8/8/2019 497625

1/11

International Journal of Rotating Machinery1999, Vol. 5, No. 2, pp. 89-98Reprints available directly from the publisherPhotocopying permitted by license only

(C) 1999 OPA (Overseas Publishers Association) N.V.Published by license under

the Gordon and Breach SciencePublishers imprint.

Printed in Malaysia.

Turbulence Amplification with Incidence at theLeading Edge of a Compressor Cascade

GARTH V. HOBSON a*, BRYCE E. WAKEFIELD and WILLIAM B. ROBERTS bCode AA/Hg, Department of Aeronautics and Astronautics, Naval Postgraduate School, 699 Dyer Rd, Rm 137,Monterey, CA 93943-5106, USA,,bFlow Application Research Fremont, California, USA

(Received 28 October 1997;In final form 19 March 1998)Detailed measurements, with a two-component laser-Doppler velocimeter and a thermalanemometer were made near the suction surface leading edge of controlled-diffusion airfoilsin cascade. The Reynolds number was near 700,000, Mach number equal to 0.25, andfreestream turbulence was at 1.5% ahead of the cascade.It was found that there was a localized region of high turbulence near the suction surfaceleading edge at high incidence. This turbulence amplification is thought to be due to theinteraction of the free-shear layer with the freestream inlet turbulence. The presence of thelocal high turbulence affects the development of the short laminar separation bubble thatforms very near the suction side leading edge of these blades. Calculations indicate that thelocal high levels of turbulence can cause rapid transition in the laminar bubble allowing it toreattach as a short "non-burst" type.The high turbulence, which can reach point values greater than 25% at high incidence, isthe reason that leading edge laminar separation bubbles ca n reattach in the high pressuregradient regions near the leading edge. Two variations for inlet turbulence intensity weremeasured for this cascade. The first is the variation of maximum inlet turbulence with respectto inlet-flow angle; and the second is the variation of leading edge turbulence with respect toupstream distance from the leading edge of the blades.

Keywords." Compressors, Boundary layers, Separation bubbles, Laser-Doppler velocimetry

INTRODUCTION

While experimental research was being conductedon the Sanger cascade blades at high inlet-airangles in the cascade wind tunnel of the NavalPostgraduate School (Sanger and Shreeve, 1986),

complementary calculations, which were initiallyunsuccessful, were performed in an attempt topredict the off-design performance. Flow visualiza-tion studies by Sanger and Shreeve, at chordReynolds number of 340,000, indicated thepresence of a laminar separation bubble near the

Corresponding author. E-mail: garth@aa.nps.navy.mil.89

8/8/2019 497625

2/11

90 G.V. HOBSON et al.

leading edge. The cascade geometry was docu-mented in Elazar and Shreeve (1989), and was thesame geometry used for the calculations. Theviscous flow features of this cascade are shownin Fig. 1.

Testing by Hobson and Shreeve (1993) indicatedthat at a very high inlet-flow angle (/31 48.4) theturbulenceintensity right at the leadi-ng edge wasamplified for the Sanger cascade. This couldexplain how a laminar separation bubble wouldbe able to reattach as a short bubble in the steeppressure gradient near the leading edge at highincidence.A deeper understanding of this phenomenon isdesirable so as to correctly compute the boundarylayer development in the leading edge, especially forcompressor blades operating near stall. This is notonly true for relatively simple inviscid-viscousmethods, but also for full Navier-Stokes calcula-tions. To gain a greater understanding of thisphenomenon, detailed LDV measurements weremade in the leading edge region of the Sangercascade for inlet-flow angles of 1 43.3, 46.4and 48.4 The order of magnitude of the leadingedge turbulence amplification, which wa g signifi-cant at high incidence, was confirmed by thermalanemometer measurements at #1 48.4.

SEPARATIONTRANSITION J_ ,.:."

,- PRESSURE SIDEINLETAR SUCTION SIDEFLOW

/FIGURE Viscous Flow features of the Sanger cascade(Elazar and Shreeve, 1989).

DESCRIPTION OF THE PHYSICS ANDANALYSIS TECHNIQUEAn inviscid-viscous scheme was used for thesubsonic calculations (Martensen, 1959; LeFoll,1965; Roberts, 1975). It was found that during thecalculation the inviscid-viscous method predicted a"burst" laminar separation bubble very near theleading edge for/31 > 38, while the data of Sangerand Shreeve (1986) indicated the presence of a shortbubble. Figure 2 shows a schematic of a shortlaminar separation bubble. A short laminar bubble"bursts" into a long bubble when reattachment isnot possible in the short state, see Fig. 3 (for adescription of the flow regimes possible withvarying Reynolds number see, Roberts, 1975). Ascan be seen from Fig. 4 the suction surface velocitydistribution is reasonably well predicted by theinviscid code. Since a long or "burst" laminarseparation bubble by definition causes a significantchange from the inviscid velocity distribution,which results in a decrease in the suction peak,this implies that the bubble present on the bladewill be short.

As stated above the application of the boundarylayer method predicted that the leading edgelaminar separation bubble would not reattach.The Reynolds number and freestream turbulenceintensity (Tu-1.5%) were taken from upstreamtest conditions as given by Elazar and Shreeve(1989). The inviscid-viscous method was calibratedby Roberts (1975) for short and long mid-chordlaminar separation bubbles, and for freestreamturbulence intensities between 0% and 5%. Themethod has been successfully applied to shortbubbles found near the leading edge of a largechord wing model ( 1. 2 m) of an NACA 6613-018section, which was experimentally measured byGault (1955) in a low turbulence wind tunnel.The only mechanism that affects transition in a

laminar bubble, for a fixed geometry and Reynoldsnumber, is the local turbulence level. The length ofthe laminar part of the bubble is decreased for anincreased value of local freestream turbulence.Therefore, very high local turbulence could cause

8/8/2019 497625

3/11

TURBULENCE AMPLIFICATION

FREE STREAH FLOW./

91

sic

FIGURE 2 Sectional view of a two-dimensional short laminar separation bubble and corresponding surface velocitydistribution.

rapid transition after separation allowing shortbubble reattachment even in a severe pressuregradient. Additional calculations were performedfor l 40 , 43.3 and 46.4 and compared to thelaminar bubble reattachment locations reported bySanger and Shreeve (1986). These were determinedfrom flow visualization studies performed at aReynolds number of 340,000. For the calculationsthe experimental velocity distribution was used forthe three inlet-air angles mentioned above. This wasdone to ensure that the laminar boundary layer wasproperly calculated so as to correctly locate thelaminar separation point.

At first the freestream turbulence of Tu 1.5%was used in the calculations resulting in theprediction of "burst" bubbles. The turbulence levelwas then increased in subsequent calculations untilshort bubble reattachment occurred for each inlet-ai r angle. Finally, the turbulence level was furtherincreased until the reattachment location agreedwith the experimental data. This is shown in Fig. 5where the turbulence required to match the data isindicated: Tu 8.5% for fl 40 , Tu 9.0% for/3 43.3 and Tu 11.0% for/3 46.4. Not onlydoes this indicate that at medium to high incidencethe leading edge turbulence is amplified, but that

8/8/2019 497625

4/11

92 G.V. HOBSON et al.

sicHigh Rc: No separation, velocity distribution approximates

inviscid distribution2 Medium Rc: Short separation bubble, distribution approximatelyinviscid outside bubble region3 Low Rc: Long bubble after bursting, distribution significantlyaffected4- Lower Rc: Long bubble with complete separation

FIGURE 3 Schematic of four flow regimes possible withvarying Reynolds number (104-106 over the suction surfaceof a compressor profile (Roberts, 1975).

0.05

04

0.03

0.02

0.01

0 I, I,38 40 42 44 46 48Inlet Flow Angle (deg.)

FIGURE 5 Separation bubble reattachment for various tur-bulence intensities (Re- 340,000).

1.8Znviscid calcullon

1o4

0.6 0.2 0.4 0.6 0.8Non-dimensional Chord, X/C

FIGURE 4 Suction surface velocity comparison/31 40 .

were performed with a two-component system,which is fully described by Elazar and Shreeve(1989). Figure 6 also shows the location of theintroduction of seeding into the bellmouth of thetunnel, the profile coordinates of the Sanger Blade,an d the cascade geometry and inlet conditions.A 20 gm (sensor diameter) hot-film probe, which

had a sensor length of mm, was used with a TS Isingle channel hot-wire anemometer system (IFA-100 and -200) connected to a personal computer. Acomplete description of the hot-film instrumenta-tion is given by Wakefield (1993).

EXPERIMENTAL PROCEDURE

the level of turbulence is also a positive functionof incidence.

TEST FACILITY AND INSTRUMENTATIONThe experiments were performed in the Low SpeedCascade Wind Tunnel (LSCWT) at the Turbo-propulsion Laboratory of the Naval PostgraduateSchool, which is shown schematically in Fig. 6.For a more detailed description of the facility seeSanger and Shreeve (1986). The LDV measurements

Inlet pitchwise LDV surveys were performed aheadof the leading edge of the blades for three differentinlet-flow angles (43.3 46.4 and 48.4). The si xaxial locations of the survey planes and respectiveorientation of the LDV were the same as thosedescribed by Hobson and Shreeve (1993). In theirstudy, they performed detailed measurementsupstream, downstream and through the passageof the blade row including around the leadingedge separation bubble at the high inlet-flow angleof 48.4

Hot-film surveys were performed at the threesurvey planes which were closest to the leading

8/8/2019 497625

5/11

TURBULENCE AMPLIFICATION 93

ACRYLIC WINDOW

ENDWALLSTATIC 0.165m0.311m

0.191m

0514m

1.524m NO. OF TEST BLADES 20NO. OF ILET GUIDE VANES 60

o Io

SEEDINGLOCATION (press. side) (suet. side)(mm) (mm) (mm)0.00 0.114 0.1140.6 0.213O. 145 0.0050.564 0. 12 0.49g1.128 0.257 0.7801.692 0.394 1.0242.256 0.$26 1.2402.819 0.648 1.4253.383 0.759 .$773.947 0.838 1.6844.51 0.889 1.7555.075 0.912 .7915.639 0.912 1.7986.203 0.894 1.7816.767 0.869 1.730"/.330 0.841 1.6517.894 0.805 1.5498.458 0.765 .4309.022 0.714 1.2959.586 0.653 1.15110.150 0.577 0.99810.714 0.485 0.843il.278 0.371 0.686.841 0.226 0.52812.405 0.048 0.36812.510 0.01012.609 0.31012.725 0.157 0.157

Blade Type Controlled DiffusionNumber of BladesBlade Spacing 76.2Chord 127.3Solidity 1.67Lding I Radius 1.14Trailing Edge Radius 1.57ThicknessSetting Angle 14.2 0.1Stagger Angle 14.4 0.1Spau 25 4

Reynolds No, 700,000(Chord)rotal Temp. 294 KTotal Press. 1.03 ATMMach Jqumber 0.25Freestream Turbulence 1.5%Static Pressure 1.00 ATM

FIGURE 6 Schematic of the low speed cascade wind tunnel, controlled-diffusion compressor blade and test conditions.

edge of the blades (2.17%, 1.10% and 0.57%axial chord ahead of the leading edges). Thehot-film probe was positioned horizontally in thetunnel (i.e. in the tangential direction, with respectto the blades, with the sensor parallel to thespanwise direction) whilst traversing across theleading edges. This was done to ensure that therewas no probe stem interference on the mea-sured turbulence level as it was traversed past theleading edge.

RESULTS AND DISCUSSIONAs can be seen from Figs. 7-9, the locus of points ofmaximum turbulence intensity, for all three inlet-flow angles considered, approaches the bladeleading edge at right angles to the approachingstagnation streamline. The approximate location ofthe stagnation streamline is shown as the locus ofpoints of minimum total velocity. Due to smallblade setting angle errors (< 0. 1 ) and the lack of

8/8/2019 497625

6/11

94 G.V. HOBSON et al.

FIGURE 7 Inlet-flow angle 43.3 deg. FIGURE 8 Inlet-flow angle 46.4deg.

-0 -6

-0 7

-0 8

-1.2

-1.3

// /it rain. velocity "-"/ /! max. turbulence -+rain. velocity --ax. turbulence (-, /max. production--- /max productioFLOW FLOW

, :!" / ".,

".."J--"x" 4 -i--0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8X/PTCS

FIGURE 9 Inlet-flow angle=48.4deg. (Locus of points of min. total velocity, max. turbulence and max. turbulenceproduction).

8/8/2019 497625

7/11

TURBULENCE AMPLIFICATION 95

perfectly two-dimensional inlet-flow there weresmall differences in trajectories of the maximumturbulence and minimum total velocity from bladeto blade. However, the overall trend as the locus ofpoints approaches the blades is periodic.The increase in turbulence intensity, ahead of the

blade leading edge, is not due to streamwisediffusion, but is caused by local shear as the flowattempts to accelerate around the leading edge.Local shear will produce turbulence, as is shown bythe production term in the transport equation forturbulent kinetic energy (Hinze, 1975),

where the second term, on the right hand side, is thework by the viscous shear stresses of the turbulentmotion or the production term.

A bi-cubic spline was fitted to the individualcomponents of the LDV data in the pitchwisedirection. This then allowed the differentiation ofthe measured velocity field, to determine each ofthe components of the two-dimensional produc-tion term

The distribution of measured turbulence produc-tion is shown in Fig. 10 for/31 48.4. A "ridge" ofhigh turbulence production exists in the regionahead of the blades, at right angles to the stagnationstreamline and parallel to the points of maximumturbulence intensity. The significant turbulentproduction is the reason for the increase inturbulence ahead of the blades, particularly atincreasing incidence.

Next the actual increase in turbulence intensityalong the line of maximum intensity is plottedfor each of the three test cases (see Fig. 11). As can

0 5O0O0

o

j 0

FLOW

X [m 1.14-0.13

-0.135 Y [m ]

FIGURE 10 Upstream turbulence production for 48.4deg. inlet-flow angle.

-0.125

8/8/2019 497625

8/11

96

35

3O

25

20-

15

10

5

G.V. HOBSON et al.

Inlet Flow Angle 43

-35 -30 -25

Inlet Flow Angle 46-20 .15

-3 5 -30 -25 -20

LDV blade 8LDV blade 7 -+--.hotfilm blade 7

3O

2o 25

2010

15

10

-15 -10 -5Inlet Flow Angle 48 deg.

-40 -35 -30 -25 -20 15 10 -5% Axial Chord

FIGURE 11 Maximum inlet turbulence intensity for varying inlet-flow angle.

be seen for all three test cases, the increase is expo-nential and highly localized around the leadingedge. Note that the hot-film probe measurementsverify the turbulence level measured by the LDVsystem.

Furthermore in Fig. 12 the maximum turbulenceintensity as measured by the LDV is plotted foreach of the three inlet-flow angles. Here the increasein maximum turbulence intensity is also seen toincrease non-linearly, with increasing slope forincreasing suction-side velocity gradient. Note thatthe trend of increasing turbulence intensity withincreasing incidence is similar to that of the datamatch of Fig. 5. However, the data presented isfrom a surface flow visualization at lower Reynoldsnumber, while that of Fig. 12 has been measured byboth LDV and a hot-film approximately 1% ofchord ahead of the leading edge.

Finally, the variation of turbulence intensityalong the blade suction surface is given in Fig. 13.

I-, 20EE

4a

Incidence Angle (deg,)

Inlet Flow Angle (deg,)

LDVHotfilm

FIGURE 12 Maximum cascade inlet turbulence near theleading edge.

The values in this figure are LDV measurementstaken in the freestream at the edge of the boundarylayer at mid-span and plotted in the streamwisedirection. It can be seen that the turbulent bound-ary layer downstream of the laminar bubble devel-ops in an elevated turbulence intensity environment

8/8/2019 497625

9/11

TURBULENCE AMPLIFICATION 97I"

Boundary layer thickness (mm)Edge turbulence (%) +20

15

t0 ++

++@ + +

0.i 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9Fraction of Chord

FIGURE 13 Boundary layer thickness and variation of edgeturbulence on the suction side at 48.4 deg. inlet-flow angle.

that varies from approximately 12% near theleading edge to approximately 4% at the trailingedge. This is much higher than the inlet freestreamvalue of 1.5% and this will have a significant effecton boundary layer development.

CONCLUSIONSThe experiments reported herein show that thefreestream turbulence can be amplified greatly nearthe leading edge of a compressor blade at mediumto high incidence angle. This increased turbulencelevel can have a strong effect on the boundary layertransition process, especially when conditions allowthe formation of a laminar separation bubble.The reasonable prediction of transition through a

laminar bubble requires at least a reasonable assess-ment of the turbulence environment approachingthe leading edge. For laminar bubbles occurringnear the leading edge of compressor blades at highincidence, as well as the continued development ofthe turbulent boundary layer, the amplification offreestream turbulence should be taken into accountin order to perform an accurate calculation of theseflow phenomena. Also, the variation of elevatedturbulence levels over the blade surface will affectboundary layer development downstream of thebubble and should be taken into account.

Finally, large scale experiments should beperformed to better define the flow phenomenonof laminar bubbles and turbulence amplification atthe leading edge suction surface of compressorblades at off-design incidence and over the down-stream surfaces.

NOMENCLATURECPq2ReSTuU/AV

chord lengthpressureu/u[, twice energy of turbulencechord Reynolds numberdistance along the chordturbulence intensity, V/u 2 + v2/Vreftangential velocitytangential fluctuating velocityaxial velocityaxial fluctuating velocitygre upstream reference velocity, v/U 2 - V2upstreamVtot local total velocity, v/U 2 + V 2x tangential direction

y axial direction/31 inlet flow angleu kinematic viscosity

AcknowledgmentsThe authors would like to thank the Propulsion andPower Engineering group at the Naval Air WarfareCenter (Trenton), and in particular Mr. StoneyMacAdams for his support of this research, whichwas funded as part of a Fan and Compressor Stallproject. The last author would also like to thank theTurbopropulsion Laboratory of the Naval Post-graduate School for his support on this project.

ReferencesElazar, Y. and Shreeve, R.P. (1989) Viscous flow in a controlled-

diffusion compressor cascade with increasing incidence,ASME Journal of Turbomachinery, 112(2), 256-266.Gault, D.E. (1955) An experimental investigation of regions ofseparated flow, NACA TN 3505.

Hinze, J.O. (1975) Turbulence, Second Edition, McGraw-Hill.

8/8/2019 497625

10/11

98 G.V. HOBSON et al.

Hobson, G.V. and Shreeve, R.P.S. (1993) Inlet turbulencedistortion and viscous flow development in a controlled-diffusion compressor cascade at very high incidence, AIAAJournal ofPropulsion and Power, 9(3), 397-407.LeFoll, J. (1965) A theory of the representation of boundarylayers on a plane, Proceedings of Seminar on AdvancedProblems in Turbomachinery, yon Karman Institute.

Martensen, E. (1959) Berechnung der Druckverteilung anGitterprofilen in ebener Potentialstr6mung mi t einer Fred-holmschen Integralgleichung, Archivesfo r Rational Mechanicsand Analysis, Truesdell, 3(3), 235.

Roberts, W.B. (1975) The effect of Reynolds number and laminarseparation on axial cascade performance, ASME Journal ofEngineering for Power, 97A(3), 261.Sanger, N.L. and Shreeve, R.P. (1986) Comparison of calculatedand experimental cascade performance for controlled-diffu-sion compressor stator blading, ASME Journal of Turbo-machinery, 108, 42-50.

Wakefield, B.E. (1993) Hotwire measurements of the turbulentflow into a cascade of controlled-diffusioncompressor blades,Master of Science in Engineering Science.

8/8/2019 497625

11/11

ENERGY MATERIALSMaterials Science & Engineering for Energy Systems

Economic and environmental factors are creating ever greater pressures for theefficient generation, transmission and use of energy. Materials developments arecrucial to progress in all these areas: to innovation in design; to extending lifetimeand maintenance intervals; and to successful operation in more demandingenvironments. Drawing together the broad community with interests in theseareas, Energy Materialsaddresses materials needs in future energy generation,

transmission, utilisation, conservation and storage.The journal covers thermalgeneration and gas turbines; renewable power (wind, wave, tidal, hydro, solar andgeothermal); fuel cells (low and high temperature); materials issues relevant tobiomass and biotechnology; nuclear power generation (fission and fusion);hydrogen generation and storage in the context of the hydrogen economy; andthe transmission and storage of the energy produced.

As well as publishing high-quality peer-reviewed research, Energy Materialspromotes discussion of issues common to all sectors, through commissionedreviews and commentaries.The journal includes coverage of energy economicsand policy, and broader social issues, since the political and legislative contextinfluence research and investment decisions.

SUBSCRIPTION INFORMATIONVolume 1 (2006), 4 issues per yearPrint ISSN: 1748-9237 Online ISSN: 1748-9245Individual rate: 76.00/US$141.00Institutional rate: 235.00/US$435.00Online-only institutional rate: 199.00/US$367.00For special IOM3 member rates please emailsubscriptions@maney.co.uk

EDITORSDr Fujio AbeNIMS, Japan

Dr John Hald, IPL-MPT,Technical University ofDenmark, Denmark

Dr R Viswanathan, EPRI, USA

For further information please contact:Maney Publishing UKTel: +44 (0)113 249 7481 Fax: +44 (0)113 248 6983 Email: subscriptions@maney.co.ukor

Maney Publishing North AmericaTel (toll free): 866 297 5154 Fax: 617 354 6875 Email: maney@maneyusa.com

For further information or to subscribe online please visitwww.maney.co.uk

CALL FOR PAPERSContributions to the journal should be submitted online athttp://ema.edmgr.com

To view the Notes for Contributors please visit:www.maney.co.uk/journals/notes/ema

Upon publication in 2006, this journal will be available via theIngenta Connect journals service.To view free sample content

online visit: www.ingentaconnect.com/content/maney

NEW

FO

R2006

Maney Publishing on behalf of the Institute of Materials, Minerals and Mining