Aircraft Propulsion

APPENDIX - SYMBO L S

_ : b equivalent sl ot width Ct , C2 co nsta nts _ di s t anc e between ce n te rs o _ adjacent i m pi ngemen t nozz l e s ep sp ecific he a t at co n stant pr e ssur e _:|, D d iam eter : fr i ctio n factor i_ _ ; ' g gr a vi ta tional co n stant !i " h heat-transfer coefficie n t i .

L

" J m ec h an i c a l eq _ , _ val e nt of heat i k th e r ma l co n d u ct i vity L surface length from le_ tdi ng edge to trailing edge LE leading edge

M ( p V ) c / CoV ) G

.... P presSure ;, . Pr Prand t l n umber _i - , _I!

Re Reynolds number _, s slot width I , i : ,- . .. " . " T te m perature - - , T average temperature , _ ' ,' ' _ ' ' _ . __ TE trailit l g edge ....

t wall thickne s s ' V velocity : ' W w eig h t flow X dist a nce f ro m downstream edge of film cooli ng slot _ ' " ' x _ distance along sdrf a ce from stagnation poil _ t Z i m pinge m edt nozzle to plate spaci ng thermal effectiveness 8 1 8 an gu l a r d is place m e nt f rom s tagnat i o n poi n t o n le a di ng e d g e , h re c overy f actor p de n sity t em perature diff e re nce rat i o _ _ G "_M TG - TC _l Sub s c r ipt s : b equiw le n t slot width C cooling air C o, _ v cor _ cc t l nn e effect i ve ..... fc film co oling • G g as h aydraul ic i i n or inside , , _ l e lea din g e d ge M m e t al o o u to r o uts i de T t o t al //

REFERENCES

1. Esg a r , ja c k B. ; Coll a day, Raymond S . ; _ nd Ka ufm a n, Alb e rt: An Analysi s of the Cap ab i l ities and L imi tat i o ns o f Tur bi ne A ir Coo lin g Met h o d s . NASA T.N D-5992 , 1970. !

2 . St e pka , Francis , S . : Considerations o / Turbine Cooling Sy s te m s fo _ Mach 3 -- Flight. NASA TN D-449 1 , 1966. i 3. Gold s tein , R. J. ; Eck e rt , E . R . G. ; and R e t t n _ e y , J. W . : Film C o ol i ng with . _ Injection T h rough a Ci r cular Hole. Rep. H TL-TR-82 , Mia a esota Univ.

_/ (NASA CH-54684) , May 14, 1968.

, 1_ 4. Metzger , D. E. ; C arper , H. J. ; a n d Swank , L. R... Heat Tra nsf vr with Film "i! C ooling Near Nontan g ent la l Injection , _ lots. J. E ng . Pow e r , vol. g0 _ no. 2 , _ Apr. 19 68, pp. 157-163.

5. Metzger, D. E . ; an d Fletc h er , D. D , : _ ur _ ace Heat Transfer Immediatel y Downst r eam oi Flush , Non - Tangential L _j ec ti on Holes and Slot s . Paper 6 9 -523 , A _ AA _ June 1 9 69.

6. Goldstetn , R. J. ; Eckert , E. R. G. ; Ertksen , V. L, ; and Rar n sey , J. W. : Film Coolin g Follow i ng Injection Throug h Inclined Circular Tube. Rep.

HTL-TR-91 , Mi nn e s o ta Univ. (NASA CR- 7 2612) , Nov. 1969.

7, Moffat , H. J. ; and Key s, W. M. : The Turbulent B ount lary Layer on a Porous P la te: Experimental H e at Tran sf er wi th U ni form BloWi,g and Suction. l _ ep.

H1VfT-1, Stanford Univ. , Au g . lg67.

8. Simpson , R. L. ; Rays , W . M. ; and Maffat , R. J. : The T u rbt t l _ni Bo u ndary Layer o n a P orous Pl ate: A n Ex pe rime nt al Study of the Fluid D _ uamics With I Inj ect i on and Suct i o n . Rep. HMT-2 , Stanfo r d Univ. , Dec. 196 7 .

9. W hitta n , D. G. ;Kays , W. M. ; a nd Moffat , R. J. : The Turbule n t Boundary Layer o n a P o r o u s P late: Ex pe rime n ta l Heat Transfer W ith Variable S ti ct i on _ Blowi ng , and Surface Tempe r ature. Rep. _ I M T - 3, Stan f ord Univ. , Dec. 196 _ ' 10. J Ul ie n , H. L.; Kay s , W. M . ; a t td Moffat, _. J.: Th e Turbul e nt Bo t m dar y Layer on a P orous P late: An E x per i m e ntal Stutly o _ the Effbcts of a F a vorable Pres ....

sure Grad i e n t. Rep. HM T -4 , St a_ or d Un iv . ( N ASA CR-104140) , Apr. 196 @ .

lI. T h ielbah _ ' , W. H. ; Rays, W. M. ; a n d Moffat , R. J . : The Tu _ 'btt| e nt B ou nd _ ' y Layer o n a P orous Plate: Experimental l _ ea t Tranaf _ r W ith Blowi ng , SucUo, , and Favorable P ressure Gradient. Rep. HMT-5 , Staa f orct Univ. (NAI _ ACR- , 104141) , Apr. 1969. ¶ 8 3 i 1 2 . H _ Tcal _ , Pete r ; Lee , David T. ; Gauater , Ja m e s W / ; a n d L l v ing ood , John N . B. : E x peri m enta l Flow C h a r acte r i s t ics o f a Single Turbulent Jet I m pi ngi ng o n a Flat Pla te. N ASA TN D-5690, 1 9 7 0.

1 3. Gaunte r , Ja m e s W. ; L i vingood , Jo hn N. B. ; an d H r y cak , Pet e r: Surv e y o f L i te r atu r e o n Flow C ha racter is t ics o f a Sing l e Turbulent Jet Imp i ngi ng o n a Flat Plate. NASA T N D-565 2, 1 9 7 0.

1 4. Kauf m a n, A. ; Berry , T. F. ; a n d Me l n er s , K. E. : Joi n ing T ec hniq ues f or Fab r icati o n of Co m posit e A ir -Coo le d Turbine B la des and V an es. To b e pre- sented at the ASME Ga s Turb i ne Confe ren ce a n d Produ c_ s Show , H ousto n , T exa s , Mar. 2 8-Ap r . 1 , 19 7 1.

15. Es g ar , Jack B . ; Schu m , Eugene F. ; and Cur r e n , A r thu r N.: Effect o f Chord Size on W eight and Coolin g C h aracteristi c s of Air-Cooled Turbine Bla d es . ! : i ' NACA TR 1354, 1958. !_ 16. Clark , John S. ; Richards , Hadley T. ; P o f erl , David J. ; and Livi ng ood , John i " N. B. : Cooling P ressu r e and Fl o W D i stribut i on Through an Air-Cooled Vane t_ for a Hi g h-Temperature Gas T u rbin e . NASA T M X _2 028 , 19 7 0. i 17. Calvert , Howard F. ; Cochran , Reeves P. ; Dengler , Robert P. ; H l ckel , Robert O. ; and N o rris , James W . : Turbine Coo l i ng Re s earch Fa m ' lity. NASA TM _,!

X-1927 , 1 970. I ' 1 8. P oU_ .ck , Fra n k G. ; and Hicke l , R o bert O . : Surface Temperature Map pi ng With I_ Inf r ared Photographic Pyro m etry fo r Turbi ne C ooli n g In v e stigations. NASA _l_ ' _. TN D-5179 , 1969. , .i_t " 19. Yeh , Frederick C. ; and Co e hr an , Reeve s P . : Co m pari s on o f Exper im en ta l and _i Ideal Leakage Flows Thro ug h L ab yrinth Seals for V er y Small Pr essure Dif- ii ferences. NASA TM X-1958 , 1970.

20. Yell , Frederick C. _ Poferl , David J. ; Coch r an , l _ eeves P . ; a n d l _i chards , Hadley T. : Ai r / l ow Study o n Air C o oled Disk and Blade. NASA TM X-2171 , i 19 7 1. I 21. Lesco , Da niel J. ; Stttr m a n , 3 o h n C. ; a nd N ieberd ing , WilLia m C. : R ota t i ng il Shaft-Mounted Microelect r onic Data Syste m . NA SA TN D-5678 , 1970. i, i 2 2 . R ohde , J o hn _...; _ ic har ds, Hadley T. ; and Mt _ tger , Ge O rge W. : Discha i_ ' g_ _ : i _ Coefficie n ts for Thick Plate Orifices Wi th App, . .oach Flow P erpendiCuta _ ? and ,_ / Inclined to tlie Orifice Axis. NASA TN D-5467 , 1969. i :

t

i

I

_' '. . c, .t._,_i_ I ...... _'" ......... "-' --ii_ _ ......... ", . "' . .............. • --, ,-- , " • 'i '" ' • - , 23. Chupp , Raymond E. ; Helm s , Ha rold E. ; McFadden , Peter W ; and B r own , T.R.. Ev e du a tt en of I n_ eru a l Heat-Tr a nSfer Coefficie n ts fo _ Impi ng e men t- _ " Cool e d Turbine Ai _ :£oil s . J . Air.e.a ' _ t , vol. 6 , n o. $ , M a y-June 1909 , 208.

2 4. G a rdou , Robert; and Cobonpue , J ohn: H eat Trans f er Betwee n a F la t P la te and Je ts o f Ai r Impin4 _ tngon It. Interna t io n a l Development s i n Heat Tran sf er.

ASM _ , 19 (3 3, pp . 464-460.

t 25 . Burg g raf, F.; M u rt aug h, J. P. ; a nd Wiltso n , M. E.: Desig n a nd A _ l y sis of Cooled Turbine Blades. P a rt I. Le nd i ng a nd Trailing Edge Co _ lgurat in n s .

Hep. R 68AEG101 , G e neral Electric Co. (NASA C R - 64 515), Jan. 1 , 196 3 .

.... 2 6. TheoclituS , G. : Heat- T ransfer and Flow-Friction Char a cteristic s o f Nine Pin- Fin _ r f aces. J . Heat Tra n sfer, vol. 88 , no. 4 , Nov. 1966 , pp. $ 35-390.

2 '/ . Ka _ma n, Albert: Steady-State Str es s Relaxation Anal Ts is of Turbi n e Blade C oo ling D e signs. NASA TN D- 52 8 2, 1969.

2 8. K a ufman, Albert: Analytical Study of Cooled Turbine Blades Consid e ring Com- bined S teady-S ta te a n d Tr ans ient Condition s . NASA TM X-1961, 19 '/ 0.

2 9. Anderson, R. D. ; and Nealy , D. A. : Evaluation of Lam i nated Porous Material for High-Te m per a ture Air-Cooled Turb in eBladeS. R e p. EDR .4 968, Ge neral MotOrs Cor p . (NASA C R - _ 2 2 81) , Jan. 16 , i96 ./ .

It, , 30. Neal y , D. A. ; An derso n , R . D. ; and Hufford , A. A. : D e sign and EXperimental Evaluation of a . Tu rbi ne Vane F abricat e d From Lamin _ tted Porous Material.

Rep. EDR-6 2 9 _ , Ge n eral Motor _ Corp. (NASA C R- '/2 6 4 9) p July 31, 1960.

.... 31. K les si g , C. E. : Investigatio n of Laminated Convection-Fihn C oo led Td rb ine •. : •: S _ to r Varies, Rep- AGC- 2 1 00 - 1_ AeroJet Liquid Ro c ket Co. (NASA CR- , '_ ' "/ 2 50 6), July $ 1, 19'/ 0.

_ 3 2 . Lo m ba r do, S. ; Moskowltz, S. L.; and $elmure _ S' A. _ Experi m en ta l Res u lts of a T r an s piratiott-C oole d Turbine Operate d in an Engi n e for 150 H OurSat i 25 00 ° F Turbine Inlet Te mpe ra nC e. P ape r 6'/-GT- 2 9 , AI _ IE , Mar. 196 '/ .

3 3 . Moskowitz, S. L. ; and Schober , T. K. : De sign and TeSt of a Small Turbine at _ _ 2_ 00° F With Tran s piratio n Cool e d Bladi ng . P aper 6900 2 6 , SAE _ Jtm _ 1069. i , _ / 34. Minth _ al _ , Joseph; Wys0 ck i , Dennis A. ; and H O Ilenbe ck , Bt , la n L. : JOS-W-1 3 _ ' .ng ine Eval na tio n of Curtiss-Wr ig ht TranspiratiOn-C oo led _ ui'b in e Blade _ and Va r i able -Geom e try T u rbifl e ConC e ptS. Rep. NAPTC-ATD-129 , Na , / al Ai r P ropulsion Test Cen te r, Nov. i 967.

t P 36. Wheeler , H. L. , J r . : T ra nspiration Coo l ing in th e High Te m pe r ature G u - Turb i ne. Be nd ix-Filter D l v. , Bendix Corp . , Mar. 1964.

36 . Cole, Fr ed W. ; P nd den , James B . ; and S penc e r , Andrew a.: Oxida t ion Re si stant Materia l s for Tr a nspiration C ooled Gas Turbi n e Bl a d e s. I. Sheet flpeeimen k reenfng Teats. NA _ IACR-950 , 1968.

3V. Cole , F red W. ; l _ dden , James B.; and Spence e, Andrew R. : Oxi da tion Res i st- i ' I ant Materials for Transpirati on Coo l ed Ge m Turbine-Blades. 1I. Wire Speci. i m en Te s ts. NAJ _ CR-1154 , 196 8 .

i 38. Kaufman, Albert: Analytical _ ud y of F-loWR e dtlc ti on Due to Ox idation of Wire- i Form Porous g heet for T ra nspiration C oo led Tu rb ine ,bides. NASA TN D- i S001 , 1969.

i C 8O z /,

HETHODS FOR TURBINE COOLING

HOT GAS ..

COOLING A!R (a) C ON VECTION C OOL I N G ( b) I M PI N GE M F N T C O O LI N G - , , :, (c)r . ILM CO O LING (d) TRAN, _ PIRATION COOLIN G C 8 - S t ., 8 2 8 ", Fig u r eIII-1 i 7 ,

_ TYPICAL AIR-COOLED VANECONFI G URATIONS

' _ , _ , , -- RADIAL AIRFLOW %%

CO NVEC ,, O N / ' _ , _t_ o c, , ,, M m ,- CONVEC T, ON

, co= . , ,, , ,,, co=

r !! : .

, i , ' (a) CONVE C TION -, IMPIN G EMENT - , ANDFILM-COOLE D BLADE CONFIGURATION .

" H ' _! T RANSPIRATION C O OLED (WIRE - • : _ WOUND POR O U S SH : ,, _ RADIAL AIRFLOW ' : INTOCHAMBERS i (b) T RANSPIRA T ION - COOLED [ILAI) E CONFI G UI _T ION.

Figure 1II-2 8'7 \ # POTENTIALS OF COOLI N G HETHOD S t' - CONVECTI On / RELATIV E CO O LANT ULL - FLOW / COVERAGE RATIO I - 7 _.. / _., . , , . 1__ i_i c - - _ , ^ ,S m A T I O,

0 , I ,I I :_ =

2000 TURBINE 25 00 _ °F 3 _ )0 t ' LOCAL INL ET T F _ P , cs . s _826 Z Figure III- 3 IHP R OVlN G CONVECTION COOLING

_o - _ _

OF OF , IN. t R _LA TIVE 2 , , -0.05

,L ow i 'o 1 ooo

RATIO 1 ..... i _.... I::: _ -- .03 COOLING INL ET O UT _ II _ E WALL AI R C O OLING ML e TAL THICKN[SS, : THERJ _ ,AL AIR TEMP , , t EFFECTIVENESS , TEMP o TM ' r / CONV TC° i c _- ,,, a - '.

Figu r e III - 4 8 8 . i !

:i o , / !

'' • .' a z 1 I F CHORD SIZE AS FUNCTIO N OF COOLANT REQUIRED FOR CONVECTION-COOLED BLADE %" 21GO ° F _ PG " 1 50P S IA TRANS V E RSE FI N S - , .O A - TC , I , 12 0 0o F _"_ ,03 I _ PINFINS F L OW T C, I o _ W cI W o J_ o o o .

, ,01 =f i o o o O " ' .'

, 1 . 5 l. O , 7 5 1 , 5 l .O. 75 _, CHORD SIZE ,I N , cs- s_ e 3 _ F ig ure ITI - 5 }!. .

_ , _ AIR-COOLED TURBI N E VANE T RAI L ING , , - L EADI N G EDGE-,, , _ / EDGE / COO L ING ' j - _' - .4_ . _ j A I R INL E T _ , __- Ti PPLA TFORM C O N VECTION- C OO L ED S PL IT TRA IL I N G EDGE-, __ ',,. . LEADING EDG E PRESSURE SURFACE Q_i : _ r H UB PLATF ORM FIUVI ' C O OL I NL_ SL O T-"""__ . . _ i _ , (:-7( , .. ' ,,,T COOL I NG AIR EX IT c_ . _,,_,,. ( L EADI N G EDGE ON LY) F igure liT..6 AIRFLOW D I STRIBUTION TEST APP A RATU S CROSS SECTION OFA IR -COOLE D VANE i_ , V I cs . s _932 AIR F LOW COLLECTO R AS SEMBLY ..... _ ' Fi g ure III- 7 ' ,

, it

2500°F STATIC CASCADE t_ A TOTAL PRESSURE o TOTAL TEMP e o STATIC PRESSURE _ t OBSERVATION PORTS " _ GA S FLOW c s - s s 83 o F i gure III - 8 9O : 1 CASCADE TEST SECTION WITH COVER REMOVED " " . 7 ' / " C O O L ING AI I: I i i OBSERVATION " ': .: c - 69 - zz 4 7 OBSERVATION . ' cs - s68 8_ G A S FLOW PORT F i gu r e T H- 9 ' ¢ , . TEMPERATURE MEASUREME N T BY I N FRARED PHOTOGRAPHY " R E FERENCE T E MP 0 I N . I • / _ THERM O C O UPLE LOCA T ION _ l , , I t . -" . . . .... O F _. // - FILM COOLING i" t / H O LES i' / • '; H ' , , ... .--TRAILING

Jrd

ISO T HERM MAP c s - s 198 z Figu r e I H -IO

!

¢--.'. ............... _ , _ ,, , _ ,- _ , t ...... ' , . . , , .. : ........ :................................. _ 1 = i f " -_. _ .u-- i i' " "---i-- , , • q \ / COHPARISON OF EXPERIHENTAL AND EXTRAPOLATED VANE TEHPERA T URES-FOR CASCADE TE S TS

Wc/ W o -O.O 5

14 0 01,.-- b , . _ / 'q:P " P O 4] _ P S IA

" " tc , i - _ F

TTE _ ' V Am 1_ I -- / Tin " 15_

1000 I--. SUCTION __ _ ,..__ PRESSURE__, j PG " 18 PSIA -

_ l- t Eo,.. _ _ , , or e t c t . _ °F

. . 4o0 1 1 v , i ¢, I I t-I I I_ ' l I 1.0.8 . 0 .4 .2 0 .2 .4 . 6 . 8 1 . 0 x / L CS - 568Z5 Figur e III - ]3 | P COHPARIS ON OF EXPERIHENTAL AND ANALYTICAL i

VANE TEHPERATURES FOR _ TESTS '

_ _G " _ F .____

TC.I" 78o F ANALYTICAL , 1800 -- W_ / W G • 0.0 3 5 o EX.PERIMENTAL ' VANE " _- i TEMP, 1400 T.M, 1200 • o , " / _

"F

1000 _ °SU C TION _ I _ ._PRESSURE _ E .

8 00 '- I I SU R FA C E'I-,I SURFACE - _ 1.0,8 .6 . 4 .2 0 . ? .4 . e .8 l .O , , x / L C8 - 5683 4 i , Figure Ill - 15 j_ , h i t ,, EXPERIMENTAL C OOLING CONCEPTS !_ (a)FI L MA N D (b)FI L MA N D .

CONVECTION IMPIN G EMENT k COOLING CO O LING _ (c)FULL-COVERAGE FI LM COO L IN G _,

F f

(P O ROUS SHEET METAL) I j ' t " " __: , ' 1t , , _ Fi g ure III-16 c s - s_e 83 1 I I 9 4 ( / t \

, 5 t

---- -- r -- q

..... tk 1 \ ALLOWABLE GAS TE MPERATURE AS FUNCTION OF COOLANT FLOW RATIO ' _ M"1800o F t T C, i" 1000o F 3 000-

_ o0- ' _ V A N E e

., . , (]AS _ ,TEMP , 24002800 _ __ " VANE A : ,I: : :

=o . J

' 0 . 0 2 .04 .06 .0 8 . 1 0 , 12 ! ' cs.s 6 sz4 ,, COOLANT FL O W RATIO, W c I W G Figure III - 18 EFF E CT OF OXIDATION ON COOLANT FLOW : _ WIRE-WOUND TRANSPIRATION COOLING MATERIAL - . , T M -1800 ° F 1 u _ O F ORI G INAL C OOLA NT _ _ " ALLOY A- &_ Cr - l & 5 F e. .gMo - 1. 5Co-Ni (HAS _ LLOY X) FLOW 40 _ _ OY B- 1 5 Cr 4AI-1Y-Fe (GE 15411 0 1 00 200 : _ 100¢0 _ 0 e 00 EXPOSUitE TIMI _ HR cs-s 6 sz 3 ,: i Figure III- 19 : " :i II i n

NTI -19455

L V. - COMBUSTION

" Jack Grobman , Robert E. Jones , Cecil J . Marek , :

andRichard W. Nledzwleckl : =

!

The t r end t ow _ t _ l hi g her temper a ture s and preS e lur _ s in prima ry combustors i has necessitat e d research on liner coolin g a nd development o! shorter comb u stors i to reduce t h e cooling-air _ equ l remen _ . O ne _ periment _ t _ hort-leu _ h combus- tor h a s demonstrated _ ficiency , pressure l osg, a _ d p _ ttern factor eq u a l to _ hose i of a co n ventional ccmbust o r which is 50 percent lo / _ ger; fur t her work is needed to improve its altitude reltght per f ormance. Another experimental comb U stor has , been operated to an outlet temperature of 3 6 1 6 ° 1 _ with 100 p _ rcent effi c iency, low pressur e loss, and a pattern fact o r of 0.11. T h e experimental short-length com- bustors p r oduce v e ry U tile exhaust smoke. R ecent tests reveal that one of these combustors emits only one-sixth as much nitric oxide as a conventional combus- for; this improvement is attributed to the reduced dwell time lu the e ombust b r.

This report revi e ws some of the present-day requir e ments _ or _ m t _ rbiue combustors and shows how the l atest en g ine operating conditions are affectin g ' , the problems in combustor design. The research approaches bein g u s e d to Solve these problem _ are discuss e d. By way of review, the vario u s criteria that are used in the de s i g n of _ as t u rbine comb u stors are c b nsidered. Combustion effi- ciency, total - preSsure loss, durability, e _ it _ emperature proflle _ an d altittMe r e light are importa n t for obvious r easons. Combustor si z e and weight are ira- • / portant because they influence the overall weight _ the engine. C o mb u stot' len gt h ' ,' affects the turbine shaft leng th ahd the bealdng require m e n tS. Reducir / g combus - for len gt h al _ o reduces the amount of air required to cool the combttsflon li n er ' by reducin g liner surface ar _ a ( r o f . 1). Minimizing exhaust emissions such as smoke and g aseous poltu _ i _ a rel a tively t _ ew require m ent . AI _ these c _ teria are used as yardsticks as the results of recent NASA Lewis combust _ r _ s _ a _ ch are discussed.

Figure IV-I ill u strat e _ the trends in c o mbustor operatin g condit[ons with time.

, / These data are repres _ tative of engines th a t are eit h er o pef _ tto _ o r in develop- ment. The d _ t _ indicated for each e n _ dn e was taken to be at t he time o f l _ mili- tary qualification test or other qt l aiifyin g acceptan c e t _ st. The d _ tes beyond 1970 _i . I are dates esti r dated for en g ines in the development sta [1 6. All da t a shown are for _ , \ p S ea -l e v e l takeoff "Conditions. It i s ap p a rent fz'o m this figur e that th e r e h as b een a Ste a dy in c r e ase in both inlet pressur e ar i d inlet t em p era ture in th e combustion chamber. Th e effect is due , of c , _rs e _ to incr e a s ing comp re ss or p ressure r a tio s .

Al so there h _ be en a S t e ady i_ t e a_ e I n c o m bu _to r ex i t te m pe ra tu r e s from a bc _ t 1600° F durin g th e early 1 9 5 0' s to proj ec ted v _ lu es of ex i t t em per a ture in excess of _ 500 ° F for engine _ currently in develop me nt. T h e trend t o hi g her c o m bU _ to t exit t em pe ra tur e h a s , of cour se , f ollo w ed the u se of tu rb ine cool ing an d ha s l u- cr eas ed as improvements ha v e b e en made in turbi n e ma t e ri als and c o oling me t h od s . Fi t_ ure IV-1 als o s llow s th a t there h a s been no st _ i f ic an t increase in r ef er e nce velocity with time. Thi s is probably due to the fa ct tha t c o m bus to r pressure drop is a function of refer e nce velocity sq uared and _ in order to main- tain good s peci f ic fuel co ns u m ption s _ .nd cy c le offlc iencie s , pressure drop h as not been per m itted to increase.

Figur e IV-0. s ho ws th e combustion efficiency o f a typical annular c o m buStor.

Co m bustion ef f iciency i s plotted against a correlating pa ram eter made up of in- let pressure, inl e t te m pera tu r e , an d reference veloc i ty. The figur e shows th at co m bustion efficie n cy i n cre as es aS pre ss u r e and t em perature incr e ase and d e - crease s as conlbu sto r velocity in cr e a se s. It ha s been m en t ion ed that the use of higher co m pressor pre S Sure ratios h _ resUlted iu incre as es i n bo th inl e t pressure and inlet t em pera tu re f o r the CO n l bus_ or. The use of higher crui s e spe l ls al so leads to an inc r ease in c o m b nsto r inlet pre ss ure an d i n l e t t e mpe ra tu re. Typical k v al ues of the correlating para me ter for Se a,level ta k eoff an d c ru iSe of present-d a y annul ar co mbu s to r s fal l far to the right of t _ e bend in the _ urve. Th e re f ore, there ar e no _ any signffic l tnt problems in obtaining good co m b _ i _ flo _ effi c i enc y foi _ take- off and c ruis e with p r esent-day co m bustor de s igns. This is qui te a different ; si tu ation than 15 y e ars ago, wh en the re was c o nc ern about hi g h- alt i tu de subsonic _ flight using low- com pre ss ol _ -pre S su r e- ra ti o e n gi ne s in which low v al ue s of c o rn - buster inl e t tetnpera tu re an d pressu re c r e ated a proble _ in obtainin g g oo d _ - .' / bus t ion effici en cy.

T hi s paper dl B cusses fl o ur main topics of co m bu _to r research tha t h _ ,. oeen inflU _ ced by recent t r ends in eng ine op e rati n g c ondi ti o ns and perfo t m at _ ce re _ quir em ents. Th e first section de S cl _ ib es the effec _ of high e r pr es s u r e s _ nd te m - pe ra ture s on liner s d H _ ce te m p e ra tu r e s an d th e p ro bl em o f coolin g the co m bu _ - to r liner. The second S ection revie _ ve th6 res Ul ts o _ re t _en t NASA Lewis research o n s ho _ t-ten C d _ co m bu _to r _ . The thi v _ section descr i b e s th e pe rfor man c _ of a s hort m od ular conl bu stor desi g ned tt _ operate _ t tie _ ti'-s to i _ hiometric exit te m - pe ra ture S . T h e _t_ tal s e c tion dt S cU _ e _ th e Je _a ircraft exlutust e m i s s i ons prO b - le m .

I

j

L INER COOLING

With the trend t o w a rd s lfi g h e r pr ess ur es al o n g with hi g h e r inlet an d exit to m - perature s , liner t em per a tur es a r e incr eas i _gp thu s affecting t he dur a b_it y a nd l ife i i o f th e liner. The liner t em per a tur es were me a s u red within a c o m tm _ tor t o d e ter _ i mine t h e s everity of the pr o ble m . Showt l in the in s et o f figu re IV _ 3 i s a s ket c h of i_ th e co m bu s t o r . The m ea s ure m ent point w a s i n the primary zone , 1 inch dow n- strea m o f a c o ol i n g s lot. The flame te m p e r a tu r e s i n thi _ zon e rang ed from 3440° to 3940° F , a s m e as ur ed with a r a d i o m eter.

The liner te m pe r ature rises propor U o na toly with in creases in inlet a ir te m_ p er a ture, as sh o w n in fi gu r e IV-3. The d ire ct d e p en d en ce o n in le t air te m p er a- lu re is caused by both th e he a t- s ink capacity of the air decre as ing and the p r i- mary zone te m peratures inc r ea s i n g. The calculated te m p e ratur e s show the t " same trend as th e experi m ent a l data. T h e calculated v al ue s w e re ad j usted ba s ed o n the 6 00° F inlet te m p e rature data to correct f or co n ductio n in the wall a n d f or the efttrance lengths f or the convective coe ff icients.

Liner cooling is a probl em a t high inlet a ir t em pex'a tu res. W hen th e inlet air te m p e rature reached 1 200° F , t h e liner te m p e ratu re at th e l _ inch- d ow nstrea m locatio n was 1470° F. Since this is only 1 i n c h f ro m the f i l m -cooling slot ex t t , the wall will have approach ed 1600° F before th e n ext coolift g slot at 4 inches dow n s t r e a m . This iS the m axi m um w all te m pera tur e that sho u ld b e us e d t n order to avoid oxidation and loss o f S tr e n gt h.

The e ff ect o i increaSi n g pressure o n r adi ant output from a flame w as obtai ned within the same co m buStor and Is show n in figure IV-4. The data fox' co m b _ S- for A show that the radiation from a flame inc r eases rapidly w ith pre s S _ tre.

• . Doublin g th e pr e ssure doubl es the m easur ed radian t output. I t m igh t be e xpected th at the lin er w al l te m perat _ re B would inc rease signific an tly with preS S ur e due to t this i n cre aS ed radi a t i on.

Figure IV-6 s how s the liner w ai l tG m pe ra tur e data plotted a gainSt iucX' 6 as ing pres s ure _ or c o m bu s tor A. Th e ri s e i n l ine r t e m pex' a tur e i_ n o t as gl 'e _ tt as mi g ht be ex pect ed f r om obt 3 ervi ng th e tt _ creas es in radiation. Dou bling th e p res- sure did not double the liner te m pertttures for the B a me percentage c o olit _ air- i flows. This iB a r e s u lt o f th e increases In convection. The c o nvective rat e irt- . _ !

c r ease s as , the density i n c r eases at a fix ed reference v _ loctty. The _alc ulated _ • ' values a _ x'ee welt with t he expe ri mental data . _!

. , From th e radiatiort outp ut curv e d _ ho q T nin fi gu re IV-4 _ it is app _ lr _ nt that there cdn be a wide variation in radi an t output at the same p re s su / 'e fo r tWoei m i- ! : t \ f a r com bu st o rs . C o m bu _ tor A h _ s a l ow sm o k e m ttput th a t I s well bel o w t he vl si- b le thr es hold _ nd th a t a t low pr e S e u r e _ a pp _ae he _ th e no nlu m l n ou _ v a lue s o f r a d i - alton o u t _ t. C om bu sto r B IS of the _me ge o me t r y b u t ha s a ri c h p rima r y z one whi c h p rod uc e _a hi g hc ar b o n -l _ i c .l a t e c o nc e ntr a ti on , Th e rad i a ti o n level of .

sm ol _ yc om tm _ to r a tl o w pr ess ure _ anbe a _ hi g h astha t ofa low sm ol _e r o per_ tt ln g at a 8 0 t o t pr e ss ure ratio. The eff ec t o f thi s hi g h r_ d i aut outl _ t i s s h o wn in figure IV-6. The cal cu la ted re sal t _ t .r com but _tor B s h o w that th e in cr e _ e i _ ra _ di a ti o n du e to the carbon particu|et es in th e p rima ry zone p red u c o S _ if O dfic a nt in _ c r e ases in line r t em p e ra tur e S , Th e only w a y to r e mov e this added heat t e by in- c rea se s in c onv ec tion. Not e th e u nus u al tr e nd that th e liner t em p e ratu r e d o- c r eas e s with in creas in g pre s sur e , Thi s i s due to the f a ct that at hish wall to m - p e r a tu re s in creases in c onvect io n a re m or e eff ec t i v e i n lower ing lin e r t e mpera- lu res . In summa ry , th e c oolt _ , g probl e m i S no t e , s S e v e r e a t high p res sur es as might be expected _ o r c l e an-burnin g c o mbu stors becaus e th e radlatt0n i n creas e ! ' with pressure is C o m pensated f o r by th e incr e as e s in conve c tio n .

With th e trend towards higher exit te m p e ratures , cooling-air r e qui P ements will increase. Thi s increase in cooling-air te m peratu r e w ill limit th e m axi mum average exit te m perature that can b e re a eh ed_ a s shown in f i gure IV- _ . With 6 000 F inle t air the m axi m u m te m perature th a t can be achi e ved at a stoich i o- fn e tric fu e l- ai r ratio f or AS T M-A1 f uel is 395 0 0 F. ii the cooling-air require- ment to _e lnta i n th e wall b e low 1600° F i s as littl e as 10 perc en t, the mg .xi m u m * temper,t _ re w hic h ca n be obtained i S $ 8000 F. Wh en the co o lin g -air r e qui r e- , fnents a re S till greater, th_ m aximu m a v erage exi t t e m perature will be less. In co m bu s tor s ope ra titt g at 2 200° F exit tempe ra tures , over 50 percent of the air coUld b e available for cooli n g. HoWever _ ff these l arge a m ounts w e re used, po o r !

exit tentpe ratu re profiles would be obtain ed . Ther e fore , th e coo l ing- ai r requii . e - _ ' ments s hould be r e du c ed to a mi n imu m . ! , , The data w hi ch h a ve been pre s e n t ed wer e f or one typ e o f slot geometry. In ord e r to opti m iz e th e co o ling- ai r require m ents, it i s im po rtant to consid e r the e ff e ct of g eo m et ry , l _ our tyPic al geometri e s ar e shown i n figure IV-8. _ n the f irst confi gura tion, conti n uous slots o f vari ou s heigl _ts are use d, W ith som e J m ai l s pacer s to m ai ntain slot height. A second c onfigu ra tio n t _ slots w ith tneteritlg hol es . Wi th th i s configuration the cooling-air flow i _ re _ a U vely tinaf f ected by w a rp ag e of the p l ate. Anoth e r i mpo rtant fe a tu re o f this slot i s t _ t _ t th _ injection angle is _ uch th a t th e J ets i m p i n g e on the upp e r l i p arid spread to f o rm a c ontitm- ous fil m of air. A th ird co n fi gura ti o n u s es a wiggle s t rip, w hic h i S placed in the s lo t to mai n tai n slot h e ight. _ 'hi s al s o givi b s th e sl o t rigidity foi" aectitate m e t er- in g of th e air. Of ou /' rent interest is th e u _ e of a channeled wall with exte _ td ed i; p s urf aces , de _l_ ed to i nc rea s e th e con vec tiv e heat t ra n s fer . The cha n n e l tn_ ru ns / t he c om p le t e len gt h of th e p lata , w hich pre ve n t _ m / x ing he tWe _ the cQu ve c t t v e ly h e a ted a i r an d t he annular a i r.

Th e f i r st t h r ee c o_ l _ u ra t lo n s w it h v ario u s s l ot het_h t _ a n d ho l e si z es w ere s tudi ed within a c om b us tor , u s in g film eoollng _ nly. A ll t he d a t _ are p lo tted as c o ol l n _ e ff ec tiven ess a_al n s t x / M s in fl _r o IV-9 , The cooling e ff ec tiv e n ess ts th e t em p e z ,_ ,t u r e d iff e ren c e b e tw ee n t h e hot gas and th e w _ l l dl v l dcd by th e maxi- mum to w4 z _ r a tnr o dif f e rence (t h e h ot 8 P .ts_ o m p v ,',t _rc . , _tn ue t h e i nlet c t _ oltng- ai r te w pe __ tu r v). T he c onvent i o n a l fil m - co oli,. _ par am ete r u s ed I s x / '_ , ,c h ore x t s t h e di s tan c e dow n s t r e am o f the s lot. e is t h e Slot h o t t _ ht , a nd M is t h e ma s s flux rat i o o f cold ga s to hot g a _ . Al l th e d a ta f or 1 0 different g eo m etrie s fa ll on a Single curve. Within a co m t _s tor t h ere w a s n o sig n ificant di ff or e ncv betwe en slot geometries at the sa m e v alue o f x / Ms. Thor _ o re , slot g eometry d oes not seem to be a majo r p a ra m e ter which affects film-cooling effectiv e nes s .

There is , h owever, a lar g e differenc e betwe e n',the co m bustor d a ta and ilia @ cooling cor r elations publis h ed i n the l i tera _ re. When a re l atively simple mixing m od e l wa s u s e _ , the difference between th e co m b U s to r data a nd literatur e c orre- lations was a r esult of t h e i n c r eased tu rbul en ce in the co m bus to r. Lite r atur e cor r elations are based on data taken i n due l ed flows , where th e tu rb ul enc e i s low and varies from 1 to 3 percent. The co m bu sto r t u rbule n ce level was a p proxt- • m_ t e ly 15 percent. Curves from r e f ere n ces 2 to 4sl)own in figure I'V-9 a r e t ypi- ' cat re sults o f literature correlations.

The solid line passi ng through the data in fig u re IV - 9 is the predicted curve g iven f o r ,.he h i g her turbule n ce level. The equ a tion which was u s ed to predict film-cooling e ffectivene s s q is given by 1 t 1 + C m x - -x- wh e re C m iS the m ixi ng c o efficient and has be e n s e t t K tu _ dtOthe turbulen ce level. The in cr e ase i n tu rbult _ ttc e r esulted i n th e l iteratur e corr e l at ion s ov e rpre- dietin g t h e fil m-c o o li ng e ff ecti venes s by a fa _ tor of 6 or more. Oth e r c o m blmtors ma y have a dif f er en t lev e l o f turbtfl enCe, Whi c h w o uld produ ce either a higher o r .. lower film-cooling _ f f ecttV en eS _ . Il l fact , ev e n Wi thin t h e Same co m bustor , t he tu r bule n ce level might be e x pe , , _ ted to v a ry.

Thus, tu rbtde n ce i S an l_ po r tartt parameter i n deterfnt n tn _ the fll _ fl-cob lln g effectiveness. One way o f r v . _ ucin g the cooling-air requi r ettle nts wo ul d be t _ 1 0i • } !! , l ow e r th e t ur bule nce le v el in th e neig hb o rhood of the w all . Oth e r m_ns are b eing inves tig a t ed t o in cre a se t h e e ff ec t iv ene ss of the cooling a ir. For ex amp l e, t h e u se o f r e v e rs e co o lin g fl a w on th e a nnUl u_ s id e of th e wall would intr o duce th e a ir fur- th e r u p s t rea m so tl _ t it cou l d b e u s ed f o r combus ti o n. A r a th e r s i m pl e w a y to sl_ tficautl y r e d uce coo l ing - air rt _ : ,tui r om ent s i s to red uc e the su rf ac e ar e a wh i c h mu s t b e cool e d by s h o rt e ning t he e _m bust o r s .

SHORT COMBUSTORS

We a re pr ese ntly studying tn. _ ny w a y s o f s hort en in g trt e co m bu s to r . Th e f ol- lowing dis cussio n r e lat e s th e S e v e r al different app r oa e ltes bei ng us ed to obtai n a hi g h-perfor m a n ce short co m bus t or an d how thes e approac h e s diff e r from th e con- v en tional turbojet co m b us tor design.

Figur e IV- 1 0 shows how the le n gth of tu rboje t engin e co m tnmtor s has be en dependent on the engi n6 siz e ; as exe m plifi ed hez'e by the engine s e a-level takeoff airfl ow rat e . Three di f fer e nt types of co m bu s tor are c o mpared: th e _a n , th e can-an n ular , and t h e annular. An n ular co m bustof _ a r e be c o m in g the co n ventional practice, as this des i gn u s es th e en gine volume avatlabl 0 f or th e co m bustor mor e efficiently th an the other typ es . The s hort-co m bustor program at NASA Lewis assu mes a fairly l arge engin e having a taR ecf f all'flow of 2 80 pounds pe r s econd.

All th e NASA co m b U_t or s ar e 20 inches long. By co m paris o n a typical pres en t- day an n ul a r co m bustor of a similar-size en g i n e Would be about 8 0 in ches long, or 60 p e r o . ent lo un g e r than th e sh ort co m b _ ,s to r s .

A typictd co m l 0 ustor is shown in fl gn re 1V-11. Thi s an nular co m bus to r has all the feat u res th a t are c ommon in pre s ent - day co m bus to ts. A relativ e ly long f diffus e r i s used to diffu s e the compr es so l " exi t airflow to a high _ tatl c pressur e , t !

Often a s aout or flow spl i tter is used to assist in dividing th e di ff us e r airfloW- _ properly. Fuel i s inj e cted by a pr e ssure a tom izing no_41e and co m bust i on is ini- tiated and s tabilized in the primary zone. The re ma i n ing airfl o W i S in jec t ed into th e c o m bustor se conda ry zone th l.ough h o l e s or slots and m ix es wi th the hot gage8 of th e p rim ary zone to complete th e co m busti o n reaction and a chieve th e desir ed e xit te m peratur e prbfile. Tile diff e r en ces be twe e n th e convention al co m bus tor ii an d th e t hort co m bu _ tors are nut e d i n the followin g discussion.

Th e first short combu s tor is c al l e d the dou b le- ann ttl a r ram-induction coal- b e stor. A c roas - s ectio n al sketch of th e cot _ bus _ o r is s ho wn hi fl _ ure IV-12. l _ This co m bustor cons is ts of two concentric co m lms to r ann ul i; th e e _ i h_ tst of ea d h 10 _ an n ul us blending together i n the ex it tr an sition regi on . Thi s co m bus to r has a very : : s hort diffuser since the ra m -induction concept requires that a high dynamic pr es - i : sure be m aintained at the diffuser exit. Therefore, the diffuser is shorter than _ , th ose conventionally design ed . T h e airflow i s captt l r ed by shrouds and ducted into !

L

the co m btis to r pri m ary and S econdary _on es through rows o f scoops. Each s c oop I has turning vanes to m in imize th e l osses in turn in g a hi g h-velocity airstrea m . I Each annuluS of th is co m bu s to r i s conServatively des ign ed i n termS of its o wn i leng th -to-an nul uS-he tff ht ratio. The co m bust o r is effectiv e ly short oned by us in g _ a do _ tble-a nnu l uS g eo m etry and th e r a m- in duct to n concept, which requires hi gh- • : velocity airstreams an d hence a short dif fu ser. Further tletafls regarding the • _. c o nstruction and per f ormance of th e doubl e -a nn ular ra m-induction co m bustor c an / b e f o un d in re _ er.e _ c es 5 to 7 . Figure IV-13 is a view of th e double- an nular co m - b _to r looking u ps tre am t o ward th e headplate. The S coopS ar e cle ar ly shown, : those furthest upstr e am be in g th e p rim ary zone s coops an d th oSe downstream being in the seco n dary z o ne. Th e turning vane in each s coop c an also b e Seen.

Another s hort oom bus to r pre s ently bein g in v es tiga t ed i S c all ed the one- s ide- en try co m bus to r. Figure I V -14 i S a cro s s- s e c tional sketch of this commas to r.

Thi s co m bustor is diff e rent from th e convention al co m bust0r in that all the pri- m ary an d S ec on dary ai r is duct ed into the co m b usto r frO m only one side; in this ......

case, the outer-dia m eter side. Thi s is also a ra m -inductioi l combtm to r, as _ S_ oops ar e us ed to turn th e hi gh - velocity air into th e combus to r. T he reqtflr e - mer i t to m aintain a hi gh dynamic p re ss ure in the atrst _ eam (ram - i nd uction ap- p r oach) _ hort ens the difftlser. The difftmer has al so b e en s ho r tened by being radi al ly extend ed and overlapping the co m bus to r. A small portion o f th e ai _ , .... i / approxi m ately 10 to 15 percent, is captured by chutes regu l al.ly spac ed aromld th e diff u ser inlet. This air is passed thrO ugh the swirler arou n d t h e fuel nozzle in the h eadp l ate _ d is also us ed to cool the i ow et liner. Th is co m bt / s - t tor is of part icular inter es t bec a use i _ s ex it te m pera tu re profile i s ex t reme l y in sensitive to rad i al airfl o w p r ofile distortions coming fr om the co m pressor.

Thi s in s ens i tivi ty r esul ts _ ro m all the flo w b e ing dticted in to th e co m bus- tor _ rom o nl y On e s id e a _ d that flow b 6ing captured b y sc o o ps that e _ tteud to : / the f ul l hei g ht of the ai Hlow passage. Therefore, regai _ ll e _ o _ the S hape of th e // r ad ial alrflo _ p ro file leaving the co / hpr _ s S or , the sa _ ne aniount of air is captured by each sc oo p a nd turn ed into th e co m bustor. Th / J s, th_ m ixing pa ttel _ ik ifisitle : the comi Ri stor are not chan ged an d th_ exit r adl al tmhpera _ re profil e is un- chan ged . Figure IV-15 is _ vi ew of a port ion of the s i d e- en try combustor. " _ e sl o ts ior the a tlm i s. s io n of t h e prirh _ iry an d second a ry a i r ca n be s e en o n fi l e out e r w al l. Th e se _ ent te sta of t his co m bustor a re d es c ri bed in refer en ce 8.

The th i rd short co m bus to r being inve s ti g ated is s till more u nco n vent i o n al i n des i gn approach. Thi s co m bu sto r is the m odular s wir l -can co m bustor , which is e hown in cro ss secti o n i n figure l _ V -16. The co m bustor consistS of 120 i ndividual i swirl-c an rnodule _ arranged in three concentric an n ular rows. The m odular co rn - ' bus te r h _ no well-defined primary or secondary zones as in the co n ven t i ona l combu s tor. Ne a r l y all the airflow, ex.cept for that requi r ed to cool th e liner, _:_ passes dir e ctly th rou g l _ or aroun d th e modules. Each m od ul e combines the func- tions o f fuel injection, vaporizatio n , and combustion stabilizatio n . The co m b _ - !

J to r can be shortened because of the rapid burnin g and mixing flint occur s down- . / : stream of e ac h modul e . The m odul a r approach tends to m axi m ize th e area be- i i it • : tween the hct gases leavin g each module and the air floWing arou n d each mod ule.

Only a small fraction of the to tal airflow i s required to c oo l transition liners aS " i they are physically displaced f ro m c o ntac t with th e h o tt e st com busti o n gases. _ ; _ , Figure IV - 17 i s a vi ew of this combust o r. The in ner t ransition liner has been _ re m oved so that the mod ul es can be s e e n m o re clearly. The three rows of swirl- i::_ can mod ules c an be seen. The inset in th e figure i s a clo s eup view of the com- ' bustion m od ul es. The combus to r is v e ry simple an d a m ini m u m of met al is ex - i_i t posed to th e flame so that comlms to r durability should be go od . The m ed ics _ the ms elve s are very si m ple. Figure IV-18 s hows th e detail ed construc ti on Of a • ty pical co m bus tor m od ul e. Each modul e cons is ts of th re e basic par ts : a car- buretor , where air and JP fuel are in tr od uced - th e ai r co m ing th rough th e inlet • and the fue l being i nj ected tangentially on the o ute r w al l; a s wirle r , where th e ' ; fu el and a i r are fur th er mix ed an d a swirl i m parted to th is mixtur e ; and fln _ dly , i ; a f ame stab il izer , w h ere the combustion is ini _ ated a n d maintain ed . Tile air _ ,.

: flowing around the outside of th e mod ul e m ixes wi th th e hot combustion gase s in , I_ th e wake of each modul e. There th e combustion reaction is complet ed an d mixing _ .

of th e g as es to the d es ired exit t e m p e ratur e is a chiev ed . Segment t e s ts of this _ ' co m lm s tor are descl'ib ed in references 9 to 12. T h e m od ul ar concept i $ not new at L ew is. The Swirl c an wa s origin al l _ develop ed in 1956. At th a t time th ere " was a g rea t de al of in terest in buring h ydrogen in turbojet combdsto rs and the T_ swirl can w as d e v e lop ed as a n ew cor _ buS to r conc e pt for that fu e l . R _ fe _ - _ encee 13 to 15 docu men t that Work. Figul - e IV- I9 _ hows one of th ese early sWirl- can co m bustol - s . T' a i s w as a quar t er-sec to r o f a eo m b _ tor us ed to si t flulate t he • perfor man ce of a full - anntfla t d d sig _ . T his pa rt ictflar comb _a s to r is of I n _ ei , est - because it w as test ed w i th va po ri z ed ,fP fuel m id demonstrat ed p e rf d rma n ee - nearly as good as that achiev ed with hydrog e n fiiel (ref. 15). That res ul t gave i m petus to the belief t ha t a s m all-dia m eter co m bustor mod ule could give good perfor man ce using liq u id fuel. ..

" ; Perfor m ance t ests have been c o nd ucte _ with all three short co m bustor s . The . _ '. following diScuss ion compar es th eir performal _ Ce with the perf o r m ance of a typical annular e o m bustor of th is B tze, whi c h would b e a bout 6 0 percent long e r. The fir s t . combustor cr i terio n m entioned in the t n t z'oduc to ry l _ ragraphe " J r th is section was I : combustion efficiency. Since these c _m bus tor s ar e design ed f o r a m odern en- g in e, th e co m bustor operati n g co n dit ion s, at t al ce off an d c r uise for it / stanc e , are ' . ' eXtr em ely favorabl e for effici e nt co _bu stion. The combustion effic i ency of all _ , thr ee short co m bu sto l _ is 100 pe rcent at these co / _ dition s . Th u s , reducing th e _ ' combus tor l en gth by 85 per c ent had no effect on co m bt lS tion efficiency. Fig- i .

' .... u _ .-e IV- _ 0 compares the pressure loss of tile three short co m b u stors with that of i th e longer a _ nular combus t or. Th e Side-entry an d s wirl-can combtlstors ha ve _ _ . i total-pressure losses si m ilar to that of th e conventio nal an n ul ar c o m bus to r. The i ' • _ double-annular co m bu s to r has a much higher pre ss_ tre loss. This co m bustor has i_" _ bee n r ed eSign ed by incr eas ing i ts open flow area an d the pres s ure l os ses have i _ been redu c ed t o 4 pet-c en t at a diffu s er inlet Ma t h number of 0. _ 5, as s h ow n by th e solid square. Th ese test res glts a re f or a quarter- s ec to r. Redu c ing th e _ .

to tal-pressure loss c an_ ed no change in th e oth e r combu sto r per f or m ance pa ra m - _ : i _ stetS. Thus , it _ tppears th e * s hort-length combustorS can have l o w p r es s ure ' I _ loss es similar to th o se of longer l en gth co m bustorS. _if .... Figure IV -21 compares the e x it te m pera tu / 'e pa ttern factor of th e short co rn - ! • I ' bu _to rs. The p atte rn factor is a measure o _ the quality ot _ the e _ t te m perature !!

.. .- dist _ butlon. It iS d e fin ed aS the di f fer e nce b e tween the m a xi m u m l oc al c om bus- I_ . ' tot exit te m peratur e and the averag e comb u stor exit tempera tu re divid ed by th e .' average e o m bustor t em perature rise. Low Values o _ th e pa tt er _ f _ tc t or ifldicl _ te i!. _ • more unifor m te m perature distri bu tion. Th e pattern fac to l' for the typic al an- . _ . : i :. _ nular co m bustor varies b _ Rwe en 0. $ and 0.3. The double-aflu ul_ r co m bustor has dcm on dtrai ed a pattern fac to r variation of 0. _ to 0. 25. Patt e rn f ac to_ valu es . _ t • .. as lo w as O. 15 have be en achiev ed in S ector t ests . The sWirl- _ an co m bus to r i_ _ m d the one- S id e -entry co m l _to r have pa tte rn f a cto rs so m ewha _ ht _ h _ r, i u dt- ' eating that m _ )re work needs to be done to impr o ve them . But _ s is it _ di _ ate a ' _ here by t h e open portions of the bars, bo th c o m btm / _ l' s h i ve demol _ st ra ted _ ' e ry low pa tt ern fa cto rs in preVious s egm en t te _ ts. Good pat[ern fa c tors should al _ o _ ' be achievable for th ese _ :o m bus to_ in th eir a nn ul a r fo _ ' m . _" Another l in portant co m bustor crite _ oil is that o | al t i tude reli g h _ . We _ re al s o conce rned with co m bu s tor blowout , U t h e c _ mbtis to r biow _ mt linllt is th_ m ini m u m point wh _ re r e ligh _ could ev ® r be ach i eved. The bed t rel t ght [ _ e / -foi _m - , is the r es ult of optim i z ing many v a / - _a ble _ , such as spai'k plts g positt@a , fuel- a ir [0 5 • , ratio , and nozzle spray angle. While not ignoring the reli gh t problem, we are devoting more effort to m easuring and improvin g th e co m bus to r blowout limits.

Figure IV - 22 compares the blowout p e rfor ma nce of the s ho rt eo m bus to rs w ith that of the m or e c onve nti o n al annular co m bustor. This is th e one area where some of th e short co m bus to rS have not yet achieved p e rformance S i _ nilar to that of longer _ o m lr a s tors. At a re f erence Mach number of 0.1 , bo th t h e double- annul a r a nd the swirl-c an co m bu S tor bloW out at pre s s u r es cousid 6 rably high e r than th e longer an nul ar co m b ns tor. The blowout li m i ts of the sid e -entry co m bus- to r are, however, practically identical to those o f a 1 . J n ger co m tnt sto r. Th e blowout perfor man ce of the double-ann ul ar and swi rt -ca n co m b usto rs lshould be i m prov ed . Many ways of i m p r ovin g the co m btmto t blowout li m i ts are b e in g • inve s tigat ed - for instance , a r ed uction in co m bustor r e ferenc e velocity. Fig- ure IV-23 ShoWs the m agnitude of improvement that can be m ad e by reduci n g the refere n ce Mach number from 0.1 to 0.05 for the d o uble- _ mular c om btl s_ r, A , i r ed u cti on in reference Mach number me i ms a l 'eduction in co m lms to r airflow rate. This air m i ght be bled from the engine or bypassed around t h e combuStor.

Some form of variable- g eometry co m bu s tor m igh t also be us ed . An y S u c h ap- p roac h would only be n_eded duri n g an act ual i n -fli g h t reli gh tin g process. Onc e th e co m bus to r was relit, t h e bleed or byp _ S fl o w would be r ed uc ed u nt il the co m bus to r w as aga in operating at the pro p er throu g h fl ow condition.

A t t e m pts to improve th e blowout li m its of the swirl-can d _ )mbus to r _ are focus ed on i m pro vi ng th e ov e l- al l combustion s tability ran g e of th e in dividual m o dul e . In the pas t , sp e cial s tart fuel nozzles hav _ been used with th e preva- i poriz in g typ e o f co m bustor. The _ e fu e l noz z l e s ha _ 'e ai s o b e en test ed W i th s# / i rl- c an m odule co m bus to rs; th e y do extend the ground S tart c _ tpability an d presuni- •: ably will al so i m p ro ve al ti tu de i.eli g ht !ilIlits. B e cauSe al ti tu d e r el ig h _ i s an im- portant combustor desig _ n criterion, col _ t in ui n g efforts will b _ made to improve I th e reli g ht limi ts o f sh o rt C o m bo s to rs.

A n other co m po n ent who s e l e ng _ affec ts the overall c om beS to r length iS th e diffuser. In order to re al ize the full b _ef it o f sliort-combustor techno l ogy, the diffuser leng th flnl S t b e kept to a lninimu m W hile the sam e i _ let- to _ xit-al'e _ t ratio is m ai n tain ed . The i-esuld ng wide-angle difftmers wil | hav e flow-sep _ tration problems unles s special t _ chniqtl _ ;are us e d t o I _ e _ l _ th e flo w attached to the walls. One such tech ul qt l e is the use of b0tindary- l_ y 6 r b leed.

* This ai r cab b e drawn otit ttirough th e Wal l s of th e combuS to r diffuser. ' i'he effect this co ul d liave on s hoi't e ri in g th e diffuser i s illustrated | ii figure 1V-24. _ " , The top sketch is th e s wirl-cab co m bustor wi th w ed g e -sh a ped i ns erts to fli af n tat u i 10 6 the proper flow s plits. The botto m sketch ill us tr a ted ho w the d if fu s e r m ig h t be S hortened by usin g bo u n dary-layer bleed on bo th walls. In this w ay, the co m pres- sor ble ed can be doubly u se ful by i m proving diffuser p erfor m ance as well U by cool in g the turbin e . A s was pointed out in the previous sections of th is r e port, : large quantifieS of co m pressor bleed m ay be n eed ed to cool th e tu rbine. _ Figure IV-2 5 Shows how small an a m ou n t of ble ed flow is requi r ed t o i ra- : prove the pe rf o r m ance of a large-ar e a-ratio sho rt diffu s er. Here ar e re s ults o f tes ts on a two-di m en _ ion al diffu s er th at had an a rea ratio o f 4 to 1. When there i was n o s ucti o n appli ed to either su rf ac e , the dif t u _ er waS in S e t flew. With only ,_ id" 4. $ percent of the total airflow bl ed away, th ere is a Very ma r ked lowering of : : th e cent e r velocity and a co r re spo n d ing incr eaSe in flow a t th e upper and l ow er wall s . Svrprisi n gly, increasing th e bleed rate to "/ per c ent did not p ro duce much . / J mol-e of an-improvement. Therefo re , Some bl eed is very benefici al ; but the rate, _ once a bo ve s o me minimum value, i S not crit i c al to diffuser pe rfo r man ce. 14 Figure IV-26 shows how the diffuser pre s sure losses ar e af f ected b y _ bo undary-lay e r bleed. Note th at the diffuser to tal-pr essur e loss d e creas es as th e perc en t of bleed flow increase s . AS before, increasi n g th e qu an tity of blee{t flow above so mv minimum value ha s a di min ishi n g improvement on diffuser perform- ance. So a s m all a m ount o f bleed flow is V ery b enef icial an d l a l _ ge qu ant ifies of ble ed air are n ot r eq u i red - thoug h the use of large ble ed flows is in no way de t - rimen t al .

: ,' HIGH-EXIT - TEMPERATURE COMBUSTOR

/ ' :_ " An i n ter es ting applic a tion of Short-co m bustor technology is t he hi gh - _ xi t - te m pera h lre combus to r re s earch program. The p rogr a m ai m iS to evolve ful l - t va mul ar co m bus to r s capable of pe rforming with fu e l-air rat i os a pproaching s toi- chiometrtc value _ . Initi al re _ ul _ ha v e been most e nco ur aging. A swtl.l-can m odula _ co mbu stor w as us ed to , extend good combt _to r p _ rfor man ce to exit tem- , , pera{ureS over $ 600 ° F.

A m odula r co mbu g _ or de s i gn ban al ways _ pp 6 ared _ be d es irable _ or eng in es req u iring ve ry hi g h tu rbin e i nl e t tehit _ era tu r _% Howe ve r, bef o r _ un d ertakhig th e , o pe r a tion of a co mbu stor at hi g h te m pe r atures , it w g J_ first nf _ c es d a ry to o b tain • , sati s facto r y pe rfor m ance lev e l s at lower tem per _ tttr es . Swiri-c an! co m busto r s have be a m investigated for m_ ny yea r s at NASA Lewi s . T he in i tial w ork in i95 5 , : (refs. 13 ahd 14) w as conducted with hyd rog en f d6 1. Later work (ref. i§) us _ l vaporized hydrocar bo n fuel. The m ore recent work de s crib ed in r e _ erehces g 1

!

to 12 w a s condu cte d with liquid J e t f uel and was devoted to i m provi ng th e d esign of swirl-can co m b u s to rs and exte ndin g their operat i ng ra nge to c on ditions typical of m odern e ngines. T h u s, it was only after co m pleting s ufficient backgrou n d work tha t co m bustor o p eration at near-s to t _ hto m etric te m peratu re s was atte m pted.

Figu r e IV- 27 is a schematic of the hi gh -te m pera tu re test co m bu s tor. It i s a m odular design quite similar to th e swirl-can co m bustor s how n earlier. A mod ular confi gura Liou Was s el e ct e d b ecau se of S ev e ral apparent advantages that m_ wlular co m bus to rs have f or hit c h-exit-te m perature applications. Thes e ad _ an- rages include a near- h o m o gen eou s m ixture o f fuel and air _ mini m al lin e r coolant- i air flow r e quire men ts, an d high atLair _ ble heat- r elease rat es , t . . Th e co m bus to r utilized 1 2 0 fuel inletS to produce go od m ixing of f u el and air. 14 A ho m ogeneous fu e l-air m ixture is i m portant since stoichio m etric bur n ing re- i quir es th e re _ tction of all the air with th e fu e l. For co m paI.iSon, a convention al !

!

combus to_ d e sign woul d c on tain abo u t $ 0 fuel noZZl es , ii A min i m um amount of air was u sed f o r co o lin g the co m bus to r l ine rs since !_ this air i s g en erally no t av a ilable fo r co m bustion. Fil m- cool ed l in e _ : d es i gn s !i were u sed . The to tal co o lant fl o w und e r the liners w as s e t at 8 pe rcen t of the r : 1 : i to tal ai rflow for a combus to r exit t e mp e ra tu re o f 1600 ° F. Due to inc r eased I m omen tu m bUrn ing losS, thi s flow in creased o to about 8 percen t at an exit t ern - _i_ perahlre o f $ 600 ° E. I ' Such r e lativ e ly s mall coolant flows could b e u t ilized because of the followit 1_ ' Combus to r fe atures: I_ h_ (1) The co m bu _to r i s s ho _ t, being only 11.4 inches long fx'o m the fu el e ntry i_ plane to the co mbU s tor e_ tt. Thus, th e r e wl aS l es s lifter to cool. Also, simple ii .. . . l in er d es igns wi th m ini m al surface area were us ed , S i n ce sc oo ps o ' ," S lots w e re i_ ii : not rSquired to supply nilu s ,it air. _ (2) The cool an t film w as utiliz ed to maximu m advantage by r ed u c ing th e tur- I_ t lzflence level at the liner s . T his occu rre d since r e c irc nl aflon zo nes of h of ga _ e s r_ w e re conf in ed in the m odul e wak es and co ul d not wash al on g th e co m bustor ' ., _.

line rs , L in ers were als o exte n ded z _ a d i al ly away f ro m the fl _ t m e st r eaml ines of these wake z on es.

(3) Addi ti onal cool an t air w as supplied betwe en the arl'a y and the lmt - g as S ide of th e l ine rs . Sinc i_ Ll _ ts air W as su pplied f ar up _tre a m_ iL had a m ple 0ppoi ' tunity . . t o enter i nto co mbUs tion reacti()fls. , .. + '" Modular co L nb _ s to r s ar e al s o cap a bl e o t high- in_ slty bU rnin g , which re- 4 Sui ts in high h¢_ t-r el e as e rat e _ . TiI _ cotabu _to r i S 6ss_ mti e dl _ _ pre.m ix s y s- " te rn , wi th m ixin g of fuel an d air occurrt ng upstre _ t m of the burMn g Z one in the carbure tors . And, S ince ne a rly all t h e ai rfl ow p _s e s thi _ tt g h the a f _ 'a _ ' _ U le _ e i08 is a b u n da n t ai r ava i l a ble d uring all s ta g e s o t th e co ml_ u s tl on procesS, / Co m bus te r perfor m ance was m onit o red by over 400 pressure a _ d te m per a ture , , r e adin gs . Av er ag e co m bu s tor exit te m pera tu res up t o 23 00° F were m easur ed by ' a circu mf erential traver s $, fiv e -poi n t probe. Higher exit temp e ratures were • calculated by using a c h oked noz z le . An iterative c al cu ! ation using the f o llowing ec i u a t i o u was employed: PACd( _ [ + da) =\ R / _ K - ' _ + i 1 , i& where ,_ ' . : ' K s peci f ic heat ratio i'- I . :. g 32. 174 ft / S ec 2 : _' ' R 53.35 ft-lb / ( l b)(°R) •.:.. W total m aS s fl ow rate ' P total presst l re at n ozzle t h roat ... T t o tal _ e mpe ra tu re A n ozzle throat area, 1 . 32.5 0 i n . 2 _ , ... Cd(1 + da) nozzle growth function " " . ".' T h e choked n ozzle was c al ibrat ed with th e traVer S il l g p robe , an d _ nozzle g rowth : .! function o f 0.98 _ [l + (Texit - Ta m blen t) (0.018 / '_500)] W aS d _ termtn ed , which ' " " produced an area g ro wth of 0.7 2 percen t per 1000° F temp e rature rise, Effects ' ." ' ' of the n oz z le on test conditions and p _ rfor m anc e are discus s e d l ater hi th i _ sec- tio n . t " Ftgttre IV- 28 is a view lookin g upstrea m in to th e co m bustor after a 3 600° F • .. . . " ', . , " ' I run. The in n er li _, :r was removed to better il l ustrate the module a r ray. Mod- ules wer e cl us t e red in groups of th ree ahd held i n plac e by fuel s _ ru t s. Thes e i .... '. : e tl _ts c an be seen protrudi ng thr ough the outer co m bustor houSing. Inst _ amen- i • .. ration l e ads can also b e observ ed ext en ding _ ro m th e centel'body mid the out e r !

combu S tor houSing. . } Co m bus t o r m odule details a r e shown of _ figure IV-29. Th e swirl-C an design !

: ,' ' was m odifi _ d fo r th eSe high , temperature tests. T h e basic coinpon _ ts of e arbu- i .... _" refer an d sWirler were retained. Howe _ er t il e cone flame stabi | i_ers _ ieScrtb ed !

" p reviotlsly, an d shoWr , ,, _ ' th e d as hed lines on th e fi gu r e , wer _ replaced by flat !

• i k, pl a tes. Plates were used to eliminate metal fro m the burn in g z on , _ . Th e circu m - ference o f the flame s t a biliz e rs was al s o i nc r e ased to m axi m iz e i n terraci al mixing area b etw ee n the hot g as es in t he module Wake s and the air flowing through th e array. Sharp ed g e s were retai n ed to induce f urther turbulenc e . This hexago nal - plate flame stabilizer i s one of sever al de s igns that ar e cu r rently being i n vesti- gated.

Co m bustio n effici en cy i s indicated in figure IV- S 0 by a plot of combus to r exit te m p e rature a g ai ns t fu e l-ai r ratio. Th e s e data were genera t ed in a s in g le test run of 6 _ hou r s duration. The co mbu stor inl e t air te m pera tu re waS 600 ° F.

Inl e t pressure w as vari ed be tw een 3 a n d 6 atmospheres and ai rflows were be- tween 49 an d 6 8 pou n ds p er second. A fuel-air ratio sp a t1 of 0.017 to 0.0 6 pro- duced an exit te m perature span of 16 5 0 o to $616 ° F. Th e highest average exit temp e ra tu r e achiev ed , 3 6 16 ° F, i S only 33_ degrees F below tile m llxi m u m theo- retical te m per a ture, which is $ 953 ° F at th e se conditions. Nearly all the data are within a 2. 5 - perce n t s pa n of t00 _pe rce n t co m bu sti on effici en cy. T he source o f _! : the 2. 5 -p e rcent va ri ation i s the choked nozzle calcula ti on. A 1-percen t var ta - I_ tion i n the choked nozzle I _ arame t ers produces app ro xi m at e ly a 2. 5 -percent de- i vati on in co m bustion e f ficiency. Th e fix ed -area choked nozzle al s o restricted the flow conditions which could be investigat ed . The ref e rence v e locity wi th i 600 ° F inlet air was 132 feet per second f or is o th erm al condi ti ons, 98 feet per i_ second at 16000 F exit te m pe ra ture, and 67 feet per secoftd at 3600° F. !_ Figure IV-31 _ hows th e co m bustor to tal-pr es sure loss as a ftmction of di f - I_ fuSer inlet Mach number for iso th e rm al _ nd bu rn in g conditio n s. Accep _ ble _ , _i levels of pres s ure loss were obtain ed . A co m pa r ison _ )f isothe r m al presedr e i!

loss o f this co mbu s to r (ope t i circles) wi th the typic al ann ul ar cbmb _ tor de- _ -. _ : s cribed p re vioU S ly (open s qtml. e )shoW S co m parable pre ss ur e lo S S valu e s. AJ s o I_ prese l Red a re burning pr ess ure los ses (solid circl es ) for thre _ combus to r exit tempera tu z _ s corr espo ndi n g to co mbu stor-exit to - inlet t e m perattlre ratios o f t 2, 3 , an d 3.8 5 .

Th e following diScus S ion cove _ addi ti on al re _ mlts Obtain ed in the high-exit- te m perature t es ts: (1) Patt e rn factol': Co mbu stor exit pa ttern fac to rs w ere 0.3 2 _ tt a eo m tms tor exit te m pera tu re of 2280 ° F. The traversing probe w as h ot us ed for lfl gh er t e tn- perature s . How e ver , patte rn fac to rs at hi g h e r e _ t te m pera tu_ 'eS c _ m b e l rl- f erred. Fo r ex ample, th e pat t e rn factor a t 3 _ 16 ° F was 0.112 o r less, _ ince th e m axi m u m theor e tic al te m pe r a _ re was only 3 5 7 d eg rees F hi gh e r.

(2) Heat _el e as e r a te: A f n ax i m tt m he a rt _ ele as e rate o f 13.3 _1 0 6 B tu pe _ hour per cubic foot of co m bustor volume per at _ lo _ p h ere w aS obtai ned . J liO ( 8 ) S m oke for m ati o n: S mok e-co n c e ntrati o ns at the c om bu s tor exit were m eas- ured Us ing a s tained - filt e r - pape r method. No smo k e could be d e t e cted for ex it . . te m per a tures below 8050° F, At 8( 3 00 ° F the smoke level was below th e visib l e thre s hold.

t (4) Resonance: A c ou s tic resonance or c o m bustio n in s tability w as no t en- c oun te red. Re s onanc e i s oft e n a problem i n hi g h-t em p e r a t ur e _ o m b l t _ t io n s y s - tem s . !

(5) Li ner t em p0r _ , tu r e : Th e m:x _ i m u m lin e r te m peratur e wa s 1680° F, This i tem pera tu re o c cur r ed on the inner li ne r about i n c h es down s tre am o f the m odu- ; lar array and pr e vented f ur the r inc r eases in co m bus tor e xit temp e ratur e , i . i ( 6 ) Damage: So m e da m age occurred to the cotnbus to r during the high- te m p e ratur e ru n . Th e da m age waS m ini m al and was restricte d to fiv e vanes fro m fiv e differ en t sw irl e rs and to a wa rp ed s pacer ring b e tw ee n th e o u t e r hou s i ng and the outer fil m - c ool ed li n er.

, (7 ) R Un ti mes : T h_ su m tot al of high-te m pera tur e run times to date ar e 4 hours an d 50 m inute s at exi t t e mperatures over 8000° F a n d, 1 h ou r and 2 6 minutes at av e rage exit tempera tu r e s over 3500° F.

A more co m plete _ l es criptlon o f these t e sts is gi ven in referen c e 1 6 .

The initial result J with this co m bus to r have b e en v ery en co u ra g ing. High e xi t te m pera _ r e S hav e b ee n attain e d for a f al l-ann u lar, larg e -dia m eter co m bus- _ .., to t w i th ou t sac r ifi c in g g ood combus to r pe rform an ce. R e aso n able run times w e re . . 4 . t ,.

' _ obtain e d With m ini m al d am age to the co m b usto r. Futu re work wi th th is co m b us - " .' , .

.... to r will e m phasiz e durability, particularly at higher inlet ai r tell _ pe r atur e S; en - ':. _ . _ { durance runs; exhaust e mis sion d a ta; an d alt itud e reltght perfox 'm a n ce. , " .

_'_ EXHA U ST EMISSIONS ;" -. , , La t ely, [he concern ov e r air po U utio t zhas draWnthe att en tion of the co m bds- tot d e sign engine e r to th e quantifi es of _ au _ t e m is s iOn S con t ributed by the gas .... turbine e ngine, l _ ollutant emisgions may be ClaS s ifi ed as ei th er pttrttc u l _ tte m att e r (i. e . , g re ek s ) or as gaseous e mi ss ions S_ tch as carbon m onoxide i pa rtiall y bttrn ed h)drocarbon B , and n i tric oxide. The first poll uta nt tO co m e ttttder atYm e k has be en _an st sm elts b e ca u se the smok e tridl s th a t a r e left by Jet aire z 'a _t on t akeoff a _ d l anding are an obvious anttoyance to the gen b r al public, k g _ ner al , exha us t s m ok e beco m es a p r oble m wh e n combust _ r s op e t _ at _ at pr e s s ure s _ reat _ r than 10 a tmospher e s an d w h e n local fuel-air rafl0 _ in the pri m ary zone o _ the _i \ e o m bu sto r are in exc e ss o f an equiv al enc e ratio o _ 1 (i. e ., _to i _ hio me tric).

These c onditio ns oc cu r durin g takeof f and appr o a ch f or l and in g .

Fig u r e IV- 3 2 illu s trat es th e effect of e om bu s tor p r e ss ur e on smoke number.

_m oke number i s a par am ete l rel a ted to th e density of the e xhau s t smoke. A s the _ sm o ke density in c r e as eS, th e sm oke number inere _ e s . In gen eral, the sm o ke be c o mes visible _ th e sm o k e n u m b e r exceeds a va l ue of about 25. Th e thresh ol d of visibl e sm o k e m ay va ry , h o w e v e r, d0p end ing o n s uch th i ngs a6 a t mos p h eric co n diti on s , the en g i n e e xh a u s t di ame t e r , and t he posit i o n f ro m which the sm o ke i s bei n g viewed. The da ta show n Fo r a J- BV c o m b us to r a r e ty pi cal o f ear l i e r e cru - bu s terS, which were pr u po s e I y de s igned with ri c h p rim ary zon0s t o _ ha nc e t he al t itud e r el i g ht characteristic s o f the engin e . S m ok e n u m b0r is . _ how n to i n cr e ase rapidly with i n c r easin g pressure. These data were obtai n ed from a si ngl e J-5 7 co m bu s tor operati n g at an inlet te m perature of 6000 F , a f u el-air ratio o f 0.01 $p an d a re f erence velocity o f 54 feet per seco n d.

More rec en tly , Pra t t & Whitney Ai r cr af t has bee n abl e _ o r educe the s m ok e n u m ber f or the J T BD e n gine from a value of abou t 60 to a v al ue below th e visible th res h old of smoke. This was done by al teri ng t h e pri m a ry zone design f or this co m bu Sto r. Si m ilarly, more recent e wgi_ es , such as the CF 6 a n d JTgD, have been de s i gn ed wi th smoke numbers tha t are well be l ow t h e visible threshold of s m o k e.

The sh o rt-length , do u ble- ann ulu s c o mbus to r and swirl-c an combt l s to r w h ich l/ were d es crib ed previously were shown to ha ve a n extr eme ly low smo ke number at a pres sur e o f about 6 atmo s pheres. Data at higher pr essu re s are not avail- !

able fo r th es e two configurations; h ow ev e r , we _ p e ct that their Smoke n _mbe r i_ at h igher p ressures Will _ emain b e low th e th res h ol d of visi ble smo ke . Com- _ bus te rs A an d B are tw o _ peri men tal s eg m e n t co m b usto rs that were d e signed to I have increased primary zone ai r flow and incre as ed m ixing inte _ ity. The dif f er-- Ii ence in smoke nu m ber be tw een combtts to r s A and B is du e to a dfff e r _ ce i n pri- i m a ry zone airflow. Combus to r A, whic h has a _ x _ tre m ely low smo ke number in i_i • rela t ion to the o th er co m bus to rs shown on t_ hec u rve, doe s not ha v e a s a tis _a c to r _ ti • : al titude relig h t capability, an d further work W oUldb e r _ q ul r _ t to improve its re- ii li g h t capability. !i A rec en t su r vey (re f . 1 ' 0 haS indicat ed th at th e cofltributio n t o the o _ , e rnll urb an air pol lution problem by Jet aircraft amounts to le _ s th an 1 percent co m - q

J

par ed to o th er contril _ tors, s uc h aS t h e au _m obile _ d s tationa ry power s o _L rc es .

• HoWever, it s cont r ibution to the local pol l tt ti on in t h e vicii _ ity of th e ai rp o r t ca r l* ' be significant. The gas eous emission leve l f or a typi cal tur b o f afi en g i n e is co m - pared to t h at oi an auto m obile eng in e in th e fo l lowi n g table: 11 2 / " , 'r ]_ n gin e Operati ng m_ l e E m i ssion i n de x_ • l b / 1 00 0 l b of fue l C O i Ilydr o c a r b on NO Turt _ fad idle 50 to 17-4- 10 to 76 3. O HI L__ App ro ac h ? to 9 | to 10 3. ?

Ta keoff 0.7 to 1, 3 0.1 to 0.6 4, 3 , , ii ill i i Auto m obile Ov e ra l ! average 104 l0 18 i p_ The data shown for the tu rbofan were obtained fro m r eferenc e 1 T a _ d a re repre- :" sentative of en gines such as the JT3D and JT S D. The da ta p r esented for the i_ au tom obile are repre s en ta tive of an auto m obile tha t m eets the 19 7 0 Federal standard for hydrocarbons and carbon m onoxid e emiS S ions an d th e 19 7 1 standard i for nitrogen oxide emission, q:he auto m obile data were c al c u lated from the Federal standardS, which a r e expressed i n grams per vehicle mile, by assum ing i_ J an average m il e age o f 12 m iles per gallon. Three dif f er e n t operating nlod e s are l isted f o r the turbofan engi n e : idle, app ro ach, an d takeoff. An ov e rall averag e J mode is shown for th e au to mobil e . The exhatlst e miS sion level is ex presS ed i n terms of pou n ds of po llutant pe r thouSaftd pounds of fu e l bu sed .

The table indicateS tha _ the g r e atest q u antity of carbon m of l cxide and un- burned hydrocarbon s produc ed by th e tu rbof an occur during idle conditi O lls _ whil e ...... _ th e qu an tity of c ar bon mo n oxide and unb u rned hydrocarbon emis sionS at apprOach f and ta k eoff conditions are relatively s m_ tll. Th e nitric oxide emis sion level pro- I t duc ed by th e tu r bo fan is large s t at takeof f , btit even here th e m axi m um value is f app ro ximately one-quarter o f that for th e au t omob il e. Fo r th e pr es ent, the m ost significant po llution from Jet aircraft occu r s in th e Vicinity of th e airport dtle to _ ,,' , th e large em is sions of carbon m onoxid e an d tmlxinled hyd ro car bo n s during opera- _ lion at idle. In the fut u re , th e r _ lative contri bu tion by J 6 t aircraft to th e over _U l , ' urba, pollution p ro ble m m ay become more s ignffic an_ as ai r tr af fic i n creas es an d as emission controls are applied t o th e auto m obil e arid to ind uSt ry.

FigUre IV-33 illust rat es t h e effect o f c om bu etoz, op era ting c on ditions on th e ' ex haust e _ l is eion level for a typic al gas turb in e co m b _ tor. Th e da ta s bo_ wer _ obtained f f-o m a s ingle J -5 _ c om busti o n |in er. TSe exh _ { e m is n ion l _ vel in pp _ n is plotted ag ainst a correlatit _ g pal'ameter (desc r ib ed previou J iy) Which co h t _ tins its o \ inl e t pr ess ure, i n let te mpe r a tur e , a nd r e f e ren ce v e lo ci ty. D _ t a were o bt ai ned for a rang e c f in l et p r essuref of l to 2 0 a t m o s ph e res , a r _l_ e of inl e t t em per a - ture s of 800 to 600 ° F , an d a ran ge o f reference v e l oc iti es o f 29 to 150 f ee t p e r se cond. The va l u e s s h o wn fo r t he p ar Ual ly b ur n ed h ydr o c a rb ons a re e xp ressed as th e total a mo u n t o f o r gani c c a rb o n in ppm car b o n . Th e ex ha ust emissions f or th e u n b u rned hydro ca rb _ n, _ and ca rbon m on ox ide a re S ho Wn to in crease as the valu e for the c o rrelating parameter is reduc ed . P l ott ed on the same figure i s a curve _ fo r co m bu s tio n ef f ici e ncy against the corr e lati ng param e t e r, A s shown in f i g - ur e IV-2, as the valu e o f the correl a ti ng parameter d ecre as ed, th e co m bu st io n _, efficiency d 6c r es sed. The increase in th e carbon mon oxid e a n d u nbu rn ed hydro- I c a rbon emission l evel is dir e ctly proportio n al to the de c reas e in com bu stio n effi- ciency. T he data shown are f o r afixed fuel-air ratio o f 0.013. At lower fuel- air ratios, the drop in co m bustion ef ficiency and the increase i n car b on monoxide and hyd ro car bo n e m issions become even greater. At idle conditions _ the val u e o f PT / V is low. Tiffs f acto r , co m bined w i th a v al ue of fuel.air ratio lower than that shown in the figure lead s to lower co m bustion e f fici en cy and caus e s e ve n higher hyd ro carbon an d ca r bon monoxide emission levels. At high values of the correlating p arameter such as occur at takeoff an d cruise, combusti o n efficiency is hi g h a n d th e emission of hydrocarbon an d ca r bo n mo n o x ide level _ off an d is at a min i m u m .

N itric o xide i s f o rmed during any co m bustion pr o ces s involvinf _ air. T h e JL , • if ' amount of nit ric oxide formed is a function of flame tempe r ature and dwell time.

" The flame t e mperatun e increases as the co m bus to r i nl et te m perature in e t ea v es i_ ! and aS th e pri m ary zo n e f u el-air rati o approaches a s to ichiome t ric value. Dwell ti m e is affecttgi by combus to r length and velocity. For t he r an ge of dat 4 s hown, _ th e nit ri c oxide emission level g rad ually increases an d levels off as the correl a t- ing parameter is increased. The variation in nit ri c oxide level th at is sllown i s _ p - i mai nl y a t tribu t ed to incre asin g inlet te m pera _ r es and to reductions in referenc e velocity , which lead to increased resid en ce time.

Nit ri c oxide e mis sio n levels have rec e ntly bee _ m e as ured in th e hi[ _ h- te m pera tu r _ m odular co mbu s to r that was d es cribed in th e preVious s ec t io n .

Thes _ data are shown i n figure f V -34. The n itric o xide c o nc en tratio n p l otted on the o rdin at e is _ pre i Jsed in p pm o n th e left ordi n ate a n d i n te r ms of an e m i s sion in dex in pou n ds of nitric oxide per thoUs an d po unds of f u el on the r i ght o rd i n at e .

Data were obta in ed fo r an inlet te m per _ ttu re of 60 0 0 F an d a ran g e of co m bu _ tor e xi t t em peratur _ fro m about 1600 ° to 3600 ° F. The emis s ion level is it) pp m at an exit te m peratur e , of 160 00 1_. The concen tration increases a n d levels nff to a v a lue of a bo ut 1 00 ppr _ at 3600 ° F. Si m ilarly, th e nitric oxide e m is s io n index in- creases from a v al ue of abo u t 0.5 pound per thousand po unds of fuel bu rned at 2600 ° F t o a peak valu e of about 2.9 pound s per thou s and pound s o f f uel at _ 400 ° F and then decrease s and appro a ch e s a v al u e of ab o ut L _ pounds per thousa v d poun dS of fuel at 3600 ° F , The curv e f o r the e m ission indeX iS diff e rent than that f or th e em is s io n expressed in pp m sinc e it is af f ected by the inc reaaiu g fue l - a ir ratio, which approa c hes S to i c hio m etric aS te m perature iner ea a es_ A refer en ce : data po i n t shown f or a co n v en tion al J - 51_co m tmstor on th e same figure has a value o f a bo ut 6 0 pp m at about 2 6 00 ° F co m par ed to the value of 1 0 p p m f or th e mo dular co m bus to r. The lower nitric o xide e miss ion level f or the modular c o m - bustor is attributed to reduced re si d en ce time.

The a p p ro a ches to reducin g Jet aircr af t exh a ust e m issions ma y be s umma - ri zed as f oll ows: !

(1) Til e e lim ina tion of visibl e smoke is obtain ed by imp rov ing p rim ary zone fuel - air mixin g and _ voidin g fuel -ri ch region s in th e p rim ary zone. Howev e r, smoke nu m b e r has been Shown to i _ c r eas e with p ressure and th e t z end to higher .... operating preSsures m ak eS the proble m mor e diffic ul t. Further m ore, m ethods used to reduce smo k e m ust not _a criflc _ the alt itud e r el i g h t capab i lity of the air - c raft.

(2) Carbon monoxide an d u n b urn ed hydr o c ar bon e miss ions appear to be mainly a problem in th _ . _ .cinity o f the ai rport durin g idle operation. • A S has been Sh o Wn _ the emi s sion lev e l f o r c a rbo n m onoxide an d un burn ed hydrocar bo ns can be r e duc e d by i m proving co m bustion eff i c i e n cy. One way o f doing this is to re- duce the number of fuel entry poiuts durin g idle operation in ord e r to improve fue l a to m ization an d to opti m ize the loc al fu el-air ratio in the p rima ry zone.

(3) The trend to higher inle t an d exit tempe ratur eS tend _ t o inc r e _ _S e th e a m ount of ni tric o xid e for m ed. It a ppe a rs , however , that ai tric o xi d _ emiss i on c an be r ed uced by minimizing combustion residence time, as has been done in th e h i gh-te m pe ra ture m od ul ar co m bus to r, t Work on th e e m issions proble m is j ust begi n ning. These app ro ac h es are not nec- essarily final solutions, but merely indicate pO s sible dir e ctions to _ in elimi- nating th e problem.

CONCLUDING REMARKS

There are four m ajor achieve m ents of th e s hort - comb u stor - t echnolo b_ y pro- grams . We have s hown that co mbu s to r _ f or ai rcraft tu rbine a_ gt _ te$ c _ be b u ilt that s hould have the s ame perfor man ce as co m bu s tors that ar e § 0 p er c_ at longer.

The short co m bustors have pe rf or m ance s i m ilar to th e longer co m lms to r s in nearly all r e s pbcts except for altitude reli g ht. This p rog r am is co n tinuin g with increased em phasi s On i m provi n g th e alti tu de r e li g h t p e rfor m ance of these c o m - bu s tor s .

One of the short co m l _ sto r s, the m odular Swirl-can co m bust o_ , _ wa _ oper- ated to give turbin e inl et te m pera tu res as high as 3 6 00 o F. P e rfor m ance of t hi s co m bus to r w as very good; howev e r, much addi t ional work writ be-required to ob- t a in s atisfac t o ry du ra bility at high inle t air te m pera tu r e s.

The lack of visible smoke has been one of the outsta / _ d ing characteri s tics of the s hort co m b nsto rs. This i s due in part to th e relatively lean f u el-air ratio and the ve ry high in t ensity o f tu rbul en ce in the pri m a ry zones of these combu S tors. ' Fi nal ly, we hav e Shown that a redu c ed level of nitric oxide e m issionS m ay be ob- tain ed by reducing the residence ti m e in . _ co m bu Sto r. While th ese r es ults must stil l be consi d er ed pr el im in a r y, c ontro l of ni t ric oxide e m issio nS even at ' ve ry high tu rbine inle t temperatures sho u ld not be difficult to obtai a . , = R EFE REN CES 1. Rou d e bush , William H.: State of the A rt in S hort Co m bus to rs. Pres en t ed _ at the Sixth Co ng res s of the Internatio n al Council of the Aeronau ti c al _ ' _ Sciences, MUnich , Germany , Se pt . 9- 1 3 , 1 968. _ , 2. Ecke rt , E. R. G. ; a n d Bi rk ebak, R. C. : The EffectS of Slot Geomet ry on i_ Fil m Cooling. Heat Tl _ ansfer Ther m odynamics and Education. H.A.

Johnson , ed. , Mc G raw-Hill Book Co. , Inc. , 19 6 4 , pp. 1 5 0-1 6 3.

: 3. S to llery , J. L. ; and E1-Ehw an y , A. A. M. : A Note on th e Use of a Boundary- Layer Model for Correlating Film-Cooling Data. Int. J. Heat Mas s Trails- fe r , vol. 8 , no. 1 , Jan. 1965 , pp. 55-65. t 4. Hatch, Ja m es E.; an d Pape H , S. S t ephen: U Se of A Th eoretic al l _ low Model

'!

to Correlate Data for Film Cooling or Heat in g an Adiabatic Wall by Tan- i gential In jection of G as es of Different FlUid Properties. NASA TN D-130 , !l 19 5 9.

5. Perkins, Po rt er J. : Comparison of Test Resul ts form a 90 ° Sector and a Full Ann ul us Ad w tced T u rbojet Co m bus to r. NASA TM X - 52 7 0 7 , 1969.

6 . Kit ts , D. L. : Develop men t of a Short-L e flgtli T U rbojet Co m bustor. Rep.

PWA-FR-2433, Pratt &Whitney Aircraft (N AS A CR- 5 4 5 6 0), Mar. 1 8 , 19 6 8.

11 6 • f , j .

/ , _ | !

t \ V . Cle m ents j T. R. : A 90-De g ree Sector Develop m ent of a Short Len g th Co mbo s - ., to r for a Sup e rsonic Cruise T urbofan Engino. Rep. P W A-FH -$_ 90 _ Pratt & , W h itney Aircraf t (NASA CR- ' 1 2_$ 4), Aug , 6 , 19 _ 0.

8. Hu men ik, Francis-M.: Perfor m ance of a S hort L en gth-Tuz g _ oJet Co m bus to r In- s en s itive to Radi a l Di s tortion of Inle t Ai z fl o w. NASA T N D-5 §_ 0 , 19 7 0. i 9. Butze, Helmut F. ; Tr o ut, Arthur M. ; and Moyer, Har z y M. : Performance _ , of _ twirl-Can Tur bo jet Co m bu s tors at Simulated Supersonic Combus to r-Inle t _ Co n diti on e r. N AS A _T D- 4 99 6 , 19 6 9. ,, 10. Niedzwiecki, Richard W. ; an d J on eS , Robert E.: Com b usti on Stability of i Single Swirl- C an Cornbus _ or Modules U s in g ASTM-A1 Liq _ / Jd . F U el.. NASA _ TN D -5436 , 1969.

_i 11. N t edzwiecki, Ri 0 1 m rdW.; an d M oyer, H arry M.: Perform an ce of a . 4 K - _ • > i Mod ul e, Swirl-Can T U rbojet Combus to r Segm en t at High Temperatures Using !

AS T M -A1 FUe l . NASA TN D-559 7 , 1969. i 12. Niedzwiecki, Ri t_ hard W. : Prel!mtnary Te s ts of a Siml _ ar Tu rbojet i: Combu sto r. NASA T N D-5 6 88 , i9 7 0. !i , • 1 5. Rayle, Warren D.; JO n es , Rob e rt E.; an d / ez ied man , Rob ert F . : Experimental i !, EValuation of " Sw irl-Can " Elemen ts for H ydrogen-Fuel Co m bus to rs. NACA i RM E5 _ C18 , 195 7 . !, ' 14. Jon es , Robe rt E. ; and Rayle , Warren D. : Pe rf o rm auce of Five Sho rt Multi- i element T urbojet Com be stors for H ydl - oge n Fuel in Q ua rt er-Ann ul us Duct. ' NACA RM E58 D 15, 1 9 58.

i,. 15. Jones , Ro bert E . ; and Paw l ik, E ugene V. : . A Preliminary In v e stigation of • _ the Performa n ce o f a S11o rt -L O ngth Tu rbojet Combtm tor U s in g Vapo ri zed H y- t droearbo n l % lel s . NACA P _ ME5 7 J0 5 , 1958.

16. Niedzwiecki, Ri chard W.; Juha _ z, Albert J.; a _ d Anderson, David N. : Per- form an c _ of a Swirl-Can P rim ary Combus to r to OUtlet Te m perahtres of .: 3 6 00° F (2 95 6 K) . NAS A TMX - 529 0 2 , | 9 7 0.

1 7 . Anon. : Nature and Co n trol o t Ai r craft En g ine Exhatmt E m issio ns . R ep.

No. 1134 - 1, N orth e rn R es . and Eng. Corp., N ov. i 968 .

: ! M i i ? ....

Q • F ' a' ,. +.• lie , . , ' "' EFFECT OF C O MB USTOR I N LET A IR TE M PERATURE .i . ." O N LINER TE M PERATURE . • 16 0 0- IO - .ATM INLE T PRESSURE CALCUL A TED ., , 0. 0 1 9 FUEL - AIR R A TIO / _ , , , ' _ 1400 __ , I / I_ PERIM E NTAL .. ; / % " L , - ,

.:.. , / _If FLOW _ i_

_ • ; ": 800 _ " - LINERTEMP ;"_ I T "_ : I I I I _ I ,1 .

i. ..600 800 1000 1200 _ - : c s.s 6 7 z 5 COM B USTOR INLET A IR TEMPo OF I_ F i gu r e 1 " V - 3 i ; i : i . 1!

ii

• EFFECT OF PRESSURE ON TOTAL RADIATION i_ 3 0x 1 0 4 _ ' _ • EWIS COMBUSTOR B -i: .: : . 25 ,: 20 _ / ' " TOTAL l : - RADIAT J ON , 1 5 -- BTU / FT ' : / HR / , , / _ :; O , MBUSTORA !

10 _ " "_ A ' _ CULATED NONLUMINOU$ RADIATION 0 I I I I I S tO 15 2O _ 30 COMBUSTOR PRESSURE, , ATM cs - s 6 vz4 Fl gOre 1 V-4 1i# i EFFECT OF COHBU S TORPRESSUREON _ LINER TEHPERATURE L ( 3 00 ° F INLET AIR TEMP 16 00- 0.019 FUEL - AlP _ ATIO , _ L ,t 1 4 n0-- L!

IZ00- * _ LINER TEMP o F _ 10oo C AL C ULATED _ t 6 00 I i I I I '_ " 5 10 15 20 25 30 ,;: _ COMBUSTOR PRESSURE, ATM c . _ - s 67 Z 3 i_ F i g u re _-5 : _ ' i! , EFFECT OF COHBU S TORPRE S SURE _ ON | " LINER T EHPERATURE i : . .... , ( _ 0° F INLE _ . _.AL R -TEMP " • 1 6 00 - . .... . ..... 0.O19FUEL-AIR RATI O ' _ '[ j % _ ' , CALC F O R EMISSI V ITY- 1 CALCCOMBB " " _ .--...----.-

=_

1000- 800 - " ° ' v EXP COMBA c oo I I I I. , J 5 1 0 1._ 2 0 25 3 0 C O MBUSTOR PRESSUR E, A T M c s - s 67 zz F igureI V - _ 1 EFFECT OF COOLING AIR ON E X IT TEMPERATURE INLE T T E MP, 6 00 o F 6O 80 _ PR E SSUR E, 6 A T M COOLING AIR, 4 0 _ f % : , 3 9 5 0 (S T OI C HIO - i : , ; 20 -- METRIC) i' I '

o I I I I

' " : ' 1800 2 2 0 0 2600 3000 3400 3 800 4200 i ), .

c s -s 6 7 o<_ MAXEXI T TEMP , OF t i Figure I V-7 i_ K . C O OL I N G SLOT GEO MET RIES Ii I ,.

# % METERED WIGGLE H O LES S TRI P Figure IV -8

l

( b i COOLINGEFFECTIVENESS IN COHBUSTOR S REF2 _ I I ! i .4 - REF 4 J COOLING EFFECTIVENESS . 2 .1 -- COMBUS T OR DATA ,0 5 - I , I I I I I ,a I I.I I 2 4 6 810 20 4O 6 0 100 x / MS c s-s 6 . 19 F I gu reI V 4 COMBUSTORLENGTH AS FUNCTION OF AIRFLOW / i 40 / LENGTH,30 A NNULAR ' : i _ IN. t SEA LEV_i TAKEO FF AIRFLOW, LI _ I $i[C cs . s_n e Flgu re I V-10 "_ , ! , r ': i CONVENTIONAL ANNULAR COMBUSTOR (, • , !!

R Y.- - -- _ ZONE Z ONE ....

, ' L; 8- 5 6 70 _ ii Figure IV - I1 t : , -

i

CROSS-SE C TIONAL SKETCH OF I DOUBLE-ANNULAR COMBUSTO R t DOUBLE-ANNULAR COMBUSTOR, . , i ' VIEW LOOKING UP S TREAM . ; " ; • ° 'i ' d ONE-SIDE - ENTRY GOMBUSTOR, VIEW LOOKING UPSTREAM C - 7 0 - 1 7 4?

C 8 - 5 67 3 | W FigUre I V-I _ i++ CROSS-SECTIONAL SKETCH OF ' SWIRL-C A N COHBUSTOR , .

P I t +,, - - 20 ;25 IN . _ c 8 ., _6_m - Figure IV -1 6 I_ ........ r ....

, p SWIRL-CAN COMBUSTOR, VIEW LOOKING UPSTREAM C-70 °2 9 23 CS-5 6 733 RgureN - 1 7 SWIRt-CAN M O DULE ' CARBURf OR SWlR L ER FLAME STABILIZER CS- 56697 F i gure ]V- I 8 . _i _ ' 1956 SWIRL-CAN COMBUSTOR "- . C-4z gz9 • . • C. S - 5 6 7 29 .... , Figure 1 V - 19 Ik -

T O TAL PRESSURE LOSSCOHPARISON

9 - " ' " D O UBLE ANNULAR '_ TOTAL 7 PRESSURE 6 LOSS, ENTRY . ., z _ . .. ,, ' . ,. . ,.,.Z _ _ _ S IDE 5 -- __ WIRL CAN

4 ill

" 3 _ '_ '_ 1 " " ANNI_LAR I .24 .26 .28 . _ .32 DI ff U S ERINLET MACH N O. c s .s6 7 zo Figure I V - 20 :_ ' PATTERN FACTOR COMPARISON • • _ A N NULAR •6 - _ TEST::; , , -;." r_ SEGMENT .... . ;; TEST S . 5 - _ ,' r / _ .

. . : ( TYPICAL . DOUBLE SWIRL SIDE , ' :- ANNULAR ANNU LA RCAN ENTRY i : i_ COMBUSTOR c _ - s 6 _19 t _ Figu r e1V - 21 L .

i r , b I: t COMPARISON OF BLOWOUTLIMIT S REFE R ENCE MACHNO. - 0 .1 : 20-- COMBUSTOR PRESSURE,10 -- _ LEANNU LA R PSIA _ SWIRL CAN ,.. ' _ ' _ , , SIDEENTRY -' _ ANNULAR COMBUST O R

I I I

_" o ZOO 4O0 6O0 INLET'AIR i"_ tP, OF cs . s 6 ?ze FigureIV - 22

i

' " _ / i I ' / • '," I M PRO V E M E N T OF B LOWOUT L I M ITS . : '"- ', DOU B LE ANNU LAR COMB U STOR "!',:. " COM BU S T O R E M A CH NO .

" " PRES SUR E , 10 15 - __ REFERENC PSIA _ O.1 " " :i" : ' ; . / "i .OP ' - , INLET - AIR TEMP , OF c s - s r, 717 .... F ig ure IV -23 DIFFUSER WITHSPLI TTER

El

• B LEE D A I_ - _ __--j / _ ' , DIFFUSER WITHBOUNDARY LAYER B LEI[ _ .D C$.5 6721 FigureIV - 24

SHORT DIFFUSER VELOCITY, PROFILES

WITH BOUNDARY LAYER BLEED BL EED R AT E, O 0 2 ix 4. 3 0 7.0 : VELOCITY R ATIO t - I , ,_ ': 0 20 40 60 80 100 DUCT HEI G HT , % cs-s671 s _ .

F i gur e I V - 25 li • D I FFUSER TOTAL PRESSURE LOSS TOT A LLoss, PRESSURE % 3 2 0 2 4 6 8 10 !

BL EED, _ cs-s 6713 F _ gure IV - 26 { 1 3 0 ] I , / ?

r. 1

\ " HIGH TEHPERATURE COHBU S TOR _ I _ 1-- _ LOCA L IZE D 42-1 N .

RECIRCU L ATIO N I ZO N ES P . { • i. _.. : -_ 11. 4IN . = • ' = 2 0. 9 IN. =- " C S - 5 6696 " ' Figure IV - 2 7 : . HIGH TEMPERATURE COMBUSTOR !_ AFTER 3600 ° F TEST " _ COMBUS T OR ... ..... " MODUL E f _ - INNE R LIN ER • RE M O V_ O UT ER LINER J " c-7o- zo ls C5- 56 7 3 Z FigureIV - 28 t 3 l i i' COHBUSTOR HODULE DETAILS i k +!

' , CARBURETORSWIRLER i " STABILIZER t •" PLATE F L AME. , _ - INCP J _ SED CIRCUMFERE N CE STABILIZER - / CS - 56 6 98 Fi gure 1V-L R COHBUSTION EFFICIENCY - , _ I ) z:COMBUSTION EFF 97 .

COMBUSTOR 2800 " , 2000 " Tt (_ 0° _ PII _ 4 3 . 5 -( £1 PSIA 49 -_ LB I SEC

.m . o_ .m ._5 .o +

FUEL ' AIR RATIO cs - s ++1 4 Figure IV-30 F COHBUSTOR PR ES SURE L08 $ ,. . _ ,, 12 - - INLET TEMP , 6 000 F "' 8 COMB USTOR 6 - PRE S SURE I0- _ e / a_ l _ M A 2722 , . 1 L LO S S, _ , 4_ - J • BU R NING_ 1'0INDI O A TE D " 2I'- _ oCONVENTIONAL COM B USTOR, ol I ..1. I I I I .15 .20 .2 5 .3 ) . 3 5 .40 .4 1 ; DI F FUSER INLET 5U k C H NO. cs,.s 6 ' t l z FigureIV - 3 1 / .

e EXHAUST EMI S SIONS OF A J, - . 5 7 C O HBUSTOR l ooo . COM B .U _ T I ON _ F E_ I O0 , , EXHAUST in FUE L / AIR RATIO, 0 . 013 EMI S SION S , 6 0 0 80 COMBU S T PFM _ FF, :, 8O0 _ tc )O .,

' l t\ - t '° ' "

r NO _ ' i 200 _ \ , ' . , ' - HYDROCARBON S . .-- ] _

o 2 o 4o e o 80 lOO _ o _ x z o 4 _ , !_

COR R ELATING PAR A METER, PT N c e . _ , Tl , : :, Rgure [V - 33

i!'

NITRIC OXIDE EHI $$ 10N ZOO . , dO " 4 t 60 3 PPM LB I IO00 LBOF FUEl.

40 i _ iNLEr T , _ P,eO0O F ._ 2 let n LB I IO00 L8OF FUEL , , _ O !.... - .1 [ J I 0 ; " Z 0 10 2000 240o 2SO0 3 200 _ 00 C OMBUSi'OR EXIT 1T _ P,OF cs .5 67 1 o Figu r e 1"V-34 I i S 4 I %

N 71 - 945 6

V, V / S TO L PROPU L SIO N !

N e well O.sanders , Jam e s H. Dl e drlch , James L. Ha sse ll , J r . ,*

David H. Hlckey , * * Roger W.t.uldens, and Warner' L. S t _ vart

ThiB p a p e rd e _c r l b ea th epr o p ul A i n n t e chnolo g y for e om m er c _ l a i rplanoa d o - s i g ned f o r short t ak eoff a nd lardi ng ( ST OL) o r f o r vort i c a l t a k e off and la ad in_ (V T OL), som e diff ere n ce s betwe en those air pla nes a nd c onv en tio nal t ak e of f an d l a nding (CTOL) a i r plane s are il l ustr at e d b y th e t ra je c tor i e s in tig _ r e V -1 .

T h e CTOL ai r plan e uses the li f t o t wi ng s c aus ed by f u r w a rd sp ee d to g e t off the g round. A r elu t i v el y lon g run wa y is necessary f or a c ce ler a tiof _ to f l y in g s p ee d, In the $TOL aLcr Mt , w in g lif _ i s a u gme nted by t h e en g ine power. Engi ne po w er c an be ; :_ u se d to increase t h e '.* . f tin _ e ff ectiveness o f t h e Wi n gs or th e engi n e ca n prodttce di - r ect rif t t o aid the takeoff , In both caSe s , th e takeo f f speed is lo wer e d, and acce ler - , a tion _ sta zm e is reduced, T h e VTOL airp lane lifts o ft a t zero air sp e ed, and th e w i n gs do a ct prov i de lil t at a ll ; e n gi n e s provi d e t h e li f t directly. In forward flig h t, i_ the wings do provide l i f t. l_ We recognize i mm ediate l y a greater de m and for en gi he th ru st, Or l if t , a s we pr o ce ed f rom CTOL to STOL and to VTOL ai rplane s . V a lues o f e n gine thrus t t o airpla ne gross weight mi gh t be as fo llo_ s: i ' ' CTOL airp lanes , 0.3 , , } BTO L ai rplan es , O.6 _, VT OL airplan e s , 1.2 T he des i red thrust o f t he STO L airplane is double t h e thru st of the CTOL ai rplan e , t : T h e VTOL airp lane requires an o t he r doublin g o f the thr u st. T ltes e valu e s _ r e g _n- era li zat to ns; they ar _ i n t e nded _ o conv e ys se ns e o f m a g_ .itud e o n_ v. Substanti al d _ - viatiott8 fro m these value s w ill b e _ vi d dnt in this pd.per.

: InCre _ tstng tlu ' usts mea n incl' e asin g engtn _ weig h ts. Un le ss the thg ' ust to e ngin e w e ight ratio is improv e d , the e ngi ne we i g ht s can beco m e too L _fge f br eco nm ttical operatio n . Today' s CTOL ai rplan e s have engin e th ru st to Weight r atios n e a r 5. F b r STOL aircraft a n i m prov e m en t in the ei l gi n e thtm s t to weight r _ tio is fle b esS a r y to _ ' , off s et an ot h er w is e 2 to 1 increas e in e ltgl n e si z e. A valu _ of 7 i _ sug g ested , btlt We *NAS A -La n gleyResearch C e n te r. i * * NA S A-Am o sResuarch Cent e r.

"I X fi t w o ul d b e happ y wi th 2 0 f or all the s e e ng i n e s , F or t h e VT OL , a largo impro v e men t I n e n g in e t h ru a t to w e ig h t r atio i n nc -_e d t o o ff ae t a n o th e rwi se 4 to 1 t n ere ane In e ngin e : i s iz e . A n engin e t h r us t t o we i ght r a t i o of 1 0 w ou ld c u t t he p o te n ti al weig h t p enal t y in h aft.

A n w e a ll know , t oday ' a con v en ti ona l a irpl an e s a re cr i ti c i z e d b e caus e o f t he n o i s e the y a re _ t o . T h e no l n _ p r o bl em w i t h t_ TOL an d t h_ VTO L airp l a ne s wi l l be mo r e _ o v o r e b e c aus e t h e s e a irp lanes wi l l b e uf _ n d a t a i r p o rt s in hi ghly p o p ul at ed _ u rb an area s , A n o i_ o limit of 0 5 eq ui v alen t p e r c e i v e d n o i s e dec i be l s (EPNdB ) a _ an al t i tud e o f 500 f oo t I ro n b een s ug g e st e d . T _n limi t IH 1 0 t o 2 0 d ec ibe l s m a r e a tr ia - g oa t th _n tl _ opro _ o z_t fed era l reg u lat ion _o r c ert if y in g now CTOL al l'pla no_ . ' rh l o _ tr inl_ o nt nol_ o limi t hrt_ a ft r t _ t o r de r effect on t h e t _ o l o c t lon ot pro p u l s i o n, ttr r ang o - t t u o nt an d t he o nl_ i n o cyc l e, The n ext i te m 1 8 s afe t y, p a rti c u la r ly wflh r es pe c t to engi ne f a ilure. T h e CTOL has an e ng i n e-o u t s af e t y p r ob l e m , a ls o, b u t ,_ b e canso af t a t tak eoff an d l a nd ing is , c riti call y de pe n dent o n t he engi n e s i n $T O L a nd V TOL a irpl _n_ , e ngine fai lur e b e- : Co me s o v e n m o r e serious a n d l_ s a muc h gre a ter e f f ect oa propuls i o n s yst 0m dos iI_n , i_ Hig h c r u is i ng spe e ds, appr o ac hin g th _ speed s of CTOL airpla n eS, a re demanded i_i11 b y pot en t ial oper a tor s of STOL a n d VT OL ai rcr a ft. These la st tWo tte m s j s peed a nti __ eco n o m y of operat i on, fa v or t h e turb o fan an d t h e a t rpu m p types of en g i n eS. Our dis - ,i_ cussi o n will be lim it e d to t h ese two e , _ gl n es. _ In t h e i m me d iate l y s u ccee di n g _ , orti o ns of th i s paper , STOL airp la ne propulsion i_ is d iscussed , and, in t h e f in a l se ct i o ns , V ' P OL airplane p r opu l sion i s discuss e d. : . _:_'_

ST O L A IRPLANE PRINCIPLES

Ther e are, a t present , pr o posal e f or a wid e v a riety o _ d _si gns for S ' I 'OL ai r- planes. One o f the m ain rea s o ns for t his i s tb af th e STOL ai rpla n e d esig n s are se ll - t striv e t o the requir e ments placed o n t h e airp lan e i n t e r m s of fie l d l engt h a n d rid e qu a l ity , and these requirements a re, at prese n t , n o t firm l y defined. W e can i ll us - tra t e this se n s i tiv i ty first wit h th e factors t h at deter n_tn e the thrtmt reqU l_ ' _ d in t h_ a i r plane. In fig u re V-2 the ratio of t h e s ea level st a tic thrust to t he a irplan e gro s s weight to p l otted a s a function o f F A A field le ngth f o r two wi ng l oadings , 70 a g O .

100 pounds per squ ar e foot. So me ai r li n es hav e a sR ed _ or air planes _v ttlta field : le n gt h of 1500 f e et, o r about tw o times a city bloc k . Conve n t to tu il ai rp lan e s require field lengt h s f ou r a n d five times t h is value. Th _ s h ort , l _ 00-foo t fi e ld w ould , o f c o ur s e, r e duce airport l and co s ts in p o pulated areas.

High wi ng l oading s like th os e m e n tioned are de s irab l e fol ' g ood rid e a n d for e c o- nomic hi g h s p e e d cruise , E ve n for a very e ffici e nt lifting system, s p e cifying a 1500- ' o • | !

\ # , t + foo t fiel d l eng th an d a 1 0 0_p o und-p e r- sq uare-f oo t w ing loading r equi r es a t hr u s t to g r oss we i g h t ra t i o of 0, B , or a bout tw i c e t ha t o f CTOL a lr pl a n e_ . If th e re q ui red field l e ngth i s incr eas ed only _ 0 _ fe e t t o 1800 f e et a nd th e w in g lo a din g i s d _ cr ease d from 100 to _0 pound s p er s q u_re fo ot, a th r u s t t o g r o s s w eig h t r atio o f 0, 38 c a n s_f ic e, This i s a sig ni f ic an t differ enc e .

Th e specific v a l u o a of thru s t t o gross weight ratio depend o n t h e a trp _ ne d v_ ign as I n di c a ted b y the w l d _ _ h a dod area, F o r exa m ple, _ he ro _ iu l re d t h ru st t o g r oss w o l g ht r atio would e xce e d 0, 0 for a l _ 00- f o ot f l old l e n gt h; T he im po r t an t pc Jl nt i _ tl _ tt STOL a irp lhn es r equi r e hi g h thru 6 t t o g ro sA wei gh t ra tio s wh o6e v alu e d e p e nds o n f ield l e ng th , win g l o a di n_ , a nd airpl a ne de s i g n, If tile $TOL a irplan e can ta _ e o ff from a _ ho r t field , it should a l s o b e able to i azz di n th e sam e fie l d l e n g th. Wheth er it w i ll be able to do s o dep en d s lar gely on th e l_ ft coe ff i c ient that t he wing ca n deve l op, This i s i llus t ra ted i n f i t_ r e V-8 w he re th e r equi r ed l ift coe ff ici en t is plotted a s a run . Ion of field l _ } n gth f or t he same two wing lo a di n gs. To a c h i0ve a 1 5 00-foot la n d ing field l e n g t h with a I00 pouncrp e r _ qua _ 'e foot w i ng l o ad i ng requires a lift coeffici e nt o f ' / . 0. If th e fi 0 1d l ength is increased to 1800 feet and the wing loadin g is re duc ® d to T0 , the r _ quired CL is rec _ ueed to abo _ lt 4.0, a l ar g e dif f0r e nce. I n ge neral, STOL airpl a n e s r eqUire hi g h lift coefficien t , w hich , as was t he ease with th e installed thrust, a _ e sensitive to field length and w in g l oadi n g.

I n conside r i ng ho w th e r e quir Q d hi g h lif t c o efficients can b e a c hieved, keep in :, ' min d th e hig h thrustS installed in th e airpl _ ne. A pow e rful aerodyna m ic p r inciple for achi e v ing hi gh lift coeffic i ent ha s b een kno wn fo r man y years and i s shown in fi g - l ' ure V-4; it i s t he J a t flap w in g co n c e pt. Some of the propu ls i _ force of the engin e i s ex ha usted dow nw ard a lo f _g t h e tr _ ailin g edge of t he w in g . In it s 0fleet o n w i ng li _ t, this J e t sheet a cts like a n ext en sive m e e haltieal flap. Also, the d e flected J e t its e lf ' , :, con tributes to t h e I nt .

l m ple m e n t a ti ng this _onc ept g i ves r is _ to a n umber of probl emS : !

(I) _ l o Wis a sp mlw iseJ e t sheet aeh l ev _ d ? I" , _ .. (2) How is flow attachment o n the wi n g u Fpe _ surface _ fla ln ta in ed at h igh lift co- I : efflcien ts ? , , (3) H oWi s the _a i l u r e of o n e o f the e n gines tll _ t p roduce s the J e t sh eet de _ t R with? _ !

T h e ai rplan e designers hav e d e vis e d m any designs in at t e m pt S tO solve th _ e pr _ blenm. Also , as dis c uss ed earlier , th e exact r e qui t 'e / n en ts for a STOL ai _ lafle ar e not firmly e s t a bli s h e d a nd th e s e have a st rong effect o n t h e airp la ne desi g n. It is fo r tll _ s e re aso n s that th e re is at pres e nt m a n y STOL,airplane de s i gns ¢ _ td e r cob- , ' side r ation.

Three t ypes o f ST O L ai rpla n e de s i g ns will be dis cu s sed, na g / e ly , the ex t e r nal l y blo wn fl ap , the au g / h ea ter W in g , and the mu ltipl e h ln ai r plane. F o r each desi gn , the concept and propu ls ior, system will be described, and s o m e of its advantages and problems will be discussed.

EXTER NA L L Y B LOWN-F L AP AIR PLA NE The externally blown-f lap concept is i l lustrated in f igure V - 5. The exhau s t flow f ro m th e underslung turbo f an e n gi n es i s directed onto the hi g hly deflected fla ps s o that the flow i s tu I'ned downward. Lift is generated not only as a result of the re - directed thrust but also as a result o f t h e jet fla p ef f ect which _ o m eS f ro m th e spreading of the flow outward to Cover most of the f lap span an d the resultin g exten - slon o f the physical flap by the jet sheet so f ormed. Flow separation on the upper surface is prevented because some o f the engine ex ha ust pas s es through th e slots in the f la ps to en e rgize the boundary layer. Moderate wing sweep, like t ha t o f conv e n - , tional jet transports, is re q uired for good s preading of t he flow. Also , t h e hi gh by ° pas s ratio fan e ngin e s currently being dev e loped are ideally suited for the externally blown fla p conc e pt becaus e of the high ma ss flow at re la tively low velocity and the r e latively low t emperature of th e ir e x haust .

A pho tograp h o f an e xt ernally blo w n - f la p model being tested in t h e Langley full I scale tunnel at a v e ry l ow tunnel s pe ed is presented in f i gu re V - 6. The str ea m o f _,_ smok e pa s sing over the wing and th en sharply downward i s clearly visible. Thi s _' flow d e f lec ti on i s an in di ca ti on of lift coef f icient s tw o or three tim e s gr ea te r than !

f or the w in g alone.

An obvious problem with th e exter na ll y blo w n - f la p concept is engin e failure be - cause o f the a _ ymmetric loss o f lift and th e re s ulting ro lling mo m ent. This problem " .., : can be mini m iz e d by loca ti ng th e en gi nes close inboard , but th is place m ent do e s no t ' elim inate th e problem. Correc ti on of engine - out rolling m oments has been inves- t tigated very th oroughly at the Langley Research Center. In fi gu re V-7 , the rolling- moment co ef ficient i s shown plotted against lift coefficient. For the cas e shown here, the ai rpla ne has four en gi nes and a thrust-to-gross weight ra ti o of 0.6. The m axi m u m li ft coefficient is 9 , an d , of course , there i s no rolling m oment asy m - m et ry . The solid curve i l lustrates the asymmetric rolling moment f rom f ai lure of the left _ utboard engi n e an d also shows that f _ r t hi s case with three en gi nes oper - ating at m axi m um th rust, the maxi m u m lift coefficient has dropped o ff to about 7.

The real ch a llenge ha s been s o d e vi _ e a n aerody nami c control t ha t provide s sufficient roll control po w er to trim out the a symmetry of th e failed engine without l ncurring _ ad di tio na l loss of lift. The rolling moment can be balanced by decre as ing th e flap _ defl e c ti on be hi nd th e tw_ active en gi nes and incr ea sing the f la p defl e c ti on on , the side i with the failed engine. This technique i s called differential flap control. Re s ults of ° , r " t \ te s ts u s ing t h is tech ni que are shown by the da s hed curve wh i ch ind i cate s that the rolling moment has bee n tri mm ed without further red u ct i on of m_ t x i m u m li ft co e ffi- cient below 7.

Safety considera ti ons di ctate that the a i rp l a n e operate at some marg i n above s tall speed. F o r a 2 0 percent speed ma r gi n , t h e saf e usable approac h li f t c oe ffic i e n t _ i s about 4.7. Th is l ift coefficient permits operat i o n into a 1500-foot field With a wing _ load i ng of _ 0 pound s per s quare foot. :i Although they wou l d not be used t o trim the en gi ne fa i lure asymmetry because of _ the l i ft p enalty t ha t would occur , spo i lers do pr o v i de a very healthy roll co n trol i n- i : - cre m e n t for n o rmal maneuver i n g cont ro l as i s in di cated by the curve at the t o p.

I n an alternate m ethod als o tested at Langley , h i gh pressure a i r is bled f rom th e eng i nes a n d i s cross ducted to the opposi te ai l e rons to provid e bo un dary - la yer con- trol. When e ngine failure occur s , t w ice as much bleed air is directed to the drooped _tl le r on on the side with the failed en gi ne a S is to the a i lero n o n the side wi th the two active enghles. T hi s bo u n dary - la yer c o ntrol S ySte m is qu i te powerful a n d prov i de s / : r o lling m oment tri m very si m i la r to that sho W n wi th the d if f erent i al fl ap control i_ technique. In add i t i on, such a syst em has the dist i nct adva n ta g e of operat i n g pas- • sively i n that m o st o f the r o lling moment asymmetry w o uld be t ri mmed wi th out acti o n on th e part o f the pilot. En gi nes capable o f supply ing sufficient bleed a i r to power the bo undary-layer control would be requ i red or pr o v i sion for auxil i a ry power from ano ther s o urce w o uld be nec e ssary.

A ca ndid a t ee n_ e f orthe blo w n -fla p ST OL airp la n e is s ho w n in f i g ureV- 8 . Th e " tre n d in en gin es f o r c on v en tio n al tra n sp ort s has b ee n to h ig her bypaSs ratio and l o wer f an pressure ra ti os. Co n ve n tio na l tw i n -spool e ng ines of today c an meet a ll require- m ents for th i s ser vi ce except a n o i se li m i t of 9 5 E PND B at an alt i tude of 500 feet.

Acous ti cal treatme n t o f exist in g t urbo fan eng i n e s w ith bypass ra tios of 6 cou ld pUt us wi th in 6 PN dB of t h i s l imi t . The ne xt secti on w il lsh o w t ha t s till l o wer core je t re- t lo c iti e s , l o w fa npr es sur e ra ti os, an d s ound a b s orbing t r eat m e n ta r e r e quir e dto m eet th e 95 E PN d B noi se limit g o a l. N e w f a n tech nolo g y e mp l oying p rop elle rt ech - ni ques migh t b e a ppliedto t h e se n ew e ngin es .

In summ a ry , w in d - t unn e l e xp e rim e n ts sh oW tha t the d es ir ed li ft c o eff i c ien ts an d c on tr olo f e n g ln e -o u t r o l ling mom e nt s ca n b e ac hi e v ed wi th an a irpl a ne t hr us tto gloss we i ght ra ti o of 0, 6. Further devel o pme n t o f th e tr e nd to lower fan press u re ratio e ngine s i s req ui red t o me e t th e n o ise l imi ts.

AUGMENTOR WING AIRPLANE The pr o pulsion req ui re m ents f o r an augm en_ or w ing airpl ane will be c or u s idered now. The question of instal l ed thrust required , core thrust - f an th rust split, and 1 39 e fan _ n d core pressure ratios will be examin e d.

Like the externally blown flap concept, the aug m enter Wing concept is a deriva- tive of t h e jet flap princ i pl e . Figur e V-9 shows a cross section of an aug m enter wing. The a ug m enter is atwo-di m ensi0nal e J ectorformed bythe biplane f la p syste m and the pr ess urized duct containing the primany nozzle. The attg m entor increases the th rus t of th e p r i m a r y Je t by about 30 percent. The secondary inlets f o r t he ejec- tor are shoWn in th e cross section. The in l e ts on t b _ top f lap remove th e bounda ry layer so that the external f low f ol l o ws the highly deflected flap. The di s charg e f rom t he augme _ or forms a Jet sheet and g e ftel _ z _ e s high lift co ef ficien ts by the J et f la p ef- fect. As shown on th e plan view, the a ngm entgr typ i cally c overs f rom the fuselage : to 7 0 percent of th e span. For additional lift, the _ ie ro ns (typically e ov e l _ l ng th e outboard 3 0 p e rcent of the w in g) would b e drooped an d have boundary layer control. = The engi n e-out e ha r g cte ri sti e s of this concept ar e o n e of i ts Stro ng points. The i int e rconnected duc ti ng _ s hown in th e f igt _ e minimizes ul c er m ome n ts d _ le to engine i ' f ailu r e by p rovi ding unifor m di s tribu ti on of air in to th e f btp s y _ e m despit e t he failu r e .

of an y one of the engines supplying th e pressuriz ed air. T hi s lea gue , alo n g wi th th e high lifl _ g efficiency _ makes a two-engine commerc ia l STOL ai r c r aft feasible.

Figure V-10 sho w s a conceptual installation of a tw o-spool, tw o- st ream e ng ine (i. e., th e core str e am and b yp ass st r eam) in a nacelle. A t ta k eoff _ d la nding , all the b yp aSs ai r pas s es u p thro u gh th e pylon in to th e w in g duct a an d out _he duc t nozzl e .

I _ braking o n t he g _oun d i s requir ed , the forward valve mig h t be opened. , eliminating the augme n te r lift and pro vi ding a r e tarding force. During e lmi _ej t h e valves to th e ! '_ wing ducts are c lo sed, the flap s are re t ract ed t o a s treamlin _ po s itio n, arid th e valve i_i ., at th e rear of th e pylon is op e ned to augme nt the fo rw ard cruise t hrus t provided by 1 , i the core Jet . I ! • The Canadi an governm e nt , deHa vi l la nd of Ca n ada , and NASA ha ve b e en wo r kin g on th e aug m enter wing c oncept for several ye ar s. Fi gu re V-11 shows a larg e- s cale i t model m ounted in th e Ames Research C e nter 40- by 80-foot wind tunnel. T h e model I_ " sp an is approxifl _ ttn.ly 40 feet. The atl g m en to r inIe t is S h oW n a long th e tl _ til in g edge !_ of the wing. Ailerons we r e droop ed an d h _ td bolmdary-layer c o nt ro l applied at th e knee. Bo t lndary - layer con t rol was a lso appl ied a t th _ sl ot across th e top of the fuse- _i_i_ lage. A special pump for the aug m en te r ai r was m ounted in th _ fuselag e . The p u mp consisted of a J - 85 en g ine d _ v iz g two Vip e r co m pressor S ; henc e , the thre e inl et s on !| the fuse la ge visible in th e photograp h . An augrn e n to r win g ai rplane would not ha Ve i_ .. these fusela g e inlets.

Data from t hi s model are s h own i t fi gu r e V-12. Lift coeffi c ient is plotted as a f unction of dra g coeffic i ent. This curve is for an aug m en te r and bo tln da ry- la y e r con- trol thrust to g ross weight ratio of 0.31 and a flap defl e ctio n of 7 0 °. Even for th is

i't

low thrust, the m axi m um lift coefficient i s a bove 7 , w hich s hoizld b e sufficient foi" a 1.

_ - o : , _ ._, ..C _? _'_ - _'_ :;i :' " "_: """_' - _ " ....... " ..... "_" -" "" l "_ ....................

1 5 00-foo t field l e ngth airplane, , .. , . Thes e dat a define an engine sp e cifically f o r an aug m entor wing ai rcraft, but ..... : first a la n di n g approach design po int must be s e lected. Figure V-13 is the same as the previous figure, but more information has been a dded. An a i rplane de s c en ds at a n angle d e ter m ined by th e ratio of d rag -to - lift. Lines o f c o nsta n t de s c en t an g les of 3°, 6 ° , and 9° are s hown, and of c o urse th e vertical axis corr es po n ds to l evel fli g ht.

A de oign point at a 5° angle of attack an d 20 percent abov e s tall s pe e d wa s chosen.

This design point provid es a 150 angl e of atta c k and a 13- kn ot s tall m argin, w hi ch _ ', exceeds th e mar gi n r e co mm ended by so m e authoritie s, At the lif t coefficient of 5.1, an ai rplane with a wing lo adi ng of 7_ po _ tds pe r square foot could lan d in a 1500- f oot field. At this design point, the descent angle of 5° and th e 5° a n_ le of atta c k gives a " , horizontal fus ela g e alt i tude.

: One of th e require m ents o f commercial ai rcrafl i s that they be able to arr est th e :rate of landing descent with a minimu m loss of altitude after the pi lo t 's decision is made to go a round. This is done conventionally by ro ta ting the ai rplan e , in cr e asin g power , and in some eaSe S r.educ i ng f lap deflection. How e ver , as attempts a r e m ad e to lower w e ather minim ums, these techniques become less ac c eptable. With the aug- •_ mentor wing, th is characteristic can be improved by designing the engi n e S o t ha t th e core thr us t will be sufficient to arrest the rate of sink (corresponding to 0.09 thrust to gros s w e ight ra tio) an d make the thrust i mm e di ately available for e m ergencieS.

To provide this ca p ability, a f as t -ac tin g thrust m odula to r for the core is nece s sary.

By including a thruSt reverser in th e m odu lato r, the ai rplane can have a descent • _ angle between 0° a n d 8° at fixed po w er. With the p ro w _ ion of the thr us t m odulator, : handlin _ qualiti e s ha ve probably been improved beca u se the modu la tor provid e s a descen t co n trol in dependent of lift or a irp lane ai rspeed.

•. _ Th e figure contains enough informat i on to define important c ha racteri s tic s of the ...... ii aug me ntor wing engine. Wit h a duct loss of 20 percent , an a ug m entor augL,entation t •, of 30 p e rcent , an d a fli g ht sp e ed of 65 knots , th e re q uired in st aiied gross thrust to : _ii g ro ss weight ratio required is 0.3. Commercial transport regula ti ons will require this perform an ce with one engine failed, thus the ins ta lled thrust for a four-engi n e ai rpla n e i s O.4. The i nf or mati on also es ta bli s hes tlmt the ra ti o of the core thrust t o the to tal engine th ru st is 30 percent. The rem ai nin g 70 percent of the thru st is de - li vered by the fan to the aug m entor. O / course, other core th ru st r a tios can be ob- tained With different airp lan e variables or t hatchin g p hi losophies, but th e case de- rived here is represe n tative.

Thus far tWoquestions ha ve been ignored. One isthe limi tati ons on duct sizes i m posed by th e wing c ro ss-sec ti ona l a re a , a nd th e other is the practic a lity o f th e engine c y c l e and its a daptability to noise reduction technology. These que s tions will now be consi d er ed , d 1 4 1 !

i• Enough air can b e passed through th e win g by resortin g to high pressure ratios , i but pressur e must be minimized because of crui s e spe c ific fuel co n sumption and n o ise fro m the augmenter. The r e lative duct areas required t o p US the au g menter flow at Mach 0.3 in the most criti ca l wing s e ction i s shown in figure V-14 as a func- i • tion of f a n pressure ratio. A du ct loss of 20 per c e nt thrust iS incl u ded. The ca- pacity of the du cting system sh ow n in the sketch was esti m ated for a wing load ing of 7 7 p ou nds per square foot an d a 12 pe r cent th ick wing With Spars at 1 5 a n d 50 percent chord. The r e lativ e duct ar eas avai la bl e are s hown f or wi n g aspe ct ratios of 6 , 8 , an d 10. At an aspect ratio of 8 , a pressure ratio of 2.3 is r eq ui red . However , th e curve s indi ca te a wide r a nge of possible co m promises b e tween th e engine, aerody- na m ics, an d noise. For example , with a w ing as pe ct ratio 6 , th e data in di ca te that , a I. 6 pressure ratio fan m ay pump th e r eq uired amount of ai r. This, as a s ingle- stage f an , wi ll reduce com press o r an d aug m enter no iS e s ignificantly bu t would in- crease the r _ quir ed ai rp la ne th ru st to gross weight ra tio by about 10 pe rcent. S tu di es to eval uat e these compromises are requir ed .

As defined here, the engine s hould haVe 30 percent of the thrust fro m the cor e and 70 percent from the fan. Further m ore , the jet exhaust velocity should be low to i I minimize core jet noise. FigUre V-15 shows the core thrust fraction a S a function of i the core jet velocity for an 18 0 0° F turbine inlet t e mperature. Fan pre s sure r at io is 2.5. The core j et v e lo c i ty m ust be a rou _A 800 f ee t p e r s econd t o keep the co r e jet n ois e down. The desired core t hrust f ra cti on i s o btai n ed at th e m oderate tur b ine inlet te m perature of 1800° F. These figt w es represent one solution to t h e engine- ai rplane ma t chi n g problem. A S With th e blow n flap , conven ti onal engines could ha ve been used, b _ ttthe core je t noise is p _ ohib i ti v e; thus , a new e n gine iS r eq _ red for n oise r e duction. T h is engine will pr o bably be w ell with in the stat e of th e art; how - ever, no eftgines like this are being built toda y . A n ew or hig hly m odified e _ l n e i ' _ ' i _ will probably be developed Specifically f or the a ugmenter Wing airp lane. ' T hi s ha s be en a review of th e augmenter wing state-of - a _ t. The augm en te r w in g t is an efficient hi gh lift system. It is possible to get the require d duct sizes into th e ai rcraft wing. A two-spool en gi ne of special design can gen e rate t h e req ui red wing duct flows and pressures Wi th a low j e t core exhaust noise. Consider a ble work is s ti ll requir e d to opti mi ze th e m atching betwe e n th e engi n e , ai rframe , and noise.

, MULTIFAN AIRPLANE

The major factor s affecting S T OL airpla n e design are no is e, efigi n e-out _ _ a l id a la tera l spreading of the Jet sheet to ac hi eve hi g h lift coeffici e h ts v f h cieiltly. Given these f actor s , the us e O f m ultip ls_ low- p r es sure- _ at l o f ar _ s i s an approach to a STOL , ai r p la ne that may have s o m e advanta g e s . One p o s sib l e airplane layout i s s hown-ln • i _ figure V -10. On the ai r plane , 16 f an s are spread alo n g the wing traili ng edge.

These f ans have a pressure ratio of 1. 2 5 s o t ha t the propulsion can be quie t , and th e i same f ans pro V ide both t he takeo ff and the cruise t h ru s t. The lar ge nu m ber o f f ans , give s a good spanw is e ex t ent o f th9 Jet sheet. The fans are m ount e d on a hing e d f la p i to a c hi e ve a downward deflected jet f o r hig h lift. The fan inlets crea te a mas s ive S uctio n to maintai n flow at ta c hm ent on th e wing upper s urface. The effect of one , _ fan-out is s mall because the t h rust o f ea ch fan is only a s mall fractio n of t he to ta l thru st .

The fan s a r e driven by hlgh-pressure air from four ai r pu m ps t h at deliver ai r • / i ' at a pr e ssure ratio of about 7. T hi s pressure ra tio allow s the duct sizes in the w ing to be s ma ll compared with the wi ng croSs sectiona l area.

To help determine ff this arran g e men t has m erit, the mode l shown in f i g u re . : V- 1 7 was built. I t is a se m ispan m o del, and , hence, ha s eight fans. They are _ d ri ven by hi gh pressure air , a n d the w ho le s y st e m i s mo u nt e d on a f o rce balanc e . _ I_ • The t es ts were run in t he Lewis 9- by 15 - f eot V / STOL test sectio n . A typical s et of results is s t _o wn in f igure V-18.

Lift co e fficient i s plotted as a function an gle o f a tta c k f or two conditions: a 30 ° fla p deflection corre s pondin g to takeoff and a 60 ° f la p deflection correspo n dL _g to landln g . The data extend t o a 30 ° an g le of attack , and th e wln g ha s no t stalled f or either flap def l e ction. These data ha ve been int erpr e ted in terms of a STOL ai rp lan e with a 100-poufld- p e r -sclua r e-foot wi ng l o ading, a 1 5 00- f oot fi e l d l e ngth , a n d an in- sta lled thru st -t o -weight ra ti o o f 0.5V. Take o ff w o uld occur at a lift coefficient of i a bo ut 4.0 at a bo ut a 10° _ m gl e of attack. At this flight con di tion, the wing can easily _ :,: tol era te a 30-f o ot-per-second ver ti cal g u st .

. , : • The landing approac h would occur at a lift c o efficient of 7.0 w ith the engin es t partly throttled and a thru st to g r oss weight ratio of 0.44 (the install e d vahte i s :, 0.97). Agai n , at t hi s flight c o n di tion , the w ing .has a 30'foot-per-Second vertic a l gust tol e rance. T his per f orman c e tog et her with th e fact th a t it i s achie v ed w ith p d - te ut lally lo w-n ol s e fans nmke_ t hi s a c o nf i guration of in t e r est.

The propulsio n s ystem is desc r ib ed in m o re detail n e xt. A d raw in g of one c f th e fa ns i s shown hi figure V-19. It ha s a s in gle-stage rotor with a pr e ssu r e r atio of 1.25. Turbine drive air enters an annu la r scro ll , pas _ es tlttou g h the tu t bin _ noz z le s and _ rb in e b lades m odri t ed on t h e tip of the f an to d ri ve the fan , a h d t h ed exits to _ rovide p a rt o f the thrust. The drive ai r t e m peratures s t r e lov _ enou gh s0 the ducflng, scroll, an d tt w bine ma ter i als c a n be tita ai u h i.

High pressure drive air is provid ed by t h e air pump sliow n in f ig u re V° 2 0. It id a two-spool e ngi n e . The f ir st spool g ene ra tes a pressure t at to of abo u t 7 .0. Half z4S of thi s air i s duct e d off to drive the fans; the remainder is routed throug h the second spool , is bur n ed, and i s us e d to drive th e turbi n e s for both s pools.

The con s equence of an airpu mp failure depends on the ducting system. It Is propos e d that the du c tin g be arran ge d so that each fan receives o ne =quarter of it s drive air f r o m each airpu m p. The lo s s of on e ai rpump caus e s a one=quart e r loss of pow er , but all the f ans continue to run u n ifo rm ly so no roll or yawin g m o m e n ts are d e v e lop e d. Al s o, b e cause thi s is a pow e r in t erconnect e d syst e m, th e o ne -qua rte r : power loSS ideally causes only a one-eighth los s in thru s t. Th us , this concept i s tolerant of ei th er a fan or ai rpu m p loss.

Studies and tests to date indic a te that a multiple low-pressure=ratio fan conftgu- .

.... ration has the potential for low noise, good e ngi n e f a ilur e tolerance, and has very i_ I hi g h usable lift C o efficients. As a Con Seq uence of th e hi g h uSable lift c o e ffic i ents, the multif an concept c an yield ai rplanes with a wing loading of about 100 pou n ds per ! ° .- s quare foot for a 1500-foot field. W ha t i S _ worth to incr eas e the Wing lo ading f _ o m i 70 to 100 ? Th e decreas e i n w ing weight c o rresp o nds t o an in crease o f about 20 per- i cent in payload weight. The crui s e lift-drag ratio at MaCh 0.75 and at an alti tu d e of i _ 25 000 feet is improved about 10 pe r cent , and there i s q ua lita ti ve i m provem e n t i n !

ride quality, i In b ri ef, the multiple f an ai rp lane appear s to ha ve a nu m b e r of 8t g niftcan t advan- i _ tage _, but th ere are also problems thitt require further s tudy such as cruise drag, i, : a _ ro o la s tic stability, an d s yste m Weight. i_ J 4 _

VTOL AIRPLANE PRINCIPLES

, . .

VTOL ai rplanes emp l oy e ngin e s fo r lift as show n in fi gul.e V-21. T h i s draw ing i_ is purely sch e matic an d is not proposed as a favored configura ti on. A multipli city of engin es a r e required to provide the e n or m ous lift and to pro vi de safe ty in case of engine failure. The e ngines also m u st perf or m new control functi onS . At zero llft- t off speed, th ere a re no ae r odynamic forc es acting on the control surfaces. Th e powe r s yst em m ust pro vi de the cont ro l. The lift engines shown in th e wing tips are fo r roll co n trol , an d fore and aft engines shown ar e f or pitch co n t rol . In additio n , en gin e s must pl _ o vi de c ru is e thru st for forward fli gh t. Se p arate en g ine s are sho w n fo r _ tch of these func ti ons; it is pro b abl e tha _ some o f t hesefunc t io ns carl be co m - bined into few e r en gi ties.

The remainder of this secti on i s limited to a di sctmsion of lift fans. Although t hel lf t _ l gi flet _ a r e requil'ed only at takeoff arid landfltg, we must drag them a lo n g for the en tire f ligh t. TO avoid t h e d ra g pefl ai ty of these inacti v e ehgines as much as pos- s ib l e, the air p l ane desig ne r tries to bui'y th e en gi n e s in the W ing, in s m all pods , in extension s o f the f _sd age _ o r in the fuselage itself. Consequently , the e n gi n e designer i44 t \ i s r e qui r ed to m ak e th e en g in e a s s hort and a _ sm all a s p ossi ble, Another pro bl em co n cerns a n u nu su al fl ow si tuat io n ar ou t i d t he e ngin es . A f t e r-- .

ltft _ ff, the engi n e s m u s t operate whi le th e airplan e is i n f orw ar d flight u n t i l c o m pl e t e wing lift i s e s tabli s hed. Th e flow of ext er nal air a c ro ss the in let c au ses severe di s - t o rti o n s o f fl o w a t the engine face. Th e e n _ ln_ and e n _ n e in s tal la tio n mu s t be d e_ signed to w i t _ tand the _ e fl o w di s to r ti ons .

The mini m u m engin e thru s t, or lift, m u s t be a t lea _ t as great a s t h e gross w e ight o f t he airplane, Ad d iti ona l lift is req u i re d for a c c e le ra tio n and ma n euver ing .

To allow for en gi ne fai l u re , th e installed t h rust m ust be g re a t enough t o p rov i d e lift a n d maneuverin g accele ra tion with an engine out. The failure of an engine causes i upsetting mom e nts; correction of these m oments requires still nlor e i n stalled lift.

When a l l of tl _ e se f a c t ors are co n s i d e r e d , th e i nstalled t h rust to gross weight ratio wil l range between 1.2 and 1. (J.

The i mportanc e of engine thrust-to-we i g h t r stio is Shown by the result of a _ , - parametric study of VTOL airplanes ( f i g . V-22). The pla n e in t hi s figu r e can c ar r y 100 passen ge rs 5 00 miles, and i t s engine thrust to weight r atio is 1.3. Ai rc raft gross weight i s plott e d as a function Ofthe rati o of englne thrust-to-engi n e weight as installed in the ai rp la ne. Curves e o _ 'r e spondin g to lift ing times of 5 a n d .10 m inutes are Shown. YOumay recall that the en gi n e thrust to weight ratios o f to day 'S airp _ e e n gines a r e n ear 5 ; such e n gi n es are t otal l y unsuita b le of th i s service. A thrust t o wei g h t rati o o f 1 0 gives a n airp lane g r o s s wei g ht cl o se to 100 00 0 pound s f or a lifting _ ti m e o f 10 mi n utes. DCg's and _ 37' s ar e a pproxi m at e ly t hi s S ize and c a rry a pproxi- t ma t ely t h e sa me n um b er o f passen g erS. W e would like, th ere / ore, to have engin es

i

with thrust t o w ei g ht ra tios of 1 0 o F bette r . New en gi n e desig n s and n ew t _ chnol _ I will be required to produce thes e eng ines . ! • The figure also shows t he e ffect o f var _ fl n g specific fuel consumption (SFC) on il airp la ne gross weight. I n creasing the S pecific fu _ l cortSu m ptio n f r o m 0.4 to 0.5 has iI j , minor effect of ai rplane gross W e ig ht b e caus e t llsse f an engines ar e alread y very el- • '_ • ficie n t an d the fuel co ns u me d in the 5 or 10 mifl u_ s o f op e rati o n is n ot gr ea t, ti The noise li m itation drives us to very _g h-byp a ss- ra tio, l ow -pre ss u re -ratio, !

arid low-tip-speed fans. !_ In summary , t h e i n s ta lled lift m ust be 1. 2 to I. 6 times th e g ros s weight of the I ai rp lane. The e ngine th ru st to weight ratio m u st b _ hi gh , p r obably abov e 10; th e engine must be compact, have low noise , an d be tolerant to the e x ternal flow a cross the inlet.

!

i4 5

EFFECT OF EX T ER N AL FL OWO N V T O L F AN PER F OR M A N C E

A pic tu re of a fan in s t a ll e d i n t he 9 _ by 1 B - f o ot V / B T O L wind - tun n el t e_ t s e ct io n is sho wn i n fig ur e V - 93 , The mo d a l lift fa n has a d iam e t e r o f 1 5 i nc he s a nd i s mo u n t e d in a w ing w hi c h spans t he t un n e l te s t s e ct ion, The w in g w as mo unt e d ve r t i - c a lly t o f a c i l i t a t e t h e m e as ur em e n t o f fo rce s. Th e fan is dr i v en b y a co mpac t , high- pr es s ure air tur b ine whic h i s mou nt e d i n t h e hub of the fan . Th_ fa n ham a de sign pr ess ure ratio of about 1 , 3 a t a t ip sp e _ d o f a ppr ox i ma tel y 1000 feet per s e cond .

T he a xi s o f th e f a n i s p er pe n d i cul ar t o t he di re cti o n o f th(_ t unn e l a i rf l o w. Con _ s 0quen t ly , t h e ai r i s f orced t o turn 90 ° t o en t er the fa n . T h is is terme d t h e c ro s s - ( tow env i ron me nt.

The on c om ing ai r s tr eam i s forc e d t o turn s harp l y to c a ter t h 0 f an on t h 0 up - stream edge of the f a n i n l et. A t h ig h crossflow ve l ocities , t h e f l o w s e par a ted a nd blocked a portion of t h e inl e t are a. T hi s is po t ential l y a s e vere probl e m; but in t hi s part i cu l ar case, th e m ax im u m i nlet area blocka g e a mo u nte d o _l y to about 6 perc e nt.

B l ockag e s li ke t h ese can be tolerated by the li ft f an a n d a i rcraft as w i l l b e shown l a ter.

!

Anot h er problem caused by crossflow i s more ser i ou s : t he c ha n g e of rotor i n - c i den C e a ng l e ( see f ig . V-24). The inset is a pl an for m of the fan in cro s sf l ow. The rotation of the fan is indic a ted by the l ong arrow. Each fan b la de advanc es a n d r e - ] !

treat s r ela tive to th e eros s f l ow as the fan rotates. This ca u ses po s it i ve and n e gative ' rela ti ve i n cide n ce angl es o f t h e flo w , a s s ho w n i n the in set. I n th e f i gure th e th e o- ret i cal change in in c idence an g l e is plott ed _ ts a function of the ctrcun _ er e ntial posi - t i on. T h e re is a cycl i c p a tter n w h ich is agg rav a t e d as th e e l 'os S f i ow V e l uctty in- creases fro m 0 to 1( _ 0 m ph. When rotor i n ci denc e an gl e ch an ges, rotor pre A sur e .. ratio chang e s. Conf i r m ing exp e ri m ental dat a are shown in fi g ure V-25.

The rotor total pressure ratio is shown as a function of circu _ lfe _ ent i al po s it i on.

t Static c o n ditions result ed in a unifor m pre s sure ratio. As e ro e s f low i n c r e a sed to 100 an d 160 m p h , the predicted cyclic beh a vior w a s obser v ed. Howeve r , t he cur V e $ were not s y mme tric a l in the advancing port i o n of the cycle. T h e da ta re a ch e d a li m iting value wh i ch w a s ind e p en dent of th e crossflow velocity. T his ef f e ¢ t p r_ u ee_ a n e t re duction of the m ean rotor to ta l pre s su r e ratio as the croesflo w velocity in- crease s . Co l mequ e ntly, we would expect t he fan thru s t to i a ll o f f Witl i i nc reas e s i n cros sf lo w velocity.

Fi gtl re V - 26 shoWs m odel l i ft fan total th ru st p e r unit fro n tal area as a function o f the tmmel air velocity. Da ta are shown f or a tip speed of 980 f ii et p e r se v ond (or desi g n conditions)• ' l _ vo add i tional curves are superi m posed for co m parison l _ t i_ - poses. The top curve i s a n ide a l fan performance curve, wh i ch a s su m es r _ cov e ry I, i: of all t h e momentum of the tunn e l a i rstrea m . The l ower curve i s the l i ft fan t hl ' u s t r e q w tr e men t of a VT O L a i rcraft conv e rti n g t o w in_ - _ upp o rt ed fl _g ht a t 1 4 0 m p h. W e have ref e re n ce d th e two curv ea t o t he B t a t i c thru s t o f , _h e model lift f a n a t d es ign speed.

F o r the- de sig n condition, th e experimental f _ tni _ ru s t f a ll _ o f f with ir, c re _s l n g er os_fl ow v e l o_ I t l e_ . Th e p e r forma nc e o F tld _ fan l a t a r f r om ide a l, T he e ro _ s t l o w onv l _ ' on me at d _g r ad c _ th e f an p e r f orm _ nc,. _ by a c o n s i d e r a ble amount , but there l_ _ B uff ici ent marg i n to prov i de t he requ ir ed t hr u s t f or t h e example sh ow n .

This l_ J ust on e caB e , a n d further t es ts a re bei n g pl a nne d With low -p r e a _ u r o ratio, l ow-tip , p eed, loW-noi s e fa n a to see ff the re tre n d s a re ma int a i ne d.

VTO L FAN PROP ULS ION SYS TE MS

As has been pointed out before, noise c onstrai n ts lt _ v e resulte d i n the r _ lui re - m e a t o f rat he r l ow pr e ssu r e ratios, 1. _ 5 o r less, for lift fa ns. Th e se faris i n them- selveS , t he n become rather we ll -d e fi ne d a n d ar e f lat and li g ht w eigh t. T he r e asons w e get into various fan syste ms lies i n the question o f how the fans a re driven.

In ge neral, the drive systems c a n be put in two ca teg or i es, a s ShoWn tn f i g - ure V-2 7 , The inte gral d rive t an e n gi n e is self - contai ne d with a hub drive turbi z te and i ts o _ n i nner spool. I t i s, in principle, a eo n ve r ltlonal tw0 - sponl , I nC h by pa ss ratio tu r bofan e ngi n e. T he lo w er part o f the figure s hows exa m pl s s of t he remote d r ie s systems. In such Systems the po w er source i s in a loeati bn remote f _m t he f an it s e lf , _ ith pneumatic e n ergy t ra nsmitte _ l to the power tu _ ' bhte thr ou gh duc ts o r pipes. These ducts can be co _ ig U re d in such a w a y th a t f _ J u _ a x id powe r S our C e d can b e i nt e rcon ne cted to reduce the e ngine failure and d o a tro i probl ems . The f ar[ ro od - i u le s are rat he r short in d e pt h . This, and the ab i lity to orient the power s ou rc e i al bn g a differ e nt _ xis results in a potent ia l r eduction i n installation proble ms as co m pared to th e integral driv e sys t em. :_ Two remote systems _ re sh own . One is a ga _ g e n _ rator syste m , ta which a conv e ntional turbo j et en g i ne is en i ploy e d. The hot pr e ssur t_ .ed g _ t s , rather than bei n g us e t _ f or t h rust, i s div erted through appropriate d uc tint _ to the f an - driv e tti r .

b in e inl e t. This fan system has been u s ed in so me of the di r e c t lift reseai _ ch ai r - _ craft that ha v e been exa m ined i n r ecent years , notably th e XV _ . The other S y S t em _ uses a n airpu m p. This ai rpump w ould be a very lo w byp[ts s ratio, hi g h- _ i ' e s sure ratio t u rbol ! a _ engine. T h e pressurized ai r , w i n c h is con s ider a b l y cooler the f t the gas g _ fl e rator exhaust, i s r outed th r ou g h dt _ cts to th _ lift t an. An a _ d l ta ry co _ l _ us - for heats the air Just befo r e it goe s i n to t h e t a _ ttl _ b i ne d r ive.

The u lUm ate choice of the f a n dri v e system t s a function of m any factors , s o m e rel a ted to the engi n e itsel f , such a s wei gh t, sy stem _e l e ohstimptiofl , v olu me, a n d 14 7 / d e pt h an d o t hers r e lat e d t o a i r c r aft op e rati o n , s u c h as e n nt r o l a nd e ngin e- o u t . Th e d ls c u s_ I o n h ere wi ll be c o nf i n e d to t h e e ngines t h e ms e lv e s a nd t h e ir f ea tur es ,

INTEGRAL DRIVE FAN

_ o m e of the fe atur e s of the i n t e gral drive fan are sh own i n figur e V - 2 8 , w hic h pre se nt s a cro ss - s ec t ion a l view of one f a n curre n t l y being co n s id e red. T h is t_ a s i ng le- s tage , relat i vel y low-spe e d fa n (du e to t h e required low pr ess u r e r a t / o). I t h a s t h e I Jo tentla l o f being very light weig h t bec a u s e of t h e pos si ble u se of adva nc e d co m pos i te s truct u re s . T h e fa n i s sl l ow n Wttl _ t he flow to t h e co m pre ss or s plit off i from that of t h e b y pa sse d floW. Such a fan geo me try i s attractive because it permits the pressure rise at t h e h ub section to b e low e r than t h at o f t he main fa n , t h ereby _ permitting a lower fan inner radiu s and , , t h er e fore , a smaller diameter en g ine. T h e inner spool Includes a m ult i stage compressor driven by a single-stage turbine. The i co m bu s tor sl _ w n iS a reverse-flow type. Such a combustor geometry i s considered to be attr a cti v e because i t s h ortens t h e l eng t h o f t h e engine , thereby mir A m izin g its _ !

dept h and volume, t The inner spool must be rather _ impl c a nd co m pa c t for this applicatio n . Fo r - i _ - tu n ately, the engin e cycle parameters, namely, pressure ra ti o a ad turbine inlet te rn - i._: perature , a re such that this i S possible. The results of en g i n_ Cycle calculatio ns "_i made to find the opti m u m cycle pre s sure ratio and turbine inlet temperature ar e shown in figur e V-29. Th _ engine W0i gh t plus 5 minutes fuel is S how n on a relative !

scale as a function of th e turbine inlet te m pera tu re at t hre _ co m p r e ss ion ra tio s.

: The optimum te m per atu re is 18000 F, w h ich is surpl'i s ingly low and a conservative value by today' s st an dards. The optimum pres s ure ratio is also surpris i n g ly low , being about 8. Low opti m um te m perature and pres s ur e ratio are obtained because ?

• the sh o rt o perating time of the e n gi n e m inimizes t he I m portance of sp e cif i c fu el co z i- su m p ti o n . These m_x i e st cycle te m p 6 rature s a nd pre s sures gr_ t tl y si m plify th e g _ s generator.

From f i gure V-28, it c an b e seen that th e fa n is dl'i V en by a m tll t ist Ag e turb i ne l B ecause o f speed re s tr i ctions o n the fan a rid because the tu rbine d i ame t e r is sub - stantia U y l ess th an that bf the fa n, the turb i ne blade sp e edJs _ ' tnd associated s tre ss es of this tur b ine a re qu i te low. Therefore, t he tu r bi ne I s pot ent ia ll y lig h t in w ei g ht thro u gh the g ene r ous u se o _ th in sheet metal in its _ orlstruct / Oa.

From a r t aerodynamic standpoi n t, this t ufl J in e ha s requirements that ar _ co , i- sidered to be beyond that of c on v en tioitai practice. The re _ s nn for this is indicated in figure V-30, where stage efficiency is plotted as a f u nc t ioh of turb ine s ta_ e Work ,' factor. (Stage work factor re lates the work being exti ' a cted pet _ potmd of air p er s ta g e t, th e ki ne tic ene rgy o f th e rotating b la d e_ . ) I n c rea sing t he w o r k fac t o r r e_ s ult _ i n d e cr ea s i ng th e turb i n e effi c i e n c y. Current cr _ji_ o e t lgl ne s u s u a l ly op e rate a t wor k f act o r s o f I to 2 b e c ause of t he des i r e to h av e : _g h turb i n e e ff i cie n cy and , t h u _ , low s p ec i f i c fuel c on s u mp t i on. S tu _es of t h e i nte g r a l e ng ine ha v e s h o w n t h at , w he n co n s i de r ing b o th turbin e weight an d ef fi c i e nc y, work fact o r s in t he ra _ e f ro m 4 to 5 b e c ome n ecess a r y. S uch high work factor s ar e _ ncountered b e cauge the b |a de _ pO e d is l o W , as p re v io u s ly me nt i o ned , an d t h e work p er p_ und is hig h because o ! the _ : _ f l o w t hr ou g h the g a s ge n era t o r. S how n is a br o ad ban d of unc e r t a ln ty in the e xo p e et e d e f fi c i e n cy at th e s e h ig h work f acto r s , thu s t ndtcat l ng that there ar e major que s t l on _ regardin g t h e ulti m at e d fi c ie ncy i n thi s r a n ge. Pro g ra ms are currently und e r wa y to attempt to a n swer these qu e stio n s.

?

T I PTURB I NE FAN SYSTEM il

The low blade speed encountered in the turb i ne tor the inte gr al engin e i s one of 1 _ i , , th e r e asons why other approaches p _ rmittin g increased blade speedS ar e desir a b le. 1 Introduci ng a gear system would be on e way of accomplishin g this. To date , how- ever, the adde _ l complexity and weight has not r esulted in this approach bei ng Sltown to ha ve a net adva n ta g e.

Another idea to _n c r ease the turbin e blade speed - also considered to be un- practical for the integral engin e - is the use of a turbine located a t the tip of the fan.

Th e se tip tu rbi n es ar e located where the b la de speed i s greatest, thereby res u lting i n more modest values of stag e work factor , ab o ut 2 . Tip turbines a r e tise d Jr, t he remote drive syste m s desc ri bed pl ' e Viously. Beca U se o _ the location of the turbine, rather large scrol l s are requi r ed to direct the high p z ' e sgure ga _ to t he tur b in e en , trance. Also , tu rbine inlet te m peratures on the order of 1400° F a l ' e b _in g cong t d- t ered fo r the drive systems, resulting in th e whole out e r Section b e ing at _ lli g h tem- peratures. These outer ele men ts, of cours e , requi re insulation and major a t tent i on _ : ! to maintai n the structural integrity _ nd li _ e of th _ h Q t part s , A cross se c tional view of o n e o f theS _ f _ ms is sho w n in figure V,31 for the gas ge r lerator system. The fan rot o r, its staler, the s c r oll , a nd t he S in g le sta g e tip t ur bine are shown. Although th e ga s gen e ra tor arid air pu m p d _ ive systems appear to be similar (fig. V-27) , there are substantial di fferences. These diff e renc e g oc- cur princ i pally because o J _ th e di ff e rence in th e avai lab le pr e ssu re ratio across the _ " _ " tttrbine. Since the g_s gene _ tor i s ba s icall y a _rho Jet e ngin e , th e pressure avail- able to th e tip turbine is mod est - perhap s off the ord e r o _ _ at m osphere s . Th e ai r pump system , on the other hand , can deliv er ' pre ssUre S c o ns id e rably gr ea te r tharl that. We hav e been exa m ining delive ry pressid'e s a s high a s 8 a t ntosph e re s ih 0i ' de r to allow th e pump to op er ate wi t h a s ingl e spool and to m inimize the size an d w e ight o f the as s ociat e d duets. These ductS could be a t a s uffic i ently low temperatu r e to ' per mi t titanimn to b e used.

A substantial i m pact o f these di f ference s in pressure raUo is felt Ont he s crOll an d tu rbine. Figure V-$: l illustrates the s e dif f erenc e s. Shown is th e Volume f low to the hot sc rO ll a s a function of turbine pre ss ure ratio. The flow for the gas generator system i s seen to be over fo u r-ti m es that f or the ai r pump system , and thi s results in a larger_and heavie r Scroll for the gas gene rato r system. On th e other hand , the low pressure ratio acro ss th e Up turbine for the gas generator syst em p e rmits a single-stage tu rbine to be USed. The high pressure ratio ac ro s s the tip turb in e for the air pump e ase has forc ed us to consider a two-stage tip tur bin e for this sy st e m .

• _ C ro s s sectional view s of th e tip regionS for these two case s are s hown in fig- ure V- $ 3. T he difference in the sc ro ll size and th e number a t tu rbine stage s can be s een. So m e o th er flsct 0 r s that affect the turbine an d fan performance can also be ' seen. First , generous axial clearanc es must be p rovi ded to accommoda te large m o- tlons of the fan tip. Second , the very short b la ding , particularly for the a / rpump tu rbine, has an adverse eff ec t On th e effic i ency. Third, leakage at the front face of the fan _ which oc curs due to th e p r eSSure differential betwe en the tu rbine stator out- let an d fan ti p inlet , reduce s effi c i e ncy. Tiff s leakage l)enalizes th e tur b in _ by di- verting the flow and penalizes th e fan by th e introduction of hot gas _ the f t tn at th e tip. Thi s i s a rather s erio us p ro blem since th e s e sea ls are in a region tha t i s § fe et • " _ in dia me ter, a , _ re investigat i on is n eeded to d e termine the best seal geom et ry for thi s application.

This diS c _si on has pointed up some of the features and pro b le m areas internal to th e lift fa _ _ themselves. Little was discussed about weight e o m pftrl S ons s inc e • "/ conside r abl e desi gu i m p ro ve m ()nts must be made before prac ti cal value s can be o h - .. tained. Th e di s cu ss ion Ires s hown , h ow ev e r, that the effl ei en e y o f high work factor turb in es m ust be determ in ed for th e i n tegral engin e . On th e o th er ltamt for the r e-.

m ot e drive fanS, m ajor questio nS ari s e regarding the tip tu rbine perfo r_ ee, par- ticularly as related to the clearances and i_ utkage _ . There i s also th e qUestion of aeeep u tbility of 2 1 th e hot com po nents that are locat ed outside the fail. T heS e and other questions m ust b e an Sw e red to d et er _ l in e which configu ra tion i s be st for the VTOL tra ns port s .

u

CONCLUDING REMARKS

For the blown-fl ap _ i*pl _ me, th e eng in e thru st tatl _ t be close t o 0.6 of Ute grosb we i g ht . The problem of e n gi ne-out r oll in g m o m en ts It _ beeitinve s t i ip #_ d wi th . ' E • # / models in a wi n d-tu nn el, a n d the results e n cour a ge _s to b e l ie ve that thes e m o m e n ts .... . can b e ma n aged by d i fferent i al flap s or boundary-laye r control a nd prop e r airplane - pro p orti o ning . En gine ss i mi l ar t o_xl s t i ng CTOL e ngin es ca n meet al l pr opul s ion r e quirements e x cep t lo w noi se. A s p e el al en g i n eha ving a h ig h byp a s s r a t i o , l o wf an , p ress u re rati o , a n d s oun d treatment will be req u ired to stay w Rhin t h e n oise g o ats.

The exa m ple aug m e n te r w ing airp lan e require d an i n stal l ed thrus t - too wei g h t ratio ne ar 0 .4. The ai rpla n e has good en g ine - ou t c ha racter i stics. HOWever, a spe- c ial e ngin e desig n is req u i r e d t o gi ve a t h rust sp li t of approximately 7 0 percent to the au gm enter a nd 30 p e rce n t to the core J et . In addit io n to achiev i ng the noi se g oal s , the core veloci t y must b e held to a low value ne ar 800 feet per s econd. A maj o r pr o blem wi th this airp lan e is the control o f the n oise fro m the aug me nter.

_! iii'I-'_ Th e no ve l mu Ri fan sy s tem i s p r om i si ng, b ut f urth e r re s earch is re qu ired fo r • ' _. e va lua tion of i ts p o te n tial, i ...... F o r the VTOL airp lan e , w e hav e show n som e o f t he features a n d pr o blems of i '.... t h e in te gr al d ri v e fa n ,t h et i p - t urb i n e fa n d ri ve n by a g a _g e n e r at o r, an d a ti p - 7, tu r bine fa ndriv e n by a c ompre S S e d a irg e n era tor.In al l cases, ad va n ce d tech no lo g y .... w i llbe r equir ed t o pr o d u ceu s ef u l e ngin e s , i i 3_ • , . _ i _ • j , " I f J_ ' i I !.

i s i

!

.r Q TYP I CA L TA K E O FF AND LANDING TRA J ECTORIES F OR CTOL, STOL AND VTOL AIRP L AN E S CTOL STOL VTO L # I / fff 1 11HH1ff_rr'r_ ' IffffI H _Iff111f1 " 1 "' J1fffllff_ ' Jr_ ',' I p' ff ,'. [ . rr1111f_ .

4 0 ,0 00 5,0 0 0 2 , 000 0 ' :' RU N WAY L ENGTH, FT , , C S - 38 6 7 _

.. Ji

STOL AIRPLANE TAKEOFF THRUST RE Q UIREMENT S i - •6 - ,' , .... W / S , _ LB / F r 2 7 ///7 K/////( _ , t S. LS .

A C -7" C ' _ . 4 THR U ST - " / % , 3 WEIGHT = S T O L---- - --- _ .2 1 I I I I m J 1000 2OO0 3OO 0 4OOO FAAFIELD LENGTH , FT c s - s6 s l z F igureV-2 / ' STOL AIRPLANE LANDING LIFT REQUIRE; 4 1ENTS 1 0 LIFT COEFF W / S , L B / F _ 2 i 4 I00 ,, STOL .... ' -- I

o I I

, 1000 2 OO O 3OOO 4 OOO FAAFIE L D L E N GTH, FT c s . 5_ s l_ Figure V - 3 HIGHLIFT COEFFICIENTS A N D THEIRPROBLEHS t Figure V-4 C5 - 56814 \ EXTE R NALL Y BL OWN - FLAP STOL A IRP L A N E /// _____I_RON Flg u r e V - 5 c s . s 6s oo i . _ , i ' BLO WN- FL A PMODELI N WI N D TU NN EL i ! , i J_ ¢ BLOW N FLAP E N GINE - OUT CONTROL T I W - O . 6 .2-- !

, , .,.- - - -'- -- " ,DI F FERENTIA L F L A P + S P OILER ........ ..I,,... . .. _ F.FER ENT IA L FLAP,,,= 0 4- - , _ , , ROLL I ' _ MOM _ / 4 ENGINES COEFF -.2- i _ ENGI N EOUT I I USEABL E i _

: - . 4 I I I I I I _ ; _

0 2 4 6 8 10 i,_ i, L IFT COEFF cs .s_79 z F i g ur e V- 7 , I: i ' ENGINE F OR BLOWN-- F LAP AIRPLANE

• l

t

I

]' i' F i gure V-8 c s.568o z 1 55 AU G HENTOR WING STOL AIRPLANE DUCTIN G --., // _ _ - _._ ' _ . '.' _ -- _ - t " "_ ,, \ AUGM _ NTOR'FLAP _ . _ BLO _ VN AILERON c s- s s 7 9s i2 I.

Figure V-9 i !'!

ENGINE FOR AUGMENTOR WING AIR P LANE AUGME NT OR ' _.'-"_ ,, .

;I C5 - 56801 Flgu reV-IO , - " '% ................. I l l! ii Ill " r _ i i r J i ....... i ..... , " " ----- AUGMENTOR WING AIRPLANE MODEL IN WI ND TU NN EL =. F ig ur e V- 11 EXPERI M E N TAL AU G MENTOR WING DATA 8 _ - I 6 OR T M - 0.21 LIFT COER: 5

3 I __l ]

O .4 .8 1 .2 DRAG C O EFF cs.s 6 ? _4 Flgure V-J2 o AU G MENTOR WIN G DESCEN T PERFORMANCE DESCENT ANGLE, 8 ,,-ARRESTED DESCENT DE G 7 R E

/ ' j; -

LIF1, 6L_ W' 0'09 / _ /// L AU G MENTORT M 'O ' 2!

C OEFF | // _" MA X / W I S- 77 LB / F'I 'z T / W• 0.4 ONEENGINE OUT i , 3 I ' 0 .4 .8 1.2 i '" DRAG COEFF cs.setg e !i ! , FigureV-B , , _ •

ii

i AUGMENT O R WING DUCT REQUIREMENT S i_ ' T / W-O.40, W I S-7 7 LB / FT 2 t / c-0.12

2 I

RELATIVE 1 - _ J RE _ UIRED "

o .... I I I

L5 2. 5 3. S FANPR E S_ ;URE RATIO c . _ - _ r,? g z FigUre V-14 g AU G MENTOR ENGINE THRUST DIVISION FANPRES SU RE R A TIO , 2 . 5 1' : 'I '- O _ T-A t -i ' H- R - US ' _ , 2 T U R BINE | INLET T_ P i_ . 1 - i.

o I I I I, 1 ,, cs.s 6,3: CORE JET VEOC _ FT I SE C FigureV - 1 5 , ! , MULTIFAN ON FL A P STO L A IRPL A NE ' ...: D U CT I " -AI ' FA S ONFLAP (16) t AIRPUMPS 141 t , % C S -5680 5 FigureV- 1 6 15o i MULTIFAN ON FLAP WIND TU N NEL MODEL ASPEC T RATIO, 5(MODEL + IMA G E) 300J E T ANGL E _: "I G _ - _6 / _ 0 7 F I gureV - 17 Jl t_ EXPERIMENTAL MULTIFAN ON FLAP DATA i_ IoW I S - 100LB / FT 2 FIELD LENGTH - 1.500 FT _ . _ ,

I"

8 30-FPS VERTICAL GUST t LIFT CO E FF 2T- ' _ 3 LANDING _ 0.44 I . L ' _ TAKEOFF30 . $ 7

"'_ i I I ' I _ I I I '

0 _ i 10 15 20 25 _ 0 ANdLE0[: ATTAC_ DEG cs -_ 6 _ o 6 Figure V-18 ,t TIP TURBINE FAN i I' "- -- --'- - -- CS- 56 8 0_ Figu reV - 1 9 _ " AIR PUHP / Figu reV - 2 0 C5- $ 6 8 04 , 161 i t \ 1 ' # VTOL FROPULSION FUNCTIONS CONTROL 7 __ _ _ CONTRO. , ,

/

uFT " _=_ \ ; F i gu r e V-21 CS - 56798 , LIFT FAN VTOJ _ TRA N SPORT GROSS WEIGHT 120x 1 03 L IF T FAN l_i,'_'\ .... 0 . 5 .._ : _ 110 \ __ , _ , _ -- . 4 , _ .

I I _ SYSTEM S. F . C. ..

• _ TOTALF AN WEIGHT , % _ % _ " -- . _ OPERATING TIME, Ir L B _. _... .

I N STALLED THRUST , , _ _"_L -. t _ ,_ 'A I C GRb SS WEIGHTJ" - " ' ' _ _"" ' _ " | ' 70 I I I I I I I 6 8 I0 12 14 16 18 cs.s6 79 7 I N STALLED L IFTFA N SYSTEM THRUST / WEIGHT Figu r e V-22

• !

c MODEL LIFT FAN - I N- WING '== m IJ i " C-69-353 1 !

_. Figu r e V-23 cs-s 6 78 9 i

i

CHANGE OF ROTOR INCIDE N CE A NGLE IN CROSSFLOW ADVANCING . V o _ (+& l ) 7 t . " MPH 2700 / 51 / _' _ , = , \N - A ,

-I o I I I J

0 90 180 270 3 60 CIRCUMFERENTIAL POSITI _ DEG CS- 5 68 11 Flgu reV - 24 ROTOR OUTLET TOTAL PRESSURE IN CROSS FLOW 15 - 1N. LIFT FAN - IN-WING X.4 TIPSPEED . 980 FT / SEC, MI D RADIUS POSITION

[- MPH _ _ _•

!.

1. 3_ ]" 1 t " _"_ ' _ - -" OU _ " / _1 80° i, 1.0 I I I I _ ' 0 90 180 270 360 _ , _ CIRCUMFEREN T IAL POSITION , DEG CS - 56810 k F l g ure V -2 5

,i

li: !: 15 INCH LiFT FAN THRUST IN CROSSFLOW ' P '"_"_m s =_3 - -_ . _ SP EE D . ! ' _ ".. _ FT / SEC I _ ' 1 00 - - _. _ 9 80 II THRUST % AREA 600- _ t !_ [ LB / FT 2 _ \ VO 5 0 0-R E QD FOR CONV ERSIO N % _-- _ !

V0 - 140MPH \ \ i" _

4OO

YJO +l I_ i Ix

.__ t i,, I I

0 25 50 75 lO0 125 1 5 0 175 cs. s t ,? 9 o T U NNEL AIRVELOCITY , Vo , M P H Flgure V - 26 _ i i p i.... i ..... ii i i i ii i i , ,, ; ,._ .

!

FAN DRIVE DESI G NS INTEGRAL DRIVE r AUXILIARY (__ ' COMBUSTOR ' , 7 ] ' " / :. _ . GASG E NERATOR AIILPUMP c s- s 6 s 77 REMOTE DRIVE Figur e V - 2 7 , : i / INTEGRAL DRIVE LIFT FAN L _: - f Figur eV-Z8 cs- _ 6sTs e INTEGRAL ENGINE CYCLE SELECTION FANPRESSURE RATIO = 1.2 WEIGHT INCLUDE S ENGINE PLUS FUEL CYCLE PRESSURE ,' 1.2 - RATIO OJ RELATIVE R A TIO1 . 1 - _ .... 8 . ; O F WEIGHT _ !_ , TOT.HRJJST1 . 0 ! _ 1_ 0 1 800 2000 2200 2400 2600 i c s- s 6 sso TURBINE INLET TEMP , OF Flgu reV - 29 _ ._

!

: , , EFFECT OF TURBINE STAGE WORK i_ : ON EFFICIENCY _' : i 9 5 I-- w f CURRENT CRUISE

STAGE EFF , _ E NGIN _ i

i' 7 5 I I I 1_ I 2 3 4 _ 6 STAGE WORK FACTOR. L _ H / UZ cs-s 6 s?S Flgu reV - 3 0 _r..._ -: ............... ,-; ................... _ ,-, , ,.p,, -,...,.,-,,,. • _. _ ........ !:'. .........

REMOTE DRIVE LIFT FAN Figure V-3 1 d CS - 5657 4 :' ?

_ . " FAN DRIVE SYSTEM VOLUME FLOW. ' .... . '_" " ' 6 - O NE' STAGE GA S " ; .!'.. . R EL A TIVE VO L UME FL OW 2 _ RATOR DRIVE . .

T W O-StAGe A IR i .. _ _ MP DRI VE " 1 3 5 7 9 , j c s- 565_9 TU R BINE PRESSURE RATIO " FigureV - 3 2 10 7 i' TIP TURBINE DRIVES i= , i_' I

i

t i} , ! ;E GASGE N ERATOR AIR P U MP !_! '_ , FigureV - 33 cs-s 6 sv 6 !i \ : ' .

N 7 1 -19 4 57

• V I. NOIS'EREDU C TION , Jam es J . Kra m er, Dav i d C h estnutt, Eugene A . Krej s a ,

James G . Lucas , andEdward J . Rice i . _

Increased use of j et transpo r t s alon g wflh a g rowin g concern f or t h e quality o f t he environ m ent has re s ul t ed in consid e rable e m phasi s o n nois e re duct i o n effo r ts in the past 5 yea r s. In order to discuss the noi se an aircra f t eng in e m akes, it is ' ' n eces s ary to define several items : wh e re the observe l` is with-respect to th e plane, i : t h e engine power settin g , and th e units o f noise m e asur e m e n t . N o ise energy levels , ar e measured on a logarith m ic scal e i n dec i bels. What we are really interes ted in, "" , _ , how e ver , is not how m uch e ner gy is in a ce x _ ain sound but how an g ry it m akes the • person w h o h ear s it. We wan t to meas ure the human response. A new unit, the pe rce i ved noise de c ibel, wa S developed t o reflect the varying hu m an response to t he I_ different kinds of sou n d , dist in gu is h ed by th eir frequ e ncy s pe ctra. So un ds which i! , have high ener gy content in the frequencies to w hi ch humans are most sens itive ate !_ weight ed so a _ to have hi gh PNdB o r pe r ce i ve d 'noise decibel v alu es. The s e fre- i:i qUencies of hi gh a n noyance ar e in the ran g e of 1000 to 5000 cyc l e s per s e cond or ii h ertz. Other fa ct ors ar e known to affe ot h um an r e s po ns e - m ost importantly the t , duration of expo s ur e and the presence o f di scr et e tones in the s pect rum . Co r rec° • tion fa ct or s have been d ev e loped and w h en th ese aJ _ e applied, til e r e s ultant levels i_ ., a x e in e ffective perceived noise deetbel_ or EI _ dB. i_ ' (: _ Both conventional tak e off and lan din g (CTOL) and vertical / short tak e o u t a n d _i_ ' . , lan din_ (V / STOL) aircraft noise i s dis cu ssed . T he sttbsonic ai rcraft fli gh t profile I i s consider ed fi rs t. Noise is a f actor for sttbsonic ai rcraft when they are landing and taki n g off. These ai rcraft land o n a 3° g lide slo pe , s o that their altitude atx _e th e co mm ttuity is f ai rly we l l define d as a function o f d ista n ce fr o m th uchdc N m. The engin e po wer s e t t ing v ar i es _s the p i l o t maneuver s d ow n th e g l ideslo pe ,bu t i _ en - erally the e ngine s are at a b out 2 5 pe rce n t of takeoff thrust. Durin g takeoff t he al - titude above th e co m munity d e pe n ds o n the o pe ra t i ng we igh t o t t he aircraft and the power settin g . It is po s sib l e fo l` th e pi l ot to l`ed uce pow e r ariel, he has ac hie v e d some sa f e altitude. This pow _ l, redu c tion res uJ ts in less noise .

_ NA S A-Langley Research Center. 4 \ CT OL AI R CR AF T Recently, the FAA issued an airc_ a R ce r t ifi cat i on s tandard for noise. The FAA- s elected measuring stations f or conventional takeo ff and l and i ng (CTOL) air - craft are below t h e a i rcraft 1 nautical mile from touchdown on approach and 3.5 n an - tical m ile s from bra k e release o n t ak eoff (fig. VI-1). Power cutback is per m itted d u r ing take o ff if o the r saf e ty req u i r ements a r e me t . F or app ro ach on a 3 ° g l id e slope, _ he a i rcraft Is 370 feet ab o ve the ground at 1 nautical mile from touchd ow n; i' an d at 3.5 na u tical miles from brake rele as e, the aircraft i s an ywbere from 600 to 1200 feet in altitude, or higher. The sideline station i s at 0.3 5 na u tical mile for four-engine ai rcraft an d 0.25 nauti ca l m ile for two- a r id three - e n gi n e aircraft. The : ' m axi m um per m itted noise levels at these stati o ns are sh - ,wn in fi gu re VI-2.

• ' Consider a 707 / DC- 8 -class lo n g - ran g e tr an sport. It produces about 118 PNdB :.

of noise a t 10 00 f e et alt it ude at full takeo f f power. If a replacement aircraft were ' _ to be buil t , i t c ou ld m ake no more t h an 103 E P NdE according to t hi s FAA _ tandard. i ', The actual permitted noise levels are a f unc t ion o f ai rcraft size, wi t h the variation I'_ • _ /r in effective perceived n o ise level (EPNL) a fun ct ion of ai rcraft gross weight as _, s how n in fi gur e VI-2. Get i er al ly , the levels req.uir od now are about 10 to 15 PNdB i_ lower th an previous levels, i ., _, A t Lewis we have been working f or several years on a low-noise engine s t titable ii for us e o n a CT O L ai rplane. The engine will in c o rporate all av ai lable s o u nd reduc- ! d , • t ion features i n an eff o rt to c on s o li d aie a n d a d va n ce t he t echn olo gy o f no i s e reduc- _,_ ti on . The e ngin e is c alle dthe Q uiet E ngine . This e ngin e in c on j un cti on wi th a n I_ ac o ustic al ly trea t ed na c e lleshouldper mi t t he a t ta inm e nt o f no i s e levelsa bo u t _ • I0 P N d B b el o w t he c u rre nt F AA ce rt ifica t i on l e ve l s. Cu r r e nt l yb_ing buIR un der co n- i_ tr actwl t h G e n eral Elect r icis a I tl g h - b yp a s s-r '_to e n g ine dev e l o p ing 22 0 00 po u nds o f ' _ .... i take o ff th rus t . A cr o ss-secti on al view o f t he e n g i ne is show n in figure VI o 3. I n t his _ ' progr am , several f ans a n d tu r bi n es _o r t he e ngine are bei n g b uil t , lin i ng s f or t he " _ c o ld an d h otdu cts a r e be ing in v e s t i g ated, an d s o me p o te nt ial no ise red u ctio n de vi ces a r e b e ing t e st e don a ha l f -s ca l e mod el f an . We exp ec tto b e t e st i ng th e engl he _t Gen eral Elec tri c in abo ut 6 m ont h s an d t o t ak e d elivery here at Lewl _ in late 1 9 7 2.

Th is engine serves a s th e b as i s for th e follow in g discussion of our noise res ear ch.

NoiseSources _ , .

Gas turbin e e ngin e no ise c an be divided i nto two g ehe r al cate go rie s : inter n ally il g e n e ra ted noise , u su ally ass o c i at e d with the rbt at t ng m achi n ery; a n d ext e rnally g en - !_ crated no ise, o r Je tno ise. Any no ise red uct i on pro gr a m m u s t co n side r alls ou rces _ , ( ' _ i " , . ; " and brin g them a ll down to an acceptable level.

T he cutaway view of the Q u iet Engi n e s how n i n f i gur e VI-4 reveals t he so u rces of internally generated n oise. These p rimary sources are t h e f an , th e com p ressor, and t h e tur b ine. In this e n gine, a s Lt m o s t h igh-by p a ss -rat io e n g i ne s , th e fan i s th e d omin ant s ou rce of t h e in ter n ally g e ne rate d noi se. T he fan no ise p ro p agates ou t both the e n gine inlet and th e fan discharge duct, th e co m pre ss or noise propa- gat es out the eng in e in l e t, and the turbine noise propagat es out th e j e t e xhaust n oz- zle.

There ar e two sources o f ex t e rnally gen e rated nois e in a tur b o f a n engine. I 1 t figu re VI-5 th e m ixing r egi ons WhiCh produce the Je t no is e ar e s h oW n . One i s l o - cated downs tr e am of t h e fan exhaust duct and th e oth e r at t h e nozzle exh _ St.

_ ' There is an i m portant di s tinction b e tWe e n th e Int e rnally g e nerated an d ext e r- na l l y generated n oi ses. Th e internal noise can be suppress e d in the engi n e p as - sages , whereas th e ext e r na l noise o bviously can not. The most ef f ective co n trol o t ; externally gener at ed noise is to keep the Jet velocities as low a s p o ssible. Fig- ure VI-6 shows how acoustic treatme n t m ight be applied to the en gi ne p a ssa g e s fo r th e reduction of in ternal noise. The fan nois e is reduced by in let and fan ex, _ m s t suppressors. The compressor noise is mostly absorbed in the core engine In le t , while the turb in e noise is suppressed i nside t h e core exh a u s t n o zzle a re a.

: The relative m agnitude o_ th e differe n t n o ise s o urces can vary fro m engin e t o engine, and even f o r a given engine they can va r y with o peratin g condition s . This is illus t r at ed in the figures VI-7 and VI-8. All noise sources hav e td be l _ ere d s in ce ov e rall sound pressure level is d e ter m ined by tlle highes t level source. Fig- ' :i ure VI - 7 show s t h e perceived n o ise level / or a Boe in g 707 or DC-8 ai r craft at tkke- _._ . off an d landing approach. Th e Pr at t & W hi tn e y JT-3D e n gines for these ai rcraft " y ; have a byp as s r atio of 1.43. A t takeoff the Jet noise is greater than the engine inlet " nois e and nearly as loud as the f an exhaust duct noise. Theref o re , in ternal n o ise t • _ . suppression would pr o Vide very littl e noise r edu cti o n, At landing approa ch, h o w' ever , the jet noise i s much less than the noise r a diated f r o m the engine inlet ar i d fan exhaust duct. I n this case, i nternal ac ou stic sttppre s sio n would res u lt in st i b- statltial r e duction in perceiv ed noise l e v e l.

i_! In flgu re VI-8 the same typ e of pl ot is sh own f o r th e Quiet E n g in e. With a b y. .

p a ss r at ioo f _. 5 t o I , the nois e r e l a tion h as cha nged c onsider a bly. F or b _ tl l t a ke - , o ff and appr o a ch the _e t no ise is c on sid e ra bl y b elow the n o is _ radiat ed front b oth the inlet an d the f an exhau s t duct. Thi s is due t o the high-bypas _ fea s ts, w hich ex.

tracts significant propulsive energy from t h e m ain Je t . Tltis f e hture has been in- corporated in some o f the newest engi n es now in product i o n . • Thtts , int e r na l n oise suppression results in s ubstantial perceived noise level redu d ions a t both ap _ i'oa ch and tak eo ff.

Jet noise. - T h ere a re tWo e xt e rnal noise sources in a t u rbo f an engine s u c h a _ t he Quiet l _n gine, The s e a r e sh ow n s chemat i cally in figu r e VI-9. One s ource o f external noise i s the tu r bu l ent mixing regi o n produced when the f an J et m ixe s with a mbi e nt air, The other so u r ce o f exte r nal noise i s t h e mix in g regi or _ pr od uced w h e n t h e core J et mixes with th e fan J e t an d ambi en t a ir.

W h en two air s tr eam s a t differ e nt v e lociti es m ix , considerable turbulence i S generated and some o f this turb u lent energy is r a diated as S ound. As t h e difference in Velocity increases, the intensity of th e turbulence increases and , hence, th e noise i ncreas es .

Figure VI-10 shows the relation of th e s ou n d powe r ge neeated by Jet m ixing to th e J et velocity. The data points shown were obtain e d fro m noise st ud i e s on J et e n gines an d on nozzles supp li ed with hot and cold ai r . These data point s were taken in the velocity range W hich is typical of turbojet and turbofan engines. Th e slope of J this line on the log plot indicat e s that the noise is proportion al to the Jet v e locity to ' the eighth power. Due to this S trong dependence , a mode r ate reduc t ion in J e t re- i locity will produce a considerable reductio n in noise. For exa m ple , a re d u ct io n ! " f l . om 0 00 0 f e et per se co nd , which i S a typical Jet vel o cit y fo r a turbo j et at take o ff , I to 1500 feet per S econd , which is a typical Jet velocity for a low-byp ass -ratio turbo- !_ f an at takeoff, produces ab o u t a 1 0-dB. reduc t ion in nois e . Furth e r red u ction in the Jet velo c ity to th at o f the Quiet Engine would result i n :further reduction in t he noise. II The Quiet Engine operates with a core jet velocity of ab ou t 11 6 0 f ee t per se cond ii an d a fan jet velocity of about 900 feet per second. Th e se J et velocitie s ar e at !!

I stati _ conditions. The tak e off co mm unity noise m eaSure m ent point is at 3. 5 n auti- !

cal m ile s from brake r el e ase. At thi _ po in _ th e airpla n e will ha _ e a fo _ al _ l veloc- i_ , / fly of ab out 2 50 feet pe r s ec o nd. S in ce th e noi s e produced d e pe nds on the relati V e i ..... Jet velocity , th e Jet noise will be deter min ed by _ ,haust velocities which a r e abbt lt : _ 250 feet per seco n d low er than those s h own .

Data have been t ak en on Jet engin e s in thi s lowe r v e locity l ' ange arid ar e s hown , , !_, t in figure VI-11. The s e data , shown a _ ci r cula r poi n ts in t h e low 0r velocity ran ge , fall ab ove the ei ghth -po We r ct i rVe. Th e reason that the data fall abov e t h e eighth- _.

power curv e is that in this velocity r an g e the no tS e is do m inated by t he in ternal _ machin e ry nois e .

l _ cent work hei'e at t h e Lewis fa n noise facili t y h as de m O n sti'ated ti C . t, ff th e Jet noise is e x tract e d f rom th e t otal noise, t h e Jet noise d oes follow t he eightlt,power , curve. These po i n ts are t he s qt i ar _ po in t s in this fig u re. They / all slightly belo w • the cu rv e but g e n e rally follow the e i gh th-pow e r htw.

P igure VI-i2 sh ow s th _ jet nois e produced by s t _ 0 000-p ou ttd-gross-w e igtlt " airplane having a t ht-tl s t of 90 000 pounds. T h e coi'e Jet pr od uces a noise indic a t e d by th e upper band. Th e n oise pt , oduced by th _ fan Jet i s _ dicat e d l _y th e lower b _ nd. 1

!

The Qui e t Engin e will operate at a pre ss u re ratio of 1.5. At this pressure rati o the J e t n o i se will b e abo L tt90 o r 9 B P N dB.

' F a n noi_e, - O n e o f the ma j or no i se p r obl e m s of the hi gh _ bypa s s -r at i o tur b ofan engine is int e rnal n o i se a s s o c i a t e d with t u rb oma_ inery. Fo r t h e purp os e o f dis - cua s iou w e w i U co ns ider th e propulsion s ys t e m for a 30 0 0 00 _p otmd subsonic air- p l ane o f the DC-8 o r 7 0 7 cla s s . A new air p lane of this size w o uld have a thrust - Weight ratio of a b out 0.3 a l_ d , hence , would r e q uir e 90 000 pounds of thrust. If four en gin es are u se d , as they ar e in these airpl an e s , they W ill each have about 22 000 pound s o f thru s t, which is the size of the Q ui e t _ ngine shown in figur e VI-3.

An engine of this type is n or m ally found to produc e most of it s i n t e rnal n o i se in the fan co m pon e n t and this is quite un d e r sta n dable for several reason s . First , noise essentially is g en e r at ed by the m oving of air , and the f an h aS a hi g h e r ai rflo w than any other engine co m po n ent, in fact six ti mes higher in an engi ne of this type. In addition , the t_ mis the component most exposed to the ob s ervers view and , hence, , its noise has th e easiest path to the observ e r. Also , th e fan is physically the largest co m ponen t , and it has quit e high velocities relative to th e air. All thi B adds up to m ak e th e fan a potentially larg e generator of nois e .

The fan produces noi se of three types: bro ad band , discrete tones, and m r(Ill - pie pure ton e s. B roadband noise is e ss e ntially the noise from turbulence in.th e air and s crubbing of the ai r across hard f an s u rfac e s. It can b e Seen from figure VI-I $ that the spectral plot of bro ad b an d noise shows essentially a str ai ght line at a con- sl an t decibel level. Discr e te tones a r e th.ose at the frequency of r otor blad e pas- s a g e an d it s har m onic S , and they a r e influ e nced g rea t ly by the f lo w co n dit i o n s over th e ro t or blades and by impin geme nt of rOtor w ak e s on th e si st ers. Multipl e pure / to n e s , sh o w n on the fi gur_ as a s mall fi e ld of fi n e spikes at a l ow frequency , a r e mo s t proba b l y gener ate d b y sho 0 .k p at t e rn _ o n the rotor bl ad es caused by sup e rsonic _ ' ...... l o cal r e lative Mach numbers. Th e y can occur an ywhere in th _ spectrtt m and at azy t l e v e l. Bro ad b a nd noise is associ at e d With any flow proc e s s an d, henc e , wil l always be present in the spe ct rum o f a fan n oise. The discrete tones or multiple pure t o nes _ - m ay or m a y not be pro m inent hi _ particular s pe ctru m , de pendi ng on t he fa n d e sign and the operat ing con d itiotm. The extettt to which th e y ar e present in the dpe ctrum can affect th e char a ct e ristics of the f an -g e nerat ed n oise gr e atly. On the typical sp e ctru m shown in figure VI - 13 th e noise w ith th e highest decibel l e vel is the di s- crete tone and this tone would do m inate the noise that th _ Qb_ erve t " h e ar s . T his , in !

fact, is the whine that is s o charact e ristic of t h e n oi st _ heard fro m ctti _ r e_ | e l air- I plan es having turbof an engin e s. _ o m_ ds do min at e d by m tfltiple pur e tones _ end to have a ra spy, buzzin g quality to them. ' i , // B e fore d e t ai ls o f our f an research program and some of the re s u lts o f the p('o - g r am are presented , mention must be made of-the other in tern a l noises, tho s e from _ I

!

t h e cor e engine and the fan turbin e . N or m ally, these produce a level of noise which is low enou _ that i t is ins i gnifi c ant in t h e total engine noise. However , if.the f an co m pon e nt Were qui e ted suffi c iently either by design or by sttppression treat m ent, the noise levels of the core engine and the turbine co uld possib l y become do minan t.

Indeed this can happen, and fo r thi s reason the Quiet E n gine has Some acoustic suppression treatment in the turb in e exhau st duct.

T h e facili t y shown in f igu r e VI-14 was buil t to ac o ustically evalua t e a Se rie s of fu li -sl _e fans. Drive is by an electric m oto r of 3 7 000 hors e power thro u gh t h e shaft t o the in le t of th e f an .r ot or. Far. _ fleld nois e is m easured by microphones on poles, some of which are sh ow n in the figu r e. A c u taway view of the fan nacelle (fig. VI - 15) sh ow s su ch in t e rnaldetail s as the drive shaft, rotor blading, stat o r blades, c ow ling, and exhaust n o zzle. Th e basic aerodynamic design of the t_ zn, such as its thrust an d bypaS S ratio , is very clo s ely det e rmined by th e performanc e re- , quirement s of the engine it s e lf. The fan des ign er th en would like to s elect an op- timum c o m bination of tip Spe ed, pressure ratio, weight flow, an d. m echani ca L arrange m en t to yield the lo w est pos s ible noise output. To do th is he m ust knoW th e relationS betwe en the s e v ar iou s de sign parameters and the t ypes an d amounts of noise pr od uced. The. _ e relations are not kn own at p.res e nt, but they sh o u ld beco m e evident as the result of a test program wi th a series o f full- S ize f an s of w , rytng de- s i gn. These design s are shown in figure VI-16, where they are d e i in_d in terms o f tip speed an d pressure ratio. Tip s pee d is u s ed bec a use it is an important noise p _ , _ ameter whi ch intl i cates th e expect ed level of re la tive Ma ch numbe r s on the blad in g. Thus, it affect s th e g e n e rat i on of all thre e type s of noise. P resSure ratio is used in th is figure simply as an Indicator of blad e load in g at a given sl _ ed, or how mtich tu rning the blading has to do o n th e air pass In g b y it. TheSe fans ar e all about '/ 2 inch es in ti p diam e ter; they are all s in gl e -s t ag e fan s ; and m os t of th e m were design ed for a b out 22 000 pounds of thrust. A pressure ratio range of abo u t e 1.4 to 1. 6 is of hlt e rest for t h e CTOL airplane and most o _ th e fan desi gns ar e con- centrated In this ar ea. In the aerodynamic desi gn of a f an a tra de oOff ca _ be m ade It between tip s pe ed and loading. An _ ong othe r thi n gs, the designer m tist ] _ y# how to balance these t w o param e ters to yield the l ow est poss i ble floi se ou tput. This c a n be determined by sys temat ically varying each p _ ramete r selz _ ra t eiy and studying the noise generated. T e st in g a se r ies of _ aris de s i gned at the sam e tip speed, for exam- pi e , wi l l In dicate ho l y th e b lade loading affect s the noise output. S im ilarly , t est s on a series of fans d e si gn ed at a particular p re s sur e ratio will show h_ v tip speed affe ct s the noise ge ne ra tion. As a r e sult of the s e tests , the in ter S ' e l at io n s _ u non _ blade loading, rotor tip speed, an d noise gener a tiorl S h ou ld beco m_ apparer i t . One f an , desi gn ed at lo# v tip speed an d l ow p r e s sure ratio, r _ pr _ sents ffie type of design whi ch m ay be required for a high-bypass STOL engine, or for a V ' i _ )L eng /I ne. In d e signing these fans, generous advantage was tak e n of Som e of the previou s ly known techniqu e s for reducing noise ge n e ration. For exampl e , these fans were designed with a wide spacing betw e en the rot or and s tator blade rows to al h _ w the r otor blade wake s to dissi pat e before hitt in g the stators , thus lowering the ge neration of diS- crete tones. In additio n , i n most o f th e f an s t he number of r o tor and stato r blade s was chosen to minimize the p ro pa _ tion of discrete t o nes.

One f an , th at designed at il00-feet-p e r-sec on d tip speed an d a pres s ure r atio of 1.5, has alr ead y been te s ted. It was found th at the maximum forward-project ed fan noise occurr ed about 40 ° off t he fan inlet axis. Figure VI-_lY p r e se nt s a _ pec- trum of sound p r ess u re le v el tak e n by a microphone at that location. This s pec- _ trum was obtained _ zi th the fan oper at ing at take o ff speed with the st an dard exhmzst _: nOzzle. Th _ base level on _ e s pect rum at " bout 7 5 to 80 dB iS the broadband part , _• of th e overall noise. The tall spike is the disc r e t e to ne which occurS at rotor blade • pas s a _ frequency (i n th i s case about 2VO0 Hz) , and th e two spikes at higher fre- i quencieS are its harmonics. At_about 400 and 1500 hert _ th ere are s mall contri- lzltions of multiple pure tones. Ob viou sly, the l oud e st s ound in this particular spectrum i s th e di scr e te tone and th e observez _ beloW the airplane would he _ tr the whin e p r e vi ously de s cribed. When th e fan is run a t the same spe ed but wi th th e ex- :_ haust nozzle area inc r e as ed abou t 10 pe rcent, thus in cre asin g th e level of r elative Mach number on the bla di ng , the spectrum (shown in fig. VI-IS) is consi d erably di t - ferent. It s t ill has about the same broadband base level bu t n ow th e discrete tone lma dropped co n siderably , as ha v e its har m on i cs. The m ost striking difference be- / . : tw een the s e two spe ct ra is th e pr e sence now betwe e n 100 0 and 1500 hertz o f a baud _. of very lottd multiple pure tones. I n this case t he ob s e rve r Would hear the raspy sottnd of a noise do m inated by multiple pure tories. The in t ere dt ing fe at ure of this / i : comparison i s that the inc reas in g presence of multiple pu re tone _ i S ac com pani ed by a co nsid e rable l ow erin g o f th e disc re te t o ne w hi ch is in _ e h i gh -annoyance f re - # qu e ncy range. The result is that the calcu l at ed perceiv ed no ise level of the fan is so m ew l zat lower ed . These two s pect ra W ere o btained a t r ot o r in let relative Mach numbers of 0.9 _ an d 0.99, respe ct ively, w hi ch in dicates a very s ensitive re lgXton of th e multiple pur e tones with Mach number.

Calculations w ere m ade to e stim at e, on t _.e baSi S of th e be st awzilable d i tt _, the fan m achinery flo i se to be e X pected fl-o m t u v o o _ engines of varying de s i g n, all s ized to px' o vide a t otal of 90 000 po un ds of thrust fronl four engin e s. Figu re VI-19 presents the cal _ ula _ d per c eived no is e level of siflgle - stage fa n s at take o ff opet,- -° atioiz as a flt n ctio n of t he fan pressure ratio. The cu i*v e is showil as a hano i o _ _ e _ - eral dB width, w hi ch re flects the uncertainty in the calcul at ion. DZz _ t po in ts fo r the f an just discuss ed and for a ha lf- sca le m odel of one of the Quiet Ezigtn _ fan s ate sh ow n to verify th e ievel of th e cu rve. The Q uiet l _ ng in e fret , Which op e rate s at a preSSUr e ra t io o f 1.4 d u ring ta k eoff, is show n to p roduce about 105 PNdB of n oise, which Will nearly s atisfy the curr e nt FAA regulations for new airplanes. An upp e r lim i t to the-pre s sure r a tio that can be obtained from a Conservative singl e -stage fan i s s how n on th e fi gu re at a r ound 1. V or 1.8. Thi s m eans that H a hi gher p r es- i sure ratio is needed f o r some p articular engine cycle a tw o -stage fa n would b e re- - qui r ed. So m e t h ought _ _ so been given to u s ing a tw o-stage t an, even at lower pressure r at io, t o decrease blade loading. For these tw o re as ons, figur e VI-20 i ' presents t h e sam e cu rve-as fig u r e VI- 1 9, with the addition of an equivalent noise ' predict i on curv e for the tw o- S tage f an . The level o _ the t _ o-stage-f an cu rve i s sup- portedby data po in t s fro m the JT S D engine and from a modified TF-$9 engine. The . tw o-stage fan is s how n t o typically p roduce ab out 7 o r 8 l _d B m o re n o ise than the • , s in gle-stage fan. This is the resu l t of in tera ct ion effe ct s cau s ed by th e addi t ional - " bladtng. Some propul s ion sch em e s d o re quir e high preSS ure rati o and , of cours e , • : m u st accept th e no is e penalty of * .he two - st ag e fan. In the l ow er px'essure ratio ra nge _ however, th e s in gle - stag e fan is clea _ ly th e better. It has been shown that the S in gle -S tage - fan engine by it s elf w il l near l y meet the FAA re gu lation S ; more - over, its noise can b e lowered still further by acoustic suppression tr e a tm e n t, which is the subject of the f ol l ow ing sec t ion.

Suppress ors

Large no i se r ed u ct io n s ca n be achiev ed by th e use of a cu ustic s u ppression ih I th e in ternal passages o_ an engine. This d is c u ssion stre ss es t h e use and results of _' a typical suppre s sion eoftfiguratioft in th _ _ tiet Eng in e. J .... T h e cu taw a y vieW in fi gu re VI _ r e veals the acoustic liners of t h e Q uiet E n gine. i The forward-radiated no is e o f th e fan is partially ab so r bed by th e inle t s U ppre S sor 1 co n sisting of the t _ ' eat ed c0wl and splitter rings. The l ong fan exhaust du ct i s ' tre ate d on t h e inner a n d outer wa i ls to pro vi de suppres s ion of the aft - radiate d_ 'an I_ no is e. The internal pass ag e leading to th e com p ressor is al _ o lined. The acousti c _ .

lifting in t he core nozzle wi l l r e m ove turbine tto is e. _ L T h e lnechan ism s beh in d the ope ra tion of the acoustic li n ers can be shown by _ observing the d u ct s in a l itt|e more detail. A small s e ctio n of the acous l tic l i n er is _ shown enlaz , ged in fi gur e VI-21. A pe _ or & t _ d plaie ove r a horieyco mb ba cking cav- . * . q _ . ity is fur th e r m agriffied in the ins e t. When t h e acoustic pressure is high on th e dt z ct I side of the pl at e, a j e t w i ll be formed i _ lowing in to th e back ca vi ty. The _ et will flow out of th e orific e wlt e n th e _ tcotmtic pl- e ssure drops b e l ow ambi e nt. T he ab sorption o f acoustic p owe r is accomplished through the turbule n t dissipation of th e kin e tic ene rgy of th e se J e ts. Th e resonant prop e rti e s of th e liner can b e adjust e d to pr0 _ i d e 1 '/ 6 ,, m aXimu m a cou s tic poW e r dissipation a t a giv e n f r e q uency, such as the frecltteney o f _ nois e pr o viding th e g r eatest annoyance. This is done by adjustment of the b a c k : ' cavity depth and the sheet thicknes s and poro s ity. Adjusting these parameter s allows us to obtain the proper wall acoustic im pedance to optim i ze the coupling be- tween the line r and the duct. _ Other wall con s tructions c ou ld also be used. The face plate could be m ade of !

• very fine-wire s or screens. The back cavity co u ld be paclted with a bulk a b s orber S uch as fiber g lass. In these c onstru ct ions the main dissip at i on m echanism would be th e v iscous s h ea rin g effe ct in the ve ry s m all passages of the m aterial.

. In the fan and c ompres S Or passage s , com m on materials , su ch as aluminum, ' are sufficien t fo r the absorber. In the core engine exhaust, m oder at ely high- :_i tem pe rature m at erials must be u s ed.

:i / A 6-foot-diameter turbofa _ similar to that of the Q uiet Engine has been te st ed her e at Lewis. Figure VI-22 show s the.inlet suppressor u _ ed wt _ rids fan. The . , _ uppre S sor iS m ade up of alined outer cowl.and three spltfler ring s lined on both sides. The out e r ring has a lined lengt h of about 3 feet. All line rs are constr u cted • of alum in um perforated _ la t e bonded , tO an alum inum honeyco m b b acking. The h det i s shown with a bellmouth which . _ s used for static t esting only.

Figure VI.- 2 3 sh ows th e eff e ct of th e inlet suppressor on the narrow-band noise spe ct rum at th e po sition of m aximtlm fr on t -end no is e. This occurs at 40 ° off the • _ nlet axis. The f an w as operated at takeoff s pe ed for thls data. The upper curv e represents the noise for the hard-co, _ l conflgUr ai to n . The lower is for th e a co u s - • tic al ly lined confi gu ration. The liner absorbs n oise over _tv e ry wide frequency rang e , with subst an tial reduction _ between 400 and 10 000 hertz. The maximum absorptio n is near the blade paS Sag e freque n cy. The har m onics of the discre t e tone • ar e al, J o r e duced. The multiple p u re ton e s r _ ear 500 an d 1500 hertz have been _m - : pletely removed.

The suppressors p rovi ded perceived nols _ level red u ctions of 12 PNdB at s lm- : ulated takeoff condition _ and 13 PNdB at app r oach. It is felt that a 15-PNdB re- .. d u cti o n could be realiz e d with proper suppressor design an d b al ance. This is re- flected in figure VI-24. T h e predi c t ed PNL versuS fa n pressure ratio is shown.

The upper curves represent the turb o nlachine ry noise f or one- and tWo-stage en- gines without suppression. The lower cu r ves indi cate the results wi th t h e exp e cted 15-1 _ dB reductions.

Choking Aerodynamic ch oking is def in ed as the acceleration of airflow to sonic velocity IV7 / , t \ by reduci ng the flow area to a critical value. T h e concept of chokin g for noise re - duction i s shown dia g ram m atically in-figure %1 - 2 5. The n oi s e bein g ge nerated propagates t o th e left against t he airflow until It reaches the choked-r e gion, where the air velocity is equal t o th e sp e ed of sound. T he sound waves b egin to reflect but are believ e d to lose their phase ident i '.y due to the i rr e gularity of t he shock wave. Thi s condition i S termed " b _r d " acoustic chokin g ; that is, none of the noise generated d ow nstream can propagat e ups t re am beyond th e choked regio n . If this phenomenon is g oin g to be used in an airc t 'aft engine inlet, it must be variable since the engine r equir es varyin g airflows for i ts operation. Figu r e %1 - 2 6 shows.a varie t y of m echal _ is m s that ca _ he use d to a c c om p l ish thi s . Th e first type o f m echanis m indicates variabl e geometry vanes , some of these vanes m ay be trans la ted, rotat e d , or exp _ d e d; the second type indica t es that the cowl m a y be contracted; th e third type includes var ia tion of the c e nterb od y by either e xpandin g it or t r anslatin g It.

• _ In figures VI- 2 7 an d VI- 2 8 are typi cal results ob eye d by r o tating the i n let guide ' v an es to cause chokin g . The outer curve in fi gu re %I - 27 shows the uachokednoi S e level at all azimu t hal locations i n a fo _ -ward quadrant. The inner curVe s h ow s the corresponding c h oked noise level. Elgure VI-28 shoWs s pectra taken at the po int of m axi m um sound pre s sure level (SPL), which was at t he 30 ° azimuth fro m the com- pressor centerline. These m easure m ent s were m ad e us ing a m odel compressor ope ra ting at high rotational S peeds, hence, the hi gh frequency noise. The upper curv _ S hows a t yp ical unchoked Sl _ etrum. The maximu m peak is that as soci at ed '.

wi th the fun damental b la de pass in g frequency. The lower cu rve shows th e choked spe ct f _ n n taken at the same azimu th . Not o nly have the fundamental frequ e n c y and its harmonics been dr a stic al ly reduced, bu t the br o adband level has been si go ffi- cantly reduc e d also.

S in ce the ch oking concept iS so pro m i s ing acoustically , we will now examine th e de s ign requirements associated W ith us in g a choked m echanis m in a production en- e gine for co mm erci al flight operation S . Of parti cu lar concern is oper a tional safety.

Whatever m echanis m is c h osen for choking m u st be f ail-safe. It is most desirable to use _ tdevic e that i s m echanically simple. A ce rt ain amo un t of effo rt is be in g di- rected toward t hi s goal. In addition, there are d e sign considerations to min i m ize the effe ct s of in let press u re recovery, inlet distortion at the fat l roto r , c oW l drag, and we i gh t.

.; . C TOL Sum m a ry

" , Based o n.t h e p r eco dinE di s_ ussio n o f the tw o types o f n oise ( je t and fan) and nois e estim a t e s for CTOL a trcr _ t , we can estimate the total p r opulsi o n system noise, Unsuppressed fan noise follows t r e n d s like tho s e shown ,n figu r e VI _ 29. A s ingl e -stag e f a n s h o u ld b e used if a t a l l p ossi b le , Noi se r e du ct i on can b e achiev e d b y l _ ssiv e li n int _ s or eholCedt n let s . These t e c h niqu es ca _ give r e duced ma ch in e r y no ise , as shown i n the figur e VI- 2 9. Th e Jet noise from th e core a _ d bypas s s t r ea _ us follows t he trend shown. Suppressed fan and Jet noise sources a r e pretty well bal an ced arotmd a fan pr e ssure ratio of 1.5, the ratio for which w e de s igned ou r Quiet Engine. Nois e is even l o w e r at lowe r fan pre s suz ' _ ratios , but en gi ne _ diam e ter in c reases and dr D . _ losses rise, A! L levels in figure VI - 29 a re for 90 000 I pound s of thrust at 1 0 00 feet alti tu d e at full takeoff pow e r. T] _= l e vels th at app e ar to be achi e vable are in the r an ge o f 90 to 95 PNdB. T he se v alu es are ab out 10 PNdB b e lo w curr e nt FAA regula t ions for n ew CTOL ai r craft.

STOL A IRCRAFT

The air transportation system of the 1980's will include no t only CTOL ai r craft but w il l als o pr ob ab l y in c lu d e su b stantia l nu m bers o f STOL and VT O L ai rcraft.

An adv an tag e o f s hort - takeoff - and - landin _ (STOL) ai rcraft is tha t . th e ai rport can be * closer to th e c e nter _ f toW n . But noise le v els must therefore be lower th _ n for , , C'TOL ai rcraft. Short - takeoff -an d - lan d ing field require m ents r e sult in consid e r - ably increased installed thru s t levels co m pared to conv e ntional ai rcraft. Atld th is incr e ased engine p ow e r produces m o re noise thatt wflh CTOL ai rcraft. The reason : . l ot " the high thrust is that part o f th e propulsion is used to augment t h e lift. ' There i :' ; are sev e ral lift augm e ntation schemes, and th e se may th e mse lves be nois e sources.

• :; Two o f th e se lift au gm entation systems - the bl ow n flap a n d the att gme ntor w in g - are exam in ed. •

. , B l own - Fla p Nois e i

In co n side r ing th e bl ow n-flap g e ome t ry, th e e mLo ha si s is on noise prod u c t io n _ d when t h e flap is lo w ei'ed into th e eng ine e X h a ust, i!

'_ A four.engine STOL ai rcraft u sing bloWn flaps fo r lift au gm enta t i u _t is s h oW = l in , _ _ f i gu r e VI-30, It lo o ks v ery simil ar to con v entional ai rcraft , ex c ept that the fl a p is a

I

m u c h la rge r i nc omp ar iso n t o th e main w i ng se ctio n.T h is ca n b e se e _ inth e wing cross s e c ti o n at t he l ow er r i g ht c orn er of t h e fib re . 4 , T wo important que s t i on s a b out bl oWn-flap n oi s e nee d to b e ans w e r e d: , (I) C an t he fl o w i nt e racting with th e f la p b e co m e a si gnifi can t e xt e rn a l n oi se source 9 i (2) C an t he fl ap redirect int e rnally g e ner a ted fa n nois e ?

I n o rder t o g e t s o me pr e l i m inary a ns wer s t o th e se qu estion s , noi se me asure m ent s were mad e on th e b l ow n_fla p mod el th at w as u se d in the 8 T eL wind tunne l t est s .

In f igure V I - 3 1 a cro s s sec tion o f th e blown-flap m odel is s hown. T he m m_ imu m flap ang le o f O0 °, to s i m u l at e approach, is s h own. No ise m e a s ur eme nt s were a ls o m ade for takeoff cond i t i o n s with 30° flap d e fl ec tion and f or crui se cond i tion s with no ,, flap deflection.

This mode l • i s s ho w n i n fi gure VI-3 2 . The model is ro t l g hly 1 / 1 0 scale ba _e d on a 100 000-pound-gross-wei g ht a i rcraft. Only on e 6-i n ch-d ia meter fan Was used. A '_ 60° fl a p defl e ction correspondin g to the approach condition is shown. Noi se w as m_ asured in t h e p lane perpendicuh t r to the wing looking at t h e underside, and in the i plane o f the wing l oo k in g at the e n d. _ : i A typical noi sespec trum pr od uce dby the mod el i ssh o w n i n figure V I-33. ii S o und press u re l eve ls v e r s us freque nc y are sh o w n f o r no flap d e flectio n . O f cou r se, : all the frequencies are high because o f the s m all scale of the model As m ight b e I expected, w h e n there is n o flap interacti on wi t h t he fl o w, the spec t rum is t Ypi ca l !_ _ o f a n y fan. At the high e r freq u encies there are inte r nally gen e rated fan n oise spikes !i [ a t bladepa s sa g efreque n cy a n dits h ar moni cs,a lo rl g w i ths o me bro ad b an d fan n o ise, i In th e expe ri me n ts n o S ign ifica n t c h , _g_sin f an noisestr engt h o r d irectiV i t y We re !

f o und as the f la p was loWered i n to th e fl _ w . !

• i .... At lower fre q ue n cies there i s some evide n ce of externally g e n era t ed J _ t n oise.

' • .. _ The s li g h t bu l ging o f th ec urv eis as s o c i ate d withthis jet n oise.The cha ng ei n th i s • _ ' :' l o w-frequ en cy hu m p w h ic ho cc u r s W h e n t h eflap is f ull y l o were di nt o thee x ha us t : fl o w is sho wn in figUre VI-34. While fa n noise rema in ed about the same , t h e l o W- frequency h um p was substant ial ly increased as the flap wa s lowered to th e 60 ° approach co n dition. Thi s in crea s e is du e to the fan flo w int eract ing W ith the flap.

The b an d indi ca tes d i fferences ass ociated with differe n t m e as ure m ent ang les. Re- sults for a 30° flap deflection at takeo f f conditio n s fell be t wee n the 0° an d 60° flap , . cases , as wou ld b e e x pected.

If the m od e l fan had be _n acoustical l y treate d o r design ed with c[ uie tii _ _a tu r es, ' the h igh- f reque n c y fa nn o i sew o ,l d be lo wered an d thelo W er-fre qu e n c y f l ap in ter- a cti o n no is e c o tfl d b ec o me th e _i o min a ni s our ce.

Th e ex pe rimen t al i n te r ac tion no i s eda t a w e r e s ca l e dtoful l sca l e, a nd pe r cei v ed noi se l e v e l s w e r e calc u lated. Su ch s ca ling i nv o lv es sh iftin g th edatato lo we r f r e - <t uen c i e s , s_ t h at some unc ert a int y res ult s . _owe v e r , s ca l ing doe s a l low th e us e of th e m ode l re_ ult _ t o m ake es ti m at es o f h o w the in tera c ti on no i s e co n t r ibut es to the t o ta l n o i se of th e bl o wn _ fl _ p p ropulsio n _ys te m .

F l rst, c on_ Id er t h e no i s e prod u ced b y th e tu rb0f_ m s alo n e, as s hown in fig .

ure v, . 3 _ . A 500-fo o t d i s tanc e is tre e d here since t he STOL ai rc raft w i ll o p e r _ tte very cl ose t o de nsel y p o pulated a r ea s . The _e t n oi s e be com e _ t h e d o mi n a t e sou rce above pr e s s ur e ratio s of about 1, 4 f_ In f it _ ttr e VI-36 a third baud i s added, which i s the flap interactio n noi se that was obtained by _ cali ng the mode l results, It inc l ud e s both th e noi s e o f th e ro d1- i r e cted J e t and the n o ise produce d in th e p r o c e ss of re dir0cti ng i t. T he int e r a ctio n . , no i s e b ec o m e s t he d o m inant noise s _ u rce at apre ssu re ratio of about 1.3, w h e rea s Je t noise alone dominated a t 1.4.

T h e two q u estions posed ear " ier have now been give n at least preliminary an- swers. First, a significant inte r a ct ion noi s e can be g ene r a t ed. I n the process of ' turnin g th e engine exhaust f l ow. Second , no significant red ir ection of the in ternal e n_ d n e noise w_ s observed.

Augmenter-Wing Noise

Shown in figure VI-37 i s _ . sche m at i c o f the augme n ter- w ing propulsion system.

High - pressure air is ducted to the wing durin g takeoff and landin g . Thi s ai r ex - ' hausts through a s lo t in the wing and flows through an ejector t nad e up of flaps o n the wing tr ai lin g edge. Tlds arran g e m ent provid e s the hi g l t lift coefficients fleeded for sho rt takeoff and l an ding.

Since the pre ss ur e rati o across the w ing slot is hi gh , about 2.5, cons i d er able .. noise could be generated when this air m ixes W ith ambient ai r . z, v t JVe r , s in c e m ost o f t he m ixi ng o ccurs b etween t he flaps, some beneficia t _ , _ ,eldin g m ay help t re d uce thi s noise. T o study this nois e source, a n au _ e t l tor - wing noise test fa - cility was built at Le w is. This facility is shown in figure VI - 38. P resS uriz ed ai r is fed to the wing throu g h the pipe s hown. Th e w ing i s a 6-foot s pan of an approxi- mately 1 / 2 - scale wing for a 1 0 0 000-pound-gross- we ight ai rp | al _ e. T h e air flow s _ out of the slot _ nd flows be tw een the tw o flap s . Noise m easurement _ were m ade over a r an ge of pressure ratio s for several slot height s , Fi gu re ' / 1 - 39 shows a typical di r ectivity pattern o f th enois _ prodt _c ed by this w ing . These data were taken witha slot h e ight of 0.7 in ch an d a pre s su r e r atio of 2. The tw o - lobed pattern centered ab out the jet exhau s t is typic a l of Jet floise. D U e to the ang l e of the flaps, the lower lobe is almost directly b 6 1 o Wth e _ing . I _ t l Hn _ , flyover , sound in thi s lobe i s radiated s traight d own ward t ow a r d the g t utmd. S ou n d 181 !, \ l . /" f ro m th e other l o be travel s f arth e r be f or e h i ttin g t he g round. Thu_ , e v e n tho ug h th e no is e pow er of th e two l o b es i s n e arly equal, th e no is e re p re s e nt e d b y th e l ower lobe appe a r s loud e r to an ob se rver on the g rotmd.

Figur _ t 7 1.40 S how s th e spect _m of th e n o i s e in this l owe r lobe. T he n oi s e ts _ I br o a d b and , w hi ch i S typical of J e t n o i s e. Ther e ar e no s pike s in th e spe ctru m , a s the x ' e were in th e s p e ct ru m s of ma chinery n o i se . Th e s e d at a W e r e al B o taken with a _ s lot h e i ght o f 0. V in ch anti a p x , e s_ ur e x ' _ io o f 2. T h e l o cation o f t he mi crop ho ne re lative to th e wing l _ sh own in th e s eh em_ ttie i n th e uppe r r / g ilt eo _e_ o f t h e ti p, r e.

T he s pectrum pe al_ at about 8000 h ertz. The fr eq u ency o f thi s pea l_ id pr e dict a b le !

from th e s l ct Z e i g ht. B in ce the s lot hei s t i s -s m all, about 1.5 i nc h e s for th e full - s cale p l ane _ th e noi se gen e rat e d by the w i ng wil l be i n the freque n cy rE m _ e that the ear i s most sen s it i ve to. _ In order to stcdy ways of r e ducing the noise p r oduced by the wing , the flaps _i were lifted wit h acous t ic ab s o r b ing m at er i _ d , a _ shown in fl _ r e VI-41. T h e m a- teri al is simila r to tha t us e d in th e f an tests. It is m ade of a perfor a t e d sheet i , bonded to a hon e ycomb b _ ckin g . The ma terial w as place d on the flaps as shown in the figure, i' The effect of t hi s m ateria l on the noise spectrum i s show n in fi gu re VI-42. i Data for th e tm li ned flaps are shown as cir cu la r points , and the s qua r e po in t s a _ e i".

the data obtain ed w i th the lin ed flaps. A b out an 8-dB r ed uction w as obtained at i 5000 hertz. EVen though t his lining produced only 2 PNdB reduction in t he overall I linin g th e flap s urfa ce s.

noise, the test did in dicate th ai the nois e pl'oduc ed by th e wing can be attenuat e d by I'i i T he re Sul t S of thes e tests we r e scal e d to a 100 O00-p ou nd-gross- w e i ght air- !i plane and are s hown in fi g u r e VI-45. For a plane of this size , about 30 000 pounds _l_ • of thrttst would be obtained from th _ _ing . The _ sti m ateS indiCate that a plane of _!_ • this size would produce about 105 to 115 PNdB wi th no linin _ on the flap surfaces, t_ ff the fla p s ar e lined, th e no is e may be r ed uc ed to about 100 to 110 PNdB. , !_ t Conventional ai r craft noise tecl _ ology an d _ xperim en f _ d work h as been applied i!

to STOL aircraft noise prediction and has led to the pre dict ion of ezt g in e inle t floi se li_ as th e do m inant nois e sour ce , as sh own in _ ig U re VI-44. Thes _ predi ct ions are scaled to the anticipat ed siz e of the aug m enter-wing prototype _ ircrafl. Iti addition , low Jet noise, in ci n der to obtain a high p r es s ti r e ratio f or small du ct s in th e whi g , _ . the u nS uppre sS ed f an tnl _ t noise i s est ima t ed to be in the t _ tn _ e o f 120 PN d B, as the engines are assum ed t o b e designed for lo _ Jet exhau s t velocities and , hence , I !\ s h ow n. Tiffs is fo r two or thr _ e stage s of compres s iol _ , ff la rg e r ducttn g with ,_ l ow e r press u r e ra tios could b e tolerated , a dist inct l _ ot s e _ dvant _ ( _ tbout10 PNdB) could b e ft ai iz ed . This i s S h ow n by the lower bafld for fan ini e _ noise at low pre _r e ratio s , This d e crea e _o i n n o i se i s asso ci a t e d with t h e use of a si ngl e - s ta ge . ' an. Thi s inl e t noi e _ c o ul d b e f urt he r r e d uced b y eith e r inle t cho king , , o r inl e t s up p re sso r s , Th e inl e t s u ppr e ss o r s y s t em i s s h o w n in f igu r e VI- 4B , Th e suppr e s s ors sho wnin thi s f i g u re ar e e sti ma t ed to be _ p a b le of no i s e r e, , A ction of th e ord e r of 10 F Nd B. If a w i n g s lo t pr ess ur e rati o o f 1 . 4 w e r e a c c ep t ab le , t h e i_ O S. P NdB go al m ig h t be reach e d , It a hi t che r wi n g s lo t p res s u re _ t io i s re q uir ed, a c hol(In g mec h a n i s m i n t h e in l e t i s i n dica t ed , A c e md i d _ te c h _ kln _ m oeh e m l e l m i s shown in fi br e VI-4 6, This o h o ltin g devis e is cal l e d a _ rid in l et, It is m_ de o f a i s eri es of air f oils arrm _ e d across th e in l et like a venetian blin d , I n t h e crui s e m ode , , those V an e s a r e i n s e ve ra l r ow s , W hen c hok i n [ _ i s requ ir o d p all the row s are ' br ought toK e t h e r to r e duce th e inle t a re a , This tech n ique h e m bee n tested s ucc ess - ful l y on a co ,scads ri g an d on 8-inch- _ d 12 - inch - di s .m e te r m odel f ans . C h okin g sh ou ld red u ce th e t an inl e t nOisOb y 2B to 3 0 dB, whi ch is within t he 9 B I _ N dB goal, How e ver, t h e nois e er e at0d at t h e win g slot will now dominat e , and som e m e t hod must be found to furt h e r red u ce this lev e l befor e 9 B PNdB can be realized, ,

i

ST O L Summe ry i

The l imited amount of b l own - flap dat a a vaila b le in dicate that this augme n tation system ma y h a ve a noise problem. S m all-sea ls m odel tests have indicat e d n o r e - di rec t ion of the f an dis cr ete frequency noise, but h e _v e s h ow n ar _am plification of _ th e jet noise. The amount of am plifi ca tion is e s timated at abo ut 5 PNdB for t he wing configu ra tion at takeoff. At a pp roach , the ampiiflda t io n i S large r beeattse o f the l a rg e r flap immersion in the Jet. How e ver, the J e t velocities are loWer at approach.

The Jet noise beco m es a donlinan t source ab ov e a fan pressure ratio of ab out 1.3.

_ Acoustic resu l ts o _ o n e m odel o f an augmenter wing s h o w rather high no ise le v els _ and poin _ to the need for further work in d evelopin g a win g Configuration with accept - I able at:oustic and aerodynamic ch ar acteristics.

For n o th t h e au gm enter win g and th e blown flap i _ iS clear tha t considerabl e _ ef f ort i s required to produc e s yste ms with lo w noise output .

i

V TO L A irc r af t

._ Several characteri s tics of t h e VTOL airplane and it s propulsio n S yst @m stand out. Fi r st, o _ co u rse , the v ery reason for exist e n ce c i the V ' I _ L airpieZe is to I al l ow it to oper a te from popu _a t l o i _ cent e rs, whic h m e an s that people will be very , close to the airpl an e durin g it s takeo f f and lan din g oper a ti on s. In addition , duri n g

!

th e s e o p e r a t ions , th e airc_ t v elo cit y wri t be q uite Iow , a n d he nce th e r ela t i v e Je t ve l o cit y wiil be a b o ut e q ua l to t he abso lut e J e t velocity, Thls i m pli es a n incr ea s e in Je t n o is e g e n era ti o n o v e r th e C TO L or S TO L air pl _ te s , Al so , li f t f a n s will be u s e d t o g e n era t e t he low-pressur e. ra ti o Jet s for v e rti _ l t h_st , an d t he y will tend t o b e lrmt _ ll od v ery c om p a ctly t o pe_ , m it good c _ -4 1 s o p orform_ m ce . As a r e su l t , it will b e dif f icu l t to su r r o und t h em with a gen e rou s a m o u nt o f ac o tmtic stt p p v e s s io n , F i nally, an eno r m ou s amount o f thrust will be i nst a ll e d on the VTOL a ir plan e s o : tha t i ts t h ru s t o t o .g ross= w oi_h t r a t i o w ill e x c e e d u ni ty, a n d th e n x ld od t h_s tw ill ' ' ' \ r e pre se nt an inc r e ase d s o ur c e of n o i s e. Tho s e ch _ tr ac t e r is tic s co mbi ne to ma ke th e VT O L no i se prob l e m d i ffi cu l t. Beca us e there i s o nl y a l imit e d a moun t o f dat a a vai l - a ble o n VTO L li f tin g s y s te ms , ma ch in o _ noi s e d a t a and Jot noise dat a have b _e n u se d to e s ti ma te th e noise per f or man ce o f $_me VTOL propulsion s y s i s , in a sim ilar for ma t a s w as us e d f or the CTOL an d STOL systems , f igur e VI - 4 '/ pr e sents c a lc u l a t e d perc e iv e d n oise l e vel s at 500 f eet from t h e tal(eo ff point o f a VTOL aircraft. The propulsion systems have b e en ge ne r aliz e d so th e y c an o per ate : ' at different pressure ratios, The calculations assu m e a 100 000-pound -g ro ss . , weight ai rplane having twelve 10 0 0 0 -pou n d - thru s t units. The fan m achi n ery noise ' _ of these 12 engine s unsuppre s sed follows th e trend shown. As discu s sed ea r lie r , ff s p ace w e r e available, this fan m achi n ery n oise could b e _ ppr e ss e d b y 1 6 P NdB. ' i_ Acoustically t h ere is n o r e ason why such suppression would not b e possible. How- ! ;i' ever, there m ay be a severe m echanical p ro ble m i n Just fi n di n g sufficient a r ea t o ' :_ treat in the small e ngine nacelle. Because of t h is , the ov e rall f e asibility of I B- PNdB suppression is problematic at this point and n e eds further study. The je t i_ ' m ixi n g nois e of th e VTO L lift fans wi l l produce t h e tr en d _ llow n , obvio us ly falling , _ off ra pidly as t h e p r e ssure ratio of t h e f an s d e creases. The overall impli ca tion of ii these cu rves is that VTOL lift fans must op e rat e at a fai r ly low press u r e ratio ff "_I_ ; they are to produce low n o ise l e vels. _ additio n , a s e rio u s effort must be m ad e to i_ in troduce sufficient ac o ustic suppression int o the housing around these lift f an s to reduce the ma chine r y tmise.

/ SUMM A RY i Fo r c o nv e n tio nal _ r c r _ the r eisa p o t e tltlal f ors ub s t_.nti_ fLtrt h et' l i o lse r _- : ' d u ct t on s b e low p r ese n t F A A c e rt ifica t .'.o n limits. C o m pl e t e s y st em _ ork is U nderway w it h in NASA in the Quiet E ngine Pr o gl-a m .

: S TO L aircra ft r e q uire eve n l o w _rno isele _c els , while _ t h esame t i m eat l d _g a not he rno i s e p1 " odu ce z ' i nt he l lft a u g menta t i o n s y s te m. Bo t h e x tt _rnMl y -blow t t.

fl ap a ndaugm e n to r- wi n g co n ce p ts n ee d c on si d e r a bly m o i 'e r esearcheff o l' t i noix_ e _ ' i • ' tOd e fine _ ccurately t he severity o f their noise and t o devise nots _ reduction scheme s . To a _ i _ ve noise levels as low as 95 PNdB at 500 feet Will r equire te ch - _ nology beyond that currently availabl e . Thi s technology, however, m ay well be de- veloped ff serious attention iS givell the p r oble m , as has been done in the case o f CTOL aircraft.

VTOL aircraft usin g lift fans h ave two no ise problems: Install at ion o t enough acoustic treat m ent to s upp re s s the hu l machinery noise is diffic u lt in the typical Co m pact in stallation. L ow -veloci ty Jet noise appears to f ollow the ei ghth -power la w; 0 but in order to have a low-Jet- nois e s i gn atur e these lift fans will have to op c_ tt e at low pres s ure ratios. _,i ! '

-!

i !

i

!

• !

FAA CT_OL NOISE REFERENCELOCATIONS , TAKEO _ ,, _ ' _ FROM . BI _ KERELEASE / _ APPROACH REF '_ POIN T ., I N MI i FROM T HRESHOLD c s- _ 6 78 z , Figu r e W -1 i7 i CTOL NOISE RESTRICTIONS 110 -- I ' _ .

APPROACH AND S IDELINE t P ERCEIVED . . ,- NOISE 100 LEVEL, EPNdB , . . ,

I I I I .... "

0 2 4 6 8X][O _ Cs-s 67 eo AIRPLANE GR O SS WEIGI'fi' , LB Figur e V] - 2 1 8 6 ", 3 -- e

OUIETENGINE

CS-S 6783 ; Fi g ure V I-3 J INTER N AL NOISE SOURCES- QUIET ENGINE .,, D ISCHARGE FA N / _\\ ' ,.-. \\\\ ', . \ ',. \\\_ ' __\ .. \\ .. \\\\ ............. _\ NOlO:It " COM-PP _ TUR BINE ' NOISI _ J C _ -$6623 FigureVI-4 ,.' ,: _ .... .

EXTERNALNOISE S OU R CE S FAN JET NOISE_ C O R E JET NOISE !'

i C8-56573 Figure VI -S OUIET ENGINE:WITH SUPPRESSION t88 • 707 / DC-8 ENGINE NOISE BYPASS RATI O ,1 .4 3 dB i _ , PERCEIVED • NOISE L EVEL I _ : _J ] D FANDISCHARGE

" Il l] = = ;;' _ ' _ '

TAKEOFF APPRO ACH CS- 5 6871 FigureVI- 7 QUIET ENGINE . NOISE .... ' BYP A SS R A TI O , 5 .5 I ; _ ] J FA N DISC HA RGE _ . _ _ FANIN L E T

_J ET

PER CEIVED N OISE LEVEL _ _ _ " lO PNdB _._.

i i' : TAKEOFF APPROACH C5-56870 FigureVI -8

i

EXTERNAL NOISE SOURCES TURBULENT ;: \\ MIXINGRE G IONS

\\

. .' : I I _ I _ . _ CO"E J E T

CS-56586 . , Figu re VI - 9 JET NOISE CORRELATION FOR VELOCITIES OF I000 TO 2000 FT / $EC _]_0dB S (TAKEOFF) t

_ OUw _ _ _ _ , _- O F ,. _ -- -

ENGINE ENGINE FAN CORE JET JE T t QUIET Q U IET

I I I I

. 500 700 tO00 1500 2 000 ' J ET VELOCITY , FT / SEC cs -s 6_ sz FigureV I -]O g ET N OISE CORRELATION FOR VELOCIT - LES : OF 500 TO 2000 FT I SEC i ' 0 E NG I NE D A T A .oJ E] SUPPRF . SS _ FANDATA , _. ., _ " - _ TUR B OJETS S (TAKEOFF) , ' : POWER d B !; . / ,; -- _ -- - - _ TURBOFA N S -- - - _ ' ,.

• _ , _ e _ _-

' ' .... , .. .. o , , '; _ [] _ QU IE T QU I ET

: __- u - E N G INE.EN G INE

,, _ - "- FA N JE T C OR E JE T

' I I I I

500 700 1000 1 500 2000 JET V E L OCI TY , FT I SEC cs-s 6 s sl-- Fi gure VI-11 ,' TURBOFAN E N GINE JET NO IS E " , 3 00 0B} - L B GROSS WEI G HT A I RPLANE ;" 1 3 0 9 0(]OB - L B THRUST , 120 t

, , ; ; .' [ IO00FT

' PERC E N E O 1 10- _\\\"" ' . .

NOISE LEVEl . , .__._.

P N dB 100 x\\ " _ .... • '_ ' _"_"F" : _ __ FAN JET , _i_"

, , 80 ._._i _._.1 I I I I I ..... I

1. 2 1 .4 L 6 L 8 2 . 0 2 . 2 2 , 4 2.6 cs. s 6 s 87 FAN PRESSURE RATIO F i gure Vi- 1 2 :" I' 1 9 1 t ,', FAN NOISE COHPONENTS DISCRET E T ONE AT ROTO R BLADE PA SS A GE FRE Q UENC Y -\ CL U S T ER O F SO UND MU L TIPLE _ _ DI SCR ETE TO N E

L E V E L 11111111 _

PR ESS UR E PUR E TONES ] L_ MON ICS \ _- BRO A D ' B AN D NOISE I , I I I I 200 500 1 0 00 2000 P 0 00 10000 FREQ UENC Y Flgu r e V I - 1 3 c s- 565 o 7 ' !: ; FAN NOISE F A CILITY i' .

I i; i : i

• i

C- 7 o . 3_ 47

il

c,s - _ 7 ,_o8 I!

Figure VI -14 = t9 2 \ _t F AN NACELLE ' ' CD 2 1 0 (, 80 ' '.' ._, : ; ' .. : I Flgu reVI-1 5 i : I

i

i F U LL-S C A L E RE S EAR C HFAN S i 1.6- 0 0 I'

.. i . _ - 0 6 ) o '

" 1 .4- 0 " PRESSURE RA TI O , ,!

, , 1 . 3 - • .'.. 1.2- 0

1 . 1 I I I _ I , I I

6 OO _ I00 0 I L _ O 1 41W) 1 60 0 c_.565 o, T I PSPEED, FT / SE C Figu r e VI- 1 6 ! . FA N INLET NOISE- STA N DARD NOZZLE TAKE O FF SP E ED 1 10 F 40° PRESSUR E LEVEL ,

d B S O

SOUND 90_ _ j __.

( 3 0.- _ I t t I 200 _ 1 000 2000 5 000 10000 FREQ UENCY, Hz Figure VI - I7 cs.s 6 so 6 FAN INLET NOISE- OPEN NOZZLE = TAKEOFF SPEED 110 - . I 00 -40o •,. . , ., _ SOU N D 90 ' PR E SSURE ' LEVEL, .. , dB 80- , " 70- " 200 5 00 1 0 0 0 2000 P000 l O000 . " FR EQ UENCY , Hz ' . C8-56505 ,-. Fi gure VI - 18 i_ lg 4 • \ /

FAN MACHINE R Y NOISE EI| TIMATE S -

S IN G LE- S TA G E FANS

:; l_ L _ S _W E i o At N E '

" , , u o , F ^ N _ : . _ , _ / # II/ I_!f . _'. .

_ v c_ / /// _i N _ . ST A G E ':I/i FOR :_ ;I N OI _; STA OEFANS :_ i

'": k0 ...... 2 . _

.. , , c s . 5 , _9 1.2 . . I. 4--. i. 6 - L _. z.O ._2. _ .4 . " " , , . F i ( NP R E SSU R E:RATIO ,_ F i g u re VI-19 . FAN MACHINERY NOISE E STIMATES- ' S INGLE-STAGE AND TWO- S TAG E FANS 300 O00-LB GROSS WEIGHT AIRPL A NE '" 90 O00- L B T HRUST 130-- " TW O -STAGE.FANS " / - -MODIFIED TF- 3 9 _ ///////////////////// _

,. _ . _ ,, / ., / // // // _ / " / "'"

._ //// . / / " , ', -J T_ .....

, . , P E R C EIV E D SINGLE - S T A GE FA N ; ' NOISE LEVEL. 100 1_ ' l t_, I I I I. I _ ....

FANPRESSURE RAT I O c _-_ , se o F lg ure VT - 20 .. I 'i FAN EXHAUST DUCT WITHWALL O_ AIL

l_r l

C8 - _ 6514 FigureVI - 21 INLET DUCT WITH ACOUSTIC TREATMENT # t / , FigureV I - 22 ' tl _ ' • :-. , FAN INLET NOISE- WITH AND WITHOUT SUPPRESSION ': " TAK E O FF SPEED STANDARD NOZZLE I

• = I

SOUND P R ESS LE VEL , . : ... . dB 80

6 o ll , i _ - 'F

200 500 10 00 2000 5000 10.0 00 FREQUENCY , Hz c s- s 6 s l z _ Figure VI- L 3 • ,.. F A N M A CHINERY NOISE ESTIMATES- ' "' WITH A ND WITHOUT SUPPRESSION -,, 3 00 O00- L B GROSS WEIGHT AIRPLANE • 9 00 0 0 -L BTHRUST .... ' _' 1 000 FI" .... P ER C E IVED l l O - _ /////////////////) N OISE _" " "\ _"_" - - -WIT H S UPPRt_SSION ":' 80 I I :1 I I 1 , 1 . 0 1 . 2 1. 4 L5 tS 2 . 0 2 . 2 2. 4 2 _ 6 FAN PR(ESSURE R A TIO cs - s 6s t l F I gureV I- 24 ' : ' " CHOKING CONCEPT ! , i t; • r : . : CHOKED i,_ .

• • R EGIO N t' • i Figure V[-25 c s- 5_5?z ;1 ,,, j INLET CHOKING HECHANISHS !i ' ' . ' VARIABLE . GEOM _ I'RY BLADES ORVANES • t VARIABLE COWl .

"VARIABLE ' CEN _ RBODY EXPANDING 01 _ T R ANS LA TING C S * _6 S ? Z ,' , , Figure V] - 26 , i g 8 _ : I b SOUND PRESSURE LEVEL- . i " WITH AND WITHOUT CHOKING UNPJ4 _ C _n__ . 10dB.

__I

., - CHOKED i _ I '--._ / f\ / , : .

• '- / I i . _ i'm c s . +++++ !+ Fi gure VI - 27 !

. + INLET SPECTRA -W ITH AND WIT H OUT CHOKING . .- 3 0 ° ANGLE • • !+ ' ; 10di_ UN C HOKED ,. SOUNb PR E SSURE LEVEL C H OKED + d _ 2000 5 000 10000 20000 I ' F REQUENCY , Hz C8-5657 0 FigureVI-28 i

I

. ,, . .. ,

EXTE R NALLY-BLOWN-FLAP STOL AIRPL A NE

It. >

FLAP

I ' i ?

F i gure VI -3 0 Cs . 56800 l O0 B L OWN- - FLAP GEOHETRY :, _ __ P AN G LE 600 FLAPAPPROACH !, • CS - , _ ( ,820 _ Flgu reV I - 3 1 I i BLOWN - FLAP MODEL !

k t Figure VI-32 201 • BLOWN-FLAP MODEL NOISE SPECTRUM - T o° FLAP SO UN D 10dB P R ESS U RE- -J -" LEVEL

I I I I I , I I I

1(30 2 0 0 500 I00 0 2000 5 000 1 0000 2 000 0 FREQUENCY, Hz c s - s 6 68s !

Figur e VI - 33 FLAP I N TERACTION NOISE _ , , _ _ o _p A / _ ,

T

lOdB j ., _ / V LEVEL , 4\ ,\\\\_ -

, I I I I I I I I

100 200 5 0 0 1000 2000 , 5 000 10000 20 000 FREQUENCY , Hz cs-s 66 s 9 Figure V] -3 4 NOI S E ESTIMATES FOR BLOWN - FLAP ENGINES - , = WITHOUT FLAP INTERACTIO N 100 000 - LB GROSS WEIGHT AIRPLANE 12 0 -- 6 0O00-LB THRUST

.500 Fr

1 1 0 -- JET S (FAN +CORE) 00 FLAP PERCEIVED NOISE 100-- SUPPRESSED FANS LEVE L , : _, PNdB ' , _

9 0 i

' : " 8 0 I I ,I I _ ' _

1. 0 1.2 1. 4 1.6 1.8 2. 0 PRE S SURE RATIO cs.s s69 o F i gure Vt - 35 NOISE ESTIMATES FOR BLOWN - FLAP S YSTEM- INCLUDING FLAP INTERACTION ,. 100 O00- L B G R OSS WEIGHT AIRPLANE ii ' : 120-- 6 0O00- L B THRUST .... . _ OFT B L OWN FLAP INTERACTION t 110-- JETS (F A N + CORE) 0 _ F LA P, : PERCE N E D NOISE LD / EL,100-- SUPPRESSED FANS PNdB 80 ___1 I 1.0 1.2 1. 4 1 . 6 i.8 2.0 PRESSURE RATIO c s .s 6691 Figure VI-)6 | _ .O3 \ AUGHENTOR - WlNG P ROPUL S ION SY S TEH Figure VI- 3 ? / AUGMENTOR-WING N OISE TEST FACILITY !, i I.

i i Figure VI - 38 2O4 AU G MENTOR-WlN G .. DIRECTIVITY PA T _TERNAT 50 F T ...

.: SLOT HEIGHT , 0,7 IN, a P RESSUR E RATIO, 2 . 0 \ P RESSURE L _WL, ., , 90 10 dB ,':

\

G6- _b S R3 Figure V 1 - 3 9 AUGMENTOR-WINO t , NOISE SPECTRUM AT 50 FT - UNLINED . SLO T HE I GHT , 0 . 7 I N . ; P R ES S URE RAT I O , 2 . 0 .... : I00 -- , - , 0 000 O0 , ,

• , : o" o

' " i _s -- o ° o

,, • : , ,,,, SOU NO o o

.Z • ..' : :. P R ESSURE 90 " - O O

: . LEVEL, O

-! dB - O

e c 1 I I I 1 I

_ 0 "_ 0 ZOO0 2000 5000 100(3 020000 " FREQUENCY, HZ cs-s6s .

• Figure VI-_ O

FLAPACOU S TIC LINER

i . , CS . s6 B9 4 i£ Flgu reVI - 4! , i i , _ i iL ' AUGMENTOR- WIN(5 ii_ NOISE S P£CTRU M AT 50 FT - LINED AND UNLINED ,_ _ ooo oo'__._, , 9S-- O I _ SOUND O O O PRESSURE I "1 [ 3 El LEVEL, 90 -- 0 dB 0 I ' 1LINED ,,

m ,, I I .... I I ,.,i,, I

zoo _ i ooo zooo 5n z ooo o2 o - ooo

FREOUEkr C y, Hz cs . sssS 4 ¶ FigU r e VI-42

,d

AU GRE NTOR-WlNGPROPUL S ION S Y S TEH WITH INL E T SUPPRES S OR Fi gure VI-4,5 AUSHENTOR-WlNG PROPULSION S Y S T E H WITH i , ' CHOKEDINLET , , , _ [ _ t , _ - C S o5 6 7 ? fl i FlgureV t -46 ' soe NOI S E E S TIHATE S FOR VTOL PROPUL S ION 100 O00-LB GROSS WEIGHT AIRPLANE • ' " TWELVE 1 0O00,-LB THRUST E N GI N ES ' UNSUPPRESSED SINGLE- 120 ST .A GE F AN S 110 F AN JE T PE R CE I V£D NO I SE L EVEL , 100 SUPPRESS E D SI N GLE-ST AG E FA N S PN d B :_ 9 O I • ' !, : 1 .2 1. 4 1.6 1.8 2.0 F AN PRESSURE RATIO cs-s6 s o 4 , : Figure V 1-4 7 ' #' / !"

_,', ,, , ......... - • • j . ' we

N71 -19458

VII. LOW - COST ENGINES F O R AIRCRAFT

Robert L Cu m mings an d Harold Gold

Gas turbine e ngines h ave now almost co m pletely taken o v e r the field of large aircraft propu l sion. Their s m all size atld weight also make them very attractiv e fo r l ight aircra _ t. A ma j or obstacle , not technica l but economic, is the very high :: cost of .c urre n t gas turbine eng i n es w h ich Subs ta ntially re s tricts thei r use. T h e approx im ate cost of Several current genera l aviation engi n es are shown in th e fol- lowing tabl e : Turbocharged pi st on engin e s and prop Direc t -dr i ve 285-hp engine ........................ $ 1 0 2 00 Dlrect-<{r t ve 290-hp engine ....................... 1 1 300 Gea r ed 425-hp e ng ine ........................... 1 _ 400 Turbine e l _ gines: L ow presS n re ratio 10 2 5 - 1bf thrust tu r b o je t ............... 2_ 2 00 Turboshaft 605 - shp eugiae ........................ 35 $ 00 _ Tur b o je t 9 850-1bf t hrust e ngine ............ 8 5 § 00 Fa n J et _ .00 - 1bf th rus t engi n e ............... 6 _ 000 It _ hould be kept I n m ind th at th e g e n e ral avia ti on light plane m u st be s u i ta ble for a ,.. _ , retail sales prt _ e o f around $30 000 f o r a singie - e ng lb e airc raf t and $ 4 _ 000 f or a lig h t twiu - etlgine aircraft. The tu rb o char g ed pi s ton e ngi n e s are th em s elves quite V costly. Fo r th e hig h fligh t speeds th at we will co n side r, a 425-hor s epower engine -._ would be required , b ut the price o f over $17 00 0 is already too high. The cu r r e ntly ava i lable t u rbojet , tu rboSha _ t , and fanjet e ngines are m uch too costly , wi th p r ice s • ' rangi ng _ rom $ 2 2 000 tO over $0 5 000. ,i Looki ng at the s e prices , we can n 0 Wund e rSta l id the com pe titive i mpactof a re a lly low -co st gas turbi n e ef_ ine w ith a 1000 - potutfl static _ ea level th ru s t having i....

a to ta l m a m ffactttring , pri ce of $5000 , or $5 Pe r pt _ u n d o _ thrttdt. S u ch an e_ in e would provide importan t perfo n hance gains | o r light ah'cr a ft and als o h _ ve a very t,_lportant price advantag e ov e r ei th er curr e nt pist0h engin e s or Ct l rrent Jet en tw ines, ( _ I t ord er to sell at 1 / 5 the pric e per poudd of thrus t t _ cui _re nt J e t 6 n _i fle s , I _ w _ ver , th is e ng i n e would require really major design si m plifications ari d manu / actur , i n g cost reduction. ,; PRECEDIN G P A G E B L A N K N O T F ILI _ IE I _ , _ : 21i i : I .......... - :_. , _. , _ , _._._ , _ . ,: ........ . _,. : _, _I .... I- F-_ - -i ........ I IIIII IIIIII1"1 I .......... i _ .......

• . { _ Thi s paper d i scu s s e s the major re s ult s o f a p rogram which is concerned wi th such low - cO s t e n gi n es. Th i s sm a ll Lewis program ha s ha d . o n ly limi ted a s sistance f rom co n tractors. I t has bee n e sse n t i a l ly exploratory , cover ing a n u mber o f pro - bl em areas: The mai n areas c o vered in thi s paper are (1) e n gi n e cycl e a n aly si s a n d airpla n e perfo rm a n ce studies r ela t ing t o the c o st - per fo rma n ce trade o ff q u estio n , (2) engine co n fi g urati on s w h i ch are being e m phasized , (3) n ew co m pre s s o r and tu r- _' b i ne r o tor c o nstruction m ethod s a n d results from spi n te s ts, (4) construct i on a n d i test results o n a low-cost annu la r co m bustor, (5) results on a new type of hydro- m echa ni ca l f uel co n trol , (6) work o n l o w-cost accessories , a n d (7) curre n t pla n s f or construction a n d test of co m plete tu l 'boJet a n d fa n Jet engines.

ENGINE DES IGN •: To achieve low cost on a device as com plex a n d critical as a gas turbine engine , • w e must be w illi ng to giv e up some p e rf ormance to li mit th e desi gn te mpe rature to a i , level low enO U g h to ob ta in re li able machi n ery , requiring o n ly low cost m aterial s , and to limit the pressure rat i o t o reduce the tip speeds , the st ress levels , and th e num- ber of expensive s i ages. The decisions as to just how far to go wi th th is appr o ach .- are a m atter of judg m ent involving a tra de of f of e n gine pe rfor m a n ce aga i nst it s ini- tta lc ost....

• The perf orma n ce tra d e o ffs on fuel co n sumpt i on an d thrust are s h own i n f i gures !_ VII - 1 a nd VH-2 , where we have th e spe cific fuel co n sump ti o n a n d s pe c i f i c thru st of tur b oj et and f a nJ et e n gine s plo t ted agai nst pressure ratio. The s e results are f o r a / 'i fl igh t s peed o f 450 mp h at a n altitude o f 2 5 00 0 f ee t , w hich are bei n g u sed a _ a ' . typ i cal adva n ced g e n era l aviatio n air craft desig n point and for a t u rb in e inlet te rn - , ' _ p eral n r e of 130 0° F. Th is choice o f a l ow tu rb in e tem pe ra tu re w a s mad e to pr o m o t e J ec onomy a nd r e li a bi l ity in th e d e si g n.

: F ro m figu re VII -I w. eca n see t h atch o ice o f a l o w pres su re ra ti oof 4. 0 d o es i n - deed cause about a 2 5 pe rc e nt h ig her f u el c on sum p ti on th an i f 12.0 w{_ re ch o s e n. The pressure r a ti o of 4. 0 ca n be ach i e ved wi t h ab out ha lf th e n u i nb e r o f sta g e s r eq ut i' ed f o r a pr e ss u re rati o 1 2 .0, however, an d th is is th e _pe o f tradeoff we m ust n _ a lt e ff w e a i ' e to achieve lo w co st . For th e tu rbo j et, it is our exp 4 _ c tati o n that the main a p- p lica tion would be for mi ssil _ and dr one en gin es. The des ig n poin t for the tt l r b ojet .. ha s th er e fore bee n ch o se n , as sh own on the fi g ure , at th e lo w est pres sur e r _ ti o th at i s consistent w ith mo d erate perf or m ance.

The fa n Jet e ng ine desi g n po i nt tak es a dvanta g e o f t he a dded fan sta g e to o p era t e at a h ig her o verall pressure ra ti o. Th e fan sta g e a ls o prov i des add iti onal propuls i ve mass f l ow. Bo th of these factors im p r o ve th e en g ine pe rfor m a n ce. Note i n fi g - 2 1 2 !

ur e VII - 2 that t he : f an J et d es ign poi nt uses a m o d erate b ypa ss r atio o f 2.5 a ad a fan pressure r a tio of 1, 3 and thereby a chi ev es a s i gnifica n t perfor m an c e i m p r ove men t , co m par e d to a simple turb oj et. At t he design poi n t a s pe c ifi c fuel eo ns u m p t ioa -of 0 .9 0 pou n ds pe r ho ur p e r pu un d o f thrust i s ac h ieved w i th o n l y o n e add iti onal s ta g e.

T h e fa n J e t engi ne , thus , h a s m oderat el y good f u el c o _ su m p tton a n d s hould b e t he most attractive for aircraft p ropulsio n .

Both th e turbojet a nd the fa n Jet eng in es in fi g ur0 VII-2 ha ve very us e ful specif i c thru s t levels, in th e range o f from 4 6 t o S 5 p ound s per poufld per second o f core en- gin e a i rflow. Here the cor e engine airflow is e o ftsidered a s th e r ele v ant pa r a m ete r, since the core en _ e contains the expensive co m po n ents. This Spe c ific th rust level m oans that 10O0 pounds static sea level thrust can oe obtaine d With either a tu r bo je t or a f anjet engine w ith an inl e t diameter of less than 10 inches. Th e gas turbine en- g in e is, thus, _ much s m aller than th e pis t on engine it would replace. A f ur t h e r factor, _ whl _ h sh ou ld be noticed in figure VII-2 , is the specific thru st advantage of the l ani er, i , , z For the same thrust level, th e fanJ e t has a size ad vantage that would tend to offset i the cost penalty due to th0 added co m plexity of the fan Stage. _ With d _ s i gn point s _ hosen, as shown, th e perf J orm an c _ o b ta i ned for both the i turbojet and th e turbofan is substantially lower than most modern j e t engines of th e i type d e scribed elsewhere in this publication. Th e critical qu es tion , however, is whether th ese pe rfor m a n ce levels are good e n ough to provi d e u s eful rang e and op- erating c o st fo r a li g ht ai rcraft. Th e performa n ce figures for a li g ht twin-engin _ ai rplan e , which is designed to cruise at 450 miles pe r hour at an altitud e of 25 50 0 feet are as followS: Desig n cruise at 2 5 000 ft, m ph ........................... 4 5 0 Airpla ne charac t eristics: : , Twi n engine gross weight, Ib m ......................... 50 00 : _ Fuel weight including 3 / 4 hr reserve, Ibm 20 0 0 t Cruise th rust, lbf / engine .................... ........ _50 Pe rf or m a n ce at c ru ise: Specific htel cons u mp ti o n .................... ....... 0.90 Takeoff dis _ nce, ft 1050 10_ 0 Useful range, mi ........................ ; , .......

Comparative fuel cost (recipro ca ting e_ lne = 1. O) ..... , .... • • • O. 81 213 _ . _ It ha s a typi c al f ue l-to- gr o ss -w e lght _ at l o and f u el r ese rv e and a wi r _ lo a dlnS l ow enou g h to provide the low tak e o ff a n d l anding s pe e d of 80 miles poe hour. For th i s a i r plane th e two 1000- peu nd takeoff thr u s L _ i ues are ca p able of achieving a low tak e off di sta nc e of only I050 f o ot. Th e very u se ful r an ge of over 1000 n _ il es i s also obtained a t a n o per ating fuel cost level le e s tha n p is to n e n gi ne s of equ iw , lent thi r st.

These co s t esti ma tes , of cOUrse , tak e into a ceount the low e r cost of Jet fuel co m- pared with aviation gasoline.

This airplan e must be stress ed to operate at flight s pe e ds at least twic e a s great as current li ght airplanes and b e c a p ab le of cab i n pre s surization. To achi e ve these qualities at reaso _ bl e cost, work will be necessary on the all ' plan e f rame as well as h :he engine. Fo r the present, / _owev er , the discussion w ill be limited to the engi ne .

Duri ng the course of thi s progra m , n um e rous tur b ojet a n d f a nJ e t d e s / gn s have been co n sidered and co m par ed . For application to light al rc r afl _ th e i m proved range and the l O wer noise levels of th e fanJet caus ed major interest to b e cente r ed i on this type engi ne . The obj e ctive of the fanj e t design st u dy was to obtai n th e i_ economy o f it s smalle r size core e ng i ne wi th out addi _ Cost s beeause o f u _ du e co rn - .

ple x i ty . !_ A config u ration bei ng s tu died for the fan eng ine is shown i n fig ur e VII- $ . This figure sh ows a g eared fanJet engin e , which i s n ot only a low noi se , low tip speed _ i f a n but also a s i ng l e -shaft , two-be a ring design, f or the cot 6 e ng ine. T h e 1000- i_ pouu _ .thru s t engi n e uses a 15-inch-diameter fan , a 10-inch-diameter , five-sta ge !

compr es sor, and a two- Sta ge axial tu rbine. It also uses a 660 _ ho rso po w e r gear il , . _ box with a speed redu c tion r a tio of 2 to 1. This g ear b ox all ow s th e turb i ne stage s to operate at high speed and to share the work and , therefore , minimize s the diam- eter a n d the number of stages requir ed .

":". Th e d e _d g n of OdS g_ box has been _ tn d i ed by th e AIH so n Division of Gefteral Motors under a NASA contl ' act, add t h eir r esults i _ di cats. that a con _ e t v al iVe _ r 650-horSepoWer , coaxial geari ng syste m can be produce d at a t o tal cost of a p proxi- mately $600. U sing such a ge ar box avoids th e co m plexity of coaxial shafti ng wi _ its ad di tional high DN bearing and s eals. The shee t m etal c oinpre _ sor _ ui d tur b i n e shown in th e figur e will be des cr i b ed later. This g _ ared fan e ng ine congigu _ tiOn ltas a n um be _ of advantage s . However, oth _ r f a n e ng ine d e_ uS are ad a p ta ble to th e low-cost fabrication ts ehniques a n d are also still raider consideration.

In ad di tion t o gene r al a v ia ti on appli ca tion s, Lewis h a s been wol ' ki a g in eoo pe ra lio n with the U. S. Navy to d et er mine the applieability of th e low-cost t leS ig ns and _ f ab ri ca tion techniq u es to e ng ine s for this / di e s and drones. S u eh ex pe ndable e ng ine s a re referred to a s ordna nc e engi nes . A typical cu rret _ t _eq ut r_e nie nt for such an e ng in e i s g iven in th e _o l l owing tabl e an d th e r es ultin _ f our- stag e comp re s s o r en - gine d esi gn i s s h own in f i gu r e VII - 4.

Cru i s e thru st r e qu i r e d, lb f . . . _ t 5 0 Sea level s tatic thr us t , Ib f . . . 6 5 0 En gin e dia me t e r l imi t , i n .... 1 2 Specifi c fuel con B ump Uo n . , . . <1. 8 Th e require me n t is f or 3 5 0 pou n ds o f th x_ St a t th e design poi nt o f Ma ch 0.8 a t an a ltitude o f 20 000 feet and a se a lev e l thrust of 690 potmds. T he weight limit is 100 pou_ ls, th e di a me ter limit is 12 i nches , a nd the specific fuel consu m ption must be below 1 . 8. Both wind m ill s tart under ra m _ o n dltiO us (cruise) and imping e - m ent st a rt at s ea level a re required. The desi g n fli g ht duration is only 15 minutes.

T he turbojet engine des i gned to meet the s e c ou dition _ uses a f o ur-stage cast axial-flow co m pressor and a single-stage in ve s t m ent- cast turbine. Fig ur e V II-4 al s o shows the si m ple sha f t and bea r ing de s ign. Th e engine out s ide dia m ete z ' is only 11 ½ inches , and it is est i mated to weigh less th an 100 pounds and to have a specific fuel cons um ption of 1.3. With these s i z e a n d perfo r m an ce figures, it will provide overall range an d pa y load m uch b e tter t han san b e aclfleved by a rocket engine. It i s also a ttractiv e in i t s promise f or l ow production c ost .

FABRICATION DEVELOPMENT

In addition to this eng ine design work, i n v es tigatio n has been made Oftechniques for the fabrication of l o w-co st axi al -flow rotor stages. Th e m ain approa ch es being t considered are (1) ca sting a n d (2) sheet metal stampi ngs . The cas ting approach is already kn ow n in the i ndustry and will no t be f ur th er discuss ed i n this paper. It al s o appears that the stamping approach, usin _ co in ed blade profil e s , m ay have a better pot ential for low cost and reliability. This construction h a s ther e fore bee n e m phasized i n a fabri ca tion an d te st progr am ; its main features are illustrat e d in f igure VII-5. Aft ax ial-flow co m pl'e _so r is co mpo sed of tWo she e t-metal di sks , W i th b la d e s for med on tabs on the edge o f th e disks. A pair of such di sk s is placed to{ _ e th er to give th e solidity re q uired for g ood t r an s o ni c ax ial-co m pressor per- formance , and fitted ihto a pair of _in gs. The ri _ s are slotted to r e cei v e the blades , or a filler t _ ateri sl is used b e tween the blad e s. A com p l ete sheet- m et al co m p re ssor rotor is shown in figure VH-6. There a 10-inch dia m eter sheet m etal Q A , .

!

co m pr esso r i s in sta ll e d in an e nd pl a t e a ssembly a s req u ir e d for spin te s ti r ,g. A coini ng or cold for g i n g proce ss h a s been use d to f or m t he b l a d es, T h is produ c e s a cc ur ate r e p r od u cible c o nt o urs with s ma ll l eadi ng- a n d trailin g - e d ge r a dii, T h e r e i s , c o n s eque n tl y , n o k n ow n a e r od y nami c pc _o r m a nc e pol ity impo_ e d by u s e ot thi s t y p e of constructi o n.

X number of these s heet- me tal c om pre ss or rot o r s h a vin g sl i gh tly differing c o n- _ s tr u cti o n h ave b e e n bu il t a n d tested i n a Lewis s pin test ri g . S t r ain me aSure m ent data , take n dur i ng sp i n t es t i ng of the sl ott e d h ub versi o n of th is compressor , ar e sh o w n i n f igu re V II -7 where th e c al c ul a t ed s tr ai n at two po in t s o n the r o t or is p lo t- ted against t h e r otational speed. Also , plotted i n t h e f igure a r e th e m ea s u rem e n ts I taken from s t ra in g a ge s attac h ed to th e rotor a t th e sa me two p o ints. The meas ured va l ues cor r espond very w e ll w i th t h e c al culated ones and the m easured strain i , I is li near. The rotor has b e en tested to 2 5 p _ reent a b ove i ts design speed , indicating a good streng th margin and safety factor at the op e rating s peeds of our fa n Jet and i turbojet e ngines. 14 The she e t-metal e onstr . ac tt o n tech n ique is also bein g evah u tted for app l ication ;_ to axial-flow turbine rotors. A m odel of th is is sh o wn in figure VH-8. He r e , th e pr o b l em is somewhat more c o mplicat e d B ecause of th e additional camber of th e fur- _.

bine blad e s , but stre ss calcu la tions show the designs to be attractive with re s pect ; ,_ to both centrifu g al stress and th e rm al stre S S. A s s h own i n the figur e , the turbi n e i .

rotor w o uld consist o f two sh e et -m e ta l pla tes f or m ed with co in ed blades o n th eir I ', !i rl m , fitted toge th e r and placed intO ri ng s i n a m ann e r S imi la r to that of th e co rn- l_i"_ p resser. A test rot o r of the type show _ i s now b e ing co n structed. !_ In additi o n to this s heet- me tal appl.Oaeh , both cast and welded turbine designs are also bei _J g i n v e stigated.

T h e fan requires an axial stage , as s h own in figure VII - 9 whic h i s lar g er than the core engine and w hi ch has subs ta ntially longer blades. Thi s stage is also th e One w hi ch woul d be th e most adversely aff e cted by dam age from foreign obj e cts . !

Therefore, for this rotor we have investigated the use o f h olloW, stamped sheet- m e ta l blades that are re m ovable. T h ese b_ d e s are stamped from two pieces of stainless-steel s h eet m etal and are joined together by welding or braz ing . T hey have a stamped or cast base to p r ovide a low-cost , freely pivot ing attachm en t to the hub.

Because of their hollow construction, th e y have the advantage of a high resoi l ant vi- bra ti o n fr equ ency a nd low ro o t s tr esses and can operate witho u t z 'equiri ng m i dspan vibration da m pers. The hollow construc ti on also f avors a l ig htwe ig ht , low-cost rotor disk. Thes e blades h ave bee / _ vibt'ation t e st e d to verifY th e h ig h be _ dt ng i re - quenc i es , and th e co m plete rotor shown ha s bee _ spi n te sted to a s pe ed 40 perce n t above th e f a n rotor design speed.

21 6 t \ A low-cost a nnttl a r e om bu _ tor ha s been con s tr u cted-fro m perfo r ate d sheet me t _ i ( f ig. Vii-10), This co m bustor us e s a c oo ling ai rfl o w htyer o bt a ined from sl0e- c tal o r i o n _ ti on of the p a tt ern of holes i n com m erc ia l pe r / or a ted she et, al ong with a simple p _ ttet' u of punched ho l e s . It als o us e s a Very simp l e a i r - a t omi Z ing fuel n o z- z l e. Thi s c o m b us to r rut s been te a ted , a nd results oft co m b us tion effic ie n c y over a r_tnge of fuo l_a i r r a ti o are show n i n f i gure VII-11. G o od d e sig n p oi n t c om b us ti on ef f ici e nc y w _ts obh t i n ed an d ign ition an d bl o wout ch a r a c te r is tic 8 wer_ v e l ' y s_ t Ls f a c- to r y. In R dditiot _ , th e co m bustor e xhibited a good t em per ature vari _ o n pattern fa ctor of a bout 0.20 an d a d es i g n pressure drop of 6 perc e nt at a r e l a tive l y high d es i gn poi nt -M _t c h nu mb er. , , , Accessories ar e also v e ry impo r t a nt to the overal l cost a nd size o f small J e t engines since t h ey tend to b e large and ex pe nsive. The most critical accessory is _ ! , t he eng i ne speed a nd fuel control , which m ust provide capability f or rapid thrust i response without sur g in g t h e c o m pre ss or or exce e din g th e al loWable tu r bine te rn - :_ ' perature. The control i S , therefore, a k ey to the safety and reliability of the e ntire !- , en gi ne. It is , furthermo r e , c r itica l to the cost an d may add up to 20 percent of the i cost of the en g ine. The fuel control has , the r efor e_ been extensively investi g at e d , u sin g a hydro m eehanic al control ba s ed on use of a zero grad i ent pump speed sensin g techniq u e. The principles on whi ch the work i s based ar _ s h ow n in fig t l _ e VII-12 wherein a parameter , consis ting of fuel flow di vided by s peed an d by the ambient pressure correction facto r , is plotted as a fu n ctio n o f t h e co m p re ssor pressure ra- tio.

: ' The value s plotted i n figure V H -1 2 ar e for a current typical turboj e t engine , a nd it will b e noted that th e steady-state ope r a ti ng li n e , t h e surge te m pe r a tu re li m it, a n d the co m bustion blowout li m i t ca n all be approx i m a te d by th e line ar re la - _ li on b etwee n the fue l- flow speed paramete r and th e presSt t re rat i o. A co ntr o l that schedule s fuel flow, following a l in ear r elat i on w i ll, th erefore , app r oximately pro- ¢ vide the correct stead y -state fuel flo w to th e en gi ne over the complete ran g e of ro- tati o t m l speed, al t it ude, and flight Math n u mber. S u ch a f t t e l-flo W sc h edule c an then be m odulated by a speed error s i gna l from a spee d g over nor to co n tro l th e speed of th e e ng i n e. The e ng i n e m ay b e kept bel bw the s u r ge line durit t g an accel e r o a lie n an d avoid blo wo ut duri ng a dece le r a tio n b y keepin _ th e fu {_l - S p _ ed pa r a m e t e l .

between th e limits shown. . , , ' Plotted in f igure VII-iS is a typical acce l e r a ti on fu e l-fl ow schedul e o f a e urr _nt •' Jet engi n e. Here th e u n corrected quantity , fuel fl o w di vi des by pre s sure , is used.

" The li m iti ng fuel flow is , th ere t o re , a function o t th /' e e vari _ tbl e s: spe ed , tef r / p e ra- lu re , a n d p ressure. Fr o m th e co m pleXity of this li m it scllt _ ldle , it m ay be ap p re- ciated that the simple linear limit of the hydrt Jm e _ hanical fu el co n ti.ol should of f e r som e importa n t simplifications in constr u ction. 4 u-- i Q \ T he t e c hn iq ues w hi c h were u se d to inco rpo r at e t hese p rin c i p les in t o a fun ct io n al fuel c on tr ol dev i ce a r _ sh o w n in fi g ure VII - 1 4 . A s ma y h e s e en , a sma ll s p ee d se n sing po s itiv e di s plac e m e nt gear pu m p is driv e n by the eng in e . T he pr essu r _ dif - f e r e n ce acr oss thi _ pu m p is k ept at zer o by a pre ss ur e r egu l a t i n g valv e whi c h by - ' .

p _ t ss e s f lo w fr o m th e m ai n fue l pum p. Sin ce n a pres su r e r is e o c c ur s a c r oss the ' sma ll p um p , it s ou tput f lo w r a t e i s d i rectl y pr o port ional t o th e eng in e s p _e d. Th is flow qu a ntity i s th e , p ass ed th ro u gh a flx e d _a r ea o r ifice to g e n e r a t e a pr e s su r e s l g - nal, w hich i s d i r e ct l y pr o portiona l to th e _ua ro of e ng i n e s peod. Th e ac_al s pee d " ; may then be co m p are d with t he pi lo t 's s peed c omma nd s ett i ng to ge n erate a s pee d error s ig na l . Th e a re a of a n orific e tha t byp ass e s th e speed s en s ing pum p i s the n controlled by the compressor inlet a nd di s ch a rg e pressures. Wi th th is circ ui t the i total fuel flow delivered to the e _ i n e at any s p e ed a nd pressure is t h en a linear func - tion c , _ speed an d pre s sur e ratio as requir e d for the fuel flow schedules previou s ly .... i_ discussed. T i: is _ z.i / ice area ma y also b e modulated by th e sp e ed error si gn al to ' control or govern the engine speed. Thus , when th e throttle i s adva nc ed , th e con - tr ol provides a d ditional fuel flow to increase the e ng in _ s pe ed. During such speed tr an s ients , however, th e cha ng es in fuel flow ar e limited by the acceleration and de ce leratio n Schedules previou s ly . _ is cu ss e d. In an a c tu _ l control , the variable i area orifice would be provided by spoo l valves, i_ _ This simple hydraulic circuit provides all th e req u ir e d functions of the Jet en- _ ' gi n e co n trol. The working parts required to construct s uch a control are s h ow n : i_ in figure V H -15: In th e figure are s hown th e small gear pump wh i ch provides the !_ speed s ig nal , th e fuel bypass valve w h ich co n tr o ls the zer o pre s sure gra d ie nt, a n d _ ; _ th e valve which adjusts the flow accordit _ g to th e speed error s ig n al. I n addition , i i * _ th ere a re two valves operated by compressor inlet a n d dis ch arg e pressures which ' pro vi de the controlled area o rifice for th_ fuel schedule.

For co m parison , br i efly look at fig u r e VII - 16 which sh ow s th e parts required !_ .

for th e fuel and spe ed controller of a c u rrent turbojet engi n e. The zero gradient t pu m p control has a st l bstant l ally smaller number of workin g parts. It is al s o slm - pler for assembly and adjus tm ent, and it is di r ect acting (i. e . , it does not require serve-actuators).

This co n trol has been extensively a t talyzed, an d its op e ra t ion has been si m u - ulated on th e analog computer. It has b ee n built and has perf o r m ed su c cessfully in actual oper a tion on a J85 e n gine. A typical op e rating seq u ence during a th _ t J t t le burst from 50 to 99 per c ent is shown in figure VII-1 7 ." Here, the sequence is initi- ated by a s u d d en ch ang e in throttle position. The initial fuel flow i n c 1 ' e a s es as the gove r nor call s for more ft l el. Then there is a fu r ther in crease f ollo wing th e surge limit schedule. This prod u ces a sfnooth acceleration of the engin e in a v er y s h o _ response ti m e. F in ally , the governor cuts back t h e fuel flow as the speed se t pbi n t i s a ppro a ch e d and t h e s pe e d level s ou t at the se t p o i n t With no ov e r sho ot or os c l ll a_ ti o n.

In a dditi on t o the w o rk o n t h e fue l co n trol , a nu m ber of othe r ac 0 O s s or y a r e a s h a ve boo n co ns id e r e d, T hey are (I ) Elec t ri c s t a rt e r g e ne r a tor s ( 2 ) Hydraulic m o tor.pump _ (3) F a n drive ge a rin g de sign ( 4 ) Fan bl oc ka g e t h ru s t control (5) Eat_ta_m _unt s a nd a cc e ssor y installat i on The mo re i m l x_ r _ at res u lt s a re s u mm a ri z e d in the fo l lowing paragraph s , The s_ rt u p t ec hniq u e f o rt he o n_ ;i n o h a s boonext ens iv ely c ons ide r ed. AR or e val u a ti on o f n ltm o rous p ossibl e s t _r t u p t Ochnlques , it h asb oo n co n c lu d e d that , sin ce a g e n er a tor a nd a b attery wi ll be ne eded fo r fl ig ht , t h e overa ll best e c on omy wil l b e a ch ieved b y usi n g t h es e s am e co m po n e n t s for e ng i n 0 cranki ng , Si n ce the cost of starter - ge n e r _ ttor sr i ses very rapidly wi th powe r output , c r an ki n g must b e achieved wit h th esma l lest un it p o s sibl e. Thist yl_ o f s t ar t er -g e n er at or l_ s , t heref o r e, b ee n test e d to v e ri f y i tsca l_tb i lit les i n b o thmodes of operation a n dto in v e st i gat e p os si b l e me cl _ m l ca l s im p lif ic ations that ma y pe rm i t cos tr e du c t ion . Th e co n c lu sio n wa s r eache d t hat the 1 5 0 - ampe r es ta rt er -g e n e rat o r , operatin g at_.4 v ol ts , s hou ld be ad e q u ate to c rankth e 1 00 0-pound- thr ust eng i n e.

The fron tal area of access ori e s has prov e n t o be a s u bstan t ial pe n alty for th e 8 - to1 0- i n c h-d la _n ete r e ngin es herec ons id er ed. T he acce ss ory p ow er takeo f f shafting told g e aring i s a lso e xp e nsiv e and lim i ts th ea llo w abl e d esig n c onfi g u r atl o ns. A hydra u l ic d r iv es ys tem u s ingpositiv e di splac e me nt gear p um ps a nd m o to rs h a s , the r ef o r e , b ee n desig n ed a nd is cur r e nti y b e i ng t e s t ed .T h iss ys t e m w o ul d a ll bw t r em ot e pl aceme nt ofacce s s ori e s.

A desig n s t ud y h as a l s ob ee n u nd erta k e n on t he f a n drives l_ e dg ear i n g .The t r e sult o f t h is st u dy was fav or a bl e on bot ha t ec hni cal a nd a co stb a sis. Th i s fan d ri v e sys t em was show n in f i g u r e VII - 3 a nd previo us ly d i scussed. The g e aring sys- .

te r n selected c on si s ted of three para l lel red u ct i on gear shafts e ach havi ng t Wo m es h es , g i vi ng an overal l speed ratio ra nge of 28 to 15. This g e ar i ng system h as the advant a ge o f a coax i al o u tp u t shaft a n d o f f it tin g a sma ll S p l ice. It also has low gear too th l o r d s andbear in g l oadss u ch tha t v _ r y goo d l if e ca n be obtaine t W i _ sm a l l , lo w - c os t b ea ring s and gea r s . B oth a lowf a bri ca tion c ost a nd a longl ife , • . t h erefo r e , app e ar possible f o r th is desi gn .

T he g eare d -fan e ng ine req u i r e s that a m e a n s be i n corporated f or r edttci Ng t h e fan torque at idl e spe e ds to p re ve n t e xcessi vel y hi g h turbi n e t ni et te m per a ktr _ $.

,t9 '"

S

CONCLUOING COMMENTS

Th e f or eg o in g ha s bri _ff ly s umm at ' ized a nd descr i bed the maj or aspects of t hi s lo w - cost - e n g i n e p r ogra m . At this tim e _ major results a nd conclusio n s ca nn ot b e li m ade b e cause t h e pr o gra m i s a t an i n ter m e di at e p oin t, i!

For th e f u ture D i t is planned to co n ti n ue the f a br ic atio n d e ve l op me nt progra m on I she et - me t a l a xi _ tl st a ges an d t h e control deve l op m ent an d i t s o .pp ll catio n to th e turbo - i' j e t a nd the f anj et engines. D e si gn work of t th e f an_ et engi n e will continue a nd a lso th e fi n a l d e sig n w ill be _ om p leted on the Navy Ordnance engi ne . Fabricat i on of proto - type Or d nan ce engi n es w ill th e n begin. This e ngi n e w il l be b u ilt S o as to simula t e tt t" productio n engi n e tully, and it will b e tes t ed at i t s full design oper a ting condfl i o ns . , I n c onn ect i o n w ith ou r overa l l interest i n low-cost aircra f t eng i n es , i t is i n - s tr u c ti ve to e x amin e the e s ti ma t es that h ave been made o n t h e productio n prices o f th e co m po nen t s of this e n git xe . T h ese a re Show t L i n table VII - 1 and ar e based on product i on rat _ s o f 2000 u M ts Per yea r where t h e tooli ng w rtt e ott costs ma y b e t n eg l ecte d . _ t_ It is no t th e p u r po se of th is paper to di scu s s these prices comp l etely. T h e t o tal cost of J u st over $3000 fo r th e 650 - poun d-thr u st e ngi ne , cle a rly ir i tlieat e s, h owever, that w e ca n expect to provide th e a d v a ntag( _S Of turbo je t propulsio n to m issiles a n d d r ones a t a price th ai is _ o m p e ittiv e with any ot h er form o f pro pulsio n . $ l m llt i r price estim a tes _ vere pr ev iousl y m ade for the 10 0 0 . pound - th r ust leve l t ttr b o t ets , which ittdi - c a ted t h at m a m ffact u i, ing costs of abottt $ 50 0 0 to $ 0 000 s h o _ l ld b e attain e d f o r pro - d u cti o n qua n tities o f 2 000 per ye;tr. T h e pro flu ctto i t c ost level of $5 . 00 per i _ o u nd of _ !

t h r u st has , th er e fore , bee n i n di cat ed toe both th rust ieW ls. We a re not co t t e k i ding that $5.00 per po und of thru s t w ottld be a fl im l m anu fa ctttr ei - t s s e |itngpr ic e sitice th er e are a number of i n dl _ ' ect costs whic h wo ul d aff ect th e s e l li ng p H _ e. For e g- a m ple , the costs fo r calibratio n runs , the wr i teoff of qti a lifi e / _ tio n e)tpenses , saMs, and ]

!

......... • - : .... . . ..

TABL E VII-I. - PRELIMINARY COST ESTIMATE S

O F S AW

........ _ ' _m _ .................................. O o s t d o l hr _ I i A x i al ¢ o m p re_a o r a nd stal er 68 0 T u r b ine rotor a nti s t a tor 2 20 Front m_ ro a r be a r in g su pport s 16 5 C o m p r e ss or h o us i n g t _ s s o m bly 55 Co m bu s tor a nd h o u sing assembly 1 85 Compressor s h aft 40 Fuel m a nifold anti nozzles 55 B e a r i ng s , springs , bo l ts , etc. 50 Fuel control, pu m p, f i lter , etc. 88 5 Ignition system an d i gn i ters 175 Subto ta l purch a s e d parts 1960 Cost of as s e m bly an d inspection 250 Ma teri a l h andlin g burden 200 Manuf a ctur e r ' S gen e ral an d administra - tive expense and profit 7_ _ _ Total cost assembled 31 6 5 , .

field engineering , and t h e distribtttor * s m ark u p. I t is pozJstble that these ite ms co u ld , a c tu al ly doub l e the pr i ce at w h i c h th e e ng in es w ou l d b e f i nally so ld to th e user. _ !

From the esti m ates made i n th e prog r a m , howev e r , It I S evid e nt th at gas tur- i ¢ bine eng i nes o f the type h ere considered wi ll be attractive and cost co mpe titi v e for ... general av ia ti o n, for missiles and drones, for addi ti onal services uses ( _ ch as reconnaisa n ce ai r plane S ) , a n d po s sibly f o r th e s m aller business category aircra f t.

If th e obstacle of h igh c o st can be eli mina ted , gas turbine engin e s w i i l m ake major perfo r mance i t np r ov _m e nts a vai la ble for the se pu rpos e s.

22i LOW COST EN(: _N ESPECIFIC FUEL CONSUHPTION 450 - M PH C RUISE AT25000 FT 1 . 3 .TURBINE INLET TEMP • 13000 F FANSTAGE . .- '-TURBOJET BYPASS PRESSURE e _ DESIGN RATIO RATIO '; :

\P O l_ ; , -o 1 o

1.1 - _ , / '1.25 1.6 ;

S P EC IFI C FuEL _ "--U ! , - _i O - 3

CONS UM PTION , 1 .2 ( L B E IR) / LBF F _ / ' .7 i'i: • 2 4 6 8 10 12 i; : cs - s 676z OVERALL PRESSURE RATIO J° Figure V II - 1 i LOW COST ENGINESPECIFICTH R UST '_ 4 50 - MPH CRUISE AT25000F T TURBINE INLE T TEMP = 1 3 000 F "_ FANSTAGE 60- FANJE'T BYPASS PRESSURE DESIGN POINT 7 RATIO RATIO e " - " "_""_ _...... _" ' _ - ... _ 1 . 20 ,, THRUST , LB / LB I SEC 50 - SPEC IFIC _ . , . _ 1.30 , C O RE . 1.

. , AIRF L OW = e ._ ' _ ' _ ' _ ' _ ' _._ I. 0 2 4 6 8 10 I2 c s.s G_ Gz OVERALL COMPRESSOR PR E SSURE RATIO FigureVII-2 _._. - ................. _ '1 I I I ll nl " " i • ' , ............ | .............. ;-- _" .... ,. , ......... J , , _ G EARED FAN dET CO N FIGURATIO N i 2 ' i , i' .... i ' i Figure VII " 3 cs - s _T z _ ' ii 650 LB STATIC SEA LEVEL THRUST _i, : _ , NAVY ORDNANCE ENGINE d t 4 Figure V I I-4 c s.s 675 9 Q \ SHEET METAL COMPRESSOR ROTOR COMPONENTS 4 B i L, : K . .: I i' I.

i_:<:! l;i t _.34 _ -- .... II f" - F "|11 r " "_'TH ..... i * _" ..... i ...... _ I ' 2= 5

t

tl

STRAIN HEAS UREHENT$

104-- // tO - IN.DIAMSHEET M ET AL COMPRE S SOR

;/

_ F DRUM 0 ' MICROSTRAI N , 1 0 3 pin. / in. " D I d .. !

i ,,, -- D ESIGN SPEED , 28000RPM _ • _ j ' _' [ .... : , , ' - YIELD P OI N T, 4200 0 RPM {' i - ' ' ! " (ESTIMATED) "

1 i

1024 6 8 10 2 0 40 60 =l_ ) x 1 _ RPM cs . s 3 vz s =i Figu reVI I - 7 _ t _2 6 d In r .......

SH E ET METAL TURBINE CONSTRUCTION ' i :i! c . 7o . 34 7 4 CS-56821 Figure VII- 8

i

!

SHEET METAL BLADED FAN ROTOR_ il , .,, F igu r e VII- 9 \ + .

LOW COST ENGINE COMBUSTOR + . ',:• i+ f i i. +i . !

i + , + _ , . : ', _i=_ . = m ';;: / + " c . 69._98o _ i C S - 5 6 7 5 4 !

..

Figure VII- I O !: r_ If ; LOW COST ENGINE !_ i ANNULAR COHBUSTOR TEST RESULTS ii.

t U SO JEr SLT O

1. 0 - l i : ' r ........ _ li ' • 6 / ,_ , _-T URBOFAN 25 0 00 - FT CRUISE i! COMBUS T ION

EFF / '-

o ,,.! I L , I,+ I '

, .004 . 006 .008 . O l O .012 .014 FUEL-AIR RATIO c s - s 67_ F i gure VII- ]l _ .... L_ . . !_ . : - ......... r, I I n 11 F_ III .... i" T *i ..... i m lll i i , .- "_ "_ "I .... • ..............

. Q ', ' .... FUEL CONTROLACCELERATIONLIMITS " "_ TY P I C A L TU R BOJET ENGINE 40- SURGE AND / OR . 3 0-

N

_ . __ u_ IuMIT 0 1 2 ] 4 5 6 cs - s 67_ C O MPRESSO R PR E SS U RERATIO , Ps I P2 F lg u r e VII - 1 2 CURRENTJ ET E N GINEACCELERATION SCHEDULE 40 - ' ' i!

# • ._ 3 2 - Ps3 "' (i P " I I I l I I I I . I h J , 10 20 3 0 4 0 50 60 7 0 80 90 Z OOI i 0 , _ N cs - s 37 z 3 F i g u r e VII - I] _ag ZERO GRADIENTPU M P FUEL CONTRO L SENSING PUMP . , SPEED SENSIN G ORIFICE FROM "4 ' "* PUMP L q BYPASS ORIFICE _ RETURN CONTROLL E D BY _ ; PRESSURES ANDSP E ED CS-56770 Flgure VII - 14 )i FUEL CONTROLCOMPONENTPARTS Figure Vli - _ 23 0 _ ' . . • _-Jt ,_ . I Ira _. IIIII I ........................................................ , i I

EXPLODED VI E WS

" CURRENT TURBO J ET F UEL - CONTROL

..,: : _

!,!

¢i +•

" J 85 E N GINE A CCELER A TIO N 50 TO 95 %

_ PERIM E NTAL FUEL CON T ROL , , SPEED " i _ I n I ."! I

, I I

.. FUEL FL OW

I I

• I i HI ..... • __ -- L , _ , ,,, THROI"I'LE POSITION

.... I I _ _

TIME , SE C C S- 56 ? 65 Fi gureVII -l / t

'. "1 9 4 59

VIII. SUPERSONICEXHAUST N O Z ZLES ii

. Milton A. Behelm, B ernhard H . Anderson, John S . Clark , . :

. . i Blak e W . Corson. Jr . , , Leonard E.Stltt , andFred A . W q cox _

A m a j or pro b le m in the de s ign of _ n exh a ust s yste m for a super _ ontc a irphmo is _, , tl_ t l t _ geometr i cal s ha p e shou ld change as fl ig ht condit i ons at0 changed. T he so ni c area and expansion zatl o must be variab l e , and it m ig h t also be required to reverse thrust a n d to suppress Jet noise or i _ rared r a di a tion. The mech _ S m s t l mt a re needed to do this can be c omp li cated mi d heavy. T h erefore, t h ere i s a trade b _ tween we i ght and perfo r man c e t ll at depends on the miss i on.

The missions for supezsonic a ir craft can be d i vided I nto the fo l lowin g types: (1) Supersonic cruise , such as the B- _ 0 _ SR- _ I, and the supersonic transport ( 2 ) Supersonic dash, suc h as the B-fiB, F-4, F-111, B-l, F-14, and F-15 ( _ ome of wl _ h are a i rc r aft of the f u ture) The supersonic dash airc r aft f ly lon[ _ distances subSo l dca U y but are also abl e to go supersonic for relative l y short distances. Exhaust nozz l e concepts fo e both types of k a i rcraft are considered in this secUon . Supersoolc cruise nozzles are d i scussed • first, and the n s o me of the supers o n i c dash pr o bl ems al.e described. • , ' " • " i

S UPERSONIC CRUISE A IRCR A FT " i

, .:: : : The first step in , analyzin g the exhaust system is to get some ind i cation of the i ; sensiti vi ty of a m i s si on to t_ design. Some results of an analysis for a sup e rsonic !

, cru i se airplane are show n i n f i gure VIII-1 . The ah'pl l tne t _ assu _ ned to have a take- i l off g ross weight of 7B 0 000 pound _ and a p _ ty l oad 6 _ percent of the _ eoff w_ i _ ht, l : / " O n the l eft side of the t igttre, the changes in rang _ _ or a 1-percent change in nozzle gross thrUSt coeffic i ent at ct _ ise and lo t te l' are conip _ red to a 1 -perc e nt change in i nozzle welght _ for a cruis e Speed of M i lch 2. 7 for 39 3 0 nauti c al miles. Getting eno ug h range out of a st _ personic _ uis o a l rp l_ m _ has always be e n a fund _ tn l entfl| prob l em, and it is e ven more crR ical _ or coiu _ fl e t _ lal operatl _ m . Sidce _ ttie _ are . ' .". fixed dtst a tnces apart and since runways will n evez' I _ flmre t ho rn _ mtleslong, rmtg _ ' in c rements have to be W atched c l osely. For this mis _ ion _ t h e cru i se nozzle efil- • NA SA - Langley Research Center.

PRECI_ING PAGE P J , _.NK N O T FILMr : ._ ' 2_S , t ,, , ci e n cy aff ec t s ra n ge by 3 _ per c e n t and I s quite Im port an t. In f a ct , a l o per cen t ga in her o I s at lea s t thr e e ti mes a, e f f e c tiv e a s a 1 - p _ rc e nt gain In p e rfor ma nc e o f a ny oth _ r c om pon e nt o f the propu ls ion s y s te m . On the right sid e o f f igure VIII-I It is pr es u m ed t ha t thi s sam e a irp la ne flys an a l l - _ ub _o n i c mission a t a crui se _ pe e d of M u c h 0.9 f or 3 2 80 r _ t U t innl mi les . A l-p e r ce nt ch a nge in _ ub _o nlc cru ise thru s t co- e f ficient a ffect s r a n ge b y about 2 per c ent, a nd th e _ cn _i t i v i ty t o l o i ter thru s t is th e i s am e a_ i t w as be f o re . F o r the _ e ass u m pt i on _ , the n , it c _ n be worth a lot of noz z le weight to keep por f or _ mcc h ig h a t a i r _ e ods .

S om e o _ha u s t no zzl e s which co u ld b e u_ ed on t h i s a i rpla l w a re _ hown In fi gure VIII- 2 . On the l e f t t _ a v a r ia ble £htp ojo c tor _ w hich Is t ile ty p e u s ed o n the J-93 en- g in e. For a turbojet e n gine the sonic a re a is v a r i ed b y about 40 percent f or af t er.

burner operation , a nd s o m e kind o f iris m _ c han i sm is required. For flight at Ma _ h n u mbers near 2. ' / t h e expan s io ll ratio f ro m t h e sonic area to th _ exit area is about 3 . 6 , but a t subso, _ lc speeds i t m ust be decreased to a va l ue near 1 . There f ore , the i divergent shroud i s made up of several cver la pp in g _ laps and seals s o that it can b e i'_ closed down at subson i c speeds. Th e s _ cond n ozz _ is like tlmt used on th e J-58 and TF-30 engines. At ltlgh sp ee ds its appearance i s similar to that of t h e va ri able flap i ejector a n d s econdary _ low is still used to coot the hot parts. For low - spee d op e ra- i_!

tion the divergent shroud also uses multiple f la ps , but i t d o es not clos e as far in i order t o simplify t h e flap m eclmais m . Auxil i ary inlets a re then o pened u p in th _ r s _ condary flow passag e a t these low speeds _ bring in additi o na l air to help f i ll up i : _ ' . the ex it area. The da s hed lines indica t e the p o sition o f th e doors when th e inlets a re il opened. The minimum di a m et e r inside t h e shroud is larger than i t is o n th e varia b le i_ flap ejector so t ha t this inl e t air c _ n _ et thro ug h. The c o nical plug shown on the !_, right is a more r e cent idea th a n th e diver g ent ej e ctor s a nd i t Ires not yet b een used on a productio n engine. The son i c area could b e varied e ither w i th a n iris primary ii f la p or by an a x ial translation of th e flap relative to the plug. At high speeds th e in- I! t ternal expan s ion occurs in th e a m m la r flow pa s sage between the plug and the cy li n - drical shroud. For low- s peed operation , the exit ar e a could be decreased b y tra _ - lut in g the cyl in drica l shroud upstream. For subsonic flight, then , the shroud would be completely retracted so that t h e primary flap is expos e d to the externa l flow.

This kind of nozzle i s hard e r to coo l t h an the divergent ej ec tors, but if that p I'oble m can be solved, it has some adv a ntag e s. Fot' on e thi ng , it w o uld not le _ k a s m uc h since th e l ength of seal betwe e n the movabl e surthc e s could be decreased fro m about , , 30 0 feet to 30 feet. In ad di tion , the m e c han t snt s appear si _ npler and m igh t b e more durable. So m e j et-noi s e tests also indicat e that I t is inherently a littl _ quieter tl u tn , the other no z zle s (r e f. 1).

i '

Supersonic CruisePerformance

Since the perf or m ance of a high- s peed nozzle i s s o i m p o rtant , sm al l d ifferences affect the opti mi zation of its design. In i ts develop m ent , a lot of configu r ations must be tested but it i S hard to m easure the experi m ental per f or m ance with enough accu- racy. F o rtunately, some m ethods of.analy sis have be e n improved so tha t th e y a r e becoming quite help ful .

T h e ejector flow model i s based o n th e in vis cld and vi s cid i n t e r action betw e en a hig h- e nergy stream (primary flow) and a low-ener gy s tream (secondary f l ow) (f ig .

VHI-3). Thes e tw o streams begi n to interact at the primary nozzle 11p. For t h e ejector operating in the s upersonic regi m e , the s econdary flow i s effectivel y " sealed off " from ambient c o nditions. When this occurs , the ejector mass -flow characteris- tics become independent o f the ambient static pre s sure. R is thi s e j ect o r operat in g co n dition that is consider e d in the theoretical analy s is. The flow regi m e s o ccurr in g within the ejector syst e m can be cate g orized on the basis of t h e p r edo m inant flow m echani s ms.

When the a m ount of s eco nda ry f lo w supplied to the e j ector i S small, the p_rin _a ry flow plum e s out and impinges on L h e shroud wall (botto m of fig. VHI-3). This causes an obli q ue shock to for m which eff e ctively "s e als off " the se c ondary f low f ro m a m - bient condi ti ons. The se c ondary flow is " dra g g ed" through th e obll _ / u e s h ock by it s mixing wlth the higher veloc i ty pri mary j et f l ow . H the s eco n dary f low is incr ea se d , t he s e condary pressure incr eas e s and th i s " pushes " the primary Jet away from the s hroud wall. The obli q ue Shock can no longer be sus t ain e d at th e s hroud wail , an d thus the s econdary flow a ccelerate s and chokes within the shroud (top of fig. VHi-3).

The aerody na mic pheno m ena t ha t det e rmine equilib ri u m condi ti o ns at " low " s e con da ry ejector flows are the s a me as tho s e t ha t dete r mine the base pressure be- hind a backward-facin g step. For " zero " secon dary flow , the mass flow entrained t by the mixi n g proc e ss m ust be e q ual to the mass flow reversed by the pressure r is e through the recompression zone ( o blique shock). This condi tio n i s satisfied when th e total pr e ssure on the d ividi ng strea mli n e in the mixing zone e quals the reco m pre s - sion static-pre s sure rise. Within this bas e flow concept, the flow which ' tl eaks o u t " past the r e compress i on zone is t ha t amount of fluid which has a total pr e ssure gr e a t- er than the pres s urc rise through the reco m pres s ion zone. Eq uili briu m con di tions are th us e stablished in the " low " ejector flow regi m e when the a m o unt of seco n dary f lo w su pplied to the ejector is equal to th e fluid which " leaks " past the recompres- sion zone associated with th e oblique shock.

For "hi gh " ejector flows , the interaction between the two st reams is such t ha t the secondary flow accelerates to a critical condition so m e w her e d o wnstr e a m of the primary nozzle lip. The viscous interaction be tw een the two streams occurs along

J

235 i the interface (dashed line i n fig. Vl _ I-3). Th e effect o f mix i ng resu l ts (1) i n a trans- fer of en ergy (Shear Work) from the primary jet flow to t h e secondary stream, a n d (2) i n a m o d i f i cat ion o f t h e p u mp in g _ harac t er i stic s d u e t o th e di s p l aceme n t t h ic k- he ss of the mi xing zone.

In figure VIII - 4 , the two effects due to m i _ ng a r e evaluat e d f or an ejecto r w i t h a large s econdary sh o u lde r dia m ete r compar ed to the pr im ary n ozzle e xi t di a m eter.

The i nv is ci d s olut i o n ( S olid curve) is obtai n ed if mixi ng betwee n th e two strea ms is neglected. The prima ry flo w field i s de t ermin e d by the me th od of charact e rist i cs , whereas the s econdary flow is assu m ed to b e on e di m ensional an d reversible adia- batic. Two con diti ons were appli e d along the i nterface boundary. (1) the local static , pressure m ust be equal for bo th str e ams at their boun da ry , and (2) continuity be- _ twee n the st r e am s must be preserved. _ he effectS Of mixi ng alo n g the i n te rfa ce !:'_ : boundary hav e been evaluated by locally su perimposing the mixin g region on the _ established inviscid flow f i e ld solution (solid curve) at th e critical secondary f lo w J area. Th e as su mption is tha _ mixing takes l _ ace as thoug h the interface were a i_ constant pre s sur e boundary (ref. 2 ). The reSult S of such a mixing C orrection are '_ repre s ented by the upper da shed-line solution i n fi gu re VIIL-_4. Inherent in this type i_ of mixing corre c tion is th e assu m p ti on tha t th e disp la ce m en t effec t s due to mixing l _ tve a negligible e ffect on pumping c ha racteri s tic s . When tht _ k in d of-correction is if " I used , the effects of mixing are viewed as an in creasei in th e secon da ry corrected _ we i ght-f lo w ratio over t ha t gi v e n by t he in v ts cid solution. Continuity is th u s p r e- t_ S e rv ed by increas in g th e initial secon dar y _ vei gh t-f lo w r a tio by th e am ount th at was J entrained as a result o _ mixin g , an d by as suming the two flo w fields a re no t appre- _ ciably chang e d by mix in g, i I n order to accoun t f or the change in shape of th e p ri mary jet bo un dary due t o i_ mixing, th e mixing co rr ectio n must be applied at each po in t a lo n g th e interface !_ boun dary . The results of this type of mixi n g sol uti on are S hoWn a s th e lower da s hed t line for low flo w rat es and dash-do t li ne for t h e hi gh-f lo_ v-r _ tte so lution. Fo r these solutions, contin ui ty w a s applied along the i nter _ ac e boundary by requiring that the sum of the tnviscid weight-flow ratio plus tile m ix in g comp o nent be the same as th e secon da ry weight-floW r a ti o supp li ed to th e ejector. This reSulted in a much la rger effect due to m ixing than was o r iginally i aicuhtted, t n gene ra l, t hese sohlUons agree quite well wi th th b d a ta i n di cated by the circ ula r $_ mbols.

This analys J_ is particu la rly use f u l in tryi ng t o find th e best shape of t h_ div e r- gent shroud. Fo r the auxiliary inlet eject or th e nflrdnitt m dis _ eter at t he shoulder must be re la tively large to accommo da te the auxiliary air at of_ -design speeds.

The position of th is shoulder downstrea m of t he pri ma r y exit m t l st then be picked to ensure high nozzle performance.

Figure VIII-5 shows calculations which help in making this choice. The nozzle r - w i z t h rust coeffic i ent i s s ho wn as.a fun ct i o n o f th e Sp ac ing ratlo _ wh ic h Is de fin ed as th e d is tance b et w ee n th e shou l der and the pri m a _ y e x it divided by t h e pr i m ary ex / t diam- ete r . W ith a relative l y sh a rp c o n tou r at th e sho u lder , a n abr up t lo s s i n thru s t o c - Curs if-th e sp_a _ lng is to o low , u s how n b y the lower curve. _ h e sho u ld e r iS more r ounded, the ejector is less senSit / ve to spacing; and hig her p e rformance re sults , as i ndicated by th e uppe r cu r ve. In most d e s 4 gns th e po s ition o f this shoulder i s f ixed I , / but the f lap s m ust be moved t o vary the divergent S ltroud exit al'ea. The s h ape o f t hese fla ps is an o th er de s ign va riable. For the sharp-shoulder pe rf or man ce curv e , I t was assu m ed that the f laps llad _ tn ls e ntropic contour. H th e f lap s we r e chang e d to a conic shape, the daSh-dot curve r e sulted, and it ma y be a better choice. Some , data points are shown o n the figure to indicate that th e theor e ti c al trends actually occur. T he circle i s for an t S entr o pi e coflt a ured shr o ud, while the s quare is for a conic f lap Confi gu ration.

The S e curveS ar e shown for th e aux i l i ary inlet ej e ct o r. A simi la r study has been made for t he variable flap ejec t or and th e r e sttlts show an even greater sensi- ti v ity to shroud geo m etry than i s shown here.

All-these eurve _ are sho w n for a 2-percent corr ec ted secondary flow ratio.

However, th es e re s ults may be sen s itive to th is flow r a te. In figure VIII-6, th e sec o ndary flow is va ried. TheOretical re s ul ts a r e S hown f or the geometry w ith a s harp sho ulder and con to ured flap s . The 9 -p e rc _ mt flow c u rve is repeated from th e previous figure. H ow eVer, for this pa _ , U cular fi gu re, the ram drag of the Sec ondary flow has been subtracted from th e gro ss thtwst. Curves for secondary flows of 4 - and 8-perc e nt a r e also sho wn. Alt ho ugh higher pe r fo r mance is r each ed at 4*percent cor r ected w eight-floW ra t i o , it rotluires a Stu dy o f the ov e r all {i e si _ n o f the p r opul- s i on syste m to dec i de whether th e s e high e r flow s s hould b e us ed . The e xp e ri m ent a l ; data point a again s h ow good col'relation with t he th eory.

A sum m a ry o f sup e_ sonic cruise nozzle p e rfo r mance t s s hown i n f igu r e VIII- _ / . t The gross thru s t coefflc t en [ i s show / z for e ach of the thre e noz z les a t 9 -p e rcent cor- te e :t ed se conda ry fl ow . The to p of t h e bar is the theoretical m_m u m peHorn _ tnce which co u ld be obtai ned with an optin U z ed des ig n. The pe r forma n ce of _ v _ri able f lap and the attx / ltary inlet ejectors were tak e n fro m curv e s s imilar to t ho se of fig- ure VIII-5. Pfelifltin _ y cale ul tttion o n the p l_ g n ozzle indi c ate that its p er f 0rmanCe could be as high as t ha t s how n for t he Va / ' ia ble flap ejector. T he aux / lia ry in let ejec to r is less b ec ause of it s r _ l ul renietlt fo _ a larger _ econdary diamete r . The best experimental results obtai n ed to da[a a r e i ndi ca t ed by th e da s h ed line B and ar e qu ite close to theory. _ ,

Subsonic Cruise P e rformance

The off - design perfo r manc e o f a nozzle i s hard er to analyze because of the in- teract i on S betwee n the internal and e xtern a l flows. Therefore, w e have to depend i i more on experimental data. The isolated nozzle model _ how n in f igure VI H - 8 i S : used for wind tunnel tests at this Center. A 21.6 - cen t i m e ter-dia m eter c y li nder is supported by a s t rut, and h ig h - pr essur e air is du st ed to the test nozzle to s imulate !

the engi n e exhaust. In the 8 - b y e - Foot Supe r so ai_ Wind Tu nne l, tests are fl l ade fro m Mach 0 to 2. However, the transo n ic pe r forma n ce of a nozzle is particu la rly ha rd to get because of tuim e l wall in t erfere nc e. A t transonic spe e ds the m os t i ra - { portan t airf r ame instal la tion effects occur. Th es e effect s r esult beca us e the exter- i I ' ' h al fl o w is dis to rted by the ai r frame, and it vari es d e pe n ding on the e ngine location, o So the t ransonic testing problem is m ade e ven more difficult since a bi g piece of the airframe mu st be test e d along W ith the e x _U st no zz le. As a r _ t l lt, th e nozzle tt m o del e n ds up b ei n g a lot smaller than w e would like whe n worki ng within th e size I limitS o f ou r prese n t wi n d tunnels. O n e appr o ach to t hi s problem that w ould h elp is Ii to have a coordinated fli g ht an d wind tunnel model p ro gra m that uses the best fen - ! : tureS o f each test techni q ue. An effo rt like t his iS in progress a t the Lewi s R es e _ arch _ C e nter using a modified F - 106 as i llust ra ted in figure VIII - 8. A new e ngi n e nacelle /i has been added under each wi ng s o that the no z zle sticks out be hi nd the Wing. This I_ kind of engine ins ta lla ti on has ho t been us ed b ef o r e , bu t it has an importan t advan - • rage since the wing car l be u sed to shi e ld the inlet fro m an gle-of - attack effects. An • aft _ rburning J-85 turboje t engine was used i n each pod. These po ds were 60.8 centi - m et e r _ in dia me ter al l d were d esi gned to accep t any of th e nozzles w hi ch ga #e good . results in th e iso la ted tests. For each noZZle design the ai r p la n _ is also flown At 4 _ low altitudes for fly-b y n oise measul'e m el _ ts. I! " In parallel t o the f light tes ts , a wi n d tunnel m odel p r ogram is being conducted L _ an effort to g et more flexibility in the nace l l e s hape and its location. Ah _ u nple 1_ t is the 1 /2 0 - sca l e model F-10 6 s h own in figure VHI-8. It wa _ s m all enot _ h to _ tvoid _ the t r anso ni c w all in terf e rence problems but the haeelle di am e t e r wa s only abotit i , ! , 3.18 centimeters , and the W in _ st ru cturewas s o t hin th a t we could not _ et pr es sur - t_ ized air to th e nacelle to s im ulat e Jet effe c ts . Thel _ d or e , _ nl y ti l e simplest exhaust nozzles could be tested with this m od e l. .. _- L Fli gh ts ha ve bee n ma de w i th t h e F - 106 at Mas h 0.4 f or nois e fly-by m eas ur e- Ii : m erits an d at Mach t iu lnbel.s froin 0.6 to 1. $ f or _ uitist no z zie e va lu _ ttttm. Math _!

nu m ber 0.9 is used i n the ensuillg figu r e _ fo r discu s siotz of subso ni c cru is e. The _ flight test program is described in references 3 to 5. I The instal la tion of the nacelles is _ ,hoWnin f i gur e VIH-9. The n _ celie $ ivere tangent to the wing lower surface at the trailing edge of th e wing and they were at- J I i tach0d by two li n k s . Axial forc e s were measured by a load cell. A S imp l e n o r mal S hoc k i n let w a s adequate for the M ach number ra ng e us e d. A m o vable r o tating valve was l o cated at - th e engine face to control flow o f s e c ondary air t o the nozzl e .

Wit h a n exhaust no zz l e located i n t h e co m bin e d flow f ields o f t h e wi n g and ua - c e lle, flow conditions diffe r from those around t h e m o del of figure VIII-8 that wa s used to e va l uat e is olat e d exhaust n o V zies. The Se fl ow field di ffe rence s a re fi rst p o inted ou t a n d th e n the effect o f i ns ta l ling v arious nozzles i n a n und e rwin _ flow field i S de s cribed. F lo w fi e ld effects are reported i n references 6 to 10.

Installed flow fields. - The variation of s tatic - pre s sure coefficient un der the wing at a spanwtse location near the n acelle location is presented i n f i gure VIII-10.

• Da ta are shown with and wit h out th e n acelle. Without th e nacelle th e value of p r e s - sure coeffi c ient dr o ps al o ng th e wi n g ch o rd a n d then rises to zero at the win g trailin g !

edge. W h en th e na celle iS add _ L to t h e wi n g, the effect of t he combined f le w field is t o raise the pr es sure i n th e region-of th e inlet. The f lo w then ove r expandS a ro und _ the Juncture of the in let and na celle; th e pressure coefficient drops to a.lower value i ' ' t ha n with th e c le an wi n g and th en it rises ba c k to zero. High pressure near the inlet inc r ea se s inlet dr ag. Some of t hi s dr ag may be ca n celle d , ho w e ve r , by pre s sure _. : forces acting o n the surface s of the n 0_ .z le if pr e ssure coefficient downstr e am of the win g tra i ling ed g e reac h es a value ab o ve z e r o . S o me n ozzles ma y b e more effective t h an others in this dr ag c an cellati o n. There is a compress i on shock-pre s e n t in the I_ pressu.re rise nea r t h e wing trailing edg e . T he p o sition o f th is shock varies With I_'_ , fligh t _ ch n umber. _ The mo v ement of the compression s h ock with f lig ht Mac h n umber is shown on i / " f igure VU / - ll . At Mac h 0 .8 ther e is a small a m oun t o _ oV e rexpa n Sion and a gradua l i_ - ; rise in p r essure c oeff icient b ac k to zero n _ r th e win g t l _ i l i ng edge and ab o v e zer _ i .... on th e boattail of the nozzle. At Mach 0.9 t here is more overexpafls i on, f o ll o _ ¢ed by _ ' a sha rp rise in pressure. The co m press i on s ho ck is loca t_ i in t h is steeply ri s in g I ' pressure regio n . A t Mat h 0 . 95 t h ¢ _ s h ock ha s mo ved r e arward n _ ar th e t _ aili ng edg _ of th_ wing and at Mach 1 i t h as m o ved o ff th e end of th e no_ .zl e . P r eSsure rentah _ low along the entire length of the na celle and n o z z le. Low pr e s su r e on the nozzl e boat _tll f or this co n dition re sti lts i n hi g h n ozz le drag. / Th e extez _ nal static-pr eSs ure coef _ ici e n _ wa s tmif0rfl 2 alon g _ le i sola ted n ozzle t est m od e l at a value of about zero for all lt ubso _ tc c o n dit ion S . There Was no c i r- cun d e_ e n tial vari ation in p res sure ar o ur _ d the ,s o lat e d m odel. " _ his _a s also th e case i n f li ght a bo ut one no zz le dia met er a h e a d of t h e Bo aflail ju n cture _ md at the z'ea r of the boattail. _ n the region J US t dow n stream of the wing t r ailii _ g ed ge, h owe v er' , t ' the external pressure was higher aro und the top o f ttle ri o zZle t l _ m at th e I _ otto m .

There were diffe r ence s i n exter na l bo unda ry layer m easur e d upstr _ tfl _ o | t h e nozzle. In fli g h t t h e bo u n d a ry layer was ge n e ra lly t hin ne r tha n on the i s olated niod e l, 23k / !

?

i q i except in th e corners betwe en the n acel l e and w ing an d over the top of the w in g.

T h e re were r e gi o ns of l ow e nergy within the b oun d ary layer a l l around the nacelle as i a result of th e m ore complex flow field and presence of shocks. Th e va r i ou s exhaust nozzles s hown in fl _ re VIII-2 hav e been te s ted i n thi s flow field. I s o la ted n ozzle i tests w e re made prior to the fli g ht tests.

Variable flap ejector. - Th e variable flap ejec t or nozzl e is shown in figure i i Vlll-1 2 in th e subsonic crui s e configuration. T h e nozzle flaps are closed down and ii I provide a low internal area ratio (Ae / A 8 = 1.40) and good inte r na l performance at i this f Ught speed (ref. 11): The bo a ttai._ drag can be a problem since the projected area o f the boattail i s 60 percent o f the r_ t c el le area a nd th e boatt all angle i s 15°. : Boattail d ra g depends on t he static-pressure distribu t ion, w hich is affected by the shape of the boattail. It would b _ desirable fro m a drag standpoint to radius th e cor - ner be tw een the cylindrical, na celle a nd th e 15 0 boatta il angle. H _ weve r , i t is i m- i practical to provide much of a radius at thi s external hi nge poi n t and Still be able to i maintain a cy li ndri c al b oattail S hape at s upersonic c ru ise when th e t r ailing-edge i _ Typical s tatic-pressure di stributions for both a sharp- an d a rounded-corner : f la ps are f _lly op en. li boattailare sho _ n in flgure VHI-I$ ataMach number o f 0.90. The flow ov e rexpands conSiderably downstream of a s harp cor n er (R / D = 0) an d results in a hi g h dr ag . i_ .

Rounding the corner can reduc e both th e i ni tial overexpansion and the drag, as dis- _i cuss ed in re f erence 12. The boattail sta t ic-pressur e distr ibu tion can al _ o be af- i fected by th e jet bolmdary that is located at the trailing edge of the boat ta il. _ his _ Jet bo undary vari es wi th no z zle pressur e rat i o and , therefore , je t effects mu st be i!l_i dup li cat e d w llen m e asuri ng the boattail drag (r e f. 13). I_ The i solat e d pressure drag of a 15° boattail is shown in f i gure VIII-14 at a Mac h _li • number of 0 .90 as a fu, c fl on of no z zl e p r e ss ure ra t i o, d efined a s the to tal pressu r e of t h e p r imary Jet di vided by the ambient static pressure. The bo a ttail pressure drag is rafloed to the ideal gross thrust of the pr ima ry fl ow . The hig h est dr_ tg was , i t obtaln _ with a s h arp-corner ed boattail (R / D = 0) and a thin boun d ary layer , t 3 / pi _ l I of a nacelle inStallation. F o r th e exa m ple show h , the botin #a ry- la yer t hi ckness was _ ' 7 pe r cent of the nacelle diamet er . At _ nozzle pressure ratio of 3.7 5 , typi c al for a , !!

• tu rbo j e t engi n e at su b so ni c cruise, the pressure dr ag of t hi s co ni cal boattail was , about 6 percent of the id ea l gross tlirt _ st of th e n ozzl e . At t hi s flig h t speed the n et li It thrust of the en gi n e i _ about on e - ha lf of th e gross thrust, S o t ha t the boat _ ti | dr ag f would be about 1 2 percent of the ai rpla ne diag. A thicke _ boun da ry layer - , typical I / of a fuselage i n stal _ tio n , re d uces th e d ra g of t h e sharp-cornered bo a ttaiI to 5 per- _ cent of the id eal gross thnl st . F or this e xam ple t he boundary-layer thiclmesS w a s 18 percent of t h e ftise _ ge dia m eter. I t is i t _ teresting that the effect of a t hi ck bound- a ry la yer is to m_ tke a sha _ -p co _ ' _ ier ap[ _ e a r rounded. It reduces the i n itial over- ' , / exp a n s io n and raises tl l e general level o f pre ss ure s over th e boat m il , a s s hown i n reference 12, The effec t o f b o u n da r y- la y e r tl _ ckne ss o n the pre ss ur e d r_ o f a s e- ries of boattall s o v e r a ran g e o f Mach number is p re s e n t e d i n re fe r e n ce 14.

A small rad i u s at the c o rner of the b o atta i l with th e thick e r boundary la y er (f ig .

VII I- 14) reduced the pr e ssure drag to 4 percent of t he ideal g r o ss _h ru s t, Thi s ra - d i u s ratio, R / D = 0.5, appe a rs t o be a reas o n able value f o r the typ _ of f la p and sea l arrangement shown o n the model in figur e VIII-12. A m o re gene r ou s radius can r _ duce t h e. pre s sur e drag even more , a s sh o w n in £ i gu _ , e VII I- 14 . K is e vid e nt that t h e i s olated bo a ttal l drag f or a va r i a ble f lap ej e ctor nozzle ca n b e s ig n if i ca n t at s ub s o n ic cruise. How this external drag i s modi f ied when the nacelle i s i v s t a ll ed on t h e F - 10 _ aircraft i s n ow examined, .... Figure VIII-15 shows a var i ab l e f la p ejector n ozzl e m ou n ted o n the F-106.

Flight performance for this noz z le typ e i s reported i n refere n c e 1 5 . A section o f the eleve n was cut out and r i gid l y att a c h ed to th e w i n g. To simplify i ts f abricatio n , t h e boattat l part of th e nozzle was so l id rather than made o f indi v i dual flaps and seals a n d t h e area ratio was fiXed at t h e appropriate value for subsonic cr u i s e.

Instal la tion effect o n boattail drag i s show . , .in f igure VIII-18 for a sharp- j unctured variable flap e j ector nozzle. The instal la tio n effect i S to g reat l y reduce boatta i l drag at t h e hi gher subsonic speeds. Drag is about z ero at Mach n umber s ..... fro m 0. 8 t o 0.9 , where the c o m p ressi on s h o ck is ahea d o f the no zz le. At Mat h 0.9 5 , where the shock is near the no zzle, th e drag goe s n egat i ve. Th e drag r mes s harp l y whe n th e compre ss ion s hock move s of f the e nd of the nozz l e at hi gh e r S peeds.

In th e i so l ated case , ro un ding the boattail junct u re was e f fe o tive i n reduc i n g drag. T h e effect o f doi ng t hi s in flig h t i s s h o w n on f i gttre VIII - 1 1 . D _ ta ar e sh o wn f o r R / D = 0 a n d R / D = 2 .5 no zzles. Little decrease in dra g b e l o w the already l o w t subso n ic va lu es was obtained. There wa s so m e reducti on i n drag above Mach 1 .0, • _" however, t Auxfllar_ inlet, ejector. - An auxiliary inlet e j ector nozz l e i s sho w n i n figure " V _ II-18. At subsonic cru is e the auxiliary inlet doors a r e open ar i d external air en- t e r s the secondary shroud to prevent overexpansio n of t he pri m ar y J_t at lo w va lu e s of nozzle pressure ratio. Since thi s air fi lls par t o _ the s hroud there i s a re d u ced requirement for exit - area variation with th e trailin g -ed ge flap _ . 1 _ thi s boattatl is co m pared with the var iable flap ejector ( _ ig. VIII-12), w _ find that the p ro ject ed ar _a _, h as been reduced fr o m 60 percent t o 45 percent o f t h e nacelle area. F O r an e q ual ; _ _ t ra iling- e dge- fla p length , th e boatta i l angl e can b e r e d u c ed fl ' b m 15° t o 10°. Th e i n - ternal expansion ratio Ag / A 8 for the auxiliary inlet ejecto r at t hi s flig h t s p _ d is about 2.0 {re f . 16 ) , co mpa red to a value of 1.4 for the vari ab le flap ejector.

The isolated performance of au xili ary inlet ejector nozzles is shc _ vn in f ib r e i VIII-19 at a Mach number of 0.90. The gross thrust coef ficient is u s ed _s a me asure / J of : loz z le perf o r m a n c e a :_ d i s d e fi n ed a _ th r u s t m intu _ dra g d i vid e d by t he l d u a l gross .... t h ru s t of t he pr i mary f l ow. T he op t im u m pe rforma nce wa s o b tained w he n th e in le t door s w e r e f i xed o pe _l and t he trailing- e dge flap s wer e h eld cl ose d ( r e f, 16), as sho wn by the upp e r curv e , It w a uld b e d es irabl e t o minimize t he a c tu a ti o n require .

m e nts an d t he me chan i cal c omp lexit y o f this n ozz l e type by a ll o wi ng s o m e o f it s co m - pon e nt s to b e p o si t i o ned by the a i r l o _ d s . At a s ub so n i c c r uibe p r ess ure r a t i o o f 3.75 th e n o zz l e p e r le : ma nce w as r e du c ed b y 4 percent wh en t h e in l e t do o r s w e re al - l oWed t o fl oa t a n d ended up bei n g cl o s ed o ver h a lf wa y ( ref , V/ ). A n a d d itiona l lo_s o f 2 ½ perce n t i n per f or manc e re s u l t e d wh en t he tr aili ng-ed g e f lap8 fl oa ted sli ght ly open from t h e cl osed p_,si ti o n (re f . 1 8). At thi s fl i ght s peed _ t h e np the use o f fl oat- ing co m po n ents r Qs ultod l a a n o z z le th at w a s oper a tin g well ev e ro x pa n ded at a l o W value o f p e rforma n c e .

T h e geometry o f th e float in g door s a n d f l aps would h ave to b e ca r efu l ly se l ected to avoid t h e prob l ems of In s tabi l it y . There is a l ot o f energy i n t he pri m ary flow. !_ I f the hin g e l ocatio n s a nd flap l engths a re not pic k ed with care , osci lla t i on s o f th e ! .

fl o a ting co m ponents m i g ht r es ult . Seve J . _ tl coa f i 6 mr a t i oa s w i t h fl oatin g co m pon e nts i " ha v e b ee n t e sted in th e wi n d t unn e ts a t the Lew i s R esearCh Ce n ter. Some o f the s e i designs proved to b e st a b l e (refs. 1 1 an d 18 ). Som e exa m p l e s o f tr a i li ng-ed g e fla p ; an d auxiliar y in le t do o r instabi li ty a re discus s ed i n re f erence 1 9. i T h e auxi li ary in l et e j ec tor n o zz l e is sh o wn i n sta ll ed o n the Fol08 in figure iI VII I - 2 0. ReSu l ts f or th i s nozz le are repor te d in re f erence 2 0 . This n ozz l e had i_ .

1 6 at tx i lt ar y in let d oo rs wh ich w ere either fixed in p la ce or free _l oati ng . Th e t op : i : thr e e o pened i nt o the tr o ugh bui l t i nt o t h e w ing. T h e boat tail p a rt of th i s n o zzle wa_ i_ a gai n fix e d a t th e subsonic c_ uis e a re a r a ti o . ! _ P er t or n m n ce o f th e a uxi lia ry i nl et ejecto r nozz le iso la ted and i n fl i gh t i _ Shown ' _I' / j_ in f igure VII I -21. Nozzle g ross thrust , coe ff icient and boattai l drag _a ttoed to ld_ai !: t hrust are prese n ted a s a f unc ti on o f d o or po s Y ' l 'he i so la ted g ro s s thrust co _ i- I : r f ic i e n t f e ll a l on g the dashed curve a nd it conti lm ed to Hse aS the doo r s were opened, i_ The _ li g ht thrust coeffi c ient was s o me w h at hi gh er an d i t also rose as the doors W ere _i _ o pened b u t it l eve l ed off and w as the sa me as i s ola te d f o r fu ll -t _ pe n doo rs . Th e I upper p_ rt of t he figure shows that boat ta il _ dra g wa s l ow er in fli g ht. Th e lower I _ _ boat ta il dr ag acc oun ts f o r t h e hi g h e r n oz z le g ro s s t h r u st wi th th e doe r s c lo sed and I_ part way ' open . Th e b e ne f it o f t h e l ower drag was los t , h uwev e r , when the doo rs I i were f ul l open. i As indicated i n f i gu re VIIi-2 1 , the per f o rm attc _ o f i l oa ti _g doors in _ light w a s near l y the o pti m u m . The doors f l o a t e d t o an a ver a ge position so m e wha t over m id- open. In the i s ola t ed case, floatin g d e al" p _ rform an ce wa s conside r ab l y low e r. Whdn th e tr ail in g f l ap s were a l l o w e d to f l o at in addi ti on, per f oimuince wa s eve i_ l ower. In f l i g ht the trailing f la ps prob a bly wou ld h a v e floated to a l ower-a r ea-ratio poslflor i 24 2 : t h _ n the y did in i so lat e d t es t s b e ca use h oa ttai l p res su r e s w er e high e r due to th e p r e s en ce of th e sh o ck . Thi s would hav e r e s u lt e d in a hi ghe r g r oss thru s t c oeffic i en t.

P o s iti o n s of the f lo ating inlet d oo r s are s how n in f igu re VIII - 9 2 for f ligh t _ t t i I M ac ll 0.9 , Th e trai l i n g e d ges o f th e doors a r e vi e w e d from the r ea r. Th e higher i e xt e rn a l pr e ss ur e around t h e t o p of t he n ozz l e h old t h e do o r s o pen around the t o p and th e o utboard s ide, T he_ o do o r _ admitted air from th e l _ w os t en e rgT r _g ton s of t h e f lo w field. T h e Do w e rextern a l pr e ss ure o n t h e b o tt om a nd t he inb o ard s i d e , , clos e d the remai n in g d oo r s. I n the c a_e wh ere _ lt t he d oors were fixed open , li ttle ai r appeared to b e e n te ri ng t h e doors on t h e bottom. T h e p e rfor m a nc e _ hown on t h e prev i ou s fi _ r e f or th i _ co ndi t io n was about t h e sa me as f or fl oating do o rs. Hig h er per f or m a n ce wa _ no t ac h iev ed wit h full-open doors becau s e very little o f the h ig h er , e nergy a i r arou n d t h e lmtto m e ntere d t h e n o zzle. '_ I P _ _ _ . - The 10w-a n g l o conical p lug n ozzl e i s s h own i n figur e V M -23 in the sub - ] s o n ic cruis e position. T h e out e r s h ro u d is r e tract ed up s tream for operation at this { flight s pe e d. Both the primary fl a p and the plug su r f ace ar e no w ex p o s e d to the e x- ter n al f l o w. T h l _ n o zzl e co n cept u tilizes s ec on dary f ie r y t o coo l so me of t h e engine pa r t s and t he pri m ary f la p , T h is cooli n g air is disc h ar g ed in t he an n ul u s between t h e pri m ary flap and the outer S h ro t z d. The major difference betWeen the two con- fi guratio ns show n i s i n t h e sh ape of t he sur f aces u p str e a m of t h e primary t h roat.

T he co z ffi g uration on t h e le f t has a cylindrical outer s h roud and a cl in i cal primary flap , whil e the othe r ha s a boattailed outer shroud and a rounded primary flap s hap e , It wa s s h own earlier in this disc u ssion that floati n g components of nozzles can be tin s table. In the plug co n cept s shown, t he plug is immersed in the high- en ergy primary strea m and can be u n stable if not supported properly (r efs. 2 1 at _ ti_. _ ).

..... From a stabi l ity st an dpoint it would be desirable to supp o rt th e plug from t he outer " ' shro u d with struts. H owe v er these st ru ts ar c no w i mme rsed i n t h e aft e rbur n er f lo w :S and present a dif f icult c o o li n g problem. Support s trut s also act as fl _m e h t _ lders for -i ?

unburned fuel I n t h e afterburner, Another suppo rt co n cept of interest is to cantil e ve r i the plug from the turbi n e frame with a sting through the c _n t e r of t he afterburlier.

Afterburner temp e rature profiles can be t ailored t o mai ntain cooler t e m peratur S s along the sting t o mini m ize the co o li n g r e qui r e m ents. ' H b wever, the sting must b b surf en ou gh t o prevent o scil la tion s o f the plug.

The is o lated per f or m an c e of these two pl u g co _ igu _ tion s is Shown in figt _ r _ !, V M -24. Da ta ar e presented both at static cond i tions and at a Mat h n um ber of 0.90. "_ At a subsonic cruise pressure ratio of 3.75, there i s a s ize c b le ef fec t of e _e rnal flow o n t h e performance of th e se tWo configU r ations. T he l o ss o f about B p e rcen t i l i / performance resulted from the combined drag o n t h e p r im ary flap an d 0n the afm ula r ': base where the s eco n dary flow is discharged. Romiding the surfaces up s t r ea m o f the prima ry th roa t d i d improve the performance, as exp e cted. A further di s cussion 24 _ c of t _e isola t e d perf o r ma nc e of a ser l e_ o f c oni c a l plug l_ozz le s ovvr _ wide r_ tn ge o f fli g ht s p e ed s is p re sen ted i n r e f e rence s 2 3 to 2 7 .

The p lug no zz le i s s h ow n inst a ll e d o n th e a ir c r a ft in figur e VIII -2 5. Thi s n o zz le was u n co o led a nd w _ts o pe r a te d with th e p r im ary thro _t t i n th e fixed pos iti o n s h ow n I .

w i th the engin e i n th e n o na R erlmrning mode. Th e l ns t _tl led perf o r m a n ce of th is n o_ ,.

zl e i s co m pared to i s o l at e d d _ ta i n "fig u r e VI I I=26. T he effect of t he m ov e m e nt of t he c o m pr ess ion s ho ck ca n b e see n on thi s f i g ure. It l fl a hea d of t he no zz l e at M as h nu mb e r s b e low 0.9. At Mat h num b er 0.95 it i s near th e p r imary flap , a nd th e h i ghes t no zz le g ro ss t h ru s t coeffici e nt i s o b t a i ne d. Th e th r us t coefficient d rop s s harply near Ma c h I.0 , w h ere t h e compre ss io n Sh ock mo ve s of fthe en d o f th e n oz - z l e a nd lo w ex t er na lpressu r e is o b ta in ed- pr im a rily o n t h e p lug sur face . Flig h t p e rfor m a n ce data f or th i s no zzl e ar e pre s e n t e d in refere n ce 20. ,_ N ozzle performance comparison. - Figure VIII-27 sho ws gr o ss t hrus tc o effici en t bot h isol ate d a n d in f lig ht.T h e v ar i a bl e f l a peje ct or w as th 0 poor e st w hen isolat ed b ecause o f its b oatt ail drag, bt l t the b est whe n i n stalled. The pl u g ha d very h i g h per- t r e f in a n ce w h en i s ol ated a n d i m pr o ve d i n f lig ht , so that it was s eco n d b e s t. I _ol at e d i_i per for ma n ce o f t h e aux ili ary i n let ejec tor was good if the do o rs a n d flaps wer e . actu- _, ate d ; bu t f li g ht per f or ma n ce s ta y ed a b o u t th e s a m e.

t '

Nozzle Cooling

In the d is c ussion so f ar, a e r o dynam i c per f o rm ance has bee n st re ss ed. N oz zl e li h • : c oolil_g is a n oth er pr ob le m , pa rt icularly with t h e p l u_ n ozz l e. E xpe r i me nt al h eat- ! : tra n sfe r st u d i es have b een mad e at Le w i s o n bot h e j ector-type n oz z les and _ l u g - type I no zz l es. A br ief d is cus sio n of these e xp eri m e n tal res u lts iS p r ese n te d i n th _ f oll o w - l n g paragraphs.

Film coolin{_. - Fig u re VIII-2 8 sh o w s so m e typica l re s ult s fr o m t h e e jector c 0 o l- !i # , ing s t- d y. A cy li ndr i cal ejector was tested on a J-85 aft e rbur n i ng turbojet eng in e in a n altit u d e f aci li ty. E jector coo l i n g w as acco m pli s h e d by film c ooling and ra d i atio t t.

Film coo ling i s a mea n s of insu lat i ng the ej e c t or wa l l fro m t h_ h o t priit m ry Jet with a la y er of the cooler seco nd a r y a i r. EJector wall temperatures ar e show n _ ts _t ftmc- : t ion o f di sta n cef rom th e primary e x itf o r a ca se w ith ma ximum afte r bur nin g ( a ppr ox - Im a t et y 3 100 ° F) a n d h ig h s e c on da ry fl o w r a te. Th e pr e dict e dtem p ei_Rir e s w e re obt ai n e d f ro m a h ea t-b a l a n c e calc ulation f orth e wa l l (r e f .28) . T he ius t ila t i t l g e ff e ct o f t he s ec ond a ry st r ea m w as ca l c u late d by u s ing a m o di f i e d Hatch- P _i p eli fi lm- c ooling corre la t i o n .Tld s corre la t ion was e m pi r ica lly developed for a fla t p la te, sub s oni c flow, and no pre s sure gradie n t. R wa s m od if ied f o r a nn u la r flo _ with Var y ..

i " I b \ i ng p r ess ure s . T he pr e dicted t _m p e r a t u r es ag r ee r e asonably well wR h t he men ..

sL :'o d v a lu es ° A c om p a ri so n of t h e v a riou s c a lculated h e a ting an d co o l i n g m echa n i sms l_ s h o wn in f igure VIII - 29, T he u pper part o f t h e f igure s how s t he p a r ame ters that he at t h e wall a nd th e lower p a rt sh o w s t he e oo l lt _g ter ms a s a f u n c t i on o f di s t a n ce from the pr ima r y ex i t. Radiation f r om th e hot g a_ t o th e wall I s a ppr o x ima tel y u nif ornt over the entire ej e et o r. TI _ s eco nd a r y ai r s tr eam f ilm c o ols the w all fo r ab o ut two-thirds of th e ej e ct o r le ngth. At this p o i n t , th e s e cond ary str ea m h a s b o c omo h o tt e r th an t he w r it a r i d , th e r e aft e r , add _ h e at t o th e wa ll . Ziadiat i ot _f ro m t he wall to t he s ur - r o undin gs is t h e o n l y cooh ng m e chan i sm o ver t h e la s t th i rd o f t h e n o z z le. 8 1 m tlar r e sult s w e re obt a i ned for other pres s ure r at i o s a n d s e c ondary flow rate s , Thu s , t h e predi c t i o n techniques develop e d ca n be used with confident0 to d e s i gn e J 0 ¢tor- ty p e no z z les.

Fig ure VIII- 8 0 s how s some typic al results f ro m th e plug fi l m-coo l ing study. A 21.6-ce n ti me t e r-d ia m e ter model wa s t e sted in a n al ti tud e f a c i lity. Coo li n g a i r c a - " tot e d t h e pl u g thro u g h a sting m ount. Three separate p l ug m od e ls were tested, T i l e first had a fi l m -coo li ng slot at th e S0-percent point (i. e., halfway betw e e n the pri- mary throat and the e nd of the p l ug). The second h ad a slot at the 1 0 -p e rc e nt po i nt and the t hi r d plug b ' ,d a Slot upstream of t h e no z z l e throat a t the -10-per ce nt point.

Tests were made at primary te m peratU re s to S 40° F. A typica l plug s t a tte-pre s_ re distribution is s h own for a hi g h n ozzle pressure ratio. Typically , the pressure d is- tr t b u tio n s downstream o f the S0 - perce nt S lot are about constant, while pressure gr adients following th e other two s l ot $ a r e first favorable an d then advers e .

In fl gm re VUI - 31, m_ asu r ed cooling effiei e ncie S a r e co m pare d with t h e H atch- P apell f i hn-cooling correlat i on. Cooling efficiency is simply a ratio of pri m ary r e- covery te m p era t u re m inus wa U temperature t o the primary recovery temperature • mlnus the coolant inlet static te m per a ture. W hen t he wall te m perature is eqtml to th e coolant temperature , this ra tio Is equal to 1.0; as th _ wall te m pera tu_ ' e in- t creases, th e r a ti o decreases. T he Hatch - Papel l para m et e r in c ludes about 10 differ- e n t ter ms , i n cl u di ng di stance fr o m slot exit a n d coolant flow rate (s e e ref. 2 8). T h e curv e o n the figure rep res e n ts t h e co r r e latio n . T h e circu la r symbols on t h e flgtlre, f or th e 50-perce n t slot, co r related w ell for all press u r _ ra ti o s , c oolan t f l o w r at e s, an d a ppropriate sec on dar y s hroud positi on s. Th e t r i e tngttJa r symb ols, for th e 10.- perc e nt slot, ge n e ra lly fel l above th _ correlation line. T h e "kn ee " in th _ da _ a was found to corresp o nd to the point whe r e the p r essure graPl e_( turns _ 1"o _ f a _ .o _ 'ableto adv e rse. The favorable pressure gradient r e tai'dS ml _ n _ of the p r t in _ try to l d c _la nt streams, keeping the wall te m p 0 ratu: o s low. The adver _ pressure gradient accel- e rate s the m ixing, caus in g a sharp i dc rease in Wall teniptrattt r e. For the - 10- percent slot, si m ilar res u lts were obtained. All the data of interest fell o n or above 1

I

24_ - - -4 t \ I • the correlati on line, ho wever. This m eanS the measured wa ll te m peratuees were l ower than w ould be p r edicted us i ng t he correlation. T h erefore _ the use of the cor- re la tio n should r e sult in the conservat i ve prediction of plug wall t e m peratur e s.

Convecttvel] r cooled plu_ nozzle. - H eat-transfer tests were also made on a convectively cooled plu g nozzle s ystem ( r ef. 29). The plug wa s strut s u pported for !

easy a t tachment to the J-8 5 engi ne (see fig. VIII-3 _ ). Coo ling air was obt a ined fro m t he co m pressor di scharge ports of t he engine. Cooling channels (shown in S ection A-A, fig. VI11 - 32) w e re formed a lo ng the s o rfac e of the plug and stru t s by at t aching i nickel fin s to the hi g h-strength outer wall. Nickel was use d beca u se of its high ther- • re al co n d u ctivity and the r e sultin g high effectiv e heat transfer fro m the out er wall to 4 the coolan t . A co ni cal extension was attached to th e 6 0-percent point on _ h e plug and was film coo le d w i th the cooling air di sc ha rging from the plu g cooling channels. The nozzle during a high-te m per a ture t e st in th e al ti tude fac _ ty i s shown in figure • ! _ VI _ -33. Th e primary gas tempera tu re was 2900 ° F - sli g htly less than maxim um _ i afterburning. The nozzle pres su re ratio was about 3.0. Th e plug an d struts were cooled with _ perc en t of the primary ai rflow.

The plu g w as designed f o r wall tempera tu res o f 1 7 40 ° F wi th t his cooling f low ra te, but the maxi m u m t em pe ra ture on th e plug was only 1500 ° F. The hi[ _ heSt tern- _i p e rature on the ext e nsion waS only 1300 ° F and the h ottest t e m pe ra tu re on the p r i- _ mary was about 1 7 00 ° F. The plug wall temperatures were lower than expected be- _ c aus e of th e ga s t emperature profile that existed in the engine. A typical profil e is _ shown on th e botto m left in figure VIII-32. A radius ratio of zero iS o n the c enter- line, arid a radi us ratio of 1.0 is on the p ri mary wall. T his profile r es ults from th e parti c ula r afte r burn er fuel nozzle d e sikql fo r this e n gine. Te m peratu re / _ were folmd to b e as much a s 500 ° F.cooler at the pl ug surface than th e maxi mum g _ ts t em pera- I!

ture. S in c e t hi s profile e _ t s es the plug cooli ng problem, it s ho uld be designed int o an y advanc ed plug n ozzle s y st e m .

The experimenta l heat-trans |e r te sts are su ml l la r tzed as follows: t , i (1) FirSt, for ejector- t ype noz z les, the Hatch-Pap e ll film-c oo ling co rr e c tion appears to yie ld a reasonabl e pre di ction of w a ll temp eratu re wh e n cofllbined with radiation terms. :.

(2) Second, for th e fi lm -c oo led pl ug , th e Hatch-Papell co rrela tion resu l ted in a cons e r vative pr edi c tion of pl u g wall temperatu r e s . I. : / (3) Finally, it wa s demonstrated that an ai r-c oo l e d p lu g could be coo l ed in _ ' /' aft e rburfll ng turboj e t w i[h a reasonabl e amount of co m presso r b _ ed air. . ; Th e pu r pos _ of th es r_ exp et i m eht al tests was to deve lo p the p redi ction methods nec- essary to evaluate advanc ed sl at te rns . , .

Supersonic cridse aircraft e h ldieS. - Tire next lo gical S t e p is t o tre e the p redic- tion m ethods deve | otSed and to extrapo la te t h e sm al | -scai e p | u g no_.zle data t O a i uli- • / i size s uper s o ni c cruise eng i ne. T he followi n g is a br _ tef s u m mary o f two t h eor e t i ca l studies that are und e rWay at the Lewis Research Center. In the first, engine fue l is us ed to regenerat i v e ly cool th e plug. In the second , compressor a i r is the coola n t.

In eac h case , a stin g -supported p l ug was selected , to e li m i n a te the struts that are immersed i n the hottest region of th e hot gas. T he fuel-coo le d nozzle app ea r s feas i ble, fr om a heat-trans f er st an dpo i nt, to cool the plug and sting s upport. There- fore, no e ngi n e cycle air would b e required to c ool t he p l ug. Also , p l ug W a ll te rn - _' : I v perature S could be kept below 1000° F Whenusing fuel coo l_ ng.

The the o retica l study using a i r cooling i ndi c ates t ha t f or maxi m u m afterburning , $ a 21 percent o f the engine cycle air would b e required to cool the plug below 1 7 40o F.

& A lso, the afterburner i s pr e su m ed to be on during S upersonic crui s e, resulting i n / : p rimary te m peratures of about 19 0 0° F. T h e plug would not have a very lo n g life at this te m perature unless i t were cooled. The calcu l atio n s Show that 1 / _ percent of [ : / i : the engine cycle air is needed to ensure reasonable plug t em pera tu res during super - sonic crui s e.

Summary

The effect of nozzle type on the range of a supersonic cruise ai rcraft , if the c oo ling r eq uire m en t s are included a l ong with the aerodyna m ic performance c ha rac - teristics, is shown in f i gure VIII-34. The airplane an d the two m iss ion _ illus t rated are the Same as t h ose presented i n fi gu re VIII - 1. T h e auxiliary inlet e j ector n ozzle • i s used as a baseline conf i guratio n and i t provi d es 3930 na u tic _ tl mil e _ of ra ng e for a typ i c al Mach 2.70 m i ss ion . Both ana l ytic al c a lc u lations an d experime n ta l data (fig.

.... V Iii - 7) h ave show n that the variable _ia p ejector ha s about 1 / 2 - percent hi g her gross :: . t hru s t during cruise a n d , ther efo re , pr o vi d es a n a d ditiona l 68 na u ti cal nfll e s o f range. T h e plug n ozzle is competitive with the va riable f lap eje ctor if the plug sur - i I face is cooled with the engine fuel at no loss in cycle ef f iciency. Howe _ ,er, a p lu g cool ed With compressor di scharge air show s little gain in ran ge over th e baseline nozzle. The range of the air - c oo led plug could be impro ve d if a n in ter s tag _ bl e ed could be used aS the source f or the coo lin g flow. The plug nozzle has So m e other f eatures t ha t make i t an a ttr acti v e nozzle conc e pt. As m entioned pr _ iously , it may be eas i er to seal , ha s less m ec ha nical complexity, _ and m ay be i nhere n tly quiete r .

The range co m parison sho w n does n ot account for afly of the _ e factors. _ ' T h e variab l e f la p ej e cto r and the plug nozzle also proVid e l_ lo r e rah g e for _ n al l -subso ni c (Mach 0.90) mission (fig. VuI-34). These tw o n o zz le concepts _t adth e highest installed performance at subsonic cruise, based on in-flig h t thr u st _ east i fe- m ents using th e F-106 aircraft (fig. VIH-27).

\

SUPERSONIC DASH AIRCRAFT

Th e e xhau s t systems for th e s e co nd t ype of air c raft - the supe r s o ni c da s h _ . , plane s t ha t cru is e primarily at subs onic s p ee d s - a re ex am ined brief l y in thi s s ee- ' r tion. Fo r this kind of mission , afterbu r ni n g turbofan engines would be used rather 'T tit an turbojets. Sub s onic performance is m o s t im portant , but it is a ccephtble to compromise s up e r s onic p e rfor m ance ti i t helps to minimi z e the nozz l e weight.

S o m e exa m ples of these nozzles are s h own on figure V] I I - 35. On t h e l e ft i s a varia- , : i ble co n vergent - diverg e nt nozzle wh i c h is shown in the subsonic c ru is e position. The tufts sticking out of the top of the nozzle were used in flig ht tests to detect flow sepa - ration. In co n cept, it i s sir c_ll ar to th e v a ri a ble flap ejector except that for u s e on a turbofan engine the seconda _ ', ; _ rflow could be eliminated. In addition , the sonic ! , _ o area variation i s la rger. In fact, during full a t terbur nin g the sonic area could be _ . _ . 1 i "1 tw i ce as large as that shown. Some amount of internal expa n si o n is need e d again for i_ , hig h - speed op e ration, but it would be l e ss than that of the va riable f la p e je ctor. By i ' _ con s tructing the boattail with over la pping f la ps and seals and by using enough actua - I_ tors , these variations could be provided. But to m inimiz e the m e c hanical probl e ms , 1.1_ these fla ps shou l d be s hort. Therefore , at s ub s oni c crui s e speeds, the boattatl on thi s type of nozzle is even larger and ha s steeper angles than the var iable f la p ejee - f _ tot. For the nozzle shown, th e pro j ected area i S about 75 per c e nt of the r_ ce Ue area and the m axi m um angle is 24°. Putting a pl u g in the n ozzle, a_ _ hown on t h e I_ Y _ right, would decrease both the boattai l area and angle. /f the plUg ha d ade q u _ t _ _ ' cooling , it could also be used to suppre ss infrared radiation from th e hot engine l_ parts. However, its structural weight would probably b e hi gher tha t l that of the other nozzle.

i

Isolated Performance

The isola ted pe r fo rman c e of t h e two su pe r son ic da _ h config u rations i s s hoW nin f i g u r e VHI - 36. Fo r r efe r e n ce , the int e rnal perfo rman c e of an i de a l co nv e rg e n t n o_ ' - zi e without any exte rna l d ra g i s p r ese n ted. The dro po t f in per t o r m a t lc e o f t h is f loO - zie with increasing pressu: e ratio is a n indi ca tio n of its tm der expan s ion losses.

The plug nozzle has the same perfor m ance cha r acter i stics as t h e ideal converg et i t nozzle. The small decrease in p e rfo rm ance with ext e rnal flow resulted from tl_ e drag on the 9° circular - arc bo a tt ai l. At a typical subsonic c _ls e pre s su r e ratio of 2, 8 for a turbofan engine, th e plug nozzle provides a high gross thrt l s t coe f ficient of better than 97 percent. At th e sa m e n ozzle pressure ratio th e v a r iabl e co n Vergent - divergent nozzle h a d a s lightly lower performance, about 96 percent. The diffe r e t ice 24 9 t \ In p e rf o r manc e w as due t o t h e hig her dra g as so c i ated w i t h th e 2 4° boa t ta i l, Some add i ti o nal discussion of th e I s olate d -p e r f or m anc e o f plug and conver ge nt- di ver g e n t nozzles fo r a pp l i c ation t o a s up e r s onic dash a i rc ra ft are pre se nted in r ef ere n c es S O a n d 31.

The hig h -a n gle b o attall Shown for t h e variabl e co n ver g e n t-div e r g e n t n o z zle (fi g .

VIII-35) i S operating near t h e limit f o r maintai n i n g attached floW. Typ i cal pressure di s tri b ution s f or this 2 4° b oattail s h a p e are sh ow n i n _ ig _r e VII1-37. If the flow re- m a i ns attached the di s tributio n Is as Shown b y th e s olid c urv e . If the f l ow separate s local l y , a distri b ution a s shown by th e dash e d cu _ 've results, with a n incr e ase in boattail drag. It i s i m portant , then , t o define the separation c ha r act e ri s t i c s o f these hi gh-an g le boattails and to d e termi n e their sensit _ vity to Reynolds number and to in- ; _ stallat l on effects. The Lew i s Research Center has recently f li ght t e sted a ser i es of these h i gh-angle boatta i l s and data are now a va i l ab l e that show some of these effects : _ on bo a ttail drag.

' N acell e Installation

The con ver g e n t- div e r.g e n t no zzles t hat were test fl o w n aresh o w n in fig u re i VIII-38. These t hr ee no zzles all had the same pr o jected boatta i l area a nd a 24° t_l _ a ngl e a t the e n d o f th e no zzle. The two o n top were the Same , except t h at th e Case 1

l

nozzle w a s moved d o wnstrea m relative to t h e Ca s e 2 nozz l e by about o ne - ha lf nozz l e dia m eter. The boattail junct u re f or th e se tw o w as f ail, l y s ha rp. The C ase 3 no zzle i!

_ / ha d a f u ll circ u lar-arc boatta i l.

_!:_ : B oa tta i l drag d i vid e d by i deal pri m al'y thrust f o l"the se n o_.z les is s hown in f i g- u re V IH - 39a s a f un c tio n o fR e yno ldsnu mb e _ . R eyn olds n umbe r w a s variedbyfly i ng the F-t0 6 at d i ffer e n t alt i_ d e s. A n gle o f attack was held const m it at 6° by fly i ng i n • ' co o r din a t ed tur n satt h e hig he r Re y n olds nu mber. T he lo west R ey noldsnu mbers wer e o b ta in ed by flyin g s trai g ht at an altitude of 45 000 feet. T h e hig he s t Rey no lds t nu mbers were o b tained by fl yi ng i n 3- g tu r n s at an alti tu d e o f 15 0 00 feet. Re y n olds n u m ber was b ased o n th e l e ngth from the inle t cowl lip to the n oz zl e atta c h in e nt poi n t.

The case n umber o n th e li n e s of da ta co r r e spo nd to th e nozzle de s i gna tio ns of figure VI H - 38. With the Case 1 n ozzle at the lowest R eYnolds n umb e_ th e boattail drag Wa s as much a s 10 perce n t of th e ideal pri m a ry g ros s thrust. The dra g w a _ l ow ered co n - si de r ably w he n this nozzle w as moved forwa _ l, g iving Ca S e 9 . T his was a result of _ the more favorab l e flow field cloSe _ to _h e W i n g . T h e ( _a se $ cir c ular-arc co n to u r ? pro v ided the lowest drag, Separation of the type descr i bed i n fi gu re VIH-$7 wa s encou n t er ed with these n ozzles. Separat ion wa s detected by pressure measur e fl i e n ts a n d tu f t s . Areas of separati o n a r e show n on figure VIii-39 by the sh aded regio n s. S e p ar ation wa s e n- 2 4 9 !i.

countered with a ll th e n o zzle s except th e c i r c u l ar arc at th e h i ghe s t R e y nol ds nu m - b e r. A s the R ey n o ld s number wa s increased, dr R g was reduced fo r all these nozzles b e c a u s e th is red u ced the tendency of the extern al flow to separate.

Two fra m e s from a m ot i on p ict u re o f t u ft s for t h e Case 2 nozzle are show n I n fig ure VIII=40. Thi s motion picture was taken by a camera mounted low i n the tai l .

The an g l e o f v i ew i s such that the tufts that appear to be o n the top are actually on the upper inboard sid e. The two frames show extreme positi o ns taken by the inboard tufts, w h ich were In a re g i on o f sepa ra ted flow. Flow o n t he outboard side and over the top of th e wing i S seen to be steady.

ii

: Fuselage Installation "

n many f i g h ter de s igns th e e n gi n es a r e packed i n to the fu s elage as shown in ! ;, _ fig u re VIII _ 41. The nozzles become more buried, th e n , w i t hi n th e airframe. The ;': _!

La ngls y R e s ea r e _ h C e n terhas b ee n st u dy lng thesepr o ble m sa ndI n o ne s e rles of test s i_ they ma d e s om e para m etri c variati on s i n the reg i o rt o f th e no z zl es ( s hown by th e i : _ Solid lin e s ). Som e o f theres u lts oft hesetests areprese n ted i n r ef e r e n ces 32 to36. I_ A po rtion are repeated here in s i n ce ex ha u st no zzlein s tal l at ion in theaftf us e lag e of I tw i n- e ngin e a i rc r a ft isc u r re n tl y pe rtin e n t t oth e d evelopment o_milit ar y a i rc raf t i_ _ a_ terbody- noz z l e d y n amo m ete rs h ownin fi gu re VHI-42 , desc ri be d i n refere n ce 37. ' The res u lts p rese n ted h er e in we r e o btained w ith th e tw i n- e ngin e-f us elage i'l ' The m a ximum di a m ete ro f n ozz l e s w a s 10 . 2 ce ntim ete r s, whi c hr e pr e s e n ted a pprox ' I ma t el y a 1 / 1 2- s ca l e m od el . The e ngin ee x ha ust je ts we r e s imul a t e d with c o m - pr esse d a ir , a ndw ere oper a t e d ov e r a w id e ra n g e o f pr e ssur e r at io.T he d a tapre - i se n te d here in we r e a n al y z ed at v a lu es o f pressure r ati o rep r ese ntati ve o ft l_ o s e f o r , a t urbo fa n e ngin e. Th e c ir c u mfe renti a l lln e a t t h em id-bodyis a t r a nsv e rs e c ut cl o sed with a flexibl e s e al. The nonm etr i c f o rebod y ser v ed as a s u pp o rt f o r th e t a _ terbody. Al l forces on the af terbody an d nozzles were m ea S ured w i th two si x - compo n e n t ba lan c e s ar ran ged to yt e] d a b r eak do w n o f t h e fo r ce s into no z zl e thru s t min us drag , an d aft _ rbody drag. The s e i n ve st i g _ tti o ns were conducted i n the 16-foot t ra nsonic tunnel at t h e Langley R esear c h Cent er . Th i s p r esentat i on d eals with the effects of m odel confi gu ration changes in the vic i nity of the ex haust nozzles.

Jet-exit axial location. - R esults of a b ri ef study of the effect of j et=e x it axial lo cat ion o n af tel ' bod y di'a g , ex tra c t ed f r o m i 'ef e re n ce 38 , are P r ese nt e d in f igu re VHI-43. The afterbody drag coeffic i ent is based on nacelle m ax im um cross- sectio n a l area , and dra g va c atio n with M ath nu mbe r i s / s h own .

The afterbodies were al l the sa m e le n gth and were _ha ped for mi n im u m w ave drag at M ach 1.00 /, us i ng the pro ce du res o i ' refere n ce 39. The lo w er sk etc h I n c U- I

!

..'..

t \ Installation effect on nozzle performance. - To this point, interest ha s been c on- _ tered o n afterb o dy drag. Figur e VIII-45, pr e pared from material in refer e nc es 33 i and 40 , is concerned with instal la ti o n effect on exhau s t nozzle perfor m a n ce. T h e-- i I , performance coefficient used for making comparisons is A I (F - Dn) / F i] , t h e s u m of I nozzle gr o ss thru s t minus nozzle boattai l dra g , taken a s a ratio to ideal gross thru s t, j The drag term i n thi s expression i s dra g o n only the nozz l e boattail s i ndi ca ted by th e i sha ded regions i n the sketches on the right. Thi s figur e shows inc re m ent s i n th e coefficient using the nozzle static pe rf or m ance a s the re f e r ence, l _ te s ul ta are p r e, se n ted f or convergent and co n verge n t-divergent nozz l es.

The lower sketch Onthe right repre s ents a m odel i n whic h the nozzles were in- sta lled in an aerodyna m ically clean aft-fu s ela g e, for Which the af terbod y contour s i faired smoothly into those of the nozzle. With thi s insta l la tion a t m _ bsontc speeds, good pressure re c overy in the eXter na l flow e xert ed a thrus t o n the nozzle boatt al l, i The vertically ha tched b a r s show that in the clean instal lati on , at Maeh 0 .8 , the noz- _ . i ' zl e per f ormance exceeded the s tatic value by 5 or 6 percent of t h e ideal gross th r us t , for both types of exhaust nozzles. !_ Another type of installa ti o n i s repre s ented in the up p er sk e t ch , in whic h the _ _ afterbody incorporated extended fatr in g s outboard of the nozzles and a f uS e lage ex- II tension between the nozzles. Th e presence of these ext e nsion s t end ed to disturb the !ii stream l ine f low over th e nozzles, and to prevent g o od pres s ure recove r y in the ex - i_ tern al f lo w . The result wa s a severe dra g on the nozzle boattail. The sha ded bar s l_ , !!

o n the lower side o f the plot at the left s h ow t ha t f o r M a ch 0. _ the incX'ea s ed drag tm !t th e nozzles resulted i n a large loss o _ p et, fo rma n ce. T h e dif f er e nce i n p erf ormance between these tw o ins tal lati o ns of 10 o r 12 percent o f ideal gross th rust at h i gh s ub- sonic speed represents app r oximatel y 20 percent diff er e n ce in net thrust. The i n - stal la t i on effects ar e important. At Mach 1. 2 , both type s of nozzles suff e red a small loss. These co m parisons show t ha t the n o zzle operating environ m ent has a t critical influence on nozzle perfor m ance at hi gh st l bsonic speed S , b u t r e lati v ely less effect at superson i c speeds.

After_dy boattail angle. - Another factor t o be considered ia t he d es ign of an e. _ haust nozzle i nstal la t i on _ namely the boattail angle o f th e af t el _ body just upst r eam : / o f t h e nozz l e attach m ent, i _ treated in f igure VII t -46. Agai n th_ basic C o _ fig ur atio n was the clean twi n- e ngine a _ erbody with closely sp aced conv e rge n t nozzl e S. The • da ta present increments in the p e rformance p ara m ete r as a function of aftel _ body _ boat ta il angle.

At a Mach n umber o f 0 .8, wi th the n ozzle i n the mi li tar y poWer setti n g , goo d pressure recover y i n the external flow wa s realized; the n ozzle in s tall e d p e rfor- mance exceeded the S tatic perform an ce in all cases, and i m p _ ved with inc , 'ea s ing i boattail angle. 1 !

cates a configurat io n with th e Jet exits at the downstream end of th e f u se lag e , for which the afterbody drag is shown b y the cir c ular points. From this down s t ream p osition the j et exits were m oved forward on the fu s elag e by o n e-lutli body w id th as shown in th e mid d le sketch, and then by one f ull body width a s s h own i n t h e top .

sketch. Fo r these la tt er tWoaft er bodt eS to have th e sa m e area progression as _ hat i n the lower s ketch , the c ros s -secti0nal ar e a o f cyli n drical Jets i s in c luded. Whi l e t h e direct thrust 0 _ t h e nozzle s t' J excl u ded from these measurementS, the results do include the effects of Jet interfer e nce on afterbody drag.

At low supersonic speed S , th e dashed c urve shows cal c ulated dr_ g for a n axi- symmetric afterbody havin g a n are a progression equal t o t ha t of the m od e l S , t he : values being the su m of wave drag and skin friction. The good agre eme nt between the calculated drag and tha t measured o n the after bo dy with downstream exits i m - pr o ves confid e nce bo t h in the theory (r ef . 39) _m di n th e exp e ri me ntal t e chniq u es i : ., used .

A t all s peeds-up t o Math 1 .3, the afterbody w ith Jet exits a t the ext r e m e af t e n d _ s h ows the lo w es t value s o f dra g . At s u b sonic sp u ds , je t in terfere n ce on t he ex L t en d ed we dg e lnt erfairin g s i n crease s the aft e rb od y dra g . At th e hi gher spe eds , however , the after bo die s wit h we d ge-shape d ext e ns i ons b e gi n to show incr e a si n g b e n ef i t of fa vor a bl e jet in t e rfere.ace. I : Interfairing shape. - The ef f ect of in terfatring shape on the comb in ed dra g of t he afterbod y an d n ozzles i s iU ustrated in figUre VIII - 44. T h e sk e tc h on th e upper rig h t is a rear vie w of th e no zzle in stallation , an d the s had _ l area s how s th e region in w hi c hth e in te rf al rin g s hape sw e r e m o d ifie d . T h e _ 1 _ et _h on t he u p pe r left _hows !!

t h e ba si c afterbo d y , whic h ha d c los e l y sp a c ed noz zle s l o c at ed in the dow n str e am ii p o s i ti on . The dashed li ne in th e s R e teh i ndicate s the c on t o u r in t he plane o f sy m - I : m etry o f t he el li pti _ al i n terfairing s ha p e . Oth e r shapes were circu l ar a r t , bl unt , _i a n d b lu n t ext en d ed . The blunt i n terfairing had a flat b as e at th e nozz le -fuSe lag e [ v jurtctu r e. T h e b l un t-exte n d ed iht e rfair in g termina t e d i n a fiat base app r oxi m at el y fl u sh w it h th e n ozzle eX i tS. _ _ T h e results prese n t ed i n t h e pl o ts Show r _ laLiv e dra g _ s _t f t mct io ft o _ Mac h !_ n umber, w h ere the h ighest value of d l -a _ is g iv {m a value of ul _ ity. At subS o nic I sp eedstheaf t erb o d , _ w it h th eel li ptic _t l i n ter f a lrln g S h ow ed t h e l egi s t d ra g fo ropera- _ ti on either at m ili t ary p o wer o r with m a _ t i mu m after b ur n er. T h ereisl i t tl e choice _ betwee n t h e e lli ptica l an d circulat'-arc shapes. The use o f a f la t ba s_ w as c onsi s - : t eR ly d etri m e n tal at subso n ic spe eds. Th _ data p o i nt d a t t h e upper right s h ow all i the tn t erfairi n g shapes t o h ave apl _ roxifltately t h e sa i _ e drag at low stq)erso ni c i spee ds ,th o ugh theflat- b ase i n te r fa ir i n gs di dsho w slightl y l o wervalues i n t hi s ii speed ra ng e. : i ¸ . The plot o n- the right s h ows that a t l o w s up e r s o n ic sp e ed, M ac h 1.2, pr ess ur e recovery i n the exter n al flow is re l at i ve l y poor, an d that the nozzle perform a nce i s n ot si gn if i ca n t ly b e tter than f o r s tatic o peration. In th is speed ran g e, a fterbody bo attail a ng l e ha s little ef f ect o n the i n s tall ed n o zzle perf o rma n c e .

Nozzle lateral s acin . The eff ec t of no z zle l ateral spacing o n no zzl e p er f o r - m ance a n d on aft e r bo d y d r a g i s illu s t :rat e d in f igu r e V M- 4 q . Th e s k e tch on th e upper left represents the basic twin- e n gine a f t e rbo dy With convergent no zz l e s. In t he mid- die sketch , wh i ch is a rear view of the no zz le i n s ta l l ation , the spacing ratio i s de- fin ed a s th e ratio o f t he di s tanc e b e twee n th e no zz le centorli nes s to t h e nozzle di- ameter dn at the nozzl e-f u s e lag e Juncture.

T h e le f t Side of fi g ur e V M -4V shows the ef fe ct o f l at er al spacing o n exha us t noz - zle performance. The small r a te of increas e i n performance with increased spacing at bot h s ubso ni c and low supersonic speeds indicates that th e m utual interferenc e drag o f the n o zzles decreases With increasing di s tance be tw een the n o zzles. The ef- fect o f increased sp a c i ng on n ozz le performa n ce, then, is be ne fici a l but small, The r i g ht side o f this f igu re Sh ow s th e effect of lateral spacing o n the drag of the J complete af t er bo dy and nozzles. Values of r ela t ive dr ag are s ho w n as a fu n ctio n of _!i the spacing ra tio. The after bo d i es having spacing rat i os of 1.59 and 2.31 ha d lower values of fi n e n ess ratio t h a n the basic afterbody , which had a spacing ratio of 1.32. i_ A _ M ach 1 . 2 the basic data in di cate a muc h m ore ra pi d increase i n a R erbo d y drag i with increasing la teral spacing t han i s shown here, because th e drag wa s Influenced by _t fterbody fineness ratio as We ll as by lateral spacing. T h e valu e s of relative • drag presented for M ach 1.2 were obt a i n ed by apply, _ n g a correc ti on for fi n en e ss ra tio t o the wave dra g of th e two afterbodi e s hav i n g the l arger va l ues o f spacing i ra tio . The values presented o n t he plot for Mach 1.2 presumab l y show only the _ f- , _ ' : : feet o f la ter a l sp a c in g on aflerbody re l ative d rag . Th e fine ne ss ra tio co r rect ion is b ased on i n f o r m atio n co nt a in ed in re f ere nc e 4 1 , a n d i so utlin e d in. refe r e n ce 33. _ t The d rag te n ds to in cre ase sli g ht l y as sp ac e bet w ee n th e n ozzles in creases at I bo th subs on ic and l o w super so nic speeds. The n _ t res u lt , the n , i S t h at late ra l t spaci ng o f the ex ha ust n ozzles appear s no t t o be a crit i cal facto r i n the a e r o dynam i c , desi gn o f a n after bod y- n ozz l e i n stallati o n, f Tail interference. - The effect on e_ ha u st no zzle p er f or ma n ce of a dd ing ta i l I surfaces t o the afterbody in the vi c ini ty of t h e n ozzle i n sta lla t ion is show n in f igu r e I Vlli-48. The basic v _ ode l was the aerodyna m ically s mt_ o th twi n - e_g l n e afterbod y I wi th c on ve rg e n t-dive rg e n t n ozzles. The pl o t pr e s en ts the vari at i o n w ith M a c h nufn - I b er o f the c hange i n no z z le perf o rma n ce w h i ch o ccurr e d w ho r l ho r i zon tal a _ d vert i cal i tail surf a ces were added to the basic c o nfigurat i o n . Th _ di r e ct drag of the tail sur- faces does not enter these measurements ; t h e cha n g e i n exhaust nozzle p e t fo f n mn ce reflects only the c ha nge i n nozzle dra g caused b y pro x imity of the tail surf _i ces.

I !

253 t A s sh ow n in th e low e r curv e i n f i gure VIII-48, w i th th e e n gi ne s o p erat i ng at n d llt _ ry p o We r adding t he t a il s urfa c e s c a u s e d a los_ i n n o z zle p e r i o r ma ne_ at e ub- so nic s p e e d s , whi c h at l W tch 0 .95 a mo unt e d t o 4 percent of th e id e al gro ss thru s t.

T he up per curve s h ows re su lt s _ bta ino d with t he n oz z l e i n t he m axi mum _ fte r bu rn e r s ett i n g . At thi s c o nditi o n , ad d ing t h e t a il s ur face_ t o _h e _ ft e rbody had a sm all favorable effect o n nozz l _ per f orm a nce a t a ll s p e _ d _ b e l o w M a th I. 3.

i

S ummary

i'l Th e inst alla tio n p r obl em s e n c ounte re d in th e se tests are S u mm arized a s follows: !1 For underwin g e n gi n e nacell e s with high - a ng l e boattails , incr e asi n g R e y n olds n it ro - : , bet tends to decrease t h e extent of f low sep a r a tion a nd , th e refore , decrease drag.

, Fo _ th e F-i06 nacelle i ns tallatio n a more f orwa r d location o f the boat ta il tend _ to !

reduce drag, A n d E tpure circular-arc , boattai l had the least dra _ . For engines i'i m ounted in the f use la ge, a di s t u rbanc e to str e amli n e flow n e ar the nozzle b _ ttail _ ' is detrimental. A sha ll ow boat _i l-angle is requi r ed for su p ers o nic speed but not f t)r _ ' _ subsonic. The best performance is ob ta ined with the nozzles dt _ vnstrea m of the ter - i!

minus of the ai r frame, and finally th e lateral spacing does not app e a r to b e critica l . !

REFERENCES ! ]

1. Darchuk, George V. ; and Balombi n , J oSeph R . : Noise Evaluatio n of FoUr _ ¢ - i; haust Noz z l e s f or Af t erb U rnt n g Tu r bo je t E n gine. NASA TM X -2 014 , 1970. i; 2. Chow, W. L. ; and Addy, A. L. : Inte r action Betw e en Primary and S e co r l d al'y i! t Streams of Sup e r S onic E jector Systems an d Their Per f or man de Cl l_ t t _ t _ t erisU cs. i AIAA J., eel. 2, no. 4, Apr. 1 9 64, pp. 686-695.

3. Crabs, C lif f ord C . ; Boy _ r, Earle O. ; anti Mikkelson, Daniel C . : Eilgifleeri n g ASpects and First Flight Results of th e NASA _ ' -106 T ra n _ o rt ic P r . ulsion Re- search Aircra _ t. N AS A TM X - 52559, 1969. !_ 4 . Wilcox, Fred A. ; S am anich , Nick E . ; an d B la ha , t _ e r nard J. : Flight and Wind !_i • Tu r inel Investigation o f h |sta llati o n Effects on Supersonic Cruise EX ha ust Noz - zles at Trafmo _ ic Speeds. Paper 69-4 2 7, AIAA, Ju i c e 1969.

5. Crabs, C U ffo rd C. ; Mlkkelson , Daniel C. ; and B o yer, Earle O. : An fflfligh t _ - M ves U gatlon o f Airfr ar he EffectS on PropuLsion S ys te m P e rJorma n_ e at Tran s oni _ Speed s . Soc. Exper. Test _ ilot s , Tech. Rev., vet. 9, n o . 4, 19 6 g, p p . 51- 6 6.

2 54 J ft. B laha , Bernard J .; _ mdMiklteleo n , D an ie l _ ?.:Wlnd Tunn e l In v e_ t l gati on o f Air- fra m e In s tallati o n E ff e e tn o n U r gl e rw la g _n gtn e N am _ tl _ a t Ma e h N umber _ from 0.56 to 1,4 ' 4 . N A SA TM X - tflfl _ , 1 908 .

% Bla h _ t , Be rn ard J .; Mt _ ;el so n , D a nl e lC.; an d U _r_ ingto n , D o u g la s I _ .: Wind Tunnt _l I n ve s t iga tion o f In _ta llg t t o n _ ff ect_ on U _ td e_ vl n g S uperso ni_ Cru ise E xhau s t Nozzle s a t Tran _o ni c S p a cd _ , NA_ TM X -[t Se 04, 19 8 9 . _ 8 . Bl _ tha , Be r na rd J.. Effect of U n d v rwln _ En gin e N a c el l e S hapt _an d Location o n Bo attail Dra g a nti W inl _ Pre ss u r e s a t Maciz Numb e r s from 0, 66 to 1, 40, NA S A TM X., 1 97 9 , 1 970.

9. Mtkk e lso n , Dani e l C,; an d B _ ha , B or nar _ i J. : F li { _ ht and Wind Tunn e l I n veetti _ - t i on O f ln s tallata on Eff e cts o n Und e rwlng S up e rsonic C ru is e E xhaust Nozzl e s at T r anso nic S p e eds. Pre s e nted a t A GARD 1970 Aerodytta m ic In t er f er en c e Spe- cia l is t s M ee ti n g , Silv e r Spr ing_ Md., S ep t. 2 8-30 , 19 _ 0.

10. Blaha , Bernard J . : W i nd Tun n e l Investigation o f t h e F l ow Field Under a 60 D e - g re e Swept Win g a t Mach Number s Fro m 0.6 to 2 .6. NA SA TM X-6 2 665, 1 9 6 9 .

II. S teffen, Fred W. ; and J ones, John _ t. : Pe r formanc e of a Wind _ lhm n el Mod e l o f a n A e rodyna m ic aUy Positioned Variabl e Flap Ej ect o r at M & ch N u m be r s Fro m 0 to 2 .0, I _ ASA T M X-1 6 39, 19 6 8.

12. Shrewsbury, George D. : Eff eCt of Boattail Jtutctur e S llap _ on Pr e ssure Deag Co e fficients o f Isolated Aftez'bodtes. N ASA TM X-151' ] , 1968. i 13. Harr i ngton, Douglas E. : Jet l _ ct S on Boattail Pr e ssu r e Dra g o _ _ solat e d _ i i Ejector Nozzles at Mas h Num b ers F rom 0.60 to 1.4 ' / . NAS A TM X-1 _ 8 5 , 1969. '_ 14. B la ha, Bernard J . ; and B _ 'esnahan, Donald L. : Wind Ttmftel InStaliatton E ffects ' on I s olat e d Afterbodie s a t Mach N utt _er_ Ft' om 0.66 to 1.6. NASA TM X- 5 2 581, 19 _ 9.

15. Mikkelson, Daniel C.; and He a d, Verlon L. : Fli g ht In vestigation o f Air f r a me Instal l ation E _ fects on a Variable Flap Ejector N oz_. l e o _ an Und e rw _ g E ngine t_ N acelle at Mach N u m bex's Fro m 0.5 to 1.3. N AS A T M X-2010, 1970. |'il 16. Johns, Albert L. ; and St i ffen, F _ 'ed W. : l _ erforntan e t _ o f all At _ i li a _ In let _ Ject o r N ozzle With Fi _ ed Doors and Si ngle-Hl _ ge Tt _ tlitq _ -gdge Flap. NASA ti: .

/ TM X-20 2_ , 1970. I . , 17. Johns, Albert L .; aft _ lSteffefl , F r ed W. : P e rf o rntance of aft Atuti li a _ y Inl e t Ejectot" Nozzle With Fixed Inlet Doors arid Triple-i { L'ige T r ailing-Edge _la p. _ NASA TM X'2034, 19 '/ 0. [ i !

_ 55

• !

18,B r ee nahan,D o naldL,: P e rf o rman ce o f a n Ae ro dyn am i c a lly P oaiti one d A uxll.

ia r y I n!_ t E j e c tor N o zzl e a t M ath Nu mb e rs F r sm 0 to 2,0 ,. NASA T M X - 20 2 3, 1 9 7 0, 19 , J o hn _ , Albert L, : Per f o rm an ce o f Au x iliar y Inlet lS jo ctor Noz z le W i t h Floati ng Inl e t Door _ a nd Floatin g S in g l e -H i nge Trailin g -Edg e F lap A , NASA TM X* _ 1 0 8, 1970, 90. S ama n l c h , Hick F , , ; and Burl e y , Ric h ard R. : Flig h t P e rfor man c e ef A uxl U a x 'y In l et l _ Jcctor an d P l ug N oz z le at Tr a n so n ic Spood _ . P a p e r 7 0- 7 01 , AIAA, June 1970, 2 1. Wa _ ko , Robert A, : S t a bil i t y C ha ract e ri stics o f a S t ing- Su pported Collapsible z Plug NO zz le a t M a c h N U m ber s From 0 to 2 .0. NASA TM X-l '/ 04, 19 6 8.

22 . B lahs, l _ er n ard J. : Eff e ct of Nozzle Total Pressure and B ase B l e ed on t h e S ia- l billt y Characteristics of a S tin g = S upported Trunc a ted Plug Nozzl e . NASA TM X- 1 776 , 1969. ,' 2 3. Bresnallan, Donald L. ; and Jo hn s , Albert L. : Cold Flow Inv e stlgatzon e r a Low i_ , Angle T urboj e t P l u g No zz l e Wit h F ixetl T h ro a t anti T r an s latiug S h roud at M ath Numbers Fro m 0 to 2 .0. NASA TM X-1 6 19 0 1968.

i' 24. Bresnahan , Donald L. : Exp e r im ental Inv e stigation of a 10 ° Con i cal T urbojet _ i . Pl u g Nozzle W i th Iris Primary and Translating S h roud at Mac h Numbers !_ From 0 to _,. 0 . NASA T M X-l _ / 0 9, 19 6 8. 1 _ I 2 5. Bresmthan , Donald L. : Exper im ental Investi g atioi _ of a 1 0 ° Conical _ urbojet !: Plug Nozzle W i th Translating P rimary and Secondary Shroud _ a t Mac h Nu m - I - i be t s Fro m 0 to 2.0. NASA TM X-lVVV , 19 6 9. I_ 2 6. Johns , Albt 3r t L. : Qtll e sc e nt-Air Pe r for m ance of a Truncated T Ur bojet Plug _ t Nozzle With Shroud a n d P lug BaSe Flo _ s F roi _ a Common Sour v e . NASA i: _i TM X-1807, 1969.

2 7. Steff e n , Fred W. : Perforn u i nceof a 10° C onic a l Plug Nozzle Witha St b w ed Thl'u st Reverser at Maeh N u n i bers F ro m 0 t o 2.0. NASA T M X. _ 2L I_ , 19 _ 0.

28. C h e t ioWeth , Fr an cis C. ; and Lieb _ r m an, Arthur: Predi ct ion of Heat-T / _ as _ er , ; !

Characteri s tics f o r E j e c t o r ] _ x ha u S t No zz l t _ s. Aftalytic Methods in Aircrat t i ! : AerOdyna m ics. NASA S P - _2 8 , 19 _ 0 , pp. 62 3- _$ 8. ' _ ' : 2 9. C _ ark, John S. ; Graber, Edwin J. ; and Straight , David M. : Expe r i m ental Heat v Transfe r and F low Results From an Air-Cool e d P i ug Nozzle Systetu. NASA T M X-5289 _ , 1970.

2 6 6 / 7 _ 0, Wasko , R o be rt A , | a nd H a r rl n gto n, D o uglas E . : Pe r forman c e of a Collapsibl e Plug No zz le Ha v ing E it he r Two- P a sl tl o n Cyli n drlc_t l o r Va r iabl e A n glo F loat- in g 8hr o _ d s_ t Ma c h Nu mbe rs F rom 0 t o_ , 0. NABA T l_ X _18 57 , 19_8 , 3 1 , H a r ri n gto n, Dou gl asE , I P_rforma nce of C o nve rg e n t and Plu gNo_z lo _a tM a ch Num b ers Fr o m 0 t o 1,9_. NAflATM X-_112 , 1 9 7 0 .

_2 , R_n ck o l , Jack F , ' Aero dy namic l n t erfm ' e n eo B e t w ee n E x haust S y s te m a n d Afro { r ome . Pros, , _t_d at th e AGARD _ p o c ial i stt_ Mo ot ing on Aerodynam i c Int e r - f oroac o , S l lwr S p ring, Md ,, Sep t, 28 -a 0 , 1 0 7 0.

3 3 . Co r D o n,l_htko W . , Jr. ; andRu nck ol , , _ nc l_ F . : Ex p lora to ry 8tu a l v _ ofAir c raft ARo r body and F , x l t a us t - N o z zlo b l / ora ct lo n. N A S A TM X - 1 92 _ , 1909, 34.Or e a thuuso , W i l l iamK . : B t ondlng P ro L _ulslon wi t hAirframe , S pa c o, / Aorunnut. , voh _0,i_o. 6 , Nov . 1 96 8,pp. _ 9 - 6 8 , 35 , Mlgd a l,D. ; Mill e r, E . l i. ; an_S c hn e I L , W. C. : An l lxpe riment a l E v a lu a tion ofExh a ustNo zz le / Airfr a nm I nt 0 rf e renc_. P ap e r 8 9-43 0 , AI A A , June 1 9 6 9 .

_ 6 . Throndson , L. W. : Cl o s e -S p a c ed Noz z lesTwin JetConfl g u_' a fi c m. P aper 7 0- 934, AIAA , July1 9 _ 0 .

3 P / . M a id e n , Don a ld U . ; a n d Rtm c kel , J ack F. : Eff e ct ofNozzleh a ter a l S pac ing on AR e rbody Dr a g and Pe rform a n c e o f_Wln-J e t AR e rbody Mo de ls w ithConv e r- gen t No z zl e s a tM ae h Nu m b e r s up to2. 2, NASA TM X-209 9 , 19 7 0, 3 8 . B a rr i e r , Bobby bee ; andWood , F r e d e ri c k H. , Jr. : E ffect o f Je t V e lo c ity and Axialho ca tion o f Noz z l eE xiton t h e P e r f orman c eo fa T V c in-J et AR e rb o dy Mode l a t M ac h N u m ber s up to 2 ,2 , N A$ A T N D=5303 _ 1 9 6 9.

3 9 . Harri s , Roy V., Jr. : An A na ly sis a n d C or r el a ti on o f Ai r craft Wa ve Drag.

N ASA TM X -947 , 19 6 4 . t 40. Mercer, C h ar l es E. ; a n d Ber r i e r , B obby L. : Ef f ect o f Afterb o dy S hap e , No z - z le Ty pe, an d E ngin e Lateral Sp a c in g on th e I ns t a lle d P e rfo r man c e of a Twi n - / J et Ai te rbo d y Mo de l . NA SA T M X .- 1 85_ , 1969 .

4 1 Morris,D e ari e M . : A S um r aa r yof t h e Su pe rso nic P re ssu r _ Dra g t J f / 3 odi e s a f Rev olution. J . Aer os p a ce _ c i . , e e l._.8 , _o .7 , Ju l y 1 9 { 3 1 , p _ . _ 8- { i72.

_ 57 ,!

HISSION SENSITIVITY FOR EXHAU S T NOZZLES SUPER SONIC C R UISE AI R C R AFT; TAKE O FF G R OSS WEIGHT = 7 _ 000 0 LB I PAYL O AD - 4 9000LB SUPE R S O NIC C R UISE MIS S ION I S UBSONIC C R UISE MISSION 3930 i _ MI I 3280 N MI I 1 36 I I I I I RANGE j INCREMENT , 61 " ... " NMI I 7 -' 7 < 6 2 . 5 I 6 >.

i

U

r "- 1 === i -. _ .CRUISE LOITER._ _ . _ CRUISELOITER_ ' 1 % C HAN G E I N = THRUST COEFF WEIGHT THRUST COEFF cs- s 6 9sz Figure VIII - I _ j ,

I '

.... i > , , EXHAU S T NOZZLE CONCEPTS '- SUPERSONIC CRUISE AIRCRAFT i<i"

l

_p 4 ' I i • " ':¢ sQ C -6 7-7 6_ C- 6 7 - 3349 i VARIABLE AUXILIARY PLUG , FLAP EJECTOR INLET EJEC T OR cs- 569 4z ! ._ Figure VHI-2 i : ii. I_ . _ U _ " " _,,,mi = l .J _ m _ * L,,_iL ' ll I II "1 -- F ' ' I F ....... i i I I III I i I Ill _1 ; l'_ I_ . " -- OIVERGENT EJECTOR FLOW FIELD HIGHSECONDARY FLOW (CHOKED) LOWSECONDARY FLOW (IMPINGEMENT) !.

CS - 5 6 9 5 ?

FigureVIII - 3 t.'

I

'_ " SHROUD CONTOUR SEN81TIVITY '

AUXILIARY , INLETE J ECTOR

CORRECTED SECONDARY FLOW RATIO 1 . 01 m_ _ s -_ - 1 --- / ,ou ,DEo s , ou, oE , Dpp

.Z o.; _

GROSS 1 . 00 - , , .. . - COEFF .9 9 - .f,,\ $HAR ...... _..,_,, . _ " CONIC FLAP

.9 8 J I ! I I I I i

- .2 O .2 .4 . 6 .8 1.0 1.2 , , SPACING RATIO,S I Dp CS -5 6939 !

Figure VIII-5 t _

WEIGHT FLOWSENSITIVITY AUXILIARY INLETE d ECTOR !_

i SHARP SHOULDER, CONTOURED FLAP ili

CORRECTED SECOND A RY

: .99 - FLOW RATIO,

GROSS ' .98 - __ - '_ " --- " -- ""

* MINUS .97 - _ ,

co. --

OP A G .96-

-. / \

.9 5 I I I I I I

".2 0 .2 .4 ,_ . 6 .8 1 .0

SPACING RATIO cs.._ e _ 3e

Fig u reV | II-6

2 60

NOZZLE PERFORHANCE AT SUPERSONIC CRUISE 1.01 _ i "-- .__.- ---- -- THEORY '" 1 . O0-- _ EXPERIMENT GROS S " THRUST.99 - COEFF .9 8 _ m , VARIABLE AUXILIARY PLU G 3 ;: FLAP I NL ET i_ . EJECTOR EJECTOR cs- s6936 Figure V I l l-7 i!, EXHAUST N OZZLE TEST PROGRAM S II _. C - 69- 1 34 2 C : . 6 7-4332 ISOLATED NO ZZ LE 1 1 20 SC A LE F-IO _ _j , .? _ . \_ , F -f06 FLIGHT C- f, q- Z S ?! C5-5700, _ Figure VI II- 8 2 61 NACELLE-ENGINE INSTALLATION ,-F O RWARD _" _ LOAD CELL R EAR UNK /- FIXEDEI E VON

" / S E_ I o N

> / / / " ACCESSORY PACKAGE cs.s 69 99 Figure VIII.@ !I' LL I I N STALL A TIO N EFFECT ON WING PRESSURES

M o - O g o

i . 4 S T ATIC ORIFICES I ' - .2 ! P R E S S U R E 0 -- COFJ:F - . 2 ' . ; - . 4 o WI1HOUT ELI t - . 6 a WiTI'I NACELLE cs -_993 F lg Ur O V| I1 - 1 0 2 63 .

' : 7 : C - 69 - 15 8 5 C_- 5694 ) Figure V I II- 12 1 2 6 3 ' .t BOATTAIL PRESSURE DISTRIBUTION VARIABLE FLAP EJECTO R

M 0 ,0.90

+ 1 2. _ i ' i , / : i '., PRESSURE COEFF 0 .__ CS- 5 6948 - !

Figure VIII-1 3 !

ISOLATED BOATTAIL DRA G _ le, i VARIABLEFLAP EJECTOR NOZZLE INSTALLATION R I D . 2 . 5 Figur e V III - 1 5 " INSTALLATIONEFFECT O N BOATTAIL DRAG VARIABLE FLAPELECTOR .20 -- R I D- 0 _ .,. _ ,

. 1 6 - j _ __ I T _ , _\ -

_b_p 1 1

DRAG .08 !

C O EFF BOATTAIL _ D R // t 0 _ " " * FLIGHT !

.o4- . . _ Z_ . I SOLA _ D

- .o4 t z I , I I , ,I I , I ,5 .6 .7 .8 . _ 1 .0 L1 1.2 1.3 M0 c s. 5?ooz , FigureVIII - 1 6 _.

't

_OB _.

EFFECT OF ROUNDED JUNCTUREON BOATTAILDRAG FLIG H T DA T A _ R I D D "-' u - " - 0

,_ _ _ 2, 5 i

,12 = R . : , ' .

BOATTAIL "" ,, 6 - r E F _ _ , _ , , DRAG COEFF .08 = I .04- I O _,1 I I ,.

-.0%. .7 .9 1.1 1.3 I_ M0 c _ - _ 7oo l i_ Fi gure VII I-l? _ :

ii

AUX I L I ARY INL E T EJE CTOR ii F C. ( , , . I. 1' _ 1'11 ( : , 'i - _ ( , 9 ,10 Fi gure VlI I -18 . 266 E F FEC T OF FLO A TI N G C O MP ONENTS ON P ERF ORMAN CE l.O0 - A UX I LIARY INLET E J ECTOR M o, O. 9 0 .9 6 - . . - _I' ° FI X E D F b_PS J G R O S S FIX E D INLE T S CO E FF _F I XED FLAPS '.

, B8 , / l 2 3 4 5 6 1 c . q . , ( , ,_ . _ NO Z ZLE PRESSURE RATIO F i gu r e V III-19 AU XILI A R Y INLET E J ECT O R NOZZLE AN D ELEVON TROUGH 4r - " !:

• !

" ,F Fi gure VIII-20 2 i 3'I 1"_ . , IN S TALLATIONEFF E CT ON AU X ILIARYINLET EJECTOR PERFORMANCE A_ ,0 . 90 ,,. ... ------ FIXED DOORS BOATTAIL AND FLA P R A TIO OF .04 F " DR A G T O ,02 _ ---- " I _O LAT E D :- Ib F A L THRUSTO " ---- FLIGHT .95 , ,'- F LOATIN G DOORS : ' i GROSS , , THRUS T COEFF. ' _ NGDOORS : , I / _ 'FLO A T . _ IN O DOORS ANDFLAPS ', .8 _ i : C LOSED OPEN 1 1 4 DOOR POSITRON c s. s_ Fi g u r e Vlll - 21 POS ITIO N OF FLOATIN G DOORS " '

/ Mo -0.9 "

. " _l O E. , G E " : CLO S ED - _ / ,' _ O F DOOR cs . s T ooa ....

/ / ., Fl _ Jre VIIii 22 / , / 1 2 B 8 ' , Imllllll e _ . _ m- _ \ e LU G NOZZLE C-69 - I c_ 40 C-69- 1 34Z : ( CONI C AL PRIMARY ROUN D ED PRIMARY _ : Figure V ]]I - 23 cs . s t , 944 ,.

F i , L _ e !"

!

EFFECT OF E XT ER N A L FLOW,ON PLUG i , , i NOZZLE PERFORMANCE i_ _ PRIMAR Y FLAP i 1,00- i_ o CONICAL _ ,.

. _ . .9 8 - STATIC a ROUND E D i i G R OSS . 96 Mo - o . 9 0 , THRUST COEFF .94 , t_ ? .

.90 I j 2 3 4 5 6 1 N OZZLE PR E SSURE RATIO cs - se9 4T Figure VIII - 24 | B, \ PLUG NOZZLE INSTALLATION ¢ Fig ureV I f 1 -25 3 ' I0 .

i i .

:,4' _ " ° _° INSTALLATIONEFFECT ON PLU G NOZZLE PERFORHANCE -..o , -- F L IGHT ..... I S O LA TED _ . GROSS LO I ' THRUST COEFF,9 .6 .8 LO l.t 1.4 MO c8.s6997 Figure VIII - _ V_ .' COHPARISON OF NOZZLE PERFOR M A N CE

.:i o-O. 9

"' _ ' _ 1.0- ---- - ISOLA T ED 4 r, p : i ' i_ii .975_.__.__. FLI G HT 965 .94 , 1 , 94 5 .94 6 -- t .0.9 _ r_ -- r '- ' • , • I | " G ROSS • • I j THRUST COEFF,9 " ,' ,' I I " " I , . / | | |

'' J

• l = I _' i • I i I _ ; !

.8 i i .... I I i • ' VARIABLE AUXILIARY PLUG _ , , F LAP INL L _ 'r EJECTOR EJ _ TO R cs . s _99_ Figure VIII-2" / = : i '.

.... _ V _ " _'_ ..... l -- " 1 I I =l i =l --" I - I I" li II ' I - i ll i f I l l J _ .... ll III I

EJE C T O R SH R O U D COOLI N G

/ - S H R O UD

/ ! '

i

SEC Ond . .. . .. - ..... _ - 1

PRIMARY ' -' _ ' . . _ _L _ L : , 1100 PREDICTED TEoMFP , 1 00 SHROUD W A LL i _ j ___ 8 MEASU R EI)

: 3 oO F ° - " - T I I I

O 1.O DIS T A NCE F R OM PRIMA R Y EXIT C5- 5678 6 F Ig u r e V I I1 - 2 8

HEAT DIST RIBUTION

HEAT IN,2 SECONDARY i BTU / HR - FT5 C ONVECTION i PRIMARY i RA D IATION ' 10000 F _ SECONDAR Y CO N VECTI O N _ , HEAT OU T_ , I _il _ BT u l , _ - rro 5_ ..... . , - , • - "' _ : _ , _ : , / . " • RAOIA T ION ' ////i//111/11__ % - .

0 1.0 DISTANCE F ROM PRIMARY EXIT F i gure ' VIII - L R cs. s67 . 7

l

FILN-COOLED PLU G NOZZLE SECONDARY_ I . - .... --_ - __ _ . _ - " -SLO T i L SUPERSONIC PLUGSTATIC TO . _ _ \ PRIMARY TOTAL _ CRUISE ' PRESSURE RATIO .25 -- 0 I _" _"_ ' ,, _ 1 -.I 0.I .5 1.0 , DISTAN C E FROMPRIMARY EXIT Figure V III - 30 c s . _ 6 7 u PLUG SURFACEFILH-COOLINGCORRELATION COOLING _ , _ E FF, , .4 _ - : O _!_ ON / • j • :_ . _ .0 5 1 ' . 5 1 3 ii • .... _ H M CH-PAPELL P _ , RAMETER _ i Figure V I I1-_1 cs.s 6T 8 4 3 " / 3 I.

\

CONVECTIVELY COOLED PLUGNOZZLE

F igu r e VI II - 33 '; ' ' 4 !"I I 2 7 4 J EFFECT OF NOZZLE TYPE ON RANGE SUPERSONIC CRUISE AIRCRAFT; TAKEOFF GROS S WEIGHT- 750 001) LB = PAYLOAD . 4 9000 LB SUP ERSONIC C RUISE MISSION SUBSONI C CRUISE MISSION , , 3 9 3 0 NMI 3 280NMI • -, "-' 183 R ANGE INCREMENT , N MI 68 68 I 0 I I 0 AUXILIARYVA RI A BLE FUEL A IR AUXILIARY V A RIABLE PLU( "" _,, INLE T FLAP• _COOLEDCOOLEDj _ , INLET I_I.APj EJECT OR PL _ IG EJ E CTORcs - s 6 9 49 Figu r eV III -34 | ' " . EXHAUST NOZZLE CO N CEPTS FOR SUPERSONIC DA S H AIRCRAFT l .!

: " C - ? O -Z b 94 ' C - 6 9-40 _ 3 : % VARIABLE PLUG C ONVERGi_ NT 'DIVERGENT cs- 56 9 41 , . . F i gure VIII-) 5

EFFECT OF EXTERNAL FLOW ON SUPERSONIC DASH

NOZZLE PERFORMANCE 1. 00 -_ - _ THRUST _

o++ .+

_ " _ IDEALCONV E RGENT . 94 I " I I _ 1 LUG , IMO • O. 90 i 2 3 4 5 6 7 NOZZLE P R ESSURE _ I _ TIO {i, F igure V III - 3 6 cs ._6946 i++ , : : _ !,'I t

BOATTAIL PRES S URE DI S TRIBUTION i_

i_ VARIAB L E CONVEROERr - OIVERGE _ i , _

eo.o.9o i_

z4o _ , , . , , "\ + _ NOSEPARATION + PRES S URE COEFI: 0 ._ : ...+.,_# : _" $ _ PARA T IO N +,

NO '

i ,, i X . CS-56945 FigUl' e V i l l i - 3?

2 7 6 i .. .

_" --' _ - {I I I I [r II Ill ...... II " I ......... I"1 ............. _ ,,, ,'_^ ,_ ",,, r ," .... ! .....

Q \ , VARIABLE CONVERGENT-DIVERGENT NOZZLES CASE 1- R I D. 0,72D(TENDED 1 / 2DIAM CA S E 2 - R I D . 0.72 '. . . ._ / . ! CASE 3 " cs - sT oo 4 ....... : _ , ; , _ , _! CIRCULAR A R C Figure V II1 -38 REYNOLDS NUHBER EFFECT ON BOATTAIL DRAG 24 o BOATTAILS Mo - O.g "10 [__ _ CASE SEPA R ATION 7 i

• /

BOATTAIL . 06 DRAG TO _ THRUST :. .02

@

. O 2 3 4 5 6 7xlO 1 REYNOLDS NO. C_ -5 699 8 Fi gure VIII-39 2 '/ 7

FRAME FROM MOTION PICTURE OF TUFTS FOR CASE 2 NO Z ZLF .- SEPA R ATED _ TUFTS

O FI X ED

SECTION OFELEVON TWlN-F -. NGINg AF T IERBODY-NO Z ZL _ DYNAMOM _ TI _ R r louro VI II ..4 _ & . 16 - f ' % %.

AFT E R BOD,Y ]2 - DRAG COEFF IClENT AXISYMMETRIC .08 - _ ------ CALCULATED WAVE D R AO PLUS .04 - SKIN FRICTION O- I 1 I 1 I I I .4 .5 .8 1.0 12 1.4 1 .6 Mo Fig u reVI T , I-43 _ , h INTERFAIRt N GSHAPE EFFECT O N AFTERBODY ............... ' PLUS N OZZLE DRAG O BLU N :T.-- A BLU N T EXT E NDED O CIRCULAR ARC 1 3 ELLIF r lCAL MILITA R Y P O WER / VP _ XI M LI M A FTE R BURNI NG i , O .O ¢ OO RELATIVE & A DRAG _ A I' o L t J I _ I , I I I I i , _ .6 .8 " 1.0 .6 .8 1.0 1. 2 I A _ Mo Mo , i : ., Fi g ure VIII-44 i IN S T A LLA T ION E FF E CT ON E XHAUS T NOZZLE P E R F ORMANC E ' it " I -- " j

_ [i

M0 - 0.8 EXTE N DE O FAIRI NGS '_ , , Mo - 1.2 " "08 1 MILITARY POWER E ] _ '_ ' MAXIMUM SMOO T H .04 AF T ERBURNING ......

.....

=. 04 -.08 CO N V. C - D CO N V. C - D Figure VIII-_ \ ' l EFFECTOF AFTERBODY APPROACH ANGLEON NOZZLE PERFORMANCE o MILIT A RY POWER _ MAXlMIIMAFTERBURNIN G

%. u %. u

) 04 e m e_ m m

\, - _" -/ o

Figure VI tI. _ 16 EFFECT OF NOZZLE LATERAL SPAC I N G j¢, i ,. O M0 - 0. 8 MILITARY POWER _ Z _ Mo • Z .2 MAXIMUM AF T ERBURNING t ' f NOZZLE PERFORMANC E RELATIVE DRAG • : .08 AFTERBOI)Y O - -,.'-C ; .... _ , . LO- _

• " ( -q- , ) o.o _ o = . ,. o , o

'!

-. 04 1 I J 0 ) - t I 1. 2 1. 6 . . 2.0 2.4 1 . 2 1. 6 2 .0 2. 4.

F i gure VIII-4 _ Set > t i.

q t > . , !> i' TAIL INTERFER E N CE ON - NOZZLE PERFORMANCE ii: o MILITARY POWER , 1 3 n . 14 0 . /' -7 1 3 . ;_ : ?

a MAXIMUM AFTERBURNING , - . _ , . _ ..

•. , . I _ n = 4o PLA N VIEW _ _ ' , .

°04 - .I I I I _1 A , - I I -.08- . 6 .7 .8 .9 1 ;. 0 1.i 1 2 '1 3 : . : Figure VIII- _ " _ " # IS >> t \ / . f •_ I •

N 71 "1 9 4 6 0

IX . SUP ERS ON I C CRUISE INLETS

David N . B owdltcll, R obert E . Coltrin , Bobby W . S and e rs ,

• Norman E . Sorensen* andJ,Jseph F. Wass e rl _ uer

The des ig n of a s upe rso nic cruise i nle t is d e pe n dento n i ts application. The best i nlet for a m ission inco _ 'porates an optim um combination of i nlet cha r a c teris ti cs such as total-press u re recovery , cowl drag , bleed f low , and weight. To d e term in e : :!!i , the best inlet , it is n ecess a ry to defin e the s ensitivity of the aircraft range to eac h • / i nl et Character i stiC. This sensi ti vi ty I S shown in figure IX-I fo r a g upersonlc trans- . ! port powered by a£terburning turbojet engines a n d wit', a no m inaLra _e of a Utt / e ; oVer-39O0 n _ mtical miles. The S e nsi ti vity iS Shown as a ra ng e decrement, or de- :_ : crease i n range, in nautical miles , for the indicated cha _ es i n each of the inlet .... characteristics. R ed ucing th e t o t ai -pressttre re co very f r o w n 91 percent to fl0 per- cen t of th e fr ee - s trea m total press u re cauSes a $2 -mile range decreme n t, i n creas- i ng th e" na celle dr ag coefficient, based on th e inle t capture area,, fx,o m 0.08 to 0.09 t r ed uces range 41 miles. This dr ag incr ea se correspot l dS to in cr easi ng the extet.nal , cowl lip angle from 3° to 7°. Increasing bleed flow 1 pe r cent of th e ca pture mass flow decreases range 23 mile s , w hile a 10-percent increase in weight decreases ' : i range 17 mile s . Th e refor e , the supe rSonic trans po rt Znl et t _ d$ t o favor high re- cove r y , l ow drag , and low bleed at the e x pe nse of inl e t w _ tght, a less impo r ta n t •, . !

_. , , .-_! , - _ : c ha ract e ristic.

..... For su perso ni c cruise at Math numb e r s g rea ter tha n _ , oils of t he in let lztra m - "" , et e rs that affects th ese importa n .', i nl et character i stics is the internal cont _ acUo | t of _ , t i_ : th e superso ni c diffuser. This paper compares the pe rformance of s ev o ral Ltlle tS _., , i to e l b ow how th e amou-tt of internal co nt r action affect s th eir character i s tiCs. No t o nl y /h o s e inlet characte ri stic s that can b e megsured in ter m_ of l _ mge ar e e of _ id- . ered , but also inlet characteristic s su ch as ang|e- _ - att _ c k tolerance whtcl _ ar e more diflictttt to evaluate. In addition, m e th ods to improve the stability of the i n - tern al c _ ntl _ acflo n ihlet s a hd so t ne typical distorflon _ t h at ha ve been m e hs _ li'ed dur- i n g wind tu n n e l testi ng are discus s ed.

J_ • NASA-Am e s Research Center.

. e: , , I Q I EFFECT OFAMOUNT OFINTERNAL CONTRACTION ON INLET CHARACTERISTICS

Performanc e Charact e ristics

To show the ef f ect of in te r nal contractio n , the perfor m anc e of th e inl e t s shown in figures IX-2 and IX 4 i s discus s ed. ( _ o m e of th g se inl e ts (as well as o th e r s) were destgn 0 d by using a c o m pute r prog _ t m W hi c h incorporates the method of _ characteri s tic S , a s d e sc-tlbed in f e t e. 1 to 3. ) The cowl lip di am eters of th e inlets _ shown In tho se fi g ures range from 14 inch e s for fhe all-exte r nal - co m pre s sion inle t to about 18 i n che s for th e th re e islets with inter na l contrac t ion. All th e i n l e ts are axt s ymmetrlc; h ow ever , so m e results of th e compariso n also apply to two- dime ns ional inlets. The de Sig na ti on e such as 100-0 and 6 0 -4 0 indica t e th e amount of- su per s onic floW area contraction tha t o c cur s ups tr eam and downst r ea m of the i ' cowl lip. Therefo re_ th e 100-0 i nl et has all tts su pe rso nic flow ar ea contt,action i_ _ ahead of the c ow l lip;, while the 20- 8 0 inlet has o nl y 20 percent of th e flow a rea co n - iI trac ti on occarring ahead of th e cowl lip, with 80 per cent c _ th e c o n traction occu _ r- ii ing inter na lly.

The Job of th e su per soni c inlet is to diff n se th e high Ma ch number p f ree- str eam i air to a n ea _ - s o ni c cab l e at th e th r oat, where a nor m al shock ca b efficiently t _ ms - i_ .

, fer th e flow from low super so Mc t o hig h sU per sO ni c speed S . Downstr ea m of t h e ii_ throat , th e flow is furth er flfffuS _ l to th e low Ma ch na lnber required by th e ei _ , r in e !_ a t crui s e. All the su pel' soni c diffusi o n m u s t be acc o m pli sh ed by tt i rni _ the flow. i Ther e fore, for th e all-ex _ er na l-co m pre Se lon i nl et, or 100-0 inlet , al l th e flow td _ n- " i n g mu st b e away from the inlet axis , leavi n g th e fl ow a t a high a _ g l e at th e co w l i ip.

A high coWl llp a n gl e of 22 0 i _ req u ir ed to c _ pt u l _ all the flow and turn it b _ ck to - i_ ward the e n gine. _ lli s h ig h cowl ai s le pr0dtt ce s h igh drag ( r M. 4), _ I Wi th 40 per _nt of th e supe rsonic ar ea con tr action dOWastre _U n Of th e c _ | li p _ _ as in th e 6 0-40 i nlet, it is pe_s ible t o red u ce t he in tern a l cowl an _ le f ro m _ 20 to 5o i th us red u cin _ th e cow l drag. This cow l a _le proflUees a _ aoder a tely s_ r ou g obiicfue ! sh ock in turni n g the flow ba ck toward the e _ ne _ but h )t al -pressure l 'e co ve r y is I_ Still high.

It th e int e r na l co ntra ci ion l c further i _ creamkt to 60 or 80 p er cent, by tisi _ tg the i_ : 40-60 inl e t (refs. _ and 6) or the 2 0- _ 0 in let, a 00 in te rnal cowl _ le will i _ r nd uce _ _ a w _ ak , highly efficient , internal oblique shock s _ stem and w ill provi d e a l oW -drag _ cowl. ', _ ring tl'an so nic a e c e lerat[0n, th e r eq uired engine fl0w _s tw i ce th e f _ b Wre- _ q ui r ed at su pe rso n ic e r _so . Therefo r e, the i _e t throat e Lrea m t is t be doub | ed dur- v 2 S4 , l ' q l n _ t h is accele r atio n . F or the inlets wi th 8 0-percent inte r nal contr a cti o n or l essj t h e ce nt e rbo d y m ust be co l lapse d to double the thr _ t a r e a , _ indicat ed by th e dashed lines i n fig ur e IX- $ . How e v e r , by u s i ng 80-p e r ce nt internal contra c t i o np th e cent e r _ bo d y c an be made sm all e nou g h _ s o t h at whe n i t is translat e d f orward _ th e _ m n ular a rea between t he cen t erbody ar i d the cowl lip w i ll p r ovide th e tr _s onic e ngin e flow.

ThtS provide s a me chanically simpler c en{ er b ody with a pote n tial we ig ht _ tdv an t age .

A n o the r eff ec t of in c reas it _ th e in ter nal co n t ra ctio n fr o m 40 percent h _ 80 pe r - cent is that th e i e n _ h of the intern a l su pers o ni c dif fu ser in c re a se s. A s will be _ seen later , thi s affects th e i nlet ope r ating c ha racteristi cs .

S The perfo rm ance of thes e fillets is co m pared in f igure IX - 4 . This figure pre- :! S e n ts the tot a l-pre s sure recovery as a f unction of t he diffus e r ma ss-flow ra tio. The !

I diffuser mass-flow ratio includes bo th t _ y pa ss and engine flows. A value of I. 0 cor- i, re sponds to th_ inlet cap t ure mass fl ow , and th e difference betwe e n any curve and I 1 .0 is th e boundary - la yer ble _ d f l ow. Inlet ope ra tion w ith the ter m inal _h ock at a downStr e am su pe rcr l ti ca l position corre s ponds to the vertic al I ow-re c 0v e ry portion of e _ ch curve. As the ter mi nal Shock iS m ov e d upstream int o th e throat r egion, where th e bo un dary-l a ye r bleed is lo ca t ed, both th e to tal- pr e s stt r e r e cov e ry and the t bleed flow increa se . If the ter m inal Sh o ck i s moved too far into th e superso ni c reducin g pressure recovery. This transient is called inle t un _ tart. T herefor e_ th e 1 / diffu se r, it becomes unsta b le and w i ll pop out in front of th e coWl lip , drasticall y left limit of e a ch curve for inlet s with internal conh ' a e t to n co r r e spo n d s to the un- i S tar t lindt. The all-ex te rnal-co m pre s sion i nl e t h a s n o u n start , but e ncounters an inlet instabi li ty called buzZ that dete rmi nes th e low m a ss . .fl o w limit o f i ts curve.

• It _ is difficult to obL _in an exact co m pari s on of inl e ts b e cau se th eir pe Ho r mance ca rl change wi th d if f ere n t bl e ed sys te ms (ref. ? ). Ho w ev e r , th e b leed sy s te m fo r each inlet pr e sented h erein provided reasotm_.blefl0 W d istortio n a t the engine face an d re a sonable g u St tol e rance. From figu re IX-4 it can b e s_ e n th a t, i n general , t as in te rn a l contr a c ti on i s increased from the a ll - _ t e rn a l- com pression 100-0 to the 2 0 - 80, bleed flow Increases and m ax im um to tal- pr e S sul'e reco v ery incr ea ses. To o bt a in a better co m parison of th e inle t perfo rm a n ce, a co m par i sOn was nlade for I th e inl e ts o pe rating at th e condition s in d icated by: th e s laSh mar ks o n the cu rves.

The S e conditions were cho s en a s fa _ from th e ttnstart or buz z li m i t as possible i w i th o t lt s a crifici ng a si g tlifi ca n t amount n f nit 'c raf t ra n ge. This C t) mpa r is ofl i s s h own in figure IX- _ . _ _ The range d ecre m en t in _ tuti ca l miles is based on th e S ensitivity Val U e S pre- .... : , _ onted in figu r e IX - 1 f or a supor _ onic transpo r t , a nd the differe n c es ifi coO / 1dr y , it b oundaf'y-layer bleed , and p _ eSsure recove r y o b tained fo _ eac h irll e t. Th e 4 0-60 i nl et obtain ed th e lo ng est range a / zd is u_ J 0 das a re i erefice. The 60*40 I nlet had a higher cowl lip angle, a n d the resulti ng drag r educ ed the ran g e 5 _ miles. The o th er si g nif i c a nt rang e l o ss o f g 3 m i l es i s du e to reduced t o tal- p r e s s ur e recovery. T h e to tal range decre m ent i s 125 m ile s . The hi g h c o wl dr a g ._ f the all-external- co m pr ess i on i nlet pr o duc es such a l a rge r a ng e pen a lty that thi s in let is not even c ompe tit i ve. Th e highe s t i ntern a l cont ra ctio n i n l et, th e 20-80 , has s li g htly l e s s t o tal - p r e ssur e r ecovery and _lig ht l y mor e bleed fl o w , for a pe na l t y o f 6 3 mi le s .

Ho wever , the red u ced ra ng e tends to be o ff s et by it s si m p l er tra nsla ting ce n t e rbod y a n d i t s de s ign Math , u m ber o f 2.65 , which i s s lightly higher tha n the 2. [ _ v a l u e f or _ ' th e 40-00 in l et. Therefore , th e ge ner _ tl trettd is th a t the i nl et s w ith hi g her inte r l _ l ii contraction ha ve l ower range decrements. A ra n ge co m pariso / _ cle a rly favors in- _e ts with high internal contraction. _ i '

Un s tart Cllaracterlstlcs i : i

As show n in figure IX-4 , all th ese m ixed - co m pressi on in l ets which afford effi- cient diffusion and m ini mum cowl drag also have an undesirab l e transie n t c ha r a c- ter i stic kno w n as inlet u n sta rt. This ti n s tart causes a sha r p red u ction in mass flow a nd pre s sur e recovery a nd a lar g e i n crease in drag. An u n start can resu l t f ro m a n inter na l disturba n ce such a s a reduction in th e engine airfl ow r _ quire m ert t o r from a n e xternal disturban ce s u c h as a guSt. _ An inlet uns ta rt for th e model installati on shown in f i gu re IX-6 w as recorded On _1 : high-speed schlieren film. Selected frames from this film o f an un s tart that was I_• caused by an internal airf low di stur bance are p l 'esented in f igure IX J] . Fo r th is uns ta rt , th e 40-60 inlet that wa s i n s ta lled W i th a wing si m u la tor was ter m i na ted by ti a coldpipe a n d choked plug. The part of the instal la tion that was visible i a tlie schlieren f or st a rted inlet operation is outlt n_ i by th e upp e r recta ng le in fi g - ure IX - 6. Sketch e s of th e m axi m um upstrea m unstarted sh ock patt e rn a / ld th e ua- t started stable sh ock pat t er n are al so show n . Time in seeo t id s after inlet unstart for th e various m o ti on - pictu r e fr am es is listed on figu r e iX - 7. Flames tip to ,: 0. 01637 second after unstart show the initial exp u lsion of the inlet sh ock sy s tem.

Figu r e JX-7(f) sh ow s th e maxi m um ups tr eam positio n to which the shock sys te m is ., : expell e d . The oblique sho ck wave from a b ou ndai ' y-laye r s _ pa r a _ / o n on th e cente _ - , b ody oscillates a t a frequ e ncy of a b o ut 130 he r tz. Franies which sho w the , dow n strea m - to - u pstr e am mo ve m e nt of th e o b lique sho ck for o n e - llalf o f a cycle o f .....

, th is oscillation are presented in figures IX - 7( g ) to (i). This ifllet aisO has a large ; i n stabili t y , or b uzZ , Which has a freq u eflcy of about 1 _ hertz f o r the coldl)ipe te _ - .

ru i n ation. The buzz f / 'equet l cy f or this inlet ter m i na ted by a n engi t ie i _ , s O n _ ewhat higher , about 22 hertz. The boundar y- layer separation m oves up s t r ea m t o t ile :: , 28 6 spike tip for t he forw a rd extent of the inlet buzz cycle.

The perfor m an c e dur i ng an inlet u n a t e a 't fo r a f r ee- st rea c n Mach n u m be r of 2 .50 is p res en te d in figure IX - 8. Variation of the total-pr es sure re_ overy mea su re d at the co m pre s sor fac e as a function of ti m e aft e r un star t is prese nted. A 40-60 inl e t ,!

' unst ar t from a n i n te rn a l disturban ce i s ill us t, a t e d by the da sh e d tra c e. T h e to ta l- p r e sets r ec overy drops ra th er r ap i dly, from a s ta r ted l e vel of abo u t 9 4 p e r ce nt to a m ini m u m le vel of about 2 5 p e r c e n t. From thi _ ; m inim um l e v e l th e 40-60 inl e t is u n stabl e . A spik e and overboard bypass door movement are requir e d to s tabilize th is i nlet at a n u n S tart pressure r e cov e ry of 48 p e rc e nt. The (J0-40 inl e t tlnstart from an internal distu r bance ha s th e sa m e initi a l trend a s th e 40- 6 0 inlet un start, a n d dx'op s to a m i nim u m recovery of abou t 16 percent. The 6 0-40 i nl e t , however, recov e rs ra th er rapidly to a high r e c o v ery and is a u to m a ti c a ll y stable at an u n - star te d stable recovery of ' / 5 pe r c e nt. De pe nding on the cause of un s t ar t , the _ ,.op , L in total-pre s sur e r e covery im me diately after inle t unstart can be qt _ ite diff e rent.

, This i s sh ow n by the trace for th e 60-40 u n s tar t from an ext e rn ai di stu r banc e, which dr ops to a m ini m u m l e vel of o nl y 68 pe rc e nt r e co very, and is a u to m atically stable a t 7 5 per cent recovery.

_,i A large dr op in pr es su r e r ecov e_ ), i mm ediately after unstart m ay cat L s e th e i / engine to stall and flame out. These data a re p resent e d in the pap e r on effect s of enghle i nlet disturbances. However, ff th e unstarted inlet has been s tabiliz ed , and if th e e ngine is runni ng, a pr opul s ion system with the 60 -4 0 inl e t loses about 50 per- _ :ent of its started thrust. This p r olk l lsio n syst em with th e 40-60 i nl et loses all it s , _ th ru _ and ends up with a d ra g compo n ent. No un S tart da ta are availabl e f ur the , _ 2 0 - 80 inlet. H ow ever, this inlet would have a perfor m_ Ulce pet _ dty _t least equal to th at o t the 40-60.

i While the larger internal ar ea c o n tractio n inl e ts hav e bett er rang _ charaet er is- .

_ ticS, th ey also have a larger tmstar ts d pe rfor m ance pe nalty. Since a typical supe r- sonic cruise aircraft would i ncorporate mu l ti pl e propul si on syst em s , t m syr a me tri- ¢ cal fo r ces du r ing a si ngle i nl et unstart ca n ca u se airc r aft contr ol probl e ms until the ' it _le t has been re starte d .

Ang l e - of-Attack a nd Ma c h Number Tol e rance . q aract e rlstlcs

_ , One possible S olution to the unst a rt proble m of intex, n ttl cont r action i nl e ts i s to design th e inlet tOaccept di s turban ce s wtthodt tlnstarti ng . Fir st , consider t h e ex- _ ternal di s turbances. These disturb a nce s ar e th e re SU lt o f air( _ r af t mane u vers attd at a ttosph e ri c gusts. The ai rcraft / _ tan e tt var i s g e n _ ra U y sl0W 6 ri ou g h that the i nl et ' s control syste m ca n anti c ipate and c o m peft sa te for it. £hl t gu _ e ar e a s u dde _ oc cur -

t

2 8 '/ re n ce and the i nl e t co n tr o l s r n _ y n o t be rapid en oug h to prev en t the inlet u nstart.

There f ore , w i th i t s geo me try fixed p th e inlet m u s t ha ve a deq ua te to le ra nc e to w i th- s t an d the sudden gust s . Gu s ts ca n approa c h fr o m all d ir e c tio ns a n d affe c t the angle o f at ta ck or y a w a n d th e free- s tr e a m M a ch number a h e ad nf th e inle t (fi g . IX - 9). In- c rea s in _ f ree -s trea m Mach number ha s little effect o n i _ det perfor m a n c e . But when the f ree- s trea m Math nu mbe r i s d e e r ea s ed _ the Mat h nu mb er s ar e redu c ed throu g h- out the super so nic diffuser and in the inlet thr o at. L . ,r ge e n o ng h reduct io_s i n the f r e e-strea m Math nu m ber will c h ok0 the inlet t h roat and c _ use an in let unst a rt.

Thi s C hok ing c an b e re li eve d only by geo m etry variat i on.

When the in let i s forced to oper a te at a n an g le of at ta elt , t he inlet ca n u ns tart in two ways. If the ter m inal shock is i n i ts most f orward stable po s ition at 0° a ng le of attac k , a s li gh t increase i n th e angle of att _ k is suffici e nt to force the s hock forwa rd !

into t h e sup e rsonic diffuser, where it becomes unstable, and th e inlet unsta _ : s .

Operating the inlet with the ter m inal shock s lig htly downs tr ea m of the diffus e r throat _ (supercritical operation) allow s the inlet to operat e _ t hig her _ les of attack. BU t !

aS the angle of attack is IncreaSed , additional compres s ion of the int e r na l sup e rsonic _ flow appears on the top, or leeward , side of th e In let an d less compression on the ! * botto m , or windward, S ide. Th e lack of co m pr es sion o n th e windwa r d s ide resultS !

from eXp _ mdingthe flow around the oowl lip i n ord e r to tu rn th e f l ow axia l ly . The I ov e rcompression on tile leeW:trd si d e is th e c riti c al regi o n wher e a local cholting ii con di tion cause s th e inlet to unstart, l _ igure IX-10 ill us trat es in more d e tai l th e _ _, local chokin g that lim its th e angle of attack for s up er c _ iti ea l op s ratiorl. The _ p a ide _ : " of the 40-60 In let with the co#! and cente rho d y co tlto ur$ is showft in th e sketch a t the _ d t op of the figure. T his iS the le ew ard side of th e i t det w hen i t bp e f _ at es at a positive angle o f att a ck. The p lo t shows the cowl S tati c pres s ure ratiOed to f r ee-strek m to ta l i p rea sure as it var ies a lo n g th e cowl sur / ace , and identifies t h e shock : 'eflections on , the cowl surface. _ . _ T he solid lines re p r esent operatio n at 0° angl e c _ _ ack and t h e da s lled li n es at _t th e angl _ o / at ta ck J us t befor e th e inlet unstart s . As th e ar _ le o f att _ t ck is incr _ tsed _1 _ f ro m 0u to th e m axi m um value Of 2 .9 ° , th e first sho ck reflectio n ott the cowl lnove _ for w ard in th e inlet. The second cowl sh ock refl 0 cflo n moves a h ead of the th roat : region to th e fo rwa r d ec _ e o _ th e po r ous bleed and c o m presses th e 'flow to pr e s- i!

sures h i ghe r than S on i c va l ues. Th i s indi ca tes a l o cal eho l dn _ co n d ition at t hi s po i n t o n _ e cowl. Further inc r eases i n the angle o f atta ck wo u ld unSta r t th e inlet. This _.i_ i_ m eans that thi s second s h ock ref l ectio n can n ot be mo ve d al t t _td of th e cov / i ble Od r e - • g le n . l _ y m ovi ng th e porott s bleed z'eg l ons furt h er u pstream , a h i, her an g l e of i atta ck Of 3. ,, o was reached b ef o i 'e an u nst _ trt oc c t i rred .

Figure I _ -11 shows th_ unst b. rt ilintt s for the 40-60 tfllet f01' both a reduction : in f r ee-s tr ea m M ach number and a _ l in c re as e in aflgie of attack. A _ o, data are \ s h o wn fo rtw o diff e rent ble ed c o_ t gu ratl ons . Th e da ta I ndi cat e tl _t , w he n th e b l ee d i s re l ocat ed t o an up s tr eam po s iti on , lar g er Ma th-nu mb e r - reducti on and at l gl e - of- * atta c k toleranc es are r e ali z ed. Also _ a s th0 fr e e- s trea m Mach number i s reduced with fix e d i n let g eom e t ry, the a ng l e- of-attack to l e rance d ecr ea s e s to zero where the m a ximum reduction i n fr _ e -et re am Mach nu mber occu rs . T he in let un star te beyond th ese condition s or to the ri gh t o f thes e tw 9 c u rv es .

F o r s o me tra ns po r t co nf t g u ratio ns_ th e a c c ept e d ml_ lluu m t ol 0r an ees ar e ab o ut - 0. 0 9 in f r ee.s trea m Mach nu m ber a n d 2.50 to 3.0 ° i n ap p le c _ at ta ck. Howeve r , _ more tol e ra n ce f o r bo th th e se para m eters i s desirable to i n c r ea s e the i n let sa f ety !i m argin. AI L o p l a rger tol e r an ce s i n these para m eters may be d e sir ab l e f or e th er s u personic dash or cr ui se a ircra f t.

The tr end in the data presented here i n f or th e 40-6 0 inlet i s typical of the da ta th at were obtained f or th e o th er i nte rna l co n traction inlets of this com pari s on . All th e inlets in th is co m parison have adeq ua te tolerance to M a ch n umber red u ctio n s.

, However, the angle-of .attack toleranc e becomes more of a concern a s tile t nl e t *s inter na l supersonic area contraction 1$ inc r eased. This is-shown i n f igur e IX.1 2, which Summarizes the ang l e-of -attack expe r ience fo r th ese inlets. The le ng th of th e su personic dif f user, from the cowl lip tothe inlet th roat for th e various inlets, was th e para m eter examined. This figure p r e s ents the unstart a ngle- of -atta _ li m it I as a function of th e length from th e cowl li p to the geo m etri c thr oa t. This length was _ ' nondime n s iol _ a li zed by the cowl lip radiu _ .

_ The values of this ratio vary fro _t 1.88 fo r the 2 0-80 i nle t to 1.19 for th e 60 .4 0 inlet. T h e shaded area represents the rang e of a ngl e -of-at ta ck to lerance fo r the i nl ets considered so f ar with g ood pe rfor m i ng ble ed sy s te ms . The upper portion o f this shad e d area is a curve faired th rough the data po int _ Which re pr esezit the maxi- ! : mu m u ns tart a ngl e of a t ta ck that h a s b e_ n d em o n strated so far for each inlet. _ l ' h e curve h a s been exte nd ed to include da ta of a n additional 6 0 -40 inlet that h _ ts com- ,, t parable pe rfor m anc e to th e previously m e n tione d 60-4 0 i nl et. Figure IX-1 3 coin- pares th e two 6 0 -4 0 inlets. Th e top sk e tch is th e i nl et C onsidered p r eViou S ly. Thi s i nl et distribu _ ed th e tsefltropic co m pre ss ion from th _ cOWlover so m e l _ flgth of centerbod y . Another way of d _S ig _ li ng th is inlet i s to focus thi _ is e ntr0pic cOmpreS- siOn at one point on th e centerbody , reducing th _ leng th f r o m the co , el lip to the geO- metric th roat. This r 6 sul ts in a l eng th ratio of 0. _ 6 for th e _ ocu se d d o m pres s iofl ' inlet. Thi s focused compression inlet has a m_ J _ utn _ le-o _ -at _ck to lera n ce of 9.4 °. The tole r ance in angle el attack for th e i nl et s _ ¢ ari e df _ om ab out 2.6 ° _ or th e i ' , ' 20 4 0 I nl6 t to about 9. 4 ° f o r the 6 0 -4 0 f ocus e d c o m p re s s io n ifllet. Th e_of ore, in- c r e as ing the le ngth of the I n_e rn _ ! portion o f t h e sdpm ' so ni c diff u ser ap pea rs to re- duce its angle-of-atta ck to leran c e.

, An o the r param e t e r th a t must be e o u slUe red l s the di s tortio n a t a ngle-of. a ttack o p e r a ti on . T h e di s torti o n i s d e f in ed as the a v e r a [ _e t o t a l p r es sur e m inu s th e mini- : mum t o tal pressure diVid e d by th e a v e rage tot a l pr essu r e _ ttth e compr ess or f ac e.

For th o se i nlet s th e di s t o rti o n w as bel o W 0, 1 0 for ang l e s of a tta c k l ess th a n 5° and below 0. 13 f o r a ngle s of a t ta ck l ess th an r / . 5o. Di Sto rt i on valu e s of 0.06 t o 0. 0 _ / a re n o r mal ly a_ eeptabl e for 0 ° - an_ l e -o f - a t tac tt op e ration.

Thus , the inl e t' s tol e ranc e to external di s turb a n ces I s li m it e d by choking in th e inlet throat. This _ o l er _ tnc e ca n b e improved by prop e r loc a t i on of the porous bleed.

Also th e inlets will hav e le ss ang l e - of - att a ck to l era n c e a s the su pers o nic dffius e r length I s increased. '_ ,' t Summary - Effect ofAmounLof Internal Contraction ( , ! .

The prece din g co m pariso n s do not produce an ob v io u s choice. Range is rela - i tlvely easy to evaluate and clearly favors hi g h - i n tern al - contraction inl e ts. Ho w ever, I_ th ese inlets are more sensitive to gust and maneuver di stu r b a n c es a n d also produce il a larger disturbance at u n star t . It th en becomes a pr oble m of choo s i ng the maxi- i mum range compatible wi th acceptable toler an ce to gusts and m arteuvers and also an _ acceptable unstart amp l itude. The trades have bee n fairly well d ef ined , b ut th e air -

ii

er t d e g.e r beg t , en thed ieultc ho ice .

4 _ [

IMPROVED STABILITY OFINTERNAL CONTRACTION INLETS "

Inlets wi t h d i ffere n t amountsof interna l su p ers . q l d c co ntra c ti on ha _ , , e been c o m - pared. However, all these i nlets are about eq u ally _ e n sitiv e to i nt ernal di s tu l '- _ t bances. The conventional m ixed-co m pression inlet mu st operate w ith the ter m inal , • shock near th e throat region for peak p _ rfor m ance. I _ this shock inad, _e r te r i tly _ I ' I 4 m ove _ ahead of the throat due to an l n t e rt _ l airflo w dis tu rba n ce , unstart w ill o _ cur, t!

Current inle ts have a rath e r limited Stabl e Ol _ral i ng ran g e th at i _ p r ovi d ed by the capac i ty of the perforll _ ance bleed sy s te m to spill increased airflow aS the ter m it _ i _ : _ ..

sho ck moves u ps tr ea m . This li m ited stable _ ang e m ay n ot b e ade q u at e to ab _ o _ many o t the transi e n t disturba n c e s th a t ere en co u llte r ed by a su ps rdo n i e propulsion syst em . Larger sta b l e margins are c u rr e ntly p r ovid e d b _, i _e t operatiol _ a _ low per - formance levels. ., _ , Th e se itll e ts ca n b e de s igued to prev e nt unstart for m o _ t l n te rn itl dis turbanc e s by utilizing a th roat byp _ tss b leed syste m (refs, 8 an d I}). A s cherdatie of this sy s- tem is preseflted in figure IX - 14. A n iritern al dis tu rb _ tnce , s u ch as a mom entary i \ ( t b red u ction lu the a mo unt o f airf l o w that i s r equi r ed by th e e_ i a e , c au ses the t erm i _ _ .

shock to move up s tr e a m . U nsia rt c an b e prevented ff the e xc eBs ai rfl ow C anbe di D sor te d overboard thr o t / gh a large o pen b l eed or byp as_ i n th e i n l e t throat r egion , Th e exit ar e a f o r thi s bl oc 0 rc z ion mu st b e contr o lled t o r e gul a te the a mount of air - flow s pill ag e a s the ter m i n . _ tl shock mo v es to a different l o c ati o n , Thi s contr o l pr o - i ve n t s pr o hi b itiv e a m ounts o _ ai rf l ow _r o m b e i _ du sted o v e rbo a rd duri n_ n o rmal i n - let op e rati o u _ but a ll ows l ar_ c a m o u nt s of bleed to e xh a u s t in o r de r to p reve n t ni t- s tart a s th e t e r mi n _ ! sh o ck moves u p s t re a m o ver th e bleed reg i on. S / x t e en f a_ t - _ , act l _ valv es were located in th _ i nlet cowl to prov i de th i s exit-a r 0a c o n t r o l .

Two type s of ble e d centre ! v a lves that wore i n ves t i g ated are pr e sented i n f i g- ure IX-IlL The mechanical valve i s e s sen t _ lly a f re e -p iSton tht.. i s pos itio n e d by dif f er e nt ia l p r e ssures. An t n ter n e.l re _e r e n ce p re ss ure , in _ i{ie th e valv e c ha mbe r , i s adjusted to J u st cl o se th e Va lve d u r ing no rm al i nlet o p eration. Cont r ol of th i s i n - ternal refe r e n ce pressure deter mi nes the p ressure at w h ic h th e va lve openS. 1 _ , th is reference pressure is exceeded by an inc r ease i n bleed plenum pressure that r e sults fr o m a fo_ tva r d movement of the ter m inal shock over th e bleed re g ion , th e valve opens a nd allows bleed flow to occur. Thi s g elf- a cting mech a nical va lve w as a simpl e design that w as i nv e s U_ate d to de m ons tr at e th e concept of a c o ns tant- pressur e bleed co nt rol.

For th e vortex w tlv e , a tangential c o n trol airfl ow c reates a _ ortex inside the _ va lve cha m ber which shuts off the ble e d at th e l ow er pre s sure s that ar e associated i , _ w ith n or mal inlet operatio n . An increase in bleed pr es sur _ br L_ ks down thi s v orteX , " al lowing th e bleed a i rflo w to exhau st th r o ugh th e va l v e exits (t o p and b o tto m ). Th e _ m axin _ u m amount of airflow i s li m ited for th es e v al v _ s _ since _o r pl' o per o per _ tti o n '!i the siz e of the exit m u st be small w h e n co m p a red to th e si z e of th e va lve chamb e r. _ .... Inle t pe rfor ma nce for three dif f e r e n t types of s th b ility bleed controls at a fre _ - i stream Mach nu m ber of 2 .50 are p _ es e nted in f i gure IX-1 6 . Suf f ici e nt psi'ref ina nce : ii bleed was removed f r o m th e cente r body to a ssur e high inlet pe _ o r man c e. This ii i_' ble e d W as dttcted th ro u gh th e c e flterbody _ uppo r t struts and c o nt r olled by f / x e fl exits, i " In this fi gu re t ota l-pl _e s Su re r e cov e ry is p resented a s a f unction of difft _ se r m ass- l : flo w r atio. Inlet perfor m ance is show n f or ftXed-ex / t bleed controls that _ ,re nor- mall y treed on m L _ ed-compreS$ion i nl e _ and |or vort eX and mt _ chant ca l v _ l _e s. The ] i stab l e m argitt th at each o f th e s e s tability b l _ d contr ol s p _ ovide s ca n be i llust r at ed _ • ff a _ i nl et-e ng i ne m atc h condition of abo u t 90- pe rcent r e covery is s ele c t e d. F r o m l th is co / _ditton _ th e f / x ed -e 0 _ it bleed control p_ 'o _ ,ides a stabl e / nar gi rl _ f §.3 pe _ ' _e nt | • th a t i S , th_ co r rect e d ai rf low i n th e in le t difftis e r can be reduced by _ . _ pe rcent _ _ pri o r to i nl et un S tal't. The V ortex valv e s p r o vi de a n in c re a s e d s _ able fi _ t t gt n to _ 10. _ Jpel'c e nt. Me e hanical va t V es w lgc h we re a _ iinii ar physic a l size to the vo _ tek valves p _ ovided a s tabl e fl _ a _ gi n of 2 / } percent. T h ese data show tltat it is pOssible _ to p ro vid e a n l n e r ea e _e d to le r a nc e t o i nter nal disturb & n ees for m_e d-e om p re e , si o n inl e t s by u tlli _i rlg a th r oat bypa ss atab i_ ty bleed syste m a nd a v ar i a b le e xit control,

DISTORT.ION MEASUREM ENT AND CONTROL . , !

( On e p r obl em t ha t i_ co mm o n to l ilt th e I n let s Is di stor tion, TEI _ i _ • r _ the r J em ot i onal e ubJ ec t p pa r tic u larly wh e n a i r cra ft and e ng i ne d es ig ne r s are att em pt l_ to decid e th e res pon si bilit y f o r propul s ion sys t em st a b i lity. It i s ni x ie a V e ry diffl oul t e_ | ee t b e cau se tim e -var y i ng d i s tortio n h as r ece ntl y b e en r e_ .ogniz e d.

To d e fin e the tim e .varyi ng distort i on a t th e e n g in e _ a ce r e q u ir e s Ord e r s of mag- nitude mar e data than is required t o d efine th e old er s tea dy-sta te or ti me -ave _ag ed i distortio n . How to process this additior _ t infor m ation into an un der s bxn d ttbl e form _ is not well defined at present.

Distortion Measurement

Distortion measur e m e,tt i s revi e wed tn f igure IX-I' / . q _ hetime-honored m ethod I I to m easure to t e d-pressure dis tor tion has been to install a l a r ge number of pr obes at i,_ the compressor face ( as shoWnin th e upper right e ar n er of f ig. lX-l' l ) p each w ith a I !

lo ng transmissio n li ne to a remot e sen sor. These lo ng lines ti me -averag e the da ta. _, A pre s s u re contour m ap collst ru cted from th ese steady -sta in pre ssures would look _ li k e th e ex a mple sh oW n in the low e r right. E _c h contour i S a l in e of cortst a nt totat- pr e s _u x'e recovery..In this particular e _ _a m ple j th e i n let w as opel'til ed at a positive angl e of att a . ck, B e that the di stortion was pritn ar ily ci r cunfferential _ wit l t high _ i_ presst _ e in th e top portion Ofrite inlet and the low Pressure in th e b 0tton _ po rtion, l t There is also a radial component pr e_ el _ t, l Re C en ti y j higlt-respo n se subtnittiatu re pre s sur e s e_ sor _ ha ve _ epht e e d the long tr ansmissio n lin _ s. With t h ese sensors, it has been foultd that there i s a fldctuat- !

i ns o r dynamic c ont po nent of total pres su re superim po sed on top e l th e dteltdy-State value, a s shown t n the lower left corner of fi gu re IX-I ? . The inst _ antaneous pres- sure Pi m easured by each s e n s o r is _ en t he sum o _ the _ te l tdy-st a t e _ r es sure Pi pl u s the dynam i c c b mpo nent APi. These fluctuaUonS C an _ Oe the regdt of one O r : more flow conditio n s w ithin t he inlet. T e r m inal shock boundary- la yer interaCUo ns_ transie n t bottndary-la _ er el eparttti0n s, d uc t re s onance s, a nd engine dynam t t _ feed- b a ck arc a fe w of th es e po s e Lible cau S e S . Wh a tev e r th e i_ tus e , th ese flti _ tuatiO n s do exist , and th ey combi n e w i th the st eady-statd pres sur e to produce a d ifferent dis - to','tton pa ff_ rn at each instant of ti m e. And since _ha t th e an il ine i' e ally _ ibe s are 29 2 t \ th ese in s ta nt an eou s d / _ t orti on s , a c o mpl e te I n l et- e n l_ in e co m l]Itlbil i ty e f f o r t-requl res t he i r meas u r e me nt.

Th e b es t way tn ge t a fee l for ' , . _ t is actually g oir _ On a t t h e _ ompre _su r f_e e i s to m ake a mo tio n picture from a _ o n tl n uo us s e ri es of i ns ta n ta n_ u s p res su r e c on- tou rs . The s t _ p s Wh ie h a re follow e d to p ro du oe _ u c h a m ot i on pi c t ure a r e a s follow s : Fir _ t p th _ i n sta nt a n e o u s pre ssu r e f ro n _ e ach _ enm _ r is recorded o n analog tape; tl _ en , the pres su re l eve l o tt e a ch _e n so r at a g iv en instant i n ti m e i s de t e r m i n ed b y _ t mul_ ta n eou _ l y dig i tiz i n g al l c l_ a un _ls from t h e a mt iog t _ po at 8000 po i nt s per s eco n d per ' ch a n n el; then , f or e a c l _ time slice , a p r es s ure contour m ap i s con s tructed us ing a c o m put e r pr og ra m ; finally, cartoon m ethods ar e u s ed to m_ e the m otion pic tur e , w her e i n each contour is re pe at e d s ix times and projected a t 2 4 fra mes per second, A motion pic t u r e w as ma d e u s i ng this procedure; selected f ra mes f ro m it ar e shown in figu r e IX-18. Th e cont o ur in tile u pp e r l e st is the st e a dy-stat e p _ ttera and iS for th e s a me data point as was shown ia f i gure IX.1 7 . In the r e pr e se n tatio n used i n this fi gu re _ each Shad e d a r e a r epre s e n t s a 6-pe r ¢ _ e n t recovery r ange with the U ghtest area being th e highest recovery (94 to 1 00 l _ rcent) and the darkest area the lowest r e covery (64 perc ent a nd b e l ow ). Th e boundary betwe e n any tw o s hade d re o glens is a consta n t-press u re con to u r . When view ing t h e m otion p i cture, th e first i m pr e ssion is that th e r e i s a g r i t deal of movement i n the pressure fi e ld. After repeated Showings, it becomes app _ t re nt that the ove r all motion is du e to th e expa n - i sion and contract i on of th e lar g e r e gio n s of high and low pre s sure. T il e movement _ I _ , of the se regio n s is then the cause of th e f lt _ ctuatio n s a t any o n e po i n t. By co m pari ng : the frames at time insta n t s 1 tO $ in figure IX -18 , th e variations i n size of th_ dif- I_ f erent pressure r e gions can be se e n. ii i From th e nlo ti on picttire , it a lso _ tppeared as t / th e a m plitude of the flu ctu ations i .... was not constant across th e co m pr ed sor face. To better dete r mine this , th4 }amp li - _ rude of the dynamic co m ponent at each sensor can be found. As a m easure o _ th e il amplitude o _ the fluctuating co m ponent APi , th e root- mea n-squar e (or RMS) lev e l t t A PRMS i S u sed, as i nd i ca ted i n figur e IX -I '/ . The RMS le vel of ea c h Se nsor w as _ ratioed to th e st eady- s tate c o _p r ess or face pr e ssur e a n d the re su lting value was used to constr u ct a dyna m ic contour m ap.

FigU r e IX-19 s ho w s the steady-state at _ l dynamic contours for th e sam e data / po int as th e instantaneous distortions of figure IX-18. These d _ ta _ ere reco _ ,ded in the 10- by t0-Foot S u per S onic Wind Tu n ne | Wi th the 6 0-40 inlet installed in front of _i " '_ the J85 tu rbojet e ngine. ( _ teady- _ ,ate interactio n s b e twee n t h e 40-60 i nle t mid th_ , J85 tur bo j e t engine ar _ disc r ibed I n l,e _ . 10. ) Th e da ta at f i _ ureIX-19 We re re- corded at Mach 2.6 and 5° at _ gl e of attac k w i th th e il,,let o perati ng $om e w. _ t stt _ r- _ ritically at a steady- s tate p r essure recovery o_ ab o ttt ' _ / pet _ eeht a r id a ste ad y-state i| distortion of about 18 percent. The st.ea dy-stat e m ap used the sante rePi'e S en ta t ion _ gS : \ as in fi b re IX-16, e _ cept the nu mer i ca l values of th e r e covery bou n d a r ie s ar _ given. Ag a in , notice the co m bi n ed circun _ fere nU al and radial pattern. Now, o n the dyna m ic map , each shaded region repr e se n ts a gi ven range of dynamic activity, Here the darkest region in dicates a high dy nam ic level of between 6 and _ perce n t o f the st eady-st ate pr e ssure, and th e ltghteSt region a very low dyna mi c lev e l of be- tween 2 and 3 per c ent. For both the s te ady-state and th e dyna m ic plots , the n , the light areas represent g ood flow and th e dark ar e as _ bad flow. Com pa ri ng th e two maps sh ow s that i n th e re g ion of high pr e ssure r ecove ry the r e is a l ow RMS level.

In th e region of low pr e ssure recove ry the I _ MS level i s So m ewha t higher. The t,e- gio n of hi gh dynamics , th en , lies in a band lo cate d between th e high- a nd I ow - pressttre - r eco very regions.

Si n ce the peak - to - peak level of the fluctua ti ons at any one spot is also a mea s- urement of the dynamic level, we should be able t o correla te the h ig h dynamic r e - gion on the P _ S plot to th e s pa tial area in_the mo ti on pi ctur e which has th e g reatest number of pr essure boundary crossings pe r u ni t of time _ _ From the m otion pictur e it was found that in th e upper high- pr essure regi on there were very few boundaries Crossing a given spot. In th e high dy na mic band there were at lea st four different boundaries pa ssi ng ove r a given spot rath e r frequently. In the low-pressure region, it was pe s si b le to See t wo diffe r en t b ou ndaries c r ossing a given spot frequ e ntly , wi th two others sneaking in oc casi onally. It must be r em embered while m aking this c om pari s on that th e RMS plot was an av e l _ ge over a lo ng time pe riod, whe r eas the motion pi cture covered only 1 5 mi lliseconds. (Two seconds of f ilm ti m e eq u al 1 msec of data time. ) By com paring th e inf or ma tion g ained from th e m otion pictttre t o th a t gait l ed from the I _ MS pl ots , i t appears that th e h igh _ mnlic l'eg i ons are areas where a greater number of pr essure boundar i es pa s s per unit o t time.

From th e s _ ne da_ t ro m w hi ch the motion pi ctur e was m ade, it is possible to find th e value of a distortion parameter fo r ea ch ins t ant fll time. Figure IX - 20 sh ow s t su ch an instantaneous pa rameter plotted over a $ 0- m illiSecOnd time pe riod ju s t be- fore and duri ng co m presso r st _U . This par am ete r is a t tlnc U on of averag _ arid m ini z nu m pr essure levels an d i S discussed in grea te r deb _ til i n th e pap e r on effects of engine inle t di sturb a nces. T h e steady-state v _ tlue of th e parameter averaged over a m uch longer period of ti me is ab o ut 0.1. The critical l _ r el i s the S tead S - Sta te value of th is param e ter which w o _ il d cau s e e n tw i n e 8 i _ 11. So it is obvious [hat th e stea dy-stat e l e vel is fairly f ar from c ri tical but that there ai' e t _ nsients w h i ch do e _ ceed this critical le v e l. _ , The dat a from _ hic h this pl o t _ vas ob tain ed Were iflter e d a t 1600 he rtz. By proper a na lo g fil teri n g or digi tal a _ erag in g, i t s houlfl be _ ible to d _t e i 'fiflne the highes t frequency content of s ignificance to the _ ngine. C _ )ncepiua _ ly, wflh

l

g th e proper f ilt e r and a valid para me ter , only th e distortio n which cau s ed engi n e stall -,_ Should exceed th e criti ca l level. These me th ods a r e being purs u ed a t the present time. HoWever , for the da ta shown here , the distortion at 10 mi lliseco n ds bare l y ex ce e ded the critic _ tl l eve l, b ut the e ng in e d i d not stal l . The dis to rtion a t 1 8.5 ra i l- 1 liseconds ex ceeded the critical level a nd did ca use an e ngi ne stall. T he e vidence of 1 : the stall ha mm er s hock i S see n a t about 27.5 m i l li se conds. T he ham ma rshOck i s a : , co m pression wa ve which i S sent forward due to th e s tall . The time b e tween 18.5 and 2_ . 5 m illiseconds i s that needed for th e bad patt e rn to move / re in the rake s ta- i ' _ tion in t o th e co m p r essor and have enough time to cause s ta ll _ a n d the n f or th e stall .... hamm e_ Shock to m ove back th roug h the co m pressor to _e rake station, i_ Selected mo t ion-picture f rames shoWi ng the signifi can t occurre n c es of fig- : ' ure IX - 2 0 are sh own i n f igure IX- 2 1. Th e U: e s of th e frames co rrespo nd to th e i i " ti mes of figu r e IX-20. (Fr am e s shown at tim e s of 30.25 and $ 2. 2 5 m ill' . second s are ii_ • beyond th e time peri od sh ow n in f ig. IX-20. ) The upper left contour J s again the i_ " steady-state pa tter n for r _ ere n c e . The f rame s a t ti m es of 18.00 a n d 18.50 m il li - seconds sh ow what ca us ed th e wo rst dis to rtio n patt e rn. Just before _ is worst pa t-. _ : i " te r..u , the l oW - pr essure region.in the bottom of the inlet became lower( tim e , 18.00 m sec) and then the aver ag e pres su re region increased in exte n t down around th e sides of the i nl et (time , 18.50 m sec). At 2 _ . 50 milliSeconds the first evidence of the ham me rshock can be se e n. It appears as a h igh-pr e sSure r e gion at the top of .

the co mpr essor face. The high - pr es su re region then rotate s around {. _ e inlet i n - - ° the direc ti on of l _ )h) r ro ta tion. At 30.9.5 m i ll fscco n ds it is about halfw a y arou nd • e co mpr e s so r face, and a t $ 2 .25 m illi s econds it ha s e m co m pa _sed the w hole co m - pr esso r face.

One conclusio n tha t can be drawn from th e motion picture is t _ at th e h ig h dy- :_ ::.' namic region is not the o nl y r egio n of im porta n ce , as t h e s tan - _ t;odu e i ng m echanis m . ;:' encompassed regions of both high and low dynamic s . I A s stated before, an att e mpt iS being made to defin e th e hi ghe s t fre q uency of importa t l ce to the engin e . It t _ po ss ible tO gain a fe eling for th s hi ghest f requency from f igure IX-20. The fas t est transient Of sl d flc t ent amplihnt e _ .o exceed [he criU '

i

cal level occurred hl abo _ 1.6 milliseconds. If this is ti se d as the periOd of a s i n e wave, the fr e qttency of _ wave Would t _ e 600 _ _ rtz. The , i t wo , _ ld appe a r from ..... : i this li mi t ed da ta tha t freqtieizeles to at le a st 600 hertz are of imp _ . nc e for thi _ particular in le t si ze.

DistOrtion Control '

N ow let's _ e some po s s i ble m eads of dtstoi _ do n eoni r o l . F i gure D _ -22 : shows that vortex gen era to rs are very _f ective in redhcing bot h _ teady-st at e _ tn _ dynamic distortion. T h e pa r a me ter s plotted ar e st e ady-sta te and dynamic distortio n versus total- p r e ssure recovery. Dynamic di s tortion is defined as the avera g e RMS pres su re level at the co m pre s so r face ratioed to th e a v e rage steady-sta te pre s sure at th e compressor face. These data were reco r ded w i th th e 60-40 inlet opera t ing at Mach 2.5 and 0° angle c _ attack, and were pl 'e ft it ered to 1 k i loh e rtz. (Simila r _ ' da ta fo r th e 40- 6 0 inlet are presented in re£s. 11 a n d 1 2 . ) The open symbols i n f ig ure 1X- 2 2 deno te da ta recorded with a cold-pi pe choked plug ter m ination and the solid sy m bols denote data recorded with th e J85 turboje t e n gi n e as a termination.

The da ta on th e l ef t were recorded wi th out vor te x generato rs and th ose on th e right i_ wi th vortex generators. First let's consider til e data wi th out vorte. _ g ene r ato rs . _ Point A is a high- pr essur e -recoVery point _ wi th th e termi na l shock at th e geo m etric i : th _oa t. P oin t B _ is at low recove r y , wi th th e shock pull ed dow ns trea m into the sub- sonic diff u s e r. To p r og r es s fr 6m poin t A to po i n t B, ei th er th e choked plug i s re- - tracted or th e e ng ine spe ed is i n cre a sed. Bo th st eady-state a nd dynamic dis to rt / on inc r ease ra pi dly as t _l e shock i s m ov ed from point A to point B. T h e 0. 1 level of dy nam ic distortion is quite high , as th e pe ak-to-peak amplitude of the fluctuations would be Somewhere between 30 and 60 percent of the a v erage co m p _ 'e sso r face pre s s u re.

T he da ta on the rig h t w ere take n wi th vortex g e n erators install ed on both th e , _ ) ii " c owl an d cen te rbody aft of th e throat region. The purp _ _ Se of the g e nera to rs is to I . _ pull h ig h-energy fl o w from th e m ain duct d ow n into th e low-en e r _ b o _ da r y layer, ii Bo th steady-state and dynamic distortion were r ed uced over th e entire r e cove ry _ r _ tnge. Thus, vo rt ex g e ne rators are a ve r y effec ti ve mea ns o f b ou ndary- la yer con- , _11 tro l and, in tu r n, ar e an e _ f e ctiv e mean s of di stortion co n t e o l . These data also sh ow that the e ng ine had li ttle effec t on e i th er st eady- sta te or dyfla m ic dis to rtio n , ii H ow ever, it must be reme mb er ed that th ese are tim e -aver ag ed da ta. Also , it appears that th e dynamic distortio n p _ rameter can be us e d to dete rm ine th e quali t y of th e dy namic fl o w at th e compressor face , Just a g th e ste _ t d y- I_ t _ tteparamet e r has _ H r been used fo r ye ars to q u alita ti vely d efine the s teady-sta te flow.

Figure IX -23 shows t he result o f a r e so p.ant co b diflon which w _ .s encountet, e d in an early version o f th e 40- 6 0 inlet ( r ef. 11). It was de te r m in ed that tiff s condi- Uon was d u e to th e overboard bypass cavities re sona ting in an organ-pi pe / izanner.

Th e main duct was th en isolate d fro m the b ypas s c _ tvi ti e s b y idst al li ng a d na l- W ne cascade at th e en han ce to each ca vity. The da ta s how n w er e reCorded f r o m a sin- gle tube at th e co mpr essor face with ou t th e ca scade pre _ e n t. Pl O tted i S the pow e r spe c tral density ver su s f r equency. The r 6so nance sh ows up as high-power spikes at _7 0 hertz a nd at the second an d thiI.d h arin o nic point S . T hid i s typical t _ th e da tR recorded w i th this in l et; howe ve r , / he se h igh-po _ ver spike s _ v o i _eh ot a l ways pre s e _ lt.

Wh e n they were prese l it , their re lati ve am p litude d e pel l ded on the exact po s it / off o _ 1

!

the probe at the co m pre s sor face. The pow e r content in th e spikes adds 130 - percen t t o th e RMS level that w ould be obtained without th e spike s In' e se n t. Th e cascade virtually eli m inat e d the r eso na n ce a t the comp r essor face for most inlet operating co n ditio n s. T h e re so na n ce w o u l d tend t o re app ear, however, ff n o f l ow was taken out of the bypas s cavity. But typi ca l _ t nl et operating c on di ti o n s would call for so m e f l ow through th i s region. Thus, th e c as cade se e m s to be an effective- a pproach to control of the res on ant co n dition.

Summar y - Distortion Measure m ent and Control

It has been shown th at ti m e - varying p ressu r es m ust be eon si derecL in distor ti on analysis and th at the i n s ta n tan eous analys i s tec h niqu e is a very ef f ective m ea n s of I qualita ti vely vis ua lizing th e flow dynamics a t the comp r essor fa c e a nd quanti ta tively i correlating the inlet di stortion to engine s tall. Unfortu na tely, th e ins ta ntan eous L di stor ti o n da ta are very time co n su m ing ar i d e xpe nsi ve to obtai n . HoWever , so m e _ corre la tion does exist between the ins ta n ta neous disto r tions and th e dynamic and I s te ady-state contour plots , which are more ea s ily ob ta i n ed. Work is co n tinui ng on _ this and o th er types of corre la tion s . _i ' The dynamic da ta i ndi cated that dynamic in f orma ti on good to at least 6 00 hertz _ is required to describe th e major vari a tions in ins ta n ta neous disto l_ don at a 16-inch - di ameter co m pressor face. A first gu e ss is that the wavel e ngth o t this criti ca l l_i • ; . frequency should scale linear l y wi th c _m pressor face diameter; in ves ti gati o ns ' I_ wi th la rger inlets planned for 1971 s h ould help to clarify thi s critic al fr eq u e ncy i_ scali n g.

.... Vortex generators wer e shown to be a n eff e ctive method to co n trol di s tor ti on.

' Also , it was demonstrated that a reso na nce i n an inlet duct wa _ _ pp _e ss e d by de- t • . coupling the resonant bypass cavity fro m th e en e rgy of th e I x iail _ stre am by a ca s - • cade. Therefore , it appears that it i s possible to control dis to r ti on, but more work is r eq uired to define op tim u m configura ti o n s.

, REFERENCES

' / ! ' 1. Anderson , B e rnhard H. : Design of S u personic Illlets by a Corllp _ te r P l , o _ ram In - _ corporat t ng th e M e thod of Cha r acteristics. NASA TN D-4960 , 19 6 _ ).

, ' . 2 . Anderson , Ber nh ard H. : Opti m ization of _ per so nic I n lets U si l _ g the M _ od of Characteristics. Analytic a l Me th ods i n Aircraft Aerody _ ia m ics. NASA S P - _ 28 , 4 1970 , pp. 569 - 581.

2 ., , 3 . A n d e rso n, Bernhard H. : Character is tics Stud y of a B ico n e Mixed. ,_ o m pr _ s s to n Inlet f or Mac h 1.80 to 2.50. NASA TN D-5084, 19 6 9.

4. Calogeras , James E. ; and Meleason, Edwa r d T , : W i nd Tunnel Investiga ti on of Tech ni cLu e s for Reduci n g Cowl Drag of an.Axisy mm etr i c External- Compression In let a t Mach 2.49. NASA TM X-1516 , 1968.

5. Cob b i so n , Robert W. ; Meleaso n , Edwa r d T. ; and Johnson , David F. : Per- formance Characteristics from Mach 2. 6 8 to 1.98 of an Axis ymme t r i _ M _e d- Compression II _ letSyste m With 6 0-Percent In ternal Contraction. NASA TM X-1 _ 39 , 1969.

• 6. Sore n sen , Norman E. ; S m eltzer , Do na ld B. ; and C u bbt so n, Robert W. : Study of a Family of Supersonic Inle t Sy s tem s. J. Aircraft , v o l. 6, n o. 3 , May- June 1969 , pp. 184-188.

7. Cubb iso n ) Robert W. ; Meleason E d ward T. ; and Johnson, David F. : Effec t of Porous Bleed in a High-Pe _ or m a n c e Axisy m metric ) Mixed-Compre s sion Inlet at Math 2.50. NASA T.MX-1692, 1 968.

8. Sa n der S) Bobby W. ; and Cubbison ) Robert W. : Effect of Bleed-Sy s te m Back Pressure and PorouS Area on the P erf o rmance of an Axisy m metric M ixed - Co m pre.ss i on Inlet at MaCh 2. 50. NASA TM X - 1710 , 1 9 68.

9. Sa nd ers , Bobby W. ; and Mitchell , Glenn A. : In c r ea s ing th e Stable O per atit _ Range of a Mach 2.5 Inlet. P aper 70-686 , AIAA , June 1970.

• 10. Colh'in , Robert E. ; and Ch ob y , David A. : Steady-State In teractio ns from M a ch 1.98 to 2.58 B etw e en a T u rbojet E n gi i _ e and a n Axisy mm etr i e lhlet with 60- • : Pe rcent In te rnal Area Contraction. NASA TM X- 17 80 ) 1969.

11. Coltrin, R ob ert E . ; arid Calog er as, James E.: Su pe rsonic Wind T tu l t tel Investi- f gation of In let-Engine C o m patib i lity. P aper 6 9 -487, AIAA , J u t l e 1969.

12. Calogera s , J ames E. : Ex pe ri m ental Investigation of Dynam i c Disto r tion in a Mach 2.50 Inle _ with 60 Percent lftte r nal C o ntraction and I t S Effec t on Turbo- Jet Stall Margin. NASA TM X-1842 , 1 969.

2 9 8

EFFECTOF INLETCHARACTERISTICS ON i_

SUPERSONIC CRUISE RANGE i

"'"' SU P ERS O NIC TR AN SPORT -AFTE R B UR N I NG_ . T U RB O JE T I 3 2 i 23 RAN G EDECR F . J V E N]',

I 1 7

i i i -

.91- - .9 o '._--'._ ._-._

, ,_ , TOTAL NACELL E B LEED INLE T PRESS DRAG FLOW WEIGHT RECOVERY COI _ F AT.30 ' _ . : . ' c s- s 6 6z9 R ECO V ER Y ' t F ig u re I X - I " SUPER S ONIC CRUISE INLETS _.

' 1 LO0 " O ISENTROPIC C OMPRE S - 60 - 40INLET i SION IN L E T I:_

i

40 - 60INLET 20 - 80INLET C S.- 56633 Figure IX - 2 3O0 SUPERSONIC CRUISE AXI S YMMETRIC I N LETS tO 0-O _ 0-4 0 COL L A PS I N C ,-CEI _ f E R BOI)' r COLL AP SING CENTE R BODY _ I J 40-60 2 0 - 8 0 , CO L LA P SING CENTE RB ODY TRA NS LA T ING CE N TERBODY CS- _663 0 Figure I X-3 INLET SUPERSONIC CRUISE PERFORHANCE t .. 1 '00 l ,._ TOTA L '" PRESSURE . 90 -- _, r 1 00-0 RECOVERY 20-80 - '- y _ BOUNDARY m

v I _ L A YER - _ 1

I

.. . e _. ....... I I ,I, I ,I I

: . 88 .9 0 .9 2 . 9 4 .9 6 . 98 l. O0 cs-s6 6 z 7 DIFFUSE R MASS F L OW R AT IO Figur e iX - 4

= o l _i _i

i ' " \

EFFECT OF INLETPERFORM A NC E ON S UPERSONIC

CRUI S E VEHICLE RANGE

10 25

/ %

' ' r C O WL WAVLDRAG 12 5 9 _I t , -BOUNDARY LAYER BLEED DRAG RANGE 53 t DECREMENT, , I r T O TAL PRESSURE RECO V ERY NMI _ " 6 3 74. x ' _ ," ; x_ :58 _ REF !'

INL E T 100-0 60-4 _ 40 - _ 20-80 DESIGN MACH NO. 2.5 2.5 2.5 2.65 i ; cs-so _ ze i ; F igure I X-5 ' _ :I

P

I

/ +0-60

I N LET IN STALLED WITH A WING SI M UL A TOR i!

. 2 . 5 o i ,'

,, -BOUNDARY LAYER ON WING SIMULATOR 7 / WING S IMULATO R " STARTED INL ET _ . . , . , -.-. --- ""L.,_ I MAXUPSTREAM ,, . .. ,- UN S TA RT ED P A l " tERN SHOCK U N STARTED I N LE T -< ' .. ' .

• " "" __ UN ST ART_'D STABLE I _ I S H OCK PATTERN F i g u r e IX-6 : ' t S CHLIERE N P H O TOGRAPHS0 8 T A INEO DURIN G I N L E . ' _ UN ST A R T RESULTINGFROM I N TER N AL, A I.B £ 1.. O -W DISTURBA N C E A T MACH 2.50 (a)TIME , 0SECOND. (d) TIME , (1010 _ SECOND . l n _ , ,,, , ., " P _ " _ . 03 _ 1 SECO N D, (b)TIME , (1 0 0 096 SECO N D. ( e )TIME , I101444 S E CO N D. (h)TIME,I1 _ S E COND.

) (C) TIME, O, 00482 SECON D. (f)i'IMI_, O. 01 _ 7 SE COND. ( 8 T IM E, O , O _ S SECON D .

_, . Figure IX-7 . , , , _ .

' UNSTART LIMIT FOR INLET ANGLE OF ATTACK 40 - 6 0INL E T • , , , " . . -'_ _..... - _ ////// ///// t"_ //// , _ 7 / x C0WL / _ /, " '<: , _"_ '/ '/ _ CENI"ERBODY _ DESIGN .... ANGLE OFATTACK LIMIT . 6! - " - - - SON IC [ ' \ ,_ COWL STA TIC l , .. _ P RE S S UR E .4 - , .... , :._: ' TO T ALP RESSU RE '2 - . e , , It 0 i_ cs - s 66 z 6 AXIA LL OCATI O N O NC O WL . i.

F i g u re IX - IO r UNSTART T O LE R A N CEOF 4 0- 6 0 INLET : ,, ,. o NORM A L BLEED I:zBLEED EXTE N DED UPS T R _ Z.M DESIGN -- - 0 - - . 04- # . UNSTART • MACH NO. - . 08 I d .... TO L E R ANCE _ , .

- . 1 2 -- _

- .z 6_ I I I

0 2 4 6 _ ' UNSTAR T ANG _ OFATTACK cs-s 6 e 73 F i gu re lX-]l

Sd5

ANGLE OF ATTACK LIHITS FOR M IXED COHPRES S ION.INLETS

_- , x-_

__-_-;?17] Z_F/////////////// /_

12 -- 60- 4 0 FOCUS E D C O MPRE S SION r-I MAXEXPERIME N TAL I N LET-- - .E] ,_ : , . + . U N START A N G L E _ U N START8 -- . ' r-I t A N GLE OF ATTACK, 4 0-6 0 i . S .

DEG 4 -- 6 0 -40 60 -_-_ +"__ , ,-20 - 80 ' _ _ oc _ sEo -ll I , . _-.+"i; "_+ 0 .4 .8 1.2 1.6 2.0 ,I ; , DISTA N CE FROM COWl. LIPTO ++' _' GEOMETRIC THROAT , AX / R cs . s6 63 z F i gure IX - J2 i_ 60- _ 0 SUPERSONIC CRUISE INLETS i"i_ DISTRIB U TED COM P RESSIO N FOCUSED COMPRES S ION F l g lJ r _ IX-13 cs- s 6 e T z 306 ,

CRO S S SECTION OF 60- _ 0 STABILITY INLET

i FAST ACTING VALVES ; , ! , .

• p, st i ; Flgur_ I×-1 4

FAST ACTINGVALVE S

r TANGENTIAL ,, _ CO N TROL : \ AI R FLOW : ,. 'i. .,1 1 ,_--,'_--'--, ........_ -_ -- ..

I . REFERENCE t - V ALVE PRESSURE EXIT , MECHANICAL V ALVE VORTEX VALVE ',,, F i gu r e I X-IS c s . s 6 8eo i" %

3 _

INLET PERFGRHANCE

WITHTHREEDIFFERENT BLEEDCONTROLS

M 0 - 2. 5 0

FIXED VORTEX MECHANICAL EX IT VALVES VALVES PRESSURE

' c°v " 'r I

.7 I I I I I , 9 1.0.8 .9 1.0.7 .8 .9 1. 0 cs. s6 ee z DIFFUSER MASS FLOW R AT IO Figur e I X - 16

INLET DISTORTION

3 08 INSTANTA N EOUS DISTORTION CO N TOURS b ; "

DISTORTION " 9NTOURS

N _ CHNO.,2. 6 ;AN G LE OFATTACK, 5 0

STEADY STATE DYNAMIC

PAVG I P0 - (1?69; A P / PAv G- 0 . 178 ( A PRMSIAVG / PAvG "0 . 046

!

APRMS / PAvo J

C8-5 6 9ZO Figur e IX-19 !

INSTANTANEOU8 DIS TO R TION

.18- C AUSE • ?CRITICAL OF J _ • 16- \ , LEVEL_ STALL / • 14 -- _ ; -v. . j I DISTORTION _ -STA L L

.08

.0_ I I I I I I

0 5 10 1 _ 20 25 3 0 TIME, mSEC c s- s_g z l Figure IX-20 INSTANTANEOUS DISTORTION CONTOURS BEFORE AND DURIN G STALL L , EFFECT OF VORTEXGENERATORSON DISTORTION i STEADY- WITHOUT VORTEXWITH VORTEX .2r GENERATORS F GEN E R AT OR S DIST O RT I O N, . B _ o .

PAVG" PMIN PAV0 .12[ B INLET TERMINATION

' DIS TO R TION, .o81 / B " JS _ 'GE - 1 3 TURBOJET

( ' 6PRM$)AV o .04 DYNAMIC I _ i ' [ o COLD PI P E-CHOKED PLUG ' PAVG O ' A, L _L _ A , .7 .8 .9 L O .7 .8 .9 1.O TOTAL PRESS URE RECOVERY cs . _ 6 77 z Figur e IX - 22 , ,, RESONANTINLET DYNAHIC DISTORTION C OWL - _ C A S CA D E- _ ,- OVERBOA R D BYPASS I BYP A SS I ' POWER SPECT RAL 10 DENSITY , 5 P SD 1 "., _ ' . . s Hz " 1 B Q Q FREQUEN C Y, H Z ' cs. 56_73 l_ Figure I X - 23 ¢

X . E FFECTSOF" ENG NLET DI ST URBANCE S

ON E NGINE STALL P ERFORMANCE

John H, Povolny , F. W . Burc h a m , Jr. ? James E. Calogeras ,

CarlL. M eyer,andRichard A . Sudsy

, : The operat in K limits ot propul s ion Systems m ay re s trict the c _ p a bflittes of both subsonic a nd s upe r s oM c a ircra f t. One such limit o f g a s turbine propu l sion systems !

is compressor s t a ll or surge. The peHorman ce a nd oper a t i ng limits o ! the engine, i * especially th e compressor System, are ildlu e nced b y th e f lo w conditions at the _ , eng ine inlet. This leads to the necessity o f ad e quately matching the inlet and en gine i S o that the c a p a bilities of the aircraft are not compromised by th e condition of th e _ f low supplied by the inlet to the _ n g ine or the su s ceptibility of th e e n gi ne compressor i , system to this flow. It should be n oted th at, in addition to the in let d e si g n para m - I, e ters, the condition o _ the flo w is also d e ter min ed by e xternal influenc e s such as !, armament firing, f light m_ meuvers, th rust reversal, in l e t unsfiart, in let buzz, etc., !: s o me of w h ich cau s e te m peratur e as we l l as pre ssu re disturba nc es. In o rder to ii_ better I mdersta n d th e f li g ht environ m e n t a l disturban c es an d their effe c ts o ff th e o pe r- _!

s t ing li m its o f tu r b in e en gi ues, re s earch prog r ams ar e be in g c ondu c t e d at b oth the i_ NASA Flight a n d L e wis R esearch Centers. This p _ per dt S l _ usl l e S data obt _ tued _i fro m these pro gram s. : _ i : I' FLI G HT ENVIRONMENT ,, .

Th e program at the Flight R _S ea rch Center in clude s work wi th th e F-Ill air- , !

p l ane show n in f igure X- 1 . M e asu r e m ents we r e mad e _ t th e el _ i n e _ ace d uring the f l ight of th is a i rplane to sho w th e a_ ture of B o rn e of th e di sturban ce s that s ta ll * engi n e. Th e flight data to be di scu s sed were obta ined by the NASA Fli g l l t R e s eal'c h Cente r durin g tests of an early m odel F - 111A airplane, w h i ch was powel' e d by TF30 P - 1 e ngi nes. Eight total pres su re rakes, e ac h wi th five pro b es, w ere iri - stalled at the engin e compressor face. _ otir diagonally o r ie n ted rakes contained *NA_ iA F l ight Res e arch C e nter.

t \ miniatur e tr ans du ce r s fo r h igh- re s pon s e pr ess ur e me_ , s ur eme nt s , an d t he r ema i n - in g fou r r a k e s w e r e f o r t he l o w- r es pon se m e as ur eme nt s , Figur e X -2 s h ow s f o ur re p r e_ , e ntativ e h i gh r es pon se c om pr esso r-fac e p re s - s ure s at 400 sam pl e_ per s e co nd ob tai n e d f or a c om p r esso r _ t al l a t a flig h t M ash n um ber o f 2 , 1 7 a nd a n a ltit u de o f 44 000 fe_ t, Tra ces A a nd B _re fro m t h e 1 :3 0 o'cl oc k r a ke, w h ic h L a l n ca te d in a n _ r ea of g e ner al ly h ig h - r e c over y , a n d l owe r tr a ce s A' _l d B ' a r e fr om t he 7;a0 o 'aluck r _e w h ic h t_ I n a l o w . r e cover y r egi_n , T h e dif f erence in pre ss ur e l evol _ betwe e n t h e two r _ ke s i _ a n ind l_ tion o f t h e e xi s ten c e o f a s te _ dy- s_ ate p re s su re di s t o rti on . The ra pid pre s su r e flu ctu a tions e vid ent in ea c h tr a c e in _ cate a varia t ion in di st or t io n t h at can b e e it he r g r e ater o r le ss than the s te a dy- s t a te value dep e nd ing on the ph as e relat i o ns h i p betwe en oppo- si t e halves o f th e d uct , All four tra c_s sh ow e s se n t ia lly rando m d yn a mi c pre ss ure varia ti on s b ef o re s tall, _ 'lth a g r eat0 r p O ak-to- p eak fluctuation at t h e " / :30 o'clock r a r e , which is in the low.re _ ov e ry r e giou. Note, h owe _ cr, that at the t im e that the stall was initiated , both 1:$0 rake pressures are hig h and both 7:30 rake pros- s u r es are l ow. The r e su l tin g in c reas e in in sta n ta n e ou s distortion ove r the steady ....

sta t e distortion was sufflc is rit to stall th e engi n e _ _ Figure X-3 shows th e compressor-face total-p r essure r ecovery maps f or t h is _ sta ll . T h Q S e maps are s hui l_ tr t o th o se presented durin g the di s cu ss io n of in l et _ p e rformanc e . The low r a spon S e rake m ap o _ the left d e_in es the steady-s ta te dis- i_ tortion a n d shows a m oderat e ly se ver e 180° dist o rtion p _ tt e ra with a 10-percent _:_ _p / p. The c o m pre S so r -face ma p on t h e righ t w _ s made _ r om the h ig h -reSpons _ ii pre s sur e s at th e time that the s tal l w a s in itiated, and it d e fin es th e in stanh _ fleous _ di storti on . N o r adica l diff e re n c e in patter n i s e vi d ent, bat th e s prea d between the _ high- and low-re e ove ry areas i s greater, and A p / p is in creas e d to 14 p _ reent. I_ This hig h er value o f insta ntaneo u s distor t ion l asted f or about on e rev ol ution o f th e _ f an and is co n side r ed to be re s pon s ible f or initiat in g the s tal l .

In the ca se Just discussed, _ er e Wer e no pred0 m ina _x t or re s o n ant frequencies present in th e in let flow. Fi£ure X - 4 shows a dif f er e nt fli g h t c o ndition, Mach 2.0 and 4 5 000 feet _ titu d e, where a strong duet r esonant m o d_ was pr e sent. I _ this f igure po w er spectral density ( _ 'SD) i s shown a s a hu lc t io_ of freqtiency Wherein " P _ D is defined a s (_Prms / Pav)2 : H z wh e r e _ Pr m s isth eroo t m ean sq u areval u eof t heflu c tu a tin g compon e n tofpr e s- sure.

| Data are pr ese n te d f or a ll 2 0 h l gh-re slm n s e pre n_ u r es. A p r ed ominan t pe a k at 2 7 h er tz I s e v i d en t i n a ll ca ses , a nd a harmo ni c a t 5 4 he rt z c an a l so be s ee n , Ex amina tio n o f the ph ase r e l ati o ns h ips be tw ee n t he c o r r es po n d ing 1:30 and _ : _ 0 o 'cl oc k r a ke p ressu re s s ho w ed _ h a t tho se pre s_ ur e_ were a bout 1 80° o ut o f ph a_ e at th e 2 7 her t z r es onanc e and h a d a n esse ntially random plla s_ rel a t i on _ hip at o th , _r fr e q uen c ie s , Ev id e nce of thi s w as a l so a eo n i n t he inlet duct, a t the cowl lip, a n d in t h e en g ine, Thu _ ther e e xi s t ed a _ it ua tio n wh e rei n e ach ha lf o f tile in..

l e t d uct w an re sona t ing l i f o ° ou t c _ ph as e with the oth er half , " B as ic all y , t hen, th e r e a ro b o t h s t eady . _ t a te a n d dynam i c di s turb an ces t o d _ a l w R h at t he e ngi n e in l e t. Th us , in order t o unde r stand what ca u ses t h e en gi ne to ' . ', s ta il , i t h _ r.oco s s a ry to l ook a t the e ffect of oath dl stu r banc _ indi v id ua lly a _ we ll a s i n c o ml;l n a tlo n . This is what is being done a t the L ew is R ese a r ch Center, !

ENGINES INVESTIGATED The work ha s been primarily with three t y pes o f engines (see f i g . X-5). The top fl l ustr a t i_ m is a sketch o f the General E l ectric J85 aflerburn in g turbojet engine th at w a s tested, The J85 h a s a single spool ro t or w i th an eight-stage axial com - pressor (includin g in l e t guide va n es) driven by a two . stage turbin e . The m iddle il - l us t r ation i s a ske t c h o f t he f ixed g e o m e try t urbo f a n tha t w as t e s ted; this engi n e was a Pr at t& Wh i tn eyTF30. T h e TF30 ha s a t wo - spoo l roto z ' witha th r ee - s t a g e , fan (in cl u d ing inl et gui do v a n es)a nda s lx - s tag e a x ia l co m p re s so rd riv en b y a th re e - s tag e turb i n e on th e lo w sp o o l ,a n da s e,e n -s t a g_ a x ial co m p re sso _ d r i v e n b y a s ingl e - s tag e t u rbi n e on the hi g h spool. The b ottom sketch il l us t rat e s the v _ riable ,.

g e o raetry tu r b ofa n that wa s t ested; this e n gi n6 w as a G en e r al Electric 1 / 10. Th e 1 / 10 ha d a two - s p o o l rotor wi th a two-stage fal _ ( n o i nlet g u ide va n e s ) dri _ , e n by a t two-stage t u r b in e on the l ow s poo l , and a seve n -s ta ge axial co m p r es s or dri ven b y a two-stage turbine on the high spoo l . Th e co m pressor of this en g ige w a s c_ luipped w i th variab l e inlet gu i de vanes anti va r ia ble flrs t-_ge s t ar er s . It _bo uld be noted th a t, alth o u g h a l l three e ngin e s a re e qui p ped w i th a f ter b urners, n o n e o f the flats presented herein ar e with afterbu rn in g , STEADY-STATE PR ESSURE DIS T URBANC ES _' Tho fir s t areatob e ex a mi n edwill b e . th e ef fect s ofvar i o us s tead y-s tat e p r e s - su r e di s tor t ion s on th est all p er f ormanc e o ftho se e n g ine s .Du ri n gt ll e_e dls ou ss io fi$ itwill b e _ hownwhere th e w ea k l l nk isi nth e co m pr esso r _ys te m , p ilrtie u l a H y |o r the mo re e omp l i_ tt e d eng i nes , Ai _o _ho wn _ t r e s o me ea_ e _ of i n ter a ct ions b e twe en t h e ta n and t h e compresso r i n on e _ t h e t u r bof a 1 1 _ a nd a sim plifie d t eo hn l q ue for eorr el _ tin g th _ effe et _ of cir c um f er en tia l d is t or ti o n _ w h ere th ese a r e t he p r edomi- nant engine se n si tivit y . R _ h o uld b e n o ted that in mea t ca n e s ti le _ tall limit r e fer re d t o i _ h ard s tall rat he r th an r o t a tin _ stall,

Eff e ct ofTyp e ofDlst o rtlon

/ F i gu r e X-6 ttlul _ tr _t o _ t h e ef f ect o f t h e s h a po p o r the t y pe p o f t h e tot _ d _ p ress ure di s torti o n o n the stall line o f the simple turb o jet e ng in e, It s h o w s a stand- ard m ap o _ c x _m pres s or total pre s sure rati o a S a funct i on of corrected airflow rate ! ' f or t h e JS B -I$ e ngin e u s ed i n this program, The d a s hed lines r e present t h e corn- _ : pr e sser ma p d e t0r m in e d with clean , or undi s torted, airflow i n to the eo m pr e rsor. _ The so l id li nes are the s tall 1 / n es f o r th e d i StO r t e d inf l ow cases shown. T h e tota l = pressure distortions were produced by plaei / _ screens, s h own as the shaded regions i in each of the patterns, app r oximate l y one compressor dia m et e r upstr eam ot t h e il inle t guid _ vanes. Fo _ a l l o f the patt e rn s shown here , th _ screens w ere o f about the i same intensity or porosity. This resulted in a b ottt the s am e distortion amplitude Ii (Pay " P m in ) / P a y for e ach of th e pat t ern s a t a giv e n engine speed, li As can b e seen, th e hub radial distortion pattern ha d little effect on the stall } i l ine of th e eo m / n _ es so r, but the tip rad ia_ dist o rtion caused a m o z 'e serious loss in ' _i th e stall compressor pre ss u r e ratio. It was the circun _ erent ia l distorti o n pa ti e rn I! ' show_ in the lower r i g ht, how e ver , that caused the most loss in comp r essor pre _ - !_ ti sure ratio at stall . On e o _ th e mor e S_ pr i s in g r e sult s occurred wh e n a 120°s e c- i : ; tot o f the h ub radial patter n w as te_ e d . Al tho u g h th e pure htib radial pattern h a d Ii little eff e ct o n th e s ta ll line, th e par t ial radial di stor t io n caused a m uch :m ore _ t S erious stall l in e degr a dation. This de g r a datio n has been directly attr / b u t ed to th e ci r cumf e r e ntial componen t of the di stortion pa tt e rn.

I n as m uch a s eirct / mfc r e n U a l distortion, or the cireu mf er en t / al eoni po nent o f a co m bi ne d di stortion, had the m o s t seriou a _ ffeet o n the conlpress0r, th _ qd eS ti ou arose as t o th e e _f ect of th e extent o _ the c i rcumferential dis t ort ion on co m pre ssor p e r fo r m ance a t stall . Data relati v e t o t hi s q u estion ar e pr es e n t e d i n figure X - _ .

This f i gu re shows th e loss in th e co m p r ess or pr e s s ur e ratio at stal l (the dH f_r ence between th e c l ean and disto _ 't e d stall pr ess ur e ratios) as a fu nction o f screen " angle / 3 . Each curve correspond s to a co n s ta nt eorrect _ i eng in e speed. All the _ screens used to g en er ate th e se data Were of the same intensity or po rosity in order | to isolate the effect of c i rct _mf e r s n t / a! extent. _ or each o f these c u rveS, the loss I m stal l compressor pressure fat l 0 ihe re as ed rapidl y as the screen _ tgie w a s /' /

!

// / , # increased from 0° to abo u t 60 °. Increa sing t he screen an gl e from 60 ° to 180 ° had _ no further effect on the stall l in e of the compressor. This in d i cates that 60 ° i s th e critical angle of distortion for the J85 co m pre s Sor. This critical angle is related i to th e time requ i red by the _ o m press o r to respond to a di s tortion in a quasi-steady- state m a n ner, i

Circum f erential DJstortlon Correlation

It would be desirable to ha ve a s in g le curve w i th wh i ch to correlate th is loss in :_ : s f_l l comp r e s sor pressure ratio w ith a Si mple c i rcumferential distortion in dex ' , w hi ch accounted for variation s i n cir c vx n ferenti al inten s ities and extent s , as well " aS en g ine speed. Sev e rhl corr e lat i on techniq u es cu r rently in use were tried, an d t he _ C I ' one which w ork e d be _t is s h ow n in fig ure X-8. In th is f igur e , loss in Sf_tl l c o mp res- ' sor pr e ssure ratio i s plott ed as a function o f th e ra ti o of min imum to maximu m • pre s sures measured at the compresso r inlet st at io n. In order to accou n t for the • va r yin _ compres s or sensitivity to c ircumf e re ntial extents less than 60°, th e ; - Pm in 60 ° us e d here has been def in ed as th e lowest mean pr e s su re in any 60 ° sec- h)r of' th e flow field. For exa m ple, wi th the s inusoi d al circumferet l ti ai d ist orti on show n in this fi g u r e, the Pmi n , 60 ° would be th e ave r a g e pr e ssure of the cross - i hatch ed re g ion. ' , . All of the stall poin t s ob tain ed w ith p t tre c ircumf e ren tial di s tortions are S ho wn , . in th i s fi g u r e. The s e in clude scr e ens o f thr _e diff e rent po r o s it i es, o f ex[ en ts from ii: • $0O to 1,90 ° , arid e ng ine c o rrec t ed spee ds r ang inl _ ft ' o m 8 7 to 100 p _ rce n t of the rated 1_ i . en gin e s pe ed . But, regar dl ess of the ex _ e ti t or in tensi ty of th e circu mf ertm t ial dis- I_ _ : : tortion, these data Correlate quite w ell wi th th is stmpl _ d istor t ion in dex as long a s I s_ ' .... th e crl tl ca l angl e of di s to r tion is t a k e n in toacc o tu l t.0 _ , c ou rs e , th e sh lg lec ul we ,_ I : ' f air ed t h ro ug h th e da ta ca n ea sil y be c otw e r tL e _ : l to a (Pay " Pm in ) / ]Pav type of in d ex li f o r m o re g e n er al application, ff that is so desir ed.

Turb o f an W e ak Link

Th e f o re v _ oi ng res ul ts we r e f o r a s i m p l e eng i ne With a single co m pres / _ Or. Tl i e / effect o f steady -s tat e pr e ssure distortions oil mt en g i n_ w ith a m oz _ e coinplicated cob1 - . t, ' ' ! pressOr sy s tem (the fixed g _om_ try tu r bofa n ) w i ll no# i b O exam in ed to see W h eri _ thi _ : w_R li n k c o mpressor c ompo ne n t is. ' this en g i n e was also p i-im _ rily susceptible to c ir cu mf erential dis t or t ions. Fi g ures X - 9 and X-10 t o g e th er sho w the op el ' stin g l im its of the fan an d low - pressure compr e ssor on maps of p r essure rt i tio = is a func - 3 1 7 I . !

tton of corrected rotor speed. O n both maps, the dashed lines represent the oper - at ing l in e a n d _ t ai l limit without d i stortion, an d the solid l in es, the st a ll limits with tWO screen- in duced f an inlet di stortions. Distortion ampl i tude s , which vary wi th airflow and therefore rotor speed for a given screen, ar e shown. Without inlet air distortion, the pressure ratio marg in between the operat in g l in e and stall limit was greater for the f an than for the low pressure compressor. As a result of th e screen in duc ed distortions, the f an stall limits were lowered but remained parallel to th e operat in g line so th at pressure ratio margin rema in ed. The low-pressure co m pres- _ sor stall limits, however, were also lowered by th e dis tortions. But they in ter- = s ect ed the operat in g l in e so that the pressure ratio margin became z ero. There- fore, wi t h s im ple steady-state 180 ° circumferent i al distortions at the f an inlet, th e _,_ weak l i nk co m ponent of t l_ s fan-compressor syste m was th e low-pressure compres- i_ .... sor ra ther than the fan . Th e high _ pres s ure co m pressor was not affect ed by th ese ! , fan i nlet di st o rtionS. They were ap pa rently att e nuated before enterin g that unit. i_ R eynolds Number andCo m ponent Interaction Effects _..:.I_' i:i Effect of Reynold s num b er in dex. - I t has been observ ed that both altitude and I . _ flight M ach n umber ca n infl u e n ce t o leranc e f o r steady - state d i sto rt i on . Th i s can be !

[ seen i n figure X - 11 which illu s trate s th e ef f ect of Reynolds n umber on the steady - I_ s ta te di sto rt ion tolerance of a TF3 0-P- 3 tu rbofan operat in g wi th its ble ed s cl o sed. I _ : : These results are for simple 18 0 ° circum f ere n tial dis to rti on s. Sta l l lim i t s at e lil shown i n terms o[ di stortion amplitud e Which resulted in stal l as a fu n c t ion of cor - i,} rected f_ n rotor speed. Stall - limit cu rv es are sho wn for various levels of Reynold s t,

I

-.. : number in dex ( ReI). Th i s in dex is th e Reynolds number at the test con di tion ex - i .

pressed as a fraction of the Reynolds number at stan da rd sea-level static condi t ions. !_ (Th e curve f o r a n Rel = I. 0 W a s e stim ated. ) Itca n be see n t h at dis t ort i on amp l i - I_ t tudes at stall decreased c on siderably a s Rey no lds numb e r wa s red u ced by g oin g to I_ hi g her al t itu des. The tre n d of decreased dis torti on amplitudes at stall wi th l o wered _: cor re ct e d fan rotor spe ed in dicates that incr ea sed fli g h t M ach nu mb er would al so . redu ce distorti on t o lera n ce. T h is isbe ca u s e In c l 'e a sed fli g ht Mach number r e _ml ts in high e r i_ e t t e mp e r a tur e and t h eref or el ow e r e d c orr ec t ed r o tor spe ed.

R is int e r es ting to not e that th e w ea k l inkon thisturbo fa n e ng i n e w as notth e fi r s t c ompon e n t to b e hitby th e st ea d y-s t ate p r e s s ur e distortion, bu t W as t h e s eco l_d on e, th a t is , th e low press u re compr e ssor. It s ho ul d al so be not ed that th ei-e did not appear to be any in teractio n s b etWeen the fa n a n d the compressors al though th e poss i bility should be recog n ized th at th e large R eynol ds number effect could in volve su ch an interaction. Whe th er this i s so re m a in s to bedeter min ed for this en gin e, \ • o b u t o ne en gi ne f o r w h ich it w a s fo u nd t o b e t _ue is th e v a r i ab le - ge o met ry t u rb o fa n that was test e d. The next two f igu re s help illustrate this situ a ti on .

We e n cou n t e r e d th e first ty p e o f i n t e r a ctio n while e . xp lor in g inlet Reynolds n u m_ b e r effects on engine p erf o r m ance with no di s tort i o n at the in let. As you would ex- i p e ct, reduci n g the eng in e i n let ReI from 0. 4 1 (n omi nal test value) to 0. 18 r es ulted I i n a loss in the operating m ar gin of the fa n . Thi s loss, howeve r ' , was smaller t ha n !

the l o ss in m ar gin that w as experienced b y the hig h - p resSure co mp r e ssor. T his is / sho w n in fi g ure X-12. A s before, compressor performa n ce is shown o n th e conVen- ti o n al pressure r _ .tio-flow rate m ap. All v a l ues _ .q normal i zed i n terms of the rated _ performance o f this co m po n ent. The stall l imit for eac h ReI wa s ob ta in e d by • th r ott l ing at t he c o mpre s s o r dischar g e un til s t all w as e r. _ o un te r ed. A s s ho wn , the - m a in effec t o f t he l o wer e ngin e i n le t ReI was to ca u se a s hif t in the compressor • o pera ting l i ne; ve ry little el f ect w as f oun d in t erm s of sta l l pre s sure ra t io. Ho w- , _ ever , the op erat ing line s hif t ca _ . _e{l a si gnif ica n t re d ucti o n i _ the _ va i lable margi n ' between th e stall limit and the o p er _t t ing l ine .

Th is operating line shift i s a ttribu t ed to a change in th e flow out of the _m which caused a c hang e in the velocity p r ofile g ohl g in to th e co m pr es sor. T his iS t_ ho w n i n f ig ure X-13. Here, the var i atio n in inle t axial vel o ci t y with pa s sa g e h eigh t is shown for both e ngin e R eI' s of 0. 41 and 0. 18. T he data were obta ined at S imilar e n tw ine operat ing co nd itions of ro tat ional speed an d exha u st n oz_.le ar e a. T he lar g e velocity g radie nt th at w a s ob tain ed at th e R eI = 0 . 41 is a resul t of th e curvin g f l o w path _ r o m th e fa n di schar g e t o th e hi g h - pres s ure co m p r e ss o r i n le t . As th e e ng ine inlet R eI _._ was reduc ed , th is v e locity gra di e n t Wa s effect ed in tw o Wa y s . First a st e eper V aria- ti on in ve lo ci ty fr o m hu b t o tipwas o b t a in e d , a n d S ec on d t he ov e r al l veloci ty level d ec r ea s e d. T h e e nd r e s ult o f both o f th ese c h ange s iS a sl ig h t shi ft in p re ssu r e rat iodu e to the c hang e in in c i de n cea ng le at th e compress o r inl et a n d a red u ct ion in airfl o w t hr o ugh the c o mpress o r d u e to th e o verall dec r ease in vel o city. These t effects w ere the pr in c i pal r eas on s f or the co mp ressor op e rat ing l in e shift. S in c e the ii dif f e r e n ce s i n th ese p r of i l es i n fi g ure X- 1 3 ar e the res ul t of a change in fl ow condi - .

ti on s out o f th e fa n h u b, th e _o ss in compressor s tall mar g in i s attrib u ted to t h e i Rey no lds nu mber effect o n the flow o ut of t he f an an d n o t _o the eff e ct on the co rn - ,; rress o r per se, because th e R el at the comp r e SSo r in let Was v er y hig h in bt) t h ca s e s. I t sh o uld b e not e d th at this type of C om p onen t in te r act i o n cab ha ve a s i gnifi - cant impact on s ta ll se n s fl iv i t y of a n eng i ne w heat op e rat ing at hi g h alti _ des ( l ow R eI). / Al thou gh t heda t a j u s t prese i_ ted dea lt wi t h R _yn o l ds fiumb e r eflect wi t hou t d l sto i .- i ti on sbe in g prese nt at the eng in e in le t , i t is an ticipated thatt h e effect W ot i ld be s im i- la rw i t h d i s t ort i on.

Effect of rotat in g stall. - An o ther co m po n e n t in ter a ction e n c o untered in vol ved r o tat ing stall in the fa n o f the var iab le geometr y t ur b o fan. T his ty l m of fl ow in sta- 1

!

319 f bfli ty w as f ound Wi t h b o t h hub and t i p rad i a l dis t o rt ions . F ig ur e X- 14 shows t he effect o f a hub radial di s tort i o n o n fan perfor m ance. H ere a g ai n, th e co n ve n t i o n al p re s sur e rat io-f l o w r a te m a p i s used, T he da s hed li n e s r e p r e sen t th e u n d i. , tort e d perf o r man c e , a n d th e so lid li nes th e di s t o rte d per fo r m a n ce, Sever a l v al u es o f d is- t o rti on a m plit u de a r e sh own a s a ref eren c e to i n d i cate d is tort ion m ag n itu d e.

T h e m a in e f fe ct of t h e hub ra di al d is t o rt ion w a s th e ind i cated reduct ion in f a n stall pr es s u re rati o , A slight c h a n ge i n t h e o perating l ine w a s als o o btai n ed. T h i s i nd icate d redu cti on i n t h e st al l lim it wa s no t ca use d by a n a br u pt stal ; i n t he f a n , H ig h resp ons e d a t a in dicated t h at t he f lo w w as b reaki n g do wn i n th e h ig h- pr ess ur e c om p re ss or and th at t h is f l ow brea kd own w as ptece d e d b y a ro t a tin g s t a l l i n t he f an.

Th is S eq uen ce o f eve n ts is il l u strated i n figu re X- 15 wh e r ein hi g h r es pon s e p r es - sures at t h e f an inlet, f an o utlet, h igh - pressure co m pressor i n let, and hig h -pressure c om pre ss or o uttet are s h own as t h ey vari ed w i th t im e. ( T h ese a r e o n ly a f e w o f the many pressures rec o rded dur in g a sta ll . ) A rotating sta ll ce l l is i n di c a ted by the p _ ri o dic f l u ctua tion in the fa n outl et p r ess ur e. As t h e fa n di sc ha rg e w as thr o t tl ed t o ward s the s ta ll li mi t, the m a g nit u de of the rotat ing stall increased, a n d an abrupt sta ll occurred in the hig h -pres s ure c omp t'essor. T l 'J, s i s sh o w n by the si mul ta n eo u s d ro p in the o utlet pres s ure and ri s e in the in let pre s sure. Th is s ta ll was the n propa- g a ted fo rward th r o u g h th e f an. T hese re s u lts illu s t r ate a c ompo nent i nteract i on f w here a s m all fl ow in s ta bility , such a s a rot a ti n g s ta ll in o n e co mpon ent, caused a I i i serious flow pro blem , s u ch as abrupt s ta ll in a d o wnstrea m co m po n ent. Th e fa n t _ , t ' r o tat in g s ta ll di d n ot ser io u s ly a f fect eng in e operati o n b y itself, b u t it di d cause the i h ig h-pre s s u re c om p r ess o rtos t a l l , th u s r ed u ci n g thes ta ll l i mit o fthee ngin e. I | Distortion a m plification. - A nothe r type of c o mpone n t in t eracti on t ha t wa s en - !i c oun tered while te s ti ng t h e var i a b le ge o metry t u rb of a n e ng i n e wa s rel a ted t o a n a m- ll! , pll licati on o fcirc u mfere n t i al d i st o rti on thro ug h th efa n . Dur in g th e in itial tes t i ng o f Ii !

thi s e ngin e, the en g in e s tall l i mi t wa s red u ced mo re t h an expecte _ ') y c i t cu m fere n - _ t .... ti al di s tortions. The li m it in g co m ponent wa s aga i n the high-pr esSu re co m pres s or.

However, th e cause of th is lower limit wa s tr a ced to higher tha n expect ed pre ss ure d is tortions at the high - presS u re compre ss or inlet f or a g i ven eng in e inl et d is tortion.

S in ce th e high-pres s ure co m pre s sor in let flow cond i t i ons a re e s tabli s h e d by the f a n h ub c ha rac t er i s ti cs, the d i st o rt e d perfor m ance of the f a n h u b wa s eval u ated , as _ i.

shown in fi gure X-16. In this figure is plotted a di st o rtion ampl i tude rat i o as a f unc - t i on of the t o tal t o s ta tic press u re rati o at the fan h ub in let. Thisa m p l it u de l _ ati o i " isdef in ed as th e rati o o f th e fa n h u bo utl e t t oin let dis t orti o n st ren g t h. Valuesof am plit u de rati o greater tha n on e in dic a te a dist o rt ion a m plifi ca ti on o r increa s ed di s - tortion str e ngth; values l ess th an o ne in dicate an attenuat i on or decrease in d is tor- ti o n s trength. The total to s tatic pr ess ure ratio i s a n in dic a tion of inlet axial veloc- it y . Lo w va lu e s imp l y l ow ve l ocit y and higher va l ue s im pl y higher velocit y . A s can be see n , t he d i st o rt io n amp l if i ca ti on w as s tr on gl y inf l u e nc ed b y i nl e t velo c i ty , v a ry- Ing fr o m 2.5 at l o w value s to a b out I . 0 a t hig h value s . Another w a y t o loo k at this v e l o cit y effec t i s wit h th e u se o f a f a n eq uiv a lent s ta g e c hara e t e rlsti e cu r ve a_ s h o wn in f i gu r e X-IV. Here, the f an hub p e r fo r m an ce is p lo tted in t e r ms of a pr e s .

su re c oe ff i ci e nt as a functi o n o f f lo w eo ef f lete nt. These e oe fficl e nt s a r e pr essu r e ratio a nd i nl et a xial velocity functions n o r mal iz e d in t e r m s o f blade s p ee d. T he c urv e r e pr ese nt s the o v e ral l pe r f or mance o f both fa n s t ages . Th e c u rv e ca n b e di vid e d int o two pe r f or m an ce re gio ns : (I) a r eg ion behi n d an inlet distortion s c r een , s hown b y t he s o l i d sy m bol s , and (2) a r e g i on no t be hind a d is tort i on scre en shown b y t he op e n sym bol s . These two r e g i ons c a n be re l ate d t o low an d h i gh inl et pressure, r es p ecti vel y . In this fi gur e , ze ro a m plif ic at i on would o c cur wh e n th e pr ess u re coef - flcle nt s i n both r e g i ons w e r e equal. If w e us e t he square symbols a s a t ypical e x. I ample of distorted perf o r m ance, it can be seen that the pr O ssure coe f fi c i e nt behind i_ the d i stortion screen is less than that not beh in d the screen. In other words, 10ss i pressure rise w aF, realized i n the low in let pressure region t ha n h i the high inlet i, pressure ,'cgion. Th i s indicates a n inc r e a se in distortion a m plitude a t the f a n hub .........

outlet. This a g re es with the amplif i cat i on shown in f igure X-16.

In a n atte mp t to chan g e thi s fan hub char a cterist i c, the en g i n e man u facturer m odified the first-s t age rotor by in creasing the blade chord length by a fac t or of 4 / 3 and redu c i n g the number of blades by 3 / 4. Solidity an d blade an gles re m ained con- stant, a nd no cha n ge in blade l o ad i n g was made. The sec on d- S ta g e rot o r a n d both , . s i st e rs were not cha ng ed. The resu lt o f this modif ica t i on i _ shown in terms of p r es s u r ea nd flowc o eff i c i e n ts in f igure X-18. l _ lere, bot h t he modi fi e d a n d th e " o ri g ina l h ub characte ris tic c u r v es are sh oW n. A s i gnific _nt cha ng e i n th e pr es sure- fl o w character is t i c i s obvio u s, espec i ally beh in d ti m dis tor t ion screen as sh o w n by the s ol i d symbols. If th e square symb o ls are again used as _ e x a m ple of d i storted ! , • perfo r m ance , we no w see th at the press u re c o efficie n t beh in d th e di stort ion scree n t i s l ar g er th a n that no tbehi n d th esc r een. Th eref o re, m o re pr es su rerise W as [ i r e ali ze d in t h e lo W i n let pr e ssur e r e gion t ha ninth eh i g l l inl e t p r ess ur e r e gi o n.Thi s indi ca t e s a dec r ea s e in d istort i on a m pli tud e a tth0 fa nh u bo u tl e t. T h 0 effec t o fth i s cha ng e iss h own in f igur e X- 1 9 int e r m s o f am pl it u d e r a ti o a s a f un c tion o f tot al to st a ti c p r es sur e r a t i o .B ot h t he mo d i f ie d a n d ori gi n al r e sul ts a r e sh ow n fo r c om pa r i- s on. R canb e see nt ha t a m a rk e d i m p rov e m e ntinthe f anh u b di st oi_te d per f orm a n ce w a s r e a l ize d. T he mod if i ed fan atte nuat ed t h ec lr c Um fo r e ht i al dis t o r t io n s , a nd the o ri gin a l fa n a m plified th e se dist o r t i o fi s .

: .... These res ul ts a ga hi ill u strate a sittmti on w ll0 r e a flow pi _ oblem in on t_ co m_nen t can adversely affect a dowristrea m c om pol _ ent. Ampl i f icat ion of distortions thro t i g h a fa n can rec r uce the engine stall limit because it effect _ the s tall _t erform tizi ce of the co m pressor b _ hind it. R ecognizin g and und e rst an ding interac t i o ns like these is 3_i I n eces sary in o rd e r to c orr e ct t he d e fi cien t c o mp o nen t. Th e case that wa s J us t descri b ed, i s o ne exa mp l e o f a c o r r ec t io n that w a s e f fective in redu cin g an int e rac- tio n problem. Two thi ngs to be le a rn ed from this study are t h a t t he re c an be very s ignifi c ant inter ac tion s between c o m pre ss or c o m ponent s and that what appear to be m_n or bl _ dl n_ c h a n ge s c _n have sl_ iftcaut e ff ect s on t h e perform an ce. B ec au s e the detailed flow processes involved in the f o reg oin g p he no m ena are not c o mplet e ly u nde r s to od , the co m pr ess or re s e a rch peopl e have some a ddition a l work to do.

DYNAMIC PRESSURE DISTORTIONS

ia th is section . the effec t s of dynamic d i sturbances on the s tall clla r acter t s U cS o f t w o engines, the si m pl e turbojet a n d th e fix e d g e o m etry turbo f an, a re oxam in ed.

Before looking a t th e e n gine data, however, it would be well to discuss briefly so m e of the research techniques that were employed.

Research Techniq ues

• Figure X-2 0 shows a ph ot ograph o f th e simple turbojet (J8 6 ) wit h _ fligh t type inlet i n s ta lled in th e L e wis 10- by 10-foot supersonic tunnel. Fo r the t e sts to t)e ' di scussed the tu n nel was operated at Mach numbers u p to 2 .5. The inlet and engine wer e operat e d so tha t inl e t u a star t s w e r _ initiated to deter m ine th eft" effect on the engine, and et _ gine stalls w _ re in itiated to det e rmi n e thei r ef f ec t ml the isl e t.

The tests on the fix e d _ eo me try tu rbofan ( TF30) were conducted in th e L eWis . propulsion system laboratory al titttde test chamber wh e rehl a uniq ue d e Vice for c r eat in g d y na mi c disturbanc e s was e m ployed. This device (o r techn i q u e) is being f used to creat e cont r ollable dynamic or stead y -state p r e s sure distt w banees at th _ inlet s of engi n es being tested in the altit u d e facilities. T h e t echniq u e us e s many s m al l alt Jets upstream of the engin e, inl e t. Figures X-21 and X- g 2 S 11o _ ¢ the m afll featt _ es of th e air jet system. T h e photograpll (f ig. X-21) i s a V i e w lo O king toward the engin _ from til e bell m ou th entranc e of th e inlet duct. l _r imary airfl vw pa BSe s through th e bell m outll and inlet duct to the engine. The ai r -j e t syster _ i ncl u des an array of s m all nozzles in t h e in l e t duct. These ai r -J e t nozzles a re tmifornll y di s- tribut e d in an axial pl ane and are a r ra n g e d in a _ att e rn _ hi _ h re p e _ ts e ve _ _ 0 ° of . ' ci r cumferential eXtent. As in dicated in th e sch em atic, second, _ ' airflow Is lnJ / _ te d fr o m these nozzles and is di re ct ed cou _t er to th e primary airf l o W . Flow contro l of the secondary airflow is provided for e_ tch 60 ° s e cto r " of ai r .J et tlozz l es b_ " sik valves e xternal to the inlet duct. These are high-resporise valves of NASA d e sign for d per _ t- 32 2 ! , t i o n over an o scil l a tor y frequency r a n ge f r om 0 to abou t 200 h e rtz. Se par a te m atched flow lines a r e pr o vi d ed fr om e a ch o f these valv es t o e ach o_ the air-jet n o z zl es of a 60 ° _ ,ctor. Se c o ndar y ai r i s su pp lie d t o the va l v es a t pr ess ur es u p t _ a bout 10 a t mos p h ere s . Th r o ugh c on tr o l with th ese valv es o f t he s ec o nd ary-a ir d is - tr i b u t i on a nd t h e f l ow rat e , a var i et y of b a si c d yn a mic o r s te a d y - s t a t e i n l et pr e ss u re d is turb a n ces ca n b e cr e ated f or engine t eats .

It h as b e en hoped for som e time that the air-Jet s y s t em c o u l d b e u sed to a p pro xi - m ately simulate a n actual f l i ght environment. Re cen tly, one series of tests to chee k that potential capab i lity was complet e d. NASA Flight R e s e arCh Center cooperated in this and made available e n g in e inlet pressure data from fl i ghts of an F111A. Repro- sentativ e pressure data f rom flight, on analog tape, were us ed to provide input si g- n al s to the high re s pons e valves of t h e air- J et system f or engine tests in the L e wis altitud e facility. Figure X _ 23 shows one co m parison of t he flight environ m ent and the atte m pted S i m ulation on the ba s i s o f power spectral density of in let pr es sures ov e r t h e fre qu e n cy ra n ge fr om 0 t o 2 00 h ertz. Thi s i s a fav o rable c om par i so n .

Bo th sets of data include PSI) s pike s a t 2 ' Land 54 hertz , with the s imul a tion data 1 , s o m ewhat attenuated (in these spikes) relative to the flight data. However, the g e n- I!

eral PSD level of the s imulation da ta is c omparable to that o f the f l ight environ m ent, i . _ I _ Th i s m ay or m ay not be a r e p resen ta ti ve co m parison, b u t is encouraging. Many !

more co mpa risons of th e data from th e flight and si m ulat e d environments must be , _ mad e . Until th a t is done, c o nchtsions cannot be reached as-to the ca pa bility of th e ill air jet system to approximate simulation of an actual flight en v iron me nt, i_

Inlet - Engine Interactions

1: ' i= _i _ NoWthat th e dyna m ic disturbance testing t e chniques have been described, some I of the results witl be examined, sta rtin g with the 10 by 10 foot SWT inlet unsta r t I! t . t e sts on the. simple tur bojet with a flight inlet. Fig u re X - 24 i s a time hi s tory of an inlet unstart with th e engine operat in g at 85 p _ rcent corrected s peed at Mach 2.5 tml _ lel condition s . (An inl e t ul _ Start caUSes a rapid drop in th _ co m p r eSsor i nl et pressure. ) Shown in th is f igure as a func tio n of ti mO are th e co m p r e ss or total pressure ratio and the com p ressor i n let and the e x it total p re ssur e s, botli ratio s of fre e- strea m total pr ess ur e . Looking first at the comp re ssor i n le t press ure , it c ab be seen that after inlet u ns t A r t this pr es sure rapid ly reaches its m ini m u m in about O.O1 second. On the other hand, there is a sigflifi _a nt lag in rite pr e ss ure drop of the compr e ssor ex i t total pr e S s tire wlitch i s a B soclated with t h e volume effect of the coinlmstor. This lag caused t h e co m p r essor pi'ess U_ - e ratio to rise fa r above the stead y -s ta te stall l in e for a time eq u al to about 2_ rotor Q revolutions, which time W as suf f ic i e n t to cause c om pr e s so r st al l, a s well as corn- b u s ter b l o w o ut, Th e effects of this particular unet a rt transi e nt, a s w e ll as c ome less s e ver e un s tart tran sien t s , _ r e summa ri ze d i t. f _ r e X- _5 , Th is f igur e s h o ws b oth the lnitl_ fl and the peak c om pre ss or pr es sur e r ati o s tha t o c c ur re d d ur i n g t h e in l e t a n .

s t a rt s . T h e ba r on the r i g ht rep r e s ent s the Ma th 2. B dat a Ju s t d e s cr i bed. T he three bar s to the le f t po rt a i n to inlet un s tart a a t Mach 2, 0 with th e engi n e oper a ting fro m three dfff e rer _ t in i t ia l p re _ a ur e r _ t i oe o n tits 81 J-p e r cen t c o rrected sp e e d char.

acter ist i e . B y c om par i n g t he re l a t ive a m p l it u de s of t he b ar s , you c a n s e e t ha t t he Mash 2 .0 unstart tr a n s ient s W ere much less severe t h a n the Mach 2. 5 transient.

But st i ll , depending on th e in it i al co m pre s sor operatin _ l x _l nt, resu l ts ran g ed f ro m n o s tall to stall and co m bustor blowout.

Figure s X-26 and X-27 show some in t etstage S ta tic pressures during the lar ge - i scale tran s ients J u s t dis ous sed. Shown in f i gu r e X - 26 ar e time histories of static pressures loc a ted _ ttthe co m pr e sso r in let a n d at the e xits of the first seven stag e s of the eight-stage compressor. These traces pertain to th e inlet un s t a rt that was in itiat e d a t Mach 2.5 du r in g engine operation. Almo s t im m ediately upon inlet un.

st a rt, the pressures at the co mp res s o r in let and all the w a y th rou g h th e co m pi'es s or drop rapidly with ti m o , until that ti m e when all pr e ssures abr u ptly Start to rise. This pr e viously mentioned, this Ma ch 2.5 unstart transient wa s s e vere e n o u gh to cause co mpr esso r stal l as well as co m b u s to r blowout. The oz _ igin of a stalled zone in a c om pressor s tag e ca us es an a b rupt loss in th e a irflow pu m ping of th a t s_ ge , Tiffs iS a nal og ou s to closing a high-r _ spo n se valv e . A co m pres s ion w ave is sent for w ard of th e v alv e, or st all e d zo ne , an d an expa ns ion iS s ent rearWard. So the ab ru pt rise in pressures Seen all th e Waythrougll th e S eVenth s ta ge in dicat e s th at stall or i ginated in th e eigh th , or th e last, sta g e of th e co m p r e s sor. Thus a di sturban c e at th e in let to th e co m F c e Sso r caus e d a stall which originated at the rear of the c o m presso r , t In contrast, f i gure X-27 shows w ha t oc uur s to th ese sa m e pressures in th e corn- presser when, at th e same in itial engh _ e a nd tunn e l conditions, th e co mpr es sor ' i s forced in to stall by clo s ing the exhaust n ozzle. In this ca s _ , th e ris e in pressure is seen at th _ c o m pressor inlet and throu gh the fl _ s t t w o stages o _ t he Co mp ressor, w hile at th e same time t he exi t of th e th i r d sth g e shows an abrupt drop in pre _ tzre, in di cati ve of stal l or i gin i _ th a t stage. Th u s In this case , a dist ur bance a t th e rea _ of th e co m pres so r caus ed s tall W hich o rig in at e d iti the _ rorlt stag e s.

, In additio n to an inlet un _t art, w h ich i s in itially asingie piflSO _e_ u|t ing in ., rapid de w, ease in ettg in e itfle t press u re, there a l ' e o th e r t ypes of rapid pr e ss u i*e di sturb an c e s en couriter ed that _ an also stal l the engine. For e xaniPle, inlet btizz, which can be undistort ed cy cl ic dist u rbance, a _ d duct r _ son _ ce, enc o unt e red by th e F-1il, whe r e in eac h half of the d u et is 180° ot i t of p' h ase wit h the o th er , _ i

Ef f ect of P ressure T rans i ent s W i thout 5 p at la l Di s t o rt i on

The eff e c ts of d ynam i c inl e t p r esan re va r ia ti on s, with o ut di s t o rt ion , w}lt b e con s id e r e d first. Th e_ e will inc l u de cyclic pr essu re vari a ti o ns t h at a r e an a l o g ou s t o in let b uzz and t he p r ess u re. d e p res si o n p u l se w h ic h i s s ilufla r to an inl e t uns t ar t.

Figur e X.28 i llu st ra te s t he to le r anc e of a TF $ 0 = P -! turbo fan f_r dy nami c t an.

inl e t pr e s s ur e variati o ns induced Wit h t he a i r Je t _ys t em . Stall li m its a re sh ow _ in t e r m s of pressur e ampli tude whic h resu lted in s t al l as a fun ction of d is turbanc e f r e.

_ ency f o r both cyc l ic and pu l se p re ss ure va r iation S . The particu l ar p ulse h a d a pr ess ur e - t im e respons e o f abo u t 40 to 6 0 he r t z , an d i _ s hown a t that fr e q uen c y .

Pressu r e a m plit u de is s che m ati c ally de f ined b y the slcetd _ es in the fig u re f or both the cyclic and pulse cases. In both cases, pressur e a mp litudt _ is de t tn e d as the &P of the pressure d e p re ssi o n expres se d as a fracti o n ot th e pressure level before the d i s turbance. The pressure a m pl i tude a t stall with cyclic variations r an ged from about 0. 165 _ .o 0. 145 in the frequency ran g e from 1 to 20 hertz, and wit h the pulse wa s about 0. 16. Y _ is apparent, tho ug h, that rapid changes in inl e t pre s sur e even without distortion can result in s ta ll.

As for the turbojet, rapid changes in inlet pressure also c an cause proble ms in those components of the fan-compressor system o f the turb of an which are followed by large volumeS. The tolerance o f a TF30 - P - 1 turbofan to cyc l ic / an - inlet pre_- sure variation without distortion was restricted by th e high pressur e c o m pressor rather than the fan or low pressure co m pr e s s or. Figur e X -2 9 shows t h e ti m e .ii . histories of s ever a l pre S s U re s d _ rin g the laSt cy c l e of a tO-h er _ c y c lic (_ .u - inlet pressure oscillation which r e su l ted in stal l . Va z 'ia t ions O _ total pre SS u _ 'e$ with tim e are shown at th e f an inl e t, th_ f a n e xit in th_ tip and hub region s , the :: m iddle and e xi t o f the lo w pressure com p ressor, an d the m i dd le a nd exit bf th _ hi gh pr e ssure compressor. Stall was found to orighlate i n th e high pr e s sure co m predsor t / which, of course, i s followed b y th e co m bus to r volum e . The su dd eh reduction i n flow associated with stall cau s ed a decrease i n pr e sstl _ ' e aft of th _ s tall (fo _ exampl% at the high pressure co m pressor exit)s ad r apid increases i n presst lr e forward of i i the stall in itiation as s ta ll progressed forw a rd through th e low p r e s sur e coni p r e$ sb t and fan, until a ha m mersh0ek appea r ed _ t th e fan t _ t.

The time histories o f total pressure variation s leadi n g to st _ ll wer e us e d t b obtain time histories of p ressure ratios acro s s variott s s ta ge grouPs o _ the fai _- compressor syste m . Figu r e X - 30 shows the time hi sto ri e s o _ _ i.i ous s ta ge-group pressure ratios that were o bta in ed. V a riations of p_e ssttre ratios with time are shown f or the f an hub and ti l_ regions, t h e hi g h presst _ r e compressor f ront ari d real" stage groups, a nd th e low pressure co m presso r front a n d r eal sta g e _ r _ ti p s. To t h e L right of t h e figur e are n o t ed th e _ t ea d y . s t a t e p re_ n u re ra tio s w hic h e xi s t e d fo r t he V ar iou s _ t ag e gr_ p _ be fore th e lO - h ertz cy cli c inl e t pre ss u r e o_ efll a ti o n w as in.

d ared, it c a n h e see n t ha t t he ma zhu um pressu r e r ati o _ wtn g_ Wer e m a d e by t h e f an a nd the r e ar . s t ag e gr ou p o t t he high p re ssu r e c om pr e ssor , b ath of w h ic h a r e f o l l ow e d by la t _ g e v oi u me n . Pr e_ ttr e ra tio _ of th e f aq hub a n d t ip o_c i l l a t e t o t . al ue _ W e |l _tbo v e the _ to ad y _ a t_ ttc lev e l_ . T h e tan, h o w e ve r , h a d u u ffietcnt p re ss ur e ra ti o mar _ n a nd wa _ n o t t he w e ak l in lt c o m p o ne n t. In t h e high p r e ss u r e c om p r e s s o r, pr ess ur e ratio s o f the front _ ta go g r o u p o _ ill a t e but rema i n be l o w rit e s t e ad y . _ t a t e lev e l, but tho _e o f the rea r. e| a go g roup osc ill a t e about a mc_ m which i s abov e th e s t e ad y - s t a te lev e l. P r e ss ure ratio s wi _lgs o( th is r e a r -st a_e g r ou p to val u e s well ab ov e th e s t ea d y - s t a t e le v e l, to gethe r with th e pre s su re .t ime t r a ces i nd icat e that i t was this stag e group in which stall in i tiat e d a nd which w a s th e woa _ lin k .

Ef fects of Pr es sure T ran s ients With S patial Di s t o rt io n

It will be r e called that a duct resonance , which r esulted in an o s cill ating dis- tortion a t th e engin e inlet , was observ e d trom th e FlllA flight data. T h e a l l ' j et s ystem ha s been used to in duce dynamic in l e t di s tort i ons o f thi s and o th er types f or engine tests. : Figur e X-31 illustrates the tolerance of a TF $ 0-P-1 tu r bofan fo r an oscillating i , * 180° ci rcumferential dis torti o n , wh i ch iS anal o gou s to that du _ t resonanc e _ situat i on, i_ Stall limits a re _ hown in t e r m g of the di stortion _m plitude W hi ch resulted in stal l as a flutct i on of f re quency of th b oscillation. For e b tt i parison, dist ortion a _ plihtdes at stal l a re noted f or a steady-st a te 1 50O dis to rtion and f or a sLA _ le pols e pressur e " depre salon Of a no m inal 1800 circun _e ren t ia l extent. Di s to rt i on a m plitud _ is de- ' f in ed aS b e fore. The fan-co m p re sso r system W a s able to tolerate larger dis t ortion # _ , amplitudes with oscillating di s tor t ion th an wi t h steady-stat e or pulse dt s tortioh s , f ' , Th e tol e rance to oscfllat in_ di s t ortions tended to d e_r ea S e with increas ing f r e que n cy of os c illat i on, and app r oa ch ed the s teady-state disto r ti o n to leranc e lev e l at a f r e - quency of a bou t 60 hertz.

Figure X- $2 illustrat es th e tolex'an ee of the sa m e _ nS ine f or r otatin g 1800 ci r - eumf er_ ntial di s tortion S . Stall l i m its ar e e ho _ ,n in t e r m s ot disto r tion amplitud e s that re s ult ed in stall a s a ftt nc ti on of th_ frequen cy o _ rotatiou for eontra i eot A ti n g a nd co r o ta tin g disto r t ion. Fo r compari s on, di _ orti on atn p litude _ at stall ar e again noted for the steady-s tat e a n d ptfls _ 1 800 d| st orti O n _ . _t can b e s een that t h e f a _ - co m pre s s o r sy s tetn W as able t o tol e rate larger dtsto r tiofl ___ , . tildes with coati'a- rotat ing than wi th c o rotat ing dis to_ ti ou . Wi th eel.orati on , to ler a n c e at low r o t _ ttiofi a l f reque n c y wtls nearly cofnparabl _ to that with st ea dy- s tate diet e r-lion. Ti tt l e Was a

I

. / s h ar pd e cr ease Int o l era n ce t o co ro t a ti ng di s t o rti o n a t ro t a ti o n a l f reque nci es a b o v e 6 0 h e r tz, w it h th e di s t o rti o n _ tmp li t ude a t stal l d e cr eas ing t oa l e v e l some w hat b e .

low t ha t assoc i a t ed w ith th e pul se d i s t o rt io n w h en e o r o t _ tti on f r equen c y w _s a b o u t 76 h _ . For th el _ t leu l _ r e n gi ne_ m d it sope r _ , t tlng (, _ n d it io n s, a e o ro _ t lon fr e_ quen cy o f a bo ut V 2 h _ rt z c orr es pon d s to h a l f rotor spe e d . T h i s i s _ p p r ox i ma t o |£ t he fr _ q u o.cy of a si n glo- zo r _e r o t a tin g s t _ , l l, T h e re su lt s pre s ented in flguro s X -2 8 to X-3 Z i ndic a te t h at thi s fan o co m pre ss or s yste m had t h e Ioa _ t tol e ra n ce f or ( I ) a eor o tat ing d is tortion at a r ot a t i o na l fre - que n c y o E about ha lf -rotor s po _ , and ( _ ) t h e pu ls e d i stortion, Th e re su l ts su g° g o _ t t ha_ d t_ tt,rt i a n t o le r an ce t B a f u n ct i o n n o t a ul y ot tu s t au t eou_ di s t o rti o n lo wl : ; but a ls o rate of change o f inlet pre s sure a _ d dwe l l ti m e o f t h e-fa n- c om pres s or bladin g in the low pres s ure regio n of an e ngh _ o-tnlot di s to r tion, : I

TE M PE R AT U RE DISTURBANCES i

I : Alth ough no i n te ns ive i n vestigation of te m perature disturba nc es was conducted , i so m e data llave been obtained s h owing th . _ effects o f steady - state t e mp e ratu r e dis- I tortions and te m per a ture transients on the stall chara c teri s tics o t th e variable g _ - o m etry turbofan. This w ork i s limited, b _ ttthe resu l ts ar e sufficiently in t ere st in g i .

to be included h er e i n , i! ' Tests were conducted to de ter m ine the effect of th _ extent a n ti m ag n itude of a l spatial temperature dis t ortio n an d t h e e f fect of extent , m a _ , _ litu de , m i d r _ .t e of ] i ch an ge for ti m e - varia n t te m p e rature di _ ttirba n ces. A gaseo t is hydro g en fuels d i heater was us( _ d to p r oduce both the spatial and ti m e - vari an t te m perat u re dist., _r b - "

• l

: ances. A photograph of this h e at e r is shown h _ figure X - 33. This view is looking , : fr o m the heater i n le t , dow n stream towards where a n e ng i n e inlet wo u ld be loc at ed, t The heater is divided into four co n trollab l e 9 0 0 sectors , each containing an array of V - gu tt e r flameholders an d sw i rl can pilot bu r ners. The e ngi ne inlet air was heated by b ur n i n g hydrogen _n_ heV- gu tte r f l ameh o lders.

Effect of 5 tlal Tempel'ature Distortions "

• _ The most se n sitive c om p on{ } nt of th e compressorsystemwitheith e r spatitLl or ti m e - variant di s tortio ns w as the high pressure c o mpress o r. This se ns itivity to spatial di s tortion is s hown In fi g ure X-34 which p l_ esents a conventional pressure rati o - flow rate m ap f o r the high - p r essure compresso r . For b o th 180° a n d g 0° circu m ferenti a l di s tortion s , shown by the symbols, the loss in stall pressu r_ ratio ' ,, 3d7 va r i e d-di rec tly with d is t ort i n n m a g n i tud e a s i n dic a te d by'th e lev e l s of A ' £ s h o wn, Th _e A _ T va l u es a r e t he d iff e r enc e betw ee n th e t em pe r atur e ot t h e he at e d and un- h e at ed p D rtion s of t h_ e n gi n _ i n le t flow. F o r t e m p e r a tur e di fferen t i al _ i n e Xc e n s o f 100 _ F _ eo m p r e _ e + o r _ tall o ccurr e d o n t he und is t o rt e d o p er_ Ll ng linep t h u s co m - plet e ly e li m i n ati n g t he op e rati ng m a r gi n of t h e en gi ne , T hes e r es ult s sho w that thi s e ngine was m ore _o n _ iti vo t o t h e ma lutitud _ o r' _ T o f a s pati al di _toP ti o n tha n it was to the clx ' c um fe r ontial extent , at ,,e as t fo r th e 90 ° a r_ 1 8 00 extent s te s t _ .

Effec t ofTemperature Transients

Th e eff e ct of time va r iant to m pC l ' a t ur o dt s turba t_ ces on t his s ame e nl _ n e a r e s hown i n fig _ t re X-3 _ which i s a plot of the r ise in e ngin e inlet t e m p era tur e durin g t _ tt'a n_ i e nt as a _ unetton of th e rate o f cha ng e t _ t Qngi_ e inlet temperature. The s y m - bol s s pe c i f y the t hr ee circu mf ere n ti al ext ents t h a t were i m po s ed at the e ngine h x let.

All of th e s talls en c ountered durin g t h e s e t r an s ient s were in itiated in the hi g h pr e s- I' _ ure co m p r es s or. For all three circu m fe re nti al extents tested, the minimum r ate , , of cha_ n .gc in engi n e i _ let t em perature as s ociated with e n gin e staU was a p proxim a te l y _+ 2500 ° F per second. A ri _e in inlet te m perat u re of app r oxim a tely _ 9° was ass o- i el at ed with stall at this rate o f c h . _ n g e, Thi _ value of AT i s s i[ _ nificantly less than t the AT required to stall U le engine on its op e rat in g line as Sh oWno n th e ordinat e for th e 90° and 180° steady-state spatial distortions. As hi gh e r rate s of change wet + e impo s ed a t t h_ e n gin e in let, high e r in l et t e m p e r at ure rises w e re as _ oei _ ed with stall. Th e r e latio ne llip betw een th e rise an d rat e o f chan ge in e ngine inle t tsm p e ra _ tt r e waS a constant for all e n gi ne stalls obtain e d. T his co n stant w as equal to O.Og _ econd.

The res u lts o t th e t r ansient tests ind icat e that fox"this partic u lal, e n gine: (1) th e r a t e of cha _g e in en gi ne inlet te m pe r atur e w as a pri m_ try factor o n the ris e in inl et , te m pel'ature or AT at stall; (2)the high pressure co m pr e ssor was th_ li m it in g co m_ po n e n t; (8) circumferentia l extent had little effect (fo r the range in v e stigated) O n the , e n gine stall limit; and (4) a cormtaflt en gine r e spon s e o r t i m e to stall wa_ ob t ained regardl es s of the rate of change in eng ine in let t em pe r at ur e a t sta U .

With regard to tile effects of co m bined _ re e s ure and te m pez'atur e eff e ct s o n e ngin e stability, it _ n probably b e said that a sudden inc re ase in t e m pe ra tu r e al mi [ _ _, with a sttddefl decreas e in pr es su r e at th e in l e t to an en gin e would hav e mo re detri- mental e ff e ct on engin e stall th ai1 eith er ' of th _ se by themselves. This concluSi _ h i S k m..Sed on these t em p e rat u re transient res u lts plus t h e p t 'es S ttr e pu l s e res ul ts pre- viously discussed. I / a sudden increase in te m perat U Pe were coup i t _ i _ #ith a sud d en increase in p re ssu r e , th e effect Would likely be less tha n a te m perature t ransient : _ by itself.

CONCLUDING REMARKS

Now that the results of so me o f the N A SA programs in the area o f inlet engine disturbances an d their effects on engi n e stall perfor m a n ce h _t Ve b een exam in ed , it can be a pp r eciated what a co m plicat ed field it is. It ca n also be app t seciated h o w : i m portant it is to d e ter min e wh at is going on inSide each type of engine and partic - • ularly to determine where stall is originat ing so that it can be ascertained where " :' the wea k link is. O n ce this is known, then , of co u rse, it is mu ch eguste r to l : e m edy the situation.

SU MMARY O FRE S UL TS

The, results o f the in v e eti &_ ,ttion of the effe ct of inlet environm _ m _ ,l f _ ctors tliat disturb engin es s how th a t, fo r s teady- s t at e p r e ss ure di stortio _ s , ( 1 ) t he ct1 - _ - fe r enttal disto z' tton e ff ect wa s most severe , (2) th e w eak link i s not _ gt _ lly t h e fi r st s ta g e o r f ir st compo n e nt t h a t sees the dtstu r b a m ce , (3) t he distortion tolerance . . . o f turbof an s is significantly affect ed by alti tu de , an d (4) interac k i o n S b e tween turbo- fan co m pressor components have s igniflca _l t effect s o n stall. ,q3so, _ or dy l la m ic pressure dis tu rbanc e s (1) rapid pressure chaflg e S withou[ distor _ o n s earl stall al l e ng ine, (2) compressor s tages adjac e nt to l ar ge volume s go thro ugh th e gr _ ee t -.... , : p r e s sure ratio e x cursions , an d (3) co m bi ll_l p 1 " essure oscillations and disto r tions . . can have a greater effe ct th an either i r l di vidually, -- t

BIBLIOGRAPHY '

!

B a _m b t c k, Robert J. : Device f o r P r od u c _g Dynamic Dis to z d t o n P _t tte Pn S at Inl et s o f Air-Breathing Eng i nes. NASA TM X-20 2_ , 1 970.

Bel l man , Don a ld R. ; an d Hughes , Donald L. : Tl l e _'l ight In' vesti _t flon of Pressure : ' ; . Phen o m ena i n the Ai r In t ak e of an F-111 A Ai_ ia h e. Paper 69-4{3_ , AI AA , • June 1969.

Br al thwatt e , Willi s M. ; Dicu s , John H. ; and Moss , John E. : Evaluation with a Turbotan Engi ue of Air Jets as a S teady-State Inlet Flow D i stort i o n Device.

NASA TM X-1 9 5 5 , 1970.

i Bratth W aite , W i llis M.; an d Voll m ar , William R. : Perfor man ce an d Stall Limits of a YTF30- P -1 Turb o fan Engine with U niform In le t F10w. N AS A TM X - 1803, 1969.

Burcham , Fran k W., J r . ; and Hughes, D o nald L. : A nalysis of In-Flight Pres- sure FluctUations Leading to E ngine C o mpressor Surge in a F - 111 A Air- plane for Math Numbers t o 2.17. Paper No. 70-624 , AIAA , June 1970.

i ' Calvert , Howard F. ; Braithwaite, Willis M. ; an d M edeiros, Arthur A. : Rotatin g - Stall and Rotor-Blade-Vibration Survey of a 13-Stage Axial-Flow Compressor _ in a Turbo j et Engine. N AC A RM E54J18 , 19 5 5. _ Harry, Dav id P . , HI; and Lubick , Ro bert J. : I n l e t- Ai r Dis tort ion Effects of Stall , Surge , an d Accel e rati o n Margin o f a Turb o jet Engine Equipped wit h V a riable Iii i Compressor Inlet Guide VaneS. NACA RM E54K26 , 1955. _ Huntle y , S. C. ; S i vo , Joseph N. ; and Walker , Cu rt i s L." Effect o r Circumferential Total -P ressure Gradi e nts Typical of Singl e -Inlet Duct Installations on Per- i for m ance o f an Axial-F l oW Turbojet Engine. NACA RM E54K26a , 1955. i . ! _ Ki m zey , W. F. : An In veSt i ga ti on of the Effe ct s o f Sh o ck- in duced Turbu l ent Infl ow I_ on a YJ93 - GE-3 Tttrbo j et Engin e . AR C ), Inc. (AEDC-TR-66-19 8 , DDC No. i : i AD-377312L) , N o _ . 1966 . I_ Langston , C. E. : D i stortion T o lerance - By Des i gn In stead o f by Acc i dent. P aper : _ • . 69 - GT-115 , ASM E , M ar. 1969. , , Lubick, Ro bert J. ; and Wallner, Lew i s E. : Stall Predi ct io n i n Gas-Turblne E n - gines . J. Basi c Eng., vol. 81 , no. 3 , Sept. 1959 , pp. 4 0 1-408. t McAulay , John E. : Effect of Dy n a m ic Variations in En gine-Inlet P ressur e on the Compressor Syste m of a Tw in -Spoo l Tttrbof an E ngi r L e. N AS A TM X-2081 , 19 7 0.

Meyer , Carl L. ; MeAulay, Jolm E. ; an d Biesiadny , Tho m as J. : T ecl m iqtte for In - duc in g C ont r olled S teady-State an d D y lmm ic In le t P r essur _ Di s tut'bances for Jet Engin e Tests. N A SA TM X-t94 6 , 1970.

Povoln y , John H . : Stall an d Disto rti on Inv e stigat i o n of a YT F 30-P-1 T u r bof an Engine. P re s ented at the U S AF A _ r o Pro _ t fl s i on Lab. AtHrame,Propulsion ., Compatibility Syniposiun _ , Miami Beach , Fla. , June 24-26, I 9 69.

Reid , C. : The Response o f Axial Flow Compre ss ors to Intake Flow Distortion.

Pap e r 69-GT-29 , A S ME , Mar. 19 6 9.

Rudey, Richard A. ; and A u tl, Robert J. : The Effect of In tet Te m pe r atul'e Di s - , tortion on the Perfor m ance of a Turbo-Fan Engine Co m pres S or S y s t em .

Paper N o . 70-625, AIAA, June 1970. i Russey, Robert E. ; and Lubick, Robert J. : S o m e E f fects of Rapid Inlet Pressure !

Oscillations on the Operation of a T ur bojet _n gine. NACA RM E58A03, 1958. Ii Smith , Ivan D. ; Br al th w aite , W. M. ; an d Calv e rt , Howard F. : Eff e c t o _ In let-Air- Flow Di s to rt ion s on St e ady-State P erforma n ce of J _6 -B- $ Turbo j et l _n gi n e.

NACA RM E55109, 1956.

Staff of the Lewis Laboratory: Central Automatic Data P r oc e ssin g Syst e m. NACA TN 4 2 12 , 1958.

Wenzel , Leon M. : EXperimental In v estigation of the Effects of Puls e P ressure Distortions Imposed on the Inlet of a Turbo f an Engin e . N A SA TM X-1928 , 1969.

Werner , R og er A. ; Abd e lwahab , M ahmood; and Braith w aite , Wil / is M . : perform - ance an d Stall Limit s of an Afterburner-Equipped Turbofan Engin e with and Without Inlet Flow Distortion. NASA TM X -1947 , 1970.

Winslow , Larry J. ; Wendland , Daniel W. ; S m ith , Brian D. ; an d Welliver , Alb e rtus i!

D. : Inlet Distortion In vestigati o n - Upstream Engitte I nfluence and Screen _!_ Simula ti on. Boeing Co. (AFAPL-TR - 140, DDC No. AD-847095) , Jan. 1969.

t r : i i .

N ASA FRC F - IlIA AIRPLANE

_ T _ : : _ !

......

Figure X-1 I q ' : TIME HISTORY OF ENGINE INLET PRESSURES " _ F-111A FLIGHT DATA ) M- 2 . 17;440 t J0 FT;LEVEL ACC EL ERATION i_ • . - . . _ - STALL INITIATION _ , i PRESSURE, 19. 5 STA _ , i; PSIA 18. 5 1:30 I_I PRE SSUI _ 1 9 , . PS _ 18 1 1 0 .1 .2 .3 .4 TIME , SEC cs-s 69 6 1 figureX- 2 ENGINE INLET PRESSURE CONTOURS AT STALL : F-IllA FLIGH r J) ATA; M- 2 . 1 7 ;44000 FT Ptev •o . s _ Pr a y . O .

i

• i i TIME A VERAGED IN S T A NT A NEOU S AT STALL INITIATION ' PRESSURE RECO V ERY MAP S cs . s 696 o Figure X - 3 "'_,, P S D FOR ENGINE INLET PRE S S URES D U RING D U CT RESONANCE . , F-111A FLI G HT DATA; M - 2.00;45000 IT; STEADY ST ATE r Ill o - i ,, . , ...... • .....

g \ ENGINES I N VESTI G ATED COMBUSTOR w _ COMPRESSOR . _ i r TURBINE SIMPL E TURBOJET WITH AFTER B URN E R C ]I _ '_ " I _ "1 _"_ AFTERBU R NER J# H . P . COMP R ES S OR 7 r "FAN DUCT : , L P . C OMPRE S SOR " 1 / / r - C OMBU S TOR , FAN - _ ,, J FIXED-GEOMETRY TURBOFAN WITH AFTERBURNER i

I'!

H.P.TURBIN E _ l _ - L P . TURBINE !

FANDUCT _ r- I.L P . COMPRE S SOR , WITH AFTER B URNER AFTE R BURNE C OM BUSTOR / u L P. TU R BIN E t L H . P.TURBINE F i gure X - 5 c s - s 699_

EFFECTOF STEADY- S TATEPRE S SURE

DISTORTION

SIMPLE TURBOJE T F UN D I S TOR T ED STALL LIM ff l OO k CORRE C TED SPEED PRESSURE RATIO, PEXIT / PINLET 5 } - - j _ NORMA _ -- _ ' 4 _ t _ 6 9 8 40 42 4 4 CORRECTED AIRFLOW, LB / SE C ce-s6963 1_ , Figure X - 6 i

• EFFECTOF EXTENTOF CIRCUMFERENTIAL S POILING

_ ' ON CO M PRESSOR PERFORHANCE _ -

S IMPLE T U RBOJET '_ C ORREC T ED _: , : SPEED.

• 10 - I! t

"" ° ° 1

LOSS IN .06 _ i STALL PRESSURE 8 7 _ , -_ RATIO .04 _

.02

s

0 _ 0 _ 90 ]20 1_ 1 80

SCR_ :£N A N G _ I _ D EG c s - 569(,4 Figure X - 7 a36 i • CIRCUHFERENTIAL DISTORTION CORRELATION .14 - SIMI & E TIIEBOJET .12 - o_ ' _ ° ,1 0- a n , , _ LOS S INSTALL , 0 8 - S C PRESSURE RATIO.06 _ OPEN ARF A o "_ .. . 0 36 0 CIRCUMFERENCE, ,04 o 5 7.4 ' l _ 4 9 8 o e , .75 .80 . 85 .90 .9P 1.00 R AT I O OFMINTOMAXPRE S S, , CS- 569 6 5 FigureX- 8 ii EFFECT O F 180 o CIRCUMFERENTIAL _,• ' DISTORTION O N FA N !

ff

3 . 0 -- FIXEO - GEOMETRYTURBOFAN _ : , BLEE D S C L OSED _, 2 . 6- ReI- 0 . 5 / _ t

i

/ FAN TIP 2.2- /

! " PRESSURE STALL / / , OPERATING LINE . . i"

L IM ff S _ / / / g A p I p . _ " ,_ .0 _ // . .

I. 40. 016 _ ' I _ " --- --U N D ISTO R T ED INL E T . . .0 _ " _ _ " ------- SCRI_EN DIST OR T IONS i 1.0 1 . ...I _ " I I I I 60 7 0 8 0 90 I00 CORREC TED F A NROTOR SP E ED , %MILITARYcs . s 696_ X'9 Flgore 1 - - _; _ ' -* _" ..... _ " _r .... i i i _ 1 1 • iiii i| "- i .... M .... r 1 - i i _ ,' , \ EFFECT OF 1 8 0 ° CIRCUHFERENTIAL DISTORTION ON L.P. COMPRE S SOR 2. O "-. FIXED _ E O ME T RY TURBOFAN + BLEE D S CLOSED / / R e !, 0. _ /// i , */ 2.2 - . _// S T ALL LIMITS ., , s L P .C. P I P • 0,040 PRESSURE 1.8- RATIO ,, _ ./ _" ' _ P I P - O, 034 I " " "" OPE R ATING LINE 1 . 4- . . , I ---- UNDIS T ORTE D INLET -'- " - S C REEN DIST O R TIONS LO _ I I I I 5 0 _ 0 10 SO 9 0 100 cs.+ 69+7 C ORRE C TED L P. C .RO T OR SPEED , + MILITARY F i gu r e X- t O k. " i EFFECT OF REYNOLD S NUMBERON D I STO R TIONTOLERANCE FIXED GEO M ETRY T URBOFAN + .10 B LEEDS CLOSED R e I C IR C UMFE R ENTIAL DI ST ORTIONS /, 1 800 F 1.0(E S T)

.0 s +-- ' / o .s

DIST O R TION .06 L .-,,. o ' _ ATSTALl., 0 I I I I I I 70 7 5 SO _ 9O 9 5 100 CS - 5 6 9 6 8 CORREC_ D FAN RO T O R S PE E D , S M I L IT AR Y Figure X - l!

33?

EFFECT. OF ENGINEINLET Re T

H,P . COM P RE S S O R P E R F O RMANCE VARIA B LE GE O ME T RY TURBOFAN 1 80 - O PERATING LINE _ , _ , ' _ _ _'' " RAT E D STALL s"_ _ % SPEED j _O F 1 0 0 - LIM ff_ - , " ,- _ ,, # PRESSURE 90 - , ,

, A T Op > r' --=

RATIO 80 = .,. _ 1 1 70 I -- - - FAN INLET ReI• O. 41 " - -- --- FAN INLET ReI- O. 18 .,1 " I l I I I 65 70 7 _ 80 85 90 9S 100 O FRATED CORRE C TED AIRFLOW c 8 . _ 6 % 9 F i gu r e X - 12 i

EFFECTOF . ENGINE INLET ReI ON VELOCITY

PROFILE S !, 14. P. COMPRESSOR INL ET ' , TIP tO0-- VARIABLE GEOMET R Y TURB O FAN , ,_ , 8 0 -- _ I , OF 6 0- ' Ret - O. 41 , PA SS AG E 1 HEIGHT 40_ _ jb, s

i

_ I I I I I , ,,,I

HUB 0 10 0 20 0 300 400 _ 60 0 cs-s 697 oHIGH PR E S S U R E c O M PR ESSO R INl. _ r VELOCffYo F ps Fi g u r e X-13 3 3 8 oft-°:."_ .,>,{ " o EFFE C T OF HUB R AD I A L INLET PRE SS URE DISTORTION ON FAN PERFORMANCE R e !, D,41 a V A RIA B LE G E O ME T RY T U RBOF A N 120 - AP I P STALL LI Mff -,, 0'139 _ , _ '_ , RAT ED _ , 1 10 " , , . ," x!_ e COR R ' RAT ED _ - j I 0 .1 1 8 - ] _"- _ _ _ 1 0 0 PRESSURE R A T I O q O , - _ " - - .,., ,, .. , "_ " I q 5 ,, , _ "_ OP E RATIN G L I NE 8 0 " _" " _ 85 "-- U N 0 1 STORTED INLET ---- -.- DIS T O RT ED INLE T

_ O- ,I I I I

7 5 80 8 5 9O 9 5 IO0 OFRATED CORRECTED AIRFL OW cs. s6971 , , F l gu re X -t4 . _ : TYPI C AL H I GH RESPONSE PRESSURE T R A C ES _ , OBTAINED DURI NG AN ENGINE STAL L ' VARIA B LE GE O ME T RY TU R B O FAN _ HUBRADIAL DI S TORTIO N _ . FAN INL E' t" PRESSURE R O TATING STA LL n S U RE + ST A LL " ' OU T LET PR I.-'S SU RE TIME _ _ O . 10SEC _, cs . 5_ , 97 z Figu r e X - 1 5 i ; . , 339 ....... I t

AHPLIFICATIO N OF CIRCUHFERENTIAL DI S TORTIONS

FANHUBREGION VARIABL E GEOM L'TRY TU R BOFA N ! , 3 . 0 - o CI R C UM FERENTIA L D I S T O RTI O N _.

CIRCUMF E R E N T I A L &TIP 2 . 5 _, _' . _ , + , R ADIAL DISTO R TION AMPLITUDE 2,0 - _P q RATIO , 1.0 - .,,.q . '; \r / FAN , IN .5 - 1 .00 1.0 5 1.10 1 . 15 1.20 1 .2 5 !

cs . s t , 973 FANHUB INLET TOTAL TOSTATIC PRESS RATIO r r' FigureX-1 6 I,

FAN HUB EOUIVALENT STAGE

CHARACTERISTI C CURVE

.7 -- VARIABLE GEOM ETRY TURBOFAN OPEN SYMBOL - NOT BEHIND DISTOR T ION SCREEN , ' S OLID SYMBO L - BEHIND DISTORTIO N SCREEN t PRESSURE .; COEFF

. 6 -

.5- / .4 I J I I l ,,,I I ....

' O .1 .2 .3 .4 . _ .6 .7 FLOW CO E l e F c s - s 6 W 4 Figure X-17 84 O I) EFFECT OF FAN MODIFICATION ON THE FAN HUB EQUIVALENT ST A G E C HARACTERISTIC CURVES VARIAB LE G EO M ETRY TURBOFAN . 7 - la MODIFIED .6 '- • ,_ P R ES S URE COEFF t ' I OPEN SYMBOL - NO T BEHIND / __ / _ _ OLIDSYMBOL. B EHIND . _ , _ : _ . I DISTOR T ION SCREEN _ " 51 ORIGINAL FAN // DI $ 1 'OR T ION SCREEN ..... 0 .1 .2 . 3 .4 .5 ,6 .7 ' FLOW COEFF cs - s697_ FigureX - 18 • EFFECT OF FAN MODIFICATION ON AMPLIFICATION OF ClRCUHFERENTIAL DISTORTIONS FANHUBR EGIO N .... _ :. VAR IA BLE GEOMET R Y TU R BOFAN . : i : : _ , 3 .0 - o CI R CUM _ RENTIAL DISTORTION _ t 2. 5 _ " , ,.. _ , _ _° a CIkCUMFE R ENTIAL & TIP AMPLITUDE o , RADIAL DI S TO R TION e, ATm, 2.0- "_ , 6 1.5 - FAN FAN, OUT FAN, IN . _ _ I_ 01FIED i ;AN _ , IJ I _ I I , I - I i ;

Z.00i. 0 .S L10 L _ L 2 0 LZ 5 : i

cs - 5697 6 FANHUBINLET TOTAL T O STA T IC P R_ [SS PA T : IO " FlgurkX-19 • ¢ TYPICAL J - 85 NACELLE I N STALLATION IN IO 'x IO ' SWT .... Figure X - 20 c '_ AIR J ETS IN ENGINE - INLET DUCT Fig ureX -21 V AIR -, JET S Y S TEM S CHEMATIC

I V A LVES I

If,0 - _ Hz)J

PRIMARY FLOW ,, e , FLOW '0 SECONDARY _ IP, TOTAL , ENGINEFACE, !

I - ( 1 ,

• I

' f i . i C8- 5 6 97?

, Figure X-22 !

I" J: L i, POWER SPECTRAL DENSITY FLIGHT ENVIRONMENT (F-11]b ANDSIMULATION lAIRJETS - PSL) q PSD , 1 0 - 5 PU T ) 10"3- _ UT t

_PAV / 1 o -7 - , - ._

" I '_ % FLIGHT E NVIRO N MENT II N PUT) io-9 i I ' , I I _.

0 ._ 100 1 5 0 2 _ 0 FRE QUE N C Y, HZ c s -_ ,69 _ e , Figure X - 2 3 :/

COMPRESSOR RESPONSE TO INLET UNSTART i

!

7_ _ • 2.5 , _ CORRECTED SPED ....

PEX ff / PINLET - - _ "V _ r..... '- '--ZE R O o l sTo ,Ts o .

3 | V,, ST A LL LINE I IUNSTART AT T 0 .8 E " * ' "_ j _ P I NLET / P A MB I ENT .4 / L.

0 .01 .02 .03 TIME, S E C c s -s 698 o Figu r e X - 2 4 / INTERSTAGE STATIC PRESSURES: INLET UNSTART - COMPRESSOR STALL . STAGE EXIT 7 " -'- "_ SIMPLE TURBOJET 3 --_.\ C OMPRESSOR INL ET - _ --- ' .- _ INLET UNSTART - ' - _ C OMPRESSOR STALL , L,,I I ;I I I I I

0 204o6 080

TIME, mSEC cs - 56 9 8 z ' , F i gure X - 26 I J : .' INTERSTAGE STATIC PRESSURES: . ,L COMPRESSOR . STALL- INL E T UN S TART ' :"_ STAGI _ EXIT 7 _ 1 ' - SIMPLE TURBOJET I I _ , _ " , _" 6 t ' 4 , COMPRESSOR INLET _ ,,,, _ J __ .. INL E T U NS T AR T --.,. _ J _ . , A / .

, COMPRESSOR STALL - - 4 "1 _ VVV , I I _ lll J J . I I t ' _

o _ 40 6 0

TIME, mSEC CS .S6983 | , _I g u re "X" 2 7 8 43 _t, q r

PRESSURE R A T IOHISTORIES

lO-HzCYCLIC WITHOUT DIS TORTION _ FIX E D GEOMETRY TUR B OF A N s BLE E DS C LO S ED _ R e ] . O , 5 2 . 2 r - f_ H U B , .... S TEADY- S 'rATE 2.0 _ /'./ "_ / _ ' _ FAL N_ EL S

I / "' ' ""

// STAGE GR O UP " x "" ' / TIP 1 . 8 : ,, PRESSURE RATI O S 1 . 6

=FRONT

- .. , _ R _R , - _

1 . 8 _ HP C ...

1 .6 I" - _.. _ _ F R O N T- __EAR

0 40 8 0 120 1 6 0 2 00 , ' TIMF _ mSE C C8-56986 ' , i Fi g ure X - 30 is

TOLERANCE TO OSCILLATING 180 ° CIRCUMFERENTIAL _

._ DI S TORTI O N I , _

i _" .08 - OUT'OF - P H A SE O _ CIL L_ ,TING ' 0 I

?i .o _ - I

.... DI S TOR 1 10N ST UD YSTATE _ " AMPLITUDE A T STALL , .04 - P ULSE " /// ,," , A p / p FIX[_ D dl t O METRY TURBOFAN . 02- BLEEDS "C LOSED

Re l - O. 5

, 1 2 4 ? 10 2 o 40 _ 100

FR EQU[ ' NCY , HZ c s- s 6 9 8 7 I Fi g ure X - 3 1 :, T 3 4 7 - TOLERANCE TO ROTATINO 180° CIRCUHFERENT _ AL DISTORTIONS .08 - CONT R AROTATING .06 - - COROTATING o DISTORTION STEADY STAT E - _" -' _ AMPLITUDE.04 - ATS T AL 4 _ p i p PU .0 2 - FIXED G E OMET RY TU R BOFAN BLE E D S CLOSED ReII "O.

o J , I I n J I ,_

' 1 2 4 7 10 20 40 70 100 ' FREQUENCY , HZ c s- s 6 98s i , J Figure X - 32

fl

r ' T E MP_ ' R A TURE DISTURBAN C E DEVICE t i !i _ t _L l ,_ C- 69 -1384 ,, i C 8- 56 9 55 I " Figui'e X - 33 EFFECT OF S PATIAL TEHPERATURE DISTORTIONS ON ENGINE S TALL VARIABLE GEOMETRY TURBOFAN _ H . P, COMPRESSOR MAP 105 - STALL LIMITS 7 I""_ ' _ , _ T" UNDISTORTED AT. // _ ,..** _\ OPERATING 9 5- D E 6 .,,j _ / .. c ; .. _ I . I N E O F ,.. . . ,, . , , " _ ' _"_ . _" _ CIRCUMF E RE N TIA L RATED PRESSURE 85 . 4 " . , _ _r' t . RATIO .. = 70 . ,,, _si " ' _ gO _ -- ",. l r _ _ <- I a 180 - 7'- "', _11 _ ------ U N OI ST ORTED DISTORTED 6 s I _, v [ I I I I I I , 6 5 70 74 78 82 8 6 90 94 98 c s-s 69_9 ° k O FRA T ED CORRECTED AIRFLOW Figure X - 34 EFFECT OF INLET TEMPER A TURE TRAN S IE N T S i , ON ENGINE STALL VARIABL E GEOMETRY TU R BOF A N " _ - 120--I S.S. L IT REQUIRED ' , : . _t_ ) FOR .H P C S TALL " lO0 -J ( 90o AND 180 o DIST O RTION) t 80 - _ CIRCU M FERENTIAL RUSE INENGI N E EXTENT , I N LET T EMP , 60- _= , 0 D EG DEG B o _0 _ - ,, " I S O a & 90 . • 2O - OP_ SY M BOl. - N O STALL . o CLOSED SYMBOL - STALL cs- s u _ o R A TE OFCHA N GE INE NG I N E INLET T _P , DE G I S E C | F l gbr e X - 3 5 ) . ,\ : 349

X I . D YNA M ICS AND CONTROL N 7 1 " 1 9 4 6 2

AaronS . Boksenbom, Gary L. Cole , Daniel I. Drain, " " _ '

Kirby W. Hill e r , Ross G. Wllloh , and John R.Zeller

Some o f th e m ajor problems Ill d y na m i cs and c o n trol s o f a s up e r s onic propul- sion s y s te m arc shown in f igu r e XI-l. The Job of t h e control (inlet control , engine control , and c o u pltno _ s ignal s between the i nlet a n d the e ng i n e) shown i t, the f i gure i s to avo i d se' , _r e , d am p. g l ng , or un s table co n ditio n s th r o ug hout the s y ste m , T h ey also m ain tai n system o p erat i on at the d e sired or opt i mum cond i tion a n d, when re- q u ired, move the system from one o p erat. 4 .n g point to a n o th er. To design st l ch con- trols requir es a n u nders tan din g o f the dyna mi c, as well as th e static , perfor m ance of all syste m co m ponent s .

A st lmma ry of the su bje ct s p resen t ed i n th is p a p e r is as follows: (1) For the su p erson i c inle t, the problem is to m aintai n h ig h inlet pe r f or m a n ce, while avoidin g inlet unstart a n d avoi di n g i n let oper at ion th at produces high dy nami c d i stortion a t th e compressor face.

(2) An Inlet u n start would be a very severe experience for a sup e rs on ic trans- port. This uns ta rt m ig ht induce combust o r flameout and compresso r stall. Th _ _ : proble m i s to recover th e system back to n or m al operation In a sho r t t im e. i!

• / (3)The inlet a n d e ng ine con trols may requir e e oup l in g to set syste m op e r ati n g { ' , _ , conditions for best overa l l pe rfor m ance.

(4)We need an u nd e r s t a n_ng o f c orn bus to r a n d co mpre ss o r dynami c s. C o m- p ress o r perfo r man ce is es pe cial ly s e n sitive to dy n a mi cs. The p roble m is to p _e _ . il dict a n d u n ders t and the c om pressor stall m argi n availa b le mlde r dy n a m ic and s_ ti c I i inl et a n d compress o r c ondi t ion s.

(5)T o c on duc t sig n ificant e x p erl m r#n tal researc h, s p ec ia lly devel o ped se nso rs a n d s e rvo ac tuators a r e n eede d . T hese d e v ices sh o u l dpermit m or e d i rec tm ea s u re - m e n t of inp u tv ar i a b lesa n d fa st er m a nlpu la tlon of ou tpu ts.

( 6)Th e c ompl e x i tyo f s yst e m_ a nd the ir requir e m e nts m ay r e q uire mor e ca pa b le a n d v e r sa til e c ontrol c omputers. The u se o f digi t a l c ompat e r s in system c ont r ol o ffe rs a so l uti o n to thisprobl em.

(7 ) Fi na l , _i ; , t o ob tain the ' u tmos t impro v e m e nt s i n vario o s asp e cts of m ea _ u _ ' e - m e n t _ mon itorin g , a n d con t ro l , th e body o f m athe m atica l c o r lt rol th e o ry m ay be ap- pl i e d and i n corporat e d i n the co n trol syst e m.

P R ECEDING PAGE BLANK NOT FILMf_D SSl • o • . • _ H

SMALL P E RTURBATION DYNAMICS AND CONTROL

For sm all p e rturbation d y _ u _ Ic s a n d co :_ t r ol _ a n _ n _ lytle or co mpu te r dy _ r v . l c mod el Is ne e dedt o help us u nd e r s t a nd Inlet b eh a vior a nd to aid Inde s igning inl et co n .

trol a nd s t _ bili _ tion d e vic es .

Dynamic Analysis

When we st ar to d thinking abou t c o n trol s f or supe rs o ni c pro p ulsion s y s tems p th e desired si mu lation mode l s Just w er e not ava i lable, This w a s e sp e ciall y true fo r i f requ e ncies abo v e 5 or 10 hertz. One of the _ irst things done w _ s to model the y a h ' i- i ous pie c es of th e propulsion syste m, beginning w i th th e inlet. Y Figur e XI-2 shows how th e inlet appears to the controls analyst. H e thi n ks of !i i t a s a pipe wi th a variable area s ection, a normal shock, s u m s di s turba nce or by- i pass _ low , and an a ppropriate ter m in al bo un dary conditio n . Physic a l l y , a perturb a o 'i tion f ron _ equilibriu m caused by the e n gine or bypass doors results in pre s s n re W ave d ji_ tions. An approximate clos ed *form fre q uency respons e solu ti o n w a s found fo _ th e m oving u p and down th e duct. Mathe m atically , it can be d es cribed by wave _ lua- tli wave eq uatio ns applied t o the subsonic diffus e r. Also fou n d we r e eq uations relati ng i shock po sition ar i d velocity to adj a cen t va riables. With this m od e l, it was pos s tbl _ ,.

to predict the dynamic beha vi o r of th e inl e t i n r esponse tOdow n stre _ diS t Urb ances, I : _ which ar e th_ primary di St u rb _ tnces conside re d in this paper. EX pe ri m enta l da _ a re needed , however, t o v e rify the reSultS.

Expe rimen tal freqttency r es po nse dat a obtained fro m an i nl et du _ ing a wind tu nnel test program a _ e presented i n f igure XI - 3. The figu re contains a m pli tu d _ ratio and phase lag characte r is ti c _ of th e inl e t nor m al _ hock po s i ti o n if; r( _ spon s e to t bypass airfl ow d i st u rba n ces. The amplitude ra ti o has b e en normaltz t _ i to u ni ty at a frequency of 1 hertz. Analy ti cal res u lts obtained with the snla li pertu rbation m ode ! ; are a lso p l ott ed f or comparison. Agre e m en t betw e e n the an al_ ica l _ experiment al data is quite good eve n out to a _ r e qu e ncy of _ 0 0 h e r t z. T he termi na tio n _ or thlb inlet cons isted o _ a choked orific e at th_ en d of a piec e o t pipe. Tlte pipe l engthwa s _ , equival e nt to th e l _ng_ it of a J-85 e ngine wi th an aft er bur ne r. O ther o _ ific e l o ca - tio n s we r e investigated, and it was found th a t a choked orifice locat ed at th e com- pre ss or face s _ti o n would give resul ts very si mila r t o th o _ e obtained wi th an a etd al e ng ine. Since th e an al y s ts a n ti th e da ta a re i n agree m en t, i t i s fe l t t h at a controls : : / de si gn tool f or i _l ve s tig _ti n g th e s _ a li one-di me nsional distu rb ance occur , ri ng wi _ i n th e Inl _ t h as be e n obtained. I 352 _ C ' _ ',

I nl et Control

The r e qu i remen t s fo r inlet cont ro l a re n o w con s i der e d . I n g_n e r a l , at t he d if .

ftts o r exit of a m ixe d c o m pr essi on supe r s o n ic inl e t , to ta l p re ssur e rec o very i s h i gh - e s t and d i s to r tion i s lowe s t when the n orm al s hock i s at the i nl e t th ro at. T hu s, fr om a n e n g i n e po_fform a n e e s ta n dpo i _x t , it i s des irabl e t o f ix t h e n o r ma l sho c k at th e t hr oat. U nf or tu na t e ly , i n l e t a i rfl ow d is tur ba nce s ca n oc c u r whi ch m ight c a u s e the sho c k to m ov e e it her u p s t re a m or d o w n s tr o atn. - A n ups t re a m di s pla c e men t o f fhe shock w o u l d r es u lt i n a n i nle t u nstar t. A d ow n s tr e a m mo v emen t i n cr e a ses d i st o r- tio n a nd d ec r e a ses p r e sm t ro rec ov ery. Th ese even t s a r e un d e s i r abl e b ee A u s¢ of a pos s ibl e e ng ine up se t such as co m p re ss o r s ta ll a n d beca u se p r ol m l _ion syst em p e r- . !

f o rm a n ce c a n be s i gn i fi ca n tly de c r e as ed. U n st ar t ca n b e avoid e d i n ma n y cases b y op er ati ng sup or c r it i c al l y - th a t i s , w i th the sh ock do w n s t r ea m oi th e t hr oat. I f t h e inlet i s s upp l i ed w i th a nor m al s h oc l' control , t h e d e g re e of s up O re r itic al operatio n i , can be reduced w ith the b enefit o f h i gher per f o rm a n ce, ii The control chost m f or investi ga tio n w as a closed l oop cont r o l t 2 tat a ttempted to Ii : hold shock positio n fixed Ju st a f t o f the thr oat. To help design such a c on trol, t h e _ , analytical m odel of the inlet w as used. Th e model gave open loop a m plitude and i i phase _ aracter i stics s u ch as th ose show n in figure XI - 3. A block diagra m o f t h e !

closed loop shock po sition control system is presented in , figur e XI-4. i_ On e syst em that w a s i n v e st i g at ed u s e d a s h o ck po s itio n sensor that relied on two stat i c pressures i n the inlet to i _e r sh ock po s i tion. T he se S tat i c pressures We r e behind the normal shock. When t h e sh o ck moved f or w a rd th e pr e ssu r e s increased.

: T he resul ti ng s hock po s i ti on signal was co m pared to a c o m m a n d val u e . T he d i / _er - ence betwee n the tw o ai gll als told w h e the r th e shock was a head o t or b _ hlnd it s de- aired l o ca t i o n . Th i s di f f erence sig nal w as th en amplifie d by a contt. _ l l e r . The con - tr ollerts g ain wa s frequen c y sensitive. It had h i g h gain a t lo w f r _l uenc i es , wh il e at I hi g h frequen c ies its g ai n wa s c onst a n t . S u ch control action is ca li ed pr o pm'tion al - plus - inte g ral. The ottt F_ t o f t he c on troller drive s t h e by p ass doors, which were po s i tioned by fast response electrohyd t aulic actu at ors. , F ig a re XI . 8 sho w s th e p e r f or m ance obtain ed f ro m thi s n or m al sho ck co ntr ol t es te d on a m L. x ed co m pre esi on i nle t in the 10 b) 10 F o ot Supe r s mze e W i nd Tu nnel: Til e fr e q uen cy r e sponse of shock l m sitio n to a _low ns t _ a t n a L_ ; ow di _tu rb an c e is given. Two responses are show t l - o ne f or t h e i n le t witho ut co n trol and on e for th e • inlet with control. O n th e y-a, _ is is the a m pli tu de of sltock m ottb n . Both a m plit u de ' respo n ses have b e en n or m al i ze d b y t h e l o w f re q t t t _t tcy a m p litu de for" the un co n trol l ed in l et. Thus th e no contt _ ol curve starts o ut with a n ampl l R _ de ratio of 1.

As fre q u enc y i ncrea se s , th e u n con trol led response d _ op $ o ff a l id the ri i 'eson a t _ s.

It i s re sm m, nt b ecause th e inlet w as ter m i tm , ted with the lo ng pipe in th is e xper im ent.

\ # The sol i d curve shows the r e sp o nse of the i nlet w ith control. Thi s co n trol i s trying to hol _ the shock pos i tion ste a dy. Thus, i t acts to hold sho c k amplitude down to zerO.

At low di s tur b ance frequ e nc i e s i t does this quite w e ll. For exa _ n pie, at a frequ e ncy of I h e rtz th e n or m alized am pli tude ratio of shock motio n is 0.03 w i th co n t r ol.

Wi th out control it i s unity. Thus , th e control has reduced shock m o ti o n by about 30 to 1. At hig h er disturbance _ r eq u en ci e s, the attenuation is not as good. Thi s is due ,, to the fr eque ncy sen si tive gain of the controller. For th e faster d ts turb a nces _ the integral action has less tim e to work. The controller ha s to quit in U l e vici ni ty of ° , th e resonance; o th erwise, it would drive the syste m in to oscillation. As can be seen , there wa s som e a m plifica ti o n o£ th e resonance. The o nl y place th e control i amplified sh ock motion was near th e re so nance; at lower frequencies it pro vi ded " .: significant atte n uation. _

In l et - Eng i n e Control

Until now the discu S Sion ha s been concern e d wi th an inl e t co n trol that maniptt - i_ t lares only the inl e t t s ov e rboard bypa s s doo rs to keep the normal sho ck fiX ed . Such a i_ contro l may not al ways result in the best overall propulslon system perfo r mance. [ : For example , an increa s e in bypass airfloW may be r eq uired tOacco m moda te a dls - _ :I : _ rbance. This would ca use a correspon di ng, i n crease in spillage drag. Thus , even _'i though th e sh ock is maintained at a high pl'es s ure rec o very po s i tio n , there may be a i.

net loss in propulsio n _ ystem efficiency due to th e increase i n spillage drag. Since engine sp e ed also aff e cts sho ck po sition, th ere may be som e b e nefit in usi ng b oth i_ th e bypass doors and th e engine to control the norma l shock. Two me th ods of c o upl- _ : i!

ing no r mal sho ck pos ition control a n d engine s peed con tr ol w e re inv e stiga te d. !

• Th e fi rs t cou pled control approach i s sh ow n in figure XI - 6 . Wi th this ap proa ch '" th e inl e t n ormal sh o ck l oo p is c ou pl ed to th e engine s pee d con tr ol loop. The it e ms added to th e ba sic i nl et control at'e show n b y th e heavy lin es. T he engine s pe ed t co ntrol l oo p sense s engine speed and compares it wi th the d e sired speed a s di ctated by the throttle setting. The error in spe ed is in pttt to th e controlle r . Th e eontr01- ler m odulat e s engine J uet to decrease th e spe ed e rror. Couphng i S acc om plish ed with th e res e t device sho w n in th e diagram. The inl e t co ntr oll er outF a t is int egrate d a n d th e resulting signal is used to trim (he spe ed _ ttii l g of, th e e ng ine. With this m odifit _ t speed setting , the e ng in e s peed loop will reset the control bypass doOrs.

This dy ste m wa s evaluat ed eXper im entally and slto _ vedgood resul ts .

Figure Xt - ? sh ows the transient perf or m ance of the syste m wllen o pe r _ ttt _ t ek- peri me ntally in th e superso _ c wi d d _ rlfl e l. As th e sh ock con tr oller S en d s that a ....... Il l n'll I " 1 .... ' - I' "l "'i , , , " J .....

d i sturbanc e has occurred, the co n trol doors open q u ickly to correct for th e airflow , d i sturbance. I n itially, th ere i s a small upstre am excurs i o n of the shoc k , but this is qu i ckly correc t ed by the action of the doo r s. The output of the shock control le r to the bypass actuat i on s yste m is inte g rated, The output o _ th e inte g rator i s inpu t to th e engine speed loop and results in th e increase l n _u el flow to accelera te th e e n- g ine to a new speed. The engine responds to th i s fu e l cha nge a n d finds a n ew operat - in g speed. As this occurs, th e bypass door 2 ar e slowly r e set to a closed condition, th ereby red u ci ng overboard by paS_ dra g . Coupl i ng th e con tr ols de m onstr_tted the a b il i ty to reset th e pro pu lsion syste m to any arbitrary operating point.

Also i n vestigated was an entirely d i ff e rent coupled control approach (see fig.

XI-8) wh i ch uses eng i ne spe e d as th e pr i m a r y control var i abl e . The ma i n feature s • _ : _ of tYJs sy ste m are o utlined wi th the heavy li , .s show er in th e figure. Th e overboard bypass sys tem in th is ca _ e is a much s i mpler d e s i gn, and th erefore la e h ;s th e fast i_ : re s ponse capabil i t i e s of pr evious sy s te m s. The e ng i ne speed l O op has been i m proved _ , " : i n respon s e by usin g a fast fu e l th rottl ing valv e . L, In th i s syste m, th e e rror i n sh ock position, after it is ac te d o n by the d yna m ic : ': ele me nts of the inl e t controller, i s sent directly to the engin e S p eed loop to den t and ie a c ha nge i n spe ed . The a ctual s pe ed is inte g rated s l ow ly , and th e out pu t of the l lt - !i te g rator is sent to th e bypass actt _tti on s ystem. The performance of this cont r ol i sys tem, w hi ch ha s been used successfully in th e w in d tunnel , c a n be s t b e under b ....... _ s to od by l ooki ng at S o me of th e ex pe ri m enta l re s u lt s, i_ _ Fi g ur e ][1 - 9 sh ow s a series of tr / tn stents obt ai ne d when th e c o nt r olled i nle t- ; . . e n gin e was di st u rb e d i n a p _t rt i c ular w ay. A s t _ p type or almost i n st R nta n eous de- : crease in the i n let d o Wn s tr ea m a i rfl owwa s i niti a t ed b y a dis t urb anc e at th e c om - pre s s o r face loca tio n . T hi s decrease i n ov e rboar d airflo w ca uSe { l th e sh ock to i!

: move very q uick lyin a n u p st r eam di rect i on or t o w ard t h eu n s tart ed mo d e . T he re- _ salti ng e r ror i n th e Sh o ck l oca t ion is se n t to the engi n e s _ eed loop. As a r es ul t , th e t re l a ti vely f ast fuel co n tr o l respo n d s by i n c r ea s ing t he engine S peed to cor re ct for th e tipset in i t flet do Wnstre am airflo W . As e ng i n e a irflow increa se s , th e shock is returned to i t s d _ sired po s ition. Th {_ ch _ e in engine spe ed is integrate d sl owl y.

T hi s outp t tt is th en used as a co mm a t ld to th e by pa ss actuation syste / n. The lo w er t ra c e sho w s tha t the doors o pe n a nd th e r eb y ' : . c t'ea B _ bY l_t ss flo w . _ This inc rea s ed by pas s ai Hl ow reduc e s the a irfl o w r equi red to De talte n by th e e ngttie , thus a l l ow l b g , S peed t o be r e set slow l y to some desired Ol _i ' ati h g p o i n t.

\ • I

LARGE PERTURBATION DYNAMICS ANDCONTROL

Thus fa r th e discus si o n ha s been restricted to small perturbatio n s and n or m a l oper a tio n . If the inlet s hould u ns tart be ca u s e o f inadequate co n trol _ c o m p ressor s tall , or u pstr e am di stu rba n ces , r eturning th e s y s te m back_to normal operatio n p res ents n ew p r oblems.

Inlet Dynamlc Analysls

Considered first is our understanding of th e basic m ec ha nis m s o f inlet unstart.

T he a l_ tlysis d e v e l oped for controls design gave only a s m all per tu r bation solutio n usef u l f or li nearized sy s te m a nalysis. T he tech ni ques could not be ap plied dire c tly to inlet u n s tarts , or a ny of th e o th er nonlinear ph e no m e na that see m character is ti c of supersoni c inlets.

An u nS tart can be ca u sed ei the r by external di s turbances (such as a gu s t or a pass ing aircraft) or by internal dis tu rbance s (such a s co m press o r stall). Wi th an external dis tu rbauc e a s e cond shock wave can f or m at the inlet th roat, m o v e up- s tr_ m a n, and stand i n t ren t of th e inlet. A down s trea m distu rbance , on the other hand , can ca t tse a pressure Wa v e to f or m at the compressor face, move up th e i n- let, join with th e normal shock, and th en carry it out of the i nlet.

In ei ther ca Se, th e u nstart occurs so rapidly th a t it i s ha _ d to ge t detailed ex- perim e ntal evid e nce o f wha t occur s . The decision Was made to try to _ lod e l such inlet b eh a vior analytically.

The meth od of Lax, one of th e finite difference S che m es that ha s been used to s tu dy f l ow field dynamics , was us e d f or th e large perhu _ ttion proble m . So m e of th e results ob tained by applying this t ech niqu e to the g eo m etl _ of an i nle t are shown # i n figur e XI - 10. In this f ig ur e inlet pre ss ure prof i le s ar e plotted at variou s times during a transie n t di B turba t t co . Th e b ott om curVe is the i ni tial p r es s u r e profil e f or an o pe rati ng condi ti on wi th a h ig h b ypa s s ai rflow. The i ni tial normal sho c k lo _ ttion is thu s well beh i nd th e inle t th roat.

The disturbanc e co n sid e red c o nsis te d of a r amp in th e inlet di scharge fl _ from its n o m i na l valu _ to ze r o i n $ m illiseconds and ba o .k in 3 additiotlal m i M_ econd s .

As th e f low drops, a pr es su re wa t_ or fro m begins to propagate ba ck tip th e dt lc t.

T h e nor m a l shock re m_ hS rela ti vely Sht ti o t m _ y un ti l the pr e ss u re f r ont _ oales c es wi th it. The n th e pl'es s ur e _ v e c _ rl _ s th e n oz _m al sh ock out of th e inlet a nd a t l un - start r e stllts.

No e xperime n tal thtta _e s emb li ng i nlet pres s ure profl |e s during an uns ta rt w ere available. However, p _ tk pressures we r e recorded a t fou r inle t stati ons duri ng an 35 6 u n start caus e d by a co mp ressor s tall. T h ese e xp e rimental da ta a p pear a s the c u rve al o ng the top of the p l ot and form an upper e n velope for th e pea k pressures. If allow; _n c e i s made for no t knowing how closely th e actual co m p re s s or stall resem - bled th e 3- m i ll i s ec on d i n let dis turb _ t,-.e, th_ agre em e n t look s g o od . i Also investigated were th e effects of a n ext e rnal di stu rbance on this sa me i n l e t.

The results ar e show n In figure XI-11. A gai n th e bott om t ra c e i s the initial s tati c pressure d is tribut i o n .

The exter na l disturb _ tnc e wa s g ene rated by reduc i ng th e cowl lip v _ I rJ city $ per- cent in 2.5 milliseconds. In this c a se, a pr essure front f o rm s at th e in l et throat and moves out o f the inlet. Th e ste e peni n g o t th e front as it moves towa r d th e coWl lip ca n be noted. Th e pre ssu re front will r e sul t in a shock wave s ta ndin g i n f ront of the inlet while a shock st.t U remains in th e inlet. Such severe and r apid-transients I .

pose spe cial require m ents for control s des ign .

InletRestart Contr o l

Unstart transien ts, like th e ones Just d e scribed, cause a drop in diffuser eXit _: " total p ressure recovery. Pressure recovery can drop transien t ly tO a valu e as low as 15 percent. Also, distortion i ncre a s e s. An un _ Rart m ay also re s ult in a buzZ.

condition where the shock oscillates unstably.

Since u ns tart ed inlet performance is poor , th e i nle t S ho u ld be r es ta rted quickl y .

Figure XI-1 2 illustrates th e f u nctions a r es tart c o ntr _ )l sh o ui d perfor m . Condi- tion 1 shows th e inl e t s ta rted a n d with the ce n t e rbody a t its nor m al positidn. The " by pa ss doors are almost closed to po sitio n th e normal Sho ck Just do w n s tr e am of the i inlet thro a t. This condi ti on re su lts in high total pr e ssure recovery a nd low di_ tol'- ti on at th e diffuser exit. If a dis tu r baace ca us e s th e inl e t to unstart, the inlet go e s _ t to c ondi ti on 2. Initially , th e sh ock may be i n a buzz co ndition . Whe n th e bypass doors are ope n ed w id e a s sh own in f igure XI-12 , the th r oa t b eco m es choked. T hen the inlet is sta b le , but unstarted, a nd th ere i S a strong norfl _ al o r bow shock ii l fr on t of the cowl lip. This shock is th e cause _f th e l _ rg e dr _p i n pres s ure r e - c overy that acco m pa ni e s un s tart. Since th e th roat is choked , a . _e cond ndrm al ....

sho ck al so exls ts dow n s tr e am of the inl e t thr oa t.

To r es ta rt inlets W i th i nte r n al con tr actioi _ , t h e r a tib O _ t h roat kr _a t o ckptui'e flow ar e a must b e increased. T his is ofted dc i n e by c o llaps i n _ or t r_ tnsJa[ing th e centerl:,ody. For this inlet , th e restart cycle begi n s W i th a foi-ward trans la tion o _ th e cent e rbody. A Condition J ust before restart wotild appear as in coil di tion $ . The c e nt e rbody i s forward of its normal po sitio n, a nti the t w o fior m al shocks st i ll exiSt.

r_ in g en er al , high e r p re ssur e r ec overy can b e achi e ved ds c ent e rbo d y position cha ng es t h rou g hout the restart cycle. Th e b ypass doors m u s t b e open enough t o p a ss the in - let airflo w n ot dema n ded by the e n gi n e. If th e d o o rs a re n ot op en en ou g h, the throat will un.c h ok e anttbuzz will o c cu r . As u s ual, if th e doors ar e o pen too much , p r e s - sure recov ery Will b e low er and di stort ion higher than nec essar y. For conditio n 3 , it was as s u m ed the de e r s could b e pa r t ly cl osed to giv e bett er per f or m anc e than at 2 . This would u s ually b e tru e_ if the en gi ne was not in a stalled or flameout con- dition.

Th e c en terbod y m ust b _ translated a lit tl_ far , her th an cond i t i on $ t o r e start th e inlet. Then c ent erbody travel is rev e r s ed and b r ought back to co n dition 4. The ex- tern al s hock has b een swal l ow ed , and as f or co n ditio n 1, there is 8 s in gl e , nor mal shock downstrea m of th e th roa t . The bypas s doors are shown to be sli g htly more ' closed to give bet t er p e rforma n ce th an at $ . However, if the doors ar e closed t oo much, the inlet will unstart agai n .

The res tart cycle is co m pleted by retracting the cen te rbody to its nor m al posi- t ion. A t the same ti m e, th e bypas s doors S hould b e position ed to i nc reaSe inlet performance as th e c en t e rbody is r e tracted.

These, then, are the 1.equir e m e n ts that must be c onsider ed when d es igni ng _ t reStart control. !

On e g oal for th e r es tart control S y s te m w as to m ake it c o mpletely au to mati c .

It w as al s o desir ed to m aintain clos ed loop c ontro l of the nor m al shock th rou g hout the r esta rt cycle. This would en able th e control system to comp ens ate f o r unex- pect ed disturbances w hich might occur during th e reStart cycle. One such disturb- an e e co ul d be an unexpec ted chang e in th e e n gi n e airflow con d itions. To acco m - pl is h th ese goalS , the r e star t co n trol system uSed feedback an d some switchi n g and s ch edul ing. _ , ' ] _ 'igur e XI-l$ iS a Sc h ematic of the resta rt control system. This _ y s te m u s es I t th e nor mal s hoc k f _ tb _ tck co ntro l sy s te m of f igure XI-4. T he light li nes in fig- :: ure Xl o l $ rep r esent th e c o mpo n en ts dtBcus s ed pre _ iou _ ly. Th e h eavy tines t n di- ,

I '

cate [h os e C ompo nen ts spe cific al ly requi red for r estart co n trol.

In this c as e th0 nor m al shock control loop used a s ingle _ta tic pressure aft of !

th e n o rm al s hock to t _ e r shock po s ition. After the inlet un _ta I _ , its p r es s u re I reco v6ry beco mes lowe r , _ tud th e level o t th e Se ns ed s tati _ pre s s ur e is s ubse _ lu _ rl tiy I ' r ed uced. Th us , the co mm and or ref 6t e _ ce hi put to th e c ontr ol has to b _ S ched u led.

Tw o s ched ul ers were us ed. On 6 supp l i ed low vaht es _ [h e c on _m and _ or th e con- dittons wltei'e th e i _ [et was uns tarted . The other s ched u ler s upplied hi _ h _ i" ralli e s o f th e co mm and a n d w as tised _ vher _ th e inlet was start ed , i t was fou n d desiz' _ b | e to vary t h ese co mm_ nd value s a s functions of cent e i , b ody l _ sitioh. Tlie _ efor e , center, i ., r

i

I body po s it i on W as fed ba c k to the schedulers a _ s hown , A r el ay c o nt a c t w as u s ed i t o se l e ct th e output o f o ne o f the s ch e d u l ers , Thi s re lay w as o pe ra t ed by th e output o f an u nstart detecti o n d e vi c e , Th e d e t e ctor operates t n the following w a y. Two p r e ss ures w _ re s ensed ah e ad o f the inlet t h roat. A ratio o f t hes e pre ss ur es , - _ s take n . When th e i nl e t u ns t a rted , the ratio i nc r e a s ed. When t h e ratio exc eed ed a p r edete rm i ned reference level , t h e r el a y wa s s w i tc h ed t o t h e unst a rted co n d i t ion .

Wl_e n t h e inlet re s t a rt e d , the ra ti o d e c r e as ed. When it d ro ppt _ be l ow t h e refer- e nc e l e vel , th e r e lay r e turn ed to th e start e d condition. : Another s e t of contacts o n th e relay s e l e ct e d on e ot two co mm ands f or th e !_ cent e t body ac tuato r. A co mm and _ or th e actu a tor to _o t o a_ , e xt e r _ d e d po s itio n w a s !

u se d wh e n th e inl e t was u ns t a rted. Th e command f or t he . q ctu a tor to r e tur _ t o the desi gn po s itio n was used after t h e in l et had re s tarted. Thu _ , i t w as possible t o exorci s e closed loop control of the inlet throughout the restart cycle b y s w i t c hing " . and schedu li ng co mman ds, i Figure XI-14 shows th e results obta in_ t w ith th i s r es tart control io r an inl e t connected to a l ong pipe. A transie n t i s shown co ns iStin g of an u ns tart followed by . a con t ro lled restart. The trac es are th e u n_ ta t t s i gnal, c en terb o dy position , by- pass door ar ea, a n d throat e x it static pressure. The co m rda n d v alu e of th r oa t exit static pressure is shown where di _ f e rent from the actu al value.

The uns ta rt is i ndicated by a drop in throat ex it s tat ic pre _ sure and de te cted by a ri se in the u nSta rt pressure ratio. The centerbody i s immedi ately co mm anded to slt _ v, or translate at m axi mum vel o ci t y, in the f orWard direction. The s ta tic . : t pres s ure co mm and to th e normal shoc k control iS s cheduled to . a l o w v al tl( _ . Thi s ca us es th e bypass doors to ope n to m axi m u m area to s uppre ss bu z z. A S th e s tatic pr es sure command is incre ased , th e doo rs come Jui c eact to u contr b l l iug static pressure to its co m mand value. When the i nlet resta rt is indicat ed by the d e tector ' th e centerbody sle w s aft. Also, th e sch ed uler c o mmands h igher values of s ta tic ' t pressure to the shock control loop. The inlet is ret u rn ed to o n -d es igu s tart ed conditions in approxi m ately 1.5 seconds. Thi s ti m e depend ed primarily on til e ..... .:,l k s lew in g sp e ed of the c en t e rbody serve. The inlet a ero dynamic s th( _ nm e lv es were quite faSt, as was the by pass door loop. No c e nt( _r body ov e rt r a _ el w as pe l. m itt ed , . and pressure recoveri es we re as hig h aS possib l e without enc o u ntering buzz. _ '

Inl e t-Engine Restart-Rellght C ontrol

This re start control a p pe ars to satis fy th e i n let re s t art r e quire m e n ts; flo w- e ver, th e r e a r e add i tional p ro blems when th e inl e t i s c o u pl ed to an engineinst e ad : T , i I \ J o f a l on g pip e . For exa m pl e , the un star t t r an s ient its e l f I S se ver e a nd ma y catm e co m pre s sor s tall and /o r c o m b us tor flameout. There I S concern , then w i th clear- ing _ . po ss ibl e s tall , r e ti gh tin g the co m bu s tor , and a void ing a po ss i b le turbin e . .

overt em perature problem i f s tall occu rs without flameout, ' To account f o r an engine in the s y s te m , a re li g ht f e atur e w a s i n co r p ora ted i n to the r es tart control system of figur e XI-l$. The igniter was energized au to - _ m ati cal ly upo n receivin g a S ignal from the U nS tart detection devic e , A t th e pr es - 'i_ ; ent ti m e n o att e m pt h as b e en m ade to m a n ipu la t e a n y oth e r' en g i n e v a riable s such ' , as th e thrott l e settin g or 0 xhaust n o zzl e area.

Figu r e XI-l{5 s h ows an exa m p le of t h e t ran si en t that occu rr ed when a r e start- ' i ' relig ht s equenc e was atte m pted tn th e tunne l. During this p a rticu l ar transie n t, an _ . _ en gi ne fla me out did occur. In th e top tr a c e , uns t a rt i s det e ct ed by the u ns tart _ S ign al exce ed ing its reference le v el. W h e n the inlet r estar ts , the u n start detec- tion signal drops below the refere n ce level. I n th e second trace, the actio n of the tra nsl atin g c en te _ )ody i s Shown. A fter the inlet unstar _ , th e c ent e rbod y travels " _ a t f u ll v e locity i n the forward direction. At th e point where th e inlet re s tartS, the centerbody reverses ' and retu r as at full v e lOcity to it s on-design co n dition.

The third trace shows the action o f th e con tr olle d b ypas s doorS. They s lo w ly ope n to make up for th e en gi n e airflo w lo s t by the slowin g down of t he fla med -out engin e . In the nex t trace, it i s n o t ed that there i s a ri se in the inl e t tltroa t exit pr ess ure as th e Shock m_v e s upst re a m . When uns tart occurs, there i _ a large decrease in th e pressure. As m e n tioned before, flameout did occur durtn g th i s tran s ient. Thi s fla m eout condit i on is indicated in the botto m tra c _ b y the de- cre as6 in turbin e exi t t em per a tur e .

The ti m e sc al e of figur e XI-I _ S ho Ws tha t the r _ tart tr _ tnSie l lt waS co m - pl ei ed in about 0. 5 S eco n d. It s hould be noted that th ere i s a bl , eak in the time __ along th e ho ri zontal axis. Re l i gh t of the co m bus to r did not occur fo r about i_ t § seconds. Acco m p a ny tu g th e r ei i g ht i _ a u ov e rshoot i n the turb ine exit tempe ra . _ lure. This ris e , {hou g h, is n orm al fo r this engine du rin g a conventio na l l i g h to ff. _ F i ll _ tl l_ , upoh relight, there was a m o me nta f _ in c re _ se in t he th roat exit p r _S - sure. As can be seen , thi s rise was not a ppreci abl e , and it was m uc h S maller titan / that which oc cu rred wh en the tule t i n i t ially u n Started. Thi s ri se i s ca u s ed b y a ' / p ressur e w av e p ro pag a ting up s trea m from the en gi n e. The inlet col _ tl,ol is pres- . . enti _ adequate to handle t his pressure ris e. . How e V e r , e rich ri ses co u ld be s U ffl- / _ _ i ei ent to reinitiat e a n i nl e t u nst a rt. I n the fu tu re, tht _ dytlamics of the re l lgh t ope r a t ion sho u ld be car e fully i n v e stig a ted t o pre _ , en t th is fro n t b e_ o _ i n g a p r ob- " |e ra area.

i Q Th e t r an sien t of ft _ tr e XI-IB is o n l y on a of m any th _ .L w e r e taken durin g this prog ra m . M an y of t h e r es u l t _ v a ried t o some ex t ent f _ ' om th a t s h o wn i n f i f _ _ ure XI _ 15, In s um ma_' y, th e f a ll o wing o b s erv R ti o n _ c o r lc e. vn i n g the res ult _ s hould b e n oted.

(1) Duri n_ a u in l et un s t a rt, a o mp r ess or e _ l I did no t _ lway s o _ cu r . When it di d oc c ur, ho wev er , it was tu itiat _ du _. n _ _;he un st art pol ' t iou of t he t ransi ent a nd n o t while restart wa _ b e i ng at t e mpted, ( 2 ) Su s t aine d s tall neve r occu rr ed _v ith o ut a n ac c ompanying co m b ns tor fla m e- o ut. T h i s , o f cour se, i s b ene fic _ t iv, th a t t t elimi n ates the problem of a turbine overt em pe ra ture co n dition durin _ _ . ¢s |l.

(3) I n all case s , th e r e s tart contr o l system w as s ucces s fu l i n restorin g the i n l et to a s tarted condi t ion, ov en w hen a co m pr es sor stall had o _ cur red . The staU alw ay s cleared it s elf by th e ti m e th e inlet was res taI_ l.

(4) Following the fl _m eout, the times r equir ed to reli _h t va ri ed co n sid e r - ably. This restart-re l ight system h as been evaluated o n two J- 8 §-1 3 e n gi n e s .

On the first J - BB, th e re l i g h t ti mes va r i ed fro m _ to 30 secondS. On the se cortd e n gi n e, relight ti m es v ari ed from 0.5 to 1.5 s _ c on d s . To da ta , it haS not b e e n poSsible t o account f or the diffe ren ce'in these reli g ht tir. _ es o n two su p po s edly ide n ti cal eng ines . Such wide variati on s in r eligh t times cann o t be tolerated for a s uper soni c transport. Th e refore, further work on this problem iS itldica t ed. ' _ '

D YN AMICS O FC O MBUS T O RS AN D C O MPRES S ORS

It has been sh o wnthat co mplet e co n trol systems can b e b _ fl t _ or t h _ inlet and engin e which can ha n dl e s mal l p e rtu r bations, inlet u ns tart, e ...... j : : _ s _ or _ ' stall, and combustor fl ame o u t. S till, to improve th e se systen _ and to _ u _ her " _ ,, understand the u n co n troll e d in let -en gi n e d y na m ic i n tera (_ tionS, n _ or e de t ailed de - l , : " scription _ are ne ed ed of dyna m ic behavior - particularly f o r co m bus to r dyna m ics l_ , :: and co m preSsor dyna m i cS . " '

. , . Co m b u sbr Oynamlc

I n the co m bustor, the c ha r a ctertsttc _ ar _ g ene r ally tUth( } hi g h ( [re O A ue _ tey range from 10 to 100 hertz. Since thes e dy i ia m ic S ca b affe ct en gi ne system per - forman ce , they are of primary concern to tile _ el co h tro| di _ stg _e r.

._ S61 t Fl_re XI -1 6 i l lu e t rate B th e b _ i s of th ea n A ly e le, T hi _m od e l co n sist s of pri ma ry _om buati o n _o n e a n d a secon d a ry m ixi n K zone . The fu_ l fl o w , t o geth e r with a port io n o f th e co m pr es s o r di s ch a r ge a ir , i e _s u me d t o b a c om p le t e ly V i a. ., p_ ri z ed a n d burned i n th e co mbus ti on _o ne. Th e product _ o f c om bu s ti o n a r e then m ixed with ad diti o n al air in th e m ixin g z o n e t o c ool t h e g a_ t o a cce pt a ble turbin e inlet cond i t to r _ .

T h e m ixing p r o cess i s d es c r i bed with c onv en tional g n _ dyn am ic _ t u_ t i on n .

Su c h ana ly e l_ r e s u l t _ in n lag fo r t h e mi xin g , _ ono g e m dy,a m ic s , Th e com bu s t i on pro c e ss is lo _s well un d ers tood. For thi s case , te c hniq u e s have b ee n borrowed i from tho s e de v el o ped fo r rocket o nlin e c om bu s tion dyna m ics. A ti m e dela y f o r f u el v a po ri z a tio n a nd _ second-orde r la g fo r the a c ti o n ! co m b usfi o _ pro c e ss ar e i_ assu m ed o r calculat ed o m pirica U y. This m odel for the co m bustor w as t hen i i , c o m pared with test r esults, i Du r in _ some o f the supersonic Wind tu n n 0 1 t0 St in g, th e en g i n e s pe e d cot _t rol i could be switched fro m the sta n dard en g il _ e fu _ l control to a h i gh response fuel control. It waS possib le using this control t o ob ta in th e frequehcy r es ponse o f a number o f p a ra met ers throughout the en gin e f or a fue l flow di s turbance The open c i rcles i n figur e XI-1 7 indicate t h e no r m ali ze d a m p l itude ratio and p h aS e of co m bus tor pressure to a fuel floWdisturban ce plotted as functions o f f r e qu e ncy. Als o shOw n ( as the solid li n e) are the r e s ul ts o f t h e an al ytic al co mbu stion m od e l. The correlation b e twe e n the a n alysis m i d the data i s quite " g ood , e v en out to 90 hertz. The inclusion of the m e thods used in rocket e n gine co m b usto r dyna m ics p ro perly accou nts f o r th e progres s i v ely i n cr e asi ng ph as e sh l _ t with freq u e n cy. This is an improvement over the m id-1950 c o m bu s tiori models, H

Compressor Dynamics

A more di, _ ftcult m atter to handle an al ytically is th _ i m por ta nt problem of , compressor dyna m ics. What must be do n_ is t o an al yz e , i n detail , the d yn a m i c cl _ racteristics of a mu ltistage co m pre _ so _ '.

_ rboJet com pre ssor d _ ma m ics. - When we started m odeling the inlet, m et hod s were c onsidered that wo u ld let us si m ulat e axia l co m p r e s sors and co m - pl e te ttir b oJet en gi ne systems. Just as with the inlet analysi s , w e had to work out ou r o W_ m eth od s to get the high freq u ency dynatfii cs .

A dy n amic m od e l was patterned afte r c _ nvention al st e ady-sta l e s tage Stacking techniques. Si m ply , each stage i s rep r es en te d by a pressure ratio m ap arid a :)62 ! _ t \ t em p er atur e r i se m ap, b o th p l o tt e d as fu nc t ions o f air fl o w , T o inc lud e g a _ d y - .: nam i c_ , a l um ped volu me w a ft add e d wtt h eac h s t ag e, I n t he volu m e _, momen tu m, c o n tinuity , a nd en e r gy bal ances w e re u_ od for t h e gm_ dyn am i c_ E a c h o f t h e ma p _ and ae soc i a t ed v olu mes were c o nnect v d to ge t an o ver M l e o m p r o _ e o r model, On ce a w o ' ,'kt ng s i m ul a ti o n existed _ a , mmpr e _ol_ map wa s gen er a ted t o O f_ - t ab li s h t h e _ to a dy-atate va li d i ty o f t he m odal. T h e a g r ee m e n t w a _ w it h in w h at !

ootdd b v expe c t e d f r o m t h e a_ ,a t la b l o d a ta a|_d a n a n a l o g si m t _la tio n , A l s o , t h e , co m p u te r s i m ul a ti o n e x h ibit e d a. a t a bilit y boundary t h at looke d vO_y m t mh l ike a i_ e o m p r o ssor s tall li n e. To investi gate f u rtho t _ an a c c u ra t e l y dete rm i n e d OX l_ e ri- i m e n t al com p r e s s o r a t e6 1 li ne wa t _ nod d ed.

Pri or t o t he d eve l o p men t o f t he com p ro e s o r mo d o l p an e xp e r i me nt al pt' og r am h a d boon run in a no n-l oy a l t es t sta nd wh0ro t he J - 8 8' s c o m p resso r p o t 4 orm tt nc0 an d s ta ll l i ne wo r e d e t ermined . The exp erime nt a l data wore obta in e t t fo r a J - 8 _ _ n # no h aving a r e duced turbi ne n o zz le a re a; how Ov e r , t h os e d a ttt c ould b e tt s ed i : , f or c h ecking t he anal_ tic al m o d el.

Fig u r e XI- | 8 ind i cat es t he r esu l ts of the te s t pr o gra m pl o tt ed in th e form o f a co m pressor pe r for m a n c e ma p. Exper im ental operat i ng poittts wer e obta i n e d a l ong consta n t corrected speed lines. T h o S e d a ta provided a method of chec ld n g _ i, the s i m ulatio n . T h e s i m u la tio n w as sc h eduled to operate with t he same s c hed, aiss !, used i n th e test pro g ra mp a nd t h e d _ ta o b tai ne d w ere s im i la r t u th e tO s t d ata. T h e Ii s olid l in es in f igu re XI-18 are co n sta n t speed lines o b tained _ ro m t im si m ul _U on; I_ the s o lid c ircles i n dicate the Si m ulati on sta b ili t y bou n dary. Co m p _ ritt g t h e s olid ! [ c ir c le s with th e test data of th e op e i _ sc[u ares shows th at the ag l 'ee m e n t w a S good.

Ther0 f or e , we f elt that w e had an an al y tt _al m e thod t h at c o td d indicate t i_ e c om - p re s sov stall li ne provided s ufflci _ n t gta g e data w ere availa b le.

, : O n e o f the o ri ginal goals of t h e si m ulation progra m was to m odel syste m g as dyna mics to 50 h ertz or b etter. Test data were provid ed to verify the m o d el's t f requency respo n Se accuracy. Figure XI - 19 sho w s th e fr equenc y l ' es po i t S e o_ th _ sixth s tag e p r e S sure wh e n fuel flo w was used a s t he dis tur bttt i ce s tg tt al . Th e solid line in di ca t e s th e re sul ts o b ta ined by t h e s imula tion. The ag_e O m_ n t b # ' _e eft d i cta an d analysis is quite good. Althou gh al l o f the d a _ 0 b t _n ed duriltg th e en gine testin g d id not correl a te this we l l, th e tt g i'e em ent w aS, i n gor ie ra l , _ ood.

T hus, w e seem to have a suita b le a n al ysi s m et h od _ o r e r i e-d i m ensio nal ef- f ects o n tur b o je t en gi n e s e ve n out to quit e hi g h f reqtle n cies.

.... Turbofan compressor dyi_amics. - Not al l prop ul sion s y d te n is u se tu rbojets.

Many use tu rbof an s as shown i n figttre X I- 20. Co m p a re d to a tttrboJet , _ turbof a n has t h e added co m plexity o f a fan eo m p r essor _ a n d th e d ivisiofl Ofthe air fl o _ v bet _ eett the fa _ bypass duct and the en gi ne core co m preSsor s . Tilts additio n a l co m pl e xtO / / / 8_ 3 tl / po _ e _ f u rt h e r a naly t ical diffl cul ti e_ . Th e flow i n th e fan byp a e s du c t ca n ex h i b it r eH o n an ces which a r e t r_ m smi tted into t he core com pr esso rs thr oug h th e d i ec h arg e c onditt _ n _ they [ m p _ e e o n th e fan. An al yti c ally , n t lea_ t, th e d uc t flow l _ e e_ n o pro ble m _| t he Sam e W a v e eq t_ R tt _n me thod _ ar e u sed h e r e as were u se d fo r t h e in l e t sut mon ic dif f u s er flew. T h e cou pl i ng wit h t he f a n, ho wev e r , is anot h e r q u e A - t i on.

The prt_r try mmtytic a l p r o b le m po e od by the a ddition of t h e fa n i e the dtvt e lon o f the f _ flow between th e f a n byp _s duct _ d the core c o m pr e_ nr _ . T he fl o w d i vi s io n i s comp l ic a t ed by the d iffe re n t di s ch arg e pre ss ures tmp os_ by the duct and core re s i o n s . One appr o a ch to t h e pro bl e m is to use a n a vera g e p re_Aur o a e the f an di s cha rg e pressure. The av e rage p r es sure i s used in a mom e n tum b ai_ an c e to d e ter m ine t h e tot a l fa n flo W . The total flow i s th _n divided betw ee n the duct and core a S a nonlinear function o f the duct to core p re ss ure r n tt o _ n d 0 or- reeled spe _ i.

Once past th e probl e m of the flow S plit , th e core conlpre s scrs can be modeled with th e techniques a pplied to the J-86 co m p r e s sor , and the duct c an be m od e l ed as w as th e inlet s ub s o n ic di ff user.

Figu r e ::I- 21 , which presents an Slytically predicted f requ e nc y r espon s es f or the TF-30 en gine , shows some of th e problems associat ed with modeling a tu r bo. Ii fan en gine. No rm lfliz ed a m p l itude ratios for various syst em pressures to a co m - pres s or f ace dis tu rbance ar e plott ed agai n st diS tu rb an c e frequ en cy.

L-tpa r ticular, the z' e so nan t c ha ract e r of th e _ an duct (dashed c u rve) can be i : _ s e e n contr as ted to the al m os t co ns tant a mpl itude rati o of the f a n discha rg e into i_ the core co m p r e ss or. In additiotl t o t h e f an , the TF- $ 0 en gin e has both low a n d _ . _ high presS U r e co m pr e ss o rs in th e e ngine core. l _ r _ qtieflcy r e spons e data torth e If " dischar ge of thes e co m preS S ors a _ e also S hown. Th e l o w compressor charac- teri s tic is quite si m ilar to the fa n core cha r acter is tic. Th e high compreSso r , t which dischar g es in to th t) co m bustor, h owev e r, e x h ibits a char a cteristic si m ilar t o the d a t a lot, the J-85 engi n e. It wo _ d appear fro m th es e an _ ytical r es u lts that th e e ffe c ts of fan d u ct r esonanc _ ar e n egligi bl e. Un / ortuna t e l y, ottr e _ p _ ri- m ental results are som e wh at different fronl ou r analytical results, To investi- gat e tu rbof an dyna m ic s _ xpe _me ntally, uni / or m sintu _ oid ai p ressqre was im- po sed a t the _ mgt n e inlet and the f re _ ltiency response charact e risti cs wer _ d e ter - m i n ed at va r iou s s tation s th r ou gh out the e ngine. Th is p r ogra _ waS do _ e with the airJ _ t _ yste m , _ l e scribed in paper X.

' _ he test r esults (s ee fi g . XI-29_) do e;tliibit s hapes s imilar to t h e aila l ytical res ul ts up to f _ -eqUencies of 20 hertz. Abo _ t ha t f _ e _ lu e ncy , the d u ct p _es stire re s onanc es appear to b e l_ iteracti ng with t ile c o re pr ess ured. This eff e c t i S o _ - ___,__ '; .. _i_ • _'_,,_ _ __ : j . ,_.': ', ,- • .

Viou s ly not indicated by the analytical m odel. It would seem that a high frequency c haracteristic is missin g , poss ibl y in the fan simulation procedure.

In an att em pt to match th es e data, the shape o f the nonlin e ar m ap u s ed to e s - tablish th e flo w split w as ch ang ed. However, to g et re a s ona bl e agr e e m en t be- tw e en an aly t ical an d exp e ri m ental resu l ts, the map ha d to be changed beyond pJ _ y s - ical real i ty. Thus, work m u st s t ill be d o ne o n th e pr o blem.

To co _ tclude, there are reso n ance _ that occur in turbofan e ngines which as yet c an not be p red ict ed anal _' tic al ly. When-m od eling the tu rbo f an_ the b asic al ly one-di m ension al technique a cor, q idered for axi al co m pressor s do not giv e good agree m ent with ex peri m ent al data. Our present on e-di m en E lio nal m e th ods-allow _ or ti m e varying spatially uniform, eor _ :pressor face c o nditions. We have s ta rt ed a two-di men sional mod el for an axial co mp ressor to try to include the effec ts of s patial di sto rtio n . Thi s effort, howeve r , is _ U U in progr esS .

S ummary

" The oth e e co mp on ents in th e p ro p ulsion s y s te m , Su c h as the tu l-bfiie , a _ e r - burner, and exhaus t nozzle m ay hav e importa n t dynamic al eff e cts. For _ t turbo- Jet engin e , where the tu rbin e nozz l e is u s ually choked , such dyfta m ic s do not couple back in to th e rt _ St of th e system. FOr other configu ra tions , such as a tu x * - br d au engine or an en gine us ed as a gas generator, the dynal _ ic coupliftg m ay be i m portan t . The s e problems are not discu s sed in this pape r.

", DEVELOPMENT OFSENSORS , SERVOACTUATORS AN D TESTING TECHNIQUES

l The e arlier portion o _ the paper ha s e m phasized re s ults of som e prop u lsion 11 dyna m ic s and c ontrols i n v es tigation s . Th e ki t ids of expe r i m ental restdt s ob _ tined imply the us e of sen s ors an d s ervo _ with ra th e r uticom m ofl capabilities. Thi d portio n of th e pa per is conc e rned with expl an a t ions of some o f th es e r esearcli tool s.

Shock Position Sensor , , I Con si de r able work h as be en done i n t h e ar ea o f di re c tly s e _ in g an t _ [et t ll normal s hoc k po s ition. Potenti al ly, gu e h a dev ic e c an fa c i litate saf e o pe r s it|oti

S6s

o f a n i nl et while reduc i ng th e margin o f s upe r critical operatio n .

I n mos t ca ses , our primary i ndi ca tor of shock po s ition h as been a s tatic p r e s - s ure. This pre s sure (show n in fig. XI- _ 3) is l o c a ted a / t of th e normal shock operat- ing r ange and is referred to a s th e throat ex i t sta tic pre s s u re. This s i g nal gav e a n adequate i n dication of shock positio n for the con di t io n at which the i nlet wa s o pera te d.

However , a more direct me a sur e of shock posi tio n i s d e sirable be ca u se shock pos i - ti on m eas ur ed w it h respect to the inl e t th r oat i s dire - .tly r e la te d to inlet p e r f orm - ance a n d stability.

A common way to de te r min e shoc k po si t i o n i s to observe th e static pre s sure pro- file i n th e vici ni ty of t h e shock. T he sho ck locat i on i s identified by a Jump in stati c pressure. Ideally , this occurs as Sh ow n i n figure XI- 2 3. The flow at the inlet throat has a Mach number g reater tha n o n e. Dow n strea m of th e th roat , th e inle t area i ncr e as e s. In supersonic floW , Mach n u m l;er increases as area increases , and thus s ta t i c pressure decreases ahead of t h e shoc k . At t h e nor ma l s h o ck th ere i s a di sconti n uous Jump in p re ssure. Also, th e flow Mach number Jt_m ps fro m super- so ni c ah e ad to sub so_ L £ aft of th e sho ck . Since the area i s still incz'ea Si ng aft of th e shock , Mach nu m b e r decreases and sta ti c pressure i ncreases.

In a r ea l inlet , the pre s su r e profile can be m ea su red wi th a series o f closely spaced s tati c ta ps. F i gure XI-24 shows two typical s ta tic pressure profiles that v ,e re mea s ured simultaneously in a r ea l inlet. These pr ofi le s were measured wi th the inlet at an angle of attack. The upper prof i le was measured by ta ps i n the up pe r half of th e inlet. The up pe r pr of ile i n di ca te s th e sho ck to be betw e en ta ps 9 and 6 be ca use of th e ste e p press u re rise betw e en tho se t a ps. The lower pro f il e _ as m ea s u red by ta ps i n the lower hrd _of the inl e t. The p ro f ile indicat e s th e shock to be be tw een ta ps 3' and 4 t . T h is i s about two taps forward of th e po si t ion indi c ated ' by the up pe r ta ps. Thus , th e sh ock does not lie in a plane nor mal to th e inlet's longitudinal axis. Nonp la nar con di t i o n s , altho u gh not as severe , were di scov e red even wheu th_ inlet was at a 0° angle o f at tack . Also , the pres su r e pr of ile a head of _ ' the sh ock does not exhibit a conti n uous decrease in pres su re as does th e ideal profile - fo r example , th e rise be tw ee n 2 a n d 3 on th e up pe r , a nd the rise be tw een _ , 1 t an d 2 t o n the lower. Such Irregu la rities were also found in pr of iles m_ sur ed at 0° a n gle of at tack .

These problems a n d o th ers make it difficult to apply si m ple logi c s c he m es tb th e pr of ile s to dete rm ine sh ock position. Despite th ese probl em s , some progress ha s been m ade here in sh ock sensi ng schemes. Two different logic Schemes have _ been tried to date. The simplest logic determined th e sho ck to be between the first tap havi ng a higher pressure than tap 1 a nd its adjacent upstream ta p. Foi" ex am ple , in the upper profile th e shock woul d be be tw een 5 a nd 6 (6 bei ng th e first tap wi th a higher pres su re tha n 1 , and 6 its adjacent upstre a m ta p). On the lower pro f ile the s h o ck woul d b e dete rmin ed to b _ b e tw een 1 ' a nd 2_. Alth m lgh it i s no t _ dmw n I n figure XI -9 4 , th_ lower tap s al w ay s ind icat e d th e sh oc k to be betwe e n t t a n d 2_ .

This is because the pr ess ur es at tap s 2v to 8 t w e re a lways g r v _ t e r tl _ n that at I t r e ga rdle ss of s h o ck po s tt _ mt. For th e 0° a v _l e o f a tt a c k co n diti on at whic h t _ t s were e o nduc te d _ thi s proble m did not exi s t. I Th e s che m e just d e scrib e d , as well as the othe r one tr i e d , were both t ro pi c - i ro osted in tw o ways: one u sing el e ctrOnic pr e ss ur e transducers a nd logic e leme nts , i .

i ' an d o n e usir _ fluori c el e m ents. W e have b e en most successfu l w i th the e le c tronic s e nsors , a lthough the f l u e t ic sen s o r s did give pro m i s i ng re s ults. T h e electroni c i . _, s e nsors ar e b e ing used in th e wind tunnel control roo m a s sho ¢ lt po s ition indica tor s, _ a nd they have be en u s ed in some of th e normal shock control i n ve s ti g ations. These _ I se n s o rs ha ve f ollowed shock position very well f or frequ e nci e s up to 6S hertz with i a m a x i m u m pha se lag o f 20°. However , in some ca se s th e SensOr had to be adj us ted : _ for the con d itions under which a test was conducted. And it has been s hown tlmt _ operating the inlet a t a n _ tngle o f a ttack c a uses proble m s. Th 6 s e _ hor t comings a re o b viou s ly not desirable f or a f light appli ca tion. '_ Another scheme w hich a ppear s could work over a wide range of i nl et c on di tions i_ is being considered. This s ch em e would use the sam e lo g ic a s d esc r ibed befor e , !i except tha t i n stead of 1 or 1' be ing u se d a s the reference pre s_ r e th e r ef e r ence !!I pressure would be based on a to ta l pressure Just ahead of the static tap region, i_ Tiffs r ef erence is indi ca ted on the pr of iles in f igu re XI-24. Ba s ed on th e uppe r pro- i_ file, the shock would be indicated between 5 and 6 _ whi c h i s correct. Based on th e _!_ l ow e r prof il e , the sh ock wo _ld be indi ca ted betwe en 5' a _ d 6 t , w hi ch is two t _ . ps _ aft of the actu a l location. Al th ough th e r e t _ so m e e rror, it may riot be una c c e p ta b le . !_ kk _ I n ad di ti on , th is sche m e can be im ple m e nte d very simply by usii l g differ en tial p res - i i . , , s _ re switc h es. The static tap pressttres would be co m pared to the r efere n ce pre s - 14 sure dlrectiy _ rather than using expensive transducers and electronic co m para to rs, t_' ,' The outputs of th e switches could be s um med to giv e an ele c tro ni c sig na l pro porti o na l Io to sho ck position.

Cer ta inly , f u rther i n vestiga ti on i B call e d for in this _ re a .

Fast Response S e r ves

Another area of development of co n trols hard w are ha s been that of fa s t r e Spo l tSe actu a tors. These hav e per m itted manil l lla ti on o _ th e experim e ntal m odels to fre- quencies i n the 100-hertz range. A ben ef it Itas b ee n that o f g e tti ng better _ xp e l'io ment al verif i cat i on o f the analytical m ode l s. It has a lso pe r m itt d d contr0 |_ i n_ e s ti- 36_ ' ga tio ns t ha t wer e li m ited not by the actuation d e v ic e s , b u t by t he b as i c dyr _ m lcs o f _ h e p ro p u l sio n s y s te ms th emsel v e s.

Applic a tion s fo r w hi ch fas t re s po nse a ct ua tio n eq u ip m e n t has boo n d e sign e d and developed a r e a s follows (al s o in cluded a r e th e a r eas in whi c h th ese h_ ve been us ed): ( 1 ) Fue l t h rottling valve - U s ed f or c omp r ess or d y nami c s e valuat ions a s w ell a s a control e l emen t i n pr o pul Oion c ontrols r e search ( 2 ) .Mr distort io n va lv e - Us ed to d e termin e c o mpr e ssor stall m argi n s u n d e r v a r i o u s d ynamic di s t o rt i o n patt e r ns ( 8 ) Inl e t bypas s va l ve - Use d f or dete r m i n ing i nl e t d yn a m ic s as w e ll a s a c o n- trol e l eme nt for h ig h per f o r mance in l et sh oc k control systems All of the fa st respo n se a ct _a tio n ha s been acco m p li shed with t he servosyste m show n i n figur e XI-25. B a sic a lly thi s syst em uses a high performa n c e two-stag e e l e ctrohydr a u li c servovalv e driv i ng a pisto n -i n -cylinder ac tua tor to which th e lo _ t d is atta c h e d. The servov at ve i s drive n by a specially designed servoa m p li f ter which, through the position feedback device, pro vi des t he closed loop operat i on of the ac tuation sys tem . To a i d in th e des ig n o i th e systems for th e various a ppl _ cations , a complete n o n li ne a r si m ulatio n of the syste m of figure XI-2 B was implemented.

One o f th e _ irs t syste ms developed w a s th e f ue l th rottling valve. F i gure XI-26 shows its res pon se fol" Various levels of desired output motio n . The curves of fig- ure XI-26 s h ow tha t th e r es ponse exceeds 100 hertz for the smaller a m plitude inputs.

As th e in pu t am plitude inerea S e s_ the sy s t em does not re s pond quite as w e ll. , During th e vari o us design programs a n d with th e aid o f th e non li near si mula - tion, it w as foun d that th e dyna m ic p _rf o r rnan c e _ or th es e p E tr U cular electro h y dr au li c serves c ould encounter re spo nse limitations in un ex p e cted re g ions. An a nalysis of the s e lim itations re s ul t ed i n a new d es ign c r it b ria for maximizing the range of fast re S ponSe capability. The detail s of this criteria are well docu m ented in seve r al of th e publi ca tions listed in the bibliography (p. 375).

The actuation Sy s te m f o r the air di StOrtion Valve was designed using the new t more optimu m cri _ r ia . The nor m alized f req u ency res po ns e curve of figur e XI-27 shows its exper im e n tal pe rfor m ance. As can be seen, it s respo n se i s _ lat to beyond 150 hertz. Also shown i s one o f th e fuel v al , _ e res pon ses of fi gu r e XI-26. It can be seen that th e new more opti m Lt t n design cr i te ria pro vi d e s some s i gnifi ca nt i m prove- m ent i n res po nse. MoreOve _ th is i m pf'ovenient is obt a ined even th ough th e distor- tion valve moves an o u tpt l t wl _ ose we ig ht is thr e e times tl _ t of th e fuel va l .ve serve.

s_ s

O th e r Techni q ues

In addition to thes e _ p ecin l types of hardwar e, sp e cial t e chniqu es hav e been i i devel o ped f o r e xp e rl meninUy obt a inin g dyn am ic data i n o u r l a r ge f = cflities. These t e c hn iq _e_ a re exp inin ed in sev e r al publi Ca ti ons li ste d in t he biblio g r _ phy (p, 3 _ 5). i_ , i ' !

FUTURE T I ENDS INAIRBREATHING PR O PULSI O N C O NTR O L

Future trends in a n a irbre a thin g propttlsion co n trol w ill follow f u tur e system con f igurations a n d requi r e me nts. Th e lo w cost en gin e with it s l ow cost and simple control was discussed in paper VII. Control probl ems in th e use o t cryogenic f ue l s are discussed in paper XII. The pro pu lsion control pt _ oble m s for VTOL- S TOL ai r - pla n es a r e . _f v ital importance _ or these systems. Stu d ies in th i s ar ea have just be gu n.

Systems are beco m in g more complex. The capabilities o _ hydro m echa n ic _ d control, always l im tted _ md y be inadequate f or the future. We ar _ also seeing t he develor p me n t at fast, l arge capacity, f light wor t hy, di g ital co m p u te rs .

So m e o f th e advantages a n d ne w proble ms that wil l r e sult f ro m t he a p plication of d i git a l control to proptflsio n systems a r e now conside r ed.

i

, D i g i ta l C ontro l ofCom p l e x S yst e ms

= Th e advanta g es at co m p u ter control W ill be felt throughoutthe a ir craft _ ont r ol _ system. As regards the o v erall all'cra f t fli g ht objectives, a more compl e x control will be able to take in t o account i n forma U on concerning air tt,affic contro|, and t weather as well a s flight range and s peed. The f light co m put er can _ he n d eter mine an optimum a rr an ge mellt ot the aircraft control co mm and s . P _ trt o f this infor m a - tion will b e concerned wi th th e operatin g condition o f t he pro pu l s io n sy _ t e nt. In thi s wa y then , the proptt l sion syst em and aircra f t will b e totall y in t e grated throdg h t he "" use of digital co m pu[er cOttt r ol.

A possib l e configu r atio n f or the p roP _ tlsion control syste m is _ hown in fig - ure XI - 28. The prop u lsion control co m put e r is required to oper _ t _ th e e n gin e I - ' satis f actorily a n d safely in I'e s ponse [o the d e mands a t t he o _ er _ tJl fli g h t object iv e s , of the aircraft. Th e _ t'op U l s ion control m= ty have it s owit s e p ara t e digl f £1 cont _J tt t _ t , or it m ay be pa rt of a m ore co m pl e x ce _ ttral compu ter-. _ [s is a m att er no t dealt with in thi s pa pe r.

With the c o m puter, t h e propulsion syste m control c an be s eparat e d into m o re di s tinct levels of o peration. E ach level will be a ss igne d it s area of responsibility or c o ntrol supervisio n . In general, th es e areas of super vi sion can be s epa ra ted a s s ho wn in figure XI-28. Of most overall concern will be the op t imu m s che d ul in g of pr o pul s ion system in pu t s ac c o r din g to some de s ir e d perfo rm ance criteria. Pro- pul Si on s ys t e m perfo rm ance is measured in terms of such items as thrust, s pe c i fic fuel consu m pt i on, and rel ia bili ty . The air c raf t a nd external dist u rb a nces will deter min e tt ad e of fs be tw e en thes e mea s ures of p e rfor m ance.

A second and more direct role of the propulsi o n control co m puter p erta ins t o system-conSt ra ints such a s stall , unstart, turbine overte m pe _ ahtre, and engine _ ov e r S peed. Sequence s for restor ing nor ma l operation follow in g the vio l ation of a .. con st ra in t can b e g reatly improved over what exi st s tod _ y. Impro vem ent at this level again can come by the controller tak in g into a ccount both.aircr af t a nd e xt e rnal • environ m ent information.

There is, the n, the most dtrect level that the co mput er control will be reqttired to npe ra t e . This is co n c el med wi th the normal dynamic or trallsient conditions of the propul s ion system pro c ess. Here the controller must be capable of m aintain in g th e propulsio n system in a s ta ble an d responsi ve m ode of operation. Aircraft p r o - pul s ion s y s t em s are hi ghly nonlinea r de vi ces in which the s t eady - State an d dyna m i c c hara cteristics ch_ e drastically ov e r the ope _tin g range, l _ resent controls crudely s ched ul e va r ious con t rol elements as a function of o pe rating p oin t to comlter- act th es e anticipat ed change S . Th e co mputa tional capabilities of the di gi h tl c o mpute r can greatly improve this S ih 2 a _ on and provid _ for a t ig hter control ttud e r all co n di- tions.

To im plement some _ th e concepts just presented, the J-85 e ngin e control is L currently being r e plac e d with a di gital computer. Th e Sh ocl _ position c _ an i n l et has alread y been c on trolled by a digital c o m pu te r . For this te st , the hllet Was t e r- t min ut ed b y a choked 0f'ffice an d not by an eng ine . Ulti m a t e ly w e inte n d to cOntr Ol an inl e to e ngine p r opulsion Systenl.

For thes e test s , the di gital co mpu ter is a p r oc es s co n t1'o l typ e of S on ,pu rer.

The co mpute r arid i ts periphe r al _ we r e sel eoted so _ u tt w e _ control a real t ime process. Sp _ ial a ttention v _ as giv en to ent _ tnc ing the spewe d with whi t :h th e c b t _ - put _ r can execut e a co nt rol concept or detect arid ace o uflt for any abnor i hal R y. The control s o p hi st ica t iofi possible by havin _ a digi tal c0 m ptit _ l _ in th _ sy s te m d O e S not c om e wi thout some as so ciate d proble m s. Some of th ese pro b le ms a r e Caused b y _ e i liardwa _ e , some b _ s of t , e , and some by b oth .

These pro b lefns are as follo w s: _ , (1) Compu ta tion s pe ed. As th e process i _ sampled a n d eomm _ mds are made at 3?O

+ 1

: t \ ' : + , +t discrete i n terv al s o f time, o n e n o longer has a co n tinuous control device. Obviou sl y, in adequate co m puter speed Will cause problems.

. .

+ ( 2 ) Qua n ttzatlon error. The input dat a a re q u antized by the number of s _ t m pl e ' levels being used for the range o f each Variable. Th i s will set the ultimate accuracy of the system.

(3) Sa m p l e rate re q u ireme nt s . For rea l tim e op e rat i on, c ar e mu s t be used i n the selection of the Sample rate of th e analog to digital converter. T o b e c onsi s te n t with s am pled data theory, t h e sample rate mu st be twice that of the highest freque n cy contained in the input data. This problem i s illustrate d in f igur e XI-29 where 1 sec - ond of random inpu t data is presented. It has been f r equency li m ited to i0 hertz.

For this case, 2 0 sa m ples per second are the m ini m u m number of point s from whi ch the O rigi _ l data m ay be th eoretically recon s tructed from th e sa m ple poin ts. For co n venie nc e, the same da ta a r e shown with three different sam ple rateS. At 10 sa m ples per Second the hig h frequency content in the da ta i S obvio us ly missed. A + ., less oscill a tory curve can also fi t these sample pohlts. This ambi gu ity will cau se err o rs, becaus e the Sa m ple po in ts resulting from the high freque n cies in the data will b e interp x_ e t ed as a l ow er freque n cy oscillati on which iS no t in the data . This ), effect is called freque n cy folding , ai _ d i t w ill occur W henthe sam pled da ta th eory rule is v i ohtted. Obviously, 40 samples per second W ill g i ve pre t ty good results and cause no problems.

(4) Round - off errors. In addition to the previous in p ut da ta errorS, there a re , I other more well known digital computer p roble m s, such as round-off errors, with- i n the c omput er's co m p u tati on pr o cesses, i (5) Nu m er i cal stab il ity. F in ite word si z e also limi t s the setting accuracy of the c o eff i c i e n t s of the control law or control al g or i th m (as it i s m ore p _ 'operly • called). This problem can have an effect on numerical s ta bility. Thu s , i t is po s- : ' sible for the co n trol algorithm to be s ta ble in fl l e o ry , yet be nu m erically un sta ble ?

when inserted into the experi m e n tal syst em . + With ca reful controller design, thes e proble ms can b e min imized. Th is has : i_ + ' been demonstrated by o ur digi ta l control pi + 0 g ra m with th e inl e t in th e l hlperso n ic i I w in d tunnel. A computer contr o l la w app r o pr iat e f or controll in g htl e t shock posi - I + "" tion has been d e vised. Figure XI - $0 is a b l ock dia gr a m o f th e . System.

p • he error betwee n th e co mman ded shock pos i tion and the actu al sens e d s hock position was periodically sampl e d and _ ed in to th e digita l c ofiti.oHe P . , , The co m puter i m pl em ents the contro l law a x i d c al dd la tes th e co mma nd ne C es - sary to correct for th e e r r'or in the shock position. This co mm afl d is outp u t to the bypass doors. I t is held cons ta nt until n e w inforniatio n is sa t e ;pied, taken iilto t h e co m puter, and operated on. Only then is the by pa ss actt _ t i oh system update d With a new command.

-+ 1 I, Figure XI - $1 shows the experi m ental results obtain e d with a n tater u nder digital and under a nalo g c o m puter co n trol. T h e freque n cy r esponse of . no rmal s ho c k posi - i tion for a down s trea m airf l ow disturbance i s s hown. T h e solid c u l , v e shows the r e - • sp ons e under analo g c om puter control, Th is is a c o ntinuous c o ntr ol s im ilar t o th e ; ones th at produ c ed the control r es ult s sh o wn b efore, Th _ a nalo g r es pol is e as i n di- cated be f ore , in cr e ase s with f requencT at the low f re q uencies because o f the int eg r al action of the controll e r. The da s h e d curv e shows the re s pon s e u _d_ r digital control with a sa m pling rate o f I000 time s per se con d , The di g ital co n trol response i s v e ry ( ' . lose to that of th e a n alo g control. At th e high distUrbance frequencies it coincides w i th the solid a nalog curve, Thu s , in spite o f th e difficulties Just mentioned, it per- formed very mu ch like the ana log co n trol. The broken c u rve is f or t h e di g ital co n - tr ol wi th a slower Sa m plin g rate o f 100 times pe r second. This curv e al s o c orre- sponds closely with th e other two curv e s over most of the frequency range. At dis- turbance frequencies above 5 0 her t z, llow e v e r, the solid curve is not int ended to also !

' represent the slower digi ta l. If in put frequencies above 5 0 hertz ha d been us e d , the sa m pling th eore m , as expla in ed previously, would have been violated, and significant eXcurs i ons in shock pos i tio n above th e solid line wo u ld have resulted. Thus, i t is noted th at a fast updated digital control behaved much l ik e a n analog cont r ol. A slow update d digital did about as well at low f r equenc i es, but it could prod u ce unde si rable effe c ts at f requencies above ha lf its Samp lin g rate.

i • A p plications o f Mod e rn Control Theory , , i . With such dig ital co m pute r c ontrol capability and our u n derstand in g o f syst e m • .*. dynamic b ehavior, we m ight hope to make significant i m pr o veme n t s in system op e r- ation. This pl'es e nt s th e _ hal l enge of f ul ly exploitb _ g th e se capabilities. As a re - I t stilt we ha ve been f orc ed to tak e a ca r _ uI look at _ ontr ol t heory, e l 'the W ay t o de - I sign cont r ol system S . When d e sign in g control systems , we are really in t e l' e sted in how they p e rfo rm dyna mi cally Or as a function oJ _ time. Mo d er n high s p _e d com- puters pe rm R the s olttt ion of cont r ol d e sign proble ms di r e c tly in th e t im _ do m a in .

In th e past, co n tro ll ers were d es igned so vn ewhat artificially thr ough th e ris e o f fre- quency do m ain techni que s. The capabil i t] ! oi s olvin g control p r obl em s dir e ctly n O w enables th e d e signer to op t i n liz e t _e cont . v ol sys t e m t o certa _ r e alistic m eastir _ s o f p e r t or m a n ce.

We have been work in g a t apply in g these new techniques to propulsion sy s tem ' control p r oble m s. OUr fi rs t application was designin g a control to mini m ize tl _ e expected f rt _ luency of unsta r ts of a su person i c inle t to a randu m airflow di s turbance.

!

Th e si gnific a nc e of thi s app l i c a ti on I s il lus tr a t e d in fi gnre XI- _2. Wh e n the inlet control w as di s cu ss ed befor e , i t w _s s p ec ifi e d that the nor m al sh o c k, when in flu.

eneed by di stu rb _ m ces, should be held as c lo s e a s po ssi bl e t o s o m e de si r e d po s it i on d o Wn s t r ea m of t h e throat, Th e di s tanc e of thi s de s ired S t ea dy- s tate sh o ck se tt in_ i h as be a n de s ign a ted as a . Un _ t a rt o c _ r _ Whenth e dis turb ance is s_ ffic i en t to I ca u s e th e s hock t o m ov e forw a rd gr e a t er t _n thi s di s tan ce a. The e x p e c t e d fre-

i

que n ey of inl et u _ S t _ rt s t o a r and ont di s tur banc e with k nown s tati _ ti oa l properties, t t h en , is a f unc t io n o f t h e ma gnitude of the se tt ing _ , It iS also a function of the mean square v a l u e of th e s hoc k movement around th e s t ead y- s tate setting. Th e con- trol desi @ ed t o m ini n dze un s tart s will a ccom pl i s h this by ma n i pulatin g _ he bypa ss , doors. Thus, the exte n t to whi c h th e f r e que n cy of unsta r ts can be minimized wil l b e dependent o l_ th e capacit y of the byp a ss s y stem p rovided.

T h e results s hown in ft [ _ ur e XI- $$ were o b ta ined w i th an an alyti cal m ode l of a m ixed co m pressio n inlet. Th e lower curve, whi ch represen _ th e s m alle s t o f the three s t eady- s tate shock settin gs _ , sh ow s how the time between un sta rts in cre a ses 'L ! as more cap a bilit y i s _r ovid e d in the control bypass s ys t em . Lik e wise , a S a in- i_ ' , , c re ase s for a fixed a m ount of con t rol e ffo rt , th e m ean tinie b etween u nsta r t s g e _s lon g er.

The lar ge r values of a m e an tha t th e steady-state shock is located f u rther _ d ownstrea m f r o m the th roat, th u s reducin g i n let presSure re cov ery or efficiency, i Wi th this type of in f er,n a ti o n, the d es igner, a s Su m in g ho w f r eq uen tly a n un start can !i _ ' be t o lerated, can decide b e t we en th e peflal t y in _ fficiency r elated to larg e r _fl u e s of ii a an d th e de s ign p e nalties associa t ed wi th lar g er byp _ tss doors. _' I t , summary, then , these data _ t r e the pr e di cte d p e r _ orl _ a n ce t o t _ f amil y of opti m al inlet control s y ste m s. Sinc e th e inlet m odeled was avail ab le for an e xpe r t- I!

m e n tal pr o gram, o ne o f the _pt imal control s y stem s wa s _ r i ed.

Figu r e XI -$ 4 s h o w s o ne of th _ re s ult _ f ro m the expe r im en ta l progra m s et upto : do frequency response testing of an inlet. The f igure pl'ese _ ts a n opti mal co n tro l and a more classic al control co m pared on a frequ e ncy re s ponse b asiS. Th e dashed curve sho w s th e pel q or man ce o _ th e cla ss ical co n trol de s ignt _ t _ i th fr _ iue n cy d_m ain technique S . This contr o l, which i s S i m ilar tO th e ones di s cu s sed _ arl i er _ produced significant low f requeflc y red uc tio n of s h ock _ o { io n due to i ts in tegral a c ti on . T.o _ / solid c u rve shows the perfor m anc e of th e optimal inlet co ntrol. The desig n of the opti mal control was ba sed o n the assu mpti o h that the t ande m distu rba t i e_S had mo s t of th eir ene r gy at lo w fr e qu e ncie S , bt _ tthat some en e rgy did e xis t in the m idfreo " quencies.

Co m pared to the classical control, the optimal c on trol, i S seen to prodti e e | a s s attenuation of low frequency di sturb an ces; iu the m id _ re qU ed c ies i [ p r oduces gr _a tel"

I

atte n uat i on of s ho o k moti o n . Thus , CO m l _ tr ed on a frequency res p o n se b asis , o n e ' m igh t C on cl ud e th at th e cla s s ica l c on tr o l iS b e tt e z.

Wtl e n t hese co _t z ' oll e rl _ were eVa l uat e d o n a n un s tart b asis , however , t he bet - t er l o w freq ue ncy a tt e n u_ tt i on of the c l ass i ca l - c ont r ol did n_ t ind ica t e that it ha d ce r _ l ain m e rits over the optimal d esi gn , The resul t s , of t hi s eva l ua t i o n a re sho wn i n fig U r_ XI .$B , The mean tim e betwe e n tm s t a t _ t S f o r ea ch contr o l w a s c al cu la ted u sing th e _ peri m e n t al fr e q uen_ r_spe n se data. A lso ca l cul a t e d was the RMS valu e of th e e entrOl bypass a i rflow e ach w o uld e xpend, The ra nd om airflow d is t u r b a n ce ass um e d for both c ontro l s h ad a f r e qtl e nc y s p o ct _ u m id en tical t o that fc , t which th e o pt imal control w as d es ign e d. Th e two data p _ tnt s on th e fit _ tr e s how that th e o p ti- m a l control giv e s c on si d e r ab ly long e r m e an time betwe e n t m_ t a rt s , In fa c t , fo r this disturb a nc e , it-i s b e tt e r, than the c la s sical control by $0 to 1. Also, it a c co m_ pliS hes this w ith abo u t t h e same amount of RMS control bypass f low. From this e x- ampl e it c an be seen how th e use of optimal c on t r ol th eory c a n give considerable improvement when a re al criterion of perfor ma nc e is used in L _e design.

c on clu s i on , the _ , th e se results encou r a g e th e inve s tigation of even more ex- ten s ive applica t ions of this typ e o f theory in f uture control design b .

SUMMARY

L_ Adequate inl e t c o ntrol ca r t b e provided and inl e t- eng in e _ o uplin g can be me ch- _" anized and used to advanta g e, f In th e (ev e nt of inlet tms ta rt, r eg ta rt cont r ols ha ve proved succ e ss f ul wh e re the I : r e light of the co m b Us tor nlay Still be a problem. I Our und e rstandi n g of combustion d y na mi c s i s good. H ow ever, our under sta nd- tug of compressor system dynamics et fll has un k nown eff e cts. I.:[i The development of sp e cial sensors and s ervoact u ators has permitted wide ii I I ._ r ange expe r h _e ntttl pro _ tms.

,, F u ture rise s of digita l ¢ _ o m p u ter control a re clear _ in dicated. Applications o f ii '. cont*'ol th eory W ill point the way to the f ull utfliz a tic _ of thig e quip me n t to re a liz e I_i th e maximu m ef _e_ tivene ss of p rop U lSi on SyStems. }_

i

3V4 t \ BIBLIOGRAPHY Ar p _si , D ale J ,; Z e !l er , J oh n 1t, ! _ d B n tt er t _ ,, i , P e te r G, : Genera l P ur po se Di gi.

t al Sy s t e m f or On- L in e Control of A l rbre ath i n g Pr opul s ion P l_ t _. NA S A TM X-2 1 0 8 , 19 70.

Batt er ton, Pe te r G , ; and Z eller , J o hn H, : Dyngm in P_ r _ o rm a n ee A n ft ly H t_ o t _ t F u O l . C o nt r o l V a lve F or Us_ i n A t r b r v a t hi n g E n g i n e R c _ ca r eh 0 N ASA _ _, Do 9 3_ 1 , 1 969 . i , i , J Datt v rton , Pet e r G ; Ze ll er , John R .: Pori or mance Ch ar acto r i s t i e a o _ hupt _ ovod ; Servoa m plfflor for Ele c t r ohydraulie ContrO l B y st0m s . NASA TM X - 2 _ 0 7 , 19 7 0. i ' Bau m btck , Rob e rt J. : Device Fo r Pro _ etn g Dynamic Di s tortion Patterns at _ nlet s of A i r - Breath i ng Eng tftes. NASA _ X-20 | 6 , 19 P / 0.

Col e, Gary L. ; Neine r , Georg e H . ; and Crosb y , Micha e l J. : An Automatic Re s t a rt Control Syste m For a n Axisy m m o tric Mixed-C o m p r ession Inlet. NASA T N D- B 590 , 1989. L Cole , Gary L. ; Nein e r, George H. ; a n d Crosb y , Mic ha el J. : Design and Perform. _ I ' ant e of a Digital Electronic Nor m ld Sho _ k Position Sensor For M ixe d - Compression Inlet s . N ASA TN D-5 6 0 6, 1 0 (JP.

Coltrin, Robert E . ; and Calo g eras, Ja m e s E. : Sttpe _ soni e Wind Tunnel hlvesti 6 a - t i on of Inlet - E ngine Co m parability. Papel' 6 9-48 _ , AIAA , Jtme 19 6 9.

Coltrin , Robert E. ; and Mitcll e l l , Glenn A. : P re liminary Invstigatio n of Distortion _: 1 ; DynamiCS in a Mach 3 Mixe d Co mp r ess i O n hdet. NASA TM X-1 '] 06, 1968. I I Cole , Gary L . ; N e in e r, Geor g e H . ; and Wal l ha g_ n , R obert E. : C oupled Supersoni _ ' 1 hllet- E ngine Control U s ing Overboard Byp _ tfJ s Doors and Erq _in_ qpeed to Co n - I t r ol Nor m al Shock. NASA TN D-6C19, 19 _ 0. i t Cr o sby, Miclu _ el J.; Neiner , George H.; and Col e , Gary L. : R e s ta rt stftdHigh 1 Response T _ r m inal Shock Control For an Axi sy_ tnetri _ Mixed-Co m P re e Jsion !_ Inl e t With 6 0 -P er c e nt I n ternal Contraction. NASA TM X.1 _ 9 2 , 19 6 9. !

D ra in _ Daniel I. ; Bruto n , Wflliatn M.; an d Patdovich , Ft _ tn _ is j. : Airbi'eathing Propulsion Sys tem Testin g fJs i n g Sw e ep F _ ' _ que n cy T e chn i qu es . NASA TN D-5488 , 1 969.

Griff in , Willia m S. : Design and _ erfor man ce o _ _ 'ltte r ic Sho ck Position Sensor for a M ix ed-Co m p r ession Supe rs onic ifllet. NASA TM X - 1755 , 1969.

L or e n zo , Carl F, : V ar iable.Sw ee p - Rat e T e_ tt u g : A T e_ lm i ctu e t o _ pr o v e t he Q u ality a u d A c qui si t ion of F r eq u a n cy Re_pon_ e and V!b r atiou Data. N AS A TN D_7 0 2_ , 1 6 7 0, M i tc h e l l, G l_ au A. ; a nd _ ub b t _on , R ob ert W. : An Exp _rimen t _ l In v e_t i _ atio n of t he R e _ t ar t A rea P _ tio o_ a Mae h 3, 0 A xl _mme tr in Mix e d C om pre ss io n In l et, N ASA TM X _ 1 5 4 _ / , I EI O B , Nci ner_ Ge o r g e H., _er v o_y_t e m De s ign o f a _ iigl _ .R e t _ po n_ c S |otted _ Pl a t e Over- _ board By p _ t. _ s_ Va l ve f or a Su por Bo n in In l et, NASA T N D _ O0 _ ll, 19 ' 1 0, Pau l ovich , Fr a nci _ J. ; Ncinor, G _ or g e It, ; a nd I t agodorfl, R alp h ]Z.: A S upe rs o ni c _ Inl et= En gin e C o ntr o l U s in _ En K in e S p _ od a _ a Primary Variab l e for C o ntrolling N ormal S h ock P o sition, N ASA TN D.60 _ I, 19 '/ 0.

Se ldner, Kurt; and Oo l d, Har o ld: C om put e r and E n g in e Per f or man c e Study of a t _ en eral ized Para me t e r Fu e l Control _o r Je t E ng in e s. N ASA TN D._ 8 _ l, 19_ 0.

Szueh, John R . ; P aulovi e h, Francis J. ; I _ ruton, W illia m M. : S tudy of Turboje t Co m bustor Dy n amics Using Sweep-Frequ e ncy Data. NASA TN D-6084, 19 _ 0. !_ Wasserbau er , Joseph F. ; and Whipple, Daniel L. : Expe r i m en ta l I nvestigatio n of _ the Dynamic Response of a Supersonic Inle t t o External _ d Inte r nal DiStur- i banc eS . NAS A TM Xo1648, 1968.

P Wa s serbaue _ , Joseph F. : Dynamic R e S po n se o _ a M a th 2 .5 Ax iay m m etri c Inl et ! / With E n gin e o r C o ld P ip e a _ ! Uti l i * in[ _ 60 Percent S uper s e d e Internal Area i ' Contt'action. NASA T N D- _ $$8, 1969. !i Wenzel, L e on M. : E_ periment al In_ , e S tigation of th_ Effects of P u lse Pre s s ure I_i D istor t ion Imposed on th e Inl e t _ a Tu r bo f an E ngine. NASA TM X-19 _ 8, 1 9 _ 9.

Wilhel m , Walte r E . : In ves t igation of Tu bing Effects on , _m plltu d e Frequency Re- !!

p s po n s e o _ Pressure Sensin _ Systems Using NOnresonant Ter m inatio n s. NASA _ '4 TM X-1988, 19 _ 0. i_ ; _ Wflloh, _ o sS G. : A Math em a t ic al A n al y s is of Super s o ni c I _ de t Dy,an _ i eS . N A S A TN _ )-4969, 1968.

Willoh, Ro ss G. ; al _ dSel d f t er, Kur t : Multi s ta g e Co m pr e s S or S i m ulati o n Applied th e Prediction of Axial Flow Instabilitie S . N A SA TM X-18S0, 1969.

, / Zeller, John R. : Design and Analysis of a Modular 8ervcamplffier for FaSt- : _ Response Electrohydraultc Control Systems. NASA TN D-4898, 1968.

' Z,-I Ze U er, John R. : Analysis of Dynamic Performauce L i mi t at i ons of Fa st Response (150 to 2 00 Hz) Electrohydraulic Servos. NASA TN D - 5 5 88, 1969.

.

' PROPULSION S YSTEH

INLET _ CONTROL ENGINE CONTROL i I I C _ -56909 t: FigureX I - 1 I_ I

.... _ IDEALIZED INLET WITH TE R MINATION

BYPASS

FLoW

SHOCKPoSr rl O N ' 1 "t ERMINATION (J OROTHER C8 - 5690Z Figure X! - 2 1 $ 't 8 FREOUENCY RESPONSE OF SHOCK POSITION CO M P A RISON OF AN AL YTIC A L A ND EXPERI M ENT A L RESULTS NORMALIZED . 4 --- ANALYTICAL A M PLffUDE RATIO ,2 o EXPERI M ENTAL ,0 I ' 4 ' / , DEG -200 : . - 3 0{ _i _l ,t I t ill I !

L 2 4 1020 4 0 1 00 4 00 , FREQUENCY, Hz cs.s 69 oo , . F igU re Xl- 3 : " SHOCK POSITION CONTROL SYSTEH I '; SHOCK POS fflON " : ' .., ; : BYPASS FLOW [. ' y TEI _ .!NATION t S E NSOI _ ACTUA TO R , COMMAND CON T R O LLI[ _ + CU-S6 8 93 Figur_ XI-4

STO

t

' i

FREOUENCY RESPO N SE OF SHOCK POSITION

C O M PARISO N WITH AND WITHOU _ CON T ROL 1 , 0 .......... -. .._ CONT ROL / _ .2 " + NO R MALIZ E O AMPLrrUDF RA!'! O . I -- ,0 6 i '.40 ! _ +' , _ .04 - . , , ' WITH CONTRO L

-/ / i

. oi _ I = II I I II I I

.4.6 I 2 4 6 10 Z O 4 0601 0 02 0 0 FREQUENCY , HZ cs - , _+ a g_ Flgur _ X I-5 , %

COUPLED CO N TROL SY STEH _

, _ B Y P A SS DOORS CO NT RO L SHO C K POSIT I ON ' " SHOCK PO S ITIO N J85 ENGINE ., _ , BYPASS FLOW F UE L F L OW i t ii : ' , .+ ."

TRAN S I E NT RESPONSE OF COUPLED SYSTEM DISTURBANCE h T 1 _ _ k OFTOTAL INLET FLOW FLOW I I BYPA SS DOO RS CONT R OL , i S HOCK / SHOCK POSITION i, POSITJO_ _ _ ' L.

CONT R OL BYPASS DOORS - - I I _ , ENGINE SPEED i , , 0 5 10 ; TIME, SEC .... cs. s6914 i Figure X] - ?

i ' ii_ COUPLED CO N TROL S Y STEM EN G INE SPEED CONTROLS SHOCK PO S ITION PO S ITION ' _i BYPASS FLOW t : J85 , _ IGINE , C O MM CONTROLLER C S - _691T t H ROTTLE Fig u r e XI-8 $ 81 TRANSIENT RESPONSE OF COUPLED CONTROL SYSTEH DISTURBANCEFLOW _ ,_T I _ %OFTOTAL INLEt FLOW J _ .J EN G INE SP E E _ CONTROLS SHOCK [_ SHOCK POSITION POSITI ON

K _ I t

J _ lO 0 RP N t ENGINE SPEED _._ r_, .,_l i i BYPASS DOORS CONTRO L _ II

, o 1 2_ !

TIME, SEC cs.s 691 o !

Figure XI - 9 I INLET UN S TART AND HAHMER S HOCK DUE TO COMPRESSORSTALL _" I!

A NALY T ICAL PRESSURE PROFILES ATFIXED T IMES J EXPER I MENTAL ,, 15 L PEAKPRESSURES' 7 ._.. . ..... . .4 - . . .. . ..... __

/

L B / I N . _ : PRE SSU R E , 10 5 O I - - *l I I I I ' " 0 1 2 3 4 C OWL LIPt AXIAL DI S TANC E , Fr COMPF A C E _ CS - 56898 FigureXI -lO 1 3<12 INLET UNSTAR T DUE TO EXTERNAL G UST , " ANALYTICAL PRESSURE .PROFILES ATFIXED TIMES 1 0 12 MS 0 ,_ LB / IN. _ # d _ I "-8 I I I I

u ° - Pt z z 3 4 : COWL

f AXIALDISTANCE, FT COMPFACE_' _ _!( c s- 5 6 897 FigureX I - 11 I_ i_ ° , !!

F INLET. RESTART REQUIREHENT8 - NORMAL t A (_ BEFOR E RESTART r _ ) i 1 ' FigureXt - ]2 cs - _ s e o 4 : r t RESTART CONTROL. SYSTEN

, _ , " - - 1 tuN ' A m °

t STARTED , /' -REF-,, ' UNSTART SIGNAL FORE " POSITION OPEN _ CLOSED __ AREA THROAT EXITSTATIC PRESSURE _ . ! II. _ r " ' _"l j , I , , , , I , , • , r , 0 .5 1.0 1.5 TIME, S EC _ C$-5689 Z FigureX | -14 ................. I I.

CONTROLLED RESTART COMB U STOR RELIGH T . ,_REF . f / !

, i I I I j -" L _". t _ .

F ORE C £ NTER B ODY POSITION UN S TAR T SIS NAL A FT ......................................... _ ; _ OPEN _ _ O R A r EA _ .;i CLOSED '- - ..... ' ' ' : _' .... THR O AT _lT STATIC P RESSURE _ . : 2 I I I I I ' ' " "

_. _NE E XIT T_ P ....... '

• 0 .2 .4 .64.8 P.O ,_ .2 51 4 P. 6 I_ ' T I ME _ cs. _69 oe Flg u reX ! -15 '" IDEALIZED COMBUSTOR "_i :' SECONDARY t • .' COOLANT AI R ; .....

COMBUSTIONI " ' _ _"_ __.

_ A IR "" FUEL _ '' COMBUSTION MIX I NO = mtURBINE FLOW : . _ .: , _ ZO NE ZONE FigureXI-16 c _ -s 6 9o 3 • , 866 , ,, ,' . .

, J RESPONSE OF COMB U S TOR PRE S SURE FU EL F LO W DIS T U RB A NCE N O RMA L IZ ED :_ 1 ' 0I--" RA T I O ...... ! .......... - ANA L YSIS A MP Lff U D E : _ i--I o EXP E RIM EN T O_ P HASE A NG LE - t O 0 _ _ l_ DEC - 2 0 0

- o o I I

2 46 10 20 40 tO O FREQUENCY , HZ c _.e_. , i z Figure XI - 17 J- 85 CO M PRESSOR PERFOR M A N CE , • STALL - ANALOG SIM CORRECTED SPE E D , n STALL - EXPERIME N TA L % ' : 9 7 . 5 _ L I00 . 0 92 , 5 t PRE SSURE 8 7. 0 e ,, ¢ :__ - 90 , 0 3 I I .....I .I t, ,1 : I . . 30 32 3 4 3 6 38 40 _ 44 ' C ORRE C TED AIRP LO W, L B / SE(; c s -5 6 8 , _o Figure XI.-]8 i l q

F REQUENCY RESPONSE O F SIXTH. S TAGEPRESSURE

FUEL FLOW DISTURBANCE _ J85

^M PL . UDE RATIO

I / i PHASE ANGL E , " 100 - i ' ; BE G -2 00 i ! : " O_ I, .

- 3 00 I I _

2 4 10 2 0 40 1 00 4 OO i ; FREQUENCY, H Z cs - _9 o = i] I / F i gure XT-1. 9 i

TF - 30 TURBOFAN SCHEMATIC

L r FANDUCT t I i ilu I , I 1 _ ( --- -- . ; _. _"

i \

I _- HIGH COMP I t t L OW COMP c s- _68_7 Figure XI-20 ) • , , . , : 38? i RE S PONSE OF TURBOFAN ANALYSIS INL E T PI _ F . SS UR E DI S TU R BANC E TE-30 r _ F A NDI S C H ARG E (C O RD /F F A N D I S C H ARG E (D U CT) // F LOWCOMP DI SC H ARGE NO R MALIZ ED , 4 A M PLr r U DE RATIO .2 t,O e . _

, I _.. HIGH co MP D I S _ . ARGE - -' \

,o 6 -i II I I il'i

I 2 4 6 10 2O 40 40 ZOO FREQUENCY, Hz cs . s6s 4 o Fi g ure X I - 2 1 RESPON SE OF TURBC _F AN EXPERIHENTAL INLET PRESSURE DIS _ URBAN C E I TF-30 , _ FAN OISCHARGE (CORE) / r FAN DISCHARGE (DU C T) s t /r , LOW COMP DISCHARGE 1.0 ..... _ '-- - --- / t

. , _

AMPLITUDE RATIO .2 --, HIGH COMP

.06 I I I I,l,.ll !

2 4 6 10 20 40 60 100 FREQUENCY, Hz c s . ses4 1 Fi g ure X1 - 22 / , . , , , . , , r . 4 ,f ' 0 IDEAL INLET STATIC PRE S SURE PROFILE :_ B

I ' LOCAL S TATIC ./

PRESSURE H I,._" .,. . ... j !,, , : M >I _ / _ _!:_ ! ! _THROATEX I T i, N<I- J STATIC PRESSURE _'_ i .

Figure X1 - 2 3 cs - 56 9 06 .j i ¸ : i v T Y PIC A L REAL I N LET S T A TIC PRESSURE P R OFILES R _ ' ._' " .... ': " _.6__(Drr__ _. 8 A N G L E O F A TT AC K, 2. 5 0 P R ESSUR E t _ :, L O CAL STATIC 12_, J 4 5 i_ ." I I111111 LOCA L STATIC I " .... )t _ T -IR E R

.. P RES S URE / 2., 14, ' -

11' 1111111 ' TAP L OCATIO N. cs -s6 9 os Figure X|- 2 4 3O9 ,; r ' t ' ELECTROHYDRAULIC SERVOS YS T EM-.

• '_ I , u r . I E l e ctro- Hydraulic output _ h ydraulic actuator

, I-t k Desired _ _ Actual

L servovalve andload _ output i mim q lib limb l ira l in t _ qllml -J ServoamD Iifler transduc e r

• I -c k l __

ii'

Flgure X I -2 5 1 FREQUEN C Y RESPON S E OF FAST RESPONSE SERVO FUEL THROTTLING VALVE IN.

0.0 9_ , 1'0 E t "]'1 _ : --:" _ _ - s ' t IN . ..- .0 _ j : - .01-- i . O0 1_ - i i l l i II I l II I I i,l l O 1® 1000 1 0 , 000 FRE Q UE N CY, HZ cs . ses9o Figure X1-26 FREOUENCYRE8PON8E.OF RESEARCH8ERVOS COMPARIS ON OFTWO S ER VOS OPERATING AT 40 % A MPLIT U DE t 4 - DISTORTION VALVE SERVO, 2 - LBLOAD 2 - O PTIMUM DESI G N _. ,,..,.

NORMALIZED AMPL IT UD E 1: -- • : ......

RATt o - F U E L V A L V E- S ER VO, O ,6 -LB LOAD / .4 N ON - OPTI M UM DESIGN .. ' / .2 I I J I I I I I I J I I .4.6 1 2 4 6 10 20 4060 100 200 400 FREQUENCY , Hz F lg u reXt - 27 c s.sG e 89 DIGITAL COM PUTERCO N TROLS Y STEM ' AIRCR A Fr . P R OPULSION " , i ll ....

:_ , . CON T ROL COMPUTE R _ " PERFORM A NCE EXTERN A L (THRUST, S .F.C.) I _ NVIRONMEN T r CON S TRAINTS (ST A LL, UN . _ ;TA R T) DY N A MR _ S (STABILITY , .RE S PONSD CS-S6919 Figure XI - 28 1 i e SAMPLE RATE REQUIREMEhT S R A NDOM SIGNAI _ BANDPASS 0-10Hz SAMPLES, S_ i , 1 0 ;' f 1 SEC - _ i' CcJ -5 68 : _ 9 FlgureXI -29 _.. ' P i' ' i S HOCK POSITION CONTROL S YSTEH i_ DIGITAL C ON T ROLLER ' _ " r SHOCK POSITIO N , BYPASS FLOW _ t INLET TERMINATION SENSOR ACTUATOR (COMM ANO) DIGITAL .

CO N T ROL L _r i " 1 - C$-56 9 16 FigureXI-30 $ g2 t \ o FREQUENCY RESPONSE OF SH OCK PO S ITION COMPARISON OF ANALOG AND DIGITAL CONTROLS ,6 - 1.0- NOR MALIZ E D , _ AMPLITUDE . 2 RATIO .06 - ' ANALO G .04 - -- .... DIGITAL 1000 1 $ EC -" ---- " DIGITAL IO0 1 SEC i_ .02 J ,I I ..I I I I I I I _j .2 .61 2 4 6 10 20 40 d O 1 0 0 200 I , FREQUENCY, Hz cs-s6 e94 )! " Figure XI - 31 I; ., IDEALIZED INLET WiTH TERMINATION

/ a

THR OAT _ / " P OSITIO N __ BYPASS ROW TERMINATION : " . ,[ Figu r eX | ' 92 c s- s 6 9 1 s 393 _ .: : // APPLICATION OF CONTROL THEORY _ _ TO A SUPERSONIC INLET tO00 - a4 o , MEA N TI M E ] _ - a 2 BET W EEN UNSTA R TS , 10

HR \

i a l .1 . 2 . 4 .6 1 2 3 c s. s 6 9 z z RMS CONTROL BYPASS FLOW, LB I SEC Figure ) <I -33 FREQUENCY RESPONSE OF S HOCK POSITION COMPARISON OFCLASSICAL AND OPTIMAL CONTROL ¢4 • _ .6 _ t N O RM AL IZED .2 , , , _ PLITUDE RA T IO. 1 _ , I '- C LASSICAL CO NT R O L .06 ' - .0 4 - .02 i i I, I I i I l ! i I I I .4 .6 l 2 4 6 1( 3 20 40_ 0 lO0 FREQUI_NCY, HZ c s - 569 , 1 Figure X I -34 3 94 INLET UNSTART PER'FORHANCE COMPARISON OFCLASS ICAL ANDOPTIMAL CONTROL lOO0 -

l oo ii

MEAN TIME I_ , BETWEEN 10 - • OPTIMAL U NSTA R TS , HR i' i - ! : i • CLASSICAL j i I IJ l i_l I; , .2 .4 . 6 1 2 3 L cs . s_gl3 RM'SCONTROL BYPASS ELOW, LB / SEC g F Figu re X1- 3 5 fi t t ;;

i

3 9 S -, !f : ' r t . i¸ X I t. CRYOGENIC FUELSFORAIRCRAFT , _ '.

JackB . Esg a r _ 7 1 " 1 9 4 6 3

I , T hi s di sc u ssi on co n c e rrm s o me e xp l oratory re s ea rch bei n g done by NA SA on th e _ .

u s e of cryo g enic fu e l s for airbr e at hia g gas turbine e n g in e _ . Some of the results o f this re se ar0h m ay not f in d i mm ediate applicat i on, b ut a po r tion of NASA 's (and i ts , : pred e cessor N ACA) research has hi stor i ca ll y been devoted to advanced c oncept s .

For exa m p l e t NACA began research on tu r b in e c ooling almost 1 5 yea rs before it wa s embraced by in dus t ry, if!

Both N ASA and i ndustry hav e developed a s ign ifi ca nt background in the use J md i!

handli ng of cryo g en i c fuels f or space ve hi cle boosters. This ba c kground lnforma - ' * tion serves a s a n a id in the exp l oration of pa y o f fs t ha t nd g ht be obtained i_ c ry oeertie I f uels are to be used i n aircraft e ngin es. T hi s paper di s cusses th e possible appliea - ' tion of liquid met "bane to a s upersoni c transpOrt type air c raft and the appli c ation o f I_ liq ui d hydro gen to th e ai rbreat hi ng engines f or recove ra ble boo s ters and orbiters f or I th e space shuttle, i_: k F U el Propertle _ , The two prime reasons for the interest in cr yogenic fU 6 1s for adva n ced aircraft are the higher heat in g val u e per pound of fuel (relat i ve to Jet A) and th e heat sin k _ t capacity that is availab l e in this fuel for cool in g hot compon e n ts in the engine or th e aircraft. TheSe advan t age s a re at l eas t partially offset by th e co m p l ications in- vo l ved by havin g to stor e this fuel at a very l ow te mpe rature and b _ t he fact thM the fuel has a lower density. A lar g er- storage vo in me arid in su _ t _ < i tan R s are requited .

Comparative fu el property trends a t e s how ' _ in t | gu r _ X] _ - 1 and tabl e XII - 1 . f n !

the figure the various f uel properties are p l otted as functions of the hydros _ n-car b on '_ ra tio H / C. The vertical das h- dot lines indicate t h e h ydro ge n-earbofl ratios _ or Je t A, propahe , m eth an e , an d h ydl'ogen fu el s . Th _s_ trend L ' df _ , esi n dicate that the ' heatin g va lu e, the he a t sink, an d the m aximutn u se te m per a hire (b _s e d on coking o r fuel c rackin g ) a ll i ncr ease w itl _ incl' e a s in 8 va l ues o f the h ydro ge fl - cai-bo n ratio .

I

P R EGEDR_G PA_I_ BLAN K NO T F I L MBI , $ 9 _ • 4 • Converse ly , t he boiling te mpe r a t u r e an d th e d e nsit y d _c r ea_ e with an i ncre aSing hy dr og en - c a r bon r a t io . The cos t a f aom e Q f the se f ue l _ , _u eh a_ p ropane and m e th- ane , a r e n o t accurately known f o r lar ge s uppl y quantiti e s a t an a irport. But th e i n- f o rmati o n that i B a vailable i n di ca t es that Jet A, propan e , and me t hane would c o st ap.

pr o x ima t e l y t h e came o n a Btu b as i s . T h e c o st o f hy d rogen i s a bo u t si x ti me_ th a t of the other _e ] _ . F i gure XII-I al _o s h Q w e th e fla mmab l e f u el - ai r rat io r an ge for th e four f u e l s . Th e fla mm ab le ra nge in it i ally d ec rea s e s w i th a n i ncrea _l n8 h y dr og en- carb o n r a tio t o t he m i nimum va l ue corres po n di ng to that f o r methan e . H y d r ogen , with a n infinite h y dro g e n-carb on ratio , is a v0ry r e , r e tir e f uel and ha s a wide f l am- mability range.

The l ow flamm ab l e f u el -a i r ratio ra nge f or methane c an be b ot h an adva _gc and a d is adva n ta g e. It is a n advantage from th e po i nt o f view o f saf e ty , b e caus e i gn i t i on ca n on l y o ccur w i th in a narro w range of fue l -a i r rat i oS. It i s somewhat of a disadv an ta ge f o r the co m bustor. HaVi ng a narrow f lamm able rai s e of fu e l-air ratios c a n caus e problems i n the co m bustor, particularly i n blowout li m it s and more diffi- cult ligh to ff.

The ig n ition an d burnin g characteristics of lar ge pools of li q uid nletliane were investigated at Lake Charles , Lou i si an a i n 19 6 1 by Conch Me th ane Services li mi ted.

These tests were conducted on open pools o f m ethane enclosed w i t hi n a dike. Th e largest pool tested was 20 by 20 feet and contained liquid m ethane to a dep t h of 1 foot. The foJlow i ng co n clusions (re _ . 1) were drawn f ro m thes e tests: " (1) The f lamm able zone of vapors was conf i ned well within the v is ibl e vapor (condensed moisture) cloud e m anati n g f z ,o m th e l _oo l.

( 2 ) Although the vapor cloud remai n_ visib l e for a di s tanc e of S evera l p b o l dia m - eters downwind, the f lamm able zone of Vapors wa s confined to a few inch e s dir e c t ly above the pool and to less than one -ha lf the pool di a m eter i n the ho ri zo z ltal downwind direction near ground level.

I (3) The radiation flux level fro m the burning LNG pool w as sufficient to cause ignition o f wood under low wind conditio ns w i thin a di stance of about 1 .2 9 pool di am- eters from the edge of the pool. " The rela ti ve safe t y of ha vht g " spil ls" o r open pools of li q u id m e thafle was i llus- trated in a mo vie of these tests w hi ch showed a ma n w alking up to t he ed g e o f a large pool of liquid m ethane ho J di ug a li ght e d tor ch in his hand. l _ I e l o wer e d th e t orcll down to wit hi n a few inches of the pool level. At this point, the p o ol slowly ig t li t ed and the m an walked away unhar med .

The data for figure XII-1 are given in table XII-1. This table giv es the m ore ' specific va l ues of the heating va l ue, li m itin g t 6mpe ratures, heat sink, boilin g te m - perature, de ns ity , flammable limits , and costs of Jet A , prop _ fle, me thane, and h y- drogen fuels. A l though propane appe a rs to ha ve many d e si r able proper t ies as illus- q tra t ed i n tabl e X I I-1 , i t ha s not b ee n eo n eidured s e r i o u s ly a s a _e l _ o r a irc raft.

One reaso n i s tha t a t pr esen t propm le l e a b yprodu c t o f petroleum p r o eees l n_ p l an t _ , and large sca l e pr o du c tion of propane wou l d require major p h tflt clz _ g e_ o r c on - struct io n o f n e w pl _ t s . Pr oba b l y more im p o rtant 1 _ t h e fac t that prOp an e vapor l_ hea v ie r th e m ai r; the ref o re , wh en a _ p l l l occu r s t he vapor h _ a t [ _ ro un d l ev e l a n d p r ov i d ed a s af e ty hazar d . I n th e c_ t se S o f m ethan e and hydroi _en , ti _ v a por de n_ lti e s a re les s tha n that o f a ir , a nd the vap or so on d is p er s e s i nto the at m o s ph e r e _ /te r a s p il l. Th e h u z a rd is thereby dcsr eas e tl.

Liquid Methane for Supersonic Transport

Mi ss i on an al y s es , as report e d i n references g to _, h _ ve s hown t ha t the p a y l oad or di r ect op0ratin _ cost of a supersonic tr ans port could be a ppreciably i m p r ove d by us i ng l iquid me t han e as fu e l i n pl _ ce oLthe more co n ventional Je t A _ e l . The gains that can be obtained depend on the type o t en g in e used in th e aircraft , t h e mission, a nd tl _ e noise c o n straints. It there w ere no n ois e constraint s_ the direct 0perafln _ cost or pay l oad could be improved by more th an 30 perce n t b y swi tc hiflg fro m J e t A to liquid m etha ne as a fuel. With nois e constraintS , which must be considered for com m er c i a l aircraft, the improve ment in payload or direc t operatin g cost i s r e - duced; but the benefits ar e enough sup er io r to those fo r Jet A fuel th _ it is wo rt h- while to continue research on the problems relat in g to the appli c ation o f m etha t ie fuel to superso ni c cruise aircraft. The gains in a / r etail per fo r m anc e are dtie to : : pr i marily two factors. FirSt, l iquid methan e has approxim _ tt e ly 1 8 p erc ent fdore energy p e r pound o f fuel than Je t A fuel. Se cond , th e heat sink o _ me thane can be u p to

tS h

71 t i mes that of the Jet A f u el. Thi n high h e at si _ capacity is very useful fo r ga speed aircraft for coolin g the engine co m pon ent s such as th e tt _b ine , c ooling t h e lu- bricatin g o il an d hydra uli c fluid , and for controlli n g t em peratures in th e pa s senger t cabin. Using the m ethane heat sink fo r re du c in g t u rbi n e cooling ai r t e m perattzre call substanti a lly r e du c e the am o unt of compressor air tha t m ust be bled f ro m the en gin e for turbine cooling. T hi s reduction i n compressor air bleed can s i gnific an tly fin- prove the en gi ne performanc e .

There are also a number o f p ro bl e ius in v o l v ed with u _ i n_ liq b td m et ha n e as an aircraft fuel. The most si g nificant problems requirid g research a t @ th e fo ll owi ng : (1) Fuel storage in ai rcraft (_) Co m bust o r develop m ent (3) Engine fuel control and pum p in _ e yst _m (4) Gro u nd ha ndling an d stora g e in lar ge q u antities (5) I m p a ct on aircraft size _ d r a g , and design 1

!

3 99 (6) Future me than e cos t an d W orl dwid e avail a hi tit y R es ear c h is pr e sen t ly un d e rw ayat NAS A L e w isRese ar c h Ce nte r on th e f i r s t t hreei t ems . R osuRs oft hi s resea r ch a re d isc u sse d la t er in t his pa pe r. I n v es tl g a _ t i o nS o f th e r e m ai n in g th r ee ite ms ar e expec t ed to be c on d uc t e d i n the futu r e. Some di s c uss i o n of t hese rem ain ing th ree i t e m s i s re q ui re d , h o w e v e r, Th e re I f_ c onsl d er_ a b l e ex p o r le nc o wit h th e gr ound h andlin g of l i q u i d me t ha t m g a s , Li qui d nat u ra l gas (ab ou t 9a p e r c en t me tt mne ) , whi c h is pr o O _cec t in s i gnifi c ant q ua n ti ti e s inNor t h A f r i c a , Is t ra n s p o rt ed by s hip t o va r i ous p_ trt s of t he Wo rld . A t t he u nlo adi ng o_ ' eas, i t isst o r e d in v a ri o u s t y p es o ffa ci lities , pr im aril y i n a bo v e - groun d i ns u l a te d tan k s .

T h 0 han dli ng an d s tor age p r ob l ems of liq u id z _ t e t h a ne ( orl i q u i d m tt ur al gas ) forair- c r aft c o uld bO s o m ewhat d if f e r e n t, Fo r som e a i rp or t location s in the w o rld , th e m O Um newou l d p r o bab ly be s tor e d as a c ry ogen i c liquid a n d c o uld be pip e d t o a n d p um ped i n to th e waiti ng a i rc raft . F or o t he r l o cati ons , it ma y be desi ra b le t o pip e th e met ha n e in th e for m of n _ tur al gas to t h e air port. At th e a irp or t a liqu ifl cati o n ' .

plant cou l d produ c e the liq t d_l me tha n e w h i c h wou l d b0 sup plie d to th e aircraft. X n- ve s tl g atio n s ar e req ui red to d et er m ine t it s t e c l mo l o g i eal a n d ti n,ra ci al e ff e cts that !

th e use o f cryoge ni c fuels wou l d ha ve o n airport fuel h an dli ng probl e ms .

Th e use of li qu i d m etha n e fu e l will ha ve a si gn ifica nt e ffect o n t h e aircraft d e - i sig n , Liqu i d m etha n e has a d ens ity _tb out on, -hal f t ha t o f con v ention al Jet A fuel s , i.

As a r e sult , th e steroi ds of Li qu i d metha ne i n the a i rcra R would require i n creased _ tank volu m e; therefore, a i rcra R size would pro babl y i n cr e ase. Thi s in creased siz e could affect the a i re r a R drag characteristic s a n d cou l d dimin i sl l s o m e of th e g ai n s i re sulti n_ from s w itchi n g to l i quid m ethane a S a f u e l. A n i n ves ti g a tion is also r e- q uire d on the stru c tu r a l pro b lems t ha t would b e i n v o lv e d by stori n g _ tc r yog e n ic fuel I with a te m pe r ature of app ro xim a tely 4 ]60 ° F in an ai rcr a_ wh e re th e outside st ruc- tur e o f the ai rcraft w o u l d b e h ea ted t o te m peratures between 40 0 ° a n d 750° F ( d e = p e n d ing o n the f lig ht M ach n umber) by aer o dy n amic heatin g . The therma l dis t o rtio n t problems that result f r o m these lar g e te m p e rature diff e re n ces r e quire a d e ta i led des i gn study.

P rel i m in ary in vestigat i ons have bee n made on th e proJect e d cost an d wo / 'ldw i de availabi li ty of l i q u i d m ethane ff i t were to b e used for a superso ni c transport f l e e t.

These pre li minary i n v e sti g ations i n dicated the cost to b e competitive o r pro bab ly _ l oW er t h an t hat for Je t A fuel in mos t o f t he wor l d' s air po rt s. This p re li min ary in- !_ h ve s tigat i on was co n ducted pri o r to th e curren t co n cern ove r th e possible shortage of f fossil fuels for both household and i n dustrial n eeds. As a r e sult , a fttr t l _e r tn v es ti - l_ gallon i s requ i red on th e feasibility of usi ng li qu i d m etha n e for a world W ide fleet _ o ' f .....

superso ni c tra n spo rt ai r cra ft .

Supe r sonic transport aircraft show the gre a test adva n ta g es from sw i tch ing f rom 4 OO J e t A fuel to liq u id me th an e, p r i mar ily b e t 'a ss o f t he need for a l a rgo h eat m ink c a- p a c i ty r e_ u l t ln g fr om bot h tur bi ne c oo ling r tul r em ent s a nd a er o dyn am ic h eat input.

In add i ti o n , th e. nu m b e r o f ai rp or t s in th e W t _ td t ha t would h a v e '_ o _ upply llqU i_ me t h an e would be mo re ' limi t e d th o r n if th e f t _ e , w er e to b e use d _o r other a i rc_ t.

Ther e a r e _om e a dv a nt a g e _ , h _ ve y er , t o u _I nt _ liqtgd me tlm ne i n s ub s o ni c a ir c raft _ tn d in he lie o pt o r _ . T he e ngi ne tr e ad fo r th o_ e : : _ rera f t I s toW A rd hi gher turbi ne i n - let te m per a t u res . With t h e hi gh er i nl e t tem pe _a_ _ re _ the heat sl nk ca p a c i ty o t li¢lu l d m etlm n e ca n h e u sed very _d v _ nt ag oo u sl y to lowo _ t urb ine coolin g _ r te m per a ture an d th u_ red uc e t he quant i ty _ f ai r r e q u ir e d. The _ m a iler c o o l i ng a ir q uantitie s i ra ° prove e ngin e per f or m anc e . The hig her heating yah , __ o f l iquid m etl-, _ .ne, compared to J ot A fu el. lfl al_ o _n ad v antag e . Although t h i B d tee , _ : _e lon o n li qu id m eth a ne r e f ers m o s t sp e c if ically to _ upors o n i c transport a ircraft, m. , _ t of tile r e se a r ch being con- d ucted is equ al ly a pplic abl e to s u b s onic ai r craft a nd h_ ' teopt e r u se .

L lquld Hydrogen for S p ace Sh tle

For cnnven tional s u b s o n i c a n d supersonic air , ra ft wifl _t s _ urbine engines, liquid hydrogen pro ba bl y would n o t b e considered a s a suita _ , _ e f, _l . The pri m ar y re a son is that i t has a n extre m el y l ow d ens it y , less t han o n e- _en_i_ that of convene tion a lJet A fuel. T his low de n s ity c re at e s a v er y diff i cult Stc _ a ge problem i n con- ve n tio n al a i r c raft. A second reason i s that hydro g en ha s very _ v i t _= fla mm ability , li m i t s. T h e dange r o f ig ni tion an d explosio n w o _ Id be si _ nlflca , _ .' i _ no r m al everyday u s e. T he space shuttl e prese n ts a u ni q u e appl i ca t io n for li q t dd , _i_ dr , _ , _ ; e n as a fu el for a i r b re a th in g e n gi ns s. Th e space shut tl e w ill use liqtiid hydrc _ e z_ as a fu e l for the propulsi o n rockets, Therefore, hy d r og en w ill a lready b e a va I_ ,: _ i n the vehicle.

The plans for the space shuttle call for airbreath ing propu l sion in _ , _ _ cover a bl e ?

b o o ster a n d recoverable orbital vehi c les. The s e e ngi n es w ould be t : _ to obtain th e required cross ra n ge a n d to ha ve t h e cap ab ility f or go-a r ou n d a t th e _._ ort upo n la n di n g. U s i ng hydr o ge n t ha t is alre a dy i n the aircraft eli m inat e s th e _:q_ : essi t y of pr ov i din .g separate ta n ks an d fuels for these engin eS. A n othe r advant t_-:e _ h ydr o g e n is t ha t it has more tha n 2 _ 2times as m u c h en e rgy pe t po u nd of fuel aS _ , ) a,. vn tio n al J et A fuel. T hese c omb in e d eff e ct s could re sult in e ztr t er a v ehiv le w ith r , _ d t ieed weight o r increased payload if the ret u rn ai rbre a thi n g pro p ulsion iS req _i , ,ed.

B ecause of th e limited locations wh ere liquid hy d ro g e n will be used i e _ _ spac e s h uttles, and t h e f a ct t h at provision s will a l rea dy r equ i r e su p plying li q uid _i :drog en into the ve h icle, t he r e a re fe wer proble m s t ha t re q u i re i nvestig at ion in th _ _. pp li ca- tio n of l iq ui d hydrog e n to the ai r breatht ng engines tlum there are f or li q uit _ : ., _ tl u hie.

Th e only problems that appear to r e quire investiga t i o n a re i,q , , . ' i ¸ ' ' : (1) Co m b usto r deVe l op m en _ (2) Engi ne fuel control and pumping syste m The c o m bu s tor develop m ent i s e xpected to b e a m inor problem becaus e h yd r og e n i s a _ extr eme ly re a ctiv e fuel and burns e a sily. NACA conducted tests on gas turbine engin es burning hydrogen in the late 1950' s . These tests were conducted b o t h in ground t est engines a n d flight test engi n e s . Th e s e t es ts did not thoroughly i n vesti - gate the liquid hydrogen pumpin g proble m s nor engine tran s ient operatio n . As a re- s uit , additional r es earch i s requi r ed on en gi ne fu el control and th e liq ui d hyd r oge n pu m pi n g sy s t ems . This researc h i s presently underway at NASA Lewi s Research Center.

HEAT SIN K UTILIZATION

Two methods of uSing the heat sink of cr yo g e ni c fuels are shown in figu r e s XII - 2 and XII-3. Figure XII-2 is a result of an analytical s tudy on the heat eXchangers re- q ui red to reduce the cool in g air temperat u re for-an engine that can be uSed i n a Mach 3 transport ai rcraf t . This engine w o uld hav _ a takeoff atrflc _ of 47 5 po un ds per second. TWo heat exchan g er s would be uS ed for reduc in g cooling a ir t e m pera- tures. These heat exchanger s would be annu la r in shape and wra pped around the inn e r and outer diameters of the combust o rs. Th e di m en s io ns of the h e at _ change r s are sh o wn in figur e XIi-2. The heat ex changer s are q uite small and W o uld have v e ry ' li ttle effect on the overall engine dim enSions. Th e co m b ine d weight of the two h e at exc han ger s is less thafl l pe r c ent of the en _ e w eigh t . Wlfl _ h ea t ex cha n g e rs of this ' size th e c ooling air temperatur e could be r edu c e d more t ha n 400° F. T _ s r e thtct _ on ":": : ": , i r e sul ts In a Si gnificant red u cti on in c o oli ng ai l_lC _ v. Th e am ount of coOlin g airfl ow ' i reduction is depe n d en t on th e turbine inlet t emperatur e level an d th e m ethod of cool - ing used in the turbine. Reductio ns of at l e ast 50 p e r cerit In th e coolant flow re- r quire m ents are eas i ly possible w i t h a 400° F t e m pera _r e red uctio n of the c oo ling ai r.

Fi gu re XII - 3 i llustra t e s a di r ett c o v ting a pplicatio n us i n g c r yogenic fuels.

Paper VI i i by Beheim et al. discu ss es t h e r esearch conducted on ai r c oo li n g an ex - haust plug n ozzle. App r oximately 3.5 percent o _ t h e co mP re S sor 9 A l'flow is requir e d .... ; to air cool th e exlla uS t nozzle under aft e rbu r iitn g conditions. ,-' Calcu la tions ha ve been mad e on d irect c oo li n g a pl u g n o zz l e using J e t A fll e l. Ifl this case, the f uel w o u ld be ci r culat e d tllrou _ s ma ll tu bes bra _ , eti to the i n t % r nal surface of the plug n o zzle. T h e temp e rature rise o f t he Jet A fuel is aPl _ i _ o ki fi mtely 560° F us ing t hi s method of cooling. A te mpe imtttre r is e of this mag nitu de c b u | d res ul t in fuel coking , which co ul d create some problems in long-tei'm operation. It 40 2 m ay be f e as ibl e, however , to period i cally re mo ve the carbon that ha s for m ed i n a ma ri ner s i m ilar t o that u s ed in s elf-cleaning ovens. T h e plug nozzle would b e op e r- ated at a temperature of approxi m ately 1200 ° F with air pass i ng through t h e fuel tubes. At th in te m peratur e the carbon would ox i dize and th e tube s wou ld be clea ne d.

Thi s operation Could be performed by run nin g the engine without af terbur ning .

Research was conducted with AiRe s earch Malmfacturi n g Company (contract NAS1-5002) under the sp o ns o rs hi p o f the NA S A Langiey Resear c h Center (ref. 5) t o investigat e t he direct cooling of s urface s of hyper s o ni c cru i se vehicles with me thane i and with hydrogen. T h e results of t hi s in_ es ti gat i o n showed thata conf i gurat i on (see inset i n fig. XII-3) could be successfully us e d f or cooli n g pan e ls with m ethane or h y- drogen a t heat fluxes equ i val e nt to th o se i n the exhaust n o_ ' zle of an alrbreathing o n - . , gin o under aft e rlmr ni n g c onditi o ns. The m ethane t em perature rise for direct Co o l - ; • tng , us in g th e qua n tity o f fuel that w o uld b e b ur ned in t h e afterburner, would be !: 65 0 ° F. T h e resu l tin g fi n al fuel te m perature wo ul d be Sufficiently low that no cokin g : wo uld be expe c t ed . If hydr og en wo re used as a c oo lan t , the _e mperatur e rise w ould be o n ly 310 ° F. There ar e no coking pro bl ems at any te m pe ra ture leve l With hydr o - gen. U n der some co n ditio n s , however , material problems could develop f ro m hy - dr og en diffus i on and e mbrittle m e n t, an d from surface decarburizatio n .

An an alytica l in v e sti ga tion was a lso c o nducted ( r ef. 6) to investi gat e th e f _ a S i- bfifty Of direct cooling starer vane s o f the t urbin e using eithe r hydrogen or me thane as the coolant. This s tator vane Cooling m ethod results in a s ome w ha t co m plex con- .. s truct i on problem in order to keep _ .' . _m o verco o ling th e vane outer _ urfac e s. Int e r- h al coolant tubes were separat ed fro m the vane shell by m eans o _ loW th ermal con- duc ti vity m etal fibers that provided adequa te , lint not e _ Ces s iv e , co n ductioh betW e en the fuel tubes and the oute r shell. The results o t t he i nve _L ig a t i on show that the " c o oling w oul d b e feasible ff the vane did n o t have a lon g thin trailing edge. Th e t ra iling edge regi o n was the only re gio n whic h W as difficult to coo l . Th e proble m i n ,, the trailing edge is in pro vi ding adeq ua t e c oo lant passag e s into the thin cross section.

CRYO GENI C FUE L SYSTEM A research program is un de _ ay t o invesflg af_ t he fttel s y st e n _ for l i quid methane and liquid hydro g en fu e le d e n gines uiid e r realistic operati n g co n ditio n s.

. . The fuel systems are being designed and fabr ic a t e d under coh tr a C t (I _ A S _ J-14 _ 19) _ ; with the General Electric Company. These fu _ l syst em s will t h er i be iflstalled o n aft l exp e ri m ent a l engine and tested i n th e altitude test f a cili t y at th e NA S A Lt _ is Re- search C enter. Fi gu re XII-4 presents a sc h ematic illust _ ttion of the cry og enic fuel I systems t ha t will be inv e stigated. The four h ea vy boxe s in figure _ I _ - 4 illust / _ e th e 1 !

/ / .... ; I co m ponents that ar e simil ar fo r c o n v entiona l fu e l s yste ms and for cryoge n ic f uel s y s t e m s . T he s ecomm on ( i n p rin c i p l e) e lem e n t sin c lu de boos tp um p , heate x c hang- e rs f orthe h ydr a uli c f luid and t h eengin e oil, a nd engine c ombu st or. F o r th ec ryo - g e ni cf u e ls y s tem s ,p arti cu larly a m et ha n es yst emf ora sup e r s oni c cruis e a ir c x_t f t , additim_tl h ea t e x c hang e rs w ouldbe u t ili z ed. A heate x c hang e r w ouldprob a bl y b e us e d for c on t rollin g th e a ir te m p er atu re in th e pa sseng e r cab in . ( A s ec ond ar y heat exc han ge m ed i u m wo uld probably b e u s ed betwee n the me t han e a n d cab in a ir to en- s u re a gains t m eth an e l eaka g e in to the cab in in the event o f a _ eat e x c han ger l eak. ) Two add i t i ona l heat excha n gers would be ut ili zed f or r educ in g the coolin g ai r te m per- atur _ for the turb in e. The s e heat e x c han gers wou l d be simi la r to tho s e s hown i n f i g u re XII-2.

One of the mo s t s i gnif i cant cha nge s i n th e fue l system wo u ld be in the fue l p u mp.

Conventiona l Jet A fuel s ystem s u se a p u mp that bypa s se s exce ss fuel b a ck to th e pump inl e t . S u c h an arran g e m ent i s no t fea sibl e wi th c r y og en i c f u els , b eca us e / i ' parto fthef u el wo uld v ap ori ze an dthep u mp wou ld bec o me va p o rl o cked. _ l , Fo r the e x per i m en ta l f u e l systemt h at isto be u se d in the N ASA in ves tig at ion , !_ a v ar i a b le s peed vane pump wiU b e u sed. S u ch a pu m p ha s b ee n c onst ructed , an d i_ p u mp i n g te s t s w i th li q uid hy drog e n have b ee n sta rte d.The p u mp op e r at io n i ssa ris - _ _ f ac to ry, but not e noug h exper i e n ce has b een ga in ed to e s t ablis h p u m p dura bili ty. !_ Past exper i e n ce in cry og e ni_ f u e l pumps f or f l i ght applicati o n s have b e e n limi ted to l,_ a few h o u r s o f u se . Moa t o f these applicat ions have bee n for space vehicle b oos te rs , and l ong Life h as n ot bee n a req uir e m e n t. F o r a i rcraft u s e, p um p li f e wou l d hav e to _i be in exce s s o f I000hou r s . Iti s e x pected that co nsi dera b le devel o pme n t wi lL b e re- qui red b ef o resuchpump li fe ca nb e ob ta in e d .

Th e fuel c on tr o l system is a l s o m o re comp li cated tha n it i s f o r a Jet A _ u e le d '_ e n gi n e.The c ryog e n ic f u el wou ldbe metered in thev a p ors tate in t o th ec ombu st o r. 1i M ete ring int h is lo ca tion wouldbe r e quir e d toobt a in a d e qu a t e e n g in ere s po ns e tim e. _ ?

T h e f u elc on t ro l m u st alsoc on tr o l thea moun t o_ l i q ui d fuel t _m t i sp ump ed s o the !0.

qu a ntity o ff u e l pump e d m a tche sth e quan t ity o f fue l m ete r e dtotne c omb t t s tor . The f u el s ystem tob e in vest iga te d in theN A S A in ves tiga t ion is s ti l l in t h edes i g n sta g e; ' c ons e qu e ntly , no e xp e rim e n_tl r e su l ts a r e a vail a bl e.

FUEL STORAGE tN THE AIRCRAFT

Saturate t t Methane Fuel

Liqui d me tha n e can be sto r ed i n either the saturated or su b co o |ed s ta te, Thb sat u ra t ed f u el state 1 _ t ha t w hi ch occ u rs wh e n the f u el i s at its boil in g tempera tu i'e.

t

t This i s th e equilibrium s tat e o f li q u id m e tha n e fu e l and is ther e for e the s tat e i n w h i c h th e fuel can b e m ore ea s ily ha ndled. I t wou,d be de s irable if th e f uel could be u se d In thl e sta te i n t h e a ircraf t from th e _tan dp o l n t of ground handling pr o b l emS .

A di s advanta ge of usin g satu r ated fuel i s tha t heat ad di tio n or r educ e d vent pr e ss u r e ( s uch a s o cc urs with i n cr e a se d alti tu de) cau s e s fuel b o iloff.

T h e a m ount of be i lo ff t ha t could o c c ur I s s h ow n I n figur e XII-5. Thi s f i gure i shows boi lo ff a s a fu n ction of altitude for tWo tank vent pre s s _ tro s . If the ta nk were !

ve zJ_ ed to ambient pressure, t h e figure s how s that by th e ti me th e airc ra f t had climbed to 70 000 feet, whi c h wottld b e a likely crui se altitude f, _r a Mach $ ai r craft, the boilo ff would be approximately 1 2 percent o f the initial fuel weight l _ til e tank.

Conventiona l tanks can W ithstand some pr essurizatio n . If these ta nk s c ould b e p ressurized to 4 pounds p e r s quar _ inch above the _ . nlbiert t pressure, the bo t loff at the end of th e climb to V0 000 f eet would b e 8 percent. This boilo ff is st i ll excessive a n d would elimina t e much of the gai n that could be obtained b y swtt _ h ing from con- ventional Jet A fuel to liquid m etha ne .

t The boiloff resul t ing from cl _ ng e in altit u de could he eliminated if t he tanks were designed t o wit _ an in ternal pre ssur e of 1 atmosphere, but botloff could still occ ur due to aerodynamic heatin g . F i gure XII-6 i llustrat e s the ta n k we ights t ha t w o uld be o bta in ed f o r a nu m ber of diffe r e n t ty _S of tanks design s capable of withstanding an i n ternal pres s u r e of 15 poun d s pe _ square inch ga ge (p s i g ). These results were obtained fro m references 7 to 9 f or wi n g and fuselag _ t _ ks. T h e or - d in ate in figure XII-6 iS the tank weight as a fra c tion of t h e fuel w e i j_ ht _ o n i _ in ed J w ithin the tank. The abs c issa is t he tank volUme as a f raction of th e t o tal available volume withi n th e c a vity in the fu s elage o r th e wing wh er e the tank would b e located.

Each data po in t is the result of a design S tudy of a l _ rticu la r c0n f igurat to ll. It is des i rable to llave th e ratio of tank vol um e to available vol ume as high as pos s ibl e : _, since m e tha ne fuel dens i ty i s low. All aval _ ble volume in spaces al lo cated to fuel tanks w ould be requir ed for fuel stor a ge. Th _ r _ s uit s _ hown i h fl _e XII- _ indica t e _ th at wi th reasonable ratio _ of tank w l u m e to available volume th e fuselage t an k weights , i ncludi n g the weight of t h_ insul _ . tto n, would be o _ th e ol'der o i $ percent of the fuel weig _ ft cont a in e d. The wing t _ will b e h _ vie t be c ause tt _e y ha ve a l e_ s _ favorable s urface to volume ra tio. The win g htnk weigh _ can be of the ord er of 5 to 6 perce n t o f th e fue l co n tai n_ d W ithin th _ wing tank S . T h e ir _ tiS n accounts i o r approxi mat ely one- t hi _ 'd o _ t he welg h_ s how n lot" bot h fuse lag e and _ Vl ng t _ ts.

Th e se tank weights are consi s t e f _ t w i th th e va l ues as s t _m ed in _h e n _ ssioh anAly S es that shoW ed th e gai _ .s to be obtained l _ y swi(ch ing to _ etl _ ane fu el from con V e _ ti0na i : Je t A fuel.

There is an advan tag e to d _ si _ ling the _ to _ ithstand _ i'e s sures g reater tli _ m 15 p si g . These g reater pre s_ ui ' es would r educ e f u el _ 0i _ off due to a e t o d y na_ di c heat- i t n g , The r e sults pr esen ted i n r eferen c e s 7 a nd 8 s h o w t ha t f o r S o m e tank c ons truc- tion m e thod s the weight p e na l ty f o r tan ks de s i g ned for 3 0 psl g is only sl ight ly g re a t e r tha n t h o se de s lgn _ l f or tank pr ess ur es o f I S ps l g beca u se th e m i nim u m g _ t ge req u i r e- me at s f or t a n k wa l ls es ta b li s h ed th e We i g h t of th e t _l k s rather than wall s tre sse s .

Su bcooled Metha n e

A liqu id is su b coo l ed ff its temperatur e iS below t he boilin g temperatur e , T he h eat s tn _ is i ncr e a s e d b y using sub c ool e d me thane inst e ad of satur a t e d me tha ne , If _ U th e me tha ne w ere su b e oo l e d a pproxi ma t ely 25° F a t 1 a t m osp he r e press u r e , its vapor pressure would be lower than that which would occur at the high a lt i tude cruise po in t with tan ks Vent e d to 4 p sJ aboVe ambient pressure. As a r e sult, the problem of b oll- off d u e to incr e as in g alti tu de could b e e li m inat e d by uSing subcoole d fuel and only _ i m od e rate tank pressurization. Subcooled fuel i n troduces new problems, however, i Ground h _ nd li ng is _ omp ll cated by the n e e d f or add _ l ref r igeration equip m e n t to keep _i; t h e fuel cooled to a temperature lower than the bo i ling t e mper a ture. A more s er i ous i problem, h owever , i s the reduced vapo r pressure o f the fuel w h en i t i s s u bcooled, i, : " S tor in g the fuel in a tank, w h er e the fuel vapor pressure is less tha n ambie n t at m o- t!

spll er tc p r e s sur e , leads to the r t sl _ of tank collapse unle s s the tank iS pressuri z ed to keep the pressure insi d e th e tank at le ast equal to the pr es sure ex t ernal t o the ta nk.

T h ere is a p robl em _ i n diftg a suitabl e pre Ss uritnt l ot" li quid m etha x te that i s insoluble and no n condenstble at liquid m ethane te m perature. Helium w ou l d be a su i table pres- surant, but ff H q uid me t hane were to be uS ed as a fuel fo r a w orldwid e fleet of s u per- sonic transport S , t he w o r ld' s su pply of heli um m ay be inadequa t e. N i trog eh _ ¢ o _ ld S eem to be a s ui table pr e s s ur ant , but nit rogell goes into s olution i n liquid me t ha n _ i n excess of 11 weight peeee n t at 25° F su be oolin g ( r ef. 10). A p ossible me thod o f usi n g li quid ni trogen a s a pressur _ mt w i ll be p r es en ted late r , v Another proble m in vol _ , ed tn th e use of s U bcooled m ethane i s to g et t he m ethan e i n to the aircraft tanks i n a subcooled s tate. Th e ta nk s Wi ll n ormally b e war m a s t he result o! aerodynamic heating. The se tanks must then be cooled by th e fuel lo aded into the ta nks. If the fue l i S in a subco ot ed sta te wh e n it is lo _ d ed t _ th e ta v d _ S, so m e or all of this s u bcooli n g will be e liminat ed by cooling th e tanks. T hi s probl e m ha s be e n invest i gated und e r a NASA e ont t _ tct (NAS $ -1241 1 ) W ith Martt_ M _ Hetta Corp o_ '- alton in Denv e r. The results of part of t his i nv es tigation are S l _ J _ w n in figure XII-' / .

T hi s f i gure shows the q u afltity of st l bco ole_ t f u el to b _ lo _ ded l _to a tan k (t _ tnk i nitially ' at roo m te m peratttr e ) to obtain a f ue l b u lk t em p e ra _ r e app _ ohching the s u pply t ern - ,, perature. T h e tank us _ in t he e xperi me tits was o n e that r epre s ent e d a wing t an k for a s u per s onic ctmise ve h icl e , incl u ding the i nsu lation system. It c an be seen in fi gu re l / ,I 406 ' X I I -7 that whe n the ta n k bec ame i n itiall y fill e d the a m ou nt of s ub co oli ng of th e bulk fu e l i n t h e t a n k w as on ly o ne -half t he sub e ooling o f the s up pl y f uel, T he f ue l su p- p lied to t he ta nk was c on ti n u ed wi t h t h e ex c ess fue l b eing fl owed ¢ _u t o f t he tan k ve n t a nd in to o t h er ta n k s . F i gu re X I I - 7 s how s t h at som ew ha t in ex c e ss of t Wo v ol u mes !

o f ta nk fuel w e re require d b efore t h e bu l k te mp e ra tu re of t h e fue l a pproached t ha t o f i t h_ sup ply t em perat u re. H avi ng to s upply tWi ce as m uc h fue l t o an ai rcraft a nd re- e i reu l,_ t e i t in o rder t o coo l a ll the f ue l t o t h e requ i re d t em p e ratu r e wo u l d be a pr o b- l em i n gr o und h a ndling t ha t s h o uld b e avo i ded if at a ll poss i b l e.

Gelled andSlushMethane

Possi b l e s o lut i ons to t h e probl em s o f n i tro g en so l ubi li ty in subcoo l ed m ethane a n d t h e t an k coolin g proble m ca n be Obta in ed b y f i rst g e l lin g t he me thane to decrease the convect i on curre n ts a n d t h e n part i a l ly freezi n g the m etha n e to make a mi xture t ha t prov i d e s a l a rg er h eat si n k. Ae r oJet L iquid Roc k et Co m pa n y i s presently i n - vest i gat ing (NA S A co n tract NAS3 - 14305) m ethods o f ge t ti ng li q u i d m etha n e. The r e- suits o b ta i ned to date have i nd i cated that a d d i ng 1 _ we i ght percent of w ater to t h e l i q uid m et h ane i n t h e form of extre m ely f in e i ce crystals results i n _ tge l led so l ut i on.

Tests co n ducted w i th su b co o led gelled m ethane have sh o wn that n i t r ogen solub i lit y i s • al m o stcomplete ly e limin ated u nderbot h s tati c a n d s lo shi ng c onditi o ns f or t i m ep e- ' .

r lo dsc onsi s t e nt with a ir c r af t stor a g e.

• F igu re XII -8 showsa ppr oxim ately 500 c u b i cce n t im eters of g elled m ef hr , n e in a g_s s D ewa r. Itisob se rv e dt ha t g el l e d meth a n e ismilky i n ap p e a ra n ce , whe r eas : no rmal liquid metha n e i sas clear a s un pollute d water. Figur e XI I - 8 s how s g a s b u b - bles s uspe n ded in the g e ll ed metha n e , w hic h in dicate s th er e i s a lac k of m ob i li ty.

Thisla ck o fm o bi li ty un d oub te d ly e xp l ains wh y nit r , J ge n so lu b ilit y isl o w. The ni t r o - ?

g e n p rob a bly stil l go e sin t oso l ution a t t h e m e th a n es u rfac e, bu td ue t o t h emi xing c urr e n t s b e ing e ss e n t i a l l y e li m in ate d int he g elle d m et h a n e , onlyth e surf ace b e- c om e s s a tur a t e d and s olubility do e s n oto cc u r i n th e re s t O fth e met ha n e wi th in th e t i meper iod s e xp ected f or su per s o ni c tra n spo rt fli g h t.

E x per im e nt s h a v e n o t as y e t bee n c on d u c t e d on pa rti ally fre e z ing g el led m e th a n e i toprovid e g e ll e d s l u sh .R w ouldb e e xp ecte d , h o we v e r , t _ ' g e l li ng w i l l s tabl l l_ e the slus h so that a h o m og e n eou s mi x ture w ill occur. In l oad ing s l us h methane in to a _ tank, t h e ta nk will be c ool ed by m e l ti n g t h e s l ush. I t 50 w eig h t p e rce t l t o f t h e _ et h - :/ an e w e r e froze n to for m sl u s h , t h e h eat s i n k o _ the a l v s h W o ul d b e e qui vale n t to a _ o U t _ _ 1 2 ° F of su b co o l ing . Th i s much of a heat s in k i s l i kely to be e n ou g h t o e e . ol t h e ta n ks dur in g f illing. The m ethane wil l t h en be left i n a s ubcooled state, and the lea d ing p ro ble m w i ll be s im plified.

i

_ 407 , !- '

Methane Tank Insulation

Insu latio n sy s t ems f o r m ethan e tat _ are bei n g i n ve s tigated un der NASA con- i tract N AS 3 - 1 2 4 2 5 with the Marti n M _ rtetta C o rporatio n in De z w e r fo r both i n t e rna l !

and e xterr _ t In s ulati on systems. F i gure XII - 9 illu s tra t e s the i nt e rnal in s u la tion _ y s- te rn under inve s ti g at io n, in tiff s s y s te m the in s u lati on i S p la ced insid e the ta nk. A _ : uni que fe atur e o f t his i n su la tio n i s t h at t h ere a r e n o p ressure f o rc es o n t he insu la - v tio n . This e l i m i n at io n of pressure forces resu l ts fro m per f orat i ons i n th e Kapton fi l m cov e ri n g the insu l atio n . These perforat i o n s permit th e m etha n e co n tai ne d in t he ta n k to permeate t h e i nsu la tio n system. Th e m ethane that perm e ates into the i n- terior is q uickly vaporized. The s urf a ce tension of m e thane provides _t li q uid - vap0 r int e rf a ce at t h e fil m perforation. T h e m i ni m u m ther m a l c onductiv i ty of th is type o f i n su la t i on syste m is that of t h e m et ha ne vapor. Tiff s conductivity is l ow enough to be sati sf actory for storage o f li quid m ethane in an aircraft. It is also satis f actory f or s torage of other c ry oge ni cs , such as l i quid hydroge n in Space boosters. I_ The components of the i n s ulation system sh o wn in figure XII-9 include a po l y- 1• i mi de flexcore f ill e d with fiber glasS. Th e ti b e t glass essenti a lly e l im inateS meth- _: ane vapor convec ti o n curre n ts, w h ich would incre as e the heat f lux through the inB u - la tio n . The fib e r gla_ is also opacff i ed to radia ti on by either c oatir, g the fiber g lass !_ f ila men t s with alu mi num o r m ixing small aluminu m particl es in with th e fiber glass, i!

T iff s opac i flcat i on to radia t ion is requ i red to obtain an effec ti v e i nsula ti on sy st em _ when the externa l surfac e is heated to te m p e ratur e s of th e order of 400° t o 700° F by aerodyannflc heating, i ; !1 _ Also shown in fi gu re X I I - 9 is a sc r i _ cloth , whi c h is used as a bo ntiin g aid be - I_ twee n the polyim i de flexcore and t he tank wall.

Th i s internal i ns ula t i on syste m is s ti ll trader develop m ent a n d onl y prel i minary : res u lts hav e be e n obta in ed. Th e results o b tai n ed thus f ar hav e b e en e n co ura g i ng, v and i t appears that th e system ca n be used in th e f u t ure af t e r de v el op me n t i s co m - p l ete.

C a lcula ti ons have been made of the boilo f f t ha t would b _ obtain e d fro m iflSulo . t ed m ethane tank s (r _s . 11 and 12). It is expec te d th a t th e se r edult s w o _ d b e applica - ble to the in su lati on syst e m s h own in flgul'e XII - 9. The bar g r _ tphs in il L, ti r e XII - 10 show m ethane boi lo ff a s a fraet to ri of the htitial f uel w e ight cont ain ed in fu S ela g e _ tanks for various in stl lat io n tlfiek n ess e $ , tahk Ve h t p r essure s , anti i n itial methan e tempera tu res. It is shown th a t with a tank pressuriz e d tO 1 5 p S ia, an in sulati on thickn e ss of t inch will result in a b _ iloff b f approxitn _ Le l y g . 5 percent of th e i ni tia l ' fuel weight. As W ould b ¢ .zp e cted, t h e boiloff is approxi m atel _ inVers _ | y p l_ op o r - _' , tio na l to the in sulation thickness. The figure a ls o shows th a t ff the t a nk could wi _h - 408 p .... m, ;: I b ' s t an d a pre ssu rizatio n of 3 0 p j _ l _ _boiloff could be e ssen t i ally eli m in a t e d w ith 1 in ch of i n s ulat io n.

F| _ re XII- I O als o ll h' S trates th e bo il off encountered w i th two d i fferent am ount s o f subcool ing usin g 1 inch of i ns u la tion and t he t an k a t a vent pr ess u r e of 4 p s i _ re a t- el " th _ n am bient p ress uce. Tht s portio n o f (h e fl _ r e sht _ v s that if 2 5 ° o f ini t i a l s ub- cooling could be ob ta i ne d, th e boll a ff would be less tha n would o e e ur if _ aturated me tha n e were stored in a tardcp r essur i zed to 1 5 p s ia. Th e figu r e shows, however, t hat t h e amount of s ubcoo ling obt a i n ed is v e ry e z ' l tlcal. If on l y g O0 of subceo ling .. , we re obtained, the bo llo f f would be g r eater th an f o r t h e pressur i zed tadks. As i l- lustr a ted in figure XII- 7 , it may be difficult t o ge t th e a mou n t of subcool ing t ha t i S desired. As a result, considerable research is still required t o determine if ade- qu _ tte subcoo li n g could be utilized i n all ' craft tanks.

Figur e XII-10 shows boiloff r e sults t o _ f usel a ge tanks. Th e boiloff for wing tanks is slightly less J .J _ an that for ftm e la g e tanks b e cause the pilo t would probably ' eXpend f u el from the wing tanks f irSt. Th _ s e tanks would therefore b e emptied d ur- _ i ng the first half of the mission. Boil of f s tops when th e tan k s become dry. The total :o bo tloff fo r b oth f uselage and wi n g tanks fo r a typical Mach 9.. _ sup e r s onic transp o rt _, mi s sion am o u nt s to l ess tlwn 9. per c ent of the total fu e l contained in th e aircraft, il Calculations in re f erence 12 indicate t hat if the f li g ht Mach numb e r were increased to 3.5 fo r the identical insulation system, the total fuel boi lo ff would in cr ea se to only ' 9. .25 percen t of t h e initial fu _ l contain ed i n th e aircraft. !i !

i

ME T HANE COMBUSTORS

! , _ ' / Research has been conducted us i ng swirl can co _ ors wi th me thafl _ as _ tfuel _ at NASA L ewis Resea r ch Center. The sw i r l _ combtt _ ors, il l ustr a t ed i n figu r es _ t XII-11 and XII-1 2 , a re _ imilar to th ose discu ssed in paper I V by Grobm an eL a l. arid i n re f erences 13 and 14. Co m bdstion effici _ ney _ J_ 'e pr es ented in fi gu re XII-I $ , | which shows th a t Co m busti on of flcienCie s of a ppr o xima te ly 100 percent c _ mbe o _a J _ e d I a t r e fer'eno s ve lo citi es of th_ c_ 'd e r o f _ 0 to 90 feet per s _ o nd. Re _ r _ nce _ ; eioci ti_ s I_ / of thi s magnitude are consi st ent with low p _ e s_ re loSs e _ i n _ iret _ tft c b n _ bl sto r S , i Co m btmtio n e fficienc i eS _ ceedifl g 100 pe r ce n t, Show n in fi gu r e X]I-I$, r e s e t from unc e rtainties in te m p er ature ri _ e nie _ sttr em ents. ' _ I_ , , T h_ figure also shows t h at the co _ tbusUO _ lefficiency iS in f lu e nced by the te _ p e t _ " _!

atur _ of the m ethane _ h t e r in _ th e eombus to r. Durin _ sUpe r so ni_ crui s e, f _ m etJ _ ie !

will be he , tiedto teflz p e r ature _ U _ h ve e n600° an _ 1 _ 000 F; th er e fore, high conibdstton i efficie n cles can be e xp e c ted . DU _ lrlgth e limited op e ratio g l of _ n _ h _ stai _ ip, idl e , and letdown at subso ni c sp e eds, t h e m etltan _ ente r tnlj th _ co _b u _ tof could be at a //

' I

!: .

l ower t em p e r a ture. F o r t he ne c o nd i tio n s spree p e nl _I t y in co mb u s tio n e ff ic i enc y i s lik e ly to oc c ur, It a pp ea r s , howev e r, t h at co m btmtlon e ffici ency in a p ro pe rly de- si g ned m e t h an e c o m b u stor n e e d no t b e a p rob l em .

F l gt lr e X II - 1 4 s ho w s b l o w ou t l imi ts f o r t wo typ e s o f c o m bu s t o r s u_ i n g J o t A arid m et h a ne a s fu e l s . T h e b lo w o ut l i mi ts ar e no t s ai ia fa ct o r y f or e it h e r c om b u at or , b ut the p r i m ary con clu sion d r a wn f r o m fl_ r e X II-t4 i_ t h at t h_ bl owou t li m its for m ot h- a n o are in fe ri or to those for Jet A. It I B in d ic a t e d, t h ere f o re , that rn oe a reh will h a ve t o conti n ue on obt ain i ng c o mbu at ors w it h a dequa t e b l owo ut L i mi t s fo r" me tlm ne .

T he actual f uel used to obtai n t h e data i n f igu r e s XlI - 13 a n d XlI-1 4 w as natur al ga s .

Natura l gas contains approximately 93 p e r'cent m etl mn e. T he r esu l ts are e xp e ct e d to be th e s a m e as i f pure m et h ane were us e d.

CONCL U S IONS

I t can b e concluded fro m the in v e stigations that there are no insur m ountable ] problems for uti li zing e ither liqu id m ethane or l i quid h ydroge n as fuels f or aircraft i gas turbi ne engi n es. Additional combus t ion research is req u i r e d f or l iq uid m ethane primarily to improv e the blowout li m its at high altitude, i En gi nes ha ve been run at NASA and at a i rcraft engine plants with b oth m e thane and hydro g en as fuels dur i ng t he last 15 years. There i s no question t ha t engi n es !

i can run on t h ese fu e ls . More researc h is requ i red, ho w u vm ., o n the f ue l c ont rol i ¢ i a n d p um ping syst e m _ w h ere the liqu i d m ethane or ftydrog e _ _ .i S pumped i n the L iqu i d ! ; state, vaporiz e d and supe r heated in heat ex d hang e rs, _ nd th e m mete r ed i n the v a por s tate to th e co m b xmto r. CtTog e nl c pu m ps th a t wi ll hi,re ad _ luate l i fe f or l ong tim e _ a f rcraft op e ratio n will undotlbtedly be a de velop m ent pr ,l ble m . ,._ Investigations ha ve not b e e n made of overall ai re d 'aft p ro ble ms resulti n g f r om t using liq ui d m ethane as a fuel. Thi s l ow d e ns i ty , c ry o g enic fu e l will aff e ct ai rcraft si z e , drag, and t he S tr U ctu r al d e sign req ui re men ts to a cco mm od a te th e lar ge te m- peratu r e diff e r e nces betwe en air c raft st r uctu r e adjac en t t o the liq ui d m ethane an d the outsid e aircr a ft surface s w hich a _ e s ubj e ct to ael'o d vnam t c h eating.

M e th an e fuel avai la bility a n d cost ar e thought to b e sati s factor y fo r a ,trcraft us e, !

but r e cent forecasts of possible sho rta ges of fossi l fu _l s ma k e i t ne v ess _ try to i l wes - i!

tt ga te f u rther the avai la bility a i td cost. 1 Th e gr ou n d handling of liquid met h ane fuel at eff orts tVould be f al ' d / ff _ r e nt fro m the p l ' e s _nt ground han dling ff / e tltods us e d i or 3 st A . fu _ l . AS a _ 'ts u lt , the re V _ ouidbe s i ght / l eant cha ng es to t h_ airport fuel B upp l y syste m s , a h d h eal t hy liq u id m etha n e liquefactio n p la nts mig M be require d .

410 : There is n o a a s ur an ee at thi s time th a t liq ui d m ethane fuel will be us ed i n f utur e _ u p er s onic tr a n s p o rt s . It a pp ea r s , h o wever, th at there a re no g reat proble ms in- h ibitin g its u_ e. It i s quite p oss ible th a t the advanta ge a gained due to it s hi g her h e atin _ value , r easo ns.hie o o s t, an d _ re a ter heat s ink m i g ht m l_ k e fut u r e _j uper so ni c tran s p o rt s m ore _ ttr ae tive t h an t hos e us in g c o nvention al J et A fuels.

REFERENCES

1. Welko r , J. R. ; Wes s o n , Ii. R. ; and S lie p cevlch , C. M. : LNG _ pills: To B urn or N ot to Bu r n. Preprint 0 9 -D- 2 3, Distribution Co l ff er o n ce of A. " _ .A. , Oper- atin _ S oction _ P h iladel p hi a, Pa. , May 12 - 15 , 1969. i 2 . Whit l ow, Joh n B. , Jr. ; E t sent) erg , J o seph D. ; and Sh o vlin, Michael D. : P o te n - i tial of L i qutd-Metha n_ Fuel for Mach-3 Co mm erci _ t S u p e r s onic Transpo r t s . _ ' NASA TN D-3471, 19 _6 .

3. Koenlg , Robert W. ; and Kraft , Gerald A. : Influence o f Kt g h-Turb i ne-lnlet- Temperature Engi n e _ in a M e tha n e-Fu e l e d SST When Takeoff Jet Noi s e Li mi ts are Considered. NASA TN D-4965, 1968.

4. Weber, Richard J. : A Review of the Pote n tia l of Liq ui d-Methane Fuel fo r Super- son _ c Transports. Pr e s e n ted at the National Academy of Sciences Cryogenic E n gi n eer ing Co nf ere n ce , Cleveland , Ohio , Aug. 19-2 1 , 1 96e.

5. Waiters, F. M. ; and Buchm a nn , O. A. : Heat Transfer and Flui d FloW Analy Si s i ' of Hydrogen-Cool ed Pan e ls an d Ma ni fold Systems. At R es e arch Mf g . Co. (NASA ' CR-66925) , 1970.

6. Collad _ y, Raymo n d S . : Thermal Fea si bility of U si ng Methane or H ydrogen Fu el for Direct Cooli n g of a First-Sta g e Turbine Stator. NASA TN D- _ 042, 1 9 7 0. t t : 7 . Cha mb ellan, Re n e E. ; L u bomski , J o seph ] _ . ; and Bev e vino, Wi lli am A.: Struc- f rural Feasibility Study o f Pressurized Tanks for Liquid-M e thane Fueled Super- sonic Aircraft. NASA TN D-4 2 95, 1967.

8. C ham bellan, R e ne E. ; a nd Be v eV i no, Wil liam A. : Co m pa r ative Study of Fuse- lags Tanks for Liqu i d-Methane-Fueled $up er sonic Ai r cri d t. N AS A TN D-4[i$ _ / , 1968.

_ 9. Eisenber g , Josep h D. ; and Chamb ellan , R en e _ . : _ e S y s te m _J for a Methane-Fue l ed Supersonic Tr an spo r t. NASA TM X-1591 , 1968.

t

/ Q 10. fftbbard, Ro b e_ t R. ; a n d _ wm _ Alb e_ t, J _ .: On th e S olubilitie s ahd R a t e_ of S olution of Gases in L i quid M e thane. NASA TN D-4V01 , 1 968.

1 1. Pieba n , Eugene J. : A,alytic al He a t T ra n_ lf er Inve s tigati o n o f _ ulated Liqu i d Met t , _fl e Wi ng Tardm for _ up e r son i e Cru ise Aircraft. N ASA TN D -_84 1, 1970.

1 2 . Pleban , l _uge ne J. : Analyt i c a l Heat T ran sf er Inve _ t l gat i o n of Ir _ ulat e d Liquid M e th _ m e Fus e la ge T m l l _ fo r _ uper$ _ ni e Cru iS e Ai r craft. N A SA TN D- 6 1[i7 , 1970.

13. Marc hi onna , Nicholas R. ; and Trout , Al,thur M, : T urboj et Combust0r Perfor - manc e w i th Na tural Gas Fuel. NASA TN D_ 55 71 , 1 9 r / 0.

14. Schultz, D o n ald F, ; Pe e ltlns , Porter J.; an d Wea r_ Jerrold D. : Comp a r iso n of ASTM-A1 and Natural Oas Fu e ls i n an Ann ula r !t urboj e t Combustor. NASA TM X- 5 2700, 19 6 9.

J ti

•

/ i t \ T ABLE XII-I, - COM P A I _ON O F F UEL P R OP ERTIE S " L .. .

" J e t A Pro p ane M e t han e IH ydr ogen L , . . . ,. - : , ........... , .... i . ,' . .", Hyd roge n - c arbon ratio 0 . t e 0 . _ U 0. aa - , tteati ng v al u e , Btu / lb 18 60 0 19 700 _1 1 0 0 4 9 9 0 0 Lim i t i ng u _o to n , p orat ur e , °F 400 a_ o 1 85 0 - He at _t_, B t u / lb T o Um lt ing u se tem p e r a t u re 17 5 7 0 5 182 0 ">7 00 0 To 1000 ° F 7 00 82 0 1085 4 950 Boi ling t e mp er a ture , OF 8 5 0 - 4 4 - 2 59 -4 2 3 De ns i t y , l b / f t 3 5 0.5 36, 2 2 6. (} 4.4 Fl am mab l e limit s , 52 to 400 [_ 1 to 2 8 0 4 6 to 104 14 to 2 5 0 percent Sto t c hl o m etr i c t Relative cost per Btu I. 0 a0, 8 to. I. 0 a0.8 to I. 0 6 , ,, . ,, aEstimates b ased on ava il ab l e inf ormation.

' FUEL PROPERTYTREND S

,ll T !

.. ... _ ,xu s_,_ , H - I___ , ,__ j _ Lj_LL_ , ., .2 .4 0 .2 : 4 _ t B H /C CS-5 6 6 1Z Figu re X II - 1 4 i3 • ' POSSIBLEHETHANE HEAT EXCHANGER INSTALLATION COMPRESSOR DISCHARGE - 1 6 520R / COOLING AIR TEMP E RATURE . 123 6 0R t_- -13 . @, -- _ 1 // -23 L B - _ co _ u . o _ , . , ,, __ :

I V _ 0. 4, , _:_r_ l

C8 - 518 7 8 FigureXII- 2 EXHAUST NOZZLE COOL ING f O G E N IC COOLANT AI R 3o _ CO M P R E SS O R FLOW J E T A 560 o F TEMP R ISE METHANE 6500F T E MP R ISE CS-S6694 HYDRO G E N 310 ° F TEMPRISE z FigureX lI - 3 \ S CHEHATIC OF A CRYOGENIC FUEL 8 - Y - BTEH I FUEL I .__ _ COOLING I " I _ G'NE I

, , "i ICONTROL] [ 1 8 AIRHX ICOMeUST OR

r J VAPOR FUEL THROTT L E . "" COOLING : . AIRHX 1 HEAT EXCHANGER BYPASS CONT R O L HXNO.1 EN V IR O NMENTAL COOLING HXNO.2 HYDRAULIC OIL HXNO,3 ENGINE OIL cs- 5 6 6 o 4 Figure X[ [ - 4 .- HETHANE BOILOFF , . F R OH REDUCED TANK P R ESSU R E _ " . . 10-- TANK PRESSUR '" ]2 -- _ t O 10 20 30 40 5 0 60 70XlO 3 ALTIT UD E, FT c s . 565_ Figur e XII-5 . , _ IN S ULATED METHANE TANK WEIGHTS : - T ANK PRESSURE . 15 PSIG _, . lO F - .

. .08 / WING ,_ , TA N KS i,,

!

I I I I I _.

. 7 5 . SO . SP . 9 0 .9 5 1.00 i' ' TAN _ VOL iT AVAIL VOL c s - _611 t . FigureXI I - 6 _ :

i!

LO A DING 8 UBCOOLEDMETHANE !, ( BOI LO FF A N D CO N DE N S ATI ON VE RYLO W ) i I . _ t : FUEL 5 !

SUBCOOLING , OF lO LKTEMPIN TANK i '" -SUPPLYTEMP =_ - ° _ . I I I I 0 1 2 3 4 _ TA N K VOLUMES OF FUEL , , % , , . o l _ I I I I I 0 10 20 30 40 5 0 T I M E ,M IN c s.s 661 o Figure XI I - I , o ! 416 GELLED METHANE , / Fig u re XII- 8 INTERNAL . IN S ULATION SYSTEH .. r KAPTON FILM,0.001 - 11¢ THI C K , _ iI , / , , - PO L YIMIDE FL EX C O RE , / / _+ • ., ,,- ,, " FIBERG LA S FI LL ED /, / / "'/ , , ,- SCRIM C LO TH Z/" _ Ci_ t " ONEPERF O RATIO N PERCE LL_ / WA LL _ , , __ : _ ' _ ii 1 l t i ; " FIBERGLAS OPACIFIED TO RADIATION / i : WITHA L UMI N UM COATING ORPARTICLES J ' _ " EXPLODED VIEW CROSSSECTIO N CS - 5 6 816 x , Figu reXII-9 e LIOUID HETHANE BOILOFF FOR H=2,7 SST HISS ION ' (FU S ELAG E TANKS ) e04 m i .0 3 - - , 02 - . 0 1I- _ - / WB O _ ........ _ ......

!il

0 i 2.9 1.0 .5 15 3) 25 20 INSUL THICKNESS, IN. VENT PRESS, PSIA DE(; SUBCOOLING , OF 15-P S IA VENT PRE S S , l - IN. INSUL THICK , AMB& 4-PSI ' SAT U RA T ED METHANE SAT U RAT E D M E TH A NE V ENT PRESS c s- 566 os I'IN. INSUL THICK Figu re XII- lO q METHANE SWIRL CA N COMBU S TOR C ROSS SE C TION cs . 5 6 66 6 Flqur eXII-11 418 __ : t_ _' " \ EXPERIMENTAL METHA N E.SWIRL CAN CO M BUS T OR i• Fi gur e XII-2 : = • M ETHANE CO M B U STOR EFFICIENCY i f It . . 110 -- GAS oF IO 0 -- COMBUSTION EF F , 1 • gO _ l 0

8 0 I I I I I I

70 80 gO 100 110 120 1 3 0 REFEREN CE VELOCITY , FT / SEC C_-56 5 9 5 FlgureXII-1 3 , !

419 _ 4 2 0 XIII. A D VA NCE M E N T S I N B E A RI N GS,

SEAL S , AND LU BR I C AN T S

E rw i n V . Zaretsky andLawrence P . Ludwig " / _ 7 _ . .' . ]. 9 4 _ _ _,, q Advances in a irbr e athi n g turbo j e t engi n es have dictated th a t beari n g mate- rials a n d lubri can ts operate at hi g her te m peratu r e s , higher s pe l ls, a nd hi g her-- loads. The first g ene ra tio n sUpe r soni _ tr a nsport (SST) tu rbine en g in e main-shaf t !

bearings will o p erate at a bearing t emperature of 42 B ° F and a maximum speed of ' _ 1 . _ mill io n DN. It i s an ticipat ed that more advanced engine , de s igns may requir e i bearin g s to operate at temp e ra tu re s betwe e n 5000 an d 6 500 F and s p eeds Of a p- proxi m ately 2 millloll DN. P l 'oje _ ti0n of th_ s e trends wou l d pl' edi ct beai.i ng tern- , : ii pera tu res to 6 00 ° F and bearing s pe eds of $ to 4 m illion DN , which woul d pl,oduee I : i: higher bearing operating stres s es ( fig. XlII - 1 ) . i Aircra f t g as tu rbine e n gines have m a n -sha f t s eals t o res t rict g as l e a k age ; .

into t he bearin g _ o m part m ents ( sumps). Various types of s haft s e als ar e us ed ; r, some engi n e s use l abyrinth s eal s e z clu S ively; others use face ahd / or ci r cum- i i .

feren ti al type contact seals. Labyri n th seals have th e di S adV a nt a geS of hi g h leak- _ , : _ - age rates and associat ed debris passage; l en g operating life attd r eliabili ty _ tre ii , t h e chief advantage s as c om pared to rubbi n_ contact iv,- - seal s . The contact seals i , ' " ",' h ave l o w leakage capa b ility but are limited in pr es_ .,,e diff e r e n t ial a n d Speed b e - '_

i '

,_, cause of ru bb ing contact. For convention al con ta ct se al s, curr en t limits a r e ,, se aled pressure differenti al _, ear 125 psi, sl id i n g speed of 350 feet per second, _ , and s eal ed gas te m per a ture of 800 ° F ( ref. I).

Main-sh af t se al c onditions a re shown in figure XIII-2. In 19 r / 0 the require- m ents a re slidi ng speeds of approxi m at e ly 400 feet p e r Second With air pressures.

of a# proxi m ately 250 psi and gas tsm pbra tu res of approx ima t e ly i000 ° F0 It is predict ed th at fu tu re require m en ts will be s lidin g s pe ed s in the range of 500 to , 600 feet per second, a ir p re ssure s o _ 4 0 0 to B 0 0 l _ i, a nd al / " tenlpera to reS near 14 0 0 ° F.

Th e wo _ k reported h erei n is a st / _ tn a l _ y of the roll i ttg- e leme n{ b e l iriit g an d seal state o f th e art with e m p h asis on NASA co nt ributiohs to s olvi n g p ro blems en- countered i n advanc ed airbreathing turbojet en g ines.

42 1

t

ROLL IN "ELEMENT B E ARIN O s

i In orde r t _ _s ure pr o per b ea ring ope r a t i on at e l e vated t em pe ra tur es , a lu b r J_ cant film I s r eq u i red to se par a t e t he c ont a ct in g c o m p on en ts . T h i s is illu s tr a t e d i n f i_ l r e XIH _ a. A bail o r ra i ler is in c ont a ct W ith a s u _ tc e . Th e b _l de fo rms b e n c au s e o i the force which i s p l ac ed o n I t and the e laSti c pr o p e rti es O f th e ma te r ial.

Und er d ynamic c ondition s an d -with t h eInt r od uc tion o f a l ub r lc a nt p a fl lrn , w h i c h i s r efe r r ed t o a 8 au e l a _ tohydrody _am ic (EHD) fi lm (du e t Q the el as t ic de f or ma - tion o f the ma te ri al a nd the hyd r odyn _ t m ic action of the lub rican t) , i _ formed be - tween the two s ud a eo s (rOL 2 ). This f i lm, which i s gen er al l y depe n dent on lu_ brt c aat base stock and vi sc osity, is of th e o r der t) f 6 to 10 m ill i o n ths of _ tnInch t h ick a t e le vat ed te m p e ratur es (r o t. $ ).

A commonly accepted minimum tolerable hardne s s f o r bearing co m ponehts _ .

i S R ockwell C 5ft. At a h a rdn e s s b e low this valu e , brin _ tlin g of the bo ar h _ g races _!i c an occur. Since hardness d e creases with te m p e ratu r e, conv e nt _ nn _ l b ea r i ng i I m at er i als such aS SAI_ 5 2100 c an be used only to te m peratu r e s o f about 3500 F. !i Much effort has _ one turn dev e lopi ng steel alloy s s uita b l e { o r nigher te m pe r a tur e s . _ ; The addition o! ele m ents such as molybdenu m , tun gsten , c h ro m iu m_ and van a - ii!

diu m pro m otes the retention of ha rdn eSs at e levat ed t em perat u re s . Other mate- _ rials which p ro m ote l_ a r due sS re ten tio n are al_ t m i n u m and silicon (re f . 4). !_ !4 ' _ ev e ral bearing alloys ar e available f or operation be tw een 3 5 0° and '1 50 ° F. i_ : The most common o f these is AI$I M-50. Ill most tu rbojet applications, this m a- I_ te r ia[ ts us ed exclusivel y . Th e m aterial has the cap a bility of m ai n t a i n i n g a h _ trd- i_i h eSs a bov _ Rocl _ we ll C 5B to t em pe r a _ r es of approxi ma tely 6000 F (ref. 4). !_ _e tainers (cag eS ) are a much more severe p rob le m in s mall-bore beari _ gs i! ; " than in lar _ e r ones. In extr em e- s p eed a pplications of nmal l- bo re b _ ari ng s, it i s :!_ f r eq u en tly n ec essa ry to us e a s ilver-pl a ted , se m ihard, t ool steel r e tai n er r a ther _ ; _ than a bronze an d oil-mist l u brication rat he r than recircu l_ tti n g oil to reduce churai t tg l o _S e s . In large-bor e bea ri n gs, retainer f ailure is mu ch less common !

than i n smal l - bor _ bearing s (re{. 4).

I

HIGH-TEMPERATURE LUBRICANT SELECTION

G r o ups of _ / _ 05-size ( _ 5-m m t _ o e e) an gul a r-contact bill[ be _ 'i n_ s made from c o ns u m able- el ectrode va cu um - mel ted AIS _ M-1 steel wer _ tested with the 1 1 l u- b r i can ts (table X _- I) in a low-oxygen e nvironff _ ent (less that _ 1 pe r c e nt b y eel - ti m e). These lubrican ts v ar i ed with respect to baS e stock, additi v e content, o r i \ ,) TABLE XI H = I , - TEST LUBRICANT PROPERT I ES Lubr i c an t Lu br i c an t Ba _o _ t o c k Add i ti ve R e pr esen tat iv e t y pe d e _t gnat ia n _on te n t vl _ co ai t y, c8_ 1 00 ° F 210 °F b6 00°F Mo difi ed A Bl o nd o f thr oe - Nat a Vail _ 9 0 4 , _ 0, 8 _ p o l y p hon yl ri n_ _ m dfou r o a h[o o t he r ri ng c om p o . , n on t _ B BlO n d of t h r oe . N o t av all _ 8 8 6, 9 ,8 _ _ t , ring and f our- ab |o r i ng com po - i .o nt o ii Po ly p ho n y l C 5P 4E N o ne $ fl8 1 3, I 1, 0 7 et h er D 6P4 _ ( e ) $ 8 5 13.1 l, 07 E 6Pg E N one 1831 24. _ 1 . 30 i Est0r F M ix e d p oly e ster- No t a v ai l - 40 8 .4 1. ?0 diester able G J Diester (c), ( d), (e) Y/ 7.8 1.80 Hydroca rbon .H [ Synt _ letic None ,31 4 33 3.9 ' p a_ aff i_I c ol l I Synt h etic (d) 314 32 2 . 9 • p a r a ff ini c oil J Superrefl n ed ( e ) , ( d) , (e) _ 9 8.4 1.1 1 t n a phthentc i minera l oil . [ . , K Sup e rr e fined N one 48 0 2 8 2. l 11,.

p a r a ff i n i c , , m i ne ral o i l " aMa nufacturers' data. , bEst i mate d .

Cox i dati o n inhibitor.

d A ntiwear additive.

e A nt t foam age n t.

i .

/ • vL _ ce 6 ity a t r oom t em p e r atu r e. T es t c ond itio na l inclu ded s p e ed _ f r o m 20 000 to 4{} 0 O 0 rp m a nd b ea r i n _ th r u_ t loa ds from iB 0 t o 9[8 poun d e , which pr od uc ed i / m axi mum Hertz e t r e eses r _ m g i n g from 1 69000t o 8 4 7000p _ ion thebearin g in ne r ,, ,_ t r e e , Th e te a t e ond l tion _ included outer-r _ tc e t em p e r a ture s o f 400° t o dO 0 ° F '_ ( ref , 6). " i, _ y p e _ o f h i gh- t e m p e r a tu r e fai lur e m o de m a re d is c ussed i n r ef e r ence _ 0 a nd r / , Tl;e _e moC l es c a n l _ e e a t _ l _orized m _ f a tigu e pit ti ng, s ur f ace gl az inl_ a nd p i ttin g , and Sur l _tc v sm e a rin g o r dv fo rm a tl e n, Th e fat i _ fai lu r e s r epo r t ed in thia pa p e r we r e u_oc i _ tted with rd o re s urfac e d i e J_r e _ l (t _ lazin _ and s uperficia l _ . , pitt l [l[ _ )t h an c la s s i cal fa t i gu e sp _ il s nor mal ly exp e rienc e dun d e r m ore conven- tion al t e m per a ture en vtroiz m ent S . An ex_t m pio of thi s f a tigu e fa ilur e i s given i, _ flT _ .tr o XIII-4. A normal a pp ea rin g race ru n under good e lasto h ydrodyn am ic (E H D ) cond i tion8 is Show n i n fi gu re XIII- B ( a ) fo r c o mparison. _"_ Sttrf a co g la zin g, which c a n b e present wit h o u t s p all in g occurring , iS ii l u- ._+ strated in figure XIII-6(b). Cont h med op e ratio n results in sup e H l ci al pitting , _ ' _ shown in figure XIII- B ( c ) and / or wear of the rolli ng -ele m en t surfaces. This con- __ dttion was pr e s e nt in all b earin g s run with t h e va ri ous lu b ricants except lubri- i _

i /

c ant I. The gl azin{ _ was m ore s e v e r e with the bearings t e a t ed wi th the polyp h enyl !

ethers, which were short lived. If operation of the beari n g is continued under _ the condition shown In figure Xlll-6 (c) , spallln g of th_ r olling-el em ent su rfac e i will occur as illustrated i n figure XllI-4. Reference s 6 and _ attribut e th e glaz- !_ la g phlmo me non to m arginal el as tohydrodyna m ic fil m thickness e s. Urtder m ar- _ ' gi nal elaStrohydrodl n m m ic lubrication , hi g h tari g enti al forces c an be i n duced ! L _ which will t e n d to relo ca t _ the m axi m u m s h e arin g stresses clo s er to th e surfac e (ref. 8). Under these conditions, more _ h al low f a tigde Sp al ls wo u ld 1 0 eeXpe c ted ' tha n nor m ally o _ cu r u nd er complete e l as tohydrodyna m ic conditions.

It is int e restin g to note that , With l u b r icant I ( the synt he tic para _ ftnic off with t h e antiwear additive), no glazin g , superfici al pittin g, f a tig _ t e spl _ li n g, o r wear occu r r _ l fo r all 10 bearin g s testt _ i. Each o _ thes e bearings was r u n fo r a t l e ast 180 _ hours. For the sa m_ l u bricant without the ad d itiv _ ( lu bric _ t H) u nder the s a m e conditions (i. e., temperatures betwe e n 5{ _ 00at _ d6000 ' F) _ o m e g lazin g was observed. It i s conceivabl e that an tiwear addi t i v es will r ed uce the t ang e _ ti al forces un der m a rg i n al e l u_ ohyd r odynamic conditions an{Ithu s i n c re ase beax _ g life.

Another mode o _ f ai l u re at el e vated temperat u re i s t hat o f " s mea ri t _g ." This mode i s ill us trat e d i n figu r e X11I _ (d). S m eari ng m an ifests i _e lf by gross me tal _ , trans fer , defori _a tto n , and / oi - g alling of t h_ r olli n_ -el em ent S urface S . It t d be- li eved that sm e a rin g occt / rs when the co n dition is p _ itflarily that of bo u nda / 'y lu- 4 ] [4 I b rication acc om panied by a high degr e e of s liding i n the bai l -rac e co n tac t . T h is m ode o f f a i lure wa s dominant f or those bear i ngs which failed ot h er tha n b y f at i gue s p ai ling.

Cage wear can be another failure mod e in high-te m perat u re bearing opera- tio n . Bas ed upon work report ed i n refer en ces 9 and 10 , cages were m anufa c tured f ro m AISI M-1 m ateria l having a n ominal hardness of Rockw e ll C 58. To further assure m i nim a l wear , the c ag es were silver plat ed . I n all bearing tes ts in this i nvest i gat i o n , cage wear was not a m ode o f fa i l u re.

Lubricant ox i datio n ca n be an o ther mode of f ai l ur e (ref. 11). Whe n a l ub ri - cant starts to fail through oxidation , the result is usual l y form a tion of bo th sol- .. uble an d insoluble co m pounds that m ay appear as resins, S ludges, or acidic co m - pou n ds. The resulting effects are th e followi n g: (1) A gr _d u ai ri s e in v is cosity of the lubric a nt (2) The formation of _d herent S urface d e po sits that may interf e re in s m _l l _ clearance s paces in t he h ea r i ng (3) Corros i on o r deterioratio n o f the m etal parts (4) General dir ty in g of the system by S ludges or other insoluble co m pounds Lubricant v is cosity changes were not a problem i n the t ests re po rted in this paper inas m uch as fresh h:br t ca n t w as peri o dically added to th e t es t rig to re- ple _ . i_h th at lost because of eva po ration or le s k s ge an d because of the low-oxy gen env iro n m ent i n most of the t e sts. Consequently , bulk viscosity change w as n ot appar en t for the lub ri c an tS t es t ed . I n addition, corrosion or deterioration of the : bea rin g elemen ts did n o t occur or man if es t itsel f with the va ri ous te _ t lub r ic an ts.

This ma y ha ve been due to the rela t ively short testing ti m es .

Formation of sludge appear ed to be a minor problem with lub ri cant C, the 5P4E polyphenyl ether. However , th is problem could be r ed uced eithe r by th e i n sert i on of a copper s cree n in th _ lub ri cant test sump or by th e addition of an _ ?

oxidation inhibitor as in lubric an t D. The ex p erie n ce wi th the o th er polyph en yl _ : eth er l u brica n t s w as si m ilar. Slu d g ln gwas n otincre ased by e xp osi n g l ub ri Ca n t D ii t o a n air enviro n me n t a t elevat ed te m pe ra ttire. N o de po sit form a ti ons app e ared on the beari n gs with th e po lyph en yl e th ers, i_ The es t ers had a te nd e n cy to f o rm hard c o ke depo s i ts on th e b e a rings a t 500 ° F u n der the l o w- o xyge n en vi r o n m ent. The hydrocar bo ns exhibited no n o tice- ably hard coke formations under the l ow-o xy gen e n vlro tl fnent , a R liou g h the fluid b eca m e bl a cken ed a n d so m e s] u d go d e p osits w ere e v id e n tal t erexte n dedrun n ing at 600 ° F.

Based upon the failure mod e s discuss ed i n th is pap e f , t h e lttb ri cant tetnper - a tu re, stre s s, a nd l ife l i m i tS w ere es ti m at ed and a l 'e summarized in table XIII -2 .

d T A B L E X II I-2 . - EST I MA T E OF L UBR I CA N T TE M PERATUR E , ST R ESS, A N D L I FE L IMI TS BASED ON OP ERAT ION O F 7 20 5 - S I ZE A N GU L AR - CO N TACT B E ARINGS IN A L O W -O XY G E N ENV IR ONM E N T Lubricant Lubri- Estimated Es ti m ated Esti m at ed App a rent minimum ty p e ca n t maxim u m m axi mum l l f_ po- viscos i ty f o r des i gna- bearing Hertz tent ia l at avoidance of ball- tion operating stress, m axi m u m race contact sur- temper- p s i temper- face damage, ature , ature and cS OF s treSs , ( a) i P o ly p he n yl A >500 > 233 x 106 Fair 0.7 to : l. 0 ether B > 5 5 0 >_ 0 Fair C > 600 >302 Poor D > 600 >_ 50 Fai r b E > 600 > 233 Fair iV Ester F >500 >233 × 106 Good 1.5 to 1.7 G >500 >250 Go od 1.5 to 1.7 Hydro- H > 550 > 250 x 106 Good 0.9 to 2.5 carbon I E >250 Excellent J >268 Good K >25 0 Go od i ill aCo m pari.sou based on per f or ma nce o f lubricant I.

b Ai r e n v i ronment. , ' The esti m at ed bearing life potential was based upon the app e arance a nd perfor m - ance with th e lubricant re l ative to the appearance a n d pe rf or m ance of those bear - ings run with lub ri cant I, the snythetic paraffi n ic oii wi th the antiwear additive. .

Damage to the be a ri ng s in th e form o f surface distresfs and p _ ttiu g o f the race surfaces occurred i n relatively short ti m e periods with the polyph en yl ether lu - ' bricants. Consequently , th e life poten t i a l with this family of lub ri ca n ts was e _ ,tt- mated to be only fair relative to the pe rf ormance O f lub ri ca n t I under the respec- t I l t i re te m p e ra t ure a n d s tr e ss conditio n s listed. H o we ve r, the beari n gs o perated with lubricant C , the 5P4E polyphenyl ether wlthout an oxidatio n inhibitor , eX- hi bit ed a p oo r life potential in the l o w-oxyg e n e n vironme n t even a t low e r stresses than that l is te d . The e ster and hydrocarb o n ba s ed lubricants s howed good life pot en t i al rel a t i ve to lubr i ca n t I w i th mi n i m al s urf a ce _ts tr 0 ss at t em pe ra tu r e s of the o rder of 5 00 ° and 550 ° F, r especti v el y , u n der th e low - o x yge n e O vt r o nm e n t.

BEARING LIFEATELEVATED T EM P ERATURE , , Te sts were c ondu c t e d wit h AB EC- 5 gra d e , s p ll t- inu e r -race 1 2 0 - m i l li m ete r - bor e an g ul a r -c ont ac t ball b ea rings ha ving a nomin a l c ont act a n g l e o f20 °. T h e i nn e r a nd out e rr ace sw e r e m a nuf act ur e d f rom on e h ea to fc onsumab le-elec trode v ac uu m- m e lt e d(CVM ) A ISI M - 50 st ee l , a n d t he b a lls we r e m a nu fac tur e d- l ro m a s ec ond h e a t. T he no m in al R oc kw e llC h a rdn e sso f t h eballs a ndr ace s wa s 8 3 a t r oo m t empera tu re. E ach b earing c on tai n e d 1 5 ba l l s of13 / 16 - i nc h diameter.The , .

ca g ewa s a on e-p i ece ou te r-l a nd-r i dlng ty pe m a de o fa n ic k e l - base al lo y (A MS 489 2 ) ha ving a no m in a l R o c k we llC h a r d n es so f33 . T h e re t ai n e d aus tenit e c on - I tent o f the ba l l a n d race m ater i a l wa s m ea su r ed at le ss tha n 3 perce n t. The I i nn er- and outer-race cu r va tu res were 5 4 and 52 percent , re s pecti v ely. All com p on e nts w it htheexcep tion of t heca g ewere ma tche d w it hin _0 . _ R o c kw ellC point. T his ma t ch ingas s ur e d a nomin al di ffe r e nti a l h a r d n e ss i nall be ar ings (i. e . , th e b_ ll h a rdn es s m inu sthe r aceha rdn e ss , co mmonly cal ledA H) of zero ....

(r ef s. 3 , 12 , a nd1 3) .

T e st c ondition s we r e a thrust lo ad o f580 0 p o u nd s , which pro d u c ed a m a x i- i m u m Hertz s t ress on the i nn er race o f 323 000 psi , an o uter - race te m perature of 425° F, a n d a spe ed of 1 2 000 rpm W i th a n adva n ced ester l ubrica n t a n d a s yn - r thetic para _ f ini c oi l i n a n a i r environment. The res ults o f th6Se t es ts are S how n in fi gu re XIII-6. The e s ter flu i d produc ed l ive s appr0x im a t ely 6 times the Anti- Friction Bearin gM a n ufac tu rers Ass o ciati on (AFB MA ) pr ed icte d (catal o g) life, i w h i le t hes ynth et i c paraffl nl c oi l p rod uce d l lv e _ o verI0 ti m esthe A FB MA ll ves.

Curr e nt - s tate -o f-t h e-a rt b ea ring s a nd lubr ica nt_ a t 35 0 ° F in f l yi n g ai rcr aft pro ,. ..... 1 _ d u c ed li v es approximately 5 t i mes the AFBM A pr ed icti on .

A t te m perature s above 4 2 5° F the viscosity of an es te r flu i d i s such that it is questi o nab l e whether the fluid ca n p r oduce an adequate EHD film. On the other ba n d, the vi s co s ity of a Synth et i c paraff tnt c lubr i cant is ad equat e t o support a n EHD fil m b u t the l ub r i cant o xi d iz es r api dl y. As a re s ult , at te m pera tur e s mu ch ii a bo ve 425° F, a relati v ely l ow - ox y g e n e n v i ro nm e n t m us t b e prov i ded.

!

427 !

• i

, l - 1 Be aring t es ts were c o ndu c t e d at outer-race te m p era ture s o f 400° , _ 00°, an d 6 00° F w i th t h e s ynthetic pa r a ffin i c fluid und e r a low -o xy g en envi r o nm a n t (less than O,1 per c ent Oxy g en by volume). There We re no s t a t is t i c al d iff ere n ce s be- twee n these l ow- _xy gen environment tests no r a n y s tatist i cal differenc e between th o se tests ru n in air with the sa m e fluid. As a result deratl n g i n bearing life is n ot n ec essary at te m pera tu re s betwe en 400° and 600° F, n o r whe n r un ni ng u n de _ a l ow-oxyge n enviro nm e n t.

Typica l fatigue spa Us occurring o n th e bails of a bearing ru n with the sy n - th e tic paraff . intc o il can be seen in figure XIII-7. Met a l l urgic a l examination of the bear i n g s in d i cated that failure was by clas s ica l rolling-element fat i gue. The f at i g u e spa l l S were o f Subsur f ace o ri g i n , tnl t i a tln g i n thez on eofr e solve d m a xi - mum shea r i ng s tresses. A n u n faile d bear in g ru nt o s u spe n s ion (500hr o f op e r a- tion) is show n i n figure Xl I I-8. _ .

Testswe r e c on duc t e d w i th t h e_ P4E p o l y phe n yl etherata n ou te r -race t em - p era tur e of600 ° F in a n a ir e nvironmen t .At a maximum Her t zs tr e ss of 323 0 0 0p si , mix edlubr i cati on e x i s ted w i thb oun dar y lubrica tion be in gthepre- domtngn t m o d e . As a result , the th rust lo ad was l o wered t o 4865 pou n ds, w hi c h i pr odu ce d a m a x i mum He rt zstre s s on thei n nerrace o f 295 0 00 pS i. A serie so f i 9.6 b eari n gs was test ed i n a n a ir en viro nm ent. Despite this apparently less hos- 1!

tile en vir onm e nt and the r edu ced l oad, most o f the t es t s w i th _l_ e p olyphe nyl i , _ e ther had to be s uspe n d ed becaus e o f a lar g e amou n t o f bal l wear. I n tests ru n - _ ' u l n g from 2 to 6 5 hourS, th e average r ed uc t ion i n ball diame te r w as app r oxl - f , , ' m at e ly 0. 001 inch. A ft e r a re latively sh o rt runn i n g time , a stabl e s u spe n si on _ o f wear partic l e s m anifest e d itSelf in the polyphe n y l e th er fluid. Thes _ particles il can act aS au abras i v e w hich m ay accelerate the Wear pr o cesses. H ow ever, if • ix 10 bea ring s ra n for over 450 h ou rs (325) < 1 0 6 tun er-race rev o lutions) w i th mini- it I , ?

rea l wear. The bal l d i ameter s on these be a ri ngs were reduced b y not m o re tha n I!

0.0003 i n ch. On all th e bea ring S tested , gl az in g of the con ta ctin g surfaces was prese n t. However, with the lo n g -l ived bearin gs , n o mi crop i tt ing o 5 the surface acco m panied the g laz ing .

Of th e 2 6 beari ngs te s t ed , o n ly 2 f ailed by fatigue. This is a n i ns ufficie n t nu m ber of failures to per m it an accurate life esti m ate. However , a rough eSti- mate of life w ith th is fluid w as m ade o n the basi s of th e fat igue data. The l ife estimate is p r ese nte d i n fi gu re XIII-9 a n d is co mp ar ed wi th th e l ives obtai ned wi th sy nth etic parafflnic oil u nder simi l ar te s t c o ndi t io n s.

ti With the polyphe n yl e th er, oil c o nsu m pti o n was quite high. It w as e stt m at ed i_ that at least 30 perc eh t of the oil w as l ost by evaporatio n duri n g a 2 4-hour op e r- 4 2 8 t \ ati ng p. _ i ed , Because of this e v a poratio n , fluid had to be p e riodi c ally added to _: , • t he s u mp , - Tests were perfor m ed with a-fluoro c arbo n fluid w h ich was not : n v es ti g ated with th e 25- _ nilli me ter beari ngs p re viously discussed. Pr e li m in a ry tests wi th this fluid at d00 ° F and a thrust-load of 5800 pound s under a low-oxy g en e nviro n - m e nt .(l e ss than 0.1 pe r ce nt by volu m e) p r oduced con s idera b le ball wear and / or s urface distr ess . These tests i n dicat ed that , a t a m axi m u m Hertz stress of 323 000 psi, m ix ed l u bricatio n ex is ted with bou n dary lubrication being the pre- domin a nt m od e . As wi t h the p o lyphenyl ether the thrust lo a d was lower ed to 4365 pou n ds, which produced a maximum Hertz streSS o n th e in n ez' race of 2 95 000 psi. The fatig u e resul ts with th is lubricant are su mm a r ized, in fig- ure XIII-.9 a n d ar e compared with th ose for the o the r tWo fluid s te sted .

Wh __o th e fatigu e life wi th the fluorocarbon was approxi m ately th r e e times that predicted b y th e AFBMA, the pe r for ma nce of th e fluid w as n ot eou sts teut.

The predominant l u b ri ca ti on mod e r an ged fro rt bo undary to el as tohydr od yua m ic at th e m axi m unl H e rtz s tress-of 295 000 ps i . Sortie ball wear was appa ren t, as _" e vi denced by a decr e as e in bail diameter of 0.005 to 0.00 2 inch within 500 hours o f operation. However, th ere were tests which h _ d to be aborted b _ c a u S e o f rath e r extens i v e wear. Co nv erS e ly, there were Some testS _ which were termi- nat ed at b OOhours, ex hi b i ti ng eXtre m ely g ood surfaces wi th n o evid en ce of incip- i en t f atigue fai l ure or m eaSurable wear. Iris sp e culat ed that this fluid m ay b e _ t very s ensitive to m in ut e va ri a ti ons i n be ar i n g geo m e t ry and / or Very minor ':_ changes in th e tes t cond i tions. ;i B as ed upon its che m ical m akeup, th e fl u orocarbo n fl u id haS th e g rea t e s t po- t en ti al as a n ex tremely high-te m pera _l ' e lubricant. It eXhi bits a n u m b e r o f p ro p- e rt ies which mak e it extre m ely at tr active to e n gine de s igne rS , cll te_ a m ong which _ • iS i ts inherent fire- S afe oper a tio n . Conversely , th e corrosive as pect mi g ht tend •_ to m ake it diffi cu lt to apply in eX istin g lubricating s yste mS , al th o u gh this sho u ld / '_ not be a major probl em in a dvanced en gines , wh e re the lubric a ting s y s te m c an be ti • designed wi th m ateri al s which will acco mmod ate th is undesirabl e chara c t er isti c . _ii However, _ s noted pre vi ously, the fluid d id n ot eydlibtt the s e vere _ o r rosi v e el- I _ fee ls which had b e en anticip at ed a n d, in fact , th e ov e rall d a mage to ei th er t h e !i test facilities or the test bearings themselves w as within acc e ptabl e lifni ts . _ The most se ri ous drawbacks of th e fluo ro ca r bon fluid appear to be its high I _ d l msity and i ts low th ez ' m al conductivity. As a result, th e fl u o roc ar bo n- i_ lub ri cat ed beari n gs, for identical oper a ti n g c on ditio:m, tend ed to stabilize at a il te m perature app ro xi m ately 100 ° to 1 2 5 ° F above that observ ed foi" th e b _ .a rin gs i lubricated with the other fluids. Additional lubricant cooling was pi . o vi ded to !

i.

o vercome this probl em . H e ace , in ord e r to ac commo d a te t h i s fl u id i n a / zfr b iu e eng i ne , it will be nec es sary to reevaluate the e ngine coo l ing system and / or to m ake chang e s in the baS i c bearing de s ign to decr e a s e heat g ener a tio n .

BEA RING MA TERIA L A N D SE L E C TIO N

Bearing R etainers (Cages )

At te m perature s above 500 ° F , tests have indicated that bearing cage wear can be a l im iting f actor in the oper a tion o f bearin gs under the severe lubrication conditions encounter ed at th es e e l evated te m peratures. Therefore, in addition to the race and th e rol U ng- e le m ent m aterial, careful con si deratio n m ust be given to t h e choice of cage (r e tainer) m aterial.

In conventional rollin g -ele m ent be a rings , be th m etallic and non m et al li o cageS have found widespread u s e. Under nor m al te m pera tu res, for nonaerosp a ce applicationS, practic al ly al l th e roller be arin gs an d a larg e pe rcentage of th e b _ l .....

bearin gs in us e ha ve b ee n equipp ed either with S ta m ped cag es of low-carbo n Steel or with m achi n ed cages of i ro n-sili C on bronze Or lead brass. PreciSion be a ri n g s , such as those used for aerospace application S , ar e usually eq u i pped with c ages m achined f ro m copp e r alloy s or uounietalli c phenolic m a t erials. Iu some a ppli- cations, where m argi nal lubrica ti on exi s ts du _ iu g operatiou, such aS at high te m perature s , silver pla ti n g on the b ro nze has been used. Phenolic m a t eri al s are li m ited to te m pe r a tur es of app r oxi m ately 2 5 0 ° F, while cop per -bl t s e al loys a x e suitable for operation to a pp ro xi m ately 600 ° F. Abov e 600 ° F, some s U c- c es s has been o btained wi th l ow -carbo n steel or c as t-i ro tl c ag es, but, g _ mer ai ly, the most Successful high-tempera tu re c ag e s have been nickel-b as e al loy S . One t of the nick e l-base al loys used is AMS-4892 (ro t . 11).

O th er m ateri al s which have shown promise are high-te m p eratu re pl aS tic s which exhibit l ow friction and wear char a cteris ti cs, high-al l oy steels capable of m aintaining th eir hot ha rdness a t e levated t e m pera tu re s , and stainleSS S t e el s .

Res ults of 30-minute tests at a t eS t te m pe r a tu re of 500 ° F with a naphth bnl c m iner al o i l as the lu b r i c an t for _ x m aterials are Sh own in figure XIII-10. At th i s set of co n d i ti o ns, four ma terial s show p _ o m i _e of oper a ting for extended period s of time: M-l, AMS-4892 _ m odifi ed 440C st a inless steel, a c id a basi _ _ olyi m ide poly m e r . Tests were al so conduct ed at 700 ° F with the same mate- _ r i als but with the 5 P 4E pol y phenyl ethe _ as the lubric an t. The results of these _: tests are shown in flgureXIII-11. At this condition , the M-I , AMS-4892 , and 43O - r --- -, .....

............ , q , ! tt ..........

440C stainless steel mat erials s how th e g r e atest p _ romi se . Th e high w _ ar with the polyi m id e i s not te t al ly unexpected inasmu c h as th e rmal degr a d a tion of th i s poly- mer b e gi ns a t V00 ° F (r ef . 1 4). Th e additio n o f 1 5 p e r ce nt by weight gr aphi t e to th e basic poly imide poly mer had n o e f fe ct on th e wear r es ults a t 500° F.

In both th e 500° and the 700° F t e St s_ th e M-1 and A _ -489 2 m at e r i als ex.

hibited the l e ast a m ou n t of w e ar r e lative to th e other m a t er i als . It _ thus , c an be concluded that fo r e levat ed -t em p e ratur e b ea rin g applic a tio n s IVi-1and A _ -489 2 - o f Rockwell A hardn e s s e s 81 and e' /, respectiv e ly, hav e po t en ti al l _ea rin g ca _e ap p l i c ati o n .

Additi o n al te sts irtdicate that f o r the M-1 and AMS-4t _ 9 2 mat e rial S , cage wear decre as e s with i ncreasin g hardnes S . Howev e r, for sm al l dif fe renceS in hardness of th e m odifi ed 440C s t Jtinl ess steel mate r i al s, th er e w as no sign i ficant diff e re n ce in w ear. Op e rating te m per atu re wil l , of cours e , decreas e materi al hot hardnesS. The effect of incr eas ing tem per a tu r e fro m 500° t o 7000 F with M - 1 and A1VIS-4892cage m aterials run with th e uaph th entc m ineral off as the lu- bricant sh o ws t ha t Wear wil l incr e as e tW o to _t ve t i mes, dependin g on the mate- ri ai an d i ts heat treatmet _ t. H ow ev e r, the l _ ockwell A ha rdness for M-1 a nd AlV a -4892 decreases approxi m at e ly 1 and 0.5 po in t, res pe ctiv e ly, because of the increase in te m pera tu re from 5 0 0° to 700° F. The se dif f eren ces i n hard n eS S a re not sufficient to accou n t for the m arked increases i a wear. It c an the refor e be conclu ded that, with th es e m aterials, te mp e ra tu re affec ts the amount of wear throu gh its effect on the lubri c at i o n proces S .

From th es e results it can al s o be conc l ud ed tha t , iu g e n eral , po tet _ ti al high- te m perature c a ge ma terials of the ty pes re po rted h e re i n should be hea t- t r eat _ [ to their m ax i mu m roo m -temperatu _ 'e hard n e sS, while suffi c ient duc t il ity to pre - vent crack in g is m aintain ed . In applic ati on, however, the rolling-eleme n t mate- ri al should b e s o m ewhat harder than the cage mate ri al to prevent damage to the t rolli ng ele m ents. Wear-re s istant pla t ings or c o atings nlay also be u s ed _ .or e - duce ca g e wear.

Bearing Steels

Three bearing material s were inves ti gat ed a t 600° P (ref. 15). T h ese were AISI M -50, WB - 49, a n d A f SI M - 1. A _ I M- § 0 iS a ma rten$itic high - speed tool steel which h as been used in critic al bearing appltcatto v _ fo r the past d e c a de.

The steel was developed p rim arily for rise as a h igh - str en gth, h ighl y Wear- resistant tool steel. It h as i nheren t ly high ha r dnes s a n d good compr e geive 4 51 J s tr e n g th. ItS o perati on al t e mp e rature capability i s in o xc es_ of 000 ° F. ; , T he material a s pr e s e ntl y avail a b le i s produ ce d by t he c o n sum a b l e -el e ctr o d e vacuu m -re m elting proc ess , utilizin g e ith e r an a i r_m elt e d or a n ind uc ti o n-vacuu m_ me lted e l e c trode. Th e M- B O i s c h a r a ct e rized b y a fin e - g ra i n e d m arten s it ic m a - tri x (with relatively uniform, sm all a nd well d i s p ers ed al l oy carbid e s) c om p a r ed to th e oth e r two m at e r i als e v al uated. Th e ma teri al ha _ g ood thr ough hardO n abil- try. Fo r th e te s t bea ri ng s , t he m at e rial h ar dne ss was c , Jn t r olled a t room t e m - pe r ature to Roc k well C 6 3 _ 1 for the rin gs and Ro ckwel l C 03 6 0.5 f o r the balls.

This- ma teri al was used fo r al l the lub r icant te s ts previously m entio ne d in this paper. The re su lts wi th bearings m a d e from this m ateri al , ar e shown in figur es XllI-6, XIH - 9 , and XIII-12.

AISI M-1 is also a h ig h-speed tool steel which has be _ under investi ga tion as a potential high-te m perature beari ng material f o r a number of years. Tl_t s m ate rial tends to have rath er large carbides which agglo mera t e or band. Since these massive carbides will act as nuclei f or fatigue failure, assuming they ar e J located in th e cr i tical stress region, the y m ay tend to reduce th e rolli ng- ele m e n t _ fati gu e life. The m ate r ial hardness for the test bearings w as co n trolled to I_ R oclcwell C 63 _ i for the rin gs and Roc k well C 65 6 0.5 for the b al l _ . The bear- ings Were of the same design as the A] _ I M-50 beari n gs. T he fatigue results ob- t!

rained with this m ate ri al a re shown i n fi gu re XIII-1 2 a n d co m pared with th e M-50 bearin gs under the sam e operating conditions. } !

' t WB-49 iS a m ater ial de v eloped specific al ly for hlgh-te m peratu r e bearing ap- _ , : pl i catio n s (ref. 16). It co n t ai ns con _ iderably more alloying ele m e n ts than either the M-50 o r th e M-1 m a teri al . The inherent di f ficulty wi th this m ater ial t _ th e _'li t e ndency toward rather massive car b ide s egregatio n w h ich , at l e ast in th e preSe n t !i c as e, could no t be S u ffic iently brok e n up during forging to eliminate its effec t as a stress raiser, t The WB- 4 9 rin gs were heat-treated to a roo m -te m perature hardness of Rockwell C 64 _ 0.5. The WB-49 beari ngs utilized M-1 to ol-steel balls fro m t he same heat as that for the A _ SI M-1 bearings. Pr e vious experience (ref. 9) h as shown that WB-49 ba l l s co u ld no t be m anufac tu red without producing incipient m icrocracking. As a result, b all s made fro t h WB - 49 had extre m ely s hort f a - tisue lives.

Fa ti gue results w ith t h e WB-49 bearin gs are co m pared with those of th e AI _ I M-50 and AISI M-I bearings in fi gur e XllI-12.

From th ese data, th e fatigue-life dfffer _ mce b e tween the M-50 ar i d M-I st ee ls can be considered stat is tic all y insignific a nt a t 600 ° F. However, the difference between the WB-49 and both the M-50 a nd M-I m ateri al s is statistic all y s iguifi- i / st_ tf t c an t. For the M-50 an d M - 1 b oa ring_ run with t he sy n th e ti c pa r afft n t o oil, t h e experi me ntal b earing 10-p e rcent lifo exceed s the AFBMA-predtct _ d (c a t alog ) rife by f a ct o r e in exces a of t 8 and 6, re spectively. A s a re sul t , for these two ma teri als n o deratin g of be a ring lifo i s req ui red. H owever , th e ;ire of WB 49 wu i ' lose th_ h al f the AFB MA _ pred i_ t _ ( c a t alo g) l ifo. H en c o _ th is m ateri al wo u ld have to b e d e r a t _ d.

SPEED EFFECTS ONB EARING LIFE

Effect ofReduced Ball M ass '

When bali bearings are operated at DN values abov e 1. S m illion _ centrifugal forces produced by the balls can become sif_ iflcant. Th e resultin g incr e as e in Hertz stresses at the oute r -race b al l contacts can se riously S llort _s be m i ri n g fa- tigue life. T h e re are sever al po s sible approach e s to solvi ng this proble m .

Th e first and most obvi ou s approach is to opti m iz e t h e b _ riu g internal g o- ,. o m etry fo r m axi mum life using _ high-Speed , b alt - beart n g-dyna m i es co m pu te r program (ref. 1 '/ ). This Will yi eld the opti m u m b al l diam e t er and nu m be r, r a e , curvatures _ and contact angle. However, thi s a p proac h will not yield more f _ an _ a s m all , incre m ental i m p ro ve m e n t in li fe over that of b e arin p n o w i n u se. A s a result , less con s ervative a ppro a ch es must be conside r ed.

Another app r oach is t o reduc e the m ass o f the ball. A 50-percent r e d uotlon • _ in ball wei g ht, w h ile m ai n taining the ball diameter consttmt, c an th eor e ticall y l , e- , s uit in a signific a t _ t i n creas e i n lif _ a t D N v alu es grea te r than $ _U tion. Th e el - . : fee ls of decreaSing ball weight in a 150- m i l li me ter - bo re ball beari _ are i l ht - strafed i n figureXIII - 13. For the 5000 - a n d 1000-pound load s , the life c a rl b e t increased more than three ti m es at the hi g h er DN values simply by red u ci ng the b al l mas s 50 pe r cent,

Sphe rically Hollow Balls

One m eans of reducin g the b al l mas _ i _ tO m an _a_ tU l 'e ho l low b_I s . _ pher - , . ic al ly hollow bails are f a bricated by lot t i n g two he m i sp he ri cal s hells by e lectro n - beam weldi n g , and finis h ing and he _ t treating the hollow bal ls by methods no r - mal ly used in th e manufact u re o f coilverl _ ion al sol i d lz _ l _ .

t \ b_ o ' H ollow bal l e o f 11 / 16-1 nch d i am e t er v / t i n O,0 6 0-in c h w al lthi ek n o Be w o re f a b _ r icat_d in t h i sm a nner an d fl tt od to 75- mil ll me te r -b or e ball b e ar ings ,T h e b e a r - i n g_ w o re o pe ra t c_ t at BOO man d 1000-po u nd _ hru s t Ioa d _ a t _pee d _ u p t o 18 0 0 0 rp m (1, 35 m ill i o n DN ) , A _u p0rr e fir, od n a phth c m l_ m i n e r a l oi l w a_ u_ ed ae t he lub r i- cant i n an ai r -o i l m t _ t w i th fl a w x a te _ ran g i ng fr om 0, 01 to 0,0V pou nd p_ r min u te ( r o f _ , 16 an d 19 ) .

P o e t-t e st in spec t io n o f t h e boa r i n g_ a f t er a few bou r n ru n ni ng s h o wed n x , t o n Bi vo d _m ago to the h a l low balln w ! th m any e pM l _ o n t h e ottt o r _ u rfa co. S oy- o r al b al t _ fr o m o ne b ea ri ng wo r e 8no , t o n ed through t h e wold an d poli s h e d to d o - !

lor ra ine t he ext en t an d s our c e o f the f_ fl u ro _ , A c ro ss _ oe ti on o f a hollow ball _ is s h own i n figu r e XIII _ 14, A b ead o n t i_ o i n ner dia m et er t H f or m ed by t he e l ectro n -bea m wold, It is sp e c u late d t hef t t h is b e a d a c t _ i as a n otc h or _ tr oB_ ii raiser i n the ba ll wall a n d i n itiat _ flcz u r o f atigue f ai lur e s after the be a ri n g w as _ , i op e rated f or o n ly 13 h o u r s , A typical f l e x u re f atigu e fai l ure is sh o wn in f i g- _ ure XlIl - 15. The u p p er ph o toi _ r a ph shows two crac k s e lu i nati n g from the W eld _ - bead tow a rd the outer surface of the ball . T h e crack show n in the l ower photo - _!_ graph has also o rigi n ated a _ the weld bead but has propa g at e d into a crack net- i : i work an d a fat ig ue sp al l at the ball outer surface. Balls from othe r b e arin g s 1 that were run u nder similar conditions were ex a m i ned . So m e S how ed s poi ls o n I " _' the outer s u r f ace; som e showed no appar en t dam ag e. Al l t h ese b all s we r e sec- /_ tinned and po l ish ed , a n d al l r e veal ed c ra c ks origi na ti n g in the vici n ity of th e weld bead at the inner surface. From t h e ex am inatio n of"all run holl o w b al ls, it was conclude d that they al l had fail ed in f l exure beca u se _ f a stress coflcet _ t r ation i n I_ , the region of the Wel d at th e in n er bal l surface, i_i The u se of ho ll o w b all s c a_ still be a d v a nta g eou s a t hi g l _ e r be ari n g DN v al- i_ '_ " ues, pro vi ded the following steps c a n be take n : (1) use a w al l tht o . _e s S S uch th a t ,_ flexure iS- m ini m iz ed a n d yet m ass reduction is appreciable , (2) control t h e weld tl penetration aro un d the periph e ry of the bal l , an d (3) m ai n ta i n a un iform wa l l _ ' _ thickness. Unle S s these steps ar e t aken , ho l low b all s wil l b a re a n un b al arice and ! : !

a different stiffness at the weld. U n der dynamic co n ditions, these factors ad- ver s ely aff e ct the lives of the balls.

DrilledBalls

A n o th e r method o f red u cing ba l l mass is to m ac h i n e a conce n tric ' hole t h rough the ball (fig . X1T[-16). The amount of m as s red u ctio n eq ual to that of a'" t h if _ , g all spherically ho l low b all c a n b e ach i e ved with several advant a ges. Usin g dr i lle d 4 _ 4 b _ ll _ a ll e vi a tc _ p on al blo p ro b lem s o f b_ ll u nb ala n c e, b e c a u n e t he ho lee o n centei - C ity c _ b e m atnt _ tlned very acc urately, Additi o nally , a ve ry Sm oot h su rf a c e f l nL_h can he a _hi eve d_ wit hn ut the l rre gul a rltlo _ pre _ o n t i n th e ar e a of t he weld (rof, 20).

To_ t _ w o r e conduc t _ wit h 7 i t -m i l l ime t er-he r e b all b ear tn g_ u e ind ll / 1 6 - inc h-diame t e r d r i ll ed h al le, T he bo a rln g_ w or e o p e r a ted a t a t h ru s t lo ad o f _ , 00 p ounfl_ at . p eodn to 21 _000 rp m (2 ,1 m i l l io n DN) wit h air- o i l- mint o r o i l-_ ot lubric ati on . T he d r i l l ha lle worema d e b y el e ct r ic di a c har go mac hi ni n g ( I _ I_M) a O,4 S- i nc h - dia me t er ho le t hrough t h e c ent er of a v o l td ball t o e f fec t a fl O o p orcorl t w ol_ , lt t r e d uc t io n, A f inl nhod ba l l i s l l lu n tr a t _ t In f i gure glII . 16, Pi n_ t h ro ufl h t h e ce n t e r of e a ch ball p oo kot a t th e pitch di am e t er of the b e a ri ng (fi _ , X I II- 1 6) w ore use d t o r os trn in t he edge o f th e Ilolo from cont a ctt n_ th e ruco _ro o-v o p r i o r to op e r a tion, D u rl n f, o per atio n the ball s p ositio n th ems elv es w h o re n o si g nificant contact occ urs betw een t h e e d ge o f t h e h o l es an d t he p i ns, T w o ' / 5- m t U t m 0tor b all bea ri ngs ope r a t e d Su c c e s sf u l ly over' a rang e o f co n - ditio n s without damage or exce s s ive wear , O ne bearing was Still oper a ti ng Sat is - factori l y after 10 7 hours of a ccu m ul a ted runnin g tim e, with 6 5 hours above 1,5 m ill ion DN, I n a n other stud y , four 1 2 5- m illi m eter - bore b _ t l l beari ng s'usin g d rill _i ball s providi n g a weight r e duction of over 50 p e rcent w e re op e rated u n d e r a _ 00- poan d thrust load a t speeds from 8000 to _ 4 000 rp m (1.0 to 8.0 mi llion DN) i n two sep- ' o r ate t e _ ts. S p e cial modt f t cati o tt s w e r _ ma d e t o th e b e arin gs that p r t _ vided f o r adequate l ubrication arid sufficient coo l in g with a type II e st er oil. All fottr b e_ t' - titu s operated satisfactorily tit thos e t m tr em e conditions for several hours ru n- nin g time witho u t dama g e or S ignificant wear. O n e of these d r t Ue d ball b ea r tn g e i s shown in figure XII _ -I Y . It can b e s e en th a t the c om po nen ts of this b _a ri ng are in very good condition after 4.7 h ours of opera t io n . T h e r _ su lts o _ these ex- / perime n t a l programs indicate th a t t h e dril l ed ball beari ng cortc _ pt sho w s a great de al of pro m ise f o r h ig h -s p eed be ari ng a p p l ica t io ns .

SeriesHybridBearing "

A n other m eans of achieving high - speed beari n g o pe ratio n a nd ton g i ife i S = thr o u g h t h e series hyb ri d be ar ing (fig. XIIi - 1 8 ). I n e s Se i tce th e se ri e s hyi _ri d bea r ing co m p r ise s an angu l ar c o ntact b_U b e ating i n se ries w tttt a flu| d film bearing. Bo th b e a rin gs han d le th_ e ase t hr u s t load o f t h e mois t s h a f t; howeve r , a s s peed i s in c reas ed , th e fluid fi l m be a rin g l if t s o_f a t it8 pad b e catl S e o f { he in- c rease d pre ssu r e c aus e d by th e c e n t ri f u g al f o r c e of th e l u b r i can t f ed t hrou gh th e m ai n sh aft, A B t h e b earing li f t B o f f, t he re i s a di f fe r o: _ tial in B pood bet w een ti l e porti on a tt a c h ed rig i dl y to tl _ o _ha ft a n d t h at a tt ach e d t _ ch v a ngu l _ tr c ontact bo a r - l ag ro s ultinR in a lo wer o p o od o f ti l e a ngu la r co n ta c t bear i ng relat i v e t o the s h a ft, T h e te s t ro _u l tB wit h t hl_ boa _ . ln g a re sh own i n f igur e XI I I-1 O , w hi ch is a plot o f t h e inner r ace flp o cd o f the ba l l heari ng a _ a f u ac ti on of t h e shaf t speed, In n or - nml o per a t io n t h e sh aft _p o L_dIn _l w ayB equa l t o t h e Inn e r- race _p oo d, AB t h e s pe e d o f t he n h nft I s in cre as ed a n d t h e p re ssu re i t _bu i l t u p in t h e flu i d f ilm b oar - ing , t h ere i_ a f luid f il m fo rm ed and a d o c_ o a_ o i n _ pood betwee n tile flui d f ilm bea rin g r i g i d ly _ ttta ch od to the s lmft a nd t ha t attac h ed to the ba l l h ear i ng. T his r_ u l to in a de cr e as e in t h e inn or or_ co _ po _ o i t h e b all b e a r ing . I n t hi s ca s e (fig, X I II - 1 9 ) a fl uid fi l m Is f or m ed i n t h e f l u i d fi l m b earin g at a s peed o f a p - p r oxi m atel y 1 million D N. T h e decrea s e i _ approxi ma t01 y $0 perce n t at the ele- vated 0pe ed o f a ppr o x im atel y 2 m i ll io n DN. T h e decre as e in speed m o ans a do - cro _ e in ce n tr i fuga l force a nd as a re s u l t an in crease in l i fe. A c o mp a r is o n o f I the drille d ba ll b earing an d the s_ r i es h yb rid ;_ v ar i_g o n bearing lif e at t h e e l e - v a t ed s peed s is sh own i n f igure XIII-90. For co m p a rison I m rp o se _ d a ta f or a s o l id b a ll b e a ring are given. In co m pari n g the dril l ed b a l l w i th the so li d ba ll at 4 milli o n DN , the li fe rat i o w i t h the dr ille d b a ll is o f th e order o f B to 1 . W i t h a s er i es h y bri d b0ari ng w i th a 30-percent s pe ed reduct i o n , the l i f e ratio i s aPprox- i m at e ly 6 to 1.

T h e d ri lled bal l c o nce pt c an b e com b i ned w ith the s eri e s hy b rid bea _ 'i ng .

Wi th this combtna t iob t h e life r a tio a t 4 _ :tll t on DN i s a p prox i ma t e l y 10 tO 1 over the conve nt io na l solid b al l b e a ri ng . W ha t is al s o i m_ ortau t i s that ther e is on ly a small decrease i n life with sp eed .

I i

S EA L F UN C TI ON S A N D DE S i 6N C ONS IDER A TIO N S

i ' i ,

Seal S y stems i

As previously stated , m ain - shaft s e_ l e ar e used i n gas turbine e n gines to i restrict gas l eakage i n t o t h e b earing co m part m ents (su m ps). The sea l s at the ( rent of the compressor ( f ig, X M -21) usually present no prob l e m since t h e su m p can be surro u nd ed by comPressor bleed _ tr o f m odera t e p _ esgure and t v m pe _ -a- _ re. Bleed air l e ak s through tae t a e I n- _ d, t h us, p l ' ess uri zeS t h e f r o n t _ ump.

For moderatebleed air pre s s u r , _ mperat u re s a single s haf t s b _t ! of m a ny typ e s is adeq u ate. A s a n examp l e, ri n g sea l s have been u sed at 54 psi a n d i • 3300 F; c irc u m fer e ntial s e al s a r e more than a d e qu a te f or thi s lo c atio n, and i n the present state of development, have be en used to 8 5 psi an d 700 ° F with a sliding sp e ed of 240 feet per se cond. F ace seals ar e al so adequate , as some are oper- ating at 125 psi, 800 ° F , and a sliding s peed of 3 £ .0 feet per second. The final selection o f the seal type for th e f ront compressor location depends o n the m any design and system consideratio n s, including preference.

At the middle of the en g ine (fi g . XIII-21) and at the tur b i n e bearin g sum p, the seal opera t ion al r eq u i reme n ts beco m e more s e v ere. Th e bearing( s ) arid se al (s) tefld to be l a rge because o f Shaft size and because the thrust bea ring must be large enough to carry the n e t th rust load on th e rotating parts. In ad di t io n , th e su m p is usu al ly surround ed by higher p r e s sur e gas (and corr es pondin g higher t e m perature) than th e front of th e engine.

F i gures Xlii-22 to XHI-24 show some s haf t seal arr an g eme nts f or th e turbin e i i , bearing sump location. Her e th e basic probl e m is pro te ction o f th e b e arin g _ u m p _ from th e turbine cooling gas. In early en gi n es th e cooling gaS pres s ure and te rn - ! .

perature were relative ly low ar i d a S ingle l a byri nth se al, which restri c ted h _ rb in e _ cooling g as l e akage i nto th e sump , was adeq u ate (fi g . XIII-22). (Some engines ii use a once-through l u b ri catio n System i n which the Sump is pressuriz ed by rel- a ti vety cool bleed air, and this air and the oil luti ric ant a re allow ed to vent over- board. ) At the s e pr ess ures, th e efficiency loss due to se al leakage was not si g - nificant. Howev e r, a dis ad va n tage of the laby rin th Sea], as co mpa red to th e close-clearance se al s (ring, circu m fer ent ial, an d fa c e), is hi g h l e akage with :;, corres po nding l a rg e r duc ti ng requirem e n t s and easier pa ss ag e of airborn e water an d dirt i nto the s u m p. In additio n , f or la by ri nth seal s , re # erse p re s s u r e dro ps must be avoided to p r eclude high oil loss. Further m ore , ra m a ir temp e rature ( m i n i m u m -te m pe ra tur e su m p-pressurizing air avail a ble) i n cre aS eS w i th i n - ' _ • " i " creas i ng fligh t s pe ed. Ther efo re , a flight speed limit exists b eyond which the t low-leakage close-clear an ce se al s are d es irable in ord e r to preclude s um p t_ i r e s .

The required pressure of the turbine cooling g as is dete r rd in t _ i by the tu r- bine inlet pressures. As turb in e coolin g gas pressure r eq uire m ents i n crea s ed with v n gi ne develop m ent, the single l a byrin th seal of figtt re XII1-22 w ais n o longel' _ sui ta ble, an d some of the seal systems such aS illufJtrat ed i n figures X M -2$ [ and XM -24 wer v used. A conve n tion al face s eal (fig. XI _ -2 5 ) ha s sati s factory I performance to pressures of 125 ps i , te m pera tu r eS o f 800 O F, and s l i _ lhg speeds = : of 350 feet per s e cond. Recen t seal studies (ref. 1) showed that operation at _" higher pre ss ures, tempera tu res , or s l iding s pe ed s w as u n satisfactory from a wear and leakage st an dpoint; subsequent analysis reve al ed that the se al limita - tions were primarily due to th ermal deformations t ha t in duc ed s e al i m b al ance, i • , wh i c h, i n turn , c aus ed high wear a n d leakage , Se e r e f er e nce 2 1 for a di s cu ssi o n : o f th e im b alance problem that li m its th e c apability of the conv e nti o n al fa ce seal .

When t he p r es s ure , s peed , and t em p er ature s exc eed the capability of c o n - v en tional co n tact seal s (s u ch a s t h at o f fi g . XIII- 2 3), a press u re S tagin g se al S ys- te m i s u sed . Such a s ystem of multilaby ri nth s i s s how n in figure XIII - 24. In thi s , i s e al system a labyri n t h seal re s tricts the l eaka ge of relatively hi g h pressure (P1) , high-te m perature tu rbine cooling gas to a n overboard vent (P b ) (or to a low - pressure re gio n). Low-pressure compressor bleed (P _ ) s urr o u n d s t h e su m p a_ d, hence, provides ther m al protection, a n d l eakage (through the labyrinth) into the su m p provid e s required su m p p r es s urizat i o n . It should b e n ot ed that, as the tur- bine cooling g as press u re (Pl) increases , the leaka g e has a sigrLific an t ef f ect on ef f iciency.

In some engineS, combinations of labyrinth and face seal arc used to form the i seal system. In these systems the labyrin th se al next t o the sump(see fi g . XIII-24) would be replac ed with a face or shaft ridi n g co nta ct se al . Such a system provides low a i r le a kage into th e su m p a n d i S used when the s e al pressurizing air (P _ ) is too hot (supersonic flight) to allow high leakag e into the su m p.

Advanced S e al R e quirements

As pr e vi o usly noted , figure XITI -2 sh o ws how th e ma ' _- shaft s e a t e n vi ron m e nt has become increasin g ly s e vere. Proj e ct ed f u tu re r eq uire m ents a re shown.

. These are bas ed on p as t his to ry of engine dev e lopment and on pres en tly desired % operating goals. Durin g the p as t 20 years se al sliding spe ed s have doubl ed a n d : se al ed air pre s sur es and te m pera tu res have had corresponding increases. ' I t is an ticipated that advanc ed e n gines wi ll have c o mpressor dischar g e pres- _ r sures i n th e range o f 400 to 500 psi a n d s eal te m peratures of approxi m at el y 1400° F. The seals for these engines will have sliding spe ed s of 500 to 600 feet per se c on d. E ngi n e s f o r supe rs onic aircraft m ay ge n erally have lower c om preS- s or d i scharge pressur es , but the turb i ne coo l i ng g as te m peratures w il t be near 1200° F. The high sliding speeds i n these ad va nced en gin es dict a te that the Seal- ing s u r f aces s hou l d not be in rubb tug contact. Thus , th e s elf- a ctin g s e als_ S ince.

they operate without rubbi n g co n tact, are potentially useful i n th e advanced en- gi.e s .

J 43 8 t

FACE SEAL WI TH SELF-AC T I NG GEOMETRY

F OR LI FT AUGMENTATION

F ig u re X III-25 is a cr o ss se ct ion of a fa ce s e al wi th se l f-a c ting lift ge o m- e t _ T. As w i th a con v en tio nal f ace se a l, it con si s t s of a rot a ti ng se a t t ha t is a t - t a c h e d to th e shaft (a l l rota ti ng par t s are shade d) and a nonro t at in g seal h e ad as* 8embly t ha t i s f r ee t o mo v e in an a x ial di r e ct i on a n d t h u s acc o mmo d ate e ng in e t he rma l e xpan s ion (a x ial). T h e se c onda ry ring (p is t_ _ r ing) is subj ecte d only t o t h e ax ial m ot i o n (no rot a ti on) of the hea d a s s e m bl y, and seve ra l springs p ro v i ded m ec hani ca l forc e to m ai n tai n c _n taet at start a n d stop. I n o p e r a tio n, the sea ling : fa ces a re separ a te d a s ligh t amoun t (i n t he ran ge o f 0. 00 0 5 i n. ) by a ct i o n of t h e se lf - a cti ng lif t g e o me try , an d gas .._ . a k a ge is from t he h ig h - pre ss ure side (in s id e _ i di a me ter o f c a rbon pr im ary r ing) _ cr oas the s ealin g dam in to t he bea r in g s u m p , i T h is g as leakage pressur i zes the su m p and assures p r oper scave n gi ng of th e _: b ear i ng l u b ric a nt. I t sho u ld be n o t ed t h at , alt ho ugh t h e s e al ed g as t em p e ratt:r _ i S i_ h i g h , te s ts h a v e s h o w n that c on siderable gas te mp er a ture dr o p occ u rs i n the l_ak - _; age flow so that the l eaka g e tri te t h e s u m p d o es n ot pose a fire haz a rd w h en the se al i s operating properly. Operati o n to 40 Sta n da rd cubic feet per m i nute a n d 1 200° F s ealed air te mp erature has bee n obtained wi thout a s u m p fire.

A more deta i led sche m at i c of the seal i s shown i n flgureXIH- 2 6. The seat i s t h er m al ly a nd part iall y structura lly is o lated f r om t h e sha ft by t h e f ollo win g m ea n s: (1) The r ad i al spa c er betw e en the seat and the sh af t serves as a sprin g that mi t ig ates the effect o f nonunifor m shaft therm al growth (radi al) due to the axial ,j ' t h ermal gr ad ient. T h e effect of t h er m al m o v ement i s further a t te n uated b y p4 lo t- / , _ i ng th e s e at ov er it s ce n tr oi d.

( 2 ) T h e bellows cla m p ing spacer a pplies a predeter m ined c l a mp i ng f o rce t ( approximately 2 0 0 0 lbf ) , wh i ch is probably an o rd er o f m agni tu de less th an clamping f o rces prod u ced by t h_ u s ual be a r i ng l o ck nut as s em bly tec h ni q ue.

Therefore , the defor m ation during asse m bly is minimized. The seat the rm al de- sign is tailored to minimize the axial ther m al gradient. U s e of a m olybdenu m alloy , instead of the convent i onal SAE 874 0 or heat-resis ta nt alloy , provides a relatively low deformation criterion which is defined as the ratio of ther mal ex- pansion to thermal conductivity (ref. 22 ). Th e cooling oil flow path (see fi g . XIII -2 6 ) is also tailored to mini m ize t he ax ial ther m al g radient. By pass ing the cooli n g off under t h e rad ia l sp a c e r , the fi rst co n t ac t o f t h e cool i n g o il with t h e seat i s near t h e h ott e r f ac e ( f a ce ma ted with no sepiece) o f t h e seat. Anal ys i s

t

had sho wnthatthis oilroute m ini m iz es th e th e r m al g radi e nt.Centrifu g al pu mp - lug action due t o rotation an d th e 4 _ radial exit holes near t h e hot face as s ure ev en di stri buti on of the cooling oil. (See ref. 2 8 for d _ cripti o n o f the s _I f- a ct i n g s_al c o m po n ent s .)

T he s elf-a c tin g lift pads c on s ist of s series o f s h al low recesses a rran g ed circu mf erenti al l y a round the se al under the s eal in g d am ag s h ow n iu f igures Xili-27 a nd XrlI - 2 8. An i m portant point i s that th e lift p ad s a r e bounded a t the-- inside diameter and outside di am et e r by th e _ e al ed preSSure, PI' (This i s ac- co m plished by fe ed slots th a t communicate with the annular gr o ove directly und e r the se al ing d am . ) There f or e, a pressure g rad ie nt due to gas leakage oc curs o nly across the seal d a rn. Thus th e ef f ect s of force change s du e to se al face deforma- tion are minimized (ref. 24).

The effect of th e s elf-ac tin g pad s on face s eal operation is illu s t r at ed in _ igur e X _ II- 2 8, which shows parallel se aling faces operatin g without rubbin g con- ta c t because of a b al ance between the closing f o r c e and the opening f o r _ e (lift pad forces plus the pres s ure ac ti ng b etw een th e Sealing fac es ). For t his parallel case, if th _ se al tends to close, the a verage pressure in the g ap does not chang e but the gas bearing force increAS e s. Thu _ , a condi ti on o f no rubbi ng contact can prevail except at startup an d Shutdown. (See refs. 24 an d 25 f or a d e ta il ed di S - i .

cUssion on self-acting pad forces a n d se aling g a p forces. ) The s _ lf-acti n g lift pads develop high f or c es a t op e ratin g s peeds to prev e nt ru bbi ng contact (ref. 2 4). For inst an ce, calcul ati on s of lift pad force(fig. XIXI - #9) r indicat e a n 80 - pou nd lift force a t a 0.000 2- i n ch gap height with 31 5- pSi a ir an d _ t : _ 500-foot-pel'-second sliding sp e ed. However , if th e g a p op ens ( e. g. , to I / 0.001 in . ), the lift force drop s m arkedly to 4 po u nds. Th u S, there is little tend- I : !

• _ ency for the lift pads to hold the se al Op en at l a rge gapS where le a kage will b e !

h ig h . The lift pad s , therefore, have a force-aga i nst- g ap-heigh t ¢flAracter is tic _ i l ?

(high-gas fil m sti f fness) which m ak es lift pad ineor po £ati on inhe r e _ lfly s uited to s eal ope r at i on. Thus th e l ift pad s provide the following i mpo rtant f e _ ttures geu - ! ' er al ly missing in convention al face se al s: (1) high-gas fil m S til l ness that al- I : !_ lows th e head to dyna m ically track the seal face m otions w ithotit rubbin g contact I and (2) the ability to operate with divel' gen t face defor m a ti on. _ It i s im portant to note that , bec a use of th e axial th e r mal graditmt of th e t _- gi n e, divergent sealing faces a r e a n a tural tendency, ff the par al l el f aces of fig.

+_ ure XIII-2 8 beco m e divergent (with respect to g as leakage direction), the avera ge gap pressure will decrease and an unb al a n ce force will tend to close t h_ gap.

However , as th e gap decreases , the lift- pa d force in credses rapidly an d actS to prevent rubbing contact. There is a limit, howeve r , to the a m oiz _ t o _ divergence 44 0 e \ th a t c a n o ccur bef o r e rUbbin g contact i s es tabli s hed a t th e insid e dia me t e r o f the li ft p ads ,

S elf - Acting S e al Performanc e

A 1 20 - hou r en d urance run wa s m ad e at 400 - foot-p e r-s ec ond s lid i r.g speed, 200 - pp t s eal e d pressure d iff erential , an d 10000 F s ealed ga s te m pe r ature. T h e g as leak ag e durin g th e 1 2 0 hours ( f i g . XIII- $ 0) a verag e d about 11. r / sta n dard cubic feet per m inute a nd this i s about one-tenth that of a well des i gned l a byriuth seal. The sh a rp increase in l eaka g e at 116 h our s WaS found to be Ca u s ed by fail - ur e of the thru s t bearing. (Bearing s pall produced Severe ri g vib r ations which th e seal acco m odated without rubbing contact. ) The s urface profile trace taRen after the 120-hour endurance run iS shown i n f igure XIII-81. Inspectio n revealed t h at t h e da m h ad c i rcu m ferenti al scratches that were probably f ro m airbor n e de - brt s passing across the se ali ng da m . Thus , the a m O u n t, pa r ticl e size, an d type ' of a i rb o rne debris ma y be a significant factor i n degr ad atio n of the s eal da m .

T he deepest scratch was 0.0002 inch. The pad wear was 0.0005 inch; origi n al polishing scracth e s were still visible on t he pad su rf a c es.

F i gures XIII - 32 and XIH - 33 show t he condition of the primary se al ri n g after the 338 hours o f running. Figure XIII - 33 S ho ws a closeup o f the portion c irc l ed i n fi gu re XIII - 32. Th _ ori g inal polishing S cra t ch es are S till visi b le on the l and are as of the pads. Thus, the se al was oper _ tting w ith out rubbing contac t (except at s tart and stop).

In order to check the effect o f startin g anti s topping , th e s eal w as subjected to 40 starts and stops. The ru n s were condu c ted wi th roo m te m perature air at 115 ps ta . T he sea l was then al l o w ed to coast to _ stop (required about 25 sec). t A comparison o f s u rface profile t ra c e s bef o re and after the 40 s ta rts an d st o ps reveal ed carbo n wear was l ess t h an 0.00005 inch. r J :

L eakage and Wear Com p arlson

A l e aka g e an d w e ar co m par i son to1" conve n tional f a c _ seats , l a by ri nth, an d !

, . self - actl n g sea l s is sho wn i n figu r eX l ll - 34. In teSts ti nder slmu la te d e i_gin e t c o nditions a c o nve n ti o nal face seal h a d 400 times the wear of a se l f - acti n g se al .

T he o pe r at i ng cond i ti o ns were s im ilar e x cept th at the con v e n tio nal face seal w as _i i " t r u n for 9.7 ho u rs and the self - acti ng s eat f or 338 h ours. A labyri n t h s eal , h ow- eve r , can be e xp e cted t o operate without wear.

I n lea k ag e co m pari s o n th e la b yrinth se al h as a bout 10 times that o f th e self- acting s e at , a n d this i s the principal di s adv a ntage of th e la by rtn g h Seal. The lea ka ge of the rubbin g -c o nta c t s eat was g r e a t er than that of the self-a c ting s eal.

T h e greater le a kage i s ca u sed b y the high wear rate (wear debri _ causes S eali ng face s e pa r a t i on) a n d seal vibration a ss ociat ed with rubbi ng contact.

CONCLU D ING REMARKS

Base d u pon technology which h as b e en advanced by N AS A , beari n g s c an be op- _ e rated at600 ° F in i n ert en virOn m ents with r e liability equal to o r in excess of that i .

exp e ri en ced in flying eng ines today. Te c hnological ad vanc em ents also allow h igh- F s peed be ari ng operation with long life. Drill ed ball an d series hybrid bearing con- 14 cep ts were de v elop ed to in c r eaSe al lowable bea ri ng spe ed s. Tests to 3 m illion DN _ confir m that these bea ri ng conceptS-have th e predict ed speed po ten tial . Finally , J , _ tes ts on the s elf - acting seal show ed higher speed an d press u re capabi l ity than i_ conv en tion al face sealS , a n d gas leakag es w e re one-t en th th ose o f co m monly used i la by rin th se al s. Additional w o rk on-th es e hi g h - S peed beari n gs a n d seals is ne ed ed _ " to explore their full potenti al and to establish their endurance life. Ii

REF 'REN C ES

1. Parks , A. J . ; McKibbe n, A. H. ; an d Ng, C. C. W. : Development of Main ? " Shaft Se at s for Advanc ed Air Breathin g Propulsi o n SyStems. Rep. PWA- i_i i $161 , Pratt & Whitney Al l 'craft (NA S A CR-7 2 338) , Au g . 14 , 1967. I_i : _ , 2. Zaretsky, E. V . ; an d Anderson , W. 3 . : EHD Lub ri cation. Mac hi ne Design, g _t eel. 40 , no. 2 6, Nov . 7 , 1968, pp . 16 7- 173. t 3. Zar e tSky , Erwi n V. ; A nd erso n , W i lliam J. ; an d Ba m be r g e r , Eric N. : Rollin g - Ele m_ mt Bearing Life From 400° to 6000 F . NASA T N D -B 002 , 1969.

]

4. Anderson , W. J. ; an d Zaretsky , E. V. : Roliin g- l, _ le m ent B e ari ngs. Machine D e sign , eel. 4 2 , n o . 15, J U n e 18 , 19 7 0 _ pp. 2 0-3 7 .

,p B. Zare ts ky , l _ rwin V. ; an d And e rson , William J . : Pr el iininary Dete rm inatiotm o f Te m pera _r e Li m itatio n s o | ] _ ster, Ether, and Hydrocarbon B ase Lubri - c ants i n 2B- m_ n Bore B al l Bearings, NASA TN D-4146, 19( _ 7.

6 . S ibl e y , L. B. : E last oh ydrodyn am l c Lubric a tion. Machin e D esign , vol. 3 8 , n o . 24 , Oct. 13, 1966 , pp. 2 2 0- 2 2 1.

7 . Give n , P. S . : L u b r ica n t Film Effects on Roll i n g- C on tact Fati g Ue. Paper pi _es ented a t D a rt m outh Co l lege Bearin gs Con( e r e nce _ H a nover , N. H. , Se pt. 196 6 .

8. S mi th , J. O. ; a nd L l u , Chang K. : S t re sse s D ue to Tan g ent i al and Nor ma l Lo a d s on An Elastic Solid With Appli c ation to Some Contact S treBSP ro b- I ctu s. J. A p pl. Mech. t v ol . 20, n o . 2t June 1 953 , p p , 16_ - 1 66 .

9, Wac h end o rfe. _ : , C. J. ; and Sib le y , L. B . : B _ arln g -Lub rlc an{ En d u rance Charact e ri s tics at High S peeds and High Te m peratur e s. R e p. AL6 5 T0 6 8 , SKF Indu s tries , I nc. (NASA CR- _ 4 0 9 7 ) , 1965.

I0. Zar e t s ky , Erwln V. ; a nd Anderson t William J. : Evaluation o f High- Temperature Bearing Cage M a teri als . N AS A 'IN D- $ 8 21, 1 98 '/ .

11. Bi ss o n, Edmon d E. ; and Anders o n , Wi ll iam J. : Adva n c ed Bearing Tech- n o l o gy. NASA SP-38, 1964 , pp. I _ 5-2 02 , 328-329.

12. Bam ber g er , E. N. : Bearin g F atigue I n ve Sti gat ion . Rep. R 6 '/ FPD2 0 9 , Ge n eral Electric Co. (NASA CR- '/ 2290) , Sep t . 1 _, 1 96 7 .

13. Ba m ber g er , Eric N. ; Zaretsky , Erwin V. ; a n d A n d e r s o n, Willia m J. : Fatigue Life of 1 2 0- ram Bore B al l Bearings a t 600 ° F W i th Fluorocarbo n , Polyphenyl Ether, a n d Synthetic Para _ inie Base Lubr i c an ts. NASA TN D-4850 , 1968.

14. Buckley , D. H . ; and Jolm S or % R. L. : D e g radation of Poly m eric Co m pos i - ti o ns in Vacuum to 10 -9 mm H g in E.vaporat t on an d S li ding Fricti o n Exp e ri- m en ts . S PE Trans. , vol. 4, no. 4, Oct. 1 964 , pp. 3 0 6-3 1 4.

t 15. B am ber g er, E. N. ; a n d Z a retsky , E. V. : Fatisu e Li veS at 60 0 ° F o f 1 20 - Mill im eter-Bore Ball Be ar i n gs o f AISI M-50, AISI M-l, and WB-49 Steels. _ , NASA TN D-6156 , 19 _ 1.

16. Ph i lip , T. V.; Nehreuberg, A. E. ; and Stevet l , G. : A Study of th e Metal- lurgical Properties That Ar _ Nece s sary for Sati s f a cto ry Bearing Pe r for m - ance and the Development of I m prov _ i Bearin g Alloys _ or Service up to 1000 ° F. Crucible Steel Co. of America (WADC TR- Y/ -343, pt. 2) Oct. 1958.

44 5

!

Q 17. Harris, T. A . : Au A na lyti c a l Met h od t o Pred ic t Sk iddi n g i n Thru s t - Lo a ded , An gu l ar -Con tact Ba l l B ea _lngs . Pap e r ' l O- L ubS _ ' r , A S M E , M a y 1 97 0 .

18. Coe , Haro l d H . ; Par k er , R ich ard J. I and S c t b b e , He r bert W. : Ev al u a t ion o t Elec tr o n -Be a m W01ded Hollow Bai ls for H igh- Speed Bal l B earin g s .

Paper 7 0-Lul l -IT, AS M E , May 1 9 _ 0.

19. Coe e Ha rold H . ; S c lbb e_ H erbert W. ; a nd Parker, R i chard J. : P e rfo rm ance o f 7 _ - M llli tnet e r- Bore Bearin g s to 1 . 8 M i ll ion DN Wi th E l ect ro _. - Be am- We l de d Ho ll o w Bal l s. N A SA TN D- 8 800, 1 9 7 0.

20. Coe , H. H . ; S c i bb _ , H. W. ; and A n der s on , W. J. : Evalu a t i on o f Cyl i u _ drically Hollow _ Drt ll ed) B a l ls in Bal l B e ' _ rtn g s a t DN V _u e _ t o 2.1 Mill io n .

N A S A TN D-7007 , 19' / 1.

21. J oh t _o n , R o b e rt L . ; a n d Ludw ig , L aWrence P . : Shaf t Face Seat Wi t h Se t[- A cting L i ft Au gm ent a tio n / or A d v anced Gas Turbine Engines. NASA TN D- 6 170, 1969.

22. Ludwig, La w r e nce P . ; atrom, Tho m as N. ; Al l e n , Gordon P. ; an d J uhnso_ , Robe rt L. : I m provin g P e r f ormanc e o f Face Contact Seal in Liquid S o diU m (400 t o 1 000° F) by Incorporat i o n of Spinal-Groove Geo m et r y. NASA 'I N D-39 4 2 , 196 '/ .

23. Ludw i g _ L. P. ; and Johnson , R. L. : Design S tudy o f Shaft Face s ea l With Sel l -Actin g Lift Augmentation , HI - Mechan i cal De s ign. NASA TN D-6164 , 1971.

2 4. Zu k , Jo hn ; LUd w i g , Lawr e nce P . ; an d J o hnson , R o bert L. : D e s ign Study of Shaft Face S eal W ith Self-Acting Lift Aug m entation. I - Se l f- A cting Pad [.

Geo m etry. NASA TN D - 5' / 44, 1970.

25 . Zuk, J oh n; Ludwig, Lawrence P. ; an d Johnson , Rob e rt L. : Design Stu dy o f t S h aft Face Se al With S _ lf -A cti ng Lift A ug m entation. H - Sealing Da m . , .

NASA TN D-7006, 1 9 '/ 0.

4 4 4 " i . L _..L _ t , - -- HH II - \ t , a ? ' ADVANCED ENGINE BEARING '" RE O UIREMENT8 BE A RIN G 3 . . ' . " _ : , _ DN " :!

i, T P , i,_

8aRVN G n_ ii

200 _ , i 1 97 0 : ' _ l ,

:!

c s - 5 6 e_ 8 i{ F i gure Xlll-I i , MAINSHAFT SEAL ENVIRO N MENT SLIDING AIR 4 _ 0 i, SPEEDo - PRESS, r FT I S EC PSI 200 _ _. .. _L___ 1970 1970 18 00 " ' TEMP, " oF _; A IR 1, _ X) _j , ' 1970 " TIME _ , , F i gur e X 1 11 -2 cs . s_n_9 4 '4 5 f E L A I_ TOH Y DROD Y N AM IC LUB RI CA TIO N F F DRY CONTACT LU B R I CATED C O NTACT i.

F i g ur e XlII-3 c_ . _ ¢, _6 T YP I CAL FA T IGUE NPALL

iJ

EFFECT OF EHD LUBRICATION ON B URFACE DAMAGE TO BEARING RACE 8 (A)NORMAL RACE APPEARANCE RESU L TING FROM FULL EH D FILM (B)RACE G L AZING R ESU L TING FR O M MAR G IN, _ _ HO FIlM (C)GLAZING AN D SUPERFICIAL PITTIN G (D)GROSS PLASTIC DEFORMATION ORSMEARING ( C ) ' ; , ' (, _" " F fg u r _ XII[ 5 ' 4 47 , L : , v. " .......

\ BEARING FAT I GUE LIF E AT k 2 $ °F IN A I R 12- 1 0, 5 8 - RELAT I V E L I FE 6 , 0 iF,. o , AF E MA ADVANCED SYNTHETIC" cs. s( , _ , _3 PREDICTED ES TER PARAFFINIC i .

LIF E OIL t ' F ig u r e Xl II - 6 , i TYPI C AL FATIGUE SPALLS ON BEARI N G BALLS WITH S Y NTHETIC PARAFFINIC OIL ' _ FI , gur e X I | I - 1 ' 448 . " _!

H

UNFAILED BEARING RU N WITH SYNTHETIC PARAFFINIC OIL .50 0 HR, _ T 6 0 00 F; I N ERT ENVIRO N MENT !

6 1 .. $ i .... i Fig ur e XIII- 8 BEARING FATIGUE LIFE AT 6OO°F 1 7, .6 _ 12-- 8-- t RELAT N E LIFE 6 -- .

4- 3 .0 2 -- 1.0 0.75 I I

I I

o l--I

AFBMA POLYPHENYL FLUORO " SYNTHE T IC : PREDICTED El'HER, CARBON , PARAFFINIC cs . 5 6 8_ 4 LIFE AIREN " I N ERT EN" OI L , I N ERT VIR(3NMENT VIRONMENT ENVIRO N ME N T Figure XlII - 9 !

t E FFECT CF VARIOU S M A TERIALS ON CA GE WEAR IN AN INERT EN VIRONMENT A TAM B IENT TEMPERATURE OF5(] 0 0 FWITH SUPERR ff lN E D NAFH T HENIC MINERAL OIL HARDNESS, ' ....

ROCKWEL L A AI S I M - ! gl I( -- " AMS'4 892 6 7 ._ AIS I _ S S 77 C O B A LT AL LO Y 76 8-- C O PPER A L LOY 4 - 9, 5 RE LATIVE AVERAGE 6- WEA R 4-- oi i , _ AISI A N TS" MODIFII:D AISI POLYIMIDE C OBALT COPPL ' R i M'I 48 92 440C STAINLESS POLYMER AL L OY ALLOY I S T E EL i j %

rig .r e x m-l o _ , !

EFFECT OF VARIOUS MATERIALS O N CAGE WEAR LI I N A N I NE RT E N VIRO N ME N T AT AMB I ENT TEM PERATURE OF70 0 o FWITH 5P4E POL Y P HEN YL ET H E R _ t ¢ 12-- (I f _ :: : : t o - t i ! ,

, _ ili

RELAtiVE AVERA G E 6- WEAR 2_ _

N '

AISI A N TS- M O DIFIED AISI' POLYIMIDE COBAL T COPPER M-I 4 892 4 40(; STAINLESS PO LYMER ALL O Y ALLOY STEEL q Figure XIII-Ii BEARING LIFE WITH THREE S TE E L S AT 6OO° F WI T H SYNTHETIC PARAFFINIC LUBRICAN T , INERT ENVIRONMENT 14 - 13 . 6 1,!- 10-, RELATIVE 8- LIFE 6 .7 6- 4- 2- L0 o /.'. I , o.4.

AFBMA WB4 9 A I S I M -I A l Sl M -50 P R EDI C TED- I , LIFE Fi gu re Xill - 12 EFFECT OF SPEED ON LIFE L _ O-MM BALL BEARING LOAD , 100 000r - x l000 .50 000 B A LL WEIGHT t col .uc , .o ,, LIFE , J

. 1 . 5 _ . 5 3.5_S x l o 6

SPEEO, O N cs.s 6 8sz Figure XII| - 13

t t

4 61 CROSS SECTION OF HOLLOW BALL P CS-56868 Figur e XIII - 1 4 4 5 2 \ DRILLED BALL BEARING INNER RACE-,__ OUTER RACE -.

, FigureXIII - B c s . s 6 s6o _ DRILLED BALL BEARING ti 4.7 HROF OPERATION I !: I: t i nk_il ; t i Figure XIII - 17 \ SERIES HYBRID BEARING B MRIN O --,' _ C _R t F U _ i _ ] BALL OILFEEO __,_i , _i i : ' I I

\ /

_ EN G INE _ M ' _ $HA _" : '__ _': . THRUST ' . __ v LOA0 Figur eX HI -18 SER I ES HYBR I D BEAR I NG i TEST RESULTS , - N I "NS / '/ _ O '

" ' L _ - /// o ° ° I "

INNER RACE _ ' SPEED , _ .0_ 0 0 0 • _ - ' / o , , O°° / t I ,I _1 /

O . 5 L0 L 5 2 . 0 2._ x l O 6

SHAF T SPEED , NS,DN C 5-56847 Figu r e XII I -19 4S4 • t EFFECT OF SPEED 150 -MM B A LLB EAR I NG W IT H5 000 - LBL OAD 5 000 - r SERIES HYBRID WITHDRI LL ED BALL HR W r REDUCTION !

100 - " S O L ID BA LL .

50 I I I I I I l.._ 2.5 . _ . 5 4 .5xl0 6 SP EE D, DN cs - s G ss 7 FigureXI I I - 20 ENGINESCHEMATIC . . F COMPRESSOR / E N DSEAL _ . + • AIR . _ _ / _ f -- _-__ j ' . F L O W _t B URN E R_ F , FRONT \ ' RI NG / "-THRUST I _ " - COMP R ESSOR BEARI N G _ S EA L , ' " BE ARING I , I DISCHA R GE CIRCUMFERENTIAL ! _ PRESSURE S EA L R EA R B EARING J L FACE S EA L -+ . +" CD - 1 0Z9Z Figu r e XZZ I -2I cs-s Gs_ LABYRINTHSEAL SU MP COMPRESSOR PR E _SSURE PRESSURE P].

, , , # m LABY RI N TH S E AL S Y S T EM SCH EM ATIC SU M P FAN VE N T C O MPRESSO R .

PRESSURE P R ESSURE PO' DISC H A R GE . -_:_.

II 1 ItI I 11 ti t i li t : . ' : l,,. :, l * % ** t t t % * t t e t % r lit o o * , ! ,:,=: : : i : I , !, 1 , ; , ',; t • • t % o% %, % %% %%, .: , : ,:, :.: , , II11t11111 _ 11 10 1 I _11t I !11 I ; . : ,.: : .,,, 1,*. , 1 . ,1, . .

,%% % , % ;.:,:,:..,,, , . . . . _ i !i ' :'""":"' ! i ITtlRBI N I ; ) tI1 I!IIIII_ 0 :ii!! : _ :! :::: : ::::. :: : .

( _ , . - .; . . :' : '' : .: **. _: , _,___ . :.:::: '.. : .. : ,...,,,. 1 ,: ,.

:,:.,...;.,:, . _ 1 " : ': ' " ' " ':: ...... . ...... , ...,,,. : , : : .......

-( " ( 1' / - ' : ':' " 1' "" ":'"'":" ":'":'"'"_':'"'__'"'"" " : ':'"":":_il J!i!i i ..... :"* : Flgu re XIII - 24 SELF-A C T I NG FA C E S EAL SUMP C OMPRESSOR I PRESSURE DISCHAR G E , , ',I / HEA D AS S EM B L Y '-A _ P I " . _ PO _l_ PR ESSU RE !i

" I sP,, , No

_ .:...::.:.:.:. :: :_ .:.:.:.::.:.:.:..:: :_ .:.:...::.: _ :.:.:..:.:.:.... _ .:.:.:. : :. _ :.:.

:::: : : : : : :::::::: :: ::::: : :: : :: : ::::: ::: ::::::: : :::::::: : :::::::::::: . : . _ . _ . ._ . _ .:._..:. _:_ ._ a ._.

SHAFT P ISTON R IN G " |i_'i'iiii

, . _ _ . _ ......... , '. : !!_

. : " _ ' '_ CS-56861 Figu re XII I- 25 t \ I

Ii-

SELF - ACTING SEAL WITH PAD S ON PRIMARY RI N G A N ER !

I L _S HALL O W RECESS PAD _ GROOVE _ / ( S ELF- A CTING GEO M ETRY) ,7 FEED S L O T _ RESILIENT RING VIEW A°A Fig u re X III - 27 SELF - ACTING SEAL WITH MECHANICAL, PNEUMATIC, A ND SELF - ACTING FOR C ES ACTING ON PRIMARY RING PR IMAR Y RI N G POSITIONS ,,- INITIAL , / ,_ "DISPLACED / SELF-ACTING FORCE MECHANI C AL CLOSING /- D ISPLACED POS ff lON t ' _ " I N ITIA L P OSITIONPO F ORC E (SPRINGS) PNE UM ATIC' I SELF- / A C TIN _ __ _ , I, .

% PO Fi gur e X III-_ l • 4 S 9 i_i , \ C AL C ULATED LOAD C APACI T Y OF L I FT PAD PORTION O F SEAL NUMBER O FSTEPPADS , 20 ; ,ABSOLUTE PRESSURE, 315PSII GAS TEMPERATURE, 130 0 0 F; SLII)ING SPEED, 50I I FEE[PERSECONr } ; PARALLEl. SEALING FACES I00 -- _0_ I , OADC A PAC ITY i O FL I FT PA D , 90 _ i L B 4 0= 60_- 1 2 0- _, 1 0 - 30-- 8 lOxlO GAPHEIGHT, h, IN. "i Figu re XIII-29 !i F AIR LEAKAGE F O R 1 2 0- HO U R E N DURA N CE RU N j _ I SLID ING SPE E D , 400 FEE T PE RSE C O ND; SEALED PR ES S URE , 2 15 P S IA{ S EAL ED TEM P ERA T U R E , t O000 F ..

2 0 - SE AL ISL ST AN D A RD LEAKAGE, __t__i_111_11/_ _ CU FT I MIN10 ' IO0_,IN C REASE IN i ....

R E A R BEA R I NG F A I L URE -- 3 _ o IN C R E ASE IN R E AR IN DI C A T O R R E A D INg ..

• B E AR I N G FAI L URE IN D I C A TOR RE A D IN G .-. I ' I /

I i i i_ _i I ! I I i i I " i I __ 1

0 8 1 6 2 4 32 40 48 " 56 6 4 ? 2 80 88 9 6 1 04 11 2 120 F igure XlII - 3 O _ ; END UR ANCE , H R ] 4 6O .... _"1 I i i " , i , , - - - i -- { - • - , ..... • • ........

r -- -- = --

\ F i gure XII I -3 3 SEAL PERFORMANCE COMPARISON 4 00FT / SEC RELATIVE F WEAR RATE i, i RELATIVE • LEAK RATE 5 1 . 0 _ 0 _ I :' RUBBIN G LABYRIN T H SEL F " !.

FACE ACTING !

SEAL FACE i , SEA l.

• . : c s - 568 so Figure X II I- _ "' I;.

t

), ' I i t [ i 4 6 3 i /: , . IL + i " 2

"' RESEARCH AND DEVELOPMEN I " C O NTRACT O RS ANDGRANTEES

I These contractors arid grante es efforts ar e supplementing NASA in-hou s e pro- _ grau,s i n aircraft propulsion res e arch and techno l ogy: Companies: Lducational and Research Institutions: AeroJet Ge n eral Corpor a tio n University o f Arizona AiResearch Man _ actur i n g Co. Battelle Memorial InStitute _ :_ Bendix Co rp oration California Institute of Technology , Boeing Compa n y Case Western Reserve University Bolt, Be r anek, and Newman, Inc. Chico State College de ttavilla n d Aircraft of Canada Ltd. Univer s ity of Cincinn _ tti i_ i E. I. du Pont de Ne m ours & Co. , Inc. Cle ms o n University i P airchil d -I _U l e r Corporat i on Ge orgia Institut e of Technology i Garrett Co rp oratio n Unive _ slty of Ill _ ois i Ge nez _ tl Electric Company Johns Hopkins UniverSity i L i Ge neral Motors Co m pany Massachusetts Institute o f _ Gruman Aer o Space Co rp . Tec hn ology I!_ ' = Indus t ri al Tectoni v s , Inc. Univers i ty of Minnesota il K opp er s Co m pany, In c. Newark College of Engineering I . !

_ : Marltn - Re ckw e ll Co. Northweste r n Unlver _i ty I!

J Martin - M _ rietta Corporation University of Oklahoma t Mecha ni c al Technology, Inc. Pr ince to n U ni ver s ity I Mobil Rese a x _ h & Development Ren s selaer Polytechnic Inst itu te . , C o rporatio n Stanford U ni verS i ty _ _ ' Monsanto Re search Co rp oration Un i ver s i ty of Southampton , i .

North American Rockwell Co rp oration England j / : Northern Re search & Engir le ering U ni ve rs ity of Te nnes see i ....

_ " Sh e ll Devel op m en t Comp an y I S lop dustries , In c .

• U ni t _ i Ai r c r aft Co rp ortttion i + : !

. ' ..... / • +: . U ni on Carbide Corporation i ; _ 4 6 4 m u m . :_ : , ,, + t --- u , -+ -+.u, , _-_ r P ..........