Flowsrdquo Experiments in Fluids Vol 15 No 1 1993pp 49ndash60

11Prasad A K and Jensen K ldquoScheimp ug Stereocamera for ParticleImage Velocimetry in Liquid Flowsrdquo Applied Optics Vol 34 No 30 1995pp 7092ndash7099

12Soloff S M Adrian R J and Liu Z C ldquoDistortion Compensationfor Generalized Stereoscopic Particle Image Velocimetryrdquo MeasurementScience and Technology Vol 8 No 12 1997 pp 1441ndash1454

13Hu H Saga T Kobayashi T Taniguchi N and Segawa S ldquoIm-prove the Spatial Resolution of PIV Results by Using Hierarchical RecursiveOperationrdquo Proceedings of 9th International Symposium on Flow Visual-ization edited by G M Carlomagno and I Grant Paper 137 2000

14Lourenco L and Krothapalli A ldquoOn the Accuracy of Velocity andVorticity Measurements with PIVrdquo Experiments in Fluids Vol 18 No 61995 pp 421ndash428

15HuH ldquoInvestigationon Lobed Jet MixingFlowsbyUsing PIV and LIFTechniquesrdquo PhD Dissertation Dept of Mechanical EngineeringUniv ofTokyo Tokyo April 2001

16Belovich V M and Samimy M ldquoMixing Process in a Coaxial Ge-ometry with a Central Lobed Mixing Nozzlerdquo AIAA Journal Vol 35 No 51997 pp 838ndash84l

17BelovichV M Samimy M and Reeder M F ldquoDual Stream Axisym-metric Mixing in the Presence of Axial Vorticityrdquo Journalof PropulsionandPower Vol 12 No 1 1996 pp 178ndash185

18Glauser M Ukeiley L and Wick D ldquoInvestigation of TurbulentFlows Via Pseudo Flow Visualization Part 2 The Lobed Mixer Flow FieldrdquoExperimental Thermal and FluidScience Vol 13 No 2 1996pp 167ndash177

J P GoreAssociate Editor