T4
TH. UNIVERSITY OF TEXASAT AUSTIN
ARL-TR-70-44 CoFy No. ,,6
December ')70TESTS OF VORTEX GENERATORS TO PREVENT
SEPARATION OF SUPERSONIC FLOW IN ACOMPRESSION CORNER
NAVAL AIR SYSTEMS 'OMMANDDavid K. 'Garting APL'JHU Subconract 2, 1734, Tcsk B
p.RCHL-J .. . ., . .
. I i-I , .. ,
NATIONAL TECHNICALINFORMATION SERVICE
.prngfo d, a/. 22111
Approved for publ,C r#eIeaOe
distriuvton j,,mi; ed
BestAvailIable
COPY
- ---.------. bOUM6MTCOT0LDA"WA,4D- -5-i-etl~ s*s 4tiIU.n of title, be~,t 0otrrswWj mina g I~~J br &e nf d whtr tht vi-iaif rei.r is ripsilhrdj
Appled Research Laboratovies !CLASU~IEDThe University of Texas at Au~tiII KUA~ustin, Texas 78712 _ _____
TESTS OF VO7.EK GEliERAOpS To) P jM SEPARAjTIChl OF SUPMRCH~IC FLOW TN~ ACCRFSION CORNER
- - 4 btCtCftIPTIV9E NO1S(7I'p#olt4vtifand C*iVive42)
tedbuical report_47' ?#41tst fltpt r,00. alih, iI wal. tt name)
David K. Garting
6- "CWQRY CATC 70. TO&% MO. OF VA4Z bS, PN. or "itWs
-Decemnber 1-970 83 -C, ONY'IACT ON GRANdT NO . OetATOWO WSOO 4UuU~mIftl
APL/JJH subcontract LqT34,h. PhOJgEC T.
Task B ARIL-TR-7O-4C.' OFNSR PtOR OIS: (A'r *O n,4... that mi., boe S5I.d
Approved for Public release; distribution unlimited.
11 SOPLECTuAIV4Y NOTES 13. IS'ONSBMLtTLRf ACTIVITY
-aal Air Systems Commma: partment of the NavyWashington, D. C. 2036
iTests were ma.de on four vrortex generaLtors to determine their ability t--3 prevtuent.flow separation. The geiherators- were of rectangular planform vitha chor~d lengthsof 0.;5 and 1. 0 in:. and spans of 0.25 and 0.50 in. Generator section shape was as ynnetric double wedge. The generators were individually mounted at a 15 de~anele of attack on a-f-iat plate upstream of a 35 deg compresaion corner. Theteuts were nade in a wind tunnel at an avi'raire Mach Nuimber ofP 4+.67. Trhe bolundarylayer was turbu'Lent and approximately 1/4 in. thick at the te.st station. StaticPressure measurements- were taken along the. centerline of the plate and on therap fao Oi I and dye flows and schlierenphotographs were used to aid indeterin I1ng- the extent of separation.- Th.- vortex generators ve-e fnmd to preventseparation if placed suff'iciently close to the compression corner. The unseparatedcregion in each case was very narrov in spanvise extent. Separtiin was prevented16y increasIng the mom~enti of the boundtry layer through the action of the tip-vortex and- a secondary vortex. (U)
0I DO "'0?.1473 UNCLASSIFIEDS/_N14S74o ~c~t ~umf'to
f I IetR I t VILV aTurbule-nt FlInw IVc~rtex GeneratorsCompressible Flow
Bt-undsx'y Lay rs
DD~ ~ ~ ~ fI"" J 73tCASYf p G.2
r L-T-04
DeeFbt17
Decernbet 1970mmt 4
APL/JH'Je Sjbotc t 2fl , rsic B
rt E -=. -
APPROVED FOR PLJRUC fi~ 4RELEASE; M!STRIUIIONUNLiMITED-
v ~APPLIED RESEARCH LABORATORIESS THE UNIV-ERSITY OF TEXAS AT AUSTIN
A&VJSTR, TEXAS -7 S-12
ABSTRACT
Tests were made on four voitex generators to determine their
ability to prevent flow separation. The generators were of rectangular
planform with chord lengths of 0.5 and 1.0 in. and spans of 0.25 and
0.50 in. Generator section shape was a symmetric duuble wedge. The
generators were individual-ly mounted t a 15 deg angle of attack on a
flat plate upstream of a 35 deg o!resaion c6rner. The tests were
made in a wind tunnel at an average Mach TH=ber of 4.67. The bloundary
layer was turbulent-and approximately 1/4 in. thick at the test *tation.
Static pressure zeasurements were taken along the centerline of the
plate 9nd on the ratp face. Oil and dye flows and siblieren photographs
were used to aid--in determning the extent of separation. Yhe vortex
generators vere found to prevent separation if placed sufficiently close
to the compression corner. The unseparated regio7. in eacb c&se was very
narrow ir spanwise extent. Separation was prevented by increasing the
momentum of the bo-undary layer through the action of the tip vortex and
a secondary vortex.
N-iii
PREAC E
The semaration, of a sup -rSonicg fl-re;. in co~sxresston0- cornmers is
a iffiul deig prodblem of imoderni flightr vehicles. The l i'o
prevent senaraetionk would be an, obvious aid in the .. Sc"lono nressure
distributions and hea*t t;raxisier.- rates for fligt syst-ez ." Vo rte x
generastors have proved useful in ontroflng separatio-n in msub so;.0 and
transonic flows. -This renor preSets tE eltsotssofvt
-generators "to prevent separationr:. a supersonic flow4 in a-comressionm
corner.
The current study is one of a series, devoted to high seed flc
-searation end its prevenqtion, conducted at Applied Besearc~h laboratories
-of The University of Tenas a~t Austin. This work has been sponredb
the N'atal Air Systems Couand under Ant/Jt! Subcontract 271173141 Task B
with th Applied Phyzsics Laboratory, The Joh. s Hon-kins U.niversity.
The author would -lke to express his anpreciation t
-Dr. John C. Westkaemaper, Research Engi-neer, Apn-lied Researchs Laborwtories,
for his continued guidance and az;sistance. Special appreciation is also
due Mr. W1liam Creamer and r.j. Gary Perser for their assi stance in
the eP=-erim-mentaI wor.
F::
Ti-tO COE
ABSTIRAC Tii
tIM OF no-UGU i
M N -- CLA I M-xii
n-. :-?.Af4EFTA IP'E~S AND PRO-CEDURE 5A. Wina unneL
B. Flt Pl-ate amd Rfaup5
C Vortex Generators7
D. Flow Visudalizatlion Teebrdc-ue BL. instrumentat ion" and Proce dure
i.- AA E 'CTVX4 AND ACC-UACYA~Data Rz2ecton I
3. Accurpayr 12
A Sepaaion Region 'ethu Gxenerators
~) Seraration Pae ters 1
~,Thrz-ee-inerasion-al1 Fects
_j7 3ej:nratod 1oato
T. S I4ARY ND COMWUJ TW 2
-vs 29SERiE
LIST OF fCU2
A#pitia Pagc
Flat Plate and iRa= w. 7.M444!1. --Generatohr
Vortex fl-Enrarors
Corner
Snawise r-ress- nCeffiintsz for Ccnresan '
Corne
Sen eren Puo a! ca 4nrr-n P'ornieiSktho Typical--~. Oi Fl- for.z; %-emu-e F
t~~a %Centsr'-ire F-assuwe teffic ents for-40
Generator '1o. 1I i ForwardPtsio
Soritw 4 1e Prssr 'oiiintz for -
Gentrator n. a o~r ozif
Geea o:.CJI ForwardAo:
CGenerator fl.2 in. _orved Psto
Ceniterline Pressure Cof~i sforGenerator No n Forwar-d Posit.ion
Generator No3i heti oiion
1±5 -nerline parure rt forGe~praorIN.. It rard ro-sitlo
i~~b Soanwise Pesr ofiin- o-Gweerator No. inFrard Psto
--kp r1, of Oil Flow for- Genieratocris in1 ForardPON t-rA.
ix
-UISrFnmum-(Cont'i
13a ;eter4-e Pressurxe C-efflcients for V
-lenerao i'o. in Aft Plht Position
=~~~~-r I h- fC 'u lef* 2 _-eats for 5
Germe3 aItor TAO. I in Aft Bih Pstio
V~a Ceterlire Pres-sure Coefficients for51GeeatrNO. -2 in Aft RMgt Posi-tion
Soanwise PreA aire Coeificlents focr >Generator NoM. 2 i.ns~ Pitlit Position
i~a Centerline Pressure -Coeffictenlts for 5Generator No. ]3 -in Aft -Right P7iii
15b SonvisePro sr're o'~iensforGenerator No14. 3' in- Aft Eight Position--
1-6a sCenterlin-e Pro vssuxra Coefficients for 5Generator no. 4~ inj Aft Bi4ght Pstu
lab Snanvwise -Pro ware Coefficienlts forGenerator Nlo. 4t in Aft REit Position
11aCenter1i- Prssr Coetfi ~as for 5Generator N~o. I in -Aft Lerft Position-5
I.b bnifl-1se Pressure ..oernfcients -forGienerator No. I in AfM Left Position
l~da Ceterline -Pressurale Cweffliers for 5Generator NO. 2 -:A Aft LeftPsio
Snpanwie IPressutre COef ficl its for 46Generator No. 2 iii Aft Left Position
netxi- Pressu~1re ;!1'aticits=w for
NO. 3 in-No. Aft left Position
h91 ISn~ese Pressur-e CofiinsforGenerati-m N 3 AinM Left PAt-Z-
L0 Centerlin me T Ctaa.'....nvs forDGenertor -Ho *
4 4 inrI AmE-m 'oiia
~.eentr R. Iin Aft Left positio
F' -EL-cP-
FNOMI CLArIRE
a =free interaction angle (see Fig. )
A= aspect ratio
b = vortex generator span
c = vortex generator chord
C = coefficient of pressure
= ratio of specific heats =1,40
I = length
NI Mach Number
P pressure
' = dynamic pressure
i SUBSCRIPTS
= free stream static
o stagnation or total
s = separation
XEL
I. INiTRODUCTION
The separationi of a viscous fluid 2"low due to an adverse pressure
gradient, can present a major problem to the designer of a modern flight
vehicle-. In the siibsoii flight regimep flow separation may cause large
R modifications in pressure distribution and aerodynemic loading. The
supersoni? regime adds the prolem of high local heaing rattes to the
problem of modified loading. In 'both cases the performaric of an entire
flight Yehicle system may become dependent on local flow, condit-ions iff
separation occuirs. Since separation causes a vrariety of design problems,
it is useful to investigate the po tsibilities of preventing zeparation.
The Aeromechanics D~ivision of Applied Eesearc tabaoiso h
Utiversity of Texas at Austin is coducting an extended investigcotion
-of flow separation nnd its prevention. It is the ob~ective ofl this
report to present the results of experimentz using vortex generators
to prevent sepiration of a sup0ersonic turbulent bo'.mdary layer in a
Compreasion corner.
The separation of a fluid flow in a compression corner is
caused by the continued depletion of the 1inetdc energy of the flid
ztream catesed by an adverse pressure gradient, ConsiLdering the imper-
sonic flow of an inviscid fluid in a comression corner, a s i~le
oblique shoek wave originating at the corner turns the flow parallel to
V ~ the downstream wall. Wrhen viscosity is included, the inviscid free
streami becomes coupled to the viscous boundary layer throu_h a pressure
interaction.
Due to viscosity the velocity and kinetiz energy of a flnid vE y
17from zero at a solici boundary to the inviscid free stream value at thE
edge of the boundary layer. At a compression corner the high pressure
resulting frm the turning of the flow diffuses upstream through the
subsonic portion of the boundary layer. The already low energy fluid
near thev wall is thus forced to exoend more energy to overcome the
additional pressure force. if the pressure gradient is sufficiently
large, the boundary layer will reach a point uhere St can no longer
overcome the pressure force and will separate from the vall.
As shown in Fig. I the separation of the flow upstream of a
compression corner establishes a new effective geometry for the
- oncoming -flow, At the separation point the fAlow is deflected through
an angle smaller than the original turning angle. This results in a
weak oblique shock wave emanating from the separation point. The
separated flow reattaches on the downstream wall, where it is turned
parallel to the wall through a second oblique shock vave. Between
the separated fluid and the compression corner a low speed vortical
flow is established by the free shear action of the boundary layer and
the upstream directed presswre gradieut.
The fundamental aspects of flow separation have been invest-igated
by numerous workers with a view toward correlating separation parameters
with local flow conditions (Ref. 1,2,3). Li the area of preventing
separation the majority of the work has been in the subsonic and tren-
sonic speed regimes. D-iing the late 1940's and early 1950's workers
2
i=at United Arcraft Compnany demontrated the use of vortex generators
to prevent separation. primarily in subsonic flows. British inves-
tigators during the ite ! 9 50's extended the use of vortex generators
to the transonic flight regime (Tef. 4,15).
Tbe mechanism by which separation was prevented in these cases
consisted of re-energizing the low momentum bun-dary layer. By placing
s nmiber of smeli, wing-shaped- anes or generators normal to the srrface,
a system of stream-wise vortices is ploduced. Th*ie vortex motion SweepS
high momentum fluid from the free stream into the low mOmentwn boundary
layer as Uhe vortex system procee", dowstream. If the boundary ls yer
momehtum is increased suff'iciently,, the adverse pressure gradient willbe overcome and separation prevented.
The utility of vortex generators in preventin separatiou in
subsonic and tnansonie flows has been demonstrate i. Some previous work
done at Applied Research Laboratories of The University of Texas at
Austin vas directed towara extending the use of vortex generators to
supersonic flows. %itten in Ref. 6 measured th- drag of several
genermtor shapes but did not demonstrate their use in preventing sepa-
ration. The work of the present report was ,ndertaken to obtain
better definition of the floa behind a vortex generator placed upstream
of a compression corner.
The experimental program consisted of two rain phases. The
first phase -onsisted of defining the separation geoxetry prou,.ced 3by
the compression corner. This vas accmelshed throough the use of static
pressure me'-surcments, oil and dy-e flows, and sehlIeren photogranhs.
3.
Therenindr o th exeriefla os ccetrated on determin-ing the
effect of a single vortex generator on thO separation regions rm-tng1 the
same techniques as in the Vlarnt phaZS
_ 2-
4
rI. EXEIMEIA AM P_ _'
Wind - -n
ThIetwr efre nx7i! idt-eIo
The test P= s ere .perohed itante prssr in .-ind unne of5 t
th0dig he test program, u was 4.7 h santonarssued varieom6 to
2Alsadrn h etprgabtwscnrle aunt~ratll et
within ±3 psi during any particular test. The stagnatioteprue
range for this program was ia-* to 79F'with an average of 176 8 Fr.
Th-is was sufficient to prevent mistumre condensation in the wind tunnelA..
For each particular test the suppl.y air was heated to the preset valuep
using electric resistance- beaters located - 0 -tr-am of11 the stillijng
chamber. The stagnation temperatulre -was constat within +-sF during
any One test.
-Flat late and RAWn-
The tests uere conducted with the. frlat plate, rampfl combination
sn-own in F ig. 2. The plate ea Ame fr om Tinthcalr nad
was 17.5-1 in-. long and 6 in. wi-de. A sharp lead1 ng eage was prvided b
-~i5deg chamifer machined an the side oppst thV etsrae
boirejary layer trip-per strip, mwae of' 80-grit eme-ry cltwas mounted
approximately- i/2- in- from the leafine edge.
The ramp was made from mild steel and macbined to provide a
35 deg turning angle frc the flat plate. The ramp test face was
1.08 in. long and rose 0.625 in. above the plate surface. When-mounted
on the plate, the turning corner was 1.2 in. from the leading edge.
A rubber sealant was used to provide an airtight joint at -the ramp-
plate interface.
The plate was manted norizontall- between the sidaealls of the
wind tunnel, with the test surface coincident with the nozzle center-
plane. A rubber gasket was placed between the model and wind tunnel
wall to prevent air flow through the manting-clearance gap.
The model contained static pressure -;aps arranged in a centerline
and soanwise pattern (Fig. 2). The taps wexe spaced every 1/2 in.
along the centerline of the flat plate from a point 7.65 iv. upstream
of the ramp to % point 4.65 in. from the ramp. From that point to the-
amp the tap spacing was 1/16 in. The quantity of taps on this portion
of the model proved more than necessary for this study. Thus, only
alternate pressure taps, beginning with the first, were used. This gave
a pattern of measurements taken 1 in. &part from 7. into 4.65 in.
upstream of the ramp, and 1/8 is. apart from 4.65 in. to the rap
(Fig. 3). A centerline nattern on the ramp provided taps 0.1 in.
apart beginnLng 0.1 in. from the compression corner. The spanwise
patten- consisted of tbree rows of taps located 0.3 in., 0.6 in., and
0.9 i Trom the compression corner. Each row extended I in. on both
sides of the center!Lne with taps spaced 0.1 in. apart in each r
( ..3).
(Fi.... :
The plate was also fitte_ witt a therrlcople, located ahe-i
of the separation region, to nmnitor plate temperatu-re. A recess
m-acLned in the back 01f the plate and O-A pro ae space for the
pressure azd thermcoue leaa A cover plate provided protection to
this area f the free -trem flow.
The flat plate was provided with three slots An t _
the vortex generators. The centerline of the forward slot was iocated
4-.9 in. upstream -ot the ramp on the centerline of the model. The
centerl ines of t to slots " 5 i upstream-
of the razp. One of these slots was located 3d46 in. to the right of
the cente-line aind the other slot was 1/'2 in. t the left of the center-
line (rig. 3). All three slots were inclined at a 15 deg angle to the
plate centerlina to provide the vortex generators w-ith a- angle of attack.
The fourth slot seen in Fig. 2 was not used in this inmestigation.
C. Vortex Generators
The vortex generators are shown in Fig. I an. their dim-nestons-and desinations are given in Table T. Te ge.tors ere made of
mild steel and were a'1 of rectangular plexiform. Four differpnt
generators were used, with chor lengths of 0.5 and 1.0 in a-nd spans
of O.P2S and 0.C in. Tiche section shape of each generator was a sym-
metric double wedge with the maxitm thiekness of W-064 in. occurring
at the mid-chord oin-t. Each generator was providei hn a mounting
tab 3/1 in. iong, 1/16 in. wide, amd If' in. -deep. The generator was
m-inted in the desired slot on the plate using a therm plastic adhestve,
Crystabond 509. All of the geeraters were mounted at an ankie of
attack of 15 der.
rvi z
D. o- Visuali zai Techniques
Ext-ensive use was made o v'arious fl- isulization techniques
to help defime tn finw patterns. Schileren photographs were taken of
each configuration. using' a sinale-pass schlieren system. The I-J4-t
so -ewas a -tungsten, ribcn-frlanerit lamp. This system alow either
black -.Lid white or color photog-raphs, t:iough.n itms proram color-was
used exclu-sively due to its increased defbnition.
Adye injection technique was also used to deine the flo on
the eilate b-rfae and- ramp face. An oil soluble dye disslved z--In -tha
was -ected throieh a resu- e orifice using the pressure differeII
between the inside of the wind tunnel enul the at-osphere.
ue to the large velocities p resent in the flow, tht dye
in- .3tion proved to be of 'iited vate and subsequently an oil flowteehuie was employed. -mixtue of 30 weight raor oil and titani-m
7- ; -dioxide powder proved satisfactory. _The oil 'as Ad'ibed on the plate.
the tunnel started, Lnd the pattern allwed to form. novever, az in
the case of the dye iLnjecticn, the entraining nalture of the flow caused
buildups of oil and d:e. Upon shut-dowm cf the tunnel, the pattern ra-
distorted by the flow of excess oil. Since the tunnel geometry precluded
use of phoogra-phic equip ent during operation, the patterns -w vere
obser ved and skctches made during each nnin,
S. instrmentation and Procedure
The =bodel static pressures were displayed on three multitube
ma- rcur y imIeter boads a nd were renrrded ph--oto PphieaLy. The pi7- e
the-coupe -- trut "' was, recorded on a Hone,-ifl-Brown strip. chart
0
recorader, as was, z ge cutiut of- the virnd t unnel~ s4 nts'qr tlefp-~Ae t ure
mhrocoupnle. The stftgchadlber *re)y- z--hSJi- WTs wtha
bor&ntype pressure gug Vail vas rercorae onotogranh-dcaly
For any partilatettetiinh br pressuire ea
tecpezrature cotrllers L'r rr, ~~4n~=A ~ifc
ieas t;hen- staet'-4 and the stagnatlon- Dre ssure Al-we to tbiI1Ze
benfore the static p-ressures 'were recorded. S- 1-ce %,h vlt as no,
betted to recovery te;erue pror toechrn the pl ate teprte
-as m~onitored peri4odccll - T1he nreselected stasnation t-enratur&-e
~~~*Aa~~~~~~ the Feovr -e-rnpr'rxn's"IeP e t-o the usuall
high ambient t-emnerature- (spproximatelvy 1-0-WOar)p the temmrature
diffterence betwgeen the plate at thec star-t of a ru ad the recov.ry
te - e was sm~all. As & result of the short rn tixtes C anproxi-
tel25sec), the plate teant-e-rature increased l-ess than A15 F during the
run monitoyred. This was felt 'to be revresentative of the enti re test
program, raidng t-he s 4 'd nearly caiabatic. Duingthe earlypae
of the testing at least fordupAlicate rn'were made ai eachcni~
ration. Good repeataility,1 during hasealowei th UM- e of rns
er cooflizuration to be reduced to t-wo in the ltrstages of the
IF
~~~g - W m. -- - -
ppm-_- - Olin-
- ?~- Z __ -- ~ ---VMS-
plotz~tt-r - --
sloyre mea,.urericriLtS were also mj,4-e 1sing the schlieren photographs.
Separation lengths and shlock wave angles were measured by projecting
the 35M.sieso screen and us-Ing a Gerber Variable Scale and a
pr.t rAor
The static pressures were displayed on three mercury manometer
boards having least scale divisions of 0.2 and 0.1 in. The photogra~Phic
records were founid to be of sufficient claritY to allow interpolation
to 0.05 in. Due to the Low pressures being recorded, an error in
manometer reading of ±0.05 in. corresponds to an error of 6.6% in static
-'pressure. H~owever, by red-action of the data to coefficienlt
form., the
variation in C was less than ±0.01, w~hich was deemed sufficient for
this sw6dy.
12
IV. RESULTS ANDi DlSCUSSI1014
The phenomenon of separation Jis hiehly deDendent or- the type
of boundary layer present and the nA~itude Of' the pressure gradient
retarding the flow. A turbulent boundary layer was assum-ed for this
program based on the presence of a boundary layer trip en-d a caloulted
6Reynolds Numtrber per foot o'& 141 x 10 The bcindar ae asaooi
rnately 1/4 in. thick in the -test, area as deduc-d from the schlieren
photographs.
A. Separation Region without Generators
The centerl-ine and suanwise pressure distributions obtained for
the compression corner without the generators are p-iotted in coefficient
form in Fig. 5(a) and (b).
1) Separation Parameters
The centerline pressure profile shown in Fie. 5Wa S -P ypce,
of the separation profile for a two-dimensional t-urbule-nt boundary
layer. Tne point o.& sepaxation is generally taken to be the first
inflection point in the p essure profile. T.hin occurs I..7 in.
upstream of the compression corner. Reattactment of the flow on the
doianstrenm wall c'orresponds to tChe third inflection point in the profle,-
which is 0.4158 in, along the rampy frice. The middle infleurtion point
is the plateau pressure rise, which is -due to the defLection. of the
external stream. The plateau pressure coefficient ani the loction of
separation and reattachment are sufficienttoalwccltin f
several separation parameters.
13
Tna free interacti-;. on angle or angle between the free stream
direction and the dividin streamline at separation was obtained by two
methos. 1he plateau pressure coefficient end free strt ach num'jer
permi itted use of the oblique shock relations or the charts of Ref. 7 to
obtain the corresponding deflection angle. This method yielded a free
interaction angle of 0*-8.7 deg. By use of the ramp geometry and the
location of separation and reat achment, a free interaction angle of
-78.-- deg was calculated. Measurements from schlIeren photographs
(Fig. 6) showed a turning angle of approxinately C--=i deg, verifying the
calculations.
The separatton length (Fig. i) was found by trigonometry using
the free interaction angle and the locaticn of separation and reattach-
ent. This yielded a value of 1 =1.77 in., while schlieren photog-raphss
showed 1 j=1.9 in. A comparison of these values for free interaction
angle and separation length with data obtained by Gillette (Ref. 3)
showed good agreement. A comparison of centerline pressure distributions
with those obtai.ed by , and Sterrett and !-ery (Ref. 8) also
showed good agreement.
It should be noted that the peak pressure on the ramp fallshrt of tb-at pedi t ed by ir'viscid 4heory. This phenomenon hat been
n-t-e ed by- a nm-er o'P irvestigators and is attributed to the p-:resn- of
the side-walls. The low speed reierae flow region of eepara-ion is
nprevented fr-m svanuise venting by t-e sidewalls. This causes a targer
seoaraton area than for -an ,Lnbounded flow The result is
1&
ftee interaidon angle, Pa s'r _nsger sM-c1k -wave at nepartion, and a
smaller' finaL pressure rise. Work by 'Kaftai et all (Ref' 9) supports
this Iynothesize
2) Three-DimenSional Zff'ec-
The ~ ~ -spnienesr itributions plotted in Fig. 5(b) show
the three-dimenisiornal effects ire flowi. SoDanwise variations in
pressure were observdo the ram face, inictganoniomfo
-'o aid in determining the ;haracter of the flow,~ a series of oil and dye
flo ests vwas made. Te flow pattern obtained fyom these tests is
sketched in Fig. 7. The separation line was essentia'.lly straight over
approximately 60% of the center portion o.f the plate. Near the side'wafls
thesepraton ine bec~ne curved and joined the cureved separation line
of the sidewalU boundary layer. Inthe center pozrtion of the plate the
flow in the separated region vwas -directed Amif ormly upstream. Close
to the sidewalls a spanvise vdlocity component, directed outward from
the centerline, caused a curvature of the streemlines, which cu~linated
in a surface vortex at the separation line near each sidewall.
In t-he reatt-achmment region on the ramp face. a series of
reualy spaced radial fliow patterns -was noted. he centers of eachI
paler 'Vem ere located at the reattachment Line a.nd -were snaced at 1 i4n.
intervals in the snanmwise direction. From each -'enter the flow was
radil i alldirctios -it a small distance from +he cienter tile flow
"became more streammise, resulting in higl curved strearni-nes At t ze
spanwise points 4here two patterns Joined, a heavier oil floW resulted
giving the appeararce of a striation parteft. This striatio f h
eprarc r mio- whr th Oi pate- 'wa llz- Mr
FZ- r- of- n -- ---zt --------- nmdre--l ------
i panrWa regiuwee the oi vratern oas tessteist.ez l
rn eeralce invesgat usin vries ae~re cs ino ik. li
andieostgaDtinof ainon ea-thretberd rerlar j stratin
oth weracconuaed by rhea vkariation inw thessreandn hea tranie
rauates d Gilette (Fe! 3 prandershk en-in those (Rtef 11nvetoedms the
flowrn for i tret ratacn!as deflows zThe :r-'s of Sezte sIa's
isutn Ierfly attruedstoth creatofsafstemie otcsi
th bunar layer, due tor thetri- the vo at entraton -Maumi
Lad a re ergt orerTePinle arey V nsedere s three-X-
intesi~onl cetraeios oned en oth-eniirex tato-dimennctal fl
on. . the traghtnsn of te sertn Hn an the'-e is ren ofy atimp
caluted sc-er atin ramet- -es wia those ofer: inetigtos, h
A fe inrodutor codt ts othvrte generator,, i cuirtio
mad rtesuts ae iodr prncpa in exerh Srl~tia- iSad
r
ctcuit;-fVA -el- tmntr 4t-ef thMn ahnfec fth
ce-reaedr*vi" te senantir=.aita. Also,-a the fiou.- ptter
creete by th gennerarvdki an ±s eztonre y C=P x,
a dt~iedan~yss s beyond thesoe of the present ivork. Thereore
Only a quai~t&tive decp~nof the! itanrrn od -- are
n"-eicte tinlit sotowinI y on ions
pr ur akw o-A -fs:U
-ice ~ ~ ~ ~ F '!6 tL ic tip *r.- rnz leM cu-wrn
vidsm3a .Hrweveri-- dtnls Zbe piresent ivtitinit tecsre vfn
that ttie: -P . -te -e~ f -was- 1Anadamate todeterr'irc- the-. extent 01P
aptj~ the presence of a vorex gemnerator. Th-e spsnwiuse- pressure
-distrbu tions vere essenlal 4-- def-r~ the region= afffect--d bay the
Etnawatorz l~4j, i ) ineit into t.a met'h'----v~ by whchsemara-
e alnae were not
s~fcttto Ateme iihte tepar-at-Lr-t 1had 1-~npreven=ted, snen
prvo crteria e.x r 4 talva ~e~o. theexeniv ve of
ct iot atarfaca mattern f or -each vormrto as def innc- *Mh'-se
the~~~I prtdcinst betue Cteno. =.re---ie that the
thirl the: bol-ia laver is otvalid. Taaid isl vitn±-ii tile
-am lthe wras ta eE2I:~ .ee-arnii e#4 L4t--0e 4n - rr t %Ar rnI
arnL ~ -ys- ass mides ~ ~anna- hcrm r s walj-- f ilJ(~
I .1) ?orward Location
The pessasure distributions for the four- generators in the forward
location are sho'wn in Figs. 8- to 11. The center!lne nrofile for each
generator shows a slight lowering of the plateau pressure though theprofile shape remains similar to the typical separation proftle.
In the spanwisa profiles all of the generators sho.ed some
deviation from the no-generator profile. Generators I and 4 had similar
patterns though they differed in the magnitude of the variations in
ooh cases the dowmstream develop-ment of two pressure peaks on either
side of the centerline was-noted. Separating each peak and outboard of
the peaks were regions of low pressure. Generator I showed a more
distinct davelovment of these variations than did the smaller chord
generator 4. 3owevei, in neither case were the changes in presesure
coefficient greater than 0.2 f1m the no-geherator values,
Generators 2 and 3 also shcwed a similarity in spanwise profiles,
though again the magnitude of the variations vas proportional to chord
length. For s short span generators a single pressure Deak. developed
to the leeward side of the generator. This peak was noted to be
unsteady fIr generator 2. To the winbtard side of the generators a
;r-y- low pressure regio;i developed.
Te sketch of t oiI flow patterns obtained for these
configations is shown in Fig. 12. None of the fl7o tests showed a
1lW of oil streaniise over the entire ramp face, indicating separation
was not pr-evented by the generators in this forward location. Tne basic
ifl
pattern in Fig. 12 was comron tlo all four generators, with' only th-e
positicen ofL the surface vortex varying from one generator to- another.
For generators -I and k the Pattern vas as- shown in Fig. 12; for genera-
tors 2 and 3I the vortex was centered on the mlodel -enterline.
Coimo 'toal four configur-atioJsEi the lxpressure area near-
the centerline in the searwv se prof-ilIes. T-he shock expanston analyjsis
tor the portio of the generato in the free strem anaicate& velocities
at the trailing edgee co-,-mparable to free stream veloci-ty. At th utr
edge o f thbe bondr layer, the hihvelocity flow would intersect-* the
ram-- a-t tie reattacbrment line or hi~ther. Since thc largest va1riations
i n pressure appear high on t-he rampq, it is; felt thrat the press-ure trough
ini the profile -is a resualt of the high vilJocity f'low off the tr_ au
edge of t15he genzerator.
The, flow hlear the surface also has- -, relatively b,:igt velo city
rgion S.the trailing edge 'wu is -cc lcated by the presence of a
streawise tortex ndthe separation Ira n the corner II'rnei by
th:,e windward stge of the generator and-ti plaste, a etrearyise vortex
As zori-ead due to the nonuniform turning-ot the boundary-- th
inerator. N-Vear thit surface the Zlowfr roing fl!i it t- wed t
a greater angle than the faster gann f 1L at the cu-er edge or th
'Nomdarr ~ ~ ~ I -er Itii..al~m g t 'xuaar~ lsnwr txt
teform~ation -of a stenievorte% ne.ar- the z-riae -~
vorex act to $nce0 the maretiz f_ thte fluid near the surfa-e at
M the ~gene-rator t.rmg edg-e. - sseen inthe oil- a- th e flrW in this
region has s'affitient mc~wentim to penetrate the sepmatlw region
cauing a cuszi the sep-a-rati line inieer. wit the 14r--+m-
to everItualWy retard the flo. n"us '- sc-pflatvfl "i farh
o:fa attache&Ad owmtTremn fla;w is ukebytime i4Orstresam edue of tine
surfawce vortex seen -
The ress.re rcearrk v-ra-i n =e-as~- I 1. 'a face a&M
asszcat ivih shr~ i*-ave = creat b he Jreewr4 side 0"1 t!he
£enenor. ot gneraors and 3- the nnres-r-n peak tz vt~~eri
thdtaien e il s r m-- w ex e -tlip mon at14t- A
the ;hitc-Lrti~
z. e =---. *IaE ztni et~k
time------------------ searae
iu -3- wave vin ---- pasm ai----
te-
*~ rlatsyl ~f&SILTl. ALb-ir 'be Um
_ ~ tr ~Urt~ atthe tj SZti tWsm
-- e-s
-I 4"-- sbm.3n~-- --
essurc --- v- -eut in-
irg e- - _ I __
_ -- n ---. cse- ~ -mtS
as- fl --- t
sksq - ag ' -- = __
___ ~tX wr~$ ~ .2 _ it
Z- tnZal,- *- -SSE
Mkltr~s ne
-he -b ms tc= l!
a 5119 m s
mr- - fl a
*Z - - -
-mom~
_ ft = -
ZrM:-
= 7 T -=z 24e=- -
-=nd= IM
40= Liam_ 4 _lw- f
a. ___ -= Am - ~ -
kt z- --w -= ~ T ~ N-
-- 4'.
[ of rotation of these vortices seem to be dependent on aspect ratio and
the trailing edge conditions for each generator. For generator 2, with
the lowest aspect ratio, the vortices rotate in opposite directions.
The remaining generators have co-rotating surface vortices and -higher
aspect ratios. The oil flow for generator 1 indicates a stalled condi-
tion at the trailing edge, while for generator 4 the pattern indicates
the onset of stall. in these two cases the surface vortex to the leeward
side of the generator was further upstream than for the unstalled genera-
tors. Stalling on generators I and 4 was believed to be the result of
the high downstream pressure, the lonC span of the generators, and the
high angle of attack.
At the outer edge of the b)oundary layer the generator is
influeni..ng a free stream flow. The shock wave system from the generator
is thus able to proceed downstream above the s-..parated area and intersect
the ramp. The two pressure peaks indicated in the spanwise profiles
correspond to the leading and trailing edge shock waves. In this aft
location there appears to be little effect due to the tip vortex,
perhaps because of the reduced streamwise distance over which the vortex
has to act. Outboard of each pressure peak the pressure tends roward
the no-generator valucs, indicating the narrow region affected by the
ogenerator.
in Fig. 25 is shown a schlieren photograph of generator 1 in
the aft location. Note that the shock wave emanating from the separa-
tion point is no longer a single compression -wave as in the no-generator
24
configuration. The breakdown of the Single shock wave into several
coripression waves is felt to be the result of the highly curved --epara-
tion line near the generator shock waves. Due to the narrow spanwise
extent of the unseparated region, schlieren photographs did not show
any detail in this area. The photograph of Fig. 25 'u, tYpia f h
for generators in the aft location.
LIM'.V. SI44NARY M~ID COICIUSIQNS
An experimental -Investigation vas made to determine the
usefulness of vortex generators in preventing the sepaxation of a
supersonic flow in a coaipreszion corner. The Nort generators were
mounted on. a flat pla~te at two stations upstream of a 35 deg ramp. The
I: average Mach Number was 4.67 and the Reynolds Numiber was 14 minlion
per foot. The boundary layer on the plate was turbulent and approxi-
mately k14 in. thick. Static press~ure measurements, oil and dye flows,[ and schlieren photographs were used to develop a qualitative description
of -the flow.
The following conclusions were made:
1) A vortex generator, if placed sufficiently close to a
compression corner, c&a, prevent the separation of a super-
sonic flow due to an adverse pressure gradient.
2) lthe vortex generator epnergizes the low momentum fluid of
the bo Nay layer by an expansion process around the
generator body, in addition to the energi2-ing action of the
tip vortex and the st-reazwl.se vortex formed along the base of
the generator. For this investigation the effects of the
two mechanisms were ztA separable.
3) Toe region of effeetiveness of the generator was limited
to a very narrow region directly behind the generator
trai'ling edge.
27
receding page blank
41 The perfonnanrce of each' geneTrator 'was related more to. Chor
lngt-h them tO generator pA-.
E) Vaticai thr - onsi al -perturbations iert found in
the turbrtent reittaching filow on the rcp- in the absence of
thhe generators.
28
wuerer, UE., and F. ia-Vtonl. rtcMl -d r lon in ragn zpe-e-arligh, -A Revie o he State-o- -the-_Artc - "epr M
2. kaufran, IL. G., atl.. !2 eview of Hlypersonic Flo-W Sepa-ration anIdControl Chraterisetics, Untd . ~Ai r Force, Aeronautical.Systeins Division., TDR6-~S 4l~ 1962
Gillette, W. B., "Sceparation- MI suretents of Sunoersonic Turbul-fentSmBoudary La-yers Over Cosuression Corners,. Defense Research LaboratoryFenortb Ncz. 51 3(n-3 ECR-26)-, Defense Research~ Laboratory,The University, ofTzs as 4 ~rxs ul-y 19665.
1.Parcyo H1. 4H., and C. IM. Stuart, *k14thcods of Boundary Lwyer Cont.rolfor Posinvcurng and Mlleviatins Buffeting and Other Effects ofShok-Thduced Separation, " AUt A FirhWI-d- rmi Pat per 22, * e 15
~,.Ed~arz, ... a,-Free Flig-lt Tests of Vortex G:enerator Ccnf-igu-rations at -Transonic Suaeds," A.A.Z. TLech. Hote-Akero 282December 19462.
6, Thitten, J. W., "T--e- Drag of Vaie-!Type Vormtex Generators inComjressibe T'"' .tt FoDefense Research Laboratory Report'oM 561DR056) Defense Researc-h Laburamcry., The Universityl ofTexas, Akutinz Texas, Augt 196-7.
y "Equationa, Tables end Charts for Comressible Flo,"AAReport 113:5, U. S. 'Governmerntit Office, Wasri.-ton, D. "
1953aSterrett, J. Ri. an J. . .=~r~ Drie Separation Studies
for Two-Dimensional Wedges and Curved Suarfaces at Mach Numbersof 4.8 tvo 6-2, HASA Vfl--1ol4, February 19 -_,2
9. Kaufm-an, L. S., ea, t ninvestigatio of Hypersonic Fo
-- Separation and Coentrol Characteristics," 1nited S tate Air Force,*Research and Technology Division AM1.TR-6-lk, January 196,5_
1-0. Ginoux, J.* J, "Thle Existence of T"'ee-D-ime'si-la' PrurA.osif teReattachment of a Tuo-Dihns-na SupeGmrsonic Boundar vlayer
after S-eparaetion," AGAPD Report A' -3.
1.Rosbko, A.,and nL J icl~r~e, "Obseriatofls of Turbulent ReattaChment
BehtndC an. Axisym etrit- Downstrean-F&CiflS Step in Supersonic P10%"
A IMk Journal '46), June 1966.
"Th T, ±.A -.i nsionas- ±nterac,-1u n Of. a. Sockav with
Turbulent BOun!dary, Layepr," m eoatclQartery Vo I VI
()August 1966.
30
RE
VOT 1GMI BML-m-ao
r, in. in
[ Genert
Iv
V xk
2 I4
3r
__a
I
j
if - Iif
I N. -
N. - Iifif NifI
if
I
A Iif == - ifif
-~ V xifI -~ r __
* So 4if =
* if.,
- z-I - a
I- ifa~a
;@ *r.t
j
1r1 .4o Izr
Wa-
3- S =*-0 3 - XI-I1 4
~=~=~~1~ 0o I- -5I-'4
N
-C,
iet~
- Ir~
I I-
I
if
if
IEIc
IPIWillt
I0
00
-JDA
+ w
0
I LL
'A Zw
I -- L
a.
I I-
co -0
0c 0j - __ _
[0 BOTTOM ROW6 MIDDLE ROW[3O TOP ROW
[0.8Q6 __ _ _ _I
0.2 0%00D
.5 1.0 0.5 0 0.5 io 1.5INCHES FROM --iNTERLINE
FIGURE 56SPAN WISE PESSURE COEFFICIENTS FOR COMPRESSION CORNER
EL
'oU
UJL
'U
NOT pE ftl
I-<
'cCI-
'-'Li
~t r
I-- IIIi :r_ ;cm U
\r <
$4 I ~ U It ill
8
U-
o
'U.
C-4~
ais~irflia 1
-~~~~~ B01T- -ROWr~==--~ ~4a MJDL R07
0.8~
1 D
0 0
_ _ Io021 -~ u -t --0
too0 I 0o 0
;tciIES FR-pLa CE#TERLLNE
FIGURE ObSPAN-W'SE PRESSURE COEFFIC1- N FURGENER ATO1R N.I IM FORWARD POITION
10 I oo i
II--
O z
Ic I I
kip-
---- h z -------
SOTT-00_1 ROrW
A MIDDLE REOwo3 TOP ROW
IIsp I oA
I IF
40 41
-OIM IF
FNE X14 W0 2- ,701 4C
0
Ii -
I t! I I I do c
. . a
r_- -
acs 3-%
Id, Iilkl
004
a T OPr
a-M -..W
MEr-
.I---I I 1iI -I "w________ I -_____ -_ i "______I
_ _ _ _ i l i
.- 61 ,.= _= _. .
_ I i F -t +-=,-- i- -i o o - i - o iI- - i
A ? I a _ _ _ _ _a
--V
_ = = r *. .
+I
+.II
I I0
S I *,I,
I I I II1 I Ql0_ 1
44
Iii - i <' j i~S I i I ai I' I I -i I I_ i 1*...*j I i I '--
Mai
i1 1 ' 0': €-
£ _3Z lil -
i = i =0i-
- I
FR
4-9f1 1 11
a-r
M Af P _
RIS
Iir.
~L -L
Iu U )
YIr
jj ILU7 LU
Uz
zz
II-ja
I0 0
to-
__I 1 __UJa
W
IW £
ZwJ2
I I __I I
ujw
ira:
OD10 a:
I - ___+
oan
o BOTTOM ROW-CtA MIXE ROWo TOP ROW
10-5 in. SPAtI, O;. iHORD
0-8.
cp
0.2 000 0 0 00 a
0T. 00 01
a25 10 0.5 0 0.5 1.0 1-5
INUiES FROM CENTERLINE
FIGURE 13kSPAN WISE PRESSURE COEFFICIENTS FORGENERATOR No. I IN AFT RIGHT POSITION
(:l :)ii° lt i~ '
01 i 1~I43I 96
I___ i YS
I -er
- - I Ia -
I I 1%'
1 1 18'____ _ __ ___ ___ __ -
ce z W SI I p ~ jo
_ _ I _ _ _ _ _C0
_____ _ I ____ _ I 1inI I t EL
cc X IIn-0 LuE
x I I7:
co ac
-- f iizT-
"13
QA'
1i(
o oSoI-IA4-OW
A ;
' IGUE lbL5 spA1-.0 oa I- j .
OCE FO CE RIK
tIGUR 1 0b
~SPANWISE PRESSURE COEFFICIENTS FORi ~ GENERATOR No. 2 IN AFT RiGHT POSITION
2E a
_____ _ - _ _____
* qa
3 S 4 - 7
1l X
S -1
* I S0=£ = = -
= 45
I -._ I wt ~~ii
iIL
s R N cm W -m
La
- I I j = U
3- -1- -S = I l,
t i & i mi
-- i ao I i I
a I!37- _ 1 ^o 4, !3 k
3a o
I G I b
cvj__V I±4j
FIURA S
SPAN W ISE FPtSSURE COEFFICIENTS FORGENERATOR No. 3 IN AFT RIGHT POSiTION
I
~ - - E
- aa
3a
3
~- w - .~ a
N
I3
9-~ . ~- .5-
I *
3 5
I-'zoI
g I - * -~-~ IL
I I1I I 3
3 3 - jjr
:Ziii~ f -~~= a -. ~. ~
I
I
i
-CSC
3 3 3
- -1 3
tat
I ~I I~
3 a 3~ II J -I
= I
K 3 £ I
a -
hi
NJ ztu
*-r-----------....-..---.-.........A: I I ULu1,
=4 1 I ~ II I~iI_____
01 11 '41 I-F
I
________________I oIJ
4 4 .g
C-
U
SPAXC C.;
C ra
lif -- ft In 041__________
a = = 3
GEEAO No t INATRIT1MT0
IF-~mar
UULM
3 me
-4--- .-
ft -
co-
0 BOTTOM ROW6 MIDDLE ROW
E3 TOP ROW
1.0
0.5 in. SPAN, 0.5 in. CHORD0,8 - - ---....
0.6 -0
a 03
0i 4 -'- --
0 00010
S 1 - 0000
1.5 1.0 0.5 0 0.5" 1.0 1.5
INCHES FROM CENTERLINE
FIGURE 17bSPANWISE PRESSURE COEFFICIENTS FOR
GENERATOR No. 1 IN AFT LEFT POSITION
5-
1 00100
IU 0
11.0
-7 w
o LLW
-44*
__If
goo-~c 6 0 0
CL
040
0 BOTTOM ROW& MIDDLE ROWo TOP ROW
1.0 -
IAj0.25 in. SPAN, 1.0 in. CHORD
061 . .... A A ,
c I 0 A
C p 9 1
0.4 O '0 ' I -
0~b~ a o o Oo10130b 0 0%0 a 0 0 000
0.2 - t-a------.-
15 1.0 0.5 0 0.5 1.0 S
INCHES FROM CENTERLINE
FIGURE 18b
SPANWISE PRESSURE COEFFICIENTS FORGENERATOR No. 2 IN AFT LEFT POSITION
60
LttLj Q LLO
Iu LL LL 6
ILIxI
Qu10flf
+ +o-
I-
I
0AO
o BOTTOM ROW4 MIDDLE ROW
o3 TOP ROW
0.25 i 1. SPAN, 050 ifl. CHORD0.8- - jillQ6 - ...
06r - - 00a
C p a 0 6 4
046
! ~ , 4 6 1o O"
0- ---- o%.. 000 0 0 00 lu -'
lasL10 0.5 1___ D_____
INCHES FROM CENTERLINE
FIGURE 19bSPANWISE PRESSURE COEFFICIENTS FORGENERATOR No. 3 IN AFT LEFT POSITION
62
F
100010 10o
orj.~~L 04- N1
at II
I I zwf-ic U
w &D
-i __ 1 10.
II
O BOTTOM ROWA MIDDLE ROWo TOP ROW
0.50 in- SPAN, 0.50 in. CHORD0.8 -_ _ __,__ _
cBpa Q4 13
0 0 0
no~0.2 0, 6^ _
020
1.5 1. 0.5 0 a s 10 1.5
INCHES FROM CENTERLUNE
FIGURE 20bSPANWISE PRESSURE COEFFICIENTS FORGENERATOR No. 4 IN AFT LEFT POSITION
C-i
I,/
I~OIL
II
cc z~z03C j; I-
dc
ul 40L0 r
-SL
5--
a I-
J &-l
oD
I-x
illi
S U -
lu tu
I-W
4C
YAI
IA
wvw% 7simm