' I

IONAL ADVISORY COMMIT

*FOR AERONAUTICS

NACA CONFERENCE ON * '

AIRCRAFT LOADS,
STRUCTURES,' AND FLUTTER,
@
MPILATION OF EP
' APERS PRESENTED

Langley Field, Virginia

 NACA CONFERENCE ON

AIRCRAFT LOADS, STRUCTURES, AND FLIJTTEB

A Compilation of the Papers Presented

Langley Aeronautical Laboratory

Langley Field, Va.

March 5 ) 6, 7, 1957
 ‘i -”

TABIX OF COND3NTS

INTRODUCTION

LIST OF CONFEREES

TECHNICAL PAPEW PRESENTED

AIRCRAFT LOADS

Session Chairman: P h i l i p Donely

Loads on Liftim Surfaces and Bodies

Factors Affecting Loads a t Hy-personic Speeds . . . by Arthur L/

Henderson, Jr., and Mitchel H. Bertram

Span Loadings Due t o Wing Twist a t Transonic and Supersonic Speeds /

&

... by Frederick C. Grant and John P. Mugler, Jr.

Flight Measurements and Calculations of Wing Loads and Load Distribu- /
t i o n s a t Subsonic, Transonic, and Supersonic Speeds by Frank ...
S. Malvestuto, Thomas V. Cooney, and E a r l R. Keener

Air Load Distributions on a Flapped Wing Resulting From Leading-Edge

/
and Trailing-Edge Blowing ...
by H. Clyde McLemore

Calculation of External-Store Loads and Correlation With Experiment //

... by Percy J. Bobbitt, Harry W. Carbon, and Albin 0. Pearson ’

Steady Loads Due t o J e t Interference on Wings, Tails, and Fuselages /

L

at Transonic Speeds ...by John M. Swihart and Norman L. C r a b i l l

Effects on Adjacent Surfaces Fromthe Firing of Rocket J e t s . . . by /

Walter E. Bressette and Abraham Leiss

Loads on S t a b i l i z i n g and Control Surfaces

Influence of Automatic Control of Roll Coupling and Pitch-Up on T a i l /

Loads ... by Ralph W. Stone, Jr.

Aerodynamic Loads on Tails a t High Angles of Attack and Sideslip /

... by J. Richard Spahr and Edward C. Polhamus

Loads Due t o Controls a t Transonic and Law Supersonic Speeds . . . by /

i,

F. E. West, Jr., and K. R. Czasnecki

iii
i
1 . 2
- *
G u s t Loads

Experimental Results on Wing Loads Due t o Blasts . . . by Harold B. -i,

/

Pierce and Donald R. McFarland I ’ n

Airplane Motions and Loads Induced by Flying Through t h e Flow Field yl’
Generated by an Airplane a t Low Supersonic Speeds ...
by Gareth
H. Jordan, E a r l R. Keener, and Stanley P. Butchart
i
Effects of Airplane F l e x i b i l i t y on Wing Bending Strains i n Rough Air b
... by Thomas L. Coleman, Harry Press, and C. C. Shufflebarger

Loads Implications of Gust-Alleviation Systems . . . by W i l l i a m H. L ‘

Phillips

Ground Loads
/

. . . by
L,
Recent Data on E r e F r i c t i o n During Landing Sidney A. Batterson

A Summary of Ground-Loads S t a t i s t i c s...

by John R. Westfall, Benjamin
i‘

Milwitzky, Norman S. Silsby, and Robert C. Dreher

STRUCTURES

Session Chairman: Richard R. Heldenfels

Aerodynamic Heatina

Aerodynamic Heat Transfer t o Wing Surfaces and Wing Leading Edges ‘v

/
... by Aleck C. Bond, W i l l i a m V. Feller, and W i l l i a m M. Bland, Jr.
,,/
Total Heat Transfer t o Blunt-Nose Shapes With hminar Boundary Layers ’
a t High Supersonic Speeds ...
by John 0. Reller, Jr.
/’
Local Heat Transfer t o Blunt Noses a t High Supersonic Speeds . . . by
W i l l i a m E. Stoney, Jr.

Heat Transfer t o Bodies a t Angles of Attack . . . by W i l l i a m V. Feller

Heat Transfer i n Regions of Separated and Reattached Flaws . . . by 11

/

Davis H. Crawford and Charles B. Rumsey

Structures a t Elevated Temperatures

F l i g h t Measurements of Airplane S t r u c t u r a l Temperatures a t Supersonic ’

Speeds ... by Richard D. Banner

3 iv
 Two Factors Influencing Temperature Distributions and Thermal Stresses /
i n Structures ... by W i l l i a m A. Brooks, Jr., George E. G r i f f i t h ,
and H. K u r t Strass
n

Effect of Transient Heating on Vibration Frequencies of Some Simple /

Wing Structures ...
by Louis F. Vosteen, Robert R. McWithey, and
i/

Robert G. Thomson

Effects of Rapid Heating on Strength of Airframe Components . . . by I / /

Richard A. Pride, John B. H a l l , Jr., and Melvin S. Anderson

The Combinations of Thermal and Load Stresses f o r the Onset of

Permanent Buckling i n P l a t e s ...
by George W. Zender and Richard
u /
A. Pride

Some Expr-iments With Insulated Structures ... by Richard J. i-/

Rosecrans, Aldie E. Johnson, Jr., and W i l l i a m M. Bland, Jr.

Some Research Results on Sandwich Structures . . . by Melvin S. c //

Anderson and Richard G. Updegraff
Recent Research on the Creep of Airframe Components . . . by Eldon E. w-
.0

Mathauser, A v r a h a m Berkovits, and Bland A. Stein

A Relation Between Stress, S t r a i n Rate, Time, and Temperature f o r

I .

Metals a t Elevated Temperatures ...

by Elbridge Z. Stowell and
George J. Heimerl

Reactions of Materials i n High-Temperature Air Flaws . . . by Joseph

G. Thibodaux, Jr., and Joseph N. Kotanchik

Fatigue-Crack Propagation and Residual S t a t i c Strength of B u i l t - V p

Structures ... by Herbert F. Hardrath and Richard E. Whaley

Some Aspects of Fail-safe Design of Pressurized Fuselages . . . by

Paul Kuhn and Roger W. Peters

Results From Random-Loading Fatigue Tests of a Full-Scale Transport

Wing ... by John B. Garvin

Studies of S t r u c t u r a l Failure Due t o Acoustic Loading ...

by
Robert W. Hess, Robert W. Fralich, and Harvey H. Hubbard

t
V
 FLUTTEZ? AND BUFFETING

Session Chairman: Arthur A. Regier

Buffeting
The U s e of Wind Tunnels t o Predict Flight B u f f e t Loads . . . by Don D. J
Davis, Jr., and Wilber B. Huston

Effects of Wing and Fuselage Modifications on Buffeting . . . by Fred J

B. Sutton and J. Walter Uutenberger, Jr.

Flutter
Theoretical and Experimental Investigations of Delta-Wing Vibrations v/
... by Edwin T. Kruszewski, Eldon E. Kordes, and Deene J. Weidman

Oscillating Air Forces and a Presentation of Some F l u t t e r Calculations

... by Charles E. Watkins, Donald J. Woolston, and Herbert J.
J
Cunningham
J
F l u t t e r a t Very High Speeds . . . by H a r r y L. Runyan and Homer G. Morgan

F l u t t e r of Wings With and Without External Stores a t Transonic and J

Supersonic Speeds ... by Laurence K. Loftin, Jr., and W i l l i a m T.
Lauten, Jr .
Aerodynamics of Oscillating Control Surfaces a t Transonic Speeds J
... by Robert F. Bmmpson and Sherman A. Clevenson
J
Status of F l u t t e r of F l a t and Curved Panels . . . by Robert W. Leonard
and John M. Hedgepeth
F l u t t e r and Divergence of Rectangular Wings of Very Law Aspect Ratio J
. . by Robert W. Fralich, John M. Hedgepeth, and W. J. Tuovila

F l u t t k r Experiments With Various Control Configurations . . . by L/

Robert W. Boswinkle, Jr., and Homer G. Morgan

vi
 This document contains reproductions of t e c h n i c a l papers on some
of t h e most recent research r e s u l t s on a i r c r a f t loads, f l u t t e r , and
structures from t h e NACA laboratories. These papers were presented by
members of t h e staff of t h e NACA laboratories a t t h e Conference held a t
t h e Langley Aeronautical Laboratory March 5 , 6, and 7, 1957. The p r i -
mary purpose of t h i s Conference w a s t o convey t o contractors of t h e
m i l i t a r y services and others concerned w i t h the design of a i r c r a f t these
recent research results and t o provide those attending an o p p o r t ~ i t y
t o discuss the r e s u l t s .

The papers i n t h i s document a r e i n the same form i n which they w e r e

presented a t t h e Conference i n order t o f a c i l i t a t e t h e i r prompt d i s t r i -
bution. “he o r i g i n a l presentation and t h i s record are considered as
complementary to, rather than as s u b s t i t u t e s f o r , the Cornittee’s more
complete and formal reports. Accordingly, i f i n f o m t i o n from t h i s
document i s u t i l i z e d it i s requested t h a t t h i s dacument not be l i s t e d
as a reference. Individual reports dealing w i t h most of t h e information
presented a t t h e Conference w i l l subsequently be published by NACA and
w i l l therefore be s u i t a b l e as reference material.

A l i s t of t h e conferees i s included.

i 6
 LIST OF CONFEREES

c
The following were registered a t the NA Conference on A i r c r a f t
Loads, Structures, and F l u t t e r , Langley Aeronautical Laboratory, Langley
Field, Va., March 5 , 6, and 7, 1957:

ABBOT", Ira H. NACA Headquarters

ABRAHAM, Lewis Subcommittee on A i r c r a f t Structures
ADAMS, Fred T. Naval A i r Development Station
ADAMS, Laurence J. Glenn L. Martin Company
ALBERI, Americo Republic Aviation Corporation
ALEXANDER, Marvin M. Convair
AZLEN, Robert C i v i l Aeronautics Administration
ANBRO, Paul A. Grumman Aircraft Engineering Corp.
. ANDERSON, Enoch A. Boeing Airplane Company
ANDEXSON, Franklin C. Bell A i r c r a f t Corporation
ANDERSON, John W. C onvair
ANDERSON, Melvin S. NACA - Langley Laboratory
ANTILL, Richard B. Raytheon Manufacturing Company
ASHLEY, Holt Subcorrnnittee on Vibration and F l u t t e r

BADGER, David M. Northrop A i r c r a f t

BAILEX, Robert A. Lockheed A i r c r a f t Corporation
BAILEY, Frederick J., Jr. NACA - Langley Laboratory
EAIRD, Eugene F. Subcommittee on Vibration and F l u t t e r
BAKER, Wilfred E. Ballistic Research Laboratory
BANAS, Conrad United A i r c r a f t Corporation
BANNER, Richard D. NACA High-speed F l i g 3 t Station
BARCUS, Ronald Bendix Aviation Research Labs.
BARD, Donald 0. Chrysl e r Corporation
EARKER, LynnM. Sandia Corporation
BARTER, John W. Douglas A i r c r a f t Company
BATTERSON, Sidney A. NACA - Langley Laboratory
MUM, C. Philemon Bureau of Aeronautics
BECKER, John V. NACA - Langley Laboratory
BEEBE, John Subcommittee on Low-Speed Aerodynamics
BEELER, De Elroy NACA High-speed F l i g h t Station
BEEXOVITS, Avraha,m NACA - Langley Laboratory
BERTRAM, Mitchel H. NACA - Langley Laboratory
BEZEATCHEM(0, John W. Goodyear Aircraft Corporation
BIRNBAUM, Sidney G l e n n L. Martin Cmpany
BITZ, Eugene A. Naval Ordnance Laboratory
BLAND, W i l l i a m M . , Jr. NACA - Langley Laboratory
BOBBITT, Percy J. NACA - Langley Laboratory
BOCK, Charles D. Arma Div., American Bosch Arma Corp.
BOGEMA, Bernard L. Glenn L. Martin Company
ix ? 7
BOND, Aleck C.
BOONE, Paul W.
BOSWINKLE, Robert W., Jr,
BOUTON, Innes
BOWMAN, John C.
BRASSAW, Lloyd L., Jr.
BREMER, Paul H.
BRESSETTE, Walter E.
BREUHAUSER, W. 0.
BREWER, Jack D.
BROOKS, Thurman P.
BROOKS, W i l l i a m A., Jr.
BROWN, Clinton E.
BROWN, Harvey H.
BUCKLEY, Edmond C.
CAHEN, George L.
CARLSON, Harry W.
CARLSON, Wendell C . Raytheon Manufacturing Company
CASS, Lorne E.
CATHAWAY, Russ G.
.
CAYWOOD, W i l l i a m C
CENTER, Kenneth W . Grumman A i r c r a f t Engineering Corp.
CHILTON, James P . Douglas Aircraft Company
CLARK, L t . John B.
CLARKE, Martyn V.
CLEVENSON, Sherman A.
COL;EsIAN, Thomas L.
COONEY, Thomas v.
COOPER, Ralph D.
COWGILL, Lee C . Lockheed Aircraft Corporation
COX, Joe F.
CRABILL, Norman L.
CRAWFORD, Davis H.
CREEL, Ralph
CROWLEY, John M.
CROWLIE, John W.
CULBERTSON, P h i l i p E.
CUNNINGHAM, Herbert J.
CYPHERS, Richard J., Jr.
CZARNECKI, K. R.
DAUM, Fred L.
DAVIS, Don D., Jr.
DAWSON, John R.
DEARING, David
DEEP, Raymond A.
NACA -
Langley Laboratory
Subcommittee on A i r c r a f t S t r u c t u r a l
Materials
NACA - Langley Laboratory
Northrop A i r c r a f t
Ryan Aeronautical Company
Bell A i r c r a f t Corporation
Lockheed A i r c r a f t Corporation
NACA -Langley Laboratory
Cornell Aeronautical Laboratory
NACA Headquarters
McDonnell A i r c r a f t Corporation
NACA -Langley Laboratory
NACA - Langley Laboratory
NACA Headquarters
NACA -Langley Laboratory
G l e n n L. Martin Company
NACA - Langley Laboratory
Lockheed A i r c r a f t Corporation
c
Lockheed A i r c r a f t Corporation
Applied Physics Laboratory
A i r Research and Development Command
C i v i l Aeronautics Board
NACA - Langley Laboratory
NACA - Langley Laboratory
NACA High-speed F l i g h t S t a t i o n
David Taylor Model Basin
Cessna A i r c r a f t Company
NACA - Langley Laboratory
NACA - Langley Laboratory
Bureau of Aeronautics
Office of Naval Research
NACA Headquarters
Convair
NACA - Langley Laboratory
Grumman A i r c r a f t Engineering Corp.
NACA - Langley Laboratory
Subcommittee on Vibration and F l u t t e r
NACA -
Langley Laboratory
NACA -
Langley Laboratory
Beech A i r c r a f t Corporation
Redstone Arsenal
DeHART, Robert C. Armed Forces Special Weapons Project
DeLANCEY, Lawrence M. Naval Ordnance T e s t Station
DEPTULA, Capt. Alfred R. Office Deputy Ch. of S t a f f ,
Development, USAF
DEUTSCKMAN, Jerome N. Bell A i r c r a f t Corporation
D I E M , Capt. Walter S. USN ( R e t . ) Committee on Aerodynamics
DOETSCH, E. K. Wright Air Development Center
DOLEN, W i l l i a m K. Sandia Corporation
DONELY, P h i l i p NACA - Langley Laboratory
DOTY, Ralph John Boeing Airplane Company
DOUGHEBTY, James E., Jr. Subcommittee on A i r c r a f t Structures
DOW, Norris General E l e c t r i c Company
DRALFY, Eugene C. NACA - Langley Laboratory
DRYDEN, D r . Hugh L. NACA Headquarters
DUBERG, John E. Aeronutronic Systems
DUKES, Wilfred H. Bell A i r c r a f t Corporation
DUNCAN, L t . Cmdr. B. B. Bureau of Aeronautics
DUNN, Maurice B. Boeing Airplane Company
DUSSAULT, John Cessna Aircraft Company
DYE, L t . F. E., Jr. Bureau of Aeronautics

EXREMY C l i f f o r d 0.
ENTZ, P h i l l i p H.
EPSTEIN, Albert
ERICKSON, A l b e r t L.
ERTHAL, John F.
EVANS, Albert J.
EVEIETT, Maj. P h i l l i p E.
Convair
Boeing Airplane Company
Subconunittee on A i r c r a f t Loads
NACA -Ames Laboratory
Naval A i r Material Center
NACA Headquarters
AF Development F i e l d Representative

FALABELLA, Gaetano, Jr. Air Force Cambridge Research Center

FANTI, Roy United A i r c r a f t Corporation
FAVOR, Ronald J. Battelle Memorial I n s t i t u t e
FEDZIUK, Henry A. NACA - Langley Laboratory
FELLER, W i l l i a m V. NACA - Langley Laboratory
FERNANDEZ, Jose Convair
FETTIS, Henry E. W r i g h t A i r Development Center
FINK, Martin United Aircraft Corporation
FISCKLZB, Jerome E. Douglas A i r c r a f t Company
FITZGERAIJ), Fred Radioplane Company
F L Y W , Richard W . Sperry-Utah Engineering Lab.
FORMHALS, Edwin J. Bureau of Aeronautics
FOSTE8, Charles R. Subcommittee on A i r c r a f t Loads
FRALICH, Robert W. NACA - Langley Laboratory
FRANCEX, Capt. R. G. A i r Force Special Weapons Center
FRICK, Charles W. Convair
FROST, Richard C. Convair
FURLONG, G. Chester Arnold Engineering and Development
Center I -
,
I 9
GARCIA, Manuel A.
GARRICK, I. E.
GARVIN, John B.
GATES, Ordway B. Jr. ,
G A W , William
GENIESSE, Major Eugene W., Jr.
GEUDTNER, Walter J., Jr.
GILLIG, Franklin J.
GILRUTH, Robert R .
GIONFRIDDO, Maurice P.
GLASSMEYER, C l i f f o r d E., Jr.
GLOVER, Louis S.
GOLDIN, Robert
GOMZA, Alexander
GORANSON, R . Fabian
GRANT, Frederick C.
GREEB, Major Edwin H.
GRIFFIN, Edward J.
GRIFFITH, George E.
GRIFFITH, Max 0.
GRINSTED, Frank
Naval A i r Missile T e s t Center
NACA - Langley Laboratory
NACA - Langley Laboratory
G l e n n L. Martin Company
Radioplane Company
Subcommittee on High-speed Aerodynamics
Convair
Cornell Aeronautical Laboratory
NACA - Langley Laboratory
Air Force Cambridge Research Center
B a t t e l l e Memorial I n s t i t u t e
Applied Physics Laboratory
Subcoinnittee on A i r c r a f t Loads
Grumman A i r c r a f t Engineering Corporation
NACA Xeadquarters
NACA - Langley Laboratory
Arnold Engineering Development Center
Bureau of Aeronautics
NACA - Langley Laboratory
Bureau of Aeronautics
B r i t i s h Ministry of Supply

HALL, Bertrand M. Douglas A i r c r a f t Company

HALL, John B., Jr. NACA - Langley Laboratory
HALSEY, Robert M. B e l l A i r c r a f t Corporation
HAMILTON, W i l l i a m T. Subcommittee on Aerodynamic S t a b i l i t y
and Control
HAMMERBERG, Fritchiof C i v i l Aeronautics Administration
HAMMIL, John P. Douglas A i r c r a f t Company
HARDRATH, Herbert F. NACA - Langley Laboratory
HARRINGTON, Joseph H. Subcommittee on A i r c r a f t Loads
HARRIS, Thomas A. NACA - Langley Laboratory
HARTLEX, Richard M. David Taylor Model Basin
HAVILAND, John K. Chance Vought A i r c r a f t
HAYDEN, Harold J. Boeing Airplane Company
HEDGEPETH, John M. NACA - Langley Laboratory
HEIMERL, George J. NACA - Langley Laboratory
HEITHECKER, Heinrich A. Holloman A i r Development Center
,
HELDENFELS Richard R. NACA - Langley Laboratory
HELLEBRAND; E m i l A. H. Subcommittee on A i r c r a f t Structures
HENDERSON, Arthur, Jr. NACA - Langley Laboratory
HENDERSON, Campbell North American Aviation
HEPPER, Richard H. McDonnell A i r c r a f t Corporation
HESS, Robert W. NACA - Langley Laboratory
HOFF, D r . Nicholas J. Subcommittee on Aircraft Structures
HOFFMAN, Niles R. Wright A i r Development Center
HOGE, H. J. Subcommittee on Aircraft Loads
HOLTZ, John R . David Taylor Model Basin .
HOOD, Manley J.
HOOVER, Isaac H.
HUBBARD, Harvey H.
HUNT, Melvin W.
HUNTLEY, H. W., Jr.
HUSIC, W i l l i a m J.
HUSTON, Wilber B.
ISWLSE3?, L t . Col. Orson A.
JACKSON, David W.
JENKINS, David R.
JOHNSON, Aldie E., Jr
JOHNSON, H. Clay
JOHNSON, Harry W . Raao-Wooldridge Corporation
JOHNSON, J. Aldridge
JOHNSON, J. B.
JOHNSTON, W i l l i a m M.
JORDAN, Gareth H.
JORDAN, Peter F.
KAECHELE, Lloyd E.
KAMEACK, Elmer L.
KAPLAN, Abner
E;AsTEN, Herman G.
KEARNS, John P.
KEENER, E a r l R.
KEIRSEX, Robert L.
KEITH, Bobbie C.
KERR, T. H.
KINNAMAN, Edward B.
KINTAS, John
KLEXSS, h d r . N. J.
KORDES, Eldon E.
KOTANCHIK, Joseph N.
KRAMER, Edward H.
KRICKENBERGER,
L t . Cmdr. C u t e r F., Jr.
K R O U , Wilhelmina D.
KRUSZEWSKI, Edwin T.
KUGEL, H. E.
KUHN, Paul
KULLAS, Albert J.
KUNZE, Frank C.
LaFRANCE, Jeremie U.,
LAIDLAW, . W i l l i a m R.
NACA - Ames Laboratory
C i v i l Aeronautics Administration
NACA - Langley Laboratory
North American Aviation
Subconrmittee on A i r c r a f t Loads
C i v i l Aeronautics Administration
NACA - Langley Laboratory
b e d Forces Special Weapons Project
Wright A i r Development Center
Batelle Memorial I n s t i t u t e
. Avco Manufacturing Corporation
Subcommittee on Vibration and F l u t t e r
Lockheed A i r c r a f t Corporation
Subcommittee on A i r c r a f t S t r u c t u r a l
Materials
Lockheed A i r c r a f t Corporation
NACA High-speed F l i g h t S t a t i o n
Glenn L. Martin Company
Rand Corporation
G l e n n L. Martin Company
Ramo-Wooldridge Corporation
Ram0 -Wooldridge Corporation
Applied Physics Laboratory
NACA High-speed F l i g h t S t a t i o n
Douglas A i r c r a f t Company
Wright A i r Development Center
B r i t i s h Ministry of Supply
Boeing Airplane Company
Beech Aircraft Corporation
Naval A i r Material Center
NACA - Langley Laboratory
NACA - Langley Laboratory
Fairchild A i r c r a f t Division
Armed Forces Special Weapons Project
National Bureau of Standards
NACA - Langley Laboratory
Wright Air Development Center
NACA - Langley Laboratory
Glenn L. Martin Company
Chrysler Corporation
Jr. Glenn L. Martin Company
Subcommittee on Vibration and F l u t t e r
xiii I
LAMPROS, Alexander F.
LANDEBS, C . B.
LANDES, Paul E.
LANDON, John M.
LAUTEN, W i l l i a m T., Jr.
LAUVER, Dean C.
LECAT, Robert
LEISS, Abraham
LEONARD, Robert W.
LEVY, Samuel
L I , Ta Chung-Heng
LIEBOWITZ, Harold
LOCKE, Frederick W. S., Jr.
LOFTIN, Laurence K., Jr.
L O W , Ted L.
LoPRESTI, Antonio
LU, Hoshen
LUCAS, John W.
MAGRATH, H. A.
MALVESTUTO, Frank S.
MARCOPULOS, George
MARTIN, James E.
MARTIN, John W., Jr.
MARTZ, Ronald B.
MATHAUSER, Eldon E .
MAY, Ralph W., Jr.
MAYE3, Frank S., Jr.
MAZELSKY, Bernard
McCASKILL, A l l a n M.
McCULLAND, Clyde M.
McFARLAND, Donald R.
McGINNESS, Capt. W. T.
McGOWAN, W. A.
McHlJGH, James G.
McLEMORE, H. Clyde
McWITHEX, Robert R.
MELLQUIST, Victor G.
MEREDITH, L t . Oliver D.
MERZ, Ernest J.
ME=, Stephen
MICHEL, Douglas
MI=, Chester W.
MILLER, L t . C m d r . Jack N.
MILWITZKY, Benjamin
MIROWITZ, L. I.
MISKAM, Frederick C.
A A
Naval Air Missile T e s t Center
Lockheed A i r c r a f t Corporation
McDonnell Aircraft Corporation
Chrysler Corporation
NACA -
Langley Laboratory
Bureau of Aeronautics
Fairchild Guided Missile Division
NACA -
Langley Laboratory
NACA -
Langley Laboratory
General E l e c t r i c Company
'
Convair
Office of Naval Research
Bureau of AeronaQtics
NACA -
Langley Laboratory
Boeing A i r p l a n e Company
Subcommittee on Aircraft Structures
Republic Aviation Corporation
J e t Propulsion Laboratory, C.I.T.
Wright A i r Development Center
NACA High-speed F l i g h t Station
North American Aviation
Chance Vought A i r c r a f t
Glenn L. Martin Company
Wright A i r Development Center
NACA - Langley Laboratory
NACA Headquarters
Douglas A i r c r a f t Company
Lockheed A i r c r a f t Corporation
Douglas A i r c r a f t Company
Douglas A i r c r a f t Company
NACA - Langley Laboratory
Subcommittee on A i r c r a f t Structures
Bureau of Aeronautics
Subcommittee on Aerodynamic S t a b i l i t y
and Control
NACA - Langley Laboratory
NACA - Langley Laboratory
A i r c r a f t Industries Association of
America
Air Research and Development Command
Glenn L. Martin Company
Convair
S u b c m i t t e e on Vibration and F l u t t e r
McDonnell A i r c r a f t Corporation
Bureau of Aeronautics
NACA - Langley Laboratory
Subcommittee on Vibration and F l u t t e r
Douglas A i r c r a f t Company
X ~ V. a
MITCHAM, Grady L.
MOL?3L;LA, Roslo J.
MOORE, Lt. Col. John T., Jr.
M O M , Herbert D.
MORDFIN, Leonard
MORGAN, Homer G.
MUGLER, John P., Jr.
MULHERN, John J.
MULTHOPP, Hans
MUNDO, Charles J.
MUNTER, Paul L.
MURRAY, James J.
MYERS, Boyd C., I1
MYKYTOW, Walter J.
NAUGHTON, S. K.
NEELY, Robert E.
NELSON, Lloyd W.
NEWBERGER, Eli S.
NEWBY, Clinton T.
NEWELL, Joseph S.
NOREM, Allan G.
NORTON, David A.
O'BRIEN, R. J.
OFFTERMATT, Lt. Cmdr. Wilbur F.
OGNESS, Arthur M.
O ' M A L m , James A., Jr.
OSBORN, Earl P.
PAIMER, Carl B.
PAYSON, Peter
PEARSON, Albin 0.
PELOUBET, Raymond P.
PERSH, Jerome
PETERS, Roger W . NACA - Langley Laboratory
PETERSON, James P.
PFAFT, George C., Jr.
PHILLIPS, Franklyn W.
PHILLIPS, William H.
PEBCE, Harold B . NACA - Langley Laboratory
PIERPONT, William G.
PINES, Samuel
POLHAMUS, Edward C.
U. S. Naval Ordnance Experimental Unit
Naval Air Material Center
Air University
Aircraft Industries Association of
America
National Bureau of Standards
-
NACA Langley Laboratory
NACA - Langley Laboratory
Naval Air Development Station
Glenn L. Martin Company
Arma Div., American Bosch h a Corp.
Republic Aviation Corporation
Office of Ordnance Research,
u. s. Army
NACA Headquarters
Subcommittee on Vibration and Flutter
Wright Air Development Center
Fairchild Aircraft Division
Lockheed Aircraft Corporation
Civil Aeronautics Administration
Subcommittee on Aircraft Loads
North American Aviation
Aerophysics Development Corporation
Boeing Airplane Company
Wright Air Development Center
Bureau of Aeronautics
Northrop Aircraft
Subcommittee on Low-Speed Aerodynamics
Grumman Aircraft Engineering Corp.
NACA Headquarters
Subcommittee on Aircraft Structural
Materials
NACA - Langley Laboratory
Convair
Naval Ordnance Laboratory
NACA - Langley Laboratory
Glenn L. Martin Company
NACA Headquarters
-
NACA Langley Laboratory
Beech Aircraft Corporation
Republic Aviation Corporation
NACA - Langley Laboratory
PRESS, Harry
PRIDE, Richard A.
RACISZ, Stanley F.
RAPHAEL, Coleman
RARING, Richard H.
RAY, George D.
RAYBURN, Leo
REEVES, Walter R .
REGIER, Arthur A.
REID, Dr. Henry J. E.
REL;LER, John O . , Jr.
RHODE, Richard V.
RICHARDSON, W i l l i a m B.
ROBERTS, W i l l i a m M.
ROBERTSON, Donald K.
ROC&, Jean A.
RODDEN, W i l l i a m P.
ROLFE, R i a l E., Jr.
ROSCHE, Melvin G.
ROSECRANS, Richard J.
ROSEN, B. Walter
ROSENBAUM, Robert
ROTH, A. L.
ROTIWYER, E a r l
ROZELLE, Harold L.
RUMSEY, Charles B.
RUNSTAD, Harold J.
RUNYAN, Harry L., Jr.
SALZBERG, Leo F.
SAUNDERS, John J .
SCRNEIDER, W i l l i a m C.
SCHNITT, Arthur
SCHREIBER, Ralph
SCHUMACHER, Ernest A.
SCHUMACHER, John G.
SCROOC, D. J.
SECKIER, Sam
SEIDE, Paul
SENTKECR, Lawrence C.
SHARPE, Lawrence W.
SHATUNOFF, Stanley
SHAW, s. L.
SHORTAL, Joseph A.
SHOWERS, Nathan
NACA - Langley Laboratory
NACA - Langley Laboratory
B e l l A i r c r a f t Corporation
Republic Aviation Corporation
NACA Headquarters
Committee on A i r c r a f t Construction
Raytheon Manufacturing Company
Convair
Subcommittee on Vibration and F l u t t e r
NACA - Langley Laboratory
NACA - Ames LaPoYatory
NACA Headquarters
Naval Ordnance Laboratory
Arnold Engineering Development Center
Glenn L. Martin Company
AF Development F i e l d Office, NACA
Langley Aeronautical Laboratory
North American Aviation
McDonnell A i r c r a f t Corporation
NACA Headquarters
NACA - Langley Laboratory
Avco Manufacturing Corporation
Subcommittee on Vibration and F l u t t e r
Douglas Aircraft Company
Goodyear A i r c r a f t Corporation
Ryan Aeronautical Company
NACA - Langley Laboratory
Boeing Airplane Company
NACA - Langley Laboratory
Wright A i r Development Center
Glenn L. Martin Company
Bureau of Aeronautics
Bell A i r c r a f t Corporation
Radioplane Company
Glenn L. Martin Company
Convair
Ram0 -Wooldridge Corporation
Bureau of Aeronautics
Ramo-Wooldridge Corporation
C i v i l Aeronautics Administration
A i r Force Cambridge Research Center
Arma Division, American Bosch Arma
Corporation
Douglas A i r c r a f t Company
NACA - Langley Laboratory
Wright A i r Development Center
E
SHUFFI;EBARGm, C. C .
SILSBY, Noman S.
SKIMMIN, Kenneth D.
c SMITH, Randall C .
SMITH, Turner L.
SOCKS,, G l e n n A.
SOULE, Hartley A.
SPAHR, J. Richard
SPAUUIING, E. H.
SP-, Robert F.
STACK, John
STAHL, Harold F.
STAUFFER, Warren A.
STEIGER, Arthur R.
STEIN, Bland A.
STEVENS, Clifton D.
STEVENS, John E .
STEVENSON, C. H.
STONE, Ralph W., Jr.
STONEY, W i l l i a m E., Jr.
STOWELL, Elbridge Z.
STRASS, H. K u r t
STRINGHAM, R. H.
, -S Rodney D.
SUTLIFF, JohnD.
SUTTON, Fred B.
SWANSON, L t . Col. Arthur R.
SWANSON, Theodore B.
SWIHART, John M.
TARGOFF, Walter
'IIARNOWER, Gerald
TETENS, Robert C.
.
THEODORSEN, D r Theodore
THIBODAUX, Joseph G., Jr.
THOMPSON, Floyd L.
THOMPSON, Robert F.
THOMSON, Robert G.
THORITE, Charles J.
THORSON, Kenneth R.
TOPP, LeRoy J.
'IIRACY, Alfred C .
TROHA, Charles C.
TUOVILA, W . J.
UJIHARA, Ben H.
UNDERWOOD, W i l l i a m J.
xvii
NACA - Langley Laboratory
NACA - Langley Laboratory
Wright Air Development Center
Fairchild A i r c r a f t Division
B a l l i s t i c Research Laboratory
Chrysler Corporation
NACA - Langley Laboratory
NACA - Ames Laboratory
Subcommittee on A i r c r a f t Structures
Bureau of Aeronautics
NACA - Langley Laboratory
Chance Vought A i r c r a f t
Lockheed A i r c r a f t Corporation
A i r Force Special Weapons Center
NACA - Langley Laboratory
Goodyear A i r c r a f t Corporation
Subcommittee on A i r c r a f t Structures
Committee on A i r c r a f t Structures
NACA - Langley Laboratory
NACA - Langley Laboratory
NACA - Langley Laboratory
NACA - Langley Laboratory
Douglas A i r c r a f t Company
Convair
Ryan Aeronautical Company
NACA - h e s Laboratory
Office of the Assistant Secretary
of Defense
Aero-General Corporation
NACA - Langley Laboratory
Cornell Aeronautical Laboratory
Chance Vought A i r c r a f t
Temco A i r c r a f t Corporation
Republic Aviation Corporation
NACA -Langley Laboratory
NACA -Langley Laboratory
NACA -Langley Laboratory
NACA -Langley Laboratory
U. S. Naval Ordnance T e s t Station
Boeing Airplane Company
Boeing Airplane Company
Douglas A i r c r a f t Company
Bureau of Aeronautics
NACA -Langley Laboratory
North American Aviation
NACA Liaison Officer,
Wright-Patterson AFB
UPDEGRAFF, Richard G. NACA - Langley Laboratory
UTTER, Emmet Beech Aircraft Corporation

VANDREY, Dr. Friedrich Glenn L. Martin Company

VETERITO, A. Michael
VOLLMECKE, Albert A.
VOLLO, Samuel D.
VOSTEEN, Louis F.
Fairchild Guided Missile Division
Committee on Aircraft Construction
Bell Aircraft Corporation
-
NACA Langley Laboratory

W A L L S , James H. Bureau of Aeronautics

WATKINS, Charles E.
WATSON, Robert E.
WEBB, Everette L.
WEIDMAN, Deene J.
WEINBERGER, Robert A.
WEISMAN, Yale
WEST, F. E., Jr.
WESTRUP, Robert W.
WESTVIG, Roger
WHALFY, Richard E.
WHELDON, Wilbert G . Northrop Aircraft
WHITE, Richard P., Jr . Cornell Aeronautical Laboratory
WILLIAMS, Walter C.
WILLIAMSON, Leonard H.
WILSON, Herbert A., Jr.
WILSON, James N.
WOLCOTT, Verne V.
WOOLSTON, Donald J.
-
NACA Langley Laboratory
Boeing Airplane Company
Boeing Airplane Company
-
NACA Langley Laboratory
Bureau of Aeronautics
Subcommittee on Aircraft Loads
NACA - Langley Laboratory
North American Aviation
North American Aviation
-
NACA Langley Laboratory
NACA High-speed Flight Station
Civil Aeronautics Administration
-
NACA Langley Laboratory
Jet Propulsion Laboratory, C.I.T.
Convair
NACA - Langley Laboratory

YATES, William B. Glenn L. Martin Company

YORK, Joseph E. Lockheed Aircraft Corporation
YOUNG, George E. Northrop Aircraft
YOUNG, Willis H. Bureau of Aeronautics

ZENDER, George W. -
NACA Langley Laboratory
ZISFEIN, Melvin B. Bell Afrcraft Corporation
ZLOTNICK, Martin Avco Manufacturing Corporation

16
xviii
AIRCRAFT
LOADS
 U

"#-Nm'1-75363
" .w
f'

FACTORS AFFECTING LOADS AT HYPERSONIC

By Arthur Henderson, Jr., and Mitchel H. Bertram

Langley Aeronautical Laboratory

SUMMARY

This paper gives a brief summary of current loads information a t

hypersonic speeds. Several methods which t h e designer can employ i n
estimating t h e loads on various a i r c r a f t components a r e discussed. The
paper deals with t h e c h a r a c t e r i s t i c s of both slender and blunt configura-
t i o n s and touches upon the e f f e c t s of boundary-layer and aerodynamic
interference.

INTRODUCTION

The calculation of loads a t hypersonic speeds requires t h e use of

techniques with which many designers are not very familiar. The methods
based on l i n e a r o r second-order theory, which were widely used at super-
sonic speeds, a r e inadequate f o r slender configurations at hypersonic
speeds and, of course, a r e completely inapplicable t o configurations
with blunt noses o r leading edges.

I n t h i s paper it is shown that c e r t a i n simplifying features which

allow good design approximations of loads t o be made with a minimum of
e f f o r t e x i s t a t hypersonic speeds. I n addition, some of t h e unsolved
problems associated with hypersonic phenomena m e pointed out.

SYMBOLS

a speed of sound

A constant

C l o c a l chord length

I? mean, aerodynamic chord

Cn section normal-force coefficient

 2

CP l o c a l surface pressure coefficient

d maximum body diameter

h height of wedge

K hypersonic s i m i l a r i t y parameter, M -
d
1

2 length of nose o r wedge

M Mach number

pressure

radius

Reynolds number

S a r c length

t thickness

v velocity

X distance from nose o r 1ea.ding edge i n body-axis system

U angle of a t t a c k

7 r a t i o of s p e c i f i c heats

6 f l a p deflection angle

A incremental value

E distance between adjacent streamlines

0 cone shock angle

h sweepback angle

a cone semiapex angle

7 time

cp meridian angle
19
a
 3

Subscripts :

02 free-stream conditions

B(W) body i n presence of wing

MAX maximum

S shoulder

t based on thickness

d based on diameter

DISCUSSION

There a r e several methods which the designer can employ i n arriving

a t an estimate of the loads on the various a i r c r a f t components. Before
discussing them, however, it i s i n s t r u c t i v e t o consider, qualitatively,
how hypersonic phenomena d i f f e r from supersonic.

Although hypersonic flow introduces many problems which were not

encountered a t supersonic speeds, it also introduces c e r t a i n simplifying
features; and aerodynamicists have not been long i n taking advantage of
them. For example, one source of simplification a t hypersonic speeds i s
t h e f a c t t h a t , i n the exact shock equations, t h e Mach number term i s
usually squared and often appears i n the denominator. Thus, as t h e Mach
number increases, these t e r m s become insignificant; thus r e l a t i v e l y
simple expressions often y i e l d accurate approximations f o r c e r t a i n flow
properties a t hypersonic speeds.

Slender Configurations

Characteristics of hypersonic flow. - Some simplifying features of

hypersonic flow a r e i l l u s t r a t e d i n figures 1 and 2. One of the character-
i s t i c s of hypersonic flow i s i t s tendency toward two-dimensionality when
i n contact with slender bodies or surfaces. (See f i g . 1.) The upper
half of figure 1 depicts a sharp-leading-edge sweptback w i n g i n a low
and i n a high Mach number flow f i e l d . There a r e two streamlines t h e same
distance E apart. A s shown by t h e dashed l i n e s , t h e f i e l d s of influence
from each disturbance point along t h e leading edge spread across t h e w i n g
i n supersonic flow, whereas they a r e confined t o a r e l a t i v e l y narrow
region i n hypersonic flow. I n addition, t h e r i g h t streamline of each
p a i r w i l l s t r i k e the leading edge l a t e r than t h e l e f t one, the time lag
being AT = E tan *.Ma
Obviously, as t h e Mach number increases, t h e time

$0
4

lag decreases; thus a t high Mach numbers t h e r i g h t streamline s t r i k e s

t h e leading edge a t almost t h e same time as t h e l e f t one. Consequently,
as far as t h e f l u i d i t s e l f i s concerned, it f e e l s as though it i s p r a c t i -
c a l l y two dimensional.

The bottom half of figure 1 depicts t h e fundamental basis of t h e

generalized shock-expansion method as applied t o slender three-dimensional
bodies. A s Eggers and Savin ( r e f . 1) have shown, so long as t h e diver-
gence of streamlines along the body i s negligible, t h e flow on the body
surface and t h e associated flow f i e l d w i l l be e s s e n t i a l l y two-dimensional
i n nature; consequently, two-dimenqional shock-expansion theory can be
used t o analyze the flow about slender bodies of revolution.

Hy-personic s i m i l a r i t y l a w . - The designer has another powerful t o o l

a t h i s disposal i n the form of the hypersonic s i m i l a r i t y law (see, f o r
example, r e f s 2 t o 4), which s t a t e s t h a t t h e pressures at corresponding
points on similarly shaped bodies a r e i d e n t i c a l i f , f o r t h e two bodies,
t h e product of free-stream Mach number and thickness r a t i o i s a constant.

The physical concept behind t h e hypersonic s i m i l a r i t y l a w i s i l l u s -

t r a t e d q u a l i t a t i v e l y i n figure 2. Two marbles a r e shown, each r o l l i n g
toward i t s own wedge. The upper marble w i l l rise a height h i n the
length 2 1 with the velocity VI, while t h e lower marble w i l l r i s e t h e
same height h i n t h e longer length 22 = A 2 1 but with t h e higher veloc-
i t y V2 = AV1. The r a t i o of lengths and v e l o c i t i e s i s such t h a t both
marbles r i s e t h e same height h i n t h e same length of time; t h a t is,
they both experience the same change of velocity and, consequently, each
marble w i l l impart the same amount of momentum t o i t s p a r t i c u l a r wedge.
If the marbles a r e thought of as a i r molecules and the wedges as cor-
responding slopes on two similar bodies, a d i r e c t analogy with t h e hyper-
sonic s i m i l a r i t y l a w i s immediately apparent.

The approximate region i n which t h e hypersonic s i m i l a r i t y l a w i s

applicable has been determined by Lees ( r e f . 5 ) t o be about as shown i n
figure 3 f o r cones. This region i s determined by the condition t h a t the
cone shock angle Os i s l e s s than 24'. Thus, t h e maximum cone angle
f o r good correlation a t hypersonic speeds w i l l be about 20°. Bodies of
revolution such as ogives are e s s e n t i a l l y conical at t h e nose and decrease
i n slopes t h e r e a f t e r . Therefore, i f t h e nose of any pointed body i s
about 20' or l e s s , it should c o r r e l a t e w e l l with t h i s l a w . For ogives,
t h i s means the fineness r a t i o should be about 3 o r more.

Figure 4 presents t h e pressure-ratio d i s t r i b u t i o n on ogives. The

s o l i d l i n e s a r e the c h a r a c t e r i s t i c solutions of Rossow ( r e f . 4), each of
which i s for a t l e a s t two d i f f e r e n t combinations of rJr, and Z/d
within t h e range shown a t t h e lower r i g h t . Although M, = 1 2 was t h e
 5

highest value of used i n t h e calculations, it should be pointed out

t h a t t h i s value is not meant t o be taken as an upper l i m i t . Also shown
m e the tangent-cone approximations of Probstein and Bray ( r e f . 6). For
K 2 1, they applied t h e tangent-cone approximation t o Lees' r e s u l t which
i s f o r the case when the shock l i e s f a i r l y close t o t h e body; and f o r
K < 1, that is, when the shock i s well removed from the surface of t h e
slender bodies, the tangent-cone approximation is applied t o Kd,rmsh's
r e s u l t i n l i n e m i z e d supersonic flow.

Van Dyke has pointed out i n h i s work on the hypersonic small-

disturbance theory (refs. 7 and 8) t h a t t h e range of a p p l i c a b i l i t y of
t h e hypersonic s i m i l a r i t y l a w can be extended t o the transonic range by
replacing the Mach number term with the Prandtl-Glauert s j i n i l m i t y fac-
tor fK. The degree t o which t h i s correlation i s successful i s
i l l u s t r a t e d i n figure 5 f o r cones with semiapex angles of 5O, loo, l5O,
and 20'. I n t h i s figure $/tan2 u is p l o t t e d against \ I F 1t a n u
f o r a Mach number range from 1.15 t o hypersonic speeds. Each curve i s
ended when sonic velocity appems on the cone surfaces. The correlation
i s s e e n t o be excellent.

The correlation f o r b l u f f cones as suggested by Newtonian theory

I
i s presented in figure 6, where Cp sin2 0 i s p l o t t e d against a. For
t h e ranges of Mach number and
pressure on the surface of a bluff cone is cp
sin2 0
-
u shown, a good approximation t o the
2.2.

Shock-expansion theory.- The use of two-dimensional shock-expansion

theory t o predict the pressures on slender bodies of revolution a t zero
angle of attack a t hypersonic speeds i s well known. Eggers and h i s
associates ( r e f s . 1 and 9 ) have shown t h a t , provided conditions a t t h e
nose are known from e i t h e r conical theory or experiment, the generalized
shock-expansion method can be used f o r slender bodies of revolution at
angle of attack.

Figure 7 shows a comparison of the shock-expansion theory d t h experi-

ment f o r an ogival 6ose a t an angle of attack of 150 and a free-stream
Mach number of 5.05. The symbols show the experimental pressure coef-
f i c i e n t s along the top, side, and bottom meridians. The t h e o r e t i c a l pre-
dictions begin with t h e assumption of conical flow at t h e nose. The s o l i d
curves use the t h e o r e t i c a l cone approximation of Savin ( r e f . 10) as t h e
s t m t i n g point f o r the shock-expansion calculations, and t h e dashed curve
uses experimentally determined conditions on the nose cone as t h e s t a r t i n g
point f o r t h e calculations. Obviously, a r e l i a b l e t h e o r e t i c a l method i s
preferable f o r design work. It i s seen t h a t on t h e bottom meridian, which
would be of most i n t e r e s t f o r loads considerations, t h e shock-expansion
6

calculations agree with experiment f o r both the t h e o r e t i c a l l y and t h e

experimentally determined s t a r t i n g .conditions; t h a t i s , Savin's t h e o r e t i c a l
cone r e s u l t s combined with shock-expansion theory w i l l give good desig:i
estimates of maximum loads on sharp noses a t angle of attack.

Effect of Blunting

I n cases where high heat-transfer r a t e s a r e expected, t h e use of

blunt leading edges and noses i s dictated. Figures 8 and 9 present
r e s u l t s f o r both two-dimensional c i r c u l a r cylinders ( r e f L . 11 and 12)
and hemispherical-nose bodies of revolution ( r e f . 1 3 ) . The r e s u l t s f o r
zircular cylinders ( f i g . 8) would be applicable both t o bodies of revolu-
t l o n a t high angles of a t t a c k and t o the leading edges of blunt sweptback
wings. Figure 8 i s e s s e n t i a l l y a double correlation of pressure r a t i o s .
It shows t h e manner i n which the pressure m t i o varies with meridian angle,
and it i s good f o r a wide range of sweepback angles. Also note t h a t , as t h e
Mach number increases, t h e band of experimentally determined pressure
r a t i o s converges toward the t h e o r e t i c a l curve of Goodwin ( r e f . 12) shown
by the dashed l i n e . Penland (ref. 11) has shown t h a t pmax can be deter-
mined on yawed c i r c u l a r cylinders f o r sweepback angles from '0 t o about
75' a t M, = 6.9 by using t h e normal component of &. Thus, t h e abso-
l u t e pressure d i s t r i b u t i o n on t h e windward side of yawed c i r c u l a s cylinders
can be obtained.

The r e s u l t s f o r hemispherical noses i n figure 9 show excellent agree-

ment with Newtonian theory. I n t h i s figure the pressure-coefficient r a t i o
i s plotted against s/r, which i s t h e a r c angle i n radians. As can be
seen, t h e pressure-coefficient r a t i o i s independent of Mach number. The
curve of C P , w against & i n t h e upper r i g h t of t h e figure shows
t h a t , f o r values of & greater than about 4, %,w i s e s s e n t i a l l y
a constant on the order of 1.8. Thus, with t h e a i d of Newtonian theory,
Cp can be closely estimated, and f o r values of M, above about 4, t h e
Cp d i s t r i b u t i o n w i l l be e s s e n t i a l l y invariant with &.

As w a s mentioned previously, many of t h e exact flow parameters can

be closely approximated with simple expressions i n t h e hypersonic-flow
regime. For example, f o r y = 1.4 and Mm >> 1, it can be shown t h a t
t h e r a t i o of free-stream s t a t i c pressure t o stagnation pressure on a
blunt-nose body i s approximately 0.777/&'. ( T h i s r a t i o i s determined

. L

\
 7

i n t h e appendix.) The percentage e r r o r in the t r u e value of p , / ~

incurred by using t h e approximation i s shown by t h e sketch. A t M, = 3,
t h e e r r o r i s only 4 percent and it
decreases r a p i d l y t h e r e a f t e r with
increasing I%,. It can a l s o be
shown t h a t f o r I%, >> 1, the r a t i o
of l o c a l pressure t o maximum pres- % ERROR IN
sure i s approximately equal t o t h e
pa
r a t i o of l o c a l pressure coefficient
t o maximum pressure c o e f f i c i e n t . PMAX
Consequently, t h e r a t i o of t h e l o c a l ~~

absolute pressure t o t h e free-stream 0 2 4 6 8 1 0

s t a t i c pressure i s given by Ma

-P= - Mco2 cp ; t h a t is, a t hyper-

P, 0.777 Cp,&m
sonic speeds, t h e absolute pressure a t any point on a blunt nose i s
d i r e c t l y proportional t o t h e square of t h e Mach number. I n p a r t i c u l a r ,
f o r any given a l t i t u d e , t h e absolute pressure d i s t r i b u t i o n on a hemi-
m

spherical nose i s given by - =

p
pw
-
McQz

0.777
cos2(s/r) f o r 0 6 s/r 6 1 . 3 radians.

I
The f a c t t h a t t h e experimental p r e s s w e s deviate from t h e t h e o r e t i c a l
pressures beyond about 1 . 3 radians i s due t o a combination of entropy,
v o r t i c i t y , and boundary-layer e f f e c t s , which, of course, Newtonian theory
does not include. For t h e Mach numbers considered herein, t h e e f f e c t s a r e
negligible as far as loads m e concerned. As t h e Mach number i s increased,
however, these e f f e c t s become increasingly important.

Figure 10 shows how, as a r e s u l t of entropy gain, t h e surface pres-

sure at t h e shoulder varies with Mach number. The model i n f i g u r e 10 i s
a two-dimensional f l a t slab with a sonic-wedge leading edge. The pres-
sures were calculated by simple inviscid shock-expansion theory. It can
be seen t h a t , as M, increases, t h e shoulder pressure increases t o very
laxge values. The pressure on the shoulder of blunt-nose bodies and blunt-
leading-edge wings would follow t h e same t r e n d with Mach number.

Figure 11presents t h e o r e t i c a l l y and experimentally determined pres-

sure d i s t r i b u t i o n s on a blunt-leading-edge f l a t p l a t e f o r a free-stream
Mach number of about 7. The t h e o r e t i c a l l y determined pressure d i s t r i -
butions were approximated by assuming sonic-wedge leading-edge conditions.
Also indicated i n the f i g u r e i s t h e value of t h e pressure r a t i o f o r no
entropy gain and zero v o r t i c i t y . The assumption t h a t free-stream s t a t i c
pressure e x i s t s on t h e f l a t p l a t e was p e r f e c t l y acceptable a t lower super-
sonic speeds. A t hypersonic speeds, however, t h e lasge entropy gain
through t h e normal shock and the large entropy and v o r t i c i t y gradients
induced in t h e flow f i e l d by t h e high shock curvature r e s u l t i n the type
of pressure d i s t r i b u t i o n shown by the s o l i d curve. T h i s curve was cal-
culated f o r t h e inviscid flow a t a Mach number of about 7 f o r the sonic-
wedge leading-edge configuration shown a t the upper l e f t . The dashed
curve i s the experimental pressure d i s t r i b u t i o n f o r t h e blunt-leading-
edge p l a t e shown on t h e r i g h t a t one value of Reynolds number. The d i f -
ference between these two curves i s due t o the presence of t h e boundary
layer, I f t h e Reynolds number had been lower, the boundary layer would
have been thicker and the separation of these two curves would have been
greater; the converse being t r u e i f the Reynolds number had been higher.
It should a l s o be pointed out t h a t , as t h e Mach number increases, not
only does t h e l e v e l of p/pw at t h e shoulder increase but a l s o t h e r a t e of
decrease with distance becomes l e s s , so that t h e entropy and v o r t i c i t y
e f f e c t s are spread over a greater distance a t higher Mach numbers.

Effect of Boundary-Layer Separation

When r e a l f l u i d e f f e c t s , including boundary layers, a r e brought

i n t o t h e picture, the consequences of boundary-layer separation must a l s o
be considered. A t hypersonic speeds boundary-layer separation i s often
important, although it can sometimes be neglected.

Figures 1 2 and 13 i l l u s t r a t e examples of boundary-layer separation

which must be considered and boundary-layer separation which may be,neg-
lected. Both the body with conical f l a r e ( " f l a r e d s k i r t " ) shown i n f i g -
ure 12 and t h e body with flapped wing shown in figure 13 were t e s t e d a t
M, = 7. (See r e f s . 1 4 and 15, respectively.) The manner i n which t h e
separation point moves rearward along the f l a r e d - s k i r t body with increasing
Reynolds number is indicated by the s o l i d l i n e in figure 12. The body-
pressure-coefficient d i s t r i b u t i o n s f o r two extreme positions are shown
above with corresponding symbols. The importance of knowing whether t o
design f o r separated or unseparated flows i s obvious. For unseparated
flow the s k i r t pressure i s about what would be expected in the absence
of viscosity, while the laminar separation region e s s e n t i a l l y protects
t h e s k i r t from d i r e c t contact with the f r e e stream.

On the other hand, a large portion of the upper surface of the w i n g

with trailing-edge f l a p ( f i g . 13) i s i n a separated-flow region and there
i s e s s e n t i a l l y no e f f e c t on the upper surface pressure coefficient. Fig-
ure 13 shows the f l a p deflected 16O; however, the same e f f e c t s would be
t r u e with a negative f l a p deflection. The loads on t h e upper surfaces
of wings a t angle of attack i n hypersonic flow are e s s e n t i a l l y negligible
whether separation e x i s t s or not; the difference between free-stream pres-
sure and vacuum i s s o small i n comparison with t h e pressures on the lower
surface t h a t , f o r a l l p r a c t i c a l purposes, t h e upper surface can be neg-
lected i n loads calculations.

25
 The separated flow on t h e lower surface i s confined t o a r e l a t i v e l y
s m a l l region. I f t h i s separation point were t o move forward t h e lower
surface loads would be affected more than shown i n f i g u r e 13. The magni-
tude of t h e loads induced would a l s o depend on t h e condition of t h e bound-
a r y layer, t h a t is, whether it is laminar or turbulent.

There i s as y e t not enough knowledge about separation a t hypersonic

speeds t o be able t o predict when o r where separation w i l l o c c w f o r
e i t h e r laminar or turbulent flow.

Aerodynamic Interference

Another f i e l d which i s r e l a t i v e l y unexplored a t hypersonic speeds

i s that of aerodynamic interference and the r o l e t h a t interference plays
i n a l t e r i n g t h e expected loads on any component.

One phase of t h e interference problem w a s investigated by building

a scale model of a configuration which had previously been t e s t e d a t
& = 3.36. (See r e f s . 16 and 17.) This model was t e s t e d a t &,, = 6.85
i n t h e Langley 11-inch hypersonic tunnel. Some preliminary r e s u l t s are
presented herein.

Figures 1 4 and 15 present t h e span-load d i s t r i b u t i o n s on t h e wing

alone and on t h e w i n g i n t h e presence of t h e body a t an angle of a t t a c k
of 15' f o r I& = 3.36 and & = 6.85, respectively. The o v e r a l l trends
of t h e r e s u l t s at & = 3.36 are about what would be expected. The
r e s u l t s a t M, = 6.85 show t h e l a r g e localized e f f e c t which t h e t h i c k
boundary l a y e r plays i n interference between adjacent components. The
indicated p o s i t i o n of t h e boundary layer was taken from schlieren p i c t u r e s
a t a = 0' on t h e sharp-nose body. The thickness and condition of t h e
boundary l a y e r a t t h e wing-body juncture a t a = l 5 O i s not known. Note
a l s o t h e e f f e c t of nose shape on t h e loadings. The blunt nose decreased
t h e wing loadings. Although t h e decrement was not appreciable a t t h i s
Mach number, it i s t o be expected t h a t t h e greater l o s s e s incurred by a
detached shock a t higher Mach numbers w i l l more seriously a f f e c t t h e
loadings not only on t h e wing but a l s o on a l l components within t h e region
of influence of t h e highly r o t a t i o n a l p a r t of t h e flow f i e l d associated
w i t h b l u n t noses.

The shock-expansion theory predicts t h e loading on the wing alone

a t & = 6.85 f a i r l y well. The .results of t h e & = 3.36 t e s t s a r e
not a f a i r t e s t of t h e adequacy of shock-expansion theory since a t
a = 15O t h e leading-edge shock i s detached at & = 3.36.

Figure 16 presents t h e interference loading on t h e body due t o t h e

presence of t h e wing a t an angle of a t t a c k of l 5 O f o r I& = 3.36 on t h e

I 245
1c.

sharp-nose body and Mm = 6.85 on t h e sharp- and blunt-nose bodies.

The o r i e n t a t i o n of t h e wing and body with respect t o t h e load-distribution
curve i s as indicated.

Mach number apparently does not play an important r o l e i n i n t e r -

ference e f f e c t s i n t h i s Mach number range, as evidenced by the f a c t t h a t
t h e general trends of t h e interference loading curves on t h e sharp-nose
body at . Mm = 3.36 and = 6.85 do not differ widely. The effect of
nose shape on body interference loadings i s evidenced by t h e r e l a t i v e
displacement of t h e curves with t h e square and diamond symbols, and, as
already mentioned, t h e significance of t h i s type of interference w i l l
probably increase with increasing a. Also of i n t e r e s t i s t h e f a c t
t h a t t h e maximum interference loading f o r each of t h e t h r e e curves w a s
50 t o 60 percent of t h e corresponding body-alone loading.

CONCLUDING RFMARKS

This paper has summarized b r i e f l y current loads information a t

hypersonic speeds. Several methods which the designer can employ i n
estimating t h e loads on various a i r c r a f t components have been discussed.
The paper has considered t h e c h a r a c t e r i s t i c s of b o t h slender and blunt
configurations and t h e e f f e c t s of boundmy-layer Ad aerodynamic i n t e r -
ference. Many problems s t i l l confront t h e designer - t h e e f f e c t on t a i l
loads of t h e wing flow f i e l d and i t s associated high-energy wake and t h e
e f f e c t of t h e body flow f i e l d and i t s highly r o t a t i o n a l f l o w f o r blunt-’
nose bodies. I n addition, t h e e f f e c t on loads a t hypersonic speeds of
t h e i n e r t degrees of freedom of t h e components of t h e a i r (molecular
vibration, dissociation, and ionization) i s e s s e n t i a l l y unknom.
 PSPENDIX

PRESSURE RELATION APPROXIMATIONS

The ratio of free-stream static pressure to stagnation pressure for

a blunt-nose body can be calculated approximately from the following exact
relation (see ref. 14, eq. (100)):
1

Since -
7 = -
1 + 1, equation (1) can be written:
7 - 1 7 - 1

As becomes large, equation-(2) is closely approximated by

For 7 = 1.4, equation (3) becomes

28
 In addition, since

cp = -(x
2
7402 pm - I)
and

the ratio of local surface pressure coefficient to maximum surface

pressure coefficient becomes

which, with the aid of equation (4), yields for large Ea,

,
 REFERENCES

1. Eggers, A. J., Jr., and Savin, Raymond C.: A Unified Two-Dimensional

Approach t o the Calculation of Three-Dimensional Hypersonic Flows,
With Application t o Bodies of Revolution. NACA Rep. 1249, 1955.
(Supersedes NACA TN 2811.)

2. Tsien, Hsue-Shen: Similarity Laws of Hy-personic Flaws. Jour. Math.

and Phys., vol. XXV, no. 3, Oct. 1946, pp. 247-251.

3. Ehret, Dorris M., Rossow, Vernon J., and Stevens, Victor I.: An
Analysis of the Applicability of the Hy-personic Similarity Law t o
the Study of Flow About Bodies of RevolutiGn a t Zero Angle of
Attack. NACA TN 2250, 1950.

4. ROSSOW, Vernon J.: Applicability of the Hypersonic Similarity Rule

t o Pressure Distributions Which Include the Effects of Rotation
f o r Bodies of Revolution a t Zero Angle of Attack. NACA TN 2399,
1951-
5. Lees, Lester: Note on t h e Hypersonic Similarity Law f o r an Unyawed
Cone. Jour. Aero. Sci. (Readers' Forum), vol. 18, no. 10, Oct. 1951,
pp. 700-702.

6. Probstein, Ronald F., and Bray, Kenneth N. C.: Hypersonic Similarity

and the Tangent-Cone Approximation f o r Unyawed B o d i e s of Revolution.
Jour. Aero. Sci. (Readers' Forum), vol. 22, no. 1, Jan. 1955,
pp. 66-68.

7. Van Dyke, Milton D.: Applications of Hy-personic Small-Disturbance

Theory. Jour. Aero. Sci., vol. 21, no. 3, Mar. 199, pp. 179-186.

8. Van Dyke, Milton D.: A Study'of Hypersonic Small-Disturbance Tbeory.

NACA Rep. 1194, 1954. (Supersedes NACA TN 3173.)

9. Eggers, A. J., Jr., Savin, Raymond C., and Syvertson, Clarence A.:
The Generalized Shock-Expansion Method and Its Application t o Bodies
Traveling a t High Supersonic Air Speeds. Jour. Aero. Sci., vol. 22,
no. 4, Apr. 1955, pp. 231-238, 248.

10. Savin, Raymond C.: Application of the Generalized Shock-Expansion

Method t o Inclined Bodies of Revolution Traveling a t High Supersonic
Airspeeds. NACA TN 3349, 1955.

ll. Penland, J i m A.: Aerodynamic Characteri8tics of a Circular Cylinder

a t Mach Number 6.86 and Angles of Attack.6p t o 9
0
'
. NACA TN 3861,
7.-

1957. (Supersedes NACA RM L54Al4.) t

..> 30
,*&*
-A.
14 .f

12. Goodwin, Glen, Craeger, Marcus O., and Winkler, Ernest L.: Investi-
gation of Local Heat-Transfer and Pressure Drag Characteristics of
a Yawed Circular Cylinder at Supersonic Speeds. NACA RM A55H31,
1956.
13. Crawford, Davis H., and McCauley, William D.: Investigation of the
Laminar Aerodynamic Heat-Transfer Characteristics of a Hemisphere-
Cylinder in the Langley 11-Inch Hypersonic Tunnel at a Mach Number
of 6.8. NACA 'I!N 3706, 1956.

14. Becker, John V., and Korycinski, Peter F.: Heat !Transfer and
Pressure Distribution at a Mach Number of 6.8 on Bodies With
Conical Flares and Extensive Flow Separation. NACA RM ~56~22,
1956.
15. Ridyard, Herbert W., and Fetterman, David E., Jr.: Aerodynamic
Characteristics of a 6-Percent-Thick Symmetrical Circular-Arc
Airfoil Having a 30-Percent-Chord !Trailing-Edge Flap at a Mach
Number of 6.9. NACA RM L56B24, 1956.

16. Hill, William A., and Kaattari, George E.: Force and Pressure- (i

Distribution Investigation to High Angles of Attack on All-Movable

Piangular and Rectangular Wings in Combination With a Body at
Supersonic Speeds. NACA RM ~56~12, 1956.
17. Kaattari, G. E.: Pressure Distributions on Piangulaz and Rectangular
Wings to High Angles of Attack - Mach Numbers 2.46 and 3.36.
NACA RM A54JI2, 1955.
18. Ames Research Staff: Equations, Tables, and Charts for Compressible
Flaw. NACA Rep. 1135, 1953. (Supersedes NACA TN 1428.)

.
TENDENCY TOWARD TWO-DIMENSIONALITY OF FLOW
AT HYPERSONIC SPEEDS

Figure 1

PHYSICAL CONCEPT FOR HYPERSONIC SIMILARITY LAW

-h= V , ~hC C Md? = K

“2 12

Figure 2

3%
 RANGE OF APPLICABILITY OF HYPERSONIC SIMILARITY
LAW (K-M,~) FOR CONES
a=oo
M,= Q)
24

20
HYPERSONIC SIMILARITY LAW
16 APPLICABLE IN THIS REGION
CONE FOR UNYAWED CONES
SEMIAPEX I2
ANGLE, 0
a
4

0 2 4 6 8 O M 1416 18

APPLICABILITY OF HYPERSONIC SIMILARITY LAW

TO SLENDER BODIES
7 QI=OO

5
METHOD OF CHARACTERISTICS
P 4
- --TANGENT-CONE
(LEES APPROXIMATION)
pa3
TANGENT- GONE
(KARMAN APPROXIMATION)
2
3SMaS12
I
I
- -
3SaS12
0 .2 .4 -6 .8 IO
VI

Figure 4
 CORRELATION OF THE,ORETICAL PRESSURE COEFFICIENTS
ON SLENDER UNYAWED CONES
COMBINED SUPERSONIC-HYPERSONIC SIMILARITY RULE

CP
TAN^ u 1.15 Mq) 10

Figure 5

CORRELATION OF THEORETICAL PRESSURE COEFFICIENTS

ON BLUFF UNYAWED CONES

0 IO 20 30 40 50 60
0-

Figure 6
1 34
 PREDICTION OF PRESSURE DlSTRl6UTlON ON OGIVE AT
ANGLE OF ATTACK
GENERALIZED SHOCK-EXPANSION METHOD; M a = 5.05

----- EXPERIMENTAL CONE

CONDITIONS

0 20 40 60 80 100
% NOSE LENGTH

Figure 7

INVARIANCE OF CYLINDER PRESSURE RATIO WITH

MACH NUMBER AND SWEEP ANGLE

M a = 6.9
M a = 3.9

0 IO 20 30
+.
40 50
DEG
60 70 80 90

!
35 Figure 8
c .
 INVARIANCE OF PRESSURE-COEFFICIENT RATIO WITH
MACH NUMBER FOR HEMISPHERICAL NOSES

1.0

.8

.6 0 4 8 1 2
GP
‘p, MAX .4 EXPERIMENT

.2

0 .2 .4 .6 .8 1.0 1.2 1.4 1.6 1.8

s/r

Figure 9

EFFECT OF ENTROPY ON SHOULDER PRESSURE OF

SONIC -WEDGE FLAT PLATE

24 r

20
16 -
t
-
PS
pa
12-

8 -

4-

I I I I I I I
0 2 4 6 8 10 12 1
4 16 18
Ma

Figure 10 36
20

EFFECT OF BOUNDARY LAYER AND ENTROPY ON

FLAT-PLATE PRESSURES
M a W 7 ; y = 1.4
LHX
t C I #
T-
SONIC- WEDGE
LEADING EDGE

0 IO 20 30
x/t

Figure 11

EFFECT OF SEPARATION ON PRESSURE DISTRIBUTION

ON BODY OF REVOLUTION

i
M a " 7 ; Q= '
0
.8r

Gp .4
D

.2 -

SEPARATION
Rd .4

I I I
0 2 4 6 8 1012

----

Figure 12
37'
 EFFECT OF SEPARATION ON PRESSURES ON
TWO-DIMENSIONAL WtNG FLAP
a=16O; 8=Is0; M a m 7 ; R=1.65XIO6

_/.VACUUM
0
UPPER SURFACE

.2

EXPANSION
.4

.6

.el
0
' 1
.2
I
A
' I '
.6
1 I
.e
I I
1.0
x/c

Figure 13

INTERFERENCE ON WING LOAD DISTRIBUTION AT a=15"

M a = 3.36

2.0 k,c SHOCK-EXPANSION THEORY

I .5

1.0

.5

0 25 50 75 100
% SEMISPAN

Figure 1 4
 INTERFERENCE ON WING LOAD DlSTRlWTlON AT Q=EO
M a = 6.8 5

I.5 EDGE OF

NOSE

0
I 25
I I
50 75
-
'q SHOCK- EXPANSlON THEORY
100
O/o SEMISPAN

INTERFERENCE ON BODY LOAD DISTRIBUTION QUE TO WING

Q =IS0

.I 2

-08

A cn ,B (w) -04
I \ 'i.9

-.04
5 6 7 8 9 10 II 12 13
x /d

Figure 16
id. 39
d
 SPAN LOADINGS DUE TO WING TWIST AT
TRANSONIC AND SUPERSONIC SPEEDS

By Frederick C. Grant and John P. Mugler, Jr.

Langley Aeronautical Laboratory

SuMMP,ziY

Two similar tapered sweptback plan forms with L e same two spanw-se
variations of t w i s t have been t e s t e d i n t h e Mach nmiber range from 0.8
t o 2.0. The t e s t r e s u l t s showed, i n general, r a t h e r good agreement with
t h e o r e t i c a l predictions of t h e incremental span loadings due t o t w i s t
f o r zero angle of attack. The measured incremental span loadings due
t o t w i s t generally diminished w i t h increasing angle of attack through
the Mach number range. A t a Mach n W e r of ,O.g, t h e incremental loadings
progressively vanished f r o m the t i p inboard with increasing angle of
attack. For the highest angles of a t t a c k (about 20') a t Mach nmiber 0.9,
there was no difference i n t h e span loadings of t h e f l a t and twisted
wings. A t t h e higher supersonic speeds, a similar vanishing a t the t i p s
of the incremental loading due t o t w i s t w a s s t a r t i n g a t the highest angles
of attack (near 20').

For angles of a t t a c k lower than about 20' a t supersonic speeds, no

important change i n the shape of t h e incremental loadings occurred,
although t h e strength of t h e loading diminished with increasing angle
of attack.

INTRODUCTION

The t h i n wings of modern high-speed airplanes deform appreciably

i n f l i g h t . The changes i n air loading due t o these deformations have
not been extensively investigated. A n aerodynamically important form
of deformation i s t w i s t , o r change i n angle of a t t a c k at a given span-
wise s t a t i o n on a wing. A s p a r t of a research program.on t h e loads due
t o wing t w i s t , t w o simple spanwise t w i s t d i s t r i b u t i o n s have been t e s t e d
a t the Langley Aeronautical Iaboratory i n t h e Mach nmiber range from 0.8
t o 2.0. For a complete airplane with stores and nacelles acting on t h e
wing, t h e t w i s t d i s t r i b u t i o n along t h e span may be r a t h e r complicated.
It i s hoped that t h e loadings due t o simple t w i s t d i s t r i b u t i o n s w i l l ,
by superposition, give t h e *loadings due t o complicated d i s t r i b u t i o n s .
$3
 a
e 0
* l e
__ 5 --•
ear
a e e . -* -e -* e*,

SYMBOLS

A aspect r a t i o

b span

C chord

cAV average chord

cn section normal-force c o e f f i c i e n t

M Mach nuuiber

dynamic pressure

thickness

X chordwise distance

Y spanwise distance

a angle o f a t t a c k

incremental normal-force c o e f f i c i e n t

iricremental l i f t i n g pressure

sweepback a t quarter chord

A taper r a t i o

MODELS

The wings t e s t e d and the t w i s t v a r i a t i o n s which were b u i l t i n a r e

shown i n f i g u r e 1. The wings t e s t e d a t transonic speeds had an aspect
r a t i o of 4, 45' of sweepback a t t h e quarter chord, and a t a p e r r a t i o
of 0.17. The semispan wing t e s t e d a t supersonic speeds had an aspect
r a t i o of 3.5, 50' of sweepback a t t h e quarter? chord, and a t a p e r r a t i o
of 0.20. The thickness of t h e transonic wings varied from 6 percent a t
t h e body center l i n e t o 3 percent a t and beyond halfway t o t h e t i p .
The thickness of t h e supersonic y&ngs w a s a constant 5 percent. A small
caaiber was b u i l t i n t o t h e transonic wings. All t h e wings t e s t e d had t h e
 b*
.a
a,
. .,- . -
* C -
a a
a
0

same 65A-series thickness d i s t r i b u t i o n and t h e same spanwise v a r i a t i o n s

of b u i l t - i n t w i s t . The t w i s t angle a t t h e t i p w a s 6 O , i n every case,
which i s a t t a i n e d by a l i n e a r and quadratic v a r i a t i o n with spanwise
position. The t i p s a r e a t a lower angle 02 a t t a c k than t h e root, or
washed out, f o r t h e positive direction assumed i n t h i s paper. F l a t wings
were t e s t e d i n each speed range t o provide a reference t o which t h e
twisted wings might be compared.

INCRFSIENTAL LOADING

Figure 2 shows the span loadings on t h e f l a t wing and t h e l i n e a r l y

twisted wing a t M = 1.6 and a t a = 12'. The difference i n these span
loadings, o r incremental span loading, i s a l s o shown. Incremental span
loadings formed i n the same manner w i l l be t h e b a s i s of comparison between
l i n e a r theory and t h e t e s t r e s u l t s at t h e other Mach numbers and angles of
attack.

The incremental loading shown i n figure 2 ' i s t h e i s o l a t e d e f f e c t

of spanwise wing t w i s t with, of course, the nonlinear influence of angle
of a t t a c k and thickness neglected. If r e a l wings behave as do t h e wings
of l i n e a r theory, the incremental loading f o r a given spanwise t w i s t d i s -
t r i b u t i o n w i l l not change with angle of attack. For t h i s case t h e incre-
ment i n normal force produced by 6 O of t w i s t i s 13 percent of t h e f l a t -
p l a t e normal-force coefficient. This i l l u s t r a t e s t h e f a c t t h a t , f o r a
given o v e r a l l accuracy i n predicting t h e loading on a twisted wing, t h e
accuracy of prediction of the incremental loading can diminish as the
angle of a t t a c k increases. -

PREDICTIONS AT ZEBO AJYGLF: OF ATTACK

I n order t o e l i n i n a t e , as far as possible, t h e influence of angle

of attack, t h e root angle of a t t a c k may be s e t t o zero. The predicted
and measured incremental span loadings due t o t w i s t with t h e root angle
of a t t a c k s e t t o zero a r e shown i n figures 3 and 4.

Wings With Linear Twist

Transonic speeds.- Figure 3 shows t h e comparative t h e o r e t i c a l and

experimental incremental span loadings for t h e transonic l i n e a r l y twisted
wing. The section loading parameter Acn C/CAV i s p l o t t e d against
.the spanwise posikion 2y/b, and t h e v e r t i c a l dashed l i n e indicates t h e
spanwise position of t h e wing-body juncture.
I
't. . 42
: 1.

.,. i+c
4

At M = 0.90, t h e agreement between the data and theory i s f a i r l y

good. The theory shown i s a l i f t i n g - s u r f a c e theory w i t h a provision f o r
approximating t h e presence of t h e body. (See r e f . 1.) The prediction
i s b e t t e r outboard than it i s nearer t h e body.

A t M = 1.20, t h e r e i s close agreement between t h e data and theory,

even though t h e v a l i d i t y of l i n e a r theory i s becoming questionable as
t h e Mach nuniber approaches one. The theory used a t supersonic speeds
f o r subsonic leading edges i s that given i n references 2 and 3. In
addition, t h e boundary conditions w e r e o n l y approximately s a t i s f i e d i n
the t h e o r e t i c a l computations f o r t h e transonic w i n g s a t M = 1.20. No
attempt was made t o account f o r t h e presence of t h e body. A feature of
t h e r e s u l t s a t M = 1.20 i s t h e apparent absence of any marked influence
of t h e body on the incremental span loadings.

Supersonic speeds.- Figure 4 shows t h e incremental span loadings

wfth zero r o o t angle of a t t a c k f o r the supersonic l i n e a r l y twisted wings
a t M = 1.6 and 2.0.

Figure 4 shows that t h e data are about 20 percent lower than pre-
d i c t e d values. A s predicted, t h e loading i s s l i g h t l y weaker a t t h e higher
Mach number. The shock waves caused by t h e thickness seem t o have no
more e f f e c t on the span loadings a t M = 2.0 than a t M = 1.6, although
t h e leading edge i s supersonic a t M = 2.0 and shock waves due t o thick-
ness must c e r t a i n l y b e more severe. The theory used a t M = 2.0 is
given i n reference 4.

Wings With Quadratic Twist

Transonic speeds.- Figure 5 shows t h e incremental span loadings

on t h e wings with quadratic twist as measured and predicted a t transonic
speeds.

The agreement with theory i s again r a t h e r good a t M = 0.90. The

agreement a t M = 1.20 i s about t h e same as it w a s i n the case of t h e
wings with l i n e a r t w i s t . Again t h e r e i s no apparent body effect
a t M = 1.20.

Supersonic speeds.- Figure 6 shows t h e predicted and measured incre-

mental loadings f o r t h e wing with quadratic t w i s t a t M = 1.6. Data
f o r M = 2.0 are not y e t available.

The agreement i s b e t t e r i n t h i s case than it w a s f o r t h e l i n e a r l y

twisted wing a t t h i s Mach number. The values are o n l y 7 percent lower
as c o q a r e d with about 20 percent f o r t h e w i n g with l i n e a r t w i s t . T h i s
must be p a r t l y due t o t h e f a c t t h a t t h e average angle of t w i s t over the
plan form i s lower than it was i n t h e case of t h e l i n e a r l y twisted wings.

I' . 43
 e$
0

LIFTING PRESSURE DISTRIBUTIONS

Figure 7 shows the chordwise lifting pressure distribution corre-

sponding to two of the incremental span loadings previously shown. The
lifting pressure coefficient Ap/q is plotted against the chordwise
position x/c; distributions are for Mach number 1.6, zero root angle of
attack, and the spanwise station at which the data were taken is 0.7
of the semispan. Distributions for both linear and,quadratic twist are
shown. These distributions are typical of other spanwise stations at
this Mach number. Linear-theory predictions of the lifting pressure are
shown for both twist variations.

For the wing with linear twist, the agreement with theory is good.
The level of agreement is comparable %o that indicated by recent pressure
measurements made on a zero-thickness delta wing. (See ref. 3 . ) Since
a zero-thickness delta wing exactly satisfies the boundary conditions
of linear theory, the agreement with theory obtained on such a wing
typifies the best that can be expected. To have similar agreement on
a wing with 5-percent thickness is surprising. The agreement for the
wing with quadratic twist is even better than that for the wing with
linear twist. The fine agreement shown here was reflected in the good
agreement observed in the integrated loadings for the wing with quadratic
twist.

R U D I C T I O N S AT A N G U OF ATTACK

All the incremental loadings that have been shown thus far were
for zero root angle of attack. According to the linear theory, the
incremental loadings will not change with angle of attack, or, in other
words, the twist will produce the same change in loading whether or not
the wing is at an angle of attack. Of course, this simple prediction
is not borne out by the data.

Transonic Speeds
Figure 8 shows the effect of angle of attack on the span loadings
at M = 0.90. In this figure, instead of incremental loadings, the total
span loadings are shown for the flat and linearly twisted wings. Data
for angles of attack of bo, 8 O , and 1
' are shown. For the transonic
2
wings at angle of attack, incremental aeroelastic twists occurred which
amounted to about 10 percent of the 6' of built-in twist at 1
2' angle
of attack. Figure 8 shows that the shape of the incremental loadings
(the vertical difference between curves) changes markedly with angle of
attack while the strength of the incremental loading greatly diminishes.

4%
6

At a = bo, b c t h wings show t h e same s o r t of span loading, t h e wing with

l i n e a r t w i s t carrying t h e l e s s e r load. Between a = bo and a = 8O,
t h e flow separates a t t h e t i p of t h e f l a t wing, and a t a = 8' the flow
i s separated outboard of about 60 percent of t h e semispan. The twisted
wing a t a = 8O, however, has much t h e same type of span loading as a t
a = 4' and t h e flow appears unseparated. A t a = 1 2' both wings a r e
separated outboard of about 40-percent semispan. The incremental loading,
already small a t a = 12O, e f f e c t i v e l y vanishes a t t h e higher angles of
a t t a c k . A t t h e higher angles, then, t h e r e i s no difference between t h e
f l a t and twisted wings. Similar results have been obtained on t h e wing
with quadratic t w i s t . A t Mach number 1.2 t h e r e s u l t s a r e consistent w i t h
those t o be shown f o r t h e supersonic wings, b u t values w i l l not be
presented.

Supersonic Speeds

I n f i g u r e 9 , t h e percent of t h e t h e o r e t i c a l loading which must be

used t o obtain a good f a i r i n g through t h e data i n t h e outboard regions
(beyond semispan), where most of t h e incremental l i f t i s located, i s
p l o t t e d against t h e root angle of attack. The most s t r i k i n g feature of
t h i s p l o t i s t h e r a p i d decrease of t h e e f f e c t i v e l i n e a r t w i s t with angle
of a t t a c k . There i s no marked e f f e c t of t h e Mach nurriber, although
t h e M = 2.0 data a r e f o r a supersonic leading edge and t h e M = 1.6
data a r e f o r a subsonic leading edge. A s was mentioned previously, a
l e s s accurate prediction of t h e incremental loading i s acceptable a t t h e
higher angles of a t t a c k . Even i f 100 percent of t h e t h e o r e t i c a l loading
f o r t h e l i n e a r t w i s t were used t o p r e d i c t t h e loading a t 1 ' angle of
2
a t t a c k , t h e &>-percent difference indicated by f i g u r e 9 would come t o
an e r r o r of about 12 percent i n predicting t h e t o t a l loading. A b e t t e r
estimate o f t h e incremental loading, such as the f r a c t i o n s of t h e theo-
r e t i c a l loading indicated by t h e curves, could lead t o a negligible e r r o r
i n t h e t o t a l loading.

For t h e wing with quadratic t w i s t , only t h e M = 1.6 data, or

subsonic-leading-edge data, are available. However, t h e r e i s no reason
t o expect t h a t the Mach number e f f e c t s w i l l be any stronger than they
were for t h e l i n e a r l y twisted wing. F o r t h e wing w i t h quadratic t w i s t ,
f i g u r e 9 shows that t h e good prediction of t h e incremental loading a t
zero angle of a t t a c k i s coupled with a slow drop i n e f f e c t i v e t w i s t as
t h e angle of a t t a c k increases. This contrast w i t h t h e r e l a t i v e l y poorer
prediction a t zero angle o f a t t a c k and more r a p i d drop with angle of
a t t a c k observed on t h e l i n e a r l y t w i s t e d wing.

There i s l i t t l e change i n t h e shape of t h e incremental loadings

from 1
2' angle of a t t a c k t o a w u t 20'. I n t h e neighborhood of 20°,
incremental loadings vanish on the outboard regions of t h e wing i n a
manner similar t o t h a t observed a t M = 0.90.
! -
I
I . . 45
 7

CONCLUDING RESIARKS

At higher subsonic speeds the theoretical predictions at zero angle

of attack of incremental span loads due to twist were fairly good.
Because of separation effects, these predictions failed as the angle of
attack increased. At the highest angles, there was no difference in the
loadings of the flat and twisted wings. At low supersonic speeds, the
predictions at zero angle of attack were better although the validity
of the linear theory is becoming questionable. At the higher supersonic
speeds, the predictions at zero angle of attack were generally larger
than the actual loadings. The prediction was better for the wings with
lower average twist. At angles of attack up to 12O, factors were applied
to the theoretical incremental loading which give good agreement with
the data. Through the Mach number range of 0.9 to 2.0 the incremental
loading steadily diminished with angle of attack.

REFERENCES

1. Crigler, John L.: Comparison of Calculated and Experimental Load

Distributions on Thin Wings at High Subsonic and Sonic Speeds.
NACA TN 3941, 1957.
2. Heaslet, Max. A., and Lomax, Harvard: Supersonic and Transonic Small
Perturbation Theory. General Theory of High Speed Aerodynamics.
V o l . VI of High Speed Aerodynamics and Jet Propulsion, sec. D,
ch. 3 , W. R. Sears, ed., Princeton Univ. Press, 1954, pp. 186-206.

3. Lomax, Harvard, Heaslet, M a x . A., and Fuller, Franklyn B.: Integrals

and Integral Equations in Linearized Wing Theory. NAC!. --p. 1054,
1951. (Supersedes NACA TN 2252.)
4. Kainer, Julian H.: Equations.for the Loading on Triangular Wings
Having Supersonic Leading and Trailing Edges Due to Various Basic
Twist Distributions. Jour. Aero. Sci., vol. 20, no. 7, July 1953,
pp. 469-476-

3 . Boatright, William B.: Experimental Study and Analysis of Loading

and Pressure Distributions on Delta Wings Due to Thickness and to
Angle of Attack at Supersonic Speeds. NACA RM ~56114,1956.

I 46
8

MODEL CONFIGURATIONS
TRANSONIC SUPERSONIC
A.4; Ac/4= 45O; X=0.15 A.3.5; AC,=5O0 j X.0.20
NACA 6 5 A 2 0 6 , ROOT NACA 6 5 8 0 0 5
NACA 6 5 A 2 0 3 , 0 . 5 b / 2 TO TIP

?$?tT, %%T,
DEG -6 DEG -6
-I 0 I 0 I
2y/ b 2y/b

Figure 1

ILLUSTRATIVE EXAMPLE OF AN INCREMENTAL SPAN LOADING

CY =12" M = 1.6

.*r
/--

.4-

.2-

1
INCREMENTAL'
I J
720 0.5 I

47 Figure 2
INCREMENTAL SPAN LOADINGS ON WING WITH LINEAR TWIST
Q=OO
M= 0.90 M11.20

0 .5 I.o
2y/b

Figure 3

INCREMENN SPAN LOADINGS ON WING WITH UNEAR TWIST

CY=OO

M.1.6 M=2.0

Figure 4

48
INCREMENTAL SPAN LOADINGS ON WING WITH QUADRATIC TWIST
a soo

M.0.90 M.1.20

AC" C
'AV
I I I
0 .5 I .o

INCREMENTAL SPAN LOADINGS ON WING WITH QUADRATIC MllST

azo0;Mz1.6

- -.I
ACnC
THEORY
'AV

-,2' I I
0 0.5 I
Zy/b

Figure 6
 INCREMENTAL LIFTING PRESSURES ON TWISTED WINGS
M.1.6; a z o o ; 2y/ba0.7

LINEAR TWIST QUADRATIC TWIST

-
AP -.2 - P
9
-.4-09'
1
I
THEORY -pb
J
I
I --e *Ll-&+---

THEORY

I
I
I I I I

Figure 7

*IL,\ 1:\
EFFECT OF ANGLE OF ATTACK ON SPAN LOADING

I1.2

.4
I

I
a =4O
M= 0.90

O-- FLAT WING

D--- TWISTED WING

~~ I
Q 580 a 1120

Fa. I
I
U
h
0 .5 I 0 .5 I 0 .5 I
2y/b 2y/b 2y/b

Figure 8
sip
 -
- -
OF
O/o
THEORETICAL 50 -
0 -
LOADING - -
- -
- -
- -
I I I I I I
 *m . e

FLIGHT MEASURENENTS AND CALCULATIONS OF WING LOADS AND

LOAD DISTRIBUTIONS AT SUBSONIC, TRANSONIC, AND
SUPERSONIC SPEEDS

By Frank S. Malvestuto, Thomas V. Cooney, and E a r l R. Keener

NACA High-speed Flight S t a t i o n

Presented i n t h i s report i s a summaxy of l o c a l and n e t angle-of-

a t t a c k wing-panel loads measured in f l i g h t on s i x airplanes. I n addition,
a comparison of these loads measured i n f l i g h t with calculations based on
simple theory i s presented.

INTRODUCTION

A t t h e High-speed F l i g h t S t a t i o n of t h e National Advisory Committee

for Aeronautics, f u l l - s c a l e research i n t h e f i e l d s of s t a b i l i t y , perform-
ance, and loads i s conducted with a v a r i e t y of completely instrumented
research and military-type airplanes.

I n t h e present paper, t h e aerodynamic loads aspect of t h i s f l i g h t

research i s considered. The presentation w i l l involve a summary of l o c a l
and n e t angle-of-attack wing-panel loads measured in f l i g h t on a v a r i e t y
of airplanes flown during t h e past 5 o r 6 years. I n addition, a prelim-
inary comparison of these loads measured i n f l i g h t and t h e corresponding
loads calculated by simple theory i s presented. The object of t h i s corn-
parison i s t o assess t h e a b i l i t y of simple t h e o r e t i c a l techniques t o pre-
d i c t t h e flight-measured loads f o r a v a r i e t y of configurations. Only a
cursory comparison of t h e f l i g h t measurements with comparable wind-tunnel
r e s u l t s has been made. I n a general sense, t h e f l i g h t results v e r i f y t h e
tunnel findings. For t h e convenience of t h e reader, a bibliography has
been added.

Figure 1 depicts with plan-view outlines t h e airplanes t o be d i s -

cussed i n t h i s report. The wing panels are darkened t o emphasize t h e
f a c t t h a t only t h e wing loads w i l l be considered. An inspection of
t h e individual sketches and geometric d a t a shows t h a t there i s a good
coverage of Wing sweep, plan form, aspect r a t i o , and thickness. In
addition, t h e X - l E wing has 2’ positive incidence and t h e D-558-11 wing
has 3’ of p o s i t i v e incidence. The free-stream Reynolds number f o r t h e
f l i g h t s of these airplanes varied,from,*; x 106 t o 6 x 106 per foot. The
a l t i t u d e varied from 25,000 f e e t t o 65,000 f e e t .
..
. a - -’
.e. .. . .
’ * e e
1
e
- - -.
.
‘ *
0.
*.

SYMBOLS

A aspect ratio
b‘ wing-panel span

bf flap span

C chord

Cav average chord

Cf flap chord

Cn section normal-force coefficient

CN net normal-force coefficient

cNa slope of wing-panel normal-force coefficient

CP pressure coefficient

KP pressure coefficient differential between upper and lower surfaces

H altitude

M Mach number

%o free-stream Reynolds number

t thickness

X distance along x-axis

Y distance along y-axis

a angle of attack

se elevon deflection

*LE leading-edge sweep

53
 3

THEORIES CONSIDERED

A few preliminary remarks regarding t h e t h e o r i e s used f o r t h e wing-

panel load calculations w i l l be made. The w i n g s a r e assumed t o be r i g i d
f l a t p l a t e s and of negligible thickness. In addition, t h e e f f e c t of t h e
fuselage interference on t h e wing loads was approximated by assuming t h e
fuselage t o a c t as a p e r f e c t r e f l e c t i o n plane located at t h e wing-
fuselage juncture. On t h i s basis, t h e w i n g load i s predicted as t h e
load on one panel of a symmetrical wing with i t s r o o t chord coincident
with t h e wing-fuselage juncture. It i s r e a l i z e d t h a t t h i s approximation
t o t h e fuselage interference i s subject t o improvement; however, it i s
f e l t t o be s u f f i c i e n t f o r t h e present study. With these assumptions i n
mind, t h e wing theories used f o r load predictions are given i n t h e f o l -
lowing t a b l e :

Theories used f o r calculation of w i n g loads -

Subsonic Transonic Supersonic
(0.5 C M < 0.85) ( M = 1.0) (M 2 1.2)
~

A l l wings: l i n e a r Swept wing: l i n e a r A l l w i n g s : linear

l i f t i n g surface l i f t i n g surface l i f t i n g surface
( r e f s . 1 t o 4) ( r e f s . 5 and 6) ( r e f s . 9 t o 16)

Unswept wing: two-

dimensional f l a t
plate; two-
dimensional double
wedge ( r e f s . 7
and 8 )

I n t h e subsonic range, f o r a l l wings, linear theory w a s applied.

:See r e f s . 1 t o 4.) These subsonic calculations were made up t o a Mach
number of 0.85, although i n the neighborhood of t h i s Mach number, tran-
sonic mixed-flow conditions no doubt e x i s t . I n t h e transonic range, cal-
culations were made only f o r a free-stream Mach number of 1.0. I n t h i s
range, f o r t h e swept wings, t h e l i n e a r theory presented by Mangler ( r e f . 5 )
which i s i n essence Jones' slender-wing theory ( r e f . 17) modified f o r
l i n e a r i z e d sonic-flow condi%ions w a s applied. For t h e unswept w i n g , a t a
Mach number of 1.0, use w a s made of t h e results of Guderley and Yoshihaza
( r e f . 8 ) f o r a double-wedge section and t h e r e s u l t s of Guderley (ref. 7)
f o r a f l a t p l a t e of negligible thickness. For t h e supersonic Mach number
range, t h e well-known l i f t i n g - s u r f a c e theories were applied.

i 54'
 4

LOADING DISTRIBUTION

I n the discussion of f l i g h t r e s u l t s , the chordwise and spanwise

loadings f o r the unswept-wing X-1E airplane, the swept-wing D-358-11 air-
plane, and the delta-wing JF-10% airplane are considered and then a
force summary f o r all s i x airplanes i s given.

Some idea of the f l i g h t Reynolds number, a l t i t u d e , and angle-of-

attack excursions f o r these airplanes can be determined from figure 2. The
Reynolds number i s given on a per-foot basis and f o r free-stream con-
d i t i o n s . The open c i r c u l a r symbol represents the m a x i m u m Reynolds num-
ber obtained. It i s noted t h a t t h i s f l i g h t Reynolds number varies from
approximately 1x 106 t o 4 x 106. The a l t i t u d e covers a range from
approximately 25,000 t o 65,000 f e e t . On t h e right-hand side of figure 2
t h e hatched boundary i s indicative of the maximum angle-of-attack excur-
sions obtained i n f l i g h t . The discussion of the angle-of-attack wing
loads w i l l be within the region shown by t h e dashed boundary.

I n figures 3 t o 6 a r e presented t h e chord loadings and span loadings

f o r the X - D w i n g panel. The s o l i d l i n e represents the theory; the open
symbol, t h e f l i g h t data. The dashed l i n e through the open c i r c l e s repre-
sents "faired" f l i g h t data. The sketches on t h e left-hand side of f i g -
ure 4 indicate the panel normal-force coefficient CN f o r t h e angles of
attack a t which the chord and span loadings a r e shown. Consider f i r s t
t h e chord loadings of figure 3, t h a t i s , t h e variation of L?Ep, the
l i f t i n g pressure, with x / c , the normalized distance from t h e leading
edge. These r e s u l t s a r e f o r a span s t a t i o n -= 0.46. The symbol b'
b'/2
denotes t h e external panel span. The chord loadings a r e shown f o r Mach
numbers of 0.8, 1.0, and 1.9. For each Mach number t h e chord loadings are
shown f o r t w o angles of attack, a low angle and a high angle. The magnitude
of the high angle of attack is limited by t h e a v a i l a b i l i t y of t h e data.
The angle of attack i s always the angle of attack of the w i n g panel. A t
M = 0.8, t h e calculated l e v e l and v a r i a t i o n of the chord loading compares
favorably with the f l i g h t data. For a Mach number of 1.0, t h e r e is no
available finite-span unswept-wing theory. The t h e o r e t i c a l variation
shown here i s the f l a t - p l a t e two-dimensional theory of Guderley. Although
t h e l e v e l of the l i f t i n g pressure i s not predicted herein, t h e variation
i s similar t o the flight-measured variation f o r both angles of attack.

A t supersonic speed and low angle of attack, the comparison of f l i g h t

and theory i s acceptable. A t the higher angle of attack, the loading d i s -
t r i b u t i o n i s not predicted by theory although t h e l e v e l of the l o c a l load
-can be calculated. The midspan chord loadings and the chord loadings a t
two additional spanwise stations,'one near t h e root and one n e w the t i p ,
are shown i n figures 5 and 6.
 a .
a .

If t h e span-load d i s t r i b u t i o n s ( f i g . 4) are considered, it is noted

t h a t , a t M = 0.8 and M = 1.9, t h e calculated span loading compares
favorably with t h e flight-measured loading. For M = 1.0, t h e span loading
was not calculated, since, as mentioned previously, t h e two-dimensional
r e s u l t s of Guderley were used; however, t h e f l i g h t data have been f a i r e d .
The shapes of t h e span-loading curves strongly resemble each other f o r t h e
t h r e e Mach numbers shown.

For t h e swept-wing D-55SII airplane t h e chordwise and span-load

d i s t r i b u t i o n s f o r t h e wing panel a r e shown in figures 7 t o 10. The
s o l i d l i n e represents t h e calculations and t h e open c i r c u l a r symbol, t h e
f l i g h t measurements. The panel normal-force c o e f f i c i e n t s corresponding
t o t h e angles of a t t a c k considered are indicated i n the sketches on t h e
left-hand s i d e of f i g u r e 8. The chord loadings presented i n f i g u r e 7
a r e f o r a spanwise s t a t i o n close t o t h e midsemispan location. For t h e
subsonic and supersonic speeds, t h e theory allows t h e calculation of t h e
l e v e l and v a r i a t i o n of the chord loading except at t h e high angle of
a t t a c k f o r t h e supersonic Mach number. A t M = 1.0, t h e measured distri-
bution of t h e l i f t i n g pressure EP i s not calculated by t h e l i n e a r
theory. Theory gives a zero loading behind t h e l i n e a r i z e d sonic shock
t h a t starts from t h e leading edge of the streamwise t i p of t h e w i n g panel.
It i s possible t o obtain a nonzero loading by minor a l t e r a t i o n s of t h e
wing-tip geometry s o t h a t , f o r t h e portion of t h e wing behind t h e l i n e -
arized shock, t h e l o c a l span increases with increasing longitudinal posi-
tion; and hence l i f t i s produced. (See ref. 17.) A discussion of t h i s
artifice i s given in t h e report by Mangler (ref. 5) mentioned e a r l i e r .
The midspan chord loadings and t h e chord loadings near t h e root and t i p
a r e shown i n figures 9 and 10.

The span loading f o r t h e swept-wing D-558-11 is presented in f i g -

ure 8. A t subsonic and supersonic speeds t h e calculated d i s t r i b u t i o n
compares favorably with t h e f l i g h t measurements. For M = 1.0, t h e cal-
culated loading, especially a t t h e high angle of a t t a c k ( l l o ) , does not
represent t h e experiment because of t h e i n a b i l i t y of t h e theory t o pre-
d i c t t h e l e v e l of t h e loads i n t h e v i c i n i t y of t h e root and t i p regions.
A t an angle of a t t a c k of lIo, t h e CN of t h e panel i s approximately 0.8.
It i s possible t h a t separation e f f e c t s a t t h e r o o t and t i p a r e important
f o r t h i s configuration. In addition, t h e simple end-plate correction
used herein f o r fuselage interferences may be approximate. I n t h i s
regard t h e application of an analysis such as t h a t reported by Crigler
(ref. 6) f o r wing-body interference at sonic speeds would improve t h e
prediction of t h e loading i n t h e v i c i n i t y of t h e root.

The flight-measured loads f o r t h e wing panel of t h e 60° delta-wing

JT-102A airplane a r e considered next. I n f i g u r e 11 is shown an exploded
view of t h e wing. Note t h e two fences located i n the forward portion of

56
6

t h e w i n g and the elevon surface which is operative during f l i g h t . This

wing has conical camber and a reflexed t i p . For the calculation of the
wing-panel loads, the e f f e c t of the fences and t h e e f f e c t s of the conical
camber and t h e reflexed t i p m e neglected; however, t h e e f f e c t of the
elevon has been considered.

I n figures 1 2 and 13 a r e shown the chord loading and the span loading
f o r t h i s airplane. For the lower angle-of-attack range (angles of attack
from 3 O t o 5O) the calculations of the chord loadings compare favorably
with the measurements. Up-elevon deflection i s negative. The f a c t t h a t
the loading a t the leading edge is not predicted i s p a r t l y due t o t h e
omission of camber e f f e c t i n the calculations. Although t h e e f f e c t of
elevon a t M = 1.0 was not calculated, an inspection of t h e low-angle-of-
attack r e s u l t s indicates t h a t the elevon load calculations a t l o w super-
sonic speeds such as those obtained a t M = 1 . 2 a r e reasonable approxi-
mations t o t h e elevon load a t M = 1.0. A t t h e high angles of a t t a c k the
remarks made f o r the low angles of attack f o r the sonic and supersonic Mach
numbers are s t i l l reasonably valid. For t h e subsonic Mach number, t h e angle
of attack i s 20° and the calculations do not predict t h e f l i g h t measurement
primarily because of leading-edge separation. For the case of leading-edge
sepazation, calculations of the loading should be made within the framework
of the approximate separation flow theories such as reported by Brown and
Michael ( r e f . 18). The panel span loadings f o r t h e JF-1O;LA are shown i n
figure 13. The i n a b i l i t y of the calculations t o produce.the f l i g h t trends
a t M = 0.8 and a = 20' i s clear from the remarks r e l a t i n g t o t h e chord
loading a t t h i s Mach number and angle of attack. A t M = 1.0, since the
elevon load w a s neglected, t h e calculations overestimate s l i g h t l y the l e v e l
of the d i s t r i b u t i o n . The e f f e c t of the fences on the span loading d i s t r i -
bution can c l e a r l y be seen a t M = 1.0 and a = . '
0
1

I n general, the o v e r a l l impression from t h i s preliminary comparison

i s what would be expected from similar comparisons with wind-tunnel r e s u l t s .
Briefly, a reasonable approximation of t h e span loadings can be determined
f o r the low and moderate angle-of-attack range. The estimation of the chord
loadings i s l e s s satisfactory, p a r t i c u l a r l y i n the neighborhood of a Mach
number of 1.0.

NORMAL FORCES

I n f i g u r e 14 i s shown t h e variation of the panel normal-force coef-

f i c i e n t with panel angle of attack. Note i n t h i s i l l u s t r a t i o n t h a t the
open c i r c u l a r symbol represents the f l i g h t measurements f o r Mach numbers
of 0.8 and 1.0. The s o l i d symbol represents the f l i g h t measurements f o r
supersonic Mach numbers. The calculations a r e again represented by the
 7

s o l i d l i n e s . For t h e unswept wing at a Mach number of 1, t h e calculated

variation is simply t h e r e s u l t of Guderley and Yoshihara ( r e f . 8) f o r a
two-dimensional wing w i t h a 4-percent-thick double-wedge section. The
theory here does not predict the magnitudes or the v a r i a t i o n f o r t h e
range of angle of attack where f l i g h t measurements a r e available. Tunnel
r e s u l t s , however, f o r a similar w i n g indicate t h a t t h e CN variation
w i t h a is not l i n e a r and i n the lower angle-of-attack range (below 4'
angle of a t t a c k ) , theory more nearly agrees w i t h t h e experimental
variation.

I n figure 15 an attempt has been made t o show t h e e f f e c t of Mach

number on t h e normal-force derivative C N ~f o r all six airplanes that
were sketched i n figure 1. The theory i s again represented by t h e s o l i d
l i n e and, i n addition, t h e inverted "V" symbol has been used t o indicate
t h e magnitude of C N ~ a t M = 1.0. The f l i g h t data are represented by a
square symbol. The s o l i d symbol represents a low CN! range; t h e open
symbol, a moderate CN range; and the half-solid, a high CN range. I n
most cases, f l i g h t d a t a were available f o r ~ n l yone of these ranges.
For t h e X-1E a t sonic speed, the difference i n t h e calculated and f l i g h t
values r e s u l t s from lack of f l i g h t data i n t h e low CN range as pointed
out i n t h e discussion of figure 14.
I n general t h e calculated normal-force-curve slopes compare favor-
ably with those obtained from t h e f l i g h t data.

CONCLUDING RENAFXS

I n general, t h e o v e r a l l impression from this preliminary comparison

i s what would be expected from similar comparisons with wind-tunnel r e s u l t s .
Briefly, a reasonable approximation of t h e span loadings can be determined
f o r t h e low and moderate angle-of-attack range. The estimation of t h e
chord loadings i s less satisfactory, p a r t i c u l a r l y i n t h e neighborhood of
a Mach number of 1.0. In general, the calculated normal-force curve slopes
compare favorably with those obtained from t h e f l i g h t data.
 8 ..

REFERENCES

1. DeYoung, John, and Harper, Charles W.: Theoretical Symmetric Span

Loading a t Subsonic Speeds f o r Wings Having Arbitrary Plan Form.
NACA Rep. 921, 1948.

2. DeYoung, John: Theoretical Symmetric Span Loading Due t o Flap Deflec-

t i o n f o r Wings of Arbitrary Plan Form a t Subsonic Speeds. NACA
Rep. 1071, 1952. (Supersedes NACA TN 2276.)

3 . Glauert, H.: The Elements of Aerofoil and Airscrew Theory. Second

ed., Cambridge Univ. Press, 1947. (Reprinted 1948.)

4. Allen, H. Julian: Calculation of t h e Chordwise Load Distribution

Over A i r f o i l Sections With Plain, S p l i t , or S e r i a l l y Hinged Trailing-
Edge Flaps. NACA Rep. 634, 1938.

5. Mangler, K. W.: Calculation of t h e Pressure Distribution Over a Wing

a t Sonic Speeds. R. & M. No. 2888, B r i t i s h A.R.C., Sept. 1951.

6. Crigler, John L.: Comparison of Calculated and Experimental Load

Distributions on Thin Wings a t High Subsonic and Sonic Speeds.
NACA TN 3941, 1957.

7. Guderley, Gottfried: The Flow Over a F l a t P l a t e With a Small Angle

of Attack a t Mach Number 1. Jour. Aero. Sci., vol. 21, no. 4,
Apr. 1954, pp. 261-274.

8. Guderley, Gottfried, and Yoshihara, Hideo: Two-Dimensional Unsymmetric

Flow Patterns a t Mach Number 1. Jour. Aero. Sci., vol. 20, no. 11,
Nov* 1953, PP- 757-7600
9. Piland, Robert 0.: Summary of t h e Theoretical L i f t , Damping-in-Roll,
and Center-of-Pressure Characteristics of Various Wing Plan Forms
a t Supersonic Speeds. NACA TN 1977, 1949.

10. Harmon, Sidney M., and Jeffreys, Isabella: Theoretical L i f t and

Damping i n Roll of Thin W i n g s With Arbitrary Sweep and Taper a t
Supersonic Speeds -
Supersonic Leading and T r a i l i n g Edges. NACA
TN 2114, 1950-

11. Malvestuto, Frank S., Jr., Margolis, Kenneth, and Ribner, Herbert S.:
Theoretical L i f t and Damping in Roll a t Supersonic Speeds of Thin
Sweptback Tapered Wings With Streamwise Tips, Subsonic Leading Edges,
and Supersonic T r a i l i n g Edges. NACA Rep. 970, 1950. (Supersedes
NACA TN 1860.)

i 59
 b 0
sm
a* 0 .
e 0
a
.
I
0 a

12. Tucker, Warren A., and Nelson, Robert L.: Theoretical Characteristics
in Supersonic Flow of Two Ty-pes of Control Surfaces on Triangular
Wings. NACA Rep. 939, 1949. (Supersedes NACA TN's 1600 and 1601
by Tucker and TN 1660 by Tucker and Nelson.)

13. Malvestuto, Frank S., Jr., and Hoover, Dorothy M.: L i f t and Pitching
Derivatives of Thin Sweptback Tapered W i n g s With Streamwise Tips
and Subsonic Leading Edges a t Supersonic Speeds. NACA TN 2294,
1951.
14. Hannah, Margery E., and Margolis, Kenneth: Span Load Distributions
Resulting From Constant Angle of Attack, Steady Rolling Velocity,
Steady Pitching Velocity, and Constant Vertical Acceleration f o r
Tapered Sweptback Wings With Streamwise Tips - Subsonic Leading
Edges and Supersonic T r a i l i n g Edges. NACA TN 2831, 1952.

15. M a r t i n , John
C., and Jeffreys, Isabella: Span Load Distributions
Resulting F r o m Angle of Attack, Rolling, and Pitching for Tapered
Sweptback Wings With Streamwise Tips -
Supersonic Leading and Trailing
Edges. NACA TN 2643, 1952.

16. Martin, John C., Margolis, Kenneth, and Jeffreys, Isabella: Calcula-
t i o n of L i f t and Pitching Moments Due t o Angle of Attack and Steady
Pitching Velocity a t Supersonic Speeds f o r Thin Sweptback Tapered
Wings With Streamwise Tips and Supersonic Leading and T r a i l i n g Edges.
NACA TN 2699, 1952.

17. Jones, Robert T.: Properties of Low-Aspect-Ratio Pointed W i n g s a t

Speeds Below and Above the Speed of Sound. NACA Rep. 835, 1946.
(Supersedes NACA TN 1032.)

18. Brown, Clinton E., and Michael, W i l l i a m H., Jr.: On Slender Delta
Wings With Leading-Edge Separation. NACA TN 3430, 1955.

I 60
.
a .
c .
*t ... 1
.
,
' e
.
. e .
.
I -
s
-
,
e
-
. , 0..
e..

10

BIBLIOGRAPHY

Banner, Richard D., Reed, Robert D., and Marcy, W i l l i a m L.: Wing-Load
Measurements of t h e Bell X-5 Research Airplane at a Sweep Angle of
58.7O. NACA RM H55Al1, 1955.
Cole, J. D., Solomon, G. E., and Willmarth, W. W.: Transonic Flow.Past
Simple Bodies. (Contract AF 18(600) 383; NAw-6154), GALGIT, 1953.

Few, Albert G., Jr., and Fournier, Paul G.: Wind-Tunnel Investigation
of t h e Aerodynamic Characteristics of a S e r i e s of Swept, Highly Tapered,
Thin Wings a t Transonic Speeds -
Transonic-Bump Method. NACA RM L56I24,
1956.
Heaslet, Max. A., and Fuller, FranMyn B.: P a r t i c u l a r Solutions f o r Flows
a t Mach Number 1. NACA TN 3868, 1956.

Heaslet, Max. A., and S p r e i t e r , John R.: Three-Dimensional Transonic

Flow Theory Applied t o Slender Wings and Bodies. NACA TN 3717, 1956.

Henderson, Arthur, Jr. : Wind-Tunnel Investigation of t h e S t a t i c Longi-

t u d i n a l and L a t e r a l S t a b i l i t y of a 1 / 6 2 - ~ c a l eModel of t h e X-33 a t
Supersonic Speeds. NACA RM L56C23b, 1956.

Henderson, Arthur, Jr. : Wind-Tunnel Investigation of t h e S t a t i c Longi-

t u d i n a l and L a t e r a l S t a b i l i t y of t h e B e l l X - l A a t Supersonic Speeds.
NACA RM L55123, 1955.

Holder, D. W.: Note on t h e Flow Near the T a i l of a Two-Dimensional

Aerofoil Moving a t a Free-Stream Mach Number Close t o Unity. C.P.
No. 188, B r i t i s h A.R.C., June 30, 1954.

Keener, E a r l R., and Jordan, Gareth H.: Wing Loads and Load Distribu-
t i o n s Throughout the L i f t Range of the Douglas X-3 Research Airplane
a t Transonic Speeds. NACA RM H56G13, 1956.

Keener, E a r l R., and Jordan, Gareth H.: Wing Pressure Distributions

Over t h e L i f t Range of t h e Convair XF-92A Delta-Wing Airplane a t Sub-
sonic and Transonic Speeds. NACA RM H55GO7, 1955.

K u h l , Albert E., and Johnson, Clinton T.: Flight Measurements of Wing

Loads on t h e Convair XF-92A Delta-Wing Airplane. NACA RM H55Dl2, 1955.

Kuhn, Richard E., Hallissy, Joseph M., Jr., and Stone, Ralph W., Jr.:
A Discussion of Recent Wind-Tunnel Studies Relating t o t h e Problem
of Estimating Vertical- and Horizontal-Tail Loads. NACA RM L55El6a,
1955 9
Lindsey, Walter F., and Johnston, Patrick J.: Some Observations on
Maximum Pressure Rise Across Shocks Without Boundary-Layer Separation
on A i r f o i l s a t Transonic Speeds. NACA TN 3820, 1956.

Mayer, John P., and Hamer, Harold A.: A Study of Means f o r Rationalizing
Airplane Design Loads. NACA RM L55E13aY 1955.

Moseley, W i l l i a m C. , Jr. : Investigation a t Transonic Speeds of t h e Hinge-

Moment Characteristics of a l/&Scale Model of t h e X-I3 Aileron. NACA
RM L55FO6ay 1955.

Neumark, S. : C r i t i c a l Mach Numbers f o r Thin Untapered Swept W i n g s a t

Zero Incidence. R. & M. No. 2821, B r i t i s h A.R.C., Nov. 1949.

Neumark, S., and Collingbourne, J.: Velocity Distribution on Thin Tapered

Wings With Fore-and-Aft Symmetry and Spmwise Constant Thickness Ratio
a t Zero Incidence. R. & M. No. 2878, B r i t i s h A.R.C., June 1951.

Nielsen, Jack N., Kaattari, George E., and Anastasio, Robert F. : A Method
for Calculating t h e L i f t and Center of Pressure of Wing-Body-Tail
Combinations a t Subsonic, Transonic, and Supersonic Speeds. NACA
RM A53GO8, 1953-

Nielsen, Jack N., Spahr, J. Richard, and Centolanzi, Frank: Aerodynamics

of Bodies, Wings, and Wing-Body Combinations at High Angles of Attack
and Supersonic Speeds. NACA RM A55L13cY 1956.

Polhamus, Edward C.: A Simple Method of Estimating t h e Subsonic L i f t

and Damping i n Roll of Sweptback W i n g s . NACA TN 1862, 1949.

Polhamus, Edward C.: Summary of Results Obtained by Transonic-Bump

Method on E f f e c t s of P l a n Form and Thickness on L i f t and Drag Character-
i s t i c s of W i n g s a t Transonic Speeds. NACA TN 3469, 1955. (Supersedes
NACA RM L5lH30.)

Robinson, Glenn H., Cothren, George E., Jr., and Pembo, C h r i s : Wing-
Load Measurements a t Supersonic Speeds of t h e Douglas D-558-11 Research
Airplane. NACA IiM H54L27, 1955.

Runckel, Jack F., and Gray, W. H.: An Investigation of Loads on Ailerons

a t Transonic Speeds. NACA RM L53E13, 1955.

Shapiro, Ascher H.: The Dynamics and Thermodynamics of Compressible

Fluid Flow. Vol. 11. The Ronald Press Co., 1954.

S p r e i t e r , John R.: On t h e Applica%&n 0% Transonic S i m i l a r i t y Rules t o

Wings of F i n i t e Span. NACA Rep. 1153, 1953. (Supersedes NACA TN 2726.)

I 62
 ...
0..

S p r e i t e r , John R., and Alksne, Alberta: Theoretical Prediction of

Pressure Distributions on Nonlifting A i r f o i l s a t High Subsonic Speeds.
NACA Rep. 1217, 1955. (Supersedes NACA TN 3096.)

Van Dyke, Milton D.: The S i m i l a r i t y Rules f o r Second-Order Subsonic

and Supersonic Flow. NACA TN 3875, 1957.

Vincenti, Walter G.: Measurements of the Effects of F i n i t e Span on the

Pressure Distribution Over Double-Wedge W i n g s a t Mach Numbers Near
Shock Attachment. NACA TN 3522, 1955.

Vincenti, Walter G., Dugan, Duane W., and Phelps, E. Ray: An Experimental
Study of t h e L i f t and Pressure Distribution on a Double-Wedge P r o f i l e
a t Mach Numbers Near Shock Attachment. NACA TN 3225, 1954.

V h c e n t i , Walter G., Wagoner, Cleo B., and Fisher, Newman H., Jr.: C a l -
culations of t h e Flow Over an Inclined F l a t P l a t e a t Free-Stream Mach
Number 1. NACA TN 3723, 1956.

Vincenti, Walter G., and Fisher, Newman H., Jr.: Calculation of t h e

Supersonic Pressure Distribution on a Single-Curved Tapered Wing i n
Regions Not Influenced by t h e Root o r Tip. NACA TN 3499, 1955.

Vincenti, Walter G., and Wagoner, Cleo B.: Theoretical Study of t h e

Transonic L i f t of a Double-Wedge P r o f i l e With Detached Bow Wave.
NACA Rep. 1180, 1954. (Supersedes NACA TN 2832.)

Vincenti, Walter G., and Wagoner, Cleo B.: Transonic Flow Past a Wedge
P r o f i l e With Detached Bow Wave. NACA Rep. 1095, 1952. (Supersedes
NACA TN's 2339 and 2588.)

Willmath, W i l l i a m W.: The L i f t of Thin A i r f o i l s a t High Subsonic Speeds.

OSR TN s4-168 (Contract AF-18(600)-383) GALCIT, June 1934.

63
 A *=e.
A,X-IE4.0 +
NACA M A W 4 WOIFIED
x-3
h,A =3
e3'.1

WMFlED IEXAGON
* - -
t h 0.045

D -558 II

A = 3.6
A,.- = Y)*
- +4
-

RMlT NACA 63-010

TIP -NACA 63-012

JF-O
I >&
NACA M A 0 0 7

m-
A ~2.2
A, .60* A,. 60.
NACA OOO4-65 MODIFIED NACA 65-006.5
6%CONICAL CAMBER. REFLEXED

Figure 1

FLIGHT REYNOLDS NUMBER AND ANGLE OF ATTACK

,R PERFT. a Dag
30 -
H, FT.
25.m 25.-

I
L_____
I

i
I I

;PRESENT STUDY
I I
I
I I
I ,

0 .4 .8 1.2 1.6 2.0 0 .4 .8 1.2 1.6 2.0

Mach number. M Mach number, M

Figure 2
I . 64
 X-IE UNSWEPT WING. ( t k =0.04)-CHORD LOADING
AT y/d/2 = 0.46
1.6 THEORY -; FLIGHT 0
M = 0.8 M4.O m=19

A c p = 2.7
lo\ n- ”0

f y
:
E
n _A0
u -.t

_ _ _ o\

0 .2 .4 .6 .8 1.0 0 .2 .4 .6 .81.0
VC x/C YC

Figure 3

N
‘
X-IE- UNSWEPT WING (V~/c=0.04)

I .o
.5
0
,

P’
Lf
,
THEORY-;

, I M=0.8
Cn-C/Cav -4
i:bx
SPAN LOADINGS
FLIGHT 0 , -- --
0
X- IE- UNSWEPT WING-CHORD LOAD DISTRIBUTIONS
THEORY-, FLIGHT 0

M= 0.8 M4.O M =1.9

a=4O a=4O a=4O
12o

Figure 5

X-IE UNSWEPT WING-CHORD LOAD DISTRIBUTIONS

THEORY -
, FLIGHT 0
M = 1.0 M = 1.0 M = 1.9
u=80 a-7" U =I 4 O

12.0

Figure 6
i
I . 66
D-558-II-SWEPT WING ( Y ~ ~ 0 . 0 9CHORD
)- LOADING AT y/b&=O4I
THEORY - ,FLIGHT 0
M 3 0.8 M 21.0 M = 1.8

a =go

A cP ~~~~

0 .2 4 .6 B 1.0 0. . .
x/C x/C x/C

Figure 7

D-558-II-SWEPT WING (+/c=0.09) SPAN LOADINGS

THEORY -; FLIGHT 0, --O--

Figure 8
G7
 D-558-II-SWEPT WING-CHORD LOAD DISTRIBUTIONS
THEORY-, FLIGHT o

M= 0.8 M = 1.0 M = 1.8

a= 4 O a=3O a = 5'
12.0 72.0 12.0

Figure 9

D-5587I- SWEPT WING - CHORD LOAD DISTRIBUTION

THEORY-, FLIGHT 0

M= 0.8 M= 1.0 M = 1.8

Figure 10
JF-102A AIRPLANE-SCHEMATIC OF WING

Figure 11

JF-IOBA-DELTA WING (t/c=O.O4)-CHORD LOADING AT y/uk ~ 0 . 3 4

THEORY- (NO CAMBER), FLIGHT 0
M= .8 M=I.O M.1.2

.8 a=5' c a=5O
ACP
.4

1.6

1.2

.8

.4

-.4

-.8
.2 .4 6 .8 1.0 0 .2 A .6 .8 1.0 0 .2 .4 6 .8 1.0
vc VC Yc

Figure 12
OE

JF-102A-DELTA WING (tkz0.04) SPAN LOADINGS

THEORY-; FLIGHT 0 , -- 0 --
1.6
M = 0 . 8 1.4
1.2
ID

.4
.2
0
M= 1.0 .8 r a

CN
Cm C.CaV*2
, , ., , ,
0 n-
0 4 8 12 16 20 0 .2 .4 .6 .8 1.0

Figure 13

VARIATION OF NORMAL-FORGE COEFFICIENT WITH ANGLE OF ATTACK

THEORY -; FLIGHT ," ~ ~ ~ $ ' . o
UNSWEPT WING SWEPT WING DELTA WING
X-IE D-558-1 I JF-102A

0 4 8 1 2 0 4 8 1 2 0 4 8 12
WING ANGLE OF ATTACK, a, DEG

Figure 14
'i 0
-
.’ .
* -* r *‘ *
.
~
*
...
I
8 0
*
8
r
.
m
.
-
or’ c - .
.9
le-

20

VARINlON OF FORCE COEFFICIENT DERIVATIVE WITH MACH NUMBER

LOW
THEORY-.h; FLIGHT CN (3 ,,,,OD
HIGH
UNSWEPT WINGS SWEPT WINGS DELTA WINGS

-
.I21 0”
cp
X-\E T 0-558 -I I JF-102A

- 0

XF-92A

05 1.0 1.5 2.0 .5 1.0 1.5 2.0 .5 1.0 1.5

MACH NUMBER

Figure 15
 0.
9. . ) . . ..
0 . .- - . .
L
.
9
c, ... ..
a *

A I R IQAD DISTRIBUTIONS ON A FLAPPED WING RESULTING FROM

LEADING-EDGE AND TRAILING-EDGE BIDWING

By H. Clyde McLemore

Langley Aeronautical Laboratory

Results of recent wind-tunnel pressure-distribution t e s t s on a

large-scale boundary-layer-control model w i t h a 4 9 O sweptback wing have
indicated t h a t blowing a i r a t low blowing r a t e s over highly deflected
trailing-edge f l a p s does not produce any unusual flap-load or wing-load
problems a t low speed. Blowing a t very high momentum r a t e s , however,
produces large increases i n f l a p loads and extremely high l o c a l negative
pressures over t h e f l a p leading edge.

INTRODIETION

The current and proposed f i g h t e r airplanes having t h i n highly loaded

wings of low aspect r a t i o place imposing demands on s u i t a b l e h i g h - l i f t
devices f o r acceptable low-speed performance. Conventional leading-edge-
and trailing-edge-flap cmbinations have proved i n wind-tunnel Eind f l i g h t
t e s t s (refs: 1 t o 3 ) t o be inadequate i n t h i s respect i n many cases.
Serious consideration has been given t o methods f o r achieving additional
l i f t augmentation by boundary-layer control or even t o concepts of u t i -
l i z i n g a l a r g e portion of t h e powerplant t h r u s t f o r developing additional
l i f t a t low f l i g h t speeds. .

The boundary-layer-control method, which u t i l i z e s only s u f f i c i e n t

blowing energy t o prevent flow separation, can be expected t o achieve,
as a maximum, the i d e a l l l u i d loading of a given wing configuration w i t h
f l a p s . The concept of u t i l i z i n g a l a r g e amount of engine t h r u s t f o r
l i f t a w e n t a t i o n , however, introduces t h e design concepts f o r short
take-off and landing and v e r t i c a l take-off and landing.

This paper presents an i n i t i a l wind-tunnel pressure-distribution

exploration of l i f t augmentation by blowing and primarily covers a ,
bomdary-layer-control system f o r a fighter-type airplane. Sane of t h e
problems associated w i t h heavily loaded f l a p s , however, are discussed
w i t h a view toward detemining t h e e f f e c t s of very high momentum blowing
r a t e s Gn t h e f l a p loads.
2

The model used i n t h i s investigation w a s a large-scale wing-body-

t a i l configuration. The wing was swept back 490 a t t h e leading edge a.nd
had an aspect r a t i o of 3.5, a taper r a t i o of .O.3, and NACA 65~006a i r f o i l
sections streamwise.

Chordwise pressure d i s t r i b u t i o n s were obtained a t spanwise s t a t i o n s

of 0.31b/2, 0.56b/2, 0.76b/2, and O.gOb/2; however, only t h e data obtainea
a t t h e 0.56b/2 s t a t i o n a r e presented herein. Tests of t h i s investigation
were conducted a t a Reynolds number of 3.2 x 106 which corresponds t o a
Mach number of 0.08.

SYMBOLS

CL l i f t coefficient

C wing chord

Cf f l a p chord

f l a p section normal-force c o e f f i c i e n t

f l a p blowing mamentun coefficient, WVj/gQ

f l a p chordwise center of pressure, percent f l a p chord

section pressure coefficient

distance along wing chord, measured fram leading edge

W weight flow of s l o t ejected a i r , lb/sec

velocity of s l o t ejected a i r

acceleration due t o gravity

s, free-stream dynamic pressure

A' aspect rat i o

b wing span

A taper r a t i o

s wing a r e a \

B.L.C bowdary-layer control

r $3
Subscript:

MAX maximum

DISCUSSION

"he geometric c h a r a c t e r i s t i c s of t h e large-scale pressure-distribution

model and a general description of t h e wing configura'tions t o be discussed
are presented i n f i g u r e 1. The 24-percent-chord f l a p extends over 50 per-
cent of t h e semispan. The leading-edge flow-control devices were a
15-percent-chord s l a t and a 17-percent-chord droop. The drooped-nose
configuration incorporated a blowing s l o t a t t h e "knee" of the droop,
formed by t h e t r a n s i t i o n curvature between t h e drooped nose and the wing
upper surface.

"he t h r e e wing configurations of figure l w e r e selected because they

represent t h r e e r e l a t i v e l y good landing configurations from the viewpoint
of having t h e l a r g e s t possible values of l i f t c o e f f i c i e n t commensurate
w i t h acceptable longitudinal s t a b i l i t y c h a r a c t e r i s t i c s up t o t h e s t a l l .
For configuration I, a 60-percent-span slat was t h e l a r g e s t spanwise
extent that would produce longitudinal s t a b i l i t y a t maximum l i f t . With
f l a p blowing applied, a full-span leading-edge device i s required t o
maintain t h e l i f t effectiveness t o high angles of attack; therefore,
configuration I1 incorporated a full-span slat. Configuration I11 i s
an a l t e r n a t e method of maintaining t h e l i f t effectiveness t o high angles
of attack.

For t h e nonblowing configuration (configuration I) t h e trailing-edge

f l a p was deflected 40°. With f l a p blowing applied, however, the f l a p w a s
deflected 6 0'. The leading-edge s l a t and t h e leading-edge droop w e r e
deflected 330 and 450, respectively, measured normal t o t h e wing leading
edge.

The chordwise pressure d i s t r i b u t i o n s a t s t a t i o n 2 (oO56b/2) f o r

configurations I and I1 a t a constant value of l i f t c o e f f i c i e n t of 1.27
a r e shown i n f i g u r e 2. The f l a p blowing r a t e used f o r configuration I1
was only s u f f i c i e n t t o preven-b flow separation over the f l a p through the
l i f t range. The blowing f l a p of configuration I1 experienced a s i g n i f i -
cant increase i n load as canpared w i t h the nonblowing case, and t h e
blowing a l s o induced some additional load over the r e a r portion of t h e
wing chord. It should be noted t h a t t h e most s i g n i f i c a n t loading phenom-
enm i s t h e l a r g e l o c a l surface load over t h e f l a p , especially i n the
region of t h e f l a p leading edge.

The load on configuration I i s concentrated near t h e wing leading

edge i n a nomLalmmler. With f l a p blowing applied, however, t h e center
of load i s seen t o be concentrated more trw d t h e r e a r portion of t h e
wing chord. t 8%
4

The chordwise pressure d i s t r i b u t i o n of configuration I1 of f i g u r e 2

i s canpared i n f i g u r e 3 w i t h t h e chordwise pressure d i s t r i b u t i o n of con-
f i g u r a t i o n I11 a t t h e same l i f t coefficient of 1.27. Configuration I11
has drooped-nose-knee and trailing-edge-flap blowing i n an amount only
s u f f i c i e n t t o control flow separation through t h e l i f t range. Mainly
because of t h e absence of a flow-control device located ahead of t h e
wing leading edge, configuration I11 has t h e center of load concentrated
s t i l l farther rearward.

Although one e f f e c t of f l a p blowing is t o move t h e wing chordwise

center of pressure rearward, which would require somewhat l a r g e r t a i l
balancing loads, experiments have shown t h a t a t a i l of normal tail-volume
c o e f f i c i e n t suitably located w i t h respect t o t h e wing d m w a s h f i e l d and
operating within i t s low-speed l i f t c a p a b i l i t i e s could supply these addi-
t i o n a l balancing loads.

The spanwise center-of -pressure c h a r a c t e r i s t i c s of configurations I,

11, and I11 a r e not presented i n t h i s paper. It has been determined,
however, t h a t t h e maximum spanwise center-of-pressure movement caused by
f l a p blowing was a maximum outboard movement a t CL = 1.27 of about
3 percent of t h e semispan. This outbowd movement w a s no greater than
t h a t experienced by t h e nonblowing configuration a t l i f t coefficients
of the order of 0.8. Flap blowing, therefore, should not introduce addi-
t i o n a l low-speed wing-root bending moments.

Probably t h e most s i g n i f i c a n t information i s presented i n figures 4

t o 6. The chordwise pressure d i s t r i b u t i o n ( s t a t i o n 2) of Configuration I11
a t an angle of a t t a c k of 12.3’ f o r a wide range of f l a p blowing r a t e s i s
shown in f i g u r e 4. The blowing momentum coefficient CIL,f of 0.031 i s
representative of t h e boundary-layer-control case; whereas, t h e value of
of 0.187 represents mass-flow r a t e s of t h e order of 40 t o pounds
%,f
of a i r per second. If t h e loading over t h e f l a p is observed, it i s noted
t h a t the f l a p loads and t h e i r contribution t o the moment on t h e r e a r - s p a r
a r e a function of t h e blowing momentum. For t h e boundary-layer-control
case (Cp,f = 0.031) t h e conventional methods of potential-flow calcula-
t i o n s a r e applicable f o r calculating the section loading c h a r a c t e r i s t i c s
( r e f . 4 ) . The other cases shown i n figure 4 (the higher blowing r a t e s )
have not been examined extensively frm t h e viewpoint of calculating t h e
section loads. The form of these pressure d i s t r i b u t i o n s ( t h e i r smoothness
and uniform buildup with increasing blowing r a t e ) suggests, however, t h a t
a method can be determined t o calculate t h e section loadings a t these high
blowing r a t e s .

It should be noted t h a t not only does the load build up over the r e a r
portion of t h e f l a p and over the main wing but a l s o an appreciable load
buildup occurs over t h e f l a p leading edge. The maximum negative pressure
 ..
0.

coefficients over t h e f l a p leading edge w e r e too l a r g e f o r t h e higher

blowing r a t e s t o p l o t in figure 4; however, t h e maximum values are l i s t e d .
For t h e curve of CCL,f = 0.187, t h e peak negative pressure c o e f f i c i e n t
over t h e f l a p leading edge w a s -26.48.

In order t o show t h e o v e r a l l perspective of t h e wing and f l a p loads

f o r a high f l a p blowing rate, t h e curve f o r Cp,f = 0.187 of figure 4
i s shown p l o t t e d t o a g r e a t l y ccarrpresBed s c a l e i n figure 5. The load
over t h e f l a p i s seen t o be a l a r g e portion of t h e t o t a l load, and t h e
peak negative pressures e x i s t i n g over t h e f l a p leadihg edge axe extremely
high a s compared w i t h t h e peak pressures e x i s t i n g a t the wing leading
edge.

In order t o obtain a more general, and perhaps more informative,

outlook on t h e significance of these l a r g e indicated f l a p loads w i t h
increasing blowing rate, the f l a p section normal-force c o e f f i c i e n t s
together w i t h t h e f l a p centers of pressure f o r configuration I11 f o r a
wide range of f l a p blowing r a t e s are shown as functions of l i f t c o e f f i -
c i e n t i n f i g u r e 6.

The f l a p section normal-force c o e f f i c i e n t s a r e seen t o increase

g r e a t l y w i t h increased blowing r a t e but, as i n the case of a normal
unseparated f l a p , increasing l i f t c o e f f i c i e n t has no s i g n i f i c a n t e f f e c t
on t h e f l a p normal-force coefficients. It should be noted here t h a t t h e
l a r g e values of c a r e associated w i t h t h e l a r g e f l a p blowing rates
n,f
CCL,f
which a r e , i n turn, inversely proportional t o the square of t h e
f l i g h t speed, i f t h e momentum of t h e blowing j e t i s assumed constant.
A t r e l a t i v e l y high flight speeds t h e momentum c o e f f i c i e n t s a r e small.
These high values of cn,f a t t h e high Cp,f r a t e s are, therefore,
necessarily r e s t r i c t e d t o low t o moderate f l i g h t speeds i n t h e l a t t e r
phase of t h e landing approach and flare.

The f l a p center of pressure i s seen t o move toward t h e f l a p leading

edge w i t h increasing blowing rate because of t h e high loading over t h e
f l a p leading edge, and t h i s movement produces some compensating r e l i e f
of t h e f l a p monents. Increasing l i f t c o e f f i c i e n t a l s o has no s i g n i f i c a n t
e f f e c t on t h e location of t h e f l a p center of pressure.

CONCLUDING REMARKS

The r e s u l t s of wind-tunnel pressure-distribution t e s t s of a large-

s c a l e bo-mdary-layer-control model with a 490 sweptback wing i n d i c a t e
t h a t , f o r t h e r e l a t i v e l y small amount of blowing necessary t o prevent
flow separation over t h e f l a p , no unusual flap-lo& or wing-load p r o b l m s
a r e anticipated. When a considerable amount of additional blowing energy
 I

i s used, however, l a r g e increases i n f l a p loads can be expected; p a r t i c -

ularly, t h e loads a r e extremely high near t h e f l a p leading edge and may
be t h e c r i t i c a l design f l a p load i n t h e landing maneuver.

REFERENCES

1. McLemore, H. Clyde, and Fink, Marvin P.: Blowing Over the Flaps and
Wing Leading Edge of a "hin 4 9 O Swept Wing-Body-Tail Configuration
i n Combination With Leading-Edge Devices. NACA RM ~ 5 6 ~ 1 61956.
,
2. Kelly, Mazk W., and Tolhurst, W i l l i a m H., Jr.: Full-scale Wind-Tunnel
Tests of a 35' Sweptback Wing Airplane With High-Velocity Blowing
Over t h e Trailing-Eage Flaps. NACA RM A55109, 2-955.

3. Anderson, Seth B., Quigley, Hervey C., and Innis, Robert C.: Flight
. MeasurerLentsof t h e Low-Speed Chazacteristics of a 350 Swept-Wing
Airplane With Blowing-Type Bomdary-Layer Control on t h e Prailing-
Edge Flaps. NACA RM A56G30, 1956.

4. Allen, H. J u l i a : Calculation of the Chordwise Load Distribution Over

A i r f o i l Sections With Plain, S p l i t , or S e r i a l l y Hinged Trailing-Edge
Flaps. NACA Rep. 634, 1938.
E

GEOMETRIC CHARACTERISTICS OF LARGE-SCALE

PRESSURE-DISTRIBUTION MODEL

65A006
A.13.5
x.0.3
S.224 SQ FT
n
II
HIGH-LIFT AND FLOW-
CONTROL DEVICES

d y z o oI
PRE 080 b/2 SLAT
OR1

FULL-SPAN SLAT

28 FT 4 FULL-SPAN 'DROOP

Figure 1

CHORDWISE PRESSURE DISTRIBUTIONS

FOR CONSTANT C ~ z 1 . 2 7
EFFECT OF FLAP BLOWING

-8

-6

cp -4

-2

0
I I I I' I I
-.2 0 .2 .4 .6 .8 I
x/c

Figure 2 1.
 CHORDWISE PRESSWE DISTRIBUTIONS
FOR CONSTANT C~z1.27

-8
SLAT DROOP
-
EFFECT OF NOSE DROOP

1
FLAP
CONFIGURATION

------- II
m
-6

cp -4

-2

2
-.2 0 .2 .4 .6 .8 I
x/c

Figure 3

VARIATION OF CHORDWISE PRESSURE DISTRIBUTION WITH

INCREASING FLAP BLOWING
CONFIGURATION m; O! 42.5
-8

-6

-4

CP
-2

2
0 .2 .4 .6 B 1
X/C

1. 79 Figure 4
VARlATl ON OF CHORDWISE PRESSURE DISTRIBUTION

-
FOR HIGH MOMENTUM RATE
CONFIGURATION a;t ~ z 1 2 . 5 ~
DROOP FLAP
-28
- 24
- 20
-16
cp -12
-8
-4
0
7
-~

0 .2 .4 .6 .8 1
WC

Figure 5

VARIATION OF FLAP NORMAL-FORCE COEFFICIENT AND FLAP

CENTER OF PRESSURE
CONFIGURATIONIIi

0 LI I 1 1 1 1
1.0 1.2 1.4 1.6 1.8 2.0
CL

Figure 6
 O
1 .
B ... . .
0..
. . e
B
I
m
.
.. .. .. .
. '

CALCULATION OF EXTERNAL-STORE LOADS AND CORRELATION

WITH EXPERIMENT

By Percy J. Bobbitt, H a r r y W. Carlson,

and Albiii 0. Pearson

Langley Aeronautical Laboratory

A theory f o r evaluating t h e mutual interference between a wing and

t i p tank has been extended t o apply t o store-pylon configurations. By
use of t h i s analysis and t h e flow-field formulas of NACA Research
Memorandum L35L30b and NACA Technical Note 3938, t h e o r e t i c a l store-pylon
side-force estimates have been made f o r a number of store-pylon configura-
t i o n s . Considered i n t h i s paper a r e t h e separate e f f e c t s of spm-wise
s t o r e position, chordwise s t o r e position, angle of s i d e s l i p , s t o r e f i n s ,
store-store interference, and fuselage indentation. The experimental
d a t a and theory indicate that t h e pylon and pylon-induced s i d e forces
a r e t h e l a r g e s t contributors t o t h e t o t a l store-pylon s i d e force. A
s h o r t bibliography of recent NACA publications dealing with s t o r e and
m i s s i l e loads i s included.

INTRODUCTION

In reference 1, t h e r e s u l t s were presented of a study made t o deter-

mine t h e a b i l i t y of linearized theory t o p r e d i c t t h e s i d e f o r c e a c t i n g
on wing s t o r e s a t supersonic speeds. Though t h e t h e o r e t i c a l and experi-
mental correlations of t h i s study were made f o r only a limited number of
configurations and one Mach number, indications were t h a t l i n e a r theory
might be capable of doing a s a t i s f a c t o r y job. It w a s apparent a l s o from
t h i s study t h a t additional calculations and c o r r e l a t i o n s were needed t o
r e f i n e t h e t h e o r e t i c a l approach and a t t h e same time further define i t s
l i m i t a t i o n s or c a p a b i l i t i e s .

The purpose of t h e present paper i s t o present t h e results of some

recent calculations and correlations m a d e t o p a r t i a l l y f u l f i l l t h i s need
and a t t h e same time indicate t h e type of data now being obtained a t t h e
National Advisory Committee of Aeronautics which are p e r t i n e n t t o t h e
store-loads problem. O f p a r t i c u l a r i n t e r e s t a r e t h e data presented
herein on pylon loads and pylon-induced s t o r e loads.

. ,
A s h o r t bibliography of recent NACA publications dealing with s t o r e
and missile loads i s included.
I $1
2

SYMBOLS

Free-stream v e l o c i t y
M Mach number,
Velocity of sound i n f r e e stream

a angle of a t t a c k

b wing span

SS maximum s t o r e cross-sectional area

9 free-stream dynamic pressure

AP aspect r a t i o of pylon

bP span of pylon

SP pylon area

Side f o r c e on pylon
qsps

Side force on s t o r e
(Cqs = q%P

Side force on s t o r e
CY,s .=
SSS

Side force of s t o r e and pylon

CY,SP = qss

DISCUSSION

In order t o gain some i n s i g h t i n t o t h e mutual interference e f f e c t s

between a s t o r e and pylon and thus i n t o t h e d i v i s i o n of loads, both a
t h e o r e t i c a l analysis and experimental program have been c a r r i e d out.
I n t h i s section, a s h o r t description of the t h e o r e t i c a l analysis just
r e f e r r e d t o i s presented followed by a discussion of the a b i l i t y of t h i s
theory, when used i n conjunction w i t h t h e flow-field formulas of r e f e r -
ences 1 and 2, t o p r e d i c t a number of experimentally measured e f f e c t s .
Several p o i n t s of i n t e r e s t discussed a t the end of t h i s section is illus-
trated by d a t a obtained f o r s t o r e s situated below t h e wing but not con-
nected w i t h a pylon. Through t h e use of theory, an estimation of what
e.-
 3

the order of magnitude of these data would be i f t h e s t o r e were connected

with a pylon i s made.

Theory For Store-Pylon Side Force

The complex, three-dimensional, nonplanar problem of determining

rigorously t h e s i d e load on a s t o r e and pylon attached t o a f i n i t e wing
i s one of insurmountable d i f f i c u l t y . Obviously, some s i m p l i e i n g assump-
t i o n s must be made and some r i g o r sacrificed i f t h e problem i s t o be
t r e a t e d by a n a l y t i c a l means. One of t h e most p r o f i t a b l e simplifications
which can be affected when evaluating t h e interference e f f e c t s between
t h e store, pylon, and wing i s t h e replacement of t h e wing by an i n f i n i t e
r e f l e c t i o n plane. Inherent i n t h i s simplification i s t h a t t h e magnitude
of t h e lateral-flow velocity beneath t h e wing be determined separately
and by other methods. (See r e f s . 1 and 2. )

The problem of determining t h e loads on a pylon-store r e f l e c t i o n -

plane configuration subjected t o a uniform lateral flow can be analyzed
i n t h e same manner as a wing tip-tank configuration a t an angle of a t t a c k
since t h e i n s e r t i o n of an i n f i n i t e plane, perpendicular t o t h e wing and
i n t h e plane of symmetry, does not a l t e r t h e flow p a t t e r n . When t h e config-
uration of wing, t i p tank, and v e r t i c a l r e f l e c t i o n plane i s ’ r o t a t e d through
Po, t h e wing semispan becames t h e pylon, t h e v e r t i c a l r e f l e c t i o n plane
becomes t h e wing, and t h e tip-tank becomes the s t o r e . Also, the uniform
v e r t i c a l velocity t o which t h e wing and tip-tank a r e subjected and which
i s equivalent t o Va becomes t h e lateral-flow velocity and i s given the
value -Vp. The wing tip-tank problem has been t r e a t e d i n reference 3;
however, t h e r a t i o s of tank diameter t o wing semispan of the configura-
t i o n s f o r which numerical r e s u l t s were obtained were not high enough t o
be of use i n evaluating pylon or s t o r e loads. Consequently t h e a n a l y t i c a l
procedure given i n reference 3 has been u t i l i z e d t o extend t h e numerical
r e s u l t s from values of D/$ up t o 0.3 t o values of D/bp up t o 1.6.

Shown i n figure 1 are t h e variations of t h e pylon and pylon-induced

s t o r e forces i n coefficient form with t h e r a t i o of t h e s t o r e diameter
t o pylon span. The nondimensionalizing area f o r these coefficients i s
t h e pylon area. Note t h a t when t h e s t o r e diameter is 1.6 times t h e pylon
span, t h e pylon-induced s i d e force on t h e s t o r e and the s i d e force on t h e
pylon a r e almost equal. It i s conceivable t h a t i n t h i s s i t u a t i o n the
pylon could contribute on t h e order of 20 t o 25 percent t o t h e t o t a l root-
bending moment.

Scope of T e s t Configurations
The curves given i n figure 1 and t h e flow-field formulas of r e f e r -
ences 1 and 2 have been u t i l i z e d t o make side-force calculations f o r a
number of configurations recently t e s t e d a t t h e Langley Aeronautical
Laboratory. Before m&ing comparisons of t h e calculated and experimental
4

r e s u l t s , it would seem appropriate t o present a f e w d e t a i l s of the tests

and t e s t configurations.

Some idea of t h e scope of tests can be obtained from the two wing-
body models and associated s t o r e and store-pylon configurations shown i n
f i g u r e 2. The wing-body and store-pylon configurations on t h e l e f t of
f i g u r e 2 were t e s t e d a t a Mach number of 1.6 i n the Langley 4- by 4-foot
supersonic pressure tunnel; those on t h e right, a t Mach numbers from
0.8 t o 1.43 i n t h e Langley 8-foot transonic pressure tunnel. In t h e
4-foot-tunnel t e s t s , s t o r e and store-pylon loatis have been obtained f o r
a v a r i e t y of store-pylon configurations and a number of spanwise and
chordwise locations as indicated by t h e c i r c l e s showing s t o r e midpoint
locations. Five-cmponent s t o r e forces and maments were measured i n
t h i s investigation w i t h t h e s t o r e i n presence of t h e pylon, three canpo-
nents were measured f o r t h e store-pylon combination and t h e f u l l s i x cam-
ponents f o r the wing body. The m o d e l s w e r e tested through an angle-of-
a t t a c k range extending f r a m -2O t o 12O and an angle-of-sideslip range
f r m - 1 2 0 t o 120.

The wing-fuselage model and the two s t o r e shapes pictured on t h e

right of figure 2 represent two separate investigations conducted i n the
Langley 8-foot transonic pressure tunnel. For the s t o r e without a pylon
systematic .investigations have been conducted using both t h e basic and
contoured fuselages shown. The c i r c l e s on t h e wing body a t t h e right
i n d i c a t e the positions a t which s t o r e forces and mments were obtained.
I n addition t o being t e s t e d singly, s e v e r a l of the s t o r e s w e r e t e s t e d i n
canbination, t h a t is, two s t o r e s on a panel. Forces and mments f o r t h e
finned s t o r e were measured only on t h e contoured fuselage airplane con-
f i g u r a t i o n a t two positions. These positions are indicated by t h e squares.
In each of t h e investigations, five-canponent s t o r e forces and maments and
three-cmponent wing-fuselage forces and mments were obtained through an
angle-of-attack range extending frm -2' t o 8O.
There i s , evideutly, a l a r g e amount of recent experimental data which
can be u t i l i z e d t o t e s t t h e worth of t h e theoreticalmethods. In subse-
quent f i g u r e s showing experimental results and calculations f o r a v a r i e t y
of e f f e c t s only a representative cross section of t h i s data w i l l be
utilized.

CONTRIBUTION OF STORE AND PYLON TO COMBINED IDAD

Given i n f i g u r e 3 a r e t h e separate contributions of t h e s t o r e and

pylon t o t h e cmbined store-pylon side-force c o e f f i c i e n t a t a Mach number
of 1.6. The angle of s i d e s l i p i s zero as it i s f o r a l l other f i g u r e s
unless otherwise noted. Data and curves shown a r e f o r a sweptforward
pylon and s t o r e combination located a t t h e 0.5>-wing-semispan s t a t i o n .
The v e r t i c a l s c a l e of f i g u r e 3, which i s labeled side-force c o e f f i c i e n t ,
i s simply t h e p a r t i c u l a r side force being considered nondimensionalized *-
E
5

by q and t h e maximum cross-sectional area of t h e s t o r e . Negative side-

force coefficient indicates t h a t t h e side force a c t i n g on t h e s t o r e o r
store-pylon located below the left-hand half-wing i s directed toward t h e
wing t i p .

In t h e calculation of t h e t o t a l store-pylon load, t h e loads on t h e

s t o r e and pylon a r e determined separately and added, The load a c t i n g
on t h e s t o r e i s i t s e l f f u r t h e r broken down i n t o two separate canponents.
One i s t h e load on t h e s t o r e determined as i f t h e pylon were not present
and the other i s t h e load induced on t h e s t o r e by t h e pylon. It should
be noted t h a t t h e l a t t e r of these two loads is determined by the use of
f i g u r e 1. Fortunately f o r the s t o r e s used i n t h e investigation i n t h e
Langley 4- by 4-foot supersonic pressure tunnel, experimental data i s
available f o r each of t h e store-load camponents i n addition t o t h e t o t a l
store-pylon load. Consider f i r s t the side force a c t i n g on t h e s t o r e when
no pylon i s present. It i s evident that f o r t h i s condition t h e rate of
change of t h e s i d e force w i t h angle of a t t a c k i s underestimated by t h e
theory. This is not surprising since a calculation of t h e l i f t - c u r v e
slope f o r t h e s t o r e alone i n a uniform flow f i e l d underestimates experi-
ment by about 50 percent. The use of s t a t i c experimental. force data f o r
t h e isolated s t o r e t o form a correction f a c t o r f o r the t h e o r e t i c a l compu-
t a t i o n s would lead t o a more accurate estimate of t h i s canponent i f
desired.

The t o t a l side force acting on t h e s t o r e when t h e pylon i s added

(see f i g . 3 ) , agrees r a t h e r well w i t h the calculated curve. Evidently
this agreement i s f o r t u i t o u s since the increment i n the s t o r e side f o r c e
caused by t h e i n s e r t i o n of t h e pylon is overpredicted; that is, t h e d i f -
ference between t h e squares and c i r c l e s i s overpredicted. Perhaps the
f a c t t h a t t h e t h e o r e t i c a l curves used t o estimate t h i s increment were
meant primarily f o r an unswept pylon and a l s o t h a t t h e t h e o r e t i c a l side-
wash acted only over t h e r e a r portion of t h e pylon are p a r t l y responsible
f o r t h e overprediction.

The t o t a l force acting on the s t o r e and pylon combination as deter-

mined by experiment and given by t h e diamonds agrees s a t i s f a c t o r i l y with
theory which i s the s o l i d l i n e . Contributing t o t h i s agreement i s t h e
overprediction of t h e slope and the underprediction of the zero angle of
a t t a c k or thickness e f f e c t s . It i s i n t e r e s t i n g t o note t h a t the pylon,
carrying about t h e same amount of side load as the store, contributes
about 25 percent of the t o t a l root-bending moment.

Before leaving figure 3, it i s of i n t e r e s t t o translate t h e canbined

store-pylon side-force coefficient i n t o terms of a load i n pounds f o r a
p r a c t i c a l s i t u a t i o n . For example, i f an airplane of t h i s configuration
having a 40-foot wing span and operating a t 40,000 f e e t and a Mach number
of 1.6 were t o experience the l a r g e s t store-pylon force indicated, it
would amount t o 6,600 pounds.

85
6

EFFECT OF STORE SPANWISE LOCATION ON THE STORE-PYIDN

SIDE-FORCE COEF'F'ICIENT

There a r e other limitations and inadequacies of t h e t h e o r e t i c a l

methods used which have not y e t been mentioned that w i l l be pointed out
a t t h e appropriate places subsequently. An excellent opportunity t o
i l l u s t r a t e one of these shortcomings i s afforded by f i g u r e 4.

Figure 4 shows t h e e f f e c t of s t o r e spanwise location on t h e angle-

of-attack v a r i a t i o n of the store-pylon side-force coefficient. The
r e l a t i v e position of t h e store-pylon configuration w i t h respect t o t h e
wing leading edge remains t h e same f o r a l l three spanwise locations as
indicated by t h e sketches. Looking first a t t h e experimental points,
it can be seen t h a t as you go from t h e inboard t o t h e outboard position
t h e r e is an increase i n the slope. Theoretical predictions using t h e
supersonic leading-edge sidewash calculated f o r t h e free-stream Mach nm-
ber of 1.6 ( t h e s o l i d - l i n e curves) a l s o show an increase i n slope as t h e
s t o r e moves outboard, but t h e r a t e at which the slope increases is
overpredicted .
Schlieren photographs taken of these models, one of which i s shown
i n f i g u r e 5 , revealed that t h e wing leading edge, instead of being s l i g h t l y
supersonic as predicted by t h e theory, w a s a c t u a l l y subsonic. I n an
attempt t o obtain b e t t e r agreement between t h e t h e o r e t i c a l and experi-
mental slopes, t h e wing angle-of-attack flow f i e l d w a s recalculated f o r
t h e Mach number giving t h e same location of t h e Mach cone emanating from
t h e wing r o o t as indicated i n the schlieren photograph. Use of this f l o w
f i e l d i n determining t h e store-pylon side forces resulted i n t h e dashed-
l i n e curves. For t h e inboard location, calculations using t h e subsonic
edge flow f i e l d did not y i e l d a noticeable change; whereas, f o r the mid-
span and t i p locations, b e t t e r agreement resulted. The agreement i n
magnitudes m i g h t be f u r t h e r improved i f t h e s i d e forces induced by t h e
wing-thickness flow f i e l d were a l s o reevaluated. Because of t h e time
required, only a very rough estimation of t h e wing thickness e f f e c t w a s
made. This is t r u e of most of t h e curves shown i n subsequent figures.

EJ?FECT OF PYLON SWEEP AND STORE CHORDWISE LOCATION

Having examined t h e e f f e c t of spanwise location on a p a r t i c u l a r store-

pylon configuration, t h e next l o g i c a l e f f e c t t o discuss i s that of s t o r e
chordwise location and accompanying pylon sweep, The s p a w i s e p o s i t i o n
f o r which d a t a w i l l be presented i s a t t h e 0.55-wing-semispan s t a t i o n .
The store-pylon configurations a t t h e d i f f e r e n t chordwise locations a r e
shown by t h e sketches i n figure 6. Actually since t h e pylon c a r r i e s or
 7

induces t h e major portion of t h e t o t a l load, t h i s figure m i g h t be more

properly t i t l e d t h e e f f e c t of pylon sweep and location. The experimental
data and t h e o r e t i c a l curve f o r t h e most forward p o s i t i o n are repeated
from t h e previous figures. When t h e s t o r e i s moved t o t h e rear from the
most forward p o s i t i o n and t h e pylon becomes unswept, t h e r e i s an increase
i n the slope and a slight decrease i n t h e thickness e f f e c t s . In t h i s
position, t h e pylon i s almost completely immersed i n t h e wing flow f i e l d
and f e e l s the maximum e f f e c t of the high flow a n g u l a r i t i e s i n t h e region
j u s t t o t h e r e a r of t h e wing leading edge. The agreement between theory
an& experiment f o r t h i s p o s i t i o n is good. For t h e s t o r e i n t h e rear
position, a considerable change i n t h e side-force c o e f f i c i e n t angle-of-
a t t a c k v a r i a t i o n fram those of t h e more forward positions i s noted.
This can be a t t r i b u t e d t o t h e f a c t t h a t t h e sweptback pylon is situated
i n a region of much lower wing angle-of-attack sidewash and a l s o t h a t it
i s subjected t o t h e inflow i n t h e region of t h e t r a i l i n g edge due t o wing
thickness. The t h e o r e t i c a l slope, though somewhat less than it w a s f o r
t h e other two locations, s t i l l overpredicts t h e experimental slope. The
change i n thickness e f f e c t s , estimated, as mentioned before, i n a r a t h e r
rough manner, were a l s o overpredicted. Generally speaking, t h e agreement
f o r t h e t h r e e positions i s s a t i s f a c t o r y .

Effect of Sideslip on t h e Store-Pylon Side-Force Coefficient

Another f a c t o r which has a major influence on the magnitude o f t h e

s i d e force a store-pylon w i l l be required t o carry is t h e angle of s i d e -
s l i p . Plotted i n f i g u r e 7 is t h e v a r i a t i o n of the store-pylon side-force
c o e f f i c i e n t w i t h angle of s i d e s l i p f o r t h e sweptforward pylon and s t o r e
combination located a t t h e 0.55-semispan s t a t i o n . The angle of a t t a c k
chosen t o i l l u s t r a t e t h i s e f f e c t i s 8'. Note t h e l a r g e magnitude of t h e
side-force c o e f f i c i e n t s a t t h e highest angles of s i d e s l i p and a l s o t h e
steepness of the slope. A comparison of experimental curves inJ;zates
t h a t more s i d e force r e s u l t s fram '1 of s i d e s l i p than from lo of angle
of a t t a c k but perhaps not as much more as m i g h t be expected. The theory
used t o obtain t h e s o l i d - l i n e curve assumes that t h e s t o r e and pylon a r e
s i d e slipped i n the presence af an i n f i n i t e f l a t p l a t e hence the e f f e c t
of t h e changing leading-edge sweep relative t o t h e free-stream d i r e c t i o n
and t h e e f f e c t of t h e fuselage a r e neglected. It i s evident t h a t the
simplified theory overpredicts the incremental load due t o s i d e s l i p by
about 70 percent. A rough calculation taking i n t o account t h e changing
leading-edge sweep yielded considerably b e t t e r agreement. It should be
pointed out here t h a t t h e r a t e of change of t h e side-force c o e f f i c i e n t
w i t h angle of s i d e s l i p varied l i t t l e from one p o s i t i o n t o another i n this
investigation.

?.,. uk. .-..

.^_ ., , .
 Effect of Fins on Store Side-Force Coefficient

There a r e many s t o r e s i n use today which employ fins. A n indication

of how important t h e f i n loads a r e r e l a t i v e t o t h e pylon and pylon-induced
s t o r e loads may be obtained from f i g u r e 8. In t h i s f i g u r e is p l o t t e d t h e
v a r i a t i o n of t h e s t o r e side-force c o e f f i c i e n t with angle of attack. The
v e r t i c a l s c a l e has been reduced by a f a c t o r of one-half from t h e previous
f i g u r e s so t h a t t h e changes caused by the addition of t h e fins can be more
e a s i l y seen. The s t o r e pylon and wing body on t h e l e f t a r e t h e same as
t h a t used f o r a number of t h e previous figures and, it may be remembered,
w a s t e s t e d i n t h e Langley 4- by &-foot supersonic pressure tunnel a t a
Mach number of 1.6. The configuration on t h e right i s completely d i f f e r e n t
i n a l l respects and w a s t e s t e d i n t h e Langley 8-foot transonic pressure
tunnel a t a v a r i e t y of Mach numbers i n addition t o t h e Mach number f o r
which d a t a i s given i n t h i s figure. For t h e s t o r e on t h e l e f t , t h e addi-
t i o n of f i n s causes an increase i n t h e slope and only negligible displace-
ment i n t h e zero angle-of-attack value. Theory adequately p r e d i c t s t h e
slope change but overestimates the displacement. For t h e s t o r e on t h e
r i g h t , t h e addition of t h e f i n s causes a displacement i n t h e curves though
l i t t l e change i n slope. Theory f o r t h e fins-off s i t u a t i o n accurately
p r e d i c t s t h e magnitude and r a t e of change w i t h angle of a t t a c k of t h e
s t o r e side-force coefficient. For t h e fins-on s i t u a t i o n only t h e increment
a t zero angle of a t t a c k is predicted. This estimate i s indicated by t h e
t i c k mark. Because t h e f i n c l o s e s t t o t h e wing is affected by t h e wing
vortex wake, no attempt has been made t o c a l c u l a t e t h e angle-of-attack
v a r i a t i o n w i t h t h e f i n s on. For t h e s t o r e on t h e l e f t , t h e e f f e c t of t h e
f i n s i s small when compared t o t h e t o t a l side-force acting on t h e s t o r e
and pylon. For the s t o r e on t h e right where t h e store-pylon s i d e forces
a r e much smaller, t h e e f f e c t of t h e f i n s assumes a more important r o l e .

Interference Effects Between Inboard and Outboard Stores

Another source of s t o r e s i d e forces i s t h e flow f i e l d created by

another s t o r e , p a r t i c u l a r l y when t h i s other s t o r e is located on t h e same
half-wing as i n a four-store o r four-nacelle configuration. To give sane
idea of t h e magnitude of these e f f e c t s and t o indicate whether t h e s t o r e
flow f i e l d s can be accurately predicted, f i g u r e 9 has been prepared. Shown
on t h e l e f t of f i g u r e 9 i s t h e e f f e c t of t h e outboard s t o r e on the s i d e
f o r c e acting on t h e inboard s t o r e and on t h e r i g h t , t h e e f f e c t of t h e
inboard s t o r e on the outboard-store side force. These s t o r e s were not
connected t o the wing with a pylon. "he dashed l i n e s ana t h e c i r c l e s a r e
the theory and experiment f o r t h e s i d e force on t h e s t o r e alone and t h e
s o l i d l i n e and t h e squares f o r the s i d e force on the s t o r e i n t h e presence
of t h e other s t o r e .

The points t o be made r e l a t i v e t o the inboard-store curves on t h e

l e f t a r e t h e l a r g e e f f e c t of t h e outboard s t o r e on t h e inboard and t h e

i 88
 .
I
I .
.I
.. .. . . .
I . , * - 0 .
.e 0. , .,
, . . .. 0

accuracy w i t h which t h i s e f f e c t i s predicted. The displacement of t h e

two curves i s primarily due t o t h e r a d i a l flow away f r o m t h e outboard
s t o r e caused by t h e outboard-store thickness and is, therefore, almost
independent of angle of attack. If t h e inboard s t o r e were connected t o
t h e wing by a pylon, t h e increment i n s t o r e side-force c o e f f i c i e n t
instead of being of t h e order of 0.08 might be as l a r g e as 0.6 or 0.7
depending on t h e geometric c h a r a c t e r i s t i c s of t h e pylon. An e f f e c t of
t h i s magnitude would most c e r t a i n l y have t o be accounted f o r t o obtain
reliable values of t h e store-pylon side force. The calculated curves
shown a r e based on linear theory corrected by use of experimental f o r c e
data f o r t h e i s o l a t e d s t o r e . A n examination of t h e right-hand plot shows
t h a t t h e change i n t h e slope of t h e side-force-coefficient curve f o r t h e
outboard s t o r e when t h e inboard s t o r e is added i s f a i r l y well predicted
by theory. The magnitude agreement between theory and experiment f o r t h e
outboard-store side-force c o e f f i c i e n t s is evidently not so good as f o r
the inboard s t o r e . The reason f o r t h i s i s thought t o be a l o c a l Mach
number loss similar t o that illustrated by t h e schlieren photograph of
f i g u r e 5. I n terms of s t o r e - s t o r e interference, t h i s means t h a t one
s t o r e will a f f e c t t h e other f u r t h e r forward than it would i f t h e l o c a l
Mach numbers w e r e the same as t h e free-stream Mach numbers. Experimental
data i s available which indicates the e f f e c t on t h e store-store i n t e r f e r -
ence of these Mach numbers losses. Some of these data a r e presented i n
figure 10.

Shown i n t h i s f i g u r e are t h e side forces and yawing moments on an

inboard s t o r e , alone and i n the presence of t h e outboard s t o r e , f o r Mach
numbers of 1.43, 1.2, and 0.8. The s t o r e s f o r the t e s t r e s u l t s pictured
were located a t t h e 0.28- and 0.70-wing-semispan s t a t i o n s . The dashed l i n e s
are f o r the inboard s t o r e alone and t h e s o l i d l i n e s a r e f o r t h e inboard i n
t h e presence of t h e outboard s t o r e . A t a Mach number of 1.43, note t h a t t h e
outboard s t o r e has a negligible e f f e c t on t h e inboard s t o r e , whereas a t
a Mach number of 1.2 t h e r e is a l a r g e e f f e c t . If t h e l o c a l Mach nmber
were supersonic, it would be impossible f o r the outboard s t o r e t o e f f e c t
the inboard s t o r e a t a Mach number of 1.2. The f a c t t h a t t h e induced
moment on t h e inboard s t o r e a t zero angle of a t t a c k i s positive, or nose
in, and t h e fact t h a t t h e angle-of-attack v a r i a t i o n s a t t h i s Mach number
are almost i d e n t i c a l t o those a t subsonic speeds, t y p i f i e d by t h e
M = 0.8 curves, indicates t h a t t h e l o c a l flow i s a c t u a l l y subsonic.

Effect of Fuselage Contouring

It i s w e l l known t h a t contoured fuselages are i n use on many of

todays production and design stage airplanes. While e f f e c t i n g a change
i n t h e pressure f i e l d t o y i e l d a lower drag than the basic fuselage
configuration, these indented fuselages a l s o have an appreciable e f f e c t
on t h e lateral flow f i e l d and hence on the loads induced on s t o r e s
attached t o t h e wing. A n indication of t h e difference between t h e loads
on a s t o r e located beneath wing-fuselage m d e l s w i t h e,nd without a fuse-
lage indentation can be obtained from figure 11. "he Mach number of
89
10

experimental r e s u l t s p l o t t e d i n t h i s figure i s 1.2. Data f o r two span-

wise s t o r e positions, 0.28- and 0.5-wing-semispan s t a t i o n s , is depicted.
Judging frm t h e difference between t h e curves f o r t h e basic and indented
fuselages t h e e f f e c t of fuselage indentation is t o cause an increase i n
t h e outward flow i n the v i c i n i t y of t h e leading edge, giving r i s e t o a
negative increment i n t h e s i d e force and negative and p o s i t i v e increments
t o the yawing manents f o r t h e inboard and outboard s t o r e s , respectively.
If t h e s t o r e s were connected t o the wing with a pylon the l a r g e s t incre-
ment shown instead of being approximately 0.05 would be of the order of
0.3 o r 0.4.
The e f f e c t of fuselage indentation on t h e surface pressures and flow
f i e l d s of configurations a t an angle of a-btack has not y e t been analyzed
t h e o r e t i c a l l y . A t zero angle of a t t a c k t h e area-rule papers of Lanax
and Heaslet ( r e f . 4) and Nielsen and P i t t s ( r e f . 5 ) permit an evaluation
of these e f f e c t s though t h e procedure i s r a t h e r involved.

CONCLUDING RENARKS

It can be said t h a t t h e o r e t i c a l methods a r e capable of predicting

s a t i s f a c t o r i l y a variety of e f f e c t s . More d e t a i l i n t h e calculations
of some of these e f f e c t s than used i n the present paper seems desirable.
The experimental data and theory indicate t h a t t h e pylon and pylon-
induced side forces are the l a r g e s t contributors t o t h e t o t a l store-
pylon side force. The separate e f f e c t s of spanwise s t o r e position,
chordwise s t o r e position, angle of s i d e s l i p , s t o r e f i n s , store-store
interference, and fuselage indentation a l l play a p a r t i n determining
t h e store-pylon side loads.
 11

REFERENCES

1. Bobbitt, Percy J., Malvestuto, Frank S., Jr., and Margolis, Kenneth:
Theoretical Prediction of t h e Side Force on Stores Attached t o
Configurations Traveling a t Supersonic Speeds. NACA RM L55L3Ob,
1956.
2. Bobbitt, Percy J., and Maxie, Peter J., Jr.: Sidewash i n t h e Vicinity
of L i f t i n g Swept Wings a t Supersonic Speeds. NACA TN 3938, 1957.

3. IEartley, D. E.:Theoretical Load Distributions on Wings With

Cylindrical Bodies a t t h e Tips. C.P. No. 147, B r i t i s h A.R.C., 1954.
4. L a a x , Harvard, and Heaslet, Max. A. : A Special Method f o r Finding
Body Distortions That Reduce t h e Wave Drag of W i n g and Body Combina-
t i o n s a t Supersonic Speeds. NACA RM ~ 5 5 a 6 ,1955.

5. Nielsen, Jack N., and P i t t s , W i l l i a m C.: General Theory of Wave-Drag

Reduction f o r Canbinations Eknploying Quasi-Cylindrical Bodies With
an Application t o Swept-Wing and Body Combinations. NACA TN 3722,
1956. (Supersedes NACA RM A55aO7.)
0
.e
.
r..
. I
. . ..
b 1
e . . * . ...
, e . * ..I
L..

12

BIBLIOGRAPHY

S i l v e r s , H. Norman, and King, Thomas J., Jr.: I n v e s t i g a t i o n a t High

Subsonic Speeds of Bodies Mounted Fromthe Wing of a n Unswept-Wing-
Fuselage Model, Including Measurements of Body Loads. NACA RM ~52508,
1952.

Robinson, Samuel W., Jr., and Zlotnick, Martin: A Method f o r Calculating

t h e Aerodynamic Loading on Wing-Tip-Tank Combinations i n Subsonic Flow.
NACA RM L53B18, 1953.

S i l v e r s , H. Norman, and O'Bryan, Thomas 6.: Some Notes on t h e Aerody-

namic Loads Associated With External-Store I n s t a l l a t i o n s . NACA
RM L53E06a, 1953.

O'Bryan, Thomas C.: F l i g h t Measurements of Aerodynamic Loads and Moments

on an E x t e r n a l Store Mounted Under t h e Wing of a Swept-Wing Fighter-
Ty-pe Airplane. NACA RM L53G22, 1953.

Alford, W i l l i a m J., Jr., and S i l v e r s , H. Norman: Investigation a t High

Subsonic Speeds of Finned and Unfinned Bodies Mounted a t Various Loca-
t i o n s From t h e Wings of Unswept- and Swept-Wing-Fuselage Models,
Including Measurements of Body Loads. NACA RM L54B18, 1954.

Alford, W i l l i a m J., Jr., S i l v e r s , H. Norman, and King, Thomas J., Jr.:

Preliminary Low-Speed Wing-Tunnel Investigation of Some Aspects of t h e
Aerodynamic Problems Associated With Missiles Carried Externally i n
Positions Near Airplane Wings. NACA RM L5kJ20, 1954.

Alford, W i l l i a m J., Jr., S i l v e r s , H. Norman, and King, Thomas J., Jr.:

Experimental Aerodynamic Forces and Moments at Low Speed of a Missile
M o F l During Simulated Launching Fromthe Midsemispan Iocation of a
45 Sweptback Wing-Fuselage Combination. NACA RM L54Klla, 1955.
Alford, W i l l i a m J., Jr.: Experimental S t a t i c Aerodynamic Forces and
Moments a t Low Speed on a Canard Missile During Simulated h u n c h i n g
From t h e Midsemispan and Wing-Tip I m a t i o n s of a 4 5 O Sweptback Wing-
Fuselage Combination. NACA RM L55Al2, 1955.

Alford, W i l l i a m J., Jr., Silvers, H. Norman, and King, Thomas J., Jr.:
Experimental S t a t i c Aerodynamic Forces and Moments a t Low Speed on
a Missile Model During S h u l a t e d h u n c h i n g From t h e 25-Percent-
Semispan and Wing-Tip Locations of a 45' Sweptback Wing-Fuselage
Combination. NACA RM L59D20, 1955.

Smith, Norman F., and Carlson, Harry W.: Some E f f e c t s of C onf igurat ion
Variables on Store Loads a t Supersonic Speeds. NACA RM L55E05, 1955.
e.-

. ..

L' ..
i' 92
Alford, W i l l i a m J., Jr.: E f f e c t s of Wing-Fuselage Flow F i e l d s on Missile
Loads a t Subsonic Speeds. NACA RM L35ElOa, 1955.

Guy, Lawrence D.: Loads on External Stores a t Transonic and Supersonic

Speeds. NACA RM L55E13b, 1955.

Smith, Norman F., and Carlson, Harry W.: The Origin and Distribution of
Supersonic Store Interference From Measurement of Individual Forces on
Several Wing-Fuselage-Store Configurations. 11.- Swept-Wing Heavy-
Bomber Configuration With Large Store (Naceile). L a t e r a l Forces and
Pitching Moments; Mach Number, 1.61. NACA RM L55E26a, 1955.

Hamer, Harold A., and OfBryan, Thomas C.: F l i g h t Measurements of t h e

Loads and Moments on an External Store Mounted Under t h e Wing of a
Swept-Wing Fighter-Type Airplane During Yawing and Rolling Maneuvers.
NACA RM L55G22, 1 9 5 .

Smith, Norman F., and Carlson, Harry W.: The Origin and Distribution
of Supersonic Store Interference From Measurement of Individual Forces
on Several Wing-Fuselage-Store Configurations. 111.- Swept-Wing Fighter-
Bomber Configuration With Large and Small Stores. Mach Number, 1.61.
NACA RM L55H01, 1957.

Guy, Lawrence D., and Hadaway, W i l l i a m M . : Aerodynamic Loads on an

External Store Adjacent t o a k 5 O Sweptback Wing a t Mach Numbers From
0.70 t o 1.96, Including a n Evaluation of Techniques Used. NACA
RM L55H12, 1955.

Morris, Ode11 A.: The Origin and Distribution of Supersonic Store I n t e r -

ference From Measurement of Individual Forces on Several Wing-Fuselage-
Store Configurations. 1V.- Delta-wing Heavy-Bomber Configuration with
Large Store. Mach Number, 1.61. NACA RM L55127a, 1955.

Carlson, H a r r y W., and Geier, Douglas J.: The Origin and Distribution of
Supersonic Store Interference From Measurement of Individual Forces on
Several Wing-Fuselage-Store Configurations. V.- Swept-Wing Heavy-Bomber
Configuration With Large Store (Nacelle). Mach Number, 2.01. NACA
L55a5, 1956.
Hadaway, W i l l i a m M.: Aerodynamic Loads on an E x t e r n a l Store Adjacent t o
an Unswept Wing a t Mach Numbers Between 0.75 and 1.96. NACA RM L55LO7,
1956.
Smith, Norman F.: The Origin and Distribution of Supersonic Store I n t e r -
ference From Measurement of Individual Forces on Several Wing-Fuselage-
Store Configurations. V I . - Swept-Wing Heavy-Bomber configuration With
Stores of Different Sizes and Shapes. NACA RM L55LO8, 1956.

93
 00 ,oo 0 . 0 - 0 a 3 0 0 0 1 0..
e . 0 I 0 0 a 0 e * . 0 0 - .I.

14

Hallissy, Joseph M., Jr., and Kudlacik, Louis: A Transonic Wind-Tunnel

Investigation of Store and Horizontal-Tail Loads and Some Effects of
Fuselage-Afterbody Modifications on a Swept-Wing Fighter Airplane.
1956.
NACA RM ~ 5 6 ~ 2 6 ,

Hadaway, W i l l i a m M.: Aerodynamic Loads on an External Store Adjacent t o

a 600 Delta Wing a t Mach Numbers From 0.75 t o 1.96. NACA RM L56BO2a,
1956.
Alford, W i l l i a m J., Jr., and King, Thomas J., Jr.: Experimental S t a t i c
Aerodynamic Forces and Moments a t High Subsonic Speeds on a Missile
Model During Simulated Launching From t h e Midsemispan Location of a
45' Sweptback Wing-Fuselage-Pylon ComSination. NACA RM L56J@, 1957.
Alford, W i l l i a m J., Jr., and King, Thomas J., Jr.: Experimental S t a t i c
Aerodynamic Forces and Moments a t High Subsonic Speeds on a Canard
Missile During Simulated hunching From t h e Midsemispan Location of a
4 5 O Sweptback Wing-Fuselage-Pylon Combination a t Zero Sideslip. NACA
RM L56Jl5a, 1957.

Alford, W i l l i a m J., Jr., and King, Thomas J., Jr.: Experimental Investi-
gation of Flow Fields a t Zero Sideslip Near Swept- and Unswept-Wing-
Fuselage Combinations a t Low Speed. NACA RM L56J19, 1957.

Alford, W i l l i a m J., Jr.: Theoretical and Experimental Investigation of

t h e Subsonic-Flow Fields' Beneath Swept and Unswept Wings With Tables
of Vortex-Induced Velocities. NACA TN 3738, 1956.
 .e
I
r .-

VARIATION O F S T O R E A N D P Y L O N FORCES
WITH RATIO OF STORE D I A M E T E R TO P Y L O N S P A N

AP
AND

Figure 1

SCOPE OF TEST CONFIGURATIONS

M = 1.6 M.0.8 TO 1.43

0.25 b/2
c * 1 0

Figure 2
 CONTRIBUTION OF STORE AND PYLON TO COMBINED LOAD
M =1.6

eT EXP THEORY
0 ---- STORE
--- STORE IN
PRESENCE
SIDE -FORCE OF PYLON
STORE -PYLON
0 - COMB IN ED

-3-
-2t I
-4 0 4 8
\e
12
LOAD

a, DEG

Figure 3

STORE -PYLON SIDE FORCE AT T H R E E SPANWISE POSITIONS

M =1.6
0 EXPERIMENT
THEORY
--- SUBSONIC
- SUPERSONCL. L.E. E.

-3
0

cy,sp-- 2I p
0 4 8
Q DE0
1 2
0

0 4 8
Q DEG
1 2 0 4 8
4 DEG
1 2
0

.-
Figure 4
t' -
I. 96
 .-
#a

SCHLIEREN PHOTOGRAPH OF TEST CONFIGURATION

M=1.6 ;a=0"

Figure 5 L-57-179

EFFECT OF PYLON SWEEP AND STORE CHORDWISE LOCATION

M = 1.6

- EXP.
0
THEORY

-4
0
\
--- 4 8
a, DEG
1 2 4

Figure
aI DEG
8

6
0

12 4
a, DEG
8 12

97
7
0.
. .
.* *c .. .1 'I

. I . 0 . .
O
..,
D -I
-..
0..

18

STORE -PYLON LOAD AT COMBINED

ANGLE OF ATTACK AND SIDESLIP
M ~1.6;a=8O

Y O

e
-12 -8 -4 0 4 8
ANGLE OF SIDESLIP, B, DEG
\ 0

12
-
0 EXP.
THEORY

Figure 7

EFFECT OF FINS ON STORE SIDE FORCE

M = 1.60 M = 1.43
<

EXP. THEORY
0 ---- FINS ON
0 - FINS OFF

CY,S

-1.2

- CY,
4
DEG
8 12
-*81 1
-l.24
0 4
a. DEO
,
8
,
1
,,

Figure 8

- --
e--. ___.^
 19

INTERFERENCE EFFECTS BETWEEN INBOARD

c.i
ANDOUTBOARDSTORES
M = 1.43
.2 EX!? THEORY
0 ----- STORE ALONE
0 - IN COMBINATION

-.I

-. 32-2
-*

2 4 6 8
U,DEG
L
-2
- 2 4 6 8
U ,DEG

EFFECT OF MACH NUMBER

ON STORE SIDE FORCE AND YAWING MOMENT
EX P ER IM ENTAL DATA

M = 1.43 M = 1.43

-.I
-. 2 -.2
M = 1.2 M = 1.2

-.2

M -0.8 M = 0.8

CY,S Cn,s
-.I
-.2 -. 2
- 2 0 2 4 6 8 - 2 0 2 4 6 8
a ,DEG a ,DEG

Figure loF, -.

33
0.
0
0.
.
. 0. .. .
I - '
0 . .
n
0 . 1
.( i
1 .
r - ...
20

EFFECT O F FUSELAGE INDENTATION ON STORE

SIDE FORCE AND YAWING MOMENT
M = 1.2
.I

-&INDENTE:" ' -.I , S ~ ~ I

-.Io~
CY ,s -.2 ---&-BASIC

INBOARD STORE -.2

- 4 0 4 8 - 4 0 4 8

.I

0
CY ,s ---
-.I -.I
OUTBOARD STORE
-.2
- 4 0 4 8
-. 2
- 4 0 4 8
a,DEG a ,DEG

Figure 11
 STEADY LOADS DUE TO JET INTERFERENCE ON WINGS, TAILS,
AND FUSELAGES AT TRANSONIC SPEEDS

By John M. Swihart and Norman L. Crabill

Langley Aeronautical Laboratory

SUMMARY
,'...

This.paper gives the results of some recent investigations of jet-

interference effects on actual airplane configurations at transonic speeds.
Data presented herein were obtained with hot jets on both wind-tunnel and
flight models. Results indicate that jet-induced effects are small at
subsonic speeds; however, at l o w supersonic Mach numbers, these effects
are Comparable to those obtained at substantially higher Mach numbers.

INTRODUCTION

In recent years there have been several investigations of turbojet-

exhaust interference effects at supersonic speeds (refs. 1 and 2); however,
only a limited amount of data has been obtained at transonic speeds. It
is the purpose of this paper to show the results of some recent investi-
gations of the jet interference on actual airplane configurations at
transonic speeds. This discuksion will be limited to steady loads inauced
by simulated jets on nearby wings, tails, and overhanging fuselages. Data
presented in this paper cover the Mach number range of 0.85 to 1.20 and
were obtained with hot jets at total pressure ratios corresponding to
current turbojet-engine pressure ratios.

SYMBOLS

b span
C local chord
-C m e a n aerodynamic chord

cn section normal-force coefficient

normal-force coefficient
ACN incremental normal-force coefficient, CNJjet on - CN,jet off
pressure coefficient, Plocal - pm
cP 9
incremental pressure coefficient, Cp,jet on - Cp,jet off

primary jet diameter

dj
P pressure

9 dynamic pressure

Y spanwise distance

2 distance below wing chord plane

Subscripts:

AV average

J Jet
R resultant, Lower surface - Upper surface

t total

T tail

m free stream

APPARATUS

me models used in these investigations are shown in figure 1. The

sting-mounted 60° delta-wing model was tested in the Langley 16-foot
transonic tunnel with one jet nacelle located at 41 and 69 percent of the
semispan on one wing panel only. At the inboard station, the jet exit
was located at 38 and 69 percent of the local chord and at 1, 2, and
4 jet diameters below the wing. At the outboard location, the nacelle
was tested only at 4 jet diameters below the wing, and the exit was at
63 percent of the local chord. Simulation of the exhaust of a non-
afterburning turbojet engine was achieved through the use of a hydrogen
peroxide gas generator exhausting through a sonic exit. !&e je€ total
pressure ratio was varied from 1 (power off) to about 7 with a jet
stagnation temperature of about 1,400° F at each test Mach number. At
six spanwise stations, pressures were measured at static orifices located
at about every 5 percent of the chord on the upper and lower surfaces.
 - -
a 0

The twin-engine model with p a r t of the fuselage overhanging t h e j e t

exhaust w a s a l s o a 16-foot-transonic-tunnel model and was supported a t
the wing t i p s by a bifurcate sting-support system. Two hydrogen peroxide
t u r b o j e t simulators were mounted i n a r m - p i t nacelles; t h e i r exhausts were
p a r t i a l l y separated by a short keel. S t a t i c pressures were measured along
t h i s keel and the shoulder of the overhanging fuselage. The horizontal
t a i l w a s mounted i n both a high position (on the v e r t i c a l t a i l ) and a
r e l a t i v e l y low position (on the boom). Pressure d i s t r i b u t i o n s were
obtained on both t a i l s d i r e c t l y above the j e t center l i n e extended.

me single-engine model with the fuselage overhang w a s f l i g h t t e s t e d

w i t h the horizontal t a i l located ahead of the j e t e x i t and a t two posi-
t i o n s downstream of the j e t e x i t . An afterburning turbojet engine was
simulated by using a solid-propellant rocket motor exhausting through a
sonic e x i t a t a t o t a l pressure r a t i o of 6.0 and a stagnation temperature
of 3,200° F. Pressures were measured on the t a p and bottom of the fuse-
lage overhang and a t two spanwise s t a t i o n s on t h e t a i l s a t a Mach number
of 1.2.

A n investigation w a s conducted i n the Langley 16-foot transonic

tunnel of a target-type thrust reverser mounted ofi a single-engine f i g h t e r
model t o evaluate the t h r u s t reverser as a speed brake. The design of
t h e reverser w a s taken from reference 3, and the j e t exhaust was simulated
by a hydrogen peroxide simulator operating a t a j e t t o t a l pressure r a t i o
of 5. During the course of t h i s investigation, s t a t i c pressures w e r e
measured around half of the fuselage a t 6 meridians and f o r about 3 j e t
diameters ahead of the base. Flaw visualization was achieved by t u f t s
on the other half of the fuselage.

Data are presented f o r the wind-tunnel models from M = 0.85 t o 1.05;

data are presented f o r the f l i g h t model a t M = 1.20.

RESULTS AND DISCUSSION

Effects of Isolated J e t s

Before discussing t h e results of these tests, a review will be given

of the phenomena pertinent t o the exhaust of an i s o l a t e d overpressure
sonic j e t i n t o subsonic and supersonic e x t e r n a l streams. A t current j e t
t o t a l pressure r a t i o s , the j e t bulges outward immediately downstream of
the j e t e x i t a t both Mach numbers ( f i g . 2). A t subsonic speeds, since
t h e i n t e r n a l s t r u c t u r e of shocks and expansions cannot penetrate the
subsonic mixing boundary (see ref. 4), the only s i g n i f i c a n t e f f e c t s i n
the e x t e r n a l flow me some compression due t o the bulge near the e x i t
and a subsequent expansion dawnstream due t o entrainment of the external
stream by the j e t ( f i g . 2). Contrariwise, f o r supersonic speeds, Leiss
 c
. e . . ...
& .t I a
.--
e*.

and Bressette ( r e f . l), Love and Grigsby ( r e f . 4), and others have shown
that the external flow i s marked by a shock a t t h e i n i t i a l bulge ( e x i t
shock), an expansion over t h e curved j e t boundary, and a shock ( j e t shock)
which occurs where the shock i n t e r n a l t o the j e t penetrates the supersonic
mixing boundary ( f i g . 2). EVen though the i n i t i a l deflection angle of
the bulge i s only about h a l f t h a t f o r subsonic speeds ( r e f . 4), the
presence of shocks i n the external flow indicates that the e f f e c t s on
the external stream may be considerably l a r g e r than i n subsonic flow and
probably extend t o greater distances from the j e t boundary.

J e t Effects on Wings

Figure 3 shows the jet-induced r e s u l t a n t pressures on the 60° d e l t a -

wing m o d e l a t M = 0.90 and 1-05 f o r a = 3° f o r one nacelle located
1 j e t diameter below the right wing and with the j e t e x i t a t 69 percent
of the l o c a l chord. Qualitatively, the r e s u l t s are about what might be
expected from the discussion given f o r i s o l a t e d j e t s ( f i g . 2 ) . A t
M = 0.90, the e f f e c t s of flow i n c l i n a t i o n near the j e t e x i t and the sub-
sequent expansion due t o entrainment produce A C p , ~ peaks of no more
than 0.09 and -0.05, respectively. A t M = 1.03, t h e e f f e c t of the e x i t
shock i s t o give a positive peak of A C p , ~ = 0.27; the negative peak i s
s l i g h t l y less than 0.10. 'Ilhe e f f e c t s diminish rapidly with increasing
spanwise distance from the j e t a t M = 0.90 and only moderately
a t M = 1.05.

Figure 4 presents jet-induced r e s u l t a n t pressures f o r the same t e s t

conditions as those f o r f i g u r e 3 except the whole nacelle has been moved
forward s o t h a t the e x i t i s now a t 38 percent of the l o c a l chord. I n
general, the j e t e f f e c t s are s h i f t e d correspondingly forward and, a t
M = 0.90, the e f f e c t s are somewhat diminished. A t M = 1.05, however,
t h e maximums and minimums a r e comparable w i t h those obtained when t h e
nacelle was located only 1 j e t diameter below the wing ( f i g . 3) as f a r
as the e x i t shock i n t e r s e c t s the wing. The d a t a indicate t h a t the e x i t
shock has passed off the leading edge of the wing s l i g h t l y outboard
= 65 percent and that the j e t e f f e c t shown f o r Y = 74 percent
Of q5 b 7
comes e n t i r e l y from the influence of t h e j e t on the wing upper-surface
pressures. Evidently, t h i s e f f e c t r e s u l t s from some i n t e r a c t i o n of the
j e t - e x i t shock and the wing leading edge.

The chordwise pressure d i s t r i b u t i o n s obtained a t a = 'J0 and

M = 0.90 and 1.05 f o r the two nacelle positions sham i n figures 3 and 4
have been integrated t o obtain the spanwise loading curves shown i n f i g -
'me 5. Additional data obtained f o r the nacelle located a t 69 percent
of t h e wing semispan w i t h t h e j e t e x i t a t 63 percent of the l o c a l chord
and 4 j e t diameters below the wing are given i n f i g u r e 6. Spanwise
loadings f o r t h e basic wing (no nacelle or pylon) are a l s o given i n
 .. .
. * . . . 0.

figures 5 a d 6 f o r comparison. It i s seen that t h e l a r g e s t e f f e c t s

occur a t M = 1.05 i n a l l three cases, and that, a t t h i s Mach number,
the e f f e c t i s smallest f o r the case where the exit-shock i n t e r c e p t passes
off the wing before reaching the t i p . Although t h e ACN = 0.010
measured f o r the nacelle positioned a t the inboard location and 1 j e t
diameter below the wing i s only about 5 percent of the wing normal force,
it i s a c t u a l l y about 1.5 times t h e t h r u s t of the j e t causing t h i s e f f e c t .
This i s comparable i n magnitude t o the e f f e c t s shown i n reference 1 a t
a Mach number of 1.8.

The wing chordwise and spanwise center-of-pressure locations f o r

t h e nacelle located a t the inboard position and f o r z/dj = 1 and 4
a r e shown i n figure 7. Jet operation has a very small e f f e c t on the
chordwise center of pressure and p r a c t i c a l l y no e f f e c t on the spanwise
center of pressure a t subsonic speeds. I n general, the data indicate
t h a t the slight forward and inboard movement of the center of pressure
i s more pronounced f o r z/dj = 1.

J e t Effects on Fuselage Overhangs and

Horizontal-!Pail Surfaces

The data shown i n f i g u r e 8 give the e f f e c t of Mach number on j e t -

induced pressures measured along the overhanging portions of two d i f f e r e n t
e x i t configurations. Although these d a t a were obtained on two d i f f e r e n t
configurations a t d i f f e r e n t j e t t o t a l pressure r a t i o s , these differences
are unimportant i n showing the e f f e c t s t h a t are discussed here. For the
data a t M = 0.85 (the twin-engine f i g h t e r ) , the increment of ACp = 0.25
a t the center l i n e indicates that the j e t i s probably i n contact w i t h
the center-line row of o r i f i c e s over the e n t i r e length of the short keel;
however, very l i t t l e j e t e f f e c t s were measured a t the o r i f i c c row around
the shoulder. Extensive pressure surveys of t h i s kind have been shown
i n reference 5.

A t M = 1.20, the data (the single-engine type) indicate t h a t the

o r i f i c e s on the bottom of t h e boom w e r e probably i n contact w i t h the j e t
a t l e a s t near the e x i t . %e pressure measurements made along the t o p of
the boom when the horizontal t a i l w a s located ahead of the e x i t indicate
t h a t moderate positive pressure increments are experienced even though
t h i s position i s "blanketed" by the boom. Additional data, not shown
here, indicate t h a t the leading edge of t h i s upper-surface j e t - e f f e c t
zone moves a f t w i t h increasing Mach number.

Thus, a t subsonic speeds, the region of fuselage overhangs subject

t o s i g n i f i c a n t j e t e f f e c t s i s newly confined t o that i n contact w i t h the
j e t . A t supersonic speeds, the region of s i g n i f i c a n t j e t e f f e c t s extends
much further away from t h e j e t boundary and can even a f f e c t blanketed
areas.
1:s
6
A.

The effect of tail position on the jet-induced tail chordwise pres-

sure distributions measured at M = 1.20 on the single-engine model is
shown in figure 9. The data indicate that the loading over the forward
third of the tail is greatest for the forward tail position and that the
loading is greater outboard. The lower loading at the inboard location
may be due to the effects of a sdsonic jet mixing boundary in the region
immediately under the boom and this subsonic mixing boundary may give way
to a supersonic mixing boundary outboard, with consequent increase in jet
effects.

Figure 10 shows the effect of tail position and free-stream Mach

number on jet-induced tail and tail-plus-afterbody normal-force coeffi-
cients. Pressure distributions obtained for the high and low horizontal
tails at M = 0.85 on the twin-engine model were integrated to obtain
the jet-induced effect on section normal-force coefficient as a function
of angle of attack. The data indicate that for the high tail there is no
jet effect on Cn at any angle of attack, whereas, in the l o w tail posi-
tion, a constant reduction of 0.07 in Cn was obtained at a l l angles of
attack. This result indicates that the flow entrainment effect of a jet
in subsonic flow diminishes rapidly with distance from the jet boundary.

Accelerometer data taken at M = 1.20 for the three longitudinal

positions of the horizontal tail on the single-engine models were reduced
to show ACN induced by the jet on the afterbody and afterbody plus tail.
The data, based on the plan-form area downstream of the jet, indicate
that ACN probably goes through a maximum as the position of the exposed
tail centroid is varied between 0 and 2.3 jet diameters downstream of the
jet exit. If the reduction in tail-section normal-force coefficient
obtained at M = 0.85 be taken as representative of the effect over the
entire tail, then the total jet-induced tail loads are seen to change
sign as the free-stream Mach number is increased from M = 0.85 to 1.20.

Jet Bfects Due to Thrust Reverser

The effect of thrust-reverser operation on afterbody pressures on a

single-engine blunt-based fighter-type configuration flying at M = 1.05
is given in figure U. The curves shown represent averages of the six
pressures measured around the periphery of the left side of the afterbody
at each fuselage station. When no reverser is present, the pressures
ahead of the base are negative and decrease to -0.27 at the base. When
the target-type reverser was extended, the pressures were increased to
positive values for about 3 jet diameters ahead of the base. The tuft
studies indicated that separation occurred on the fuselage forward of
3 jet diameters and this separation was very unstable. The resulting
large lateral oscillations of the model on the relatively rigid support
system used in the wind-tunnel test indicate that the operation of this
 * *
.- * *
.* *
, * * . e m , *a I

device on an airplane in free flight might render the airplane unflyable.

Although it is now known that this reverser is not a good design, it is
believed that the magnitude of these local pressure changes is typical
of what should be expected with most thrust reversers.

CONCWSIONS

Examination of the data from recent investigations of jet-interference

effects on actual airplane configurations at transonic speeds has led to
the following tentative conclusions:

1. At subsonic speeds, jet-induced effects on wings, tails, and

fuselages are small and decrease rapidly with distance from the jet
boundary.

2. At law supersonic Mach numbers, jet-induced effects comparable

to those obtained previously at substantially higher Mach numbers can be
realized. Generally, these effects do not diminish as rapidly with
distance from the jet boundary as those induced in subsonic flow.

1. Leiss, Abraham, and Ekesette, Walter E.: Pressure Distribution Induced

on a Flat Plate by a Supersonic and Sonic Jet Exhaust at a Free-
Stream Mach Number of 1.80. NACA RM ~56106,1957.
2. Rainey, Robert W.: The Effects Upon Body Drag of Jets Exhausting
From Wing-Mounted Nacelles. NACA RM LsA09, 19%.

3. Povolny, John H., Steffen, Fred W., and McArdle, Jack G.: Swrlmary of
Scale-Model !&rust-Reverser Investigation. NACA TN 3664, 1956.
4. Love, Eugene S., and Grigsby, Carl E.: Some Studies of Axisymmetric
Free Jets Exhausting F r o m Sonic and Supersonic Nozzles Into Still
Air and Into Supersonic Streams. NACA RM L$L3l, 1955.

5. Cornette, Elden S., and Ward, Donald H.: Transonic Wind-Tunnel

Investigation of the Effects of a Heated Propulsiye Jet on the
Pressure Distribution Along a Fuselage Overhang. NACA RM ~ 5 6 ~ 2 7 ,
1956
 CONFIGURATIONS INVESTIGATED
60" DELTA WING OVERHANGING FUSELAGE
SINGLE
NACELLES TWIN JET

SINGLE JET
THRUST REVERSER
=

Figure 1

FLOW PHENOMENA ASSOCIATED WITH JET EXHAUST

P+,j/PW=2TO4

M, = 0.6 TO 0.9 F

Figure 2
 9

JET-INDUCED RESULTANT PRESSURES

Cr.5"; z/dj=I
.2
.I
ACP,R 0 SPANWISE LOCATION OF NACELLE
-. I

M=1.05
3
.2
PANWISE LOCATION OF
Acp, R ' I NACEUE
0
-. I

JET-INDUCED RESULTANT PRESSURES

W 5 O ; Z/dj=4
M =0.90
I h
Acp,R 0 LOCATION OF NACELLE
-.I

90

3 M.1.05
.2
AcP, R ' I
0
-. I
-.2
10

JET- INTERFERENCE EFFECT ON SPAN LOADING

a.50
M.0.90 M.1.05
.4r r

JET-INTERFERENCE EFFECTS ON SPAN LOADING

a = 5"; -I- = 4; NACELLE AT
dl
-&= 0.69
.4 - -

.3- -

.2- -

.I- --- JET ON -

0 .2 .4 .6 .8 1.0 0 .2 .4 .6 .8 1.0
-Y
b/2

Figure 6
..
9
e.-A
'0
 -
JET-INTERFERENCE EFFECT ON CENTER OF PRESSURE
a=50

-------
.50 I ----

.40
.- I LI
0 .80
1dj-
-1
I
.90
-------
--- JET
M
I
1.00
BASIC WING
JET OFF
ON
I
1.10

Figure 7

MACH NUMBER EFFECT ON AFTERBODY PRESSURES

UPPER SURFACE

CENTER LINE, LOWER SURFACE

I I I I I J

Ma=0.85; pt,j/pa= 5
TWIN ENGINE

CENTER LINE

O-0 I 2 3 4 5 6
AFTERBODY LENGTH, X/dj

Figure 8
JET EFFECTS ON TAIL CHORDWISE PRESSURE DISTRIBUTIONS
TWO SPANWISE STATIONS; TWO TAIL POSITIONS
M-=1.20; Pt, j/P,=6.0

x/d j = 2.3
0

.I
AcP,R 0 54
.2

- ' t @ L H, -
''0 IO 20 30 40
-
I I I I

% CHORD

Figure 9

J E T EFFECT ON AFTERBODY AND TAIL

NORMAL-FORCE COEFFICIENTS

M, = 0.85 M,=1.20
TWIN ENGINE SINGLE ENGINE
.24 r r
6
l
.FJET ON ANDOFF,,-,!I,.,

.O 8
TAIL C'n
0 ACN -08
JET OFF
-.08 LOW TAIL .04
6
-.160 l4lLU-l
4 8 12 0
u 1 2 3 4
a,DEG TAIL POSITION, x/dj

Figure 10
 -.3- Mm=1.05i Pt,j/p,=5

-.2 - NO REVERSER

-.I -\ ,
\
AVERAGE Cp 0 \.
\
\
.I - \ REVERSER EXTENDED

.2 -
I I
'\

I
'.---/-+.
AFTERBODY LENGTH AHEAD OF JET EXIT, x/dj

Figure 11
 e . . D
.. e
e 0
.
e
, ,e em

EFEECTS ON ADJACENT SURFACES FROM THE FIRING OF ROCKET JETS

By Walter E. Bressette and Abraham hiss

Langley Aeronautical Laboratory

SUMMARY

This paper is a prelbdnary and brief account of some research

currently being conducted to determine the jet effects on adjacent sur-
faces fromthe firing of rocket jets. Measurements of jet-effect pres-
sures on a flat plate as well as shadowgraphs are presented that were
obtained when a rocket jet at a Mach number of 3 was exhausted downstream
and upstream into free-stream flow at a Mach number of 2 located from 2
to 4.7 rocket-jet-exit diameters fromthe plate. The jet effects on the
flat plate with the rocket jet exhausting downstream are of the same order
of magnitude as those previously obtained from sonic exits with a total
pressure 10 times lower. A maximum pressure coefficient on the plate
of 1.35 was obtained when the rocket jet was exhausted upstream at 2
rocket-jet-exit diameters below the plate, and an integration of the
measured jet-effect pressures at this position resulted in a normal force
on the plate equal to 2 . 3 times the thrust output of the rocket jet.

INTRODUCTION

During the past few years the use of air-to-air rocket-propelled

missiles for supersonic aircraft has become prominent. Also, it is
reasonable to expect that, in the future, large-sized aircraft will be
used as platforms for the launching of large-sized guided rocket-propelled
missiles. These missiles might be fired not only forward but also rear-
ward and sideways. It is important, then, to know the short-duration
loads on adjacent wings and body surfaces of the aircraft from the rocket
jet of the missile. Therefore, in the Langley preflight jet of the Pilot-
less Aircraft Research Station at Wallops Island, Va., a testing program
has been initiated to determine the jet effects on adjacent surfaces from
high-pressure short-duration rocket jets exhausting downstream, upstream,
and normal to the free-stream flow. Up to the present time o n l y prelim-
inary tests from rocket jets firing downstream and upstream at a, Mach
number of 2 have been completed. The results from some of these tests
are discussed in this paper.
 SYMBOLS

pressure c o e f f i c i e n t , PW - Pm
s,
diameter of rocket-jet e x i t

normal force

Mach number a t exit of rocket j e t

free-stream Mach nuIriber

s t a t i c pressure a t exit of rocket j e t

t o t a l pressure a t exit of rocket j e t

measured s t a t i c pressure on p l a t e

Pm free-stream s t a t i c pressure

free-stream dynamic pressure, 7 pmMm

2
m
'

X chordwise distance on p l a t e

Y spanwise distance on p l a t e

7 r a t i o of s p e c i f i c heats (1.4 f o r a i r )

Presented i n f i g u r e 1 i s a sketch of the t e s t setup showing the

l o c a t i o n of a rocket-jet e x i t having a Mach number 3 nozzle below a f l a t
p l a t e a t t h e exit of t h e blowdown tunnel having a Mach number 2 nozzle.
A l s o shown i n t h e f i g u r e i s t h e r e l a t i o n of t h e rocket-jet e x i t r e l a t i v e
t o t h e f l a t p l a t e f o r t h e four t e s t p o s i t i o n s t h a t are discussed i n t h i s
paper. These p o s i t i o n s a r e f o r one rocket j e t exhausting downstream that
i s located a t 3.4dj below t h e p l a t e and 6.5dj from t h e end of t h e p l a t e
and f o r three rocket j e t s exhausting upstream a t 2, 3.4, and 4.7dj below
t h e p l a t e , a l l of which were located a t 0.74dj upstream from t h e end of
t h e p l a t e . These upstream f i r i n g positions a r e henceforth r e f e r r e d t o
..
 3

as positions A, B, and C, respectively. The rocket-jet v a r i a t i o n of t o t a l

pressure with time w a s similar during each of t h e t e s t s . The a c t u a l t o t a l
f i r i n g time of t h e rocket j e t was approximately 0.6 second with the rocket
j e t s exhausting a t a t o t a l pressure of approximately 1,000 pounds per
square inch during the f i r s t 0.3 second and then reducing t o zero pres-
sure during t h e remaining 0.3 second. The data as presented i n t h i s paper
were taken with t h e rocket j e t exhausting a t a t o t a l pressure of approxi-
mately 1,000 pounds per square inch.

The l o c a t i o n on t h e f l a t p l a t e of 47 static-pressure o r i f i c e s r e l a -
t i v e t o t h e rocket-jet e x i t s i s shown i n figure 2 as well as t h e area on
t h e p l a t e i n t e e r a t e d t o obtain t h e j e t - e f f e c t normal load. A s can be
seen i n t h i s figure, t h e p l a t e was pressure surveyed on only one s i d e of
t h e rocket- j e t center l i n e .

RESULTS AND DISCUSSION

Presented i n f i g u r e 3 i s t h e rocket-jet-off and rocket-jet-on center-

l i n e pressure d i s t r i b u t i o n on t h e f l a t p l a t e p l o t t e d over t h e j e t - e f f e c t
shock p a t t e r n that was obtained from shadowgraphs w i t h t h e rocket j e t
exhausting downstream a t pt ./pm= 68 and p j p m = 2.1. The shadow-
,J
graphs revealed t h e now-familiar two-shock j e t - e f f e c t p a t t e r n that i s
obtained w i t h a sonic t u r b o j e t e x i t . These two shocks are commonly
c a l l e d t h e exit shock and t h e j e t shock. "he maximum pressure coef-
f i c i e n t of approximately 0.1 that w a s obtained d i r e c t l y behind t h e i n t e r -
section on t h e p l a t e of t h e e x i t shock i s about t h e same that would be
obtained from a sonic t u r b o j e t e x i t at p = 3 and pt,jpm = 6.
j P m

Figure 4, which presents a t y p i c a l shadowgraph with t h e rocket j e t

exhausting upstream a t p o s i t i o n A, shows a bow shock, upstream from t h e
rocket-jet exit, standing i n f r o n t of t h e maximum penetration of the
rocket-jet exhaust which i s burning. This bow shock i s strong enough
a t t h e i n t e r s e c t i o n with t h e p l a t e t o cause separation of t h e p l a t e bound-
a r y l a y e r . Figure >, which presents a t y p i c a l shadowgraph a t p o s i t i o n C,
shows a bow shock, upstream from t h e rocket-jet exit, standing i n f r o n t
of t h e maximum penetration of t h e rocket-jet exhaust. This bow shock a t
p o s i t i o n C i s f a r t h e r upstream from t h e rocket-jet exit than the one
obtained a t p o s i t i o n A; also, it has a g r e a t e r curvature from i t s normal
p o s i t i o n a t t h e center, and a t t h e i n t e r s e c t i o n point on the p l a t e it i s
reflected.
4

Presented in figure 6 is the rocket-jet-off center-line pressure

distribution as well as the rocket-jet-on center-line pressure distribu-
tion on the plate, and presented in figure 7 is the rocket-jet-on spanwise
pressure distribution obtained at position A. When the rocket jet was
fired at this position, figure 6 shows that the bow shock in front of
the rocket-jet exit with the rocket.jetoff moved from a position approxi-
mately O.7dj to a position 2dj to 3dj upstream from the exit. This bow
shock stands in front of the most forward pentration of the rocket-jet
fuel-rich exhaust which is mixing with the incoming free-stream air and
is burning in the zone designated as the mixing zone. At the same time,
this bow shock induced separation of the plate boundary layer upstream
of its intersection on the plate, The resulting initial rise in the
rocket-jet-on center-line pressure distribution on the plate is similar
to that obtained in references 1 and 2 from the separation of a turbulent
boundary layer by the step technique; the first peak value of 0.35 as
obtained in this test agrees very favorably with the step-technique value
obtained in reference 1 on a flat plate at M, = 2. Also, as illustrated
in reference 2, the overall pressure rise for incipient separation can
be considerably greater than the first peak pressure rise. In this test
the overall pressure rise resulted in a maximum pressure coefficient
of l.,35, which is 83 percent of the pressure coefficient that might be
obtained directly behind a normal shock. This maximum pressure coeffi-
cient is reduced rapidly as the spanwise distance is increased in fig-
ure 7 as would be expected from the intersection of a bow shock and a
flat plate, whereas the first peak value is essentially constant with
the same increase in spanwise distance. A normal force on the plate
of 2 . 3 times the rocket-jet thrust output was obtained at this position
of 2d. below the plate by integrating the pressure field that was measured
J
on one side of the rocket-jet center line and multiplying the result
by 2 on the assumption that the pressure distribution on the opposite
side of the rocket-jet center line was the same.

In figures 8 and 9 is shown the center-line and spanwise pressure

distribution on the plate with the rocket jet exhausting upstream at
position C. When the rocket jet was fired at this position, the bow
shock in front of the rocket-jet exit with the rocket jet off did not
intersect the plate at this position and moved from approximately O.7d
j
to approximately 9dj upstream from the exit. This large upstream movement
of the bow shock resulted in the bow shock intersecting the plate upstream
of the pressure-orifice field. A pressure coefficient of approxhmtely 0.2
was obtained at the most forward orifice location, and this value grad-
ually reduced on the plate in the downstream direction as well as in the
spanwise direction.

.”
 .
0 0
0 .
.
0
.
I

An investigation is being made to determine the effects on adjacent

surfaces from the firing of rocket jets. The following conclusions were
obtained from observations of jet-effect pressures measured on a flat
plate in conjunction with shadowgraphs when a rocket jet at a Mach number
of 3 was exhausted downstream and upstream into free-stream flow at a
Mach number of 2:

1. The jet effects on the flat plate with the rocket jet exhausting
downstream are of the same order of magnitude as those previously obtained
from sonic exits with a total pressure 10 times lower.

2. A maximum pressure coefficient on the plate of 1.35 was obtained

when the rocket jet was exhausted upstream at two rocket-jet-exit diame-
ters below the plate.

3. An integration of the measured jet-effect pressures on the plate

with the rocket jet located at 2 rocket-jet-exit diameters below the plate
resulted in a normal force on the plate equal to 2.3 times the thrust
output of the rocket jet.

1. Lange, Roy H.: Present Status of Information Relative to the Predic-

tion of Shock-Induced Boundary-Layer Separation. NACA TN 3065, 1954.

2. Chapman, Dean R., Kuehn, Donald M., and Larson, Howard K.: Preliminary
Report on a Study of Separated Flows in Supersonic and Subsonic
Streams. NACA RM A55Ll4, 1956.
'

118
I
3 .
8 .*
.
I
e . . , . -
. I . 0 . . * .--
e I .*e

TEST SETUP

FLAT PLATE
ROCKET BLOWDOWN TUNNEL
M - =3
1 Ma= 2

TREAM FLOW

Figure 1

LOCATION OF PLATE PRESSURE ORIFICES

AREA INTEGRATED

Figure 2
 .. ea re
. e * * e
e ( * -
-
1 * **
a *0 ) .
1

CENTER-LINE PRESSURE DISTRIBUTION

ROCKET EXHAUSTING DOWNSTREAM; Mj=3; pt,j/pm=68; pj/pm=2.1
.2

,;
ROCKET ON
ROCKET OFF
p i pa
=-.IF-".
CP
0

-.I I I

OCK

0 2 4 6 8 IO
X/dj

Figure 3

TYPICAL SHADOWGRAPH FOR POSITION A

ROCKET EXHAUSTING UPSTREAM

Figure 4 -. L-57-161
.A f - 3
8

TYPICAL SHADOWGRAPH FOR POSITION C

ROCKET EXHAUSTING UPSTREAM

L-57-160
Figure 5

CENTER-LINE PRESSURE DISTRIBUTION

~SUR
~E
,,- ; ; ~ ~ I-
ROCKET EXHAUSTING UPSTREAM: POSITION A
1 - 5 P[RES

ROCKET ON
I .o
ROCKET OFF
cP STEP-TECHNIQUE
VALUE (LANGE)

6 5 4 3 2 I 0-1
X/d j

Figure 6
 SPANWISE PRESSURE DISTRIBUTION
ROCKET EXHAUSTING UPSTREAM; POSITION A

Figure 7

CENTER-LINE PRESSURE DETRIBUTION

ROCKET EXHAUSTING UPSTREAM; POSITION C

1.0

cp .5
-
--- ROCKET ON
ROCKET OFF

9 8 7 6 5 4 3 2 1 0 - 1
X/dj

222
Figure 8
 SPANWISE PRESSURE DISTRIBUTION
ROCKET EXHAUSTING UPSTREAM ; POSITION C

1.0
C
p.5

Figure 9
7E

INFLUENCE OF AUTOMATIC CONTROL OF ROLL COUPLING

AND PITCH-UP ON TAIL LOADS

By Ralph W. Stone , Jr .
Langley Aeronautical Laboratory

SUMMARY

An a n a l y t i c a l study has been made of t h e e f f e c t s of automatic aug-

mentation o r controlling systems on the t a i l loads experienced i n r o l l i n g
maneuvers and i n pitch-up. The r e s u l t s were calculated on an analog com-
puter and t h e equations of f i v e degrees of freedom were used f o r t h e
r o l l i n g maneuvers and t h r e e degrees of freedom f o r pitch-up. The r e s u l t s
of t h i s report a r e not intended t o be of general application, but r a t h e r
t o point out some of t h e problems t h a t may be encountered on any s p e c i f i c
design and t o indicate some probable trends.

The r e s u l t s indicate t h a t , f o r t h e r o l l i n g cases calculated, most

automatic systems tend t o reduce not only t h e violence of t h e maneuver
but a l s o t h e t a i l load encountered. The results show, however, t h a t i f
automatic systems a r e t o be used, they must be considered i n the i n i t i a l
design t o obtain acceptable motions i n rolls and t o evaluate properly t h e
t a i l loads.

For t h e pitch-up problem, the m a x i m u m t a i l loads a r e predominantly

t h e r e s u l t of control deflection, and systems which e s s e n t i a l l y reduce
t h e input o r pull-up deflections w i l l generally improve t h e acceleration
overshoots and reduce t h e h o r i z o n t a l - t a i l loads. I n general, systems
used f o r pitch-up may be compatible w i t h problems of roll coupling. On t h e
other hand, of the systems studied f o r roll coupling, only t h e p i t c h damper
would be h e l p f u l for pitch-up.

INTRODUCTION

Trends i n performance and design of airplanes have brought about

some very serious s t a b i l i t y deficiencies i n recent years. I n problems
involving such deficiencies, aerodynamic changes, of course, should f i r s t
be considered. There i s a major trend, however, toward t h e use of auto-
matic augmentation systems or controllers i n meeting these deficiencies.
This paper i s concerned with t h e e f f e c t s of such systems on t h e aerodynamic-
loading conditions f o r two of t h e more c r i t i c a l deficiencies, these being
2

divergencies i n rolls and pitch-up. The fundamentals of these problems

have been discussed i n numerous publications (refs. 1 t o 10 and 11t o 21,
respectively). This report, therefore, i s confined primarily t o t h e
horizontal- and v e r t i c a l - t a i l loads encountered when automatic systems
are used. The r e s u l t s presented herein are indicative of only some of
t h e problems and trends t h a t may be expected and a r e not necessarily of
general application.

SYMBOLS

cIJ l i f t coefficient

CD drag coefficient

CY side-force coefficient

rolling-moment coefficient

Cm pitching-moment coefficient

"I
Cn yawing-moment coefficient

moments of i n e r t i a about t h e X, Y, and Z body axes, respec-

t i v e l y , slug-ft2
IZ

product of i n e r t i a (positive when p r i n c i p a l axis i s inclined

below X body a x i s ) , slug-ft2

Io, angular momentum of engine r o t a t i n g p a r t s , ft-lb-sec

Xe e

W weight, l b

m mass, w/g, slugs

acceleration of gravity, 32.2 f t / s e c 2
g
S wing area, sq f t

b wing span, f t
-.
 3

-
C mean aerodynamic chord, f t

x~~ longitudinal distance from center of gravity t o ;/4 of

horizontal t a i l , f t

longitudinal distance from center of gravity t o C/4 of

vertical tail, f t

ZVT v e r t i c a l distance from center of gravity t o ?/4 of v e r t i c a l

tail, f t

P a i r density, slugs/cu ft

v velocity, f t / s e c

M Mach number

pressure a l t i t u d e , f t
kp
6a a i l e r o n deflection, deg

iT s t a b i l i z e r deflection, positive when t r a i l i n g edge i s down,

deg
rudder deflection, positive when t r a i l i n g edge i s t o t h e l e f t ,
deg
a angle of attack, deg

P angle of s i d e s l i p , deg

E downwash angle, deg

P r o l l i n g angular velocity, radians/sec

9 pitching angular velocity, radians/sec

r yawing angular velocity, radians/sec

t time, sec

k T h o r i z o n t a l - t a i l load, lb

Lvr v e r t i c a l - t a i l load, l b

nZ normal acceleration, g u n i t s
4 /
-

nY l a t e r a l acceleration, g units

f damping r a t i o

-
a
da
rate of change of damping r a t i o with angle of a t t a c k

A incr ernen t

acm cn = -
dcrl
cma = aa p

c, =
acm
- c - acY
“&g YP - ap

Subscripts :

0 i n i t i a l value

HT horizontal t a i l

VT vertical t a i l

Dot over a symbol indicates derivative.

-.
 5

METHODS

The results discussed herein are based primarily on calculations

f o r hypothetical airplanes t y p i c a l of contemporary f i g h t e r s . Table I
l i s t s the c h a r a c t e r i s t i c s of t h e airplane used f o r t h e roll cases.
Table 11 l i s t s those of the airplane used f o r t h e pitch-up cases. I n
the calculations, f i v e degrees of freedom were used f o r t h e roll cases
and three degrees of freedom f o r t h e pitch-up cases. The equations f o r
f i v e degrees and three degrees of freedom appear i n many references, f o r
example, references 11 and 22, respectively. I n t h e three-degree-of-
freedom calculations, however, CL, CD, and Cm were introduced as
functions of angle of a t t a c k and Mach number, these functions being non-
l i n e a r with angle of attack. For some r o l l i n g maneuvers, calculations
were made of t h e loads from motions obtained i n a c t u a l f l i g h t s . The
equations used f o r calculating t h e horizontal- and v e r t i c a l - t a i l loads
are :

and

DISCUSSION

R o l l Coupling

Divergences i n rolls are caused generally by r o l l i n g t o o r a p i d l y

f o r the d i r e c t i o n a l s t a b i l i t y t h a t exists (ref. 1). Thus, r o l l i n g veloc-
i t y and v e r t i c a l - t a i l s i z e are dominant f a c t o r s i n r o l l i n g divergences.
The e f f e c t s of these f a c t o r s a r e shown i n figure 1. Here, t h e maximum
maneuvering h o r i z o n t a l - t a i l loads and t h e maximum v e r t i c a l - t a i l loads
encountered i n rolls a t d i f f e r e n t r o l l i n g v e l o c i t i e s are p l o t t e d against
the average r o l l i n g v e l o c i t y of each maneuver.

Results a r e shown f o r an o r i g i n a l t a i l s i z e and f o r a t a i l s i z e

optimum for roll coupling. Also shown i s t h e c r i t i c a l r o l l i n g velocity,
which, as defined by P h i l l i p s i n reference 1, i s t h a t r o l l i n g frequency
which equals t h e lower of t h e pitching o r yawing ( i n t h i s case t h e yawing)
natural frequency. %..Z'8
6

The results show a very large increase i n t h e t a i l loads encountered

as t h e r o l l i n g velocity approaches c r i t i c a l and then a drop-off i n t h e
loads beyond t h i s value. Limiting t h e r o l l i n g v e l o c i t y t o some value
less than c r i t i c a l w o u l d c l e a r l y solve the problem. How fast an airplane
must r o l l , however, i s a controversial subject and i s not one f o r present
discussion. It i s s u f f i c i e n t t o say herein only that p i l o t s generally
i n s i s t and (because of t h i s insistence) the services require t h a t air-
planes r o l l a t v e l o c i t i e s l a r g e r than c r i t i c a l f o r many current and
planned configurations.

Changing the t a i l s i z e i s t h e second d i r e c t approach t o t h e problem.

There is, however, an optimum t a i l size; smaller or l a r g e r t a i l s lead t o
more v i o l e n t motions and l a r g e r loads. (See ref. 3 . ) For t h e case i n
t h i s paper the optimum t a i l i s somewhat l a r g e r than the o r i g i n a l t a i l .
The optimum t a i l s i z e i s t h a t f o r which t h e pitching and yawing n a t u r a l
frequencies are the same. For t h i s s i t u a t i o n no r o l l divergence, as such,
i s possible. A resonance condition exists, however, when the r o l l i n g f r e -
quency i s about equal t o the n a t u r a l frequencies i n p i t c h and yaw, and
the loads increase near t h i s r o l l i n g v e l o c i t y with the optimumtail; thus,
r e l a t i v e l y l a r g e loads a l s o exist and t h e motions s t i l l may be rather
violent .
Because of the wide v a r i e t y of f l i g h t conditions, speeds, and a l t i -
tudes now possible, a solution such as an optimm t a i l s i z e may not be
s u f f i c i e n t o r feasible f o r any s p e c i f i c design. (This i s t r u e f o r nor-
mal s t a b i l i t y as w e l l as f o r r o l l i n g divergencies.) Thus, the trend
towards the extensive use of automatic systems prevails. For the problem
of r o l l coupling, several systems are possible. This report treats
b r i e f l y f i v e types of systems which a r e shown i n table 111.

The first system i s a perfect controller which maintains zero side-

s l i p and zero changes i n angle of a t t a c k (ref. 10). This i s a rather
complex system requiring the sensing of attitude angles. The next system
i s a more p r a c t i c a l representation of t h i s system (ref. 10). The t h i r d
system i s c a l l e d a coupling-moment canceler (ref. 5 ) . This canceler i n
e f f e c t balances or cancels the i n e r t i a coupling p a r t s of t h e pitching
and yawing moments, which, as indicated i n reference 5 , are the primary
cause f o r r o l l i n g divergences. The last two systems are dampers, a p i t c h
damper and a yaw damper.

The choice of systems presented does not imply t h a t they a r e t h e most

promising controlling or augmentation systems but i s intended only t o show
t h e influence of some t y p i c a l systems on the loads encountered. For the
calculations shown, t h e autamatic systems are assumed t o have no l a g s and
a l l proper gains. For any s p e c i f i c design t h e influence of these f a c t o r s
must, of course, be obtained.
r-.
*
&I.=
fl.Q
4J
 Some t y p i c a l r e s u l t s of two of
t h e perfect c o n t r o l l e r ) are shown.
pling, the predominant influence of
(refs. 1, 3, 4, and 9, f o r example)
cated as a simple and d i r e c t way t o
these systems (the p i t c h damper and
I n many investigations of r o l l cou-
pitching v e l o c i t y has been evident
and t h e p i t c h damper has been i n d i -
influence t h e motions encountered.

Figure 2 shows the e f f e c t of a p i t c h damper on t h e t a i l loads. Here

a r e p l o t t e d t h e maximum maneuvering h o r i z o n t a l - t a i l loads and 'the maximum
v e r t i c a l - t a i l loads encountered as functions of average r o l l i n g v e l o c i t y
w i t h and without t h e p i t c h damper operating. The damping r a t i o of t h e
airplane with t h e damper operating was 0.5, and t h e maximum s t a b i l i z e r
deflection allowed f o r the damper (the control authority) was 1.8O.
These results w e r e obtained from flight t e s t s of a contemporary fighter
which except f o r a l a r g e r v e r t i c a l t a i l i s similar t o t h e hypothetical
airplane used i n the other r o l l i n g calculations. A s noted before, these
are not measured loads but they have been calculated from motions encoun-
t e r e d i n a c t u a l r o l l s and aerodynamic loading c o e f f i c i e n t s measured during
other flights.

A considerable improvement i n t h e loads encountered with t h e damper

operating i s shown. Rolling v e l o c i t i e s much i n excess of c r i t i c a l were
not obtained, however, and t h e e f f e c t s of l a r g e r r o l l r a t e s w i t h the dam-
per operating have not been established i n f l i g h t . Some calculations of
t h i s nature have been made, however, and a summary of such r e s u l t s i n
comparison with t h e results f o r other automatic systems a r e discussed i n
t h i s section.

The p i t c h damper i s more e f f e c t i v e i n reducing the v e r t i c a l - t a i l

loads than the h o r i z o n t a l - t a i l loads; t h i s indicates the dominant i n f l u -
ence of pitching velocity through i n e r t i a coupling. The r e s u l t s shown
here f o r the p i t c h damper are t y p i c a l of those obtained f o r most of the
other systems except, of course, the magnitude of reductions v a r i e s f o r
each system. These differences are discussed subsequently.

Some results f o r t h e most complex of the various systems i n t a b l e 111,

t h e perfect controller, are shown i n figure 3. For t h i s c o n t r o l l e r no
variations i n normal acceleration or l a t e r a l acceleration exist. P l o t t e d
are the maximum maneuvering h o r i z o n t a l - t a i l load and the v e r t i c a l - t a i l
load as functions of average r o l l i n g velocity. The controlled and uncon-
t r o l l e d cases a r e compared and the results are shown f o r two i n i t i a l nor-
mal accelerations, l g and 2g flights.

These r e s u l t s show that t h e c o n t r o l l e r not only eliminates t h e vio-

lence of the maneuver but reduces t h e t a i l loads encountered, except a t
t h e l a r g e s t r o l l i n g v e l o c i t i e s when t h e c o n t r o l l e r tends t o cause the
h o r i z o n t a l - t a i l loads t o be l a r g e r than are otherwise encountered. The
uncontrolled loads occur primarily from angle of a t t a c k and s i d e s l i p ,
whereas t h e controlled loads occur primarily from control deflections.
. .
*I . .
-.
a.
. .' -
0'

- .. ' . ( ..a
..a
0.
.-
?.
.

These control deflections (ref. 10) are a d i r e c t function of the r o l l i n g

velocity, i t s square, and i t s derivative; therefore, increasing r o l l i n g
v e l o c i t y requires l a r g e r deflections and, as a consequence, l a r g e r t a i l
loads. For the uncontrolled case t h e loads drop off beyond the c r i t i c a l
r o l l i n g v e l o c i t y i n t h a t t h e motions eventually become s t a b l e again
(ref. 1). Thus, loads with t h e c o n t r o l l e r tend t o become l a r g e r than
those without a t r o l l i n g v e l o c i t i e s beyond c r i t i c a l .

Another pertinent point regarding the perfect c o n t r o l l e r i s shown i n

figure 3, t h a t is, the increase i n loads both w i t h and without the con-
t r o l l e r a t the higher i n i t i a l acceleration. The loads are, thus, a func-
t i o n of t h e i n i t i a l angle of attack, and even l a r g e r i n i t i a l accelera-
t i o n s w i l l lead t o l a r g e r loads.

Before t h e r e s u l t s of a l l the calculations are summarized, a point

of s i g n i f i c a n t i n t e r e s t which e x i s t s , p a r t i c u l a r l y w i t h t h e perfect con-
t r o l l e r , merits some a t t e n t i o n here. Because the automatic systems (par-
t i c u l a r l y t h e perfect c o n t r o l l e r ) reduce t h e excursions i n angle of a t t a c k
and s i d e s l i p , t h e airplane with t h e systems operating tends t o r o l l f a s t e r
a t any a i l e r o n dfllection than without t h e system operating. I n f i g u r e 4
a r e shown t h e maximum maneuvering h o r i z o n t a l - t a i l loads plotted against
average r o l l i n g v e l o c i t y and a i l e r o n input f o r 2g f l i g h t with and without
t h e perfect c o n t r o l l e r operating.

For l g f l i g h t (not shown here) t h e loads are smaller with the con-
t r o l l e r operating than without a t any a i l e r o n deflection as w e l l as any
r o l l i n g v e l o c i t y except f o r t h e l a r g e s t r o l l i n g v e l o c i t y as sham i n
figure 3. For 2g f l i g h t , however, the loads, although smaller at a given
r o l l i n g v e l o c i t y with t h e controller, are always l a r g e r f o r any'given
p i l o t o r a i l e r o n input. Thus, there is a tendency i n r o l l s from greater
t h a n l g flight f o r the h o r i z o n t a l - t a i l loads with t h i s type of c o n t r o l l e r
t o be l a r g e r than without t h e controller f o r - a n y amount of applied aileron.

I n f i g u r e 5 is shown a summary of the maximum maneuvering horizontal-

t a i l loads and the maximum v e r t i c a l - t a i l loads calculated i n 360' r o l l s
a t a l l average r o l l i n g v e l o c i t i e s up t o about 2.2 radians per second.
Each bar represents t h e magnitude of the maximum load f o r each of several
conditions :

The original t a i l
The optimum-sized t a i l
The perfect c o n t r o l l e r (A)
The coupling-moment canceler (c )
The p i t c h damper (0.7 c r i t i c a l l y damped) (D)
The y a w damper (1.246 c r i t i c a l l y damped) (E).

The p r a c t i c a l c o n t r o l l e r i s omitted here because it resulted i n loads

and motions quite similar t o those of t h e perfect controller.
8E
9

For l g f l i g h t , these r e s u l t s show the l e a s t h o r i z o n t a l - t a i l load

f o r t h e coupling-moment canceler; t h e p i t c h damper shows similar r e s u l t s
The yaw damper showed l i t t l e improvement over the o r i g i n a l unaqpented
' .
case. For t h e v e r t i c a l t a i l , t h e perfect controller has t h e l e a s t load
w i t h reductions shown f o r t h e coupling-moment canceler and t h e p i t c h dam-
per, but again l i t t l e improvement i s shown f o r the y a w damper. The opti-
mum t a i l shows t h e improvements previously discussed ( f i g . 1).

For 2g flight, t h e h o r i z o n t a l - t a i l load f o r t h e perfect c o n t r o l l e r

has the increase i n load t h a t was discussed previously ( f i g . 4). Sizable
reductions i n t a i l load with t h e coupling-moment canceler and t h e p i t c h
damper exist, however.

For t h e v e r t i c a l - t a i l load, t h e least load i s encountered u i t h t h e

perfect c o n t r o l l e r as f o r the l g case, only f a i r reductions i n loads a r e
obtained w i t h t h e coupling-moment canceler, and a somewhat l a r g e r reduc-
t i o n w i t h t h e p i t c h damper. No data a r e available f o r the optimum t a i l
and t h e yaw damper f o r t h i s f l i g h t condition.

It i s clear, however, t h a t f o r a l l systems the loads increase mark-

edly with t h e i n i t i a l normal acceleration of t h e f l i g h t .

The t a i l loads without an automatic system are, of course, caused

primarily by t h e angles of a t t a c k and s i d e s l i p . With an automatic sys-
tem, however, the loads are caused by the s t a b i l i z e r and rudder deflec-

.
t i o n s as w e l l . The amount of control deflection required by any system
i s theref ore of extreme significance

Figure 6 shows t h e maximum control deflections required f o r each of

t h e various systems. Each bar represents the magnitude of t h e maximum
deflection required i n r o l l s of various average r o l l i n g v e l o c i t i e s up t o
about 2.2 radians per second. The basic airplanes w i t h the o r i g i n a l and
optimum t a i l s of course use no controls as noted by the zeros. No data
a r e available f o r t h e optimum t a i l or t h e y a w damper i n 2g f l i g h t . The
deflections used i n 2g f l i g h t are appreciably l a r g e r than those used i n
l g f l i g h t . The l a r g e s t s t a b i l i z e r deflection i s used by the perfect con-
t r o l l e r and the l e a s t , by t h e p i t c h damper. The l a r g e s t rudder deflection
i s used by t h e coupling-moment canceler and t h e least, by t h e yaw damper.

S t a b i l i z e r deflections of t h e order of 1l0, used by t h e perfect con-

t r o l l e r , may be a l a r g e r portion of the t o t a l available deflection than it
i s desirable t o use. The rudder deflections used by t h e perfect control-
l e r and t h e coupling-moment canceler are extremely excessive and c e r t a i n l y
could not be used. They are l a r g e r than t h e t o t a l available deflection of
40°. For s p e c i f i c cases, therefore, the e f f e c t s of l i m i t i n g t h e amount of
control deflection used must c e r t a i n l y be investigated.
13:-
10

It i s not s u f f i c i e n t t o evaluate an automatic system on t h e b a s i s

of t h e t a i l loads encountered or the c o n t r o l deflections required alone
because, except f o r t h e p e r f e c t c o n t r o l l e r , v a r i a t i o n s i n t h e normal and
l a t e r a l accelerations a l s o exist. With some automatic systems operating,
these accelerations s t i l l may be i n t o l e r a b l e t o t h e p i l o t .

I n figure 7 are shown t h e maximum normal and l a t e r a l accelerations

calculated i n 360° r o l l s a t a l l average r o l l i n g v e l o c i t i e s up t o about
2.2 radians per second. Each bar represents t h e magnitude of t h e maxi-
mum accelerations calculated f o r each of t h e various systems previously
discussed. "he accelerations shown occurred during r o l l i n g maneuvers
which were i n i t i a t e d from 1g and 2g flight .
The r e s u l t s ( f o r both 1g and 2g f l i g h t ) show that t h e v a r i a t i o n s i n
normal acceleration below the i n i t i a l values of l g or 2g a r e only s l i g h t l y
improved by any of the systems except, of course, t h e p e r f e c t c o n t r o l l e r
f o r which no changes occur. Negative accelerations are experienced f o r
a l l other systems. The p o s i t i v e v a r i a t i o n s i n normal acceleration are,
however, appreciably reduced by a l l systems but t h e yaw damper. For t h e
l a t e r a l accelerations, improvement i s obtained by a l l systems, although
accelerations of g f o r 1g f l i g h t and l g f o r 2g f l i g h t s t i l l a r e
experienced.

"here i s a s i z a b l e increase i n t h e accelerations encountered i n

rolls frm 2g f l i g h t over those from 1g f l i g h t .

It appears t h a t , i n coping w i t h t h e roll-coupling problem, a con-

siderable c m p r m i s e must be made between the motions or accelerations
t h a t t h e p i l o t must t o l e r a t e , t h e t a i l - l o a d s encountered, and t h e con-
t r o l deflections required by a system.

The other dominant s t a b i l i t y deficiency i s the problem of pitch-up.

.
Pitch-up occurs, of course, from n o n l i n e a r i t i e s i n t h e pitching-moment
c h a r a c t e r i s t i c s of an a i r p l a n e

The pitching-moment c h a r a c t e r i s t i c s f o r the hypothetical airplane

used f o r the calculations of t h i s paper are t y p i c a l of those of airplanes
with swept wings and high horizontal t a i l s having moderate n o n l i n e a r i t i e s
with angle of a t t a c k and are shown i n figure 8. The calculations were
made from a Mach number of 1, and changes i n aerodynamic-center p o s i t i o n
( f i g . 8) with Mach number as w e l l as pitching-moment n o n l i n e a r i t i e s w i t h
angle of a t t a c k influence t h e results.

It i s important t o r e a l i z e t h a t t h e dangers from pitch-up are not

only those occurring i n the p i t c h plane, but a l s o those which may occur -.
4 ?3
A, L.3 I
 0.. , I
I
,
I
.
,..
a
0 . 8 .
4.

- -- --
0 .
0 .
(.

ll

i n t h e l a t e r a l modes of motion when t h e large angles of a t t a c k r e s u l t i n g

from pitch-up may cause i n s t a b i l i t i e s i n s i d e s l i p , violent wing dropping,
and spinning. Aerodynamic cures a r e most desirable, of course, but i f
not possible, automatic augmentation of some s o r t appears necessary. For
t h e cases shown herein, the nonlinearities occur i n the range of angles
of a t t a c k and normal acceleration f o r which it i s desirable t o operate
t h e airplane. Thus, autamatic systems which abort a maneuver rather than
allow it t o progress reasonably are not desirable.

For the r e s u l t s presented herein, only two automatic systems are

t r e a t e d . These systems a r e shown i n t h e following t a b l e :

SENSING
SYSTEM DESCRIPTION
REQUIRED

4
5 =DAMPING RATIO
VARIABLE-
PITCH DAMPER
q AND a
OR nz
4
L a .

The first system i s a r a t h e r complex p i t c h damper, a variable p i t c h

damper. It i s representative, however, of systems which do not abort
maneuvers. The variable damping system i s one which becomes operative
beyond some predetermined angle of a t t a c k so t h a t t h e motion i s not slug-
gish i n the normal operating range of angles of attack, but a l s o s o that
t h e damping increases r a p i d l y as the angle of a t t a c k of pitch-up i s
approached, as shown by t h e small sketch of t h e v a r i a t i o n i n damping
rat io.

The second system i s a s t i c k pusher which does abort t h e maneuver.

The pusher used, however, i s one which senses only angle of a t t a c k o r
normal acceleration and becames operative only after t h e desired normal
acceleration i s reached. For the results considered herein the pusher
returned t h e s t a b i l i z e r only t o t h e o r i g i n a l t r i m position as shown by
t h e small sketch.

I n figure 9 a r e shown the results f o r t h e variable p i t c h damper. The

maximum normal acceleration, the maximum h o r i z o n t a l - t a i l load (which i s
negative and increases i n magnitude downward on the figure), and t h e maxi-
mum s t a b i l i z e r deflection required are shown as functions of rate of
change of damping r a t i o w i t h angle of attack. For these cases, t h e aug-
menter became a c t i v e when the angle of a t t a c k exceeded t h e i n i t i a l trim
.
. 0
L
0

12

angle of a t t a c k f o r l g . Also shown a r e t h e values of t h e loads f o r t h e

airplane with a l i n e a r pitching-moment curve. The results show an appre-
ciable reduction i n t h e normal-acceleration overshoot (accelerations
greater than 4g) w i t h increasing r a t e of change of damping r a t i o , t h e
large values leading t o less acceleration overshoot than e v e n t h e l i n e a r
case. The t a i l loads are s i m i l a r l y reduced i n magnitude and again, a t
t h e l a r g e r damping r a t i o s , smaller loads a r e obtained than f o r the l i n e a r
case. Rather sizable s t a b i l i z e r deflections a r e required by t h i s system;
however, values of as much as 5 O are required and t h i s i s somewhat g r e a t e r
than t h a t used by a moderate authority system. This may be considered an
excessPrc amount of control.

Some results f o r the s t i c k pusher a r e shown i n f i g u r e 10. Here a r e

shown t h e maximum normal acceleration and t h e maximum h o r i z o n t a l - t a i l
load as functions of push rate, t h a t is, the r a t e of change of s t a b i l i z e r
deflection w i t h time. The results are shown f o r three d i f f e r e n t s t a b i l i z e r
input rates, o r rates of pull-up. The results show l i t t l e improvement f o r
a s t i c k pusher of t h i s type. The normal acceleration overshoot i s reduced
only s l i g h t l y and t h e t a i l loads are e s s e n t i a l l y unchanged. Both the loads
and accelerations a r e appreciably l a r g e r than t h e values f o r the l i n e a r
pitching-moment case.

It appears t h a t a pusher of t h i s type, which allows t h e maneuver t o

reach i t s desired acceleration before operating, not only aborts t h e
maneuver but does l i t t l e good f o r t h e maximum loads encountered. A pusher
which operates e a r l i e r would, of course, produce l e s s loads but would
a l s o stop the maneuver much sooner. A pusher w i t h a n t i c i p a t i o n based on
pitching velocity o r acceleration i n conjunction w i t h a p i t c h damper
undoubtedly would prove u s e f u l on a l l counts,

Compatibility of Systems f o r Rolling Maneuvers and Pitch-Ulp

Inasmuch a s airplanes may be a f f l i c t e d by both pitch-up and r o l l i n g

divergences, t h e c m p a t i b i l i t y of an automatic system f o r one deficiency
w i t h t h e needs of t h e other deficiency i s important. Because of t h e
nature of pitch-up, a l l systems, used as a cure, require nose-down
pitching moments. I n addition, a l l roll-coupling systems require nose-
down pitching moments a t the onset of the r o l l i n g motion, primarily
because of t h e i n i t i a l positive pitching v e l o c i t y t h a t e x i s t s . A s a
matter of f a c t , a l l systems but t h e p i t c h damper require only nose-down
moments. Thus, systems f o r pitch-up would generally not have detrimental
e f f e c t s i n rolls and may be helpful. Systems used for roll-coupling which
require a sensing of r o l l i n g velocity would not operate i n a pitching
maneuver and thus would have no e f f e c t on pitch-up.

It must be pointed out t h a t i n r o l l i n g maneuvers f o r which t h e prin-

c i p a l a x i s i s below the f l i g h t path, i n f r o n t of t h e center of gravity,
i n i t i a l negative pitching v e l o c i t i e s are developed r a t h e r than p o s i t i v e
 ..I .- ** .
-_
v
a

values which occur f o r t h e cases discussed herein. Thus, f o r such cases

nose-up r a t h e r than nose-down pitching moments would be required by an
automatic system a t t h e onset of a r o l l . An automatic system f o r pitch-up
such as a pusher might thus be detrimental, whereas a pitch-damper would
s t i l l be e f f e c t i v e f o r r o l l i n g maneuvers.

CONCLUDING REMARKS

I n summary, a cursory study has been made of t h e e f f e c t s of auto-

matic augmentation and controlling systems on t h e t a i l loads and accel-
erations i n rolls and pitch-up. The r e s u l t s of t h i s study a r e not neces-
s a r i l y of general application but primarily show some of t h e problems and
trends t h a t may be expected. Calculations similar t o those presented
herein must, of course, be made f o r any specific design.

The r e s u l t s f o r rolls indicate the existence of an optimumtail

s i z e from t h e standpoint of the loads encountered, t h i s s i z e being most
naturally t h a t which i s l e a s t l i k e l y t o cause divergences. Automatic
systems ranging from a simple p i t c h damper t o a perfect controller show
reductions i n the violence of t h e motions and, i n general, a reduction
i n t h e t a i l loads. A t average r o l l i n g v e l o c i t i e s somewhat l a r g e r than
c r i t i c a l , t h e h o r i z o n t a l - t a i l loads obtained with c o n t r o l l e r s which sense
r o l l i n g v e l o c i t y may be l a r g e r than a r e obtained with no controller. I n
any event, i f automatic systems a r e t o be used, they must be considered i n
the i n i t i a l design t o obtain acceptable motions i n rolls and t o evaluate
properly t h e loads encountered.

For t h e pitch-up problem t h e maximum t a i l loads are primarily t h e

r e s u l t of control deflection and thus systems which e s s e n t i a l l y reduce
t h e input or pull-up deflections w i l l generally improve t h e acceleration
overshoots and reduce t h e h o r i z o n t a l - t a i l loads encountered.

Finally, i n general, systems used f o r pitch-up may be compatible with

t h e problems of roll coupling and generally should not be detrimental.
On the other hand, of t h e systems studied f o r roll coupling, o n l y the
p i t c h damper would be helpful f o r pitch-up.

136
 WERENCES

1. P h i l l i p s , W i l l i a m H.: Effect of Steady Rolling on Longitudinal and

Directional S t a b i l i t y . NACA TN 1627, 1948.

2. White, R. J., Uddenberg, R. C., Murray, D., and Graham, F. D.: The
Dynamic S t a b i l i t y and Control Equations of a Pivoted-Wing Supersonic
P i l o t l e s s Aircraft, With Downwash, Wake and Interference Effects
Included. Doc. No. D-8510, Boeing A i r c r a f t Co., Jan. 9, 1948.

3 . Weil, Joseph, and Day, Richard E.: A n Analog Study of t h e Relative

Importance of Various Factors Affecting Roll Coupling. NACA
RM ~ 5 6 ~ 0 61956.
,
4. Gates, Ordway B., Jr., Weil, Joseph, and Woodling, C. H.: Effect of
Automatic S t a b i l i z a t i o n on t h e Sideslip and Angle-of -Attack D i s -
turbances i n Rolling Maneuvers. NACA RM L55E25b, 1955.

5 . P h i l l i p s , W i l l i a m H.: Analysis of an Automatic Control To Prevent

Rolling Divergence. NACA RM L56AO4, 1956.

6. Gat?s, Ordway B., Jr., and Woodling, C. H.: A Theoretical Analysis

of t h e Effect of Engine Angular Mmentum on Longitudinal and
Directional S t a b i l i t y i n Steady Rolling Maneuvers. NACA RM L55GQ5,
1955
7. Finch, Thomas W., Peele, James R., and Day, Richard E.: F l i g h t Inves-
t i g a t i o n of t h e Effect of Vertical-Tail Size on t h e Rolling Behavior
of a Swept-Wing Airplane Having Lateral-Longitudinal Coupling. NACA
RM H55L28a, 1956.

8. Sisk, Thomas R., and Andrews, W i l l i a m H.:

Flight Experience With a
Delta-Wing Airplane Having Violent Lateral-Longitudinal Coupling
i n Aileron Rolls. NACA RM H55H03, 195.

9. Stone, Ralph W., Jr.: Some Notes on t h e Violent Lateral-Longitudinal

Coupling Motions of the Douglas X-3 Airplane i n Aileron Rolls. NACA
1956.
RM ~ 5 6 ~ 1 5 ,
10. Woodling, C. H.: Theoretical Investigation of t h e Effect of Rudder
and S t a b i l i z e r Deflections on t h e Angles of Attack and Sideslip i n
Rapid Rolls. NACA RM L57A30a, 1957.

11. Shortal, Joseph A., and Maggin, Bernard: Effect of Sweepback and
Aspect Ratio on Longitudinal S t a b i l i t y Characteristics of Wings at
Low Speeds. NACA TN 1093, 1946.
12. Donlan, Charles J., and Weil, Joseph: Characteristics of Swept Wings
a t High Speeds. NACA RM L52A15, 1952.

13. Weil, Joseph, and Gray, W. H.: Recent Design Studies Directed Toward
Elimination of Pitch-Up. NACA RM L53123c, lB3.
14. Toll, Thomas A.: Longitudinal Characteristics of Wings. NACA
RM L53121b, 1953.

15. Polhamus, Edward C., and Hallissy, Joseph M., Jr.: Effect of Airplane
Configuration on S t a t i c S t a b i l i t y a t Subsonic and Transonic Speeds.
NACA RM ~ 5 6 ~ 0 g a1936.
,
16. Curfman, Howard J., Jr.: Theoretical and Analog Studies of t h e
Effects of Nonlinear S t a b i l i t y Derivatives on t h e Longitudinal
Motions of an A i r c r a f t i n Response t o Step Control Deflections and
t o t h e Influence of Proportional Automatic Control. NACA Rep. 1241,
1955. (Supersedes NACA RM L50111.)
17. Oswald, Telford W.: The Effect of Nonlinear Aerodynamic Character-
i s t i c s on'the Dynamic Response t o a Sudden Change i n Angle of
Attack. Jour. Aero. Sci., vol. 19, no. 5, May 1952, pp. 302-316.

18. Bielat, Ralph P., and Campbell, George S.: A Transonic Wind-Tunnel
Investigation of t h e Longitudinal S t a b i l i t y and Control Character-
i s t i c s of a 0.09-Scale Model of t h e B e l l X-5 Research Airplane and
Comparison With F l i g h t .NACA RM ~ 5 3 ~ 1 8 1953.
,
19. Campbell, George S., and Weil, Joseph: The I n t e r p r e t a t i o n of Non-
l i n e a r Pitching Moments i n Relation t o t h e Pitch-Up Problem. NACA
RM L53102, 1953.

20. Sadoff, Melvin, Matteson, Frederick H., and Havill, C. Dewey: A

Method f o r Evaluating t h e Loads and C o n t r o l l a b i l i t y Aspects of t h e
Pitch-Up Problem. NACA RM A55D06, 1935.

21. Sadoff, Melvin, Stewart, John D., and Cooper, George E..: A Method
f o r Predicting Relative Severity of Pitch-Up Based on A Correla-
t i o n of t h e Computed Dynamic Behavior of Several Airplane Configu-
r a t i o n s With P i l o t Opinion. Prospective NACA paper.

22. Bihrle, W i l l i a m , Jr., and Stone, Ralph W., Jr.: Analytical Studies
of t h e Response t o Longitudinal Control of Three Airplane Configu-
r a t i o n s i n Landing Approaches. NACA RM L53B10, 1953.
16
+!I-- -'

C ~ l caefficients
l ana aerivatives are besea on w i n g areal]

IX,slug-ft2 .................................................
$,Slug-ft2 .................................................
I&$lug-ft2 ..................................................
IxZ,Slug-ft2.................................................
~ l b / s q. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
S,aqft
b,ft
E,ft
.. ... ... ... ... ... ........,............................... ... ... ... ... ... ....................................................'..........................*....* . .
W,lb . .. .. .. .. .. .. .. ..................... .. .. .. .. .. .. .. .. .. .. ..................... .. .. .. .. .. .. .. .. .. ...........
~ , f t / s e c. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
rn,SlUgS

h p f t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ~

M . . . . . . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ......................................
P,SlugS/Cuft

I me, ft-lb-sec . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17,554

xe
C2 ,perradian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -0.0528
. . . . . . . . . . . . . . .. ... . . . . . . . . .. . . . . . . . . . . . . . . . . . .
68
cZp, per'radian -0.255
C2,perradlan ............................................... 0.042
<,perradian ................................................ -1.0

~ p e r r a dan................................................
i -3.5
~ p e r r a dan......................
i . ......................... -1.5
C%,perradi an................................................ -0.36
c ,perlgdian . . . . . . . . . . . . . . . . . . . . e . . . . . . . . . . . . . . . . . . . . . . . . . . 0
"88
C,
s,
,perradian ............................................... -0.03

Cp,perradi an................................................ -0- 095

CDp,perrW a n . . . . . . . . . . . . . , . . . . . . . . . . . . . . .................... 0
c%, per radian ( o r i g i n a l tail) . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . 0.057
cnP, per radian (optimum tail) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.114
c y g , p e r r a d l a n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -0.50
CLkperradian................................................ 3-85
Ck,HT'perradlan .............................................. 0.755
. . . . . . . . . ............................................. 0.43
CYa vT, per radian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
da
-0.23
,
, p e r r ~ a n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.074
%s,m
CzP(a), per radian (shown i n following plot):

0
Clg, per radian

-0.139 - -

L
E 111.‘ .....
., 17

TABU SI

MASS CHATUCTERISTICSj STABILITY DERIVATIVES7 AND 0- FACTORS

USED I N THE C-IONS OF PITCH-UP

5,slug-ft2 . . . . . . . . . . . . . . . . . . . . . . . . log7 172

1 pV%,
2
lb/sq f t . . . . . . . . . . . . . . . . . . . . . . 348
-c , fsqt f t . .. .. .. .. .. .. ....... .. .. .. .. ..... .. .. .. .. ....... .. . . .
s, 530
W71b . . . . . . . . . . . . . . . . . . . . . . . . . . . 347 45016
m, slugs . . . . . . . . . . . . . . . . . . . . . . . . . .
V, f t / s e c . . . . . . . . . . . . . . . . . . . . . . . . . .
1 7 070

$7 ft . . . . . . . . . . . . . . . . . . . . . . . . . . 357 971 000

slugs/cu f t . . . . . . . . . . . . . . . . . . . . . . . 0.000737
Mach number . . . . . . . . . . . . . . . . . . . . . . . . .
p7
1.00
C , per radian . . . . . . . . . . . . . . . . . . . . . . -0 773
%T
Gq, per radian . . . . . . . . . . . . . . . . . . . . . . . -2.27
C
ma, per radian . ...................... 0.867
Ck7 per radian ....................... 4.41

TABU I11
SON3 TYPES OF ROLL-COUPLING CONTROIUBS AND AUGMENTERS

I
- System Sensing required Controls used

A Perfect controller

B Practical controller iT and 6r

-
P7 9 7 and
C Coupling -moment canceler o r pq iT and 6,
and pr

D P i t c h damper 9 iT
-
E Yaw damper r s,
-
18

EFFECT OF ROLLING VELOCITY ON MAXIMUM TAIL LOADS

360" LEFT ROLLS FROM I g FLIGHT

-ORIGINAL VERTICAL TAL

---OPTIMUM vmnw TAIL
MAXIMUM MANEUVERING MAXIMUM VERTICAL-
HORIZONTAL-TAIL LOAD TAIL LOAD
IO - X I 0 3
-
8 -
-
6-
ALHT,MAX, -
4-
-
2-
-
0 I 2 3 0 I 2 3
AVERAGE ROLLING VELOCITY, RAMANSISEC

Figure 1

EFFECT OF A PITCH DAMPER IN ROLLING MANEWERS FROM

FLIGHT TESTS
360" LEFT ROLLS FROM I g FLIGHT

-NO AUGMENTATION
- -- WITH AUGMENTATION ( DAMPING RATIO, 0.5;
MAXIMUM CONTROL AUTHORITY, 1.84)

MAXlMUM MANEUVERING MAXIMUM VERTICAL-

HORIZONTAL-TAIL LOPD TAIL LOAD

c" rD3

0 I 2 3 0 I 2 3
AVERAGE ROLLING VELOCITY, RADIANS/SEC

Figure 2
-.
 EFFECT OF PERFECT CONTROLLER ( @ = A a = O )
360’ L E F T ROLLS

-
--- WITHOUT CONTROLLER
W I T H CONTROLLER

MAXIMUM MANEUVERING I FLIGHT MAXIMUM VERTICAL-

HORIZONTAL-TAIL LOADS TAIL LOADS

CRITICAL
8rx103 ROLLING VELOCITY arx103

2 9 FLIGHT
12rx103

LU-HT, MAX 1
LB 4 LB

0 I 2 3 0 I 2 3
AVERAGE ROLLING VELOCITY, RADIANSISEC

Figure 3

MAXIMUM MANEWERING HORIZONTAL-TAIL LOADS WITH PERFECT

CONTROLLER

360° LEFT ROLLS FROM 2 9 FLIGHT

- WITHOUT CONTROLLER
- -- WlTH CONTROLLER

0 I 2 3 0 5 IO 15 20 25 30 35
AVERAGE ROLLING VELOCITY, TOTAL AILERON
RADIANS/SEC
DEFLECTION, DEG
 MAXIMUM TAIL LOADS IN 360" LEFT ROLLS
TYPICAL FIGHTER; M.0.7; hp=32,000 FT; ROLL RATES UP TO 2.2 RADIANSISEC

MANEUVERING VERTICAL-
HORIZONTAL-TAIL LOAD I TAIL LOAD

10 IO
8 8
MAX, 6
ALM, Lvr, 6
4 L B 4
2
0
WWTA C D E
2
0
ORKi-A
- '
TCtILS
- - C
SYSTEMS
D E

Figure 5

MAXIMUM CONTROL DEFLECTIONS REQUIRED IN 360" LEFT ROLLS

TYPICAL FIGHTER; M.0.7; hpa32,OOO FT; ROLL RATES UP TO 2.2 RADIANSISEC
STABILIZER DEFLECTION RUDDER DEFLECTION
I g FLIGHT

30 r 30r

2 g FLIGHT
70
60
50
ix MAX940
DEG 30
20
IO
0
OW. A C D E ORIG. A C D E
V
OPT
b -'
m.
TAILS SYSTEMS TAILS SYSTEMS

Figure 6
 MAXIMUM ACCELERATIONS IN 360" LEFT ROLLS
TYPICAL RGHTER; M.0.7; hp =32pOOFT ;
ROLL RATES UP TD 2.2 RADIANS/SEC
NORMAL ACCELERATION LATERAL ACCELERATION
1 g FLIGHT

AI-
", MAX^
g UNITS 0
-2 % , M A0 X * z L , 0 ,
g UNITS -
2 g RIGHT
6
4
"Z,MAX,
g UNITS "'t, MAXp2

--
0 g UNITS4
-2
ORKi
OPT
TAILS
A C D

SYSTEMS
E
0

--
OWG.
OPT
TAU
A C D

SYSTEMS
E

Figure 7

PITCHING-MOMENT CHARACTERISTICS
FOR P I T C H - U P CALCULATIONS

-.IO

Cm -20

-.30

-40
0 8 16 24 32 40
a, DEG

Figure 8
22

EFFECT OF MAGNITUDE OF VARIABLE PITCH DAMPING ON LOADS

IN PITCH-UP
M4.Q hp=%,OOO FT, PULL-UP TO 4 9; INPUT RATE, 2.5O/SEC
MAXIMUM NORMAL ACCELERATION

MAXIMUM HORIZONTAL-TAIL LOAD

MOMENT CURVE

0I L

MAXIMUM CONTROL REQUIRED

Air MAX*4
DEG :PI
0 .05 .IO .I5 20
,
0 .05
dUda
.IO .I5 .20

dt/da

Figure 9

EFFECT OF STICK PUSHER ON LOADS IN PITCH-UP

M4.Q %=35QOO FT; PULL-UP TO 4 9; INPUT RATE, 2.5YSEC
INPUT RATE, DEG/SEC
0
- - 2.5
---
--I-_--
- 5.0
A - 8.0
MAXIMUM NORMAL ACCELERATION MAXIMUM HORIZONTAL-TAL LOAD

L
-30 ,xi03 I
I
0 IO 20 0 IO 20
PLlSH RATE, DEG/ SEC

Figure 10
 AERODYNAMIC LOADS ON TAILS AT HIGH

ANGLES OF ATTACK AND SIDESLIP

By J. Richard Spahr
Ames Aeronautical Laboratory

and Edward C. Polhamus

Langley Aeronautical Laboratory

SUMMARY

Results are presented f o r the loads and moments acting on t h e indi-

vidual t a i l surfaces of a body-tail combination over a wide range of
angles of a t t a c k and s i d e s l i p . The e f f e c t s of forebody length and
panel-panel interference on t h e c h a r a c t e r i s t i c s a r e included. It i s
shown that large nonlinear variations i n these loads and moments occur
a t some combinations of angle of a t t a c k and s i d e s l i p which cannot be
predicted by low-angle theory. A r e l a t i v e l y simple, but general, theo-
retical method f o r calculating these load and moment c h a r a c t e r i s t i c s i s
described, and t h e results from t h i s method are found t o be i n good
agreement with experiment provided t h e i n i t i a l positions of t h e forebody
vortices are known.

It i s shownthat a simple application of slender-body theory can

be used t o predict the side loads due t o s i d e s l i p t h a t are contributed
by a v e r t i c a l t a i l on a w i d e variety of wing-body-tail combinations a t
low angles of attack. Several configuration changes a r e indicated which
reduced the v e r t i c a l - t a i l loads per u n i t yawing moment of the config-
uration a t l a r g e angles of attack.

Some results are presented on the e f f e c t of high angle of a t t a c k

on the induced-flow f i e l d and t a i l loads due t o a wing a t supersonic
speed.

INT3ODUCTION

Airplanes and missiles are frequently required t o operate over a

wide range of angles of a t t a c k and sideslip. The aerodynamic loading
on most a i r c r a f t configurations a t small angles of a t t a c k and s i d e s l i p
can be adequately predicted by linearized theories. However, a t combined
angles where e i t h e r the angle of a t t a c k or s i d e s l i p i s large, serious
n o n l i n e a r i t i e s i n t h e load c h a r a c t e r i s t i c s occur f o r which no general
t h e o r e t i c a l method has heretofore been developed. The results of
2

reference 1 have shown that, at subsonic Mach numbers, such nonlinear-

ities in the tail loads are caused by the effects of the body and wing
vortices and that these effects can be estimated if the strength and
positions of the vortices are known.

The purpose of this paper is threefold: (1) to sunat\Rsize the

results at supersonic Mach numbers of recent f’und.agental and systematic
measurements of the loads and moments on the exposed panels of body-
panel combinations through a wide range of angles of attack and sideslip;
(2) to present a general theoretical. method for calculating these load
characteristics at any Mach number, which requires only a knowledge of
the initial positions of the forebody vortices; and (3) to show the
influence of several configuration changes in reducing the vertical-
tail loads required for a given yawing moment of the configuration.

a maximum body radius

b
C local chord
-
C mean aerodynamic chord

Cn section normal-force coefficient

‘b bending-moment coefficient,
Bending moment about root of exposed panel
2qSsv

ch hinge-moment coefficient,
Hinge moment about centroid of exposed panel
2qSE

Cn yawing-moment coefficient of configuration

normal-force coefficient on exposed vertical tail, No-1 force

‘N, t qs
CN normal-force coefficient on tail in presence of wing

normal-force coefficient on tail alone

cNm
CY side-force coefficient, Side force on vertical panel
as
XY change in side-force coefficient of configuration due to
vertical tail

a t total differential horizontal-tail incidence

M free-stream Mach number

Q free-stream dynamic pressure
S spanwise distance from body center to panel tip

SV span of one exposed panel

S area of one exposed panel

V free-stream velocity
Z vertical coordinate with origin at tail-body juncture
U angle of attack, radians unless otherwise specified

P angle of sideslip

E semiapex angle of plan form

E' angle of average dawnwash acting over span of tail

Subscripts:

U upper panel
L lower panel

V due to body vortices

W wing alone

DISCUSSION

Basic Panel-Load Characteristics

The loads and moments acting on the individual lifting surfaces
of a cruciform combination are considered first. These surfaces can
4

be considered wing panels of a body-wing-tail combination o r t h e t a i l

surfaces of a wingless configuration. Systematic wind-tunnel t e s t s of
t h e configuration shown i n figure 1 have been made a t t h e Ames
Aeronautical Laboratory t o measure the forces and moments acting on t h e
four individual panels over a wide range of combined angles of a t t a c k
and s i d e s l i p . Representative r e s u l t s of these t e s t s a r e presented i n
figure 2 i n which t h e loads and moments acting on t h e upper and lower
v e r t i c a l panels are given f o r combined angles of a t t a c k and s i d e s l i p .
The c o e f f i c i e n t s of side force CY on each panel and hinge moment about
t h e panel centroid ch are shown a s functions of s i d e s l i p angle p f o r
angles of a t t a c k of Oo and 20°. It i s seen t h a t t h e e f f e c t of angle of
a t t a c k on the lower panel i s t o increase progressively t h e force while
t h e hinge moment remains unchanged. In contrast, t h e load and hinge
moment on the upper panel are both decreased by angle of a t t a c k . The
important c h a r a c t e r i s t i c t o note here i s that t h i s decrease i s not pro-
p o r t i o n a l t o angle of s i d e s l i p but reaches a maximum a t low values of
s i d e s l i p and r e s u l t s i n a large rearward s h i f t i n t h e panel center of
pressure and a highly nonlinear variation of t h e loads and moments on
t h i s panel w i t h angle of s i d e s l i p . It i s t h i s l o s s i n load on t h e
upper v e r t i c a l panel when serving as a t a i l f i n which i s one of t h e
causes of t h e serious decay i n d i r e c t i o n a l s t a b i l i t y of most airplanes
a t large angles of attack. It i s apparent that t h i s undesirable char-
a c t e r i s t i c can be a l l e v i a t e d by t h e use of a lower v e r t i c a l ( v e n t r a l )
f i n , because such a surface does not l o s e effectiveness w i t h angle of
a t t a c k but a c t u a l l y gains effectiveness. Results f o r the panel r o o t
bending moment a r e presented i n figure 3. The bending-moment r e s u l t s
i n conjunction w i t h t h e side-force results show that t h e l a t e r a l center
of pressure of the lower panel remains fixed w i t h changes i n angle of
attack, whereas that of t h e upper panel moves outboard with an increase
i n angle of attack. It i s apparent from symmetry that the r e s u l t s of
figures 2 and 3 apply d i r e c t l y t o horizontal panels i f the angles of
a t t a c k and s i d e s l i p a r e interchanged. The left-hand curves apply t o
t h e l e f t horizontal panel and t h e right-hand curves apply t o the right
panel.

Consider now t h e basic cause of these e f f e c t s of cross coupling

between angle of a t t a c k and s i d e s l i p . The loading due t o sideslip on
a v e r t i c a l panel i n t h e presence of a body varies with angle of a t t a c k
because of two d i f f e r e n t e f f e c t s : (1)t h e change i n e f f e c t i v e sweepback
of t h e panel and, ( 2 ) t h e change i n t h e influence of the forebody vor-
t i c e s on t h e panel loading.

Consider, first, the sweepback e f f e c t . It can be shown by use of

slender-body theory t h a t t h e lift effectiveness of a wing panel i n t h e
presence of a body decreases with increasing sweepback as i n t h e case
of a wing alone. This e f f e c t i s i l l u s t r a t e d i n f i g u r e 4 which shows
t h e side view of t h e v e r t i c a l panels i n combination with t h e body a t an
a r b i t r a r y angle of a t t a c k and s i d e s l i p . The e f f e c t of angle of a t t a c k
r
149
 ..

i s t o increase the leading-edge sweepback of t h e upper panel and t o

decrease t h e sweepback of t h e lower panel. The r e s u l t i n g changes i n t h e
loading due t o s i d e s l i p over each wing panel i s indicated i n t h e r i g h t -
hand sketch of figure 4. Here, it i s shown t h a t the symmetrical span-
wise load d i s t r i b u t i o n a t zero angle of a t t a c k i s changed t o an asym-
metrical d i s t r i b u t i o n a t positive angle of attack. The loading on t h e
lower panel i s increased and that on the upper panel i s decreased.

Consider next t h e second f a c t o r influencing t h e panel loads a t com-

bined angles of a t t a c k and s i d e s l i p , that i s , t h e influence of t h e fore-
body vortices. A t moderate and large angles of a t t a c k or s i d e s l i p , the
flow over a body i s characterized by a p a i r of symmetrically disposed
vortices on t h e leeward side caused by crossflow separation. The pres-
ence of a vortex i n t h e v i c i n i t y of a wing or t a i l surface changes t h e
loading on t h e surface by v i r t u e of the induced flow f i e l d created by
the vortex. Two c r i t i c a l conditions are indicated i n figure 5 f o r which
a vortex passes close t o one panel and thus has t h e g r e a t e s t e f f e c t on
t h e loading: first, t h e combination of high angle of a t t a c k and low
s i d e s l i p i n which t h e upper panel i s primarily affected by t h e vortex
and, second, high s i d e s l i p and low angle of a t t a c k i n which t h e l e f t
panel i s most affected. It w i l l be recalled from a previously given
figure that these were t h e two conditions f o r which t h e l a r g e s t nonlinear
changes i n panel loading occurred. The e f f e c t of a vortex on the panel
loads i s i l l u s t r a t e d by sketches which show t h e changes i n t h e span-
wise load d i s t r i b u t i o n , from s t r i p theory, due t o a vortex passing
near t h e t i p of each panel. It i s observed t h a t , above t h e vortex on
the upper panel, a loading t o the l e f t occurs and below t h e vortex a
loading t o t h e right occurs. This loading corresponds t o t h e d i s t r i -
bution of sidewash induced by t h e vortex along t h e span of t h e panels.
The magnitude of t h e loading increases with the strength of t h e vortex
and decreases w i t h t h e distance of the vortex &om the panel. With t h e
vortex located near t h e t i p of t h e panel, t h e net force due t o t h e vor-
t e x i s t o t h e right and reduces t h e l o a d and bending moment e x i s t i n g on
t h e panel without t h e vortex present. If t h e vortex i s moved toward
the body, i t s e f f e c t on t h e net panel load would diminish, because the
two regions of opposite loading would become more nearly comgensating.
Ekpressions based on simple t h e o r e t i c a l concepts have been derived
f o r t h e prediction of t h e forces and moments acting on a panel a t arbi-
t r a r y angles of a t t a c k and s i d e s l i p i n which the e f f e c t s of both sweep-
back and body vortices have been taken i n t o account. These expressions
a r e i l l u s t r a t e d i n figure 6 which gives t h e equations f o r t h e side force
acting on t h e t w o v e r t i c a l panels a t a given s i d e s l i p angle. The f i r s t
term i n each of these expressions represents t h e side force a t zero
angle of attack; t h e second term represents the e f f e c t of sweepback due
t o angle of attack; and t h e l a s t term represents t h e contribution of
the forebody separation vortices. The f a c t o r Cy i s the side force of
W
6

t h e panel alone a t zero angle of a t t a c k and i s evaluated from experiment,

where available, or from a s u i t a b l e wing theory. The f a c t o r s Kw and
K$ are both computed by slender-body theory and, f o r a c i r c u l a r body,
depend only on t h e r a t i o of the body radius a t o t h e panel s e d s p a n
s as shown by t h e curves i n figure 6. It i s observed t h a t increases
from 1 t o 2 as the configuration changes f’rom an a l l - w i n g configuration
t o a body with no wings. The f a c t o r 9, on the other hand, increases
t o a maximum f o r combinations with r e l a t i v e l y small bodies and then
decreases t o zero as the wings vanish. It i s a l s o noted that, because
of panel-panel interference, K$ i s l a r g e r f o r t h e planar configuration
than f o r t h e cruciform arrangement, but that t h i s e f f e c t becomes smaller
as a/s increases. Although t h e f a c t o r s I$J and 9 have been com-
puted from slender-body theory, these equations a r e not necessarily
restricted t o combinations having slender panels because I$, and
J
K$
a r e simply load r a t i o s which m o d i a t h e load on a v e r t i c a l - t a i l panel
alone t o take i n t o account the presence of the body and of t h e angle of
attack. The t h e o r e t i c a l and experimental comparisons of Nielsen and
Kaattari ( r e f . 2) have established t h e v a l i d i t y of the f a c t o r Q f o r
e s s e n t i a l l y any plan form or aspect r a t i o .

The importance of the panel leading-edge sweepback on the side force

i s seen f r o m t h e equations where t a n E, which i s proportional t o the
aspect r a t i o f o r a t r i a n g u l a r plan form, appears i n the denominator.
Thus, the lower t h e aspect r a t i o the l a r g e r the e f f e c t of angle of a t t a c k
on t h e v e r t i c a l panel loads due t o sideslip.

The evaluation of the l a s t terms i n these equations requires the

computation of t h e e f f e c t of each of t h e forebody vortices and t h e i r
images on the wing-panel loading. The strength and paths of these vor-
t i c e s have been calculated by means of a stepwise procedure based on
incompressible vortex theory. Such calculations, however, require a
knowledge of t h e normal-force d i s t r i b u t i o n along the forebody and t h e
i n i t i a l vortex position. The work of Jorgensen and Perkins ( r e f . 3 )
has demonstrated the v a l i d i t y of t h i s method f o r bodies of c i r c u l a r
cross section, and subsequent work has indicated i t s a p p l i c a b i l i t y t o
bodies of a r b i t r a r y cross section. The calculation of the e f f e c t of
a vortex of known strength and position on t h e loading of a wing panel
i n t h e presence of a body i s most simply made by means of s t r i p theory.
I n the application of the s t r i p theory, it i s assumed t h a t t h e strength
and path of each vortex remain unchanged by t h e addition of t h e wing
panel t o t h e body. The downwash f o r an ideal vortex i s then calculated
and t h e a d d i t i o n a l spanwise load d i s t r i b u t i o n due t o t h e vortex flow
f i e l d i s computed. The l i f t i n g effectiveness of each longitudinal s t r i p
i s taken as t h e two-dimensional value f o r a wing having t h e same leading-
edge sweepback. A n a l t e r n a t e method, based on slender-body theory, i s
 7

currently being investigated a t t h e Ames Aeronautical Laboratory f o r t h e

calculation of t h e e f f e c t s of a vortex on a wing panel i n the presence
of a body. I n t h i s method the influence of t h e wing-panel flow f i e l d
on t h e vortex paths i s taken i n t o account and, thus, one of t h e assump-
t i o n s of the strip-theory method i s avoided. Preliminary results of
t h i s study indicate t h a t t h e e f f e c t s of such changes i n t h e vortex'path
on t h e wing load may be important f o r plan forms of low aspect r a t i o .

Calculations have been made by means of these t h e o r e t i c a l methods

t o predict t h e experimental panel-load c h a r a c t e r i s t i c s discussed earlier.
Figure 7 shows a comparison between experiment and theory a t an angle
of attack of 200 f o r t h e v a r i a t i o n of side force and hinge moment with
s i d e s l i p angle f o r t h e two v e r t i c a l panels.

I n addition t o t h e experimental values shown by the syaibols three

t h e o r e t i c a l curves are shown i n each case t o i l l u s t r a t e the importance
of the two aerodynamic e f f e c t s j u s t discussed: first, a low-angle
theory, given by t h e f i r s t term i n t h e equations of f i g u r e 6; second,
the theory including the sweepback e f f e c t , given by the first two terms;
and, t h i r d , the theory including both t h e sweepback and vortex e f f e c t s
(based on t h e experimental i n i t i a l vortex positions) as given by t h e
complete equations. It i s seen f r o m t h e s e comparisons t h a t t h e load
c h a r a c t e r i s t i c s of e i t h e r panel can be predicted adequately by the com-
p l e t e theory. However, it i s noted t h a t f o r t h e lower panel t h e vortex
e f f e c t i s negligible, a result of t h e body v o r t i c e s being a r e l a t i v e l y
large distance f r o m t h i s panel.

The e f f e c t s of forebody length on t h e vertical-panel loads are indi-

cated i n figure 8. I n t h i s figure i s presented a comparison of t h e load
c h a r a c t e r i s t i c s of the o r i g i n a l combination w i t h a combination having
one-half t h e o r i g i n a l forebody length. The importance of forebody length
i s evident from t h i s comparison which shows t h a t t h i s change i n t h e fore-
body v i r t u a l l y eliminates the n o n l i n e a r i t i e s due t o t h e body vortices.
T h i s result i s caused by the reduction i n t h e vortex strength and by
t h e inboard movement of t h e vortices, both of which tend t o reduce the
e f f e c t of the v o r t i c e s as pointed out earlier. These experimental
r e s u l t s are i n good agreement with those predicted by the t h e o r e t i c a l
method described easlier.

The e f f e c t s of the presence of adjacent surfaces on the loads acting

on a wing o r t a i l surface are i l l u s t r a t e d i n figure 9. I n t h i s figure
i s shown the v a r i a t i o n of t h e side force and hinge moment of each v e r t i -
c a l panel w i t h angle of s i d e s l i p a t a high angle of a t t a c k i n t h e pres-
ence of each of t h e other panels of a cruciform arrangement. These
curves show that the addition of the opposite v e r t i c a l o r the l e f t hori-
zontal panel has no e f f e c t on the loads of either v e r t i c a l panel, but
t h e addition of t h e r i g h t horizontal panel increases the load on the
lower panel and decreases t h e load on the upper panel. These panel-panel
00
I .
t a
L -
ma
- 0 0
- - - a 0 0 00 a.

interference e f f e c t s a r e associated p a r t l y with the cross coupling of

t h e sidewash v e l o c i t i e s i n p o t e n t i a l flow and p a r t l y w i t h interference
e f f e c t s of the forebody vortex flow.

The loads on the t a i l surfaces of complete configurations a t zero

or small angles of a t t a c k a r e next considered. Under these conditions,
t h e side load contributed by a wing or t a i l surface t o an a r b i t r a r y con-
f i g u r a t i o n a t any s i d e s l i p angle o r Mach number can be computed by means
of a simple t h e o r e t i c a l method. I n t h i s method t h e side-force contri-
bution of a v e r t i c a l - t a i l surface i s given by the product of the side
force a c t i n g on t h e surface alone and an interference f a c t o r which i s
a function only of the cross-sectional shape of t h e combination a t the
t a i l location. Slender-body theory i n conjunction with apparent-mass
r e l a t i o n s h i p s a r e used t o evaluate t h i s interference f a c t o r f o r a given
configuration. T h i s t h e o r e t i c a l method has been used t o p r e d i c t t h e
side-force c h a r a c t e r i s t i c s of a number of current aircraft configurations
shown i n f i g u r e 10, f o r which such experimental information w a s a v a i l -
able. Most of these configurations represent recent fighter airplanes
and include a wide v a r i e t y of wing-body-tail arrangements. I n addition,
a research model from t h e National Advisory Committee f o r Aeronautics
i s included. The cross-sectional arrangements used i n t h e t h e o r e t i c a l
calculations t o approximate t h e a c t u a l configurations are shown i n f i g -
ure 11. The c o r r e l a t i o n between experiment and theory i s given i n terms
of t h e change i n side force due t o t h e addition of the surface designated
by the s o l i d l i n e i n each of t h e sketches. The results f o r subsonic
speeds are indicated by t h e flagged symbols and those f o r supersonic
speeds by the p l a i n symbols. Although the change i n s i d e force K y
shown i n t h i s figure includes t h e load induced on t h e body by the sur-
face, t h e load on the surface i t s e l f can a l s o be calculated by the
t h e o r e t i c a l method. The good agreement shown here between experiment
and theory f o r such a wide v a r i e t y of configurations a t both subsonic
and supersonic speeds shows t h e general usefulness of t h e theoretical.
method i n t h e prediction of s i d e loads and v e r t i c a l - t a i l loads due t o
s i d e s l i p a t small angles of a t t a c k where i n some cases these loads a r e
maximum.

The e f f e c t s of angle of a t t a c k on the t a i l loads of wing-body-tail

combinations a r e now considered. Figure 12 shows some r e s u l t s obtained
a t t h e Langley Aeronautical Laboratory which show t h e e f f e c t of angle
of attack on the spanwise load d i s t r i b u t i o n due t o s i d e s l i p on the v e r t i -
cal. t a i l of an airplane configuration having a low, midposition, or high
wing a t a Mach number of 0.8. It i s observed that i n a l l cases an
increase i n angle of a t t a c k causes an increase i n t h e loading along the
outer portion of the span and a decrease near t h e root. From t h e pre-
vious discussions it can be recognized t h a t t h i s e f f e c t i s associated
w i t h the forebody crossflow separation v o r t i c e s . It can a l s o be noted
by comparing the low- and high-wing r e s u l t s t h a t t h e e f f e c t of wing
e -

a
Le
‘3
 9

height i s considerably greater a t an angle of a t t a c k of 16O than a t a n

angle of a t t a c k of Oo; thus, the probability of a rather strong e f f e c t
of t h e wing on the forebody vortices i s indicated. Figure 13 shows s i m -
i l m results f o r a Mach number of 1.4 but, i n t h i s case, the e f f e c t s of
angle of a t t a c k and wing height are somewhat less.

Effect of Directional S t a b i l i t y

T h i s paper so far has dealt w i t h methods of estimating wing or t a i l

loads and comparisons of t h e v e r t i c a l - t a i l loads encountered by various
configurations f o r a given s i d e s l i p angle where it has been shown t h a t
f o r conventional configurations the v e r t i c a l - t a i l load generally decreases
w i t h angle af attack. T h i s reduction i n t a i l load i s , however, usually
accompanied by a l o s s i n d i r e c t i o n a l s t a b i l i t y which r e s u l t s i n l a r g e r
s i d e s l i p angles being encountered before t h e r e s t o r i n g moment necessary
t o counteract a given disturbance i s developed. If, as i n t h e usual
case, t h e wing-fuselage combination i s d i r e c t i o n a l l y unstable, then the
increased sideslip angle would require more t a i l load t o counteract t h e
i n s t a b i l i t y of t h e wing-fuselage combination and the t o t a l t a i l load
would increase w i t h angle of a t t a c k despite t h e f a c t t h a t t h e t a i l load
per u n i t sideslip decreased with angle of attack. 'Rowever, i f the wing-
fuselage combination has n e u t r a l d i r e c t i o n a l s t a b i l i t y throughout t h e
angle-of-attack range, t h e v e r t i c a l - t a i l load would ( i f any v a r i a t i o n of
carryover w i t h angle of a t t a c k i s neglected) be independent of overall
d i r e c t i o n a l s t a b i l i t y . Unfortunately, t h e i n s t a b i l i t y of t h e wing-
fuselage combination often increases w i t h angle of a t t a c k (see ref. 4)
and results i n an increase i n t h e vertical-tail loads encountered.

I n view Qf t h e importance of the wing-fuselage d i r e c t i o n a l s t a b i l i t y

c h a r a c t e r i s t i c s , it i s t h e purpose of t h i s portion of the paper t o i l l u s -
t r a t e t h e e f f e c t , on the v e r t i c a l - t a i l load per u n i t restoring moment,
of several wing-fuselage-configuration changes which appear a t t r a c t i v e
from s t a b i l i t y considerations. Figures 14 t o 16 i l l u s t r a t e the effect
of s t a b i l i t y by presenting, as a function of angle of attack, the
vertical-tail-load coefficient f o r a u n i t yawing-moment coefficient of
t h e complete configuration. Inasmuch as the only purpose of these f i g -
ures i s t o indicate the e f f e c t of angle of a t t a c k and compare changes
t o a given configuration, and since t h e wing areas and spans and moment
reference points are involved when making comparison between d i f f e r e n t
configurations, t h e scales have been omitted. One such change i s t h e
use of narrow horizontal s t r i p s , o r strakes, on t h e fuselage forebody
t o a l l e v i a t e t h e l o s s i n directional. s t a b i l i t y w i t h angle of attack.
(See r e f . 5.) The e f f e c t t h a t these strakes have on t h e measured load
of the exposed v e r t i c a l t a i l per u n i t yawing moment of the complete con-
f i g u r a t i o n are shown i n figure 14 f o r a Mach nuniber of 0.6. The strakes
had spans equal t o 10 percent of the maximum fuselage diameter and
extended over t h e f r o n t 27 percent of t h e fuselage. The wing, which
2 - 0 ,# .
a 0
..I
0. - .. 0 0 0 0. .. I..

was an aspect-ratio-4 delta wing that was clipped to an aspect-ratio-3

delta wing, differed from that in reference 5 and was selected because,
for this wing, the strakes also provide a beneficial effect on the longi-
tudinal stability characteristics. The strake-off condition is repre-
sented by the circular symbols and it can be observed that the tail load
per unit yawing moment increases rapidly. With the strakes on (denoted
by square symbols), however, the vertical-tail load decreases with angle
of attack and is considerably less than that for the basic configuration.
This rather large improvement is associated with a rather large reduction
in the wing-fuselage instability at the higher angles of attack.
The effect of a change in fuselage-afterbody shape on the vertical-
tail loads, as indicated by the tail contribution to side force, is pre-
sented in figure 15 for a Mach number of 2.0. Two configurations,were
investigated, one having a cylindrical afterbody and the other having
an elliptical afterbody. Both fuselages had the same volume and the
same longitudinal distribution of cross-sectional area. The elliptical
afterbody is used as an attempt to reduce the fuselage instability over
the entire angle-of-attack range while maintaining the necessary fuselage
volume. The desired stability results were obtained and were accompanied
by large reductions in the vertical-tail load for a given yawing moment.
This is illustrated by the two variations with angle of attack. The
solid cmve represents the tail load for the configuration with the
cylindrical afterbody, and the dashed curve represents the tail load for
the elliptical afterbody configuration. The results indicate a large
reduction in the tail load over the entire angle-of-attack range. This
reduction in vertical-tail load results fromthe decrease in fuselage
instability and a transfer of load from the tail to the fuselage after-
body which may, in many cases, be a desirable trade.
The effect of wing height on the vertical-tail loads at a Mach nwn-
ber of 2.9 is shown in figure 16. In this figure are shown the tail-
load results (as indicated by the tail contribution to side force) for
a configuration having a triangular wing of aspect ratio 4 in a low and
high position as indicated in the upper left-hand sketch. From a com-
parison of these curves, it is observed that raising the wing f’romthe
low to high position effects a significant reduction in the tail loads,
especially at large angles of attack. This reduction is caused by the
influence of the wing pressure field acting on the body as indicated in
the sketches on the right. As pointed out earlier, the n o m 1 force on
the right panel is greater than that on the left panel, and thus a lateral
pressure difference across the body is created by the wing. It is appar-
ent that the difference in the‘positivepressures fromthe high wing
causes an increase in the body side force and, hence, a stabilizing
moment, but that the low wing produces a decrease in side force. The
yawing moment induced by the low wing is small because of the short
moment arm. Thus, the favorable effect of a high wing on the tail loads
is the resdt of the improvement in the tail-off directional stability. - I
 I , .. ... . P e .. .. .C ..
1

11

I n the e s t i m t i o n of t a i l bending moments, however, t h e adverse e f f e c t

of a high wing on t h e span loading, as previously pointed out, must be
considered.

Effect of Lateral Control

One l a t e r a l - c o n t r o l device which has been found t o induce rather

l a r g e loads on the v e r t i c a l t a i l i s t h e d i f f e r e n t i a l l y deflected hori-
zontal t a i l . This device i s i l l u s t r a t e d i n f i g u r e 1-7 where the measured
normal force induced on the exposed portion of t h e v e r t i c a l t a i l by a
d i f f e r e n t i a l l y deflected horizontal t a i l i s presented (denoted by t h e
square symbols) as a function of Mach number. The v e r t i c a l - t a i l loads
presented f o r the d i f f e r e n t i a l l y deflected horizontal t a i l were obtained
a t zero s i d e s l i p w i t h a t o t a l d i f f e r e n t i a l d e f l e c t i o n of 30' ( r i g h t
down 15O, l e f t up 13O). These deflections produce a p o s i t i v e roll and
a wing-tip h e l i x of approximately 0.07 throughout the Mach number range
investigated. The v e r t i c a l - t a i l load, which, of course, OCCUI'S only
instantaneously as the roll control i s applied, increases r a t h e r r a p i d l y
w i t h Mach number and approximately doubles between a Mach number of 0.6
and 0.9. Although experimental results are not available f o r supersonic
speeds, t h e o r e t i c a l considerations of the e f f e c t of Mach number on the
effectiveness of t h e horizontal and v e r t i c a l t a i l s and t h e amount of
v e r t i c a l - t a i l area within t h e Mach cone f r o m t h e horizontal t a i l indi-
cate that a r a t h e r rapid reduction i n t h e v e r t i c a l - t a i l load might be
expected a t supersonic speeds. It would appear, therefore, that the
maximum loads induced on the v e r t i c a l t a i l probably occur a t high sub-
sonic speeds. I n order t o give a b e t t e r indication of the magnitude of
t h e normal force induced on t h e v e r t i c a l t a i l , t h e normal force encoun-
t e r e d a t an angle of s i d e s l i p of 60 w i t h no d i f f e r e n t i a l d e f l e c t i o n of
t h e horizontal t a i l i s presented by t h e c i r c u l a r symbols as a function
of Mach number f o r comparison. T h i s comparison i n d i c a t e s that a t high
subsonic speeds v e r t i c a l - t a i l normal forces corresponding t o an angle
of s i d e s l i p of approximately 7 O are induced by a t o t a l d i f f e r e n t i a l
deflection of 300 of the horizontal t a i l .

E f f e c t s of Wing on Horizontal-Tail Loads

The influence of a wing on t h e horizontal-tail-load c h a r a c t e r i s t i c s

a t two angles of a t t a c k i s indicated i n f i g u r e s 18 and 19. I n these
figures are shown some recent wind-tunnel results f o r t h e induced flow
f i e l d behind a rectangular wing a t zero s i d e s l i p and t h e influence of
t h i s flow f i e l d on the load a c t i n g on a rectangular h o r i z o n t a l - t a i l sur-
face. The l o c a l downwash and sidewash angles, dynamic pressure, and
Mach numiber were surveyed i n a v e r t i c a l plane a t the t a i l location. I n
these figures t h e v a r i a t i o n s i n downwash and i n t h e t a i l load are shown
as functions of the t a i l height i n wing chords. The t a i l normal f o r c e
CN i s computed f r o m t h e average downwash, dynamic pressure, and Mach nun-
ber e x i s t i n g a t t h e t a i l plane. The r e s u l t s f o r an angle of a t t a c k of 6'
( f i g . 18) show that, when t h e t a i l i s located between t h e t w o t r a i l i n g -
edge waves, the average downwash i s small, t h a t i s , 1 E' i s nearly 1,-a.
and t h e l o s s i n t a i l load i s small, t h a t is, t h e lift r a t i o i s close t o 1.
The e f f e c t of t h e viscous wake f r o m t h e blunt t r a i l i n g edge of the wing
i s evidenced by t h e variations i n t h i s region d i r e c t l y behind t h e t r a i l i n g
edge. If t h e t a i l surface passes through e i t h e r of t h e trailing-edge
waves, a sudden downwash occurs which actually reverses t h e load on t h e
t a i l , a s i t u a t i o n obviously t o be avoided. A s t h e t a i l i s r a i s e d f a r t h e r
above t h e wing plane, the downwash diminishes t o zero and t h e t a i l load
r e t u r n s t o i t s free-stream value a t t h e point f o r which t h e t a i l i s
located a t t h e shock wave f r o m t h e wing leading edge. Similarly, t h e
curves f o r negative values of t a i l height would be expected t o r e t u r n t o
free-stream conditions when t h e t a i l passes below the lower shock wave
f r o m t h e wing leading edge. The close agreement of the downwash and t a i l -
load curves i s an indication of the small deviations i n t h e average
dynamic pressure and Mach number i n t h e wake f r o m t h e i r free-stream values
a t t h i s angle of attack. The r e s u l t s f o r an angle of a t t a c k of 20°
( f i g . 19) a l s o show a sudden downwash and l o s s i n t a i l load as the t a i l
moves through t h e upper wave f r o m t h e wing t r a i l i n g edge. However, it
i s noted that, i n contrast t o the r e s u l t s f o r an angle of a t t a c k of 6 O ,
the t a i l - l o a d curves do not follow t h e downwash curves i n the r e s t of
t h e wake, and despite the increase i n downwash below t h e wing, the t a i l
load i s preserved. This e f f e c t r e s u l t s from changes i n dynamic pressure
and Mach number i n the wing wake a t t h i s angle of a t t a c k and demonstrates
the necessity of accounting f o r these changes i n the prediction of loads
on both horizontal- and v e r t i c a l - t a i l surfaces behind a wing a t large
angles of attack.

CONCLUDING RESIARKS

The r e s u l t s of wind-tunnel tests have shown t h a t variations i n t h e

loads and moments on a l i f t i n g surface i n t h e presence of a body can
exhibit l a r g e nonlinearities a t combined angles of a t t a c k and s i d e s l i p
which cannot be predicted by low-angle theory. A r e l a t i v e l y simple, but
general, t h e o r e t i c a l method f o r calculating these loads under such con-
d i t i o n s has been described i n which both t h e e f f e c t s of leading-edge
sweepback and of the forebody vortex flow a r e taken i n t o account. This
calculative method, however, requires a knowledge of t h e i n i t i a l positions
of t h e forebody vortices. It w a s found that t h e results of t h i s theo-
r e t i c a l method were i n good agreement w i t h experiment. Results have
been presented which show t h e e f f e c t s of forebody length and panel-panel
interference on t h e load c h a r a c t e r i s t i c s .
-.
It w a s demonstrated t h a t a simple application of slender-body theory
could be used successfully t o predict t h e side loads contributed by a
v e r t i c a l t a i l on a wide v a r i e t y of wing-body-tail combinations a t small
angles of a t t a c k and a t both subsonic and supersonic Mach numbers.

The three configuration changes described which reduced the v e r t i c a l -

t a i l loads required f o r a given y a e n g moment of t h e configuration a t
large angles of a t t a c k were the addition of horizontal strakes along
the forebody, t h e change i n t h e afterbody cross-sectional shape f r o m
c i r c u l a r t o e l l i p t i c a l i n which t h e major axis i s v e r t i c a l , and t h e
movement of t h e wing from a low t o a high position.
 ..

REFERENCES

1. Kuhn, Richard E., Hallissy, Joseph M., Jr., and Stone, Ralph W., Jr.:
A Discussion of Recent Wind-Tunnel Studies Related t o t h e Problem of
Estimating Vertical- and Horizontal-Tail Loads. NACA RM L55E16a,
1955
2. Nielsen, Jack N., and Kaattari, George E.: Method f o r Estimating L i f t
Interference of Wing-Body Combination a t Supersonic Speeds. NACA
RM ~ 5 1 ~ 0 4i
s, i
.

3 . Jorgensen, Leland H.,and Perkins, Edward W.: Investigation of Some

Wake Vortex Characteristics of an Inclined Ogive-Cylinder Body a t
Mach Number 1.98. NACA RM A55E31, 1955.

4. Polhamus, Edward C., and Hallissy, Joseph M., Jr.: Effect of Airplane
Configuration on S t a t i c S t a b i l i t y a t Subsonic and Transonic Speeds.
NACA RM L56AO9a, 1956.

5. Sleeman, W i l l i a m C., Jr.: Investigation a t High Subsonic Speeds of

t h e Effects of Various Horizontal Fuselage Forebody Fins on t h e
Directional and b n g i t u d i n a l S t a b i l i t y of a C o q l e t e Model Having
a 4 5 O Sweptback Wing. NACA RM L56J25, lB7.
 TEST CONFIGURATION

REAR VIEW
I
PLAN VIEW

Figure 1

LOADS ON VERTICAL PANELS

M=2.0

.05r
0

-.05 c ch

0 IO 20 30 0- IO 20 30
8, DEG 8,DEG
, 0.. 1 . . . - -. - ,. . . .
* 1
. ?e 7. .
C

16

c;.v:;
BENDING MOMENTS ON VERTICAL-TAIL PANELS
M = 2.0

UPPER PANEL LOWER PANEL

CY

7
?2 3 b : ,

cb 20°
-. I -.I ..

0' IO 20 30 0' IO 20 30
B, DEG B, DEG

Figure 3

EFFECT OF ANGLE OF ATTACK ON LOADING

DUE TO SIDESLIP
 EFFECT OF FOREBODY VORTICES

HIGH 0

I
HIGH 8, LOW 0

Figure 5

PANEL FORCES FOR COMBINED a! AND p

K4
6 CY"'CY&W - a )+CYV"
C Y L = c Y w ( K W +Kb
~a)+cYvL
2.0 -

1.8 -

1.6 -

I
0 .2 .4 .6 .8 1.0 0 .2 4 .6 .8 1
.0
a/s

Figure 6
18

COMPARISON OF EXPERIMENT WITH THEORY

i.i*/
M=2.0; a = 2 0 "
---
--_------
LOW-ANGLE THEORY
LOW-ANGLE THEORY + SWEEPBACK EFFECTS
LOW-ANGLE THEORY + SWEEPBACK AND VORTEX EFFECTS
0 EXPERIMENT
r UPPER PANEL I -.6 r LOWER PANEL

-.4 o /
CY CY
0 /'
-.2

-4.- -.
0-

Ch .04r
-.04 Ch ,
0' IO 20 30

Figure 7

EFFECTS OF FOREBODY LENGTH

M=2.0; a=20"

o +

"r UPPER PANEL

O - = +

-06 i

cy -*2k,
0- .
. . CY
-4 -

-.2 -

ch

-.04
0

0
- IO
r8, DEG
20 30
ch
.04-

10
r8,
20
DEG
I
30

Figure 8
2E

EFFECTS OF PANEL-PANEL INTERFERENCE

Mz2.0; ~ = 2 0 ”
-.4r UPPER PANEL
I

CY -.2

,.-. A.

‘6-Y--
CY
,,-.
‘1’‘

0 IO 20 30

Figure 9

CONF IGURAT I0 NS ST UDI ED

20

SIDE-LOAD CORRELATION FOR SMALL ANGLE OF ATTACK

1
b -7I "\ 8
L*
I
I
0 .2 A .6 .8 ID
THEORETICAL ACy

Figure 11

EFFECT OF WING HEIGHT ON V E R T I C A L - T A I L LOADING

M.0.8 i 8.5'

:k-
I.o

.4

.2

0
/0

I 2 0
WING

I 2 0 1 2

8
2.65
..
Figure 12
 2i

EFFECT OF WING HEIGHT ON VERTICAL-TAIL LOADING

M z 1 . 4 ; 8=5"

HIGH WING

Figure 13

EFFECT OF FUSELAGE STRAKES ON VERTICAL-TAIL LOAD

M = 0.60

ST R AKES

STRAKES
ON

STR A KES
OFF

I I I I I I
0 5 10 15 20 25
a, DEG
2.66
Figure 14
EFFECT OF AFTERBODY SHAPE ON VERTICAL-TAIL LOAD
M=2.0

I I I I
0' 5 IO 15 20
a,DEG

Figure 15

EFFECT OF WING HEIGHT ON SIDE FORCE DUE TO

VERTICAL TAIL M=2.9

V f--V SINB

01

L I 1 I I
0 5 IO 15 20
a, DEG

Figure 16
 23

EFFECT OF TAIL ROLL CONTROL

ON VERTICAL-TAIL LOAD
a=oo
n

-.3I I I I I I I
.4 .6 .a I.o
MACH NUMBER

Figure 1 7

EFFECT OF WING ON TAILLOADS

Mz2.46; /3=OO; a=6O

'.

Figure 18
EFFECT OF WING ON TAIL LOADS
Mz2.46; p=Oo; Q= 20'
/ 1.5
/
/
/
/
/

/
/ / ' t /
/ / z

\
\
\
\
-I
-
I I

Figure 19
 LOADS DUE 'Po CON'IEOLS AT TRANSONIC

Low SUPERSONIC SPEEDS

By F, E. West, Jr., and K. R. Czarnecki

Langley Aeronautical Laboratory

Some r e s u l t s of recent experimental investigations a t supersonic

and transonic speeds are presented t o show t h e present status i n the
estimation of load d i s t r i b u t i o n s on controls and adjacent wing surfaces
r e s u l t i n g frm t h e d e f l e c t i o n of f l a p controls and s p o i l e r controls.
The results indicate t h a t the development of methods f o r p r e d i c t i n g loads
associated w i t h controls has not kept pace w i t h the a c q u i s i t i o n of experi-
mental data. A t low supersonic speeds sweeping t h e hinge l i n e induces
strong three-dimensional-flow c h a r a c t e r i s t i c s which cannot be t r e a t e d by
t h e simplified methods previously developed f o r controls without sweep.
A t transonic speeds t h e estimation of loads associated w i t h controls must
usually be dependent upon experimental i n f o m a t i m inasmuch as t h e l a t e s t
attempts t o p r e d i c t chordwise and spanwise loadings have m e t with only
limited success.

IN'IXODWTION

Two problems must be considered i n t h e estimation of loads associated

w i t h controls: t h e d i r e c t loads on t h e controls themselves, and those loads
induced on adjacent surfaces. In the last f e w years a f a i r l y l a r g e amount
of experimental information has been obtained on these types of loads.
S m e success i n t h e prediction of these loads has a l s o been r e a l i z e d a t
supersonic speeds, but a t transonic speeds success i n t h e prediction of
these loads has been much more limited.

In t h i s paper sane r e s u l t s of t h e most recent investigations a t

supersonic speeds pertaining t o f l a p s w i t h a swept hinge l i n e and t o
s p o i l e r s a r e discussed, Ho*ever, most of t h e paper p e r t a i n s t o transonic
results, since t h e transonic range i s where the loads problems of most con-
cern s t i l l appear t o e x i s t . An attempt i s made t o indicate t h e present
s t a t u s of knowledge i n t h e estimation of loads due t o controls f o r t h i s
speed range. The r e s u l t s a t supersonic speeds were obtained i n the Langley
4- by 4-foot supersonic pressure tunnel, and t h e r e s u l t s a t transonic speeds
were obtained i n t h e Langley 16-foot transonic tunnel, I n t h e t e s t s a t
2

supersonic speeds t r a n s i t i o n w a s fixed near t h e leading edge by s t r i p s of

carborundum grains, while i n t h e t e s t s a t transonic speeds t h e Reynolds
numbers were always s u f f i c i e n t l y high t o insure a turbulent boundary l a y e r
over most of t h e wing chord.

SYMBOLS

increment i n pressure c o e f f i c i e n t due t o c o n t r o l d e f l e c t i o n

or projection
r e s u l t a n t difference i n pressure c o e f f i c i e n t s between upper
and lower surfaces due t o c o n t r o l d e f l e c t i o n

normalized section normal-force loading parameter due t o

control deflection

section normal-force c o e f f i c i e n t due t o control d e f l e c t i o n

wing panel normal-force c o e f f i c i e n t due t o control d e f l e c t i o n

(based on semispan wing area extended t o fuselage center l i n e )

aspect r a t i o

f l a p hinge l i n e

span
l o c a l chord

average chord

Mach number

chordwise distance

spanwise distance

angle of a t t a c k

angle of a i l e r o n deflection

d e f l e c t o r projection

s p o i l e r projection
E 3

Ac/4 sweep angle a t q w t e r -chord l i n e .

sweep angle of f l a p hinge l i n e

Subscripts :

S spoiler

d deflector

RESULTS AND DISCUSSION

Supersonic Speeds

A t the supersonic speeds simple configurations were i n i t i a l l y

studied, such as unswept-wing and delta-wing configurations, and con-
siderable success was attained i n predicting t h e loadings, except pos-
s i b l y f o r t h e high-angle-of-attack, high-control-deflection case.
(See r e f . 1.) Now configurations a r e being investigated f o r which more
d i f f i c u l t y i s expected i n t h e prediction of loads; f o r example, a
sweptback-wing model w i t h a swept-hinge-line f l a p was recently i n v e s t i -
gated, and some of t h e r e s u l t s are presented i n figure 1.

Effects of hinge-line swee2.- Figure 1 shows t h e e f f e c t of f l a p

hinge-line sweep and wing sweep a t a Mach number of 1.61 on the chord-
wise loading due t o f l a p deflection. On t h e v e r t i c a l scale of the p l o t s
i n t h i s f i g u r e a r e shown the changes i n pressure c o e f f i c i e n t on t h e
upper and lower surfaces t h a t a r e due t o f l a p deflection. m e s e values
a r e plotted against f r a c t i o n of t h e l o c a l wing chord. The curves i n
these p l o t s represent loadings a t t h e various spanwise s t a t i o n s indicated
on t h e model sketches. The p l o t on the l e f t represents t h e unswept-
hinge-line configuration f o r which it was previously shown ( r e f . 1) t h a t
t h e f l a p load as sham behind t h e hinge l i n e i s about 0.7 of t h e value
predicted by l i n e a r theory i f t h e carryover on t h e wing ahead of t h e
hinge li.ne i s not too large. It w a s also previously indicated i n r e f e r -
ence 1 t h a t the r e s u l t s f o r t h i s unswept-hinge-line case can be used f o r
prediction of control loadings a t higher Mach numbers, perhaps up t o Mach
numbers of 3.5 or 4.0. For t h i s case t h e r e was very l i t t l e e f f e c t of
spanwise location on t h e f l a p loading or on the wing carryover load as
shown ahead of the hinge l i n e . There was no carryover load on t h e wing
upper surf ace ..
In f i g u r e 1 i n t h e pressure p l o t f o r t h e swept-hinge-line model it
can be seen that spanwise location had a considerable e f f e c t on t h e
loading, p a r t i c u l a r l y on t h e carryover load ahead of t h e hinge l i n e .
4

Both angle of a t t a c k and control deflection a r e more important f a c t o r s

f o r t h i s configuration than f o r t h e unswept-hinge-line configuration.
For t h e swept case t h e c m p l e x i t i e s of three-dimensional flow a r e of
g r e a t concern. Also t h e e f f e c t of sweep may have reduced t h e l o c a l Mach
numbers normal t o t h e f l a p hinge l i n e s u f f i c i e n t l y t o cause c m p l i c a t i o n s
of mixed subsonic and supersonic flows such as a r e t y p i c a l i n t h e t r a n -
sonic range. A t present no r e l i a b l e general procedure f o r estimating
f l a p o r carryover loads f o r t h e swept-hinge-line case is available. There
a r e indications t h a t a solution a t higher Mach numbers w i l l be simpler.
For example, it was shown i n a paper on loads a t hypersonic speeds ( r e f . 2)
t h a t t h e flow over configurations of t h i s sweep may generally be regarc-.d
as two-dimensional.

Prediction of s p o i l e r loads on three-dimensional wings.- In addition

t o the t h e o r e t i c a l approach t o t h e problem of loads due t o controls,
another approach which is sometimes useful i s t o apply data f o r simple
configurations t o more ccanplex configurations. For supersonic speeds
pressure data a r e available on a f l a t p l a t e with an unswept s p o i l e r
( r e f . 3 ) and a l s o on a three-dimensional wing with a similar unswept
s p o i l e r ( r e f . 4). A preliminary evaluation of t h i s approach w a s made,
and the r e s u l t s a r e shown i n f i g u r e 2.

The top p a r t of f i g u r e 2 shows t h e wing and i t s s p o i l e r , which had

a height of 5 percent of t h e mean aerodynamic chord. The lower p a r t of
f i g u r e 2 shows ACp, which represents the change in loading on t h e upper
surface due t o a s p o i l e r deflection, p l o t t e d against chordwise distance
f r m t h e s p o i l e r i n terms of s p o i l e r height. The lower-surface pressures
were not affected by s p o i l e r deflection. The symbols represent the wing
d a t a f o r angles of a t t a c k of '0 and 12O which were obtained a t t h e span-
wise s t a t i o n indicated in t h e sketch of t h e wing. Spanwise location had
no e f f e c t s on these data, except near t h e wing t i p . The s o l i d l i n e s
represent t h e f l a t - p l a t e data which, i n t h i s case, a r e two-dimensional
and which were obtained a t t h e same Mach number as t h e wing data. For
t h e 12O angle-of-attack case, however, it was necessary t o adjust t h e
f l a t - p l a t e d a t a by a method developed by Lord and Czarnecki i n reference 4.
In simple terms, t h i s method adjusts the data by taking i n t o consideration
t h e changes i n Mach number on t h e basic wing which a r e due t o angle of
attack..

The agreement between t h e wing data and t h e adjusted f l a t - p l a t e data

i s generally very good. In other words, it appears t h a t r e s u l t s f o r a
f l a t p l a t e can be used t o p r e d i c t t h e load, including t h e e f f e c t s of
angle of attack, on a wing where t h e flow is approximately two-dimensional.
It i s a l s o believed t h a t there may be a p o s s i b i l i t y of obtaining agrement
i n cases where t h e flow over t h e wing tends t o be more three-dimensional,
although s u f f i c i e n t data t o t e s t t h i s belief a r e lacking. For any case,
however, t h e flow over t h e wing must not be s t a l l e d , and there m u s t be no
sharp changes in$he pres t h e wing without t h e spoiler. e-
 5

Transonic Speeds

Estimation of chordwise 1oadinm.- For transonic speeds t h e problems

of predicting loads associated with controls a r e generally more d i f f i c u l t
than f o r supersonic speeds. One of t h e most d i f f i c u l t problems is t h e
prediction of t h e chordwise loads due t o control deflection. Sme idea
of the present s t a t u s of t h i s problem f o r a t y p i c a l swept-wing configura-
t i o n is shown i n figure 3 .

I n the top p a r t of f i g u r e 3 is shown a swept-wing configuration which

has an inboard flap. In t h e p l o t s a t the bottaan of t h e figure, Q,R
represents t h e t o t a l change i n wing and f l a p loading due t o f l a p deflec-
t i o n and i s shown on t h e v e r t i c a l scales. These coefficients a r e p l o t t e d
against f r a c t i o n of t h e l o c a l wing chord a s t h e abscissa. The s o l i d l i n e s
i n these p l o t s represent experimental pressure data which were obtained
a t a Mach number of 0.98 f o r an angle of a t t a c k of '0 and a f l a p deflec-
t i o n of - 15'. The dashed l i n e represents calculations f r m a l i f t i n g -
surface theory f o r a Mach number of 1.0. The theory w a s very recently
developed f o r t h e flap-deflected case by Keith C. Harder and E. B. Xlunker
of t h e Langley Theoretical Mechanics Division, and these unpublished
, r e s u l t s a r e based on t h e slender-wing theory by Jones ( r e f . 5 ) as ex%ended
t o the case w i t h trailing-edge sweep by Mirels ( r e f . 6).

This theory predicts t h e chordwise loading f a i r l y well a t t h e inboard

s t a t i o n ; however, at t h e outboard s t a t i o n the t h e o r e t i c a l l y determined
loading is located forward of most of t h e experimentally determined
loading. Perhaps t h e main reason f o r t h i s difference i n load d i s t r i b u t i o n
i s t h a t i n t h e t h e o r e t i c a l case t h e shocks and Mach l i n e s extend normal
t o the f r e e stream, whereas i n t h e experimental case it i s known that t h e
shocks and Mach l i n e s a r e swept. Better agreement could probably be
obtained f o r a configuration having a lower aspect r a t i o ; however, since
t h e theory is very new, no attempt has been made t o define i t s range of
application.

Other t h e o r e t i c a l approaches t o the problem of chordwise loading

a t transonic speeds such as t h e use of t h e hodograph technique a r e a l s o
being considered. In t h e meantime, some experimental loads data a l s o
have been obtained f o r a f a i r l y large range of thin-wing (thickness-to-
chord r a t i o s of 0.03 t o 0.06) configurations. These c o n f i g u r a t i m s a r e
indicated i n f i g u r e 4.

New experimental loads data.- The upper p a r t of f i g u r e 4. shows con-

figurations having f l a p controls. The configurations shown i n t h e bottan
p a r t of t h e f i g u r e have s p o i l e r controls. The sketch on t h e lower right
shows an enlarged cross section of t h e spoiler-slot-deflector control.
For some of these configurations only force data were obtained, including
t h e forces and moments on t h e controls; however, f o r most of t h e configura-
t i o n s pressures were obtained over both t h e wings and controls. All of
6

these configurations were investigated a t angles of a t t a c k up t o about

2 3 O with generally a f a i r l y large range of control deflection. The
maximum Mach number w a s usually 1.03. The r e s u l t s f o r t h e configura-
t i o n a t t h e upper l e f t are t h e only r e s u l t s t h a t have so far been
reported (see r e f . 7).

An attempt has been made t o use the data f o r these various config-
urations t o develop empirical methods f o r estimating chordwise load dis-
t r i b u t i o n due t o control deflection; however, l i t t l e success was r e a l i z e d
i n deriving any simple correlations f o r even t h e loading Over flaps. Thus,
for t h e chordwise pressures or loadings, experience indicates t h a t , i n
general, recourse must be made t o data f o r configurations t h a t approximate
t h e configuration being designed.

Fortunately, f o r transonic speeds the prediction of spanwise load

d i s t r i b u t i o n s due t o control deflection i s not always as d i f f i c u l t as the
prediction of chordwise loadings. Studies indicate t h a t t h e shape of
these span-load d i s t r i b u t i o n s is e s s e n t i a l l y unaffected by Mach number
i n t h e transonic range f o r s m e configurations. This lack of Mach number
e f f e c t suggests t h e p o s s i b i l i t y of using subsonic theory f o r prediction.
Figures 5 t o 7 show r e s u l t s of a study made of t h i s p o s s i b i l i t y and a l s o
of the e f f e c t s of Mach number f o r some o f . t h e configurations shown i n
f i g u r e 4. Most of these r e s u l t s axe f o r s p o i l e r controls since such
controls a r e of considerable i n t e r e s t , and fewer loads data e x i s t f o r
s p o i l e r controls than f o r f l a p controls.

Flap and s p o i l e r a t wing t r a i l i n g edge.- Figure 5 which pertains t o

spanwise load d i s t r i b u t i o n indicates t h e span loadings f o r an inboard
f l a p configuration and an inboard s p o i l e r configuration. The load d i s -
t r i b u t i o n s shown i n t h e top p a r t of t h e f i g u r e a r e due t o t h e deflection
of t h e inboard f l a p , and t h e load d i s t r i b u t i o n s shown i n the b o t t m p a r t
of t h e f i g u r e a r e due t o t h e deflection of t h e inboard trailing-edge
s p o i l e r . Both of these controls were t e s t e d on t h e same swept-wing-body
canbination. On t h e v e r t i c a l s c a l e i s shown the weighted section normal-
force c o e f f i c i e n t due t o control deflection divided by t h e wing-panel
normal-force c o e f f i c i e n t due t o control deflection, The c o e f f i c i e n t s a r e
p l o t t e d against f r a c t i o n of wing semispan. The curves show t h e incremental
load d i s t r i b u t i o n over the wing and control, but because the curves have
been normalized t h e actual magnitude of t h e load, of course, i s not indi-
cated. The experimentally determined data a r e indicated by the symbols
and a r e presented f o r an angle of attack of 4'.

For both configurations these symbols show that increasing the Mach
number f r m 0.80 t o 0.98 had very l i t t l e e f f e c t on t h e incremental load
d i s t r i b u t i o n . These results a r e representative of those found t o e x i s t
f o r angles of a t t a c k from about 0' t o 6 O . Because t h e inboard end of
t h e f l a p was a c t u a l l y slightly outboard of t h e fuselage, the loadings
f o r t h e f l a p configurations show a large decrease i n t h i s region.
4 L

I T '5
 7

The s o l i d l i n e s i n figure 5 represent calculations f o r a Mach num-

ber of 0.80 which a r e based on a subsonic theory as presented by DeYoung
( r e f . 8) i n which t h e inboard f l a p and t h e inboard trailing-edge s p o i l e r
were d i f f e r e n t i a l l y deflected. This theory w a s devised t o apply t o
f l a p configurations as long as flow separation does not occur on t h e
wing. Since no theory i s available f o r t h e spoiler case, calculations
f o r a f l a p a t an azbitrary d e f l e c t i o n and w i t h t h e same span as t h e
trailing-edge s p o i l e r a r e compared with t h e s p o i l e r results. For t h e
f l a p configuration t h e calculations show very good agreement with the
experimental data; however, for the s p o i l e r configuration the agreement,
as would be expected, i s not as good.

For outboard controls l a r g e r e f f e c t s than those shown i n f i g u r e 5

can probably be expected. It has been previously shown (ref. 9) t h a t
Mach nmber has f a i r l y l a r g e e f f e c t s on t h e l a t e r a l p o s i t i o n of the
center of load due t o t h e d e f l e c t i o n of outboard f l a p s ; thus, it appears
t h a t subsonic theory cannot generally be extrapolated t o as high Mach
numbers f o r the outboard-control case as f o r t h e inboard-control case.

Spoiler controls ahead of wing t r a i l i n g edge.- Figure 6 shows span-

wise load d i s t r i b u t i o n f o r a wing w i t h s p o i l e r controls located ahead of
t h e t r a i l i n g edge. These controls were more highly swept than t h e s p o i l e r
control shown i n f i g u r e 5 and extend over more of t h e wing semispan. The
d i s t r i b u t i o n s shown i n t h e t o p portion of f i g u r e 6 are due t o the deflec-
t i o n of a f l a p spoiler. The d i s t r i b u t i o n s shown i n t h e lower p a r t of
figure 6 are due t o t h e d e f l e c t i o n of a spoiler-slot-deflector on the
same model. This type of control has been of p a r t i c u l a r i n t e r e s t because
of i t s good effectiveness a t high angles of a t t a c k .

The r e s u l t s of experimental investigations represented by t h e symbols

i n f i g u r e 6 show a somewhat l a r g e r e f f e c t of Mach nmber than i s shown i n
f i g u r e 5 . An increase i n Mach number f r a p n 0.80 t o 1.00 caused an outboard
s h i f t i n the center of additional load f o r both configurations. These
results a r e t y p i c a l of r e s u l t s found t o e x i s t over a range of angle of
a t t a c k f r m about Oo t o 60. The experimental span-load d i s t r i b u t i o n s f o r
t h e spoiler-slot-def l e c t o r configuration a r e considerably d i f f e r e n t fran
those f o r t h e s p o i l e r configuration. These differences a r e mainly due
t o the influence of t h e d e f l e c t o r on the flow. The s o l i d l i n e s shown
here represent subsonic calculations similar t o those shown i n figure 5
except these calculations are t h e average of curves frcan references 8 and
10 since t h e controls were r e s t r i c t e d t o one wing panel. It i s seen t h a t
t h e agreement between the calculations and experiment i n figure 6 i s
about the sane as that shown f o r t h e s p o i l e r configuration i n figure 5.

Loads a t a hi& anal e of attack.- Thus f a r i n t h e discussion of

spanwise load d i s t r i b u t i o n s a t transonic speeds, the high-angle-of -attack
case has not been considered. In f i g u r e 7, therefore, are shown distri-
butions a t a high angle of a t t a c k f o r t h e spoiler-slot-deflector
t ** 0
**
.
** 0.
- - -00 0. 0 1

configuration shown i n f i g u r e 6. Subsonic calculations f o r a Mach num-

ber of 0.60 a r e indicated by t h e s o l i d line. The experimental data a r e
indicated by t h e dashed l i n e s .

The r e s u l t s f o r t h i s p a r t i c u l a r configuration indicate t h a t Mach.

number has a large e f f e c t and t h a t theory no longer p r e d i c t s t h e shape
of the loadings. These e f f e c t s a r e mainly the r e s u l t of flow separation
i n the region of t h e wing t i p ; therefore, as i s well known, when separated
flow has a l a r g e e f f e c t on loading then recourse must be made t o experi-
mental data f o r loads information.

Loads on control surfaces.- Now t h a t the discussion i n t h i s paper

on t h e recent overall wing-loads information has been concluded, t h e
control loads on the f l a p s p o i l e r and the spoiler-slot-deflector a r e
considered. Scene of these control loads a r e shown i n f i g u r e 8.

I n t h e lower p a r t of f i g u r e 8 i s presented a canparison of the span-

wise load d i s t r i b u t i o n f o r t h e s p o i l e r and deflector ccanponents of these
controls. On t h e v e r t i c a l s c a l e is p l o t t e d the weighted control section
normal-force coeffiqient divided by t h e t o t a l control normal-force coef-
f i c i e n t . These coefficients are plotted against f r a c t i o n of t h e wing
semispan. The span-load d i s t r i b u t i o n s a r e f o r an angle of a t t a c k of 0'
and a Mach number of 1.00. Other unpublished data f o r t h e same config-
urations show very l i t t l e e f f e c t of Mach number f o r Mach numbers f r m
0.60 t o 1.03. The r e s u l t s show t h a t the type of control component has
very l i t t l e e f f e c t on the d i s t r i b u t i o n s ; therefore, these d i s t r i b u t i o n s
appear t o be primasily affected by t h e wing geometry and t h e control
location.

I n the upper right p a r t of f i g u r e 8 i s shown t h e v a r i a t i o n of t o t a l

control normal-force c o e f f i c i e n t w i t h angle of a t t a c k f o r the s p o i l e r
and deflector components of t h e spoiler-slot-deflector. The s o l i d and
dashed l i n e s indicate t h e e f f e c t of increasing Mach number from 0.60 t o
1.00. For both the s p o i l e r and deflector cmponents t h i s Mach number
e f f e c t i s not very large over the range of angle of attack. !be normal-
f o r c e c o e f f i c i e n t s f o r t h e s p o i l e r portion of the control decrease i n
magnitude w i t h increasing angle of attack, whereas those f o r t h e deflec-
t o r are not affected. The data i n the present case and other available
d a t a (ref. 11) indicate t h a t f o r large control deflections t e s t s need be
made only a t low angles of a t t a c k i n order t o e s t a b l i s h t h e maximum con-
t r o l loads.

CONCLUDING €ENARKS

The r e s u l t s of t h i s study show t h a t success i n predicting loads

associated with controls has not kept pace with t h e acquisition of
 9

experimental data, A t supersonic speeds d a t a f o r loads a r e now a v a i l a b l e

f o r s w e p t - f l q configurations, but methods of p r e d i c t i o n f o r these cases
must s t i l l be develaped. A t transonic speeds more experimental d a t a a r e
available, but only limited success i n c o r r e l a t i n g experimental d a t a w i t h
theory has been realized.
10

REFERENCES

1. Czarnecki, K. R., axid Lord, Douglas R.: Simplified Procedures f o r

Estimating Flap-Control Loads a t Supersonic Speeds. NACA
RM L55E12, 1955.

2. Henderson, Arthur, Jr., and Bertram, Mitchel H.: Factors Affecting

Loads a t m e r s o n i c Speeds. (Prospective NACA paper. )

3. Lord, Douglas R., and Czarnecki, K. R.: Aerodynamic Loadings

Associated With Swept and Unswept Spoilers on a F l a t Plate a t
Mach Numbers of 1.61 and 2.01. NACA RM L55L12, 1936.

4. Lord, Douglas R., and Czarnecki, K. R.: Pressure Distributions and

Aerodynamic Characteristics of Several Spoiler-Type Controls on a
Trapezoidal Wing a t Mach Numbers of 1.61 and 2.01. NACA
RM 256322, 19%.

5. Jones, Robert T.: Properties of Low-Aspect-Ratio Pointed Wings a t

Speeds Below and Above t h e Speed of Sound. NACA Rep. 835, 1946.
(Supersedes NACA TN 1032.)

6. Mirels, Harold: Aerodynamics of Slender Wings and Wing-Body Combina-

t i o n s Having Swept Trailing Edges. NACA TN 3105, 1954.

7. Hieser, Gerald: Transonic Investigation of t h e Effectiveness and

Loading Characteristics of a Flap-Type Aileron With and Without
Paddle Balances on an Unswept-Wing-Fuselage Model. NACA RM ~ 5 6 ~ 0 2 ,
19569

8. DeYoung, John: Theoretical Antisymmetric Span Loading f o r Wings of

Arbitrary Plan Form a t Subsonic Speeds. NACA Rep. 1056, 1951.
(Supersedes NACA TN 2140.)

9. Hammond, Alexander D., and West, F. E., Jr. : Loads Due t o Flaps and
Spoilers on Sweptback Wings a t Subsonic and Transonic Speeds. NACA
fiM L53D29a, 1953.
10. DeYoung, John: Theoretical Symmetric Span Loading Due t o Flap Deflec-

-
t i o n f o r Wings of Arbitrary Plan Form a t Subsonic Speeds. NACA
Rep. 1071, 1952. (Supersedes NACA TN 2278.)

11. Hanmond, Alexander D.: Low-Speed Pressure Distribution Investigation

of a Spoiler and a Spoiler-Slot-Deflector on a 30° Sweptback Wing-
Fuselage Model Having an Aspect Ratio of 3, a Taper Ratio of 0.5,
and NACA 65A004 A i r f o i l Section. NACA RM L55129, 1956.

! 179
4E

EFFECT OF HINGE-LINE SWEEP AT M a = 1.61

(2 -6O i 6a=20° ;A.3.1

Ac/4 =40°

-.4- .83
UPPER SURRICE
0

.4 -
ACP
.8- .83
LOWERSURFACE LWER SURFACE
I
0 .5 1.o
x/c

Figure 1

CHORDWISE LOADING D U E TO SPOILER AT M,=1.61

ORIFICE STATION SPOILE

A

SPOl LERa a =oo - WING

0
___-
A -A

-
FLAT PLATE a-12'

DOWN-
-UPSTREAM STREAM.,
-L
'8-12 -8 -4 0 4
-12 -8 -4 8O 4 8
CHORDWISE DISTANCE ,SPOILER HEIGHTS

Figure 2
la.>
r-'
 CHORDWISE LOADING DUE TO FLAP
&=LO; O ! = O O ; 6a=-15"

-Y
b/2 EXP. (M,=0.98)
0.75 ----- THEORY
0.25

ACP,R o p - (
-. 8
-1.6
--b/2
Y
- 0.25

::
I'
py -= 0.75

0 .5 I .o 0 .5 1.0
-X
C
-XC

RECENT TRANSONIC LOAD STUDIES

FLAPS

SPOl LERS

Figure 4
181
 SPANWISE INCREMENTAL LOAD DISTRIBUTION
FLAP AND TRAILING-EDGE SPOILER ;a = 4 O

0 0.80
0 .98
- CALC.

__.
8s =-O.IOC

Figure 5

SPANW ISE INCREMENTAL LOAD DlSTRl BUTlON

SPOILER CONTROLS ; a=4O

2 - 0 Ma=O.BO
Ma=1.00
ACnC
I-
Ac N ‘AV

ACn c
A CN CAV

0 .5 I .o
Y
-
bR
182
Figure 6
 SPANWISE INCREMENTAL LOAD DISTRIBUTION
SPOILER-SLOT-DEFLECTOR
A = 4.0
4994'45" dP
NACA 6 5 A 0 0 6 8 s = - 0.078C
8d =-O.O55C

--- a = I 5 O
Ac" C CALC.
AcN cAV

-
Y
b 12

Figure 7

SPOILER AND DEFLECTOR LOADING

Ma

0 -5 1.0
-
Y
b/2
283 ...
Figure 8
 MpmIMENTAL RESULTS ON WING WADS DUE TO BLASTS

Eiy Harold B. Pierce and Donald R. McFarland

Langley Aeronautical Laboratory

Recent data obtained i n t h e investigation of t h e loads caused by

blast-induced gusts a r e presented. These data indicate t h a t , when t h e
gusts cause angle-of-attack changes t o a point well beyond t h e s t a l l
angle, a load peak t r a v e l s along t h e chord which i s caused by a vortex
formed by t h e d i f f r a c t i o n of t h e b l a s t wave around the a i r f o i l . When
t h e angle caw-ed by t h e gust i s below the s t a l l angle of the a i r f o i l ,
it appears t h a t the loading may be calculated by existing unsteady-lift
theory f o r a wide range of t h e b l a s t position w i t h respect t o the wing.

INTRODUI:TION

This paper i s concerned primarily w i t h same recent information on

t h e traveling-load peak which w a s found t o occur when an intense b l a s t -
induced gust increases t h e angle of attack of the wing t o a point well
above the steady-flow s t a l l angle ( r e f s . 1 and 2 ) . In addition, some
d a t a w i l l be presented f o r t h e conditions where weak b l a s t waves s t r i k e
f r m directions normal t o the wing surface and nearly p a r a l l e l t o . t h e wing
span, producing angle-of-attack changes t o angles less than the s t a l l angle.

SYMBOLS

r e s u l t a n t pressure, measured f'rm conditions existing p r i o r

t o blast-wave a r r i v a l , lb/sq in.

dynamic pressure, lb/sq in.

airstream movement, or aerodynamic time, chords

angle of attack, deg

. .. . ....
'
.
.1
--
t .a
- ..
. - .
,
2

DISCUSSION

Figure 1 i s presented as a review of the information previously

available on the traveling-load peak. Shown are t h r e e s e t s of load d i s -
t r i b u t i o n s p l o t t e d as load c o e f f i c i e n t %/q against percent wing
chord. These d i s t r i b u t i o n s a r e f o r three successive times after b l a s t -
wave a r r i v a l as represented by the values of 0.33, 1.24, and 2.31 chords
of t r a v e l of t h e model. The dashed l i n e s with t h e points represent t h e
m o d e l f l i g h t data, whereas the s o l i d lines w e r e calculated by simple
potential-flow theory ignoring s t a l l and unsteady-lift e f f e c t s . Two
unusual features m e apparent i n the model f l i g h t data. The f i r s t i s
t h a t a loading peak w a s found t h a t traveled rearward as time passed, and
the second i s t h a t this peak showed a larger load than t h e potential-
flow calculations which would normally be considered an upper l i m i t t o
t h e loading.

The traveling-load peak w a s believed t o r e s u l t f r o m a vortex t h a t

had been formed when the b l a s t wave d i f f r a c t e d around the leading edge
of t h e a i r f o i l and subsequently had been pushed rearward over t h e upper
surface by the ste.ady airflow over t h e wing. However, no d i r e c t evidence
w a s available t o substantiate this premise. Accordingly, an experimental
investigation w a s undertaken i n the laboratory t o visualize the t r a v e l i n g
vortices by me-ms of schlieren techniques and t o determine how t h e vortices
would move along t h e chord. The manner i n which t h i s w a s accmplished i s
shown schematically i n f i g u r e 2. A t t h e right, an a i r f o i l having one-
twentieth the chord of t h e wing of t h e model used i n the free-flight tests
w a s placed i n t h e airstream issuing from the s l o t i n the f l a t surface.
The a i r f o i l w a s struck by t h e b l a s t wave from t h e a c t u a l explosion a t t h e
l e f t . The airstream velocity and the gust velocity induced by the b l a s t
wave were set a t t h e values obtained f o r the free-flying model, b u t t h e
duration of t h e b l a s t wave w a s one-twentieth of that t o which t h e free-
f l y i n g model w a s subjected i n order t o account f o r t h e reduction i n the
chord length of t h e a i r f o i l .

A s e r i e s of t e s t s w a s made w i t h t h e schlieren appazatus adjusted t o

take photographs of t h e a i r f o i l and flow a t various time i n t e r v a l s after
the b l a s t wave struck. A sequence of three photographs, w i t h t h e air-
f o i l oriented t h e same as shown i n figure 2, i s given i n f i g u r e 3 . The
p i c t u r e a t t h e l e f t , which w a s taken shortly a f t e r the blast-wave f r o n t
passed, shows t h a t vortices have actually been formed by t h e d i f f r a c t i o n
of t h e b l a s t wave around t h e a i r f o i l and t h a t the one formed a t t h e
leading edge has moved s l i g h t l y back along t h e chord. The second and
thirci pictures, taken a t l a t e r times, show t h a t t h i s vortex has moved
s t i l l farther back along t h e chord and successively become l e s s d i s t i n c t
and more spread out. The reduction i n d i s t i n c t n e s s i s an indication of
l o s s of i n t e n s i t y and conforms t o the results from t h e f l i g h t tests shown
i n figure 1 where t h e loading peak reduces and spreads out as it moves .
I

along t h e chord. "'P* 3E;

 3

In reference 1, it was found t h a t t h e load- peak traveled along t h e

chord a t a velocity slower than t h a t of t h e airstream. This i s i l l u s t r a t e d
i n figure 1 by the f a c t that t h e load peak i s s t i l l on t h e a i r f o i l i n t h e
t h i r d load d i s t r i b u t i o n shown although t h e airstream has moved more than
two chords since t h e blast wave struck. The schlieren photographs of
f i g u r e 3 a l s o show t h e same thing f o r the vortex. For example, i n t h e
middle photograph i n figure 3 , t h e vortex which was formed a t the
leading edge i s seen t o be only about a t h i r d of t h e way along t h e chord
although the airstream has moved more than 0.9 of a chord. Note t h a t
t h e vortex formed a t t h e t r a i l i n g edge appears t o move a t airstream
velocity, for it i s nearly 1 chord away from t h e a i r f o i l . Th movement
of t h e load peak and the vortex are canpared in figure 4. d e positions
of each along t h e i r respective a i r f o i l s are shown i n terms of percent
chord p l o t t e d against the chord-length movement of t h e airstream across
t h e two a i r f o i l s . The c i r c l e s represent t h e load peak and t h e squares,
t h e vortex. It i s obvious that, despite t h e twenty-to-one difference i n
scale, t h e movement of t h e load peak and t h e vortex is t h e same function
of chord-length t r a v e l of t h e airstream. The average velocity of t r a v e l
of t h e load peak and vortex appears t o be about one-third stream v e l o c i t y
since, i n t h e time they take t o pass over t h e a i r f o i l , the airstream has
moved nearly 3 chords.

Recent experimental r e s u l t s a r e now considered from another phase

of t h e investigation, t h a t of determining t h e loadings when a weak b l a s t -
induced gust s t r i k e s a wing and causes an angle change less than t h e
s t a l l angle. E a r l i e r unpublished r e s u l t s have inJicated t h a t when a
w e a k blast wave strikes normal t o the wing, t h e loading can be calculated
by using e x i s t i n g unsteady-lift theory. Questions have arisen as t o
whether t h i s procedure would apply i f t h e b l a s t wave struck fran a direc-
t i o n well t o t h e side, say a t an angle of 20° t o 300 t o t h e wing surface.
Figure 5 presents the load d i s t r i b u t i o n along the wing chord from one
t e s t f o r such a blast direction together w i t h the r e s u l t s previously men-
tioned f o r t h e b l a s t wave s t r i k i n g normal t o t h e wing. Both represent
t h e load d i s t r i b u t i o n existing a t t h e same very short time a f t e r t h e blast;
wwes struck. Again, the data are plotted as load coefficient @ / q as
a function of percent of wing chord. The c i r c l e s a r e t h e experimental
points associated w i t h the orientation a t t h e r i g h t where the blast wave
s t r i k e s normal t o t h e wing, and t h e squares are t h e results obtained when
t h e model rolled as shown so that t h e b l a s t wave struck a t an angle of
about 28' t o the wing surface. The b l a s t wave and the induced gust
v e l o c i t i e s were the same f o r each condition. In making t h e unsteady-
l i f t calculations, t h e f u l l gust velocity imposed by the b l a s t wave w a s
used i n determining t h e d i s t r i b u t i o n shown by t h e s o l i d l i n e f o r t h e
condition where t h e b l a s t wave struck normal t o t h e wing. For t h e r o l l e d
condition, however, a simple resolution of vectors w a s used t o determine
t h e component of t h e gust velocity which acted normal t o t h e wing, and
t h e r e s u l t of that unsteady-lift calculation is p l o t t e d as the dashed
l i n e . Both load-distribution calculations a r e seen t o agree well with
4

t h e experimental data along most of t h e chord. Near t h e leading edge,

t h e c i r c l e representing the f l i g h t r e s u l t f o r t h e b l a s t normal t o t h e wing
i s somewhat lower than t h e potential-flow calculation, as often happens,
but the square representing t'rre r o l l e d condition i s w e l l above i t s corre-
sponding calculated value. The flow phenomena which caused t h i s are not
understood, and, unfortunately, the pressure c e l l a t the 12-percent-chord
s t a t i o n d i d not operate on t h i s f l i g h t so it could not be determined i f
t h e e f f e c t w a s confined only t o the leading edge. The square at t h e
28-percent-chord s t a t i o n i s a l s o high, and again it i s not known whether
it represents some carryover from t h e condition found a t the leading
edge or if i t s position r e f l e c t s instrument malfunction. It appears,
however, t h a t by a simple resolution of the gust velocity vectors, t h e
loading over t h e major portion of t h e chord can be calculated over a
wide range of b l a s t incidence angles.

STMMARY OF RESULTS

Recent data obtained i n t h e investigation of the loads produced by

blast-induced gusts have shown t h a t :

1. When t h e angle-of-attack change caused by t h e gust is much

greater than t h e s t a l l angle of t h e a i r f o i l , the load peak traveling
along t h e chord i s caused by a vortex formed by t h e d i f f r a c t i o n of t h e
b l a s t wave around t h e leading edge and i t s movement i s a function of
aerodynamic time or chord-length t r a v e l .

2. For weak b l a s t waves causing small angle-of-attack changes, it

appears t h a t t h e chord loading may be calculated by existing unsteady-
1i.f-ttheory f o r a wide range of t h e b l a s t position w i t h respect t o t h e
wing by making a simple resolution of the gust velocity vector t o a
plane normal t o the wing.

REFERENCES

1. Pierce, Harold B., and Reisert, Thomas D.: I n i t i a l Experimental

Investigation of t h e Aerodynamic Load on the Wing of a Model Caused
by a Blast-Induced Gust That Increases the Angle of Attack Into t h e
S t a l l Region. NACA RM L35H22b, 1973.

2. Pierce, Harold B., and Spahl, Raymond J.: Experimental Investigation

To Determine the Loads on a Horizontal T a i l of a Model Caused by a
Blast-Induced Gust. NACA RM ~ 5 7 ~ 2 8 a1957.
,
 EXPERIMENTAL AND CALCULATED LOAD DISTRIBUTIONS

POTENTIAL-FLO W
CALCULATIONS
-4
.
0
-
*PR
4
-81- r

4
0 20 40 60 80 1000 20 40 60 80 100
%WING CHORD

Figure 1

ARRANGEMENT FOR VORTEX-FLOW STUDY

/BLAST
WAVE

FIELD OF
SCHLIEREN-

SUPPC"' -

Figure 2
 e me
e
.
)
a e
. I - - . ., 0 8
- - -e

VORTICES FROM BLAST-WAVE DIFFRACTION

IIC
~ 0 . 2 8CHORDS t=0.93 CHORDS ~31.31 CHORDS
L-57-178
Figure 3

LOAD-PEAK AND VORTEX TRAVEL ALONG CHORD

0 LOAD PEAK
0 VORTEX

AIRSTREAM
MOVEMENT, 2
T,CHORDS
i o
0
no
00
0
0
o o 0

I I I I
0 20 40 60 80 IO0
DISTANCE TRAVELED, % CHORD

Figure 4
EFFECT OF BLAST -GUST DIRECTION ON LWD DISTRIBUTION

Oh WING CHORD

Figure 5
 AIRPLANE MOTIONS AND LOADS INDUCED BY FLYING THROUGH

TIlE FLOW FIELD GENERATED BY AN AIRPLCWE AT

LOW SUPERSONIC SPEEDS

By Gareth H. Jordan, E a r l R. Keeaer,

and Stanley P. Butchart

NACA High-speed Flight Station

An exploratory f l i g h t investigation was conducted t o determine t h e

disturbances t o an airplane while flying i n formation with another air-
plane a t low supersonic speeds. The most s i g n i f i c a n t motions were
encountered as a r e s u l t of flying through t h e flow f i e l d of t h e lead
airplane. Several of these supersonic passes were made using two swept-
wing fighter-type airplanes i n order t o evaluate the gross e f f e c t s of
time t o pass through t h e flow f i e l d , l a t e r a l distance, and a l t i t u d e
within a Mach number range f r m 1.2 t o 1.3.

Significant airplane motions and v e r t i c a l - t a i l loads can be experi-

enced as a r e s u l t of close-proximity side-by-side passes at supersonic
speed. The most severe motions and v e r t i c a l - t a i l loads were experienced
during passes made a t separation distances less than 100 f e e t and a t a
time t o pass near and s l i g h t l y greater than the airplane n a t u r a l period
i n yaw. The passing airplane experienced maximum s i d e s l i p angles of
about 5.4' and maximum v e r t i c a l - t a i l loads of approximately 50 percent
of design l i m i t i n shear, bending moment, and torsion. Maximum v e r t i c a l -
t a i l loads can be determined e s s e n t i a l l y from t h e maximum airplane side-
s l i p angle and the v e r t i c a l - t a i l l i f t - c u r v e slope. Increasing the l a t e r a l
separation distance w a s shown t o decrease t h e maximum s i d e s l i p angle and,
thus, t o reduce the maximum v e r t i c a l - t a i l load.

INTRODUCTION

Military p i l o t s have reported t h a t severe disturbances a r e encountered

while f l y i n g i n formation a t supersonic speed. The implication of these
reports w a s t h a t severe loading conditions might be imposed on an airplane
by f l y i n g i n t h e flow f i e l d generated by another airplane a t supersonic
speed. A s a r e s u l t of these reports, the NACA High-speed F l i g h t S t a t i o n
has conducted an exploratory f l i g h t investigation t o determine the nature
and severity of the disturbances t h a t might be encountered. During t h e
2

investigation, it w a s found t h a t t h e most s i g n i f i c a n t airplane motions

and loads were imposed on t h e airplane as a r e s u l t of f l y i n g through t h e
flow f i e l d i n a passing maneuver.

The purpose of this paper i s t o summa;rize the results of this Lnves-

t i g a t i o n and t o point out some of the f a c t o r s believed t o be important i n
an assessment-of the o v e r a l l problem of supersonic passes. The experi-
mental data w e r e obtained a t Mach numbers f r m l . l . t o 1.3 a t a l t i t u d e s
from 20,000 t o 40,000 feet.

SYMBOLS

C chord, f t ( f i g . 1)

Fa a i l e r o n s t i c k force, l b

Fr rudder pedal force, l b

FS s t a b i l i z e r s t i c k force, lb

FY,V v e r t i c a l - t a i l s t r u c t u r a l load, lb

acceleration due t o gravity, f t / s e c 2

hP pressure a l t i t u d e , f t

*X moment of i n e r t i a about X-axis, slug-ft2

IY moment of i n e r t i a about Y - a x i s , slug-ft2

IZ moment of i n e r t i a about Z-axis, slug-ft2

1x2 product of i n e r t i a , slug-&

it incidence angle of all-movable s t a b i l i z e r , p o s i t i v e when

leading edge up, deg

Mach number as measured by airspeed head mounted on nose boom

v e r t i c a l - t a i l s t r u c t u r a l bending moment, in-lb

transverse-load factor, g units

 0.1 0 1
. 1 a

nz normal-load factor, g u n i t s

Pn period of n a t u r a l frequency of airplane i n yaw, sec

P s t a t i c pressure as measured by airspeed head mounted on nose

boom,' lb/sq f t , or r o l l i n g angulm velocity, radians/sec

approximate v a r i a t i o n of s t a t i c pressure i n flaw f i e l d from.

f r e e stream

r o l l i n g angular acceleration, radians/sec 2

free-stream dynamic pressure, lb/sq f t , or pitching angular

velocity, radians /sec

pitching angular acceleration, radians/sec2

r yawing angular velocity, radians/sec

r yawing angular acceleration, radians /s ec2

Tv v e r t i cal-t a i1 st r u c t u r a1 t o r que, in-1%

t time, sec

Y l a t e r a l separation distance between f l i g h t paths of airplanes, f t

U airplane angle of attack as measured at nose born, deg

P airplane angle of sideslip as measured a t nose boom., deg

6a t o t a l aileron deflection, deg

5, rudder deflection, deg

5. damping r a t i o '

Subscript :

max maximum

DESCRIPTION OF ADPLANE AND INSTRUMENTATION

The t e s t airplane used for t h i s investigation w a s a swept-wing

fighter-type airplane capable of supersonic speed i n l e v e l f l i g h t . A
. three-view drawing of t h e airplane giving o v e r a l l dimensions i s presented
i n figure 1. The physical c h a r a c t e r i s t i c s are given i n t a b l e I.
2-33
4

The t e s t airplane used f o r these s t u d i e s w a s instrumented t o measure

quantities p e r t i n e n t t o these investigations. S t r u c t u r a l loads on t h e
t e s t airplane were measured by s t r a i n gages located a t the root 'of' each
surface. The strain-gage s t a t i o n and torque axis f o r t h e v e r t i c a l - t a i l
loads a r e shown i n f i g u r e 1.

The airspeed head fram which static-pressure measurements were

obtained and t h e angle-of-sideslip vane were moun-bed on the nose boom
of the t e s t airplane. Pertinent d h e n s i o n s of t h i s i n s t a l l a t i o n are
shown i n f i g u r e 2.

The lead airplane f o r these formation f l i g h t s w a s furnished by the

A i r Force F l i g h t Test Center and w a s an uninstrmented airplane of t h e
same ty-pe as t h e t e s t airplane.

TESTS

The motions and loads associated with supersonic passes were i n v e s t i -

gated by t h e t e s t procedure shown i n f i g u r e 3 . The instrumented t e s t
airplane w a s flown through t h e flow f i e l d generated by t h e lead airplane
i n a side-by-side passing maneuver. These supersonic passes were made
a t various lateral separation distances and a t various passing rates
within a Mach number range of 1.1 t o 1.3 and at a l t i t u d e s fram 20,000
t o 40,000 f e e t . The speed d i f f e r e n t i a l s f o r t h e passing r a t e s i n v e s t i -
gated varied from about 5 f e e t per second t o 50 f e e t per second o r
incremental Mach numbers less than 0.05.

RESULTS AND DISCUSSION

Factors Involved

Some of t h e f a c t o r s t h a t should be considered i n a general assessment

of the supersonic pass problem a r e as follows:

Mach number
altitude

rate of passing
separation aist ance

r e l a t i v e f l i g h t paths

relative size
i n i t i a l conditions
stability
configuration

CONFIDENTIAL =
The dynamic pressure, which has a d i r e c t bearing on t h e loads, i s deter-
mined by Mach number and a l t i t u d e . Mach number and a l t i t u d e also deter-
mine the physical c h a r a c t e r i s t i c s of the flow f i e l d and the strength of
t h e shocks i n the flow f i e l d f o r a given configuration. The e f f e c t of
r a t e of passing and l a t e r a l separation distance on t h e response of t h e
airplane t o t h e flow-field disturbance i s discussed i n t h e presentation
of the experimental r e s u l t s t o follow. Only t h e side-by-side pass has
been investigated; however, many variations i n r e l a t i v e f l i g h t paths are
possible, such as overhead passes, head-on passes, and curved f l i g h t
paths. No experimental r e s u l t s a r e available t o indicate t h e e f f e c t of
r e l a t i v e s i z e of the lead - a d passing airplanes, i n i t i a l conditions before
passing, s t a b i l i t y , or configuration; however, these f a c t o r s must be
considered before a complete assessment of the problem can be m a d e . It
should be pointed out that, although t h i s investigation concerns t h e
motions and loads experienced by t h e passing airplane, a similar disturb-
'ance w a s f e l t by the other airplane as the flow f i e l d of t h e passing
airplane swept over it.

Time History of a Supersonic Pass

I n order t o have a more complete understanding of the supersonic

pass problem and of t h e r e s u l t i n g airplane motions and loads, a more ,
detailed sketch of t h e flow f i e l d and a time history of a supersonic
pass are shown i n figures 4 and 5. The f l i g h t conditions f o r t h i s super-
sonic pass axe M % 1.3, hp * 32,400 feet, y ss LOO f e e t , and a speed
d i f f e r e n t i a l of about I 2 f e e t per second. I n f i g u r e 4, t h e lead airplane
and a portion of t h e shocks generated by t h i s airplane are shown. For
simplicity, t h e shocks a r e represented by p a r a l l e l l i n e s at the Mach
angle f o r M = 1.3. The flow f i e l d i s f u r t h e r assumed t o be one fuselage
length i n a direction p a r a l l e l t o t h e f l i g h t path. The incremental pres-
sure changes within the flow f i e l d were obtained from t h e airspeed record
during t h i s supersonic pass and are shown i n t h e lower p a r t of the figure.
The three stronger shocks resulted i n abrupt pressure jumps of 15 t o
17 pounds per square foot. I n addition, t h e approximate path of a stream-
l i n e (somewhat exaggerated) through t h i s flow f i e l d i s indicated i n f i g -
ure 4 and t h e t e s t airplane i s shown j u s t entering the flow f i e l d . From
a consideration of t h e flow direction, it may be seen t h a t t h e t e s t air-
plane experiences a l o c a l l a t e r a l velocity over i t s fuselage and then
over i t s t a i l as t h e flow f i e l d i s traversed; hence, t h e t e s t airplane
experiences a yawing-moment input.

From t h e time h i s t o r i e s of t h e measured q u a n t i t i e s i n figure 5 , it

i s seen t h a t t h e t e s t airplane experienced an excursion i n s i d e s l i p i n
which a maximum s i d e s l i p angle of 4 . 7 O w a s measured. The maximum l a t e r a l
acceleration associated w i t h t h i s airplane motion w a s kO.7g. The
v e r t i c a l - t a i l loads i n figure 5(c) are mainly a function of the airplane
6

Sideslip angle; md the maximum v e r t i c a l - t a i l shear, bending moment, and

torque measured during t h i s pass were about 50 percent of design limit.

Maximum Sideslip Angle

A summary of t h e supersonic passes performed,within a Mach number

range of 1.1t o 1.3 a t a l t i t u d e s from 20,000 feet t o 40,000 f e e t i s
shown i n f i g u r e 6. The maximum s i d e s l i p angles measured during these
passes are shown as a function of time t o pass. Time t o pass i s defined
as t h e time i n seconds required f o r any p a r t of t h e t e s t airplane t o
traverse t h e distance from t h e t r a i l i n g shock t o t h e bow wave of t h e
generating airplane and was determined f o r these tests from t h e airspeed-
head static-pressure record. The time-to-pass values i n figure 6 repre-
sent speed d i f f e r e n t i a l s from about 50 f e e t p e r second a t the fast r a t e
of passing t o about 5 feet per second a t the slow rate of passing. Also
shown i n t h i s f i g u r e i s the n a t u r a l period of the airplane i n yaw f o r
the range of Mach number and a l t i t u d e of these passes.

The s o l i d symbols i n f i g u r e 6 represent t h e maximum s i d e s l i p response

measured during passes made a t separation distances less than 100 f e e t ,
t h e half-solid symbols represent passes a t distances frm LOO feet t o
300 f e e t , and t h e open symbols represent passes m a d e a t distances greater
than 400 feet. The e f f e c t of distance may be seen by comparing the
response of t h e airplane a t various distances. As would be expected,
the l a r g e s t response w a s obtained a t distances l e s s than 100 feet and
t h e response is seen t o decrease as t h e separation distance i s increased.

A noticeable e f f e c t of rate of passing may be seen by comparing t h e

response of t h e airplane a t distances less than 100 f e e t . For a time
t o pass considerably less than the n a t u r a l period of t h e airplane i n yaw,
very l i t t l e airplane motion was encountered. A t times t o pass near and
greater than t h e n a t u r a l period, s i g n i f i c a n t airplane s i d e s l i p angles
w e r e obtained. In addition, a small reduction i n s i d e s l i p response i s
evident a t the higher time-to-pass values.

The e f f e c t of Mach number on t h e airplane motions could not be

determined because of the limited range of Mach number investigated. No
e f f e c t of a l t i t u d e on the airplane motion w a s found within t h e range
investigated (20,000 t o 40,000 f e e t ) .

It i s apparent from these data t h a t t h e r e i s some dynamic response

t o t h e flow-field disturbance. Some preliminary calculations were m a d e
i n an e f f o r t t o p r e d i c t t h e maximum airplane motions. These calculations
were based on t h e simple dynamic response of a l i n e a r single-degree-of-
freedom system t o a yawing-moment input. (See, f o r example, r e f . 1.)
The r e s u l t s of these calculations are compared with t h e f l i g h t r e s u l t s
i n f i g u r e 7. The data i n t h i s figure have been corrected t o a lateral
 7
" * -
separation distance of 100 f e e t by t h e expression pmax&)0. This
correction i s based on some t h e o r e t i c a l work by G. B. Whitham (ref. 2)
i n which it is shown t h a t t h e lateral v e l o c i t y i n t h e flow f i e l d of a
body at supersonic speed varies inversely with t h e square root of t h e
lateral separation distance. The curve of the calculated response i s
based on a damping r a t i o of 0.07 which corresponds t o the f l i g h t -
determined damping r a t i o of t h e airplane a t low supersonic Mach numbers
( r e f . 3 ) . These calculations predict extremely large s i d e s l i p angles
near t h e airplane n a t u r a l period t h a t have not been realized i n f l i g h t .
It i s evident t h a t more refined calculations are necessary before r e l i -
able estimates of t h e maximum airplane motions can be predicted. Addi-
t i o n a l calculations employing an analysis based on t h e recent traveling-
gust concept are presently under ww.

Maximum Vertical-Tail Loads

A surmnary of t h e maximum v e r t i c a l - t a i l loads (measured during these

supersonic passes) i s shown i n figure 8. The maximum v e r t i c a l - t a i l loads
have been corrected t o an a l t i t u d e of 30,000 f e e t by a dynamic-pressure
correction. On t h e left-hand s i d e of t h e figure, t h e maximum v e r t i c a l =
t a i l loads measured during passes at distances less than 100 f e e t are
shown as a function of time t o pass. The v e r t i c a l - t a i l loads vary i n a
m a n n e r similar t o that previously sham f o r maximum s i d e s l i p angle, and
v e r t i c a l - t a i l loads of 9 percent of design l i m i t were experienced at a
time-to-pass slightly greater than t h e n a t u r a l period of t h e airplane
i n yaw.

A s a matter of i n t e r e s t , t h e fast r a t e of passing resulted i n negli-

gible airplane motion and, thus, i n r e l a t i v e l y small v e r t i c a l - t a i l loads.
Same evidence of s t r u c t u r a l excitation w a s v i s i b l e , however, i n the
f l i g h t records f o r t h i s pass. This excitation i s a r e s u l t of an impact
type of loading imposed on t h e v e r t i c a l t a i l during high passing rates.

On t h e right-hand s i d e of figure 8 t h e maximum v e r t i c a l - t a i l loads

from a l l of t h e supersonic passes are shown as a function of t h e maximum
s i d e s l i p angle. Also included i s the variation of t h e v e r t i c a l - t a i l
load based on t h e experimental lift-curve slope of t h e v e r t i c a l t a i l
from other experimental r e s u l t s . A comparison of t h e v e r t i c a l - t a i l loads
experienced during these passes w i t h t h e experimental lift-curve slope
indicates t h a t t h e maximum v e r t i c a l - t a i l loads can be determined from
the maximum airplane s i d e s l i p angle and the v e r t i c a l - t a i l lift-curve
slope.

A s previously mentioned t h e magnitude of t h e airplane motions was

found t o be r e l a t i v e l y unaffected by a l t i t u d e within t h e range of t h i s
investigation. It might be expected t h a t loads i n excess of design
a

l i m i t would be experienced as a r e s u l t of close-proximity passes a t an

a l t i t u d e lower than t h e range investigated because of t h e increase i n
dynamic pressure.

P i l o t ' s Comments

Formation f l y i n g a t supersonic speeds does not d i f f e r from subsonic

formation work as long as t h e wingman remains i n a position behind t h e
flow f i e l d of t h e lead airplane. If t h e wingman gets out of position,
e i t h e r by overshooting during j o i n up or by drawing abreast of the lead
plane as a r e s u l t of a turn-in mmeuver of t h e lead plane, control of
t h e airplane can become quite d i f f i c u l t .

If the wingman moves ahead slowly, t h e first encounter w i t h t h e

flow f i e l d of the lead airplane causes an abrupt yaw toward t h e lead
plane. (The f i r s t impression is t h a t you will fly r i g h t through the
middle of t h e lead airplane.) A l i t t l e d i f f i c u l t y i s experienced i n
holding v e r t i c a l position at t h i s point also, not unlike flying i n t h e
wing wake of another airplane at slower speeds. It i s possible Lo
retrim and t o hold formation position a f t e r the airplane has penetrated
some distance i n t o t h e flow f i e l d . Moving ahead a l i t t l e f a r t h e r
releases t h e yaw toward t h e lead plane and sets up a s l i g h t yaw i n t h e
opposite direction. It i s possible t o move out frm t h e lead airplane
f o r a distance of 500 t o 600 f e e t o r more and s t i l l remain i n t h i s
yawed position.

The most s t a r t l i n g experience occurs when t h e wingman slldes up

past. the lead a i r c r a f t a t a speed where t h e yaw toward the lead plane
and then away from t h e lead plane comes i n phase w i t h t h e n a t u r a l
frequency of t h e airplane i n yaw. Under t h i s condition the airplane
ends up i n a yawing o s c i l l a t i o n t h a t can be four o r f i v e times g r e a t e r
than t h a t a t t a i n a b l e when rudder alone i s used a t t h a t same speed. "he
airplane controls a r e not e f f e c t i v e i n preventing the o s c i l l a t i o n and
the airplane f e e l s out of control f o r a f e w seconds.

CONCUTDING REMARKS

Results of an investigation of airplane motions an8 loads induced

by f l y i n g through t h e flaw f i e l d of an airplane a t law supersonic speeds
have indicated t h a t s i g n i f i c a n t airplane motions and v e r t i c a l - t a i l loads
can be experienced as a r e s u l t of close-proximity side-by-side passes at
supersonic speed. The most severe motions and v e r t i c a l - t a i l loads were
experienced during passes m a d e at separation distances less than 100 f e e t
 9

and at a time t o pass near and s l i g h t l y greater than the airplane natural
period i n yaw. The passing airplane experienced maximum s i d e s l i p angles
of about 5.4' and maximum v e r t i c a l - t a i l loads 'of approximately 50 percent
of design l i m i t i n shear, bending moment, and torsion. Maximum vertical-
t a i l loads can be determined e s s e n t i a l l y from the maximum airplane side-
s l i p angle and the v e r t i c a l - t a i l lift-curve slope. Increasing t h e l a t e r a l
separation distance was shown t o decrease t h e maximum s i d e s l i p angle and,
thus, t o reduce the maximum v e r t i c a l - t a i l load.

A general assessment of t h e problem of supersonic passes requires

t h a t several additional f a c t o r s be considered -namely, e f f e c t s of Mach
number, s t a b i l i t y , configuration, r e l a t i v e size, and r e l a t i v e f l i g h t path
on the airplane motions and loads.

REFERENCES

1. T r i p l e t t , W i l l i a m C.,
Brown, Stuart C., and Smith, G. A l l a n : The
Dynamic-Response Characteristics of a 3 5 O Swept-Wing Airplane A s
Determined From Flight Measurements. M C A Rep. 1.250, 1-95?.
(Supersedes NACA RM A51G27 by T r i p l e t t and Smith and RM A52117
by T r i p l e t t and Brown.)

2. Whitham, G. B.: The Flow Pattern of a Supersonic P r o j e c t i l e .

Communications on Pure and Appl. Math., vol. V, no. 3 , Aug. 1952,
pp. 301-348.

3 . Wolowicz, Chester H.: Time-Vector Determined Lateral Derivatives of

a Swept-Wing Fighter-'Sype Airplane With Three Different V e r t i c a l
T a i l s a t Mach Numbers Between 0.70 and 1.48. NACA RM ~ 5 6 ~ 2 01956.
,
10

PHYSICAL CHARACTERISTICS OF THE TEST AIRPLANE

Wing :
....................
A i r f o i l section NACA 64(,6)AOO7
Total area (including aileron and 83.84 sq f t
................
covered by fuselage), sq f t 385.21
Span,ft .......................... 38.58
.................
Mean aerodynamic chord, f t 11.16
.......................
Root chord, f t 15.86
......................
Tipchord, f t . . 4.15
........................
Taper r a t i o . 0.262
........................
Aspect r a t i o 3.86
.............
Sweep a t 23-percent-chord l i n e , deg 45
Incidence,deg ....................... 0
Dihedral, deg........................ 0
....................
Geometric t w i s t , deg 0
Aileron:
.........
Area rearward of hinge l i n e (each), sq f t 19.32
...............
Span a t hinge l i n e (each), f t 7.81
.....
Chord rearward of hinge l i n e , percent wing chord 25
....................
Travel (each), deg +15
Leading-edge s l a t :
...................
Span, equivalent, f t u. 71
Segments ......................... 5
...
Spanwise location, inboard end, percent wing semispan 23.3
..
Spanwise location, outboard end, percent wing semispan 89.2
Ratio of s l a t chord t o wing chord ( p a r a l l e l t o
............
fuselage reference l i n e ) , percent 20
..................
Rotation, maximum, deg 15
Horizontal t a i l :
. . . . . . . . . . . . . . . . . . . . . NACA
A i r f o i l section 65AOO3.5
Total area (including 31.65 sq f t covered by fuselage),
sqft ........................... 98.86
span, f t ................ .......... 18.72
Mean aerodynamic chord, f t ................. 5.83
Root chord, ft ....................... 8.14
Tip chord, f t........................ 2.46
. . . . . . . . . . . . . . ...........
Taper r a t i o 0.30
.............. ..........
Aspect r a t i o 3.54
.............
Sweep a t 25-percent-chord l i n e , deg 43
Dihedral, deg ........................ 0
Travel, leading edge up, deg ................ 5
Travel, leading edge down, deg ............... 25
 -1 -a, - ...........

TABLE I.- Continued

PHYSICAL CHARACTERISTCCS OF TRE TEST AIRPLANE

Vertical t a i l :
A i r f o i l section .................... NACA 65A003.5
Area (excluding d o r s a l f i n and area blanketed by .
fuselage), sq f t ..................... 42.7
Area blanketed by fuselage (area between fuselage contour
l i n e and l i n e p a r a l l e l t o fuselage reference l i n e through
i n t e r s e c t i o n s of leading edge of v e r t i c a l t a i l and
fuselage contour l i n e ) . .................. 2.45
Span (unblanketed) , f t................... 7.93
.................
Mean aerodynamic chord, f t 5.90
Root chord, f t ....................... 8.28
Tip chord, f t ........................ 2.49
Taper r a t i o ......................... 0.301
Aspect r a t i o ........................ 1.49
.............
Sweep a t 25-percent-chord l i n e , deg 45
Rudder :
.............
Area, rearward of hinge l i n e , sq f t 6.3
..................
Span a t hinge l i n e , f t 3,33
Root chord, f t ...................... 2.27
Tip chord, f t ..... .................
; 1.30
Travel, d e g . . ...................... +20
Spanwise location, inboard end, percent
...................
v e r t i c a l - t a i 1 span 3.1
Spanwise location, outboard end, percent
...................
v e r t i c a l - t a i l span 44.8
............
Chord, percent v e r t i c a l - t a i l chord 28.4
Aerodynamic balance ............. Overhanging, unsealed

Fuselage :
...........
Length (afterburner nozzle closed), f t 45.64
Maximum width, f t ...................... 5.58
Maximum depth over canopy, f t ................ 6.37
..................
Side area ( t o t a l ) , sq f t 230.92
.........
Fineness r a t i o (afterburner nozzle closed) 7.86
Speed brake:
Sarface area, sq f t ..................... 14.14
Maximum deflection, deg ................... 50

P owerpl a n t :
Turbojet engine ..... One P r a t t & Whitney J57-P7 with afterburner
........
. Thrust (guarantee sea l e v e l ) , afterburner, l b 15,000
bfilitary, lb ........................ 9,220
Normal, l b .......................... 8,000
12 ..

TABLE I.- Concluded

PHYSICAL CHARACTERISTICS OF 'ME TEST AIRPLANE

Airplane weight, lb:

Basic (without fuel, oil, water, pilot) ............ 19,662
Total (full fuel, oil, water, pilot). ............ 24,800
Center-of-gravity location, percent mean aerodynamic chord:
Total weight - gear down .................. 31.80
-
Total weight gear up ................... 31.80
Moments of inertia (estimated total weight) :
rX, slug-& ........................ 11,103
Iy, slug-ft2 ........................ 59,248
IZ, slug-ft2 ........................ 67,279
In, slug-ft2 ........................ 941

Inclination of principal axis (estimated total weight):

Below reference axis at nose, deg .............. 6.8
7E

TEST AIRPLANE
I
Torque reference oxi

I 4 6 2 I

Figure 1

INSTRUMENTATION ON NOSE BOOM OF TEST AIRPLANE

Test ocplane fusela

Airspeed hea

L+a+icMif/ces
Angle-of-sideslip vane
14

DESCRIPTION OF TESTS

LEAD
AIRPLANE

SKETCH OF SIMPLIFIED FLOW FIELD

M 1.3; hp= 32,400 FEET

Figure 4
 I . 1 -a. 1 . I. 0. .. 0 *.

TIME HISTORY OF SUPERSONIC PASS; M ~ l . 3 ; hDlru32,400; YNIOO

AIRPLANE ANGULAR MOTIONS AND LOAD FACTORS
1.4

M 1.3

1.2
580

p, Ibkq ft 560

540

Right 8
UP

a r P Ideg 4

4
Right 4
UP

ny. nz, g units 2

2
0 2 4 6 8 IO 12 14 16 18
t, sec

Figure 5 ( a )
2.35
 16

TIME HISTORY OF SUPERSONIC PASS; Ma1.3; hpdE,400; ySYlO0

CONTROL POSITIONS AND FORCES; ANGULAR VELOCITIES AND ACCELERATIONS
.4
Right
UP
p, q, r , radians/sec 0

.4

Right
UP

p, 4, i, radians/sec2

8
Right
Airplane nose up

0 2 4 6 8 10 12 14 16 18
t , sec

Figure 5(b)
 I - -z -.. . . *. ** 3 .. . ** I

TIME HISTORY OF SUPERSONIC PASS; Mfil1.3; h@2,400; ~~100

VERTICAL-TAIL STRUCTURAL LOADS

L.E. I
Right

T v , in--Ib 0I

L.E.
Left I

Right

MbYv,in-lb

Left

Right

FY,@

Left
'0 2 4 6 8 10 12 14 16 18
t, sec I
18

SUMMARY OF SUPERSONIC PASSES

IO y -e 100FEET
0 y = 100-300 FEET
8 0 Y > 400FEET

0 0 .
2

0 I 2 3 4 5 6 7 8
TIME TO PASS,SEC

Figure 6

COMPARISON OF CALCULATED AND EXPERIMENTAL

MAXIMUM SIDESLIP ANGLES

Figure 7 ..
 8- 0.5 DESIGN LIMIT
------------I_-- --_ _ _ _ _ -
--------*
6- a
q30,OCd
Fy%st 0

e
0

2-

1 I I

Figure 8
8E

EFFECTS OF AIRPLANE FLEXIBILITY ON WING BENDING

STRAINS I N ROUGH AIR

By Thomas L. Coleman, Harry Press,

and C. C. Shufflebarger

Langley Aeronautical Laboratory

INTRODUCTION

Previous investigations ( r e f s . 1t o 3) have indicated t h k t wing

f l e x i b i l i t y could cause s u b s t a n t i a l amplification i n t h e wing s t r a i n s
i n rough a i r f o r such straight-wing airplanes as t h e Douglas DC-3,
Martin 2-0-2, and t h e Boeing B-29. Also, by application of power-
s p e c t r a l methods of analysis, good correlation has been obtained between
t h e measured root bending s t r a i n s and calculated results f o r these t h r e e
airplanes (ref. 3). A s a continuation of the work i n t h i s area, t h e
B-29 investigation has been extended t o cover t h e bending-strain ampli-
f i c a t i o n s a t several other wing stations. I n addition, a f l i g h t inves-
t i g a t i o n , involving t h e swept-wing Boeing B-47A airplane, has been under-
taken i n order t o assess t h e significance of t h e wing sweep on t h e
e l a s t i c response i n rough air.

I n t h e present paper, an e f f o r t i s made t o summarize some of t h e

more important r e s u l t s obtained i n recent studies. The material t o be
covered includes experimental r e s u l t s on t h e bending-strain responses
a t various s t a t i o n s f o r both t h e Boeing B-29 and the Boeing B-47A a i r -
planes. For t h e B-29 airplane investigation, comparisons a r e made
between t h e experimental results and calculations. Unfortunately, simi-
l a r r e s u l t s a r e not yet available f o r t h e B-47A investigation. Finally,
some indirect evidence obtained i n t h e B-47A investigation on t h e e f f e c t s
of spanwise variations i n turbulence on t h e s t r a i n and acceleration
responses i s described.

AIRPLANES, INSTRUMENTATION, AND TEST CONDITIONS

Figure 1 shows plan-form views of t h e two t e s t airplanes and t h e

locations of some of t h e primary instrumentation. The basic instrumenta-
t i o n of t h e two airplanes included s t r a i n gages (denoted by t h e small
+ signs) t o measure bending- and shear-strain indications on the f r o n t
and rear spars a t 5 spanwise stations. Accelerometers were i n s t a l l e d a t
a number of locations on t h e airplanes. The accelerometer measurements
pertinent t o the present paper a r e sham by t h e shaded c i r c l e s on t h e
figures and were a t the nodal points of t h e fundamental wing bending
mode of the B-29 airplane and a t the center of gravity of t h e B-47A air-
plane. I n addition t o t h e accelerometers and s t r a i n gages, t h e i n s t r u -
mentation included control position recorders and a t t i t u d e and r a t e gyros,
One s i g n i f i c a n t addition t o t h e B-47A instrumentation w a s a flow-direction
vane from which a time h i s t o r y of t h e v e r t i c a l gust velocity could be
derived by correcting f o r airplane motions.

Relative t o the present investigations, these two airplanes differed

i n a number of respects. The airplane weights w e r e about the same with
t h e t e s t weight of t h e B-29 airplane at 105,900 pounds and t h e weight of
t h e B-47A airplane a t 113,200 pounds. However, t h e B-29 airplane had a
f a r greater proportion of i t s t o t a l weight (about 80 percent) i n the
wings since almost a l l t h e f u e l was i n w i n g tanks. The B-47A c a r r i e s a l l
i t s f u e l i n t h e fuselage, and the r a t i o of t h e wing w e i g h t t o t o t a l weight
f o r t h i s airplane was 36 percent. I n addition t o t h e differences i n t h e
proportion of weight i n the wing, it should be mentioned t h a t t h e w e i g h t
d i s t r i b u t i o n of t h e B-29 wing was r e l a t i v e l y uniformly d i s t r i b u t e d over
t h e span, whereas t h a t f o r the B-47A wing was t y p i f i e d by l a r g e concen-
trated masses due t o engine nacelles. The airplanes a l s o d i f f e r e d sig-
n i f i c a n t l y i n s t i f f n e s s d i s t r i b u t i o n with t h e wing of t h e B-29 airplane
being about twice as stiff as t h a t of the B-47A airplane. A f u r t h e r d i f -
ference, .which w i l l be of some consequence i n the results t o be presented,
arises from t h e swept wing of the B-47A. Because of the swept wing, t h e
lower s t i f f n e s s , and the concentrated masses, the B-47A airplane has
rather large s t a t i c a e r o e l a s t i c e f f e c t s i n contrast t o t h e absence of
s i g n i f i c a n t a e r o e l a s t i c e f f e c t s on t h e B-29 airplane and on t h e other
airplanes studied.

The flight t e s t s f o r both airplanes were i n c l e a r a i r turbulence a t

a n a l t i t u d e of about 2,000 feet above t e r r a i n . The f l i g h t t e s t speeds
were 250 m i l e s per hour f o r t h e B-29 airplane and 478 miles per hour f o r
t h e B-47A airplane. The basic t e s t results were obtained i n f l i g h t t e s t
runs of several minutes duration i n continuous rough a i r f o r each of these
t e s t airplanes. I n addition t o the rough-air tests, slow maneuver p u l l -
ups were a l s o made i n smooth a i r a t various t e s t conditions i n order t o
determine quasi-static reference s t r a i n s .

METRODS FOR DETERMINATION OF FREQUENCY RESPONSE

I n discussing the r e s u l t s of these investigations, t h e basic quantity

t h a t w i l l be used is t h e frequency-response functions of t h e wing s t r a i n s
o r accelerations t o gust disturbances. This function describes the a i r -
plane response i n s t r a i n or i n normal acceleration, as t h e case may be,
t o u n i t sinusoidal gusts of various frequencies. The airplane response
*.

213.
 3

t o continuous rough a i r depends e s s e n t i a l l y on t h e product of t h e ampli-

tude squared of t h e frequency-response function and the power spectrum
of t h e gust velocity. Two basic approaches may be used f o r t h e experi-
mental determination of t h e frequency-response functions t o gust inputs,
and these a r e given as follows: The first method, called t h e spectizun
method, i s based upon the r e l a t i o n between the power spectrum of a random
disturbance and t h e power spectrum of the response of a l i n e a r system t o
t h e disturbance. From t h i s relation, t h e amplitude squared of the
frequency-response function i s given by

where

amplitude squared of frequency-response function

Oo(f) power spectrum of airplane response

a.1( f ) power spectrum of disturbance o r gust input

The application of t h i s method simply requires t h e measurement of t h e

response power spectrum rPo and t h e power spectrum of ,the gust input Qi.

The second method designated as the cross-spectrum method i s based

upon t h e relationship f o r l i n e a r systems between a random input disturb-
ance and t h e cross spectrum between t h e input disturbance and the system
response t o the disturbance. From t h i s relationship, the frequency-
response function i s given by the following expression:

where @i,o( f ) i s t h e cross spectrum between t h e disturbance input and

t h e airplane response. Note i n t h i s second method t h a t both the amplitude
and phase of t h e frequency-response function H ( f ) a r e obtained since
~ ~ , ~ is, ( f i )n general, complex.

I n the material t o be presented, both of these methods w i l l be

applied. It should be noted that i f t h e measurements of turbulence such
as t h a t obtained from the angle-of-attack vane adequately represent t h e gust
input, then t h e two methods should yield i d e n t i c a l r e s u l t s f o r t h e ampli-
tude of t h e frequency-response function. I f , however, complicating fac-
t o r s a r e present, such as significant variations i n turbulence across t h e
airplane span, t h e results obtained by the two methods w i l l d i f f e r (see
appendix). I n the following material, t h i s property w i l l be used i n
assessing t h e significance of spanwise variations i n the turbulence on
the t e s t r e s u l t s .

RESULTS FOR THE B-29 AIRPLANE

The measurements obtained i n t h e course of t h e B-29 f l i g h t i n v e s t i -

gation have been used t o estimate t h e frequency-response function f o r
t h e bending s t r a i n s a t t h e various s t a t i o n s by application of the spec-
trum method described i n the preceding section. I n estimating the
frequency-response function, it was necessary t o assume a shape f o r t h e
gust spectrum since the gust spectrum w a s not measured i n t h e B-29 inves-
t i g a t i o n . The gust spectrum used f o r t h i s purpose i s given by t h e f o l -
lowing expression:

where R = 2sf/V and V i s the airplane speed. This spectrum has been
found t o approximate t h e atmospheric conditions covered by flight meas-
urements of gust spectra when a value of about 1,000 i s used f o r the
scale of t h e turbulence L. The i n t e n s i t y of t h e turbulence which i s
described by the root-mean-square gust velocity (I is, however, not
known f o r t h e B-29 measurements.

I n order t o permit d i r e c t comparisons of the frequency-response

functions a t the various wing stations, the results were converted t o
"equivalent accelerations" by dividing t h e s t r a i n s a t each s t a t i o n by
t h e s t r a i n per g measured a t t h e s t a t i o n during slow pull-ups. The
results obtained f o r t h e amplitude of t h e frequency-response function
on t h i s b a s i s a r e shown i n figure 2. The ordinate represents the ampli-
tude of t h e bending-strain response f o r a u n i t sinusoidal gust velocity
a t the various frequencies. The four solid-line curves shown i n the
figure a r e t h e r e s u l t s f o r t h e four spanwise s t a t i o n s whose locations
a r e designated i n terms of t h e f r a c t i o n of t h e semispan b%. Also
shown i n figure 2, by t h e dashed-line curve, i s t h e frequency-response
function obtained from t h e masured nodal accelerations and which can
be considered as a reference f o r s t a t i c loading. Thus, t h e difference
between t h i s reference and the other curves gives an indication of t h e
e f f e c t s of wing f l e x i b i l i t y . ( F l e x i b i l i t y a l s o has an e f f e c t on t h e
nodal accelerations, but t h i s e f f e c t i s small. )

Comparison of t h e s o l i d curves of figure 2 w i t h the reference curve

of nodal- acceleration s e f f e c t s of f l e x i b i l i t y
f 23
 5

a r e associated w i t h a large peak a t the f i r s t bending mode, which i s a t

2.7 cps. The amplification i s highest f o r the 15-percent-semispan sta-
t i o n and progressively decreases w i t h outboard wing span s t a t i o n . Above
about 4 cps, t h e curves a r e somewhat disorderly, probably because of t h e
e f f e c t s of higher s t r u c t u r a l modes.

I n order t o see how w e l l t h e frequency-response functions could be

determined analytically, calculations were a l s o made. The calculated
s t r a i n frequency-response functions were converted t o equivalent accel-
erations f o r d i r e c t comparison w i t h the resul%s of f i g u r e 2 and a r e shown
i n figure 3. These r e s u l t s a r e f o r three modes or degrees of freedom,
that is, airplane v e r t i c a l motion and the f i r s t and second symmetrical
bending modes. Comparison of t h e r e s u l t s i n figures 2 and 3 shows t h a t ,
i n general, t h e character and trends of t h e experimental and calculated
frequency-response functions a r e i n f a i r l y good agreement i n regard t o
f l e x i b i l i t y e f f e c t s . Both figures indicate t h a t t h e p r i n c i p a l e f f e c t s
of f l e x i b i l i t y a r e associated w i t h t h e f i r s t bending mode and that t h e
amplification i s l a r g e s t a t t h e root s t a t i o n and progressively decreases
w i t h outboard wing span station.

It i s of i n t e r e s t t o note that the calculated and measured frequency-

response functions ( f i g s . 3 and 4) d i f f e r below about 1 cps. This d i f -
ference i s due p r i n c i p a l l y t o the omission of t h e pitching motion i n t h e
calculations. Because of t h e high gust input power a t these low frequen-
cies, it would thus appear t h a t t h e inclusion of t h e pitching mode i s
quite important i n determining t h e a c t u a l output response.

The e f f e c t which the frequency-response functions have on the over-

a l l b e n d i n g - s t r a i n amplification i s shown i n figure 4. I n t h e upper p a r t
of t h e figure, t h e ordinate i s t h e r a t i o of t h e root-mean-square s t r a i n
f o r the f l e x i b l e airplane t o the root-mean-square s t r a i n f o r
t h e reference condition Q€,REF' The reference s t r a i n s were based on t h e
nodal-point accelerations measured i n rough a i r and the s t r a i n per g as
measured a t t h e various spanwise s t a t i o n s i n slow pull-ups. This refer-
ence condition represents the s t r a i n s f o r t h e s t a t i c application of loads,
and thus t h i s r a t i o provides a measure of dynamic-strain amplification.
The c i r c l e s represent t h e measured amplification f a c t o r s , and t h e curves
represent calculated r e s u l t s . The dashed curve was obtained by consid-
ering two modes': airplane v e r t i c a l motion and t h e f i r s t symmetrical
mode i n bending. The s o l i d curve w a s obtained by using t h r e e modes:
airplane v e r t i c a l motion and the f i r s t two symmetrical bending modes.
The measured results show t h a t t h e overall s t r a i n amplification i s approxi-
mately 10 percent a t t h e root s t a t i o n and i s somewhat lower a t t h e out-
board wing s t a t i o n s . Both t h e two-mode and three-mode c'alculations a r e
i n rather good agreement w i t h t h e measured r e s u l t s except a t t h e far-
t h e s t outboard s t a t i o n .
214
 A second s e t of strain-amplification f a c t o r s is shown i n t h e lower
p a r t of figure 4. This ordinate i s t h e r a t i o of t h e s t r a i n s that would
occur with equal frequency f o r the f l e x i b l e airplane fiFmXand f o r t h e
reference airplane condition E ~ "he
. frequency l e v e l a t which t h i s
s t r a i n r a t i o was taken corresponds t o a value of s t r a i n equal t o about
twice t h e root-mean-square s t r a i n . These experimental amplifications
are about twice as large as those shown i n the t o p figure based on root-
mean-square s t r a i n s . This i s a consequence of f l e x i b i l i t y having a
greater e f f e c t on t h e number of peak s t r a i n s than on the root-mean-square
values. The calculations which use only two modes underestimate the
s t r a i n amplifications a t t h e outboard s t a t i o n s , whereas the analysis
using three modes gives a good approximation t o t h e measured results.

The foregoing results Fmply that, f o r t h i s straight-wing airplane,

calculations using only v e r t i c a l motion and t h e fundamental bending
mode are adequate f o r determining t h e e f f e c t s of f l e x i b i l i t y on the roat-
mean-square s t r a i n values but that the second bending mode must a l s o be
included when considering peak s t r a i n s .

RESULTS FOR TI-IE B-47A AIRPIANE

Frequency-Response Functions

The r e s t of t h i s paper w i l l cover test results obtained from the

B-47A airplane investigations . Figure 5 shows t h e measured frequency-
response functions f o r t h e rear-spar bending s t r a i n s f o r various wing
s t a t i o n s . Similar results were a l s o obtained f o r t h e s t r a i n s on the
f r o n t spar but are not included. The results sham are f o r both t h e
amplitude and t h e phases and were obtained by the cross-spectrum m e t h o d .
The gust input spectrum was obtained from t h e vane angle-of-attack meas-
urements corrected f o r airplane motions according t o t h e method given i n
reference 4. For present comparisons, the measured s t r a i n s at the var-
ious s t a t i o n s were converted t o equivalent accelerations by dividing by
the s t r a i n s p e r g i n pull-ups a t the same t e s t conditions. The reference
curve shown i n t h e figure was based on t h e center-of-gravity accelerations
which appear t o be r e l a t i v e l y uninfluenced by the fundamental airplane
v i b r a t i o n mode and provides a measure of t h e s t r a i n s f o r s t a t i c loading.
(The use of the center-of -gravity accelerations f o r the reference loading
i s discussed i n more detail later.) The differences between the reference
curve and t h e other curves t h u s provide a d i r e c t measure of t h e e f f e c t s of
t h e dynamic f l e x i b i l i t y on t h e l o c a l strains. A s i n the case of the B-29,
t h e e f f e c t s of f l e x i b i l i t y show up p r i n c i p a l l y as a l a r g e peak at t h e
first bending mode frequency which i s around 11 - cps. I n contrast t o t h e
2
B-29, however, t h e magnitude of t h e peak i s now smallest a t t h e root sta-
t i o n and progressively increases f o r t h e outboard s t a t i o n s . Another
 7

s i g n i f i c a n t difference between the frequency-response functions f o r the

B-47A and t h e B-29 airplanes i s the closer proximity of t h e f i r s t bending
mode t o t h e short-period mode which i s a%about 0.6 cps. A s a consequence,
t h e gust input, which decreases rapidly w i t h frequency, w i l l be r e l a t i v e l y
much higher a t t h e f i r s t bending mode i n t h i s case, and l a r g e r o v e r a l l
amplification e f f e c t s may be expected.

The phases shown i n figure 5(b) indicate a l i n e a r increase i n phase

l a g with increasing frequency, as i s t h e case f o r a simple system with a
moderate amount of damping. Above 2 cps, t h e curves appear t o be e r r a t i c .
The r e s u l t s above 2 cps a r e not, however, considered t o be r e l i a b l e .

Strain-Amplif i c a t i o n Factors

Figure 6 shows t h e v a r i a t i o n s i n the bending-strain-amplif i c a t i o n

f a c t o r s w i t h spanwise positions. Again, the r e s u l t s a r e shown based on
both root-mean-square values and s t r a i n s having an equal frequency of
occurrence. Note t h a t t h e ordinate scale i s compressed and covers a
wider range of amplification values than t h e scale used earlier f o r t h e
corresponding B-29 results.

For t h e swept-wing airplane, t h e e f f e c t s of f l e x i b i l i t y are com-

p l i c a t e d by large s t a t i c a e r o e l a s t i c or twist e f f e c t s which a c t t o reduce
t h e gust loading and the s t r a i n response. Thus, the derivation o f ampli-
f i c a t i o n f a c t o r s f o r t h i s case i s not straightforward, and s e v e r a l pro-
cedures m i g h t be used. For present purposes, three s t r a i n responses a r e
considered. These are t h e a c t u a l measured s t r a i n s , t h e numerator of t h e
ordinate, and two reference s t r a i n conditions. These reference s t r a i n s
are

(1)The s t r a i n s obtained by t h e s t a t i c application of t h e same

loads t o t h e airplane.
(2) The s t r a i n s obtained by t h e s t a t i c application of the loads
t o an e s s e n t i a l l y "rigid" airplane, t h a t is, an airplane
embodying no s t a t i c a e r o e l a s t i c e f f e c t s .

The r a t i o of t h e measured s t r a i n s t o t h e s t r a i n s i n the s t a t i c a l l y e l a s t i c

airplane y i e l d s t h e solid-line curves and provides a measure of the purely
dynamic s t r a i n amplification. (See f i g . 6.) The r a t i o of the measured
s t r a i n s t o t h e s t r a i n s f o r the airplane without s t a t i c a e r o e l a s t i c e f f e c t s
yields t h e dashed-line curves and provides a measure of the combined
e f f e c t s of t h e s t a t i c a l l e v i a t i o n and the dynamic amplification.

The procedure used i n t h e determination of the reference s t r a i n s

was based upon the use of the a c t u a l measured center-of-gravity acceler-
a t i o n s as a measure of t h e airplane loading i n rough air. Examination
of the power spectra of t h e normal accelerations had indicated t h a t t h e
8

f i r s t mode had only a minor e f f e c t on the center-of-gravity accelerations.

A s a f u r t h e r check, t h e average airplane acceleration w a s determined f o r
a short section of the t e s t run by using the accelerometer measurements
from 22 locations along the wing and fuselage of t h e test airplane along
w i t h t h e i r associated masses. The results obtained indicated t h a t , except
f o r t h e presence of high-frequency fluctuations associated w i t h t h e higher
s t r u c t u r a l modes, the center-of-gravity acceleration provided a good meas-
ure of t h e airplane acceleration. On t h i s basis, t h e center-of-gravity
acceleration measurements were faired t o remove t h e e f f e c t s of t h e higher
s t r u c t u r a l modes and used as a measure of the airplane loading. The loads
obtained on t h i s b a s i s were then converted t o s t r a i n s f o r the various sta-
t i o n s on a s t a t i c basis by using t h e s t r a i n s per g as measured i n slow
pull-up maneuvers a t t h e t e s t dynamic pressure. "he s t r a i n s obtained on
t h i s basis provide a measure of the s t r a i n s f o r a s t a t i c a l l y e l a s t i c air-
plane and were used i n obtaining t h e solid-line curves of f i g u r e 6.

I n order t o provide a measure of the s t a t i c - a e r o e l a s t i c e f f e c t s on

t h e s t r a i n s , t h e s t r a i n s per g i n pull-ups were a l s o determined f o r t h e
condition of l o w or zero dynamic pressure. A t low dynamic pressure, t h e
s t a t i c a e r o e l a s t i c e f f e c t s tend t o be minimized, and thus t h e s t r a i n s
per g obtained a t low dynamic pressure provide a basic "rigid" airplane
reference condition. Figure 7 i l l u s t r a t e s t h e values of bending s t r a i n
per g obtained at one s t a t i o n f o r various values of dynamic
pressure. The v a r i a t i o n of th; ; t r a i n per g appears l i n e a r over the w i d e
range of dynamic pressure represented. A l i n e a r extrapolation t o a value
of dynamic pressure of 0 was therefore used. The difference between t h e
s t r a i n per g (0.65) f o r a dynamic pressure of 0 and the value (0.51) f o r
t h e t e s t dynamic pressure provides a measure of the s t a t i c a e r o e l a s t i c
s t r a i n allevia'tion f o r t h i s s t a t i o n . The amount of t h i s a l l e v i a t i o n is

.
about 22 percent f o r t h i s s t a t i o n and varied somewhat f o r t h e other
stations

The s t r a i n s per g f o r t h e condition a t a dynamic pressure of 0 were

used with t h e airplane center-of-gravity acceleration t o obtain the sec-
ond set of reference s t r a i n s . These reference s t r a i n s constitute s t r a i n s
f o r an airplane embodying no dynamic f l e x i b i l i t y or s t a t i c a e r o e l a s t i c
e f f e c t s and w e r e used t o obtain t h e dashed curves of figure 6.

I n t h e upper p a r t of figure 6, the dynamic amplification, shown by

the s o l i d curve, i s about 10 percent a t the root but increases rapidly
along the span and reaches a value of about 2 a t the 60-percent-semispan
s t a t i o n . The amplification f a c t o r s obtained by considering both t h e
s t a t i c a l l e v i a t i o n and dynamic e f f e c t s , shown by the dashed l i n e , are
below 1 a t the inboard s t a t i o n s but reach values of about 1.5 a t the
60-percent -semi span sta t ion.

-.
E 9

The amplification f a c t o r s based on t h e r a t i o s of s t r a i n s f o r equal

frequency of occurrence, shown i n t h e lower p a r t of f i g u r e 6, follow the
same p a t t e r n as those based on the root-mean-square values but are every-
where higher. It thus appears t h a t , f o r t h e swept-wing B-47A airplane, t h e
e f f e c t s of d y n d c f l e x i b i l i t y are quite large, p a r t i c u l a r l y a t t h e m i d -
span stations, but t h a t favorable s t a t i c a e r o e l a s t i c a l l e v i a t i o n moderates
t h e large dynamic amplifications. I n addition, it is evident t h a t several
values f o r the amplification f a c t o r may be obtained depending upon t h e
p a r t i c u l a r d e f i n i t i o n or reference used. The p a r t i c u l a r amplification
f a c t o r of significance depends upon the s p e c i f i c application.

EFFECTS OF SPANWISE VARIATIONS I N TURBULENCE

A s a f i n a l point, some i n d i r e c t evidence of t h e e f f e c t s of spanwise

variations i n turbulence on airplane gust response w i l l be presented.
A s mentioned previously, two methods could be used i n determining the
frequency-response functions f o r t h e B-47A: t h e spectrum and t h e cross-
spectrum methods. A s indicated, i f no spanwise v a r i a t i o n s i n turbulence
existed, then t h e gust input would be adequately r e f l e c t e d by t h e point
measurements of t h e angle-of-attack vane. For t h i s case, both methods
should yield i d e n t i c a l r e s u l t s f o r the frequency-response function. Fig-
ure 8 shows t h e frequency-response functions obtained by t h e two methods
f o r the airplane center-of-gravity normal acceleration. It i s quite
c l e a r t h a t the cross-spectrum r e s u l t s are consistently lower than those
obtained by t h e spectrum method. Similar differences a l s o have been
found between s t r a i n frequency-response functions obtained by the two
methods. The differences between these two results are suggestive of
the e f f e c t s of spanwise variations i n gusts which have received consider-
able a t t e n t i o n recently i n a n a l y t i c a l studies by Diederich, Drischler,
and Liepmann. (See r e f s . 2, 6, and 7.)

By making use of some r e s u l t s obtained by Diederich (ref. 5 ) , f i r s t -

order adjustments f o r the e f f e c t s of spanwise v a r i a t i o n s of turbulence
f o r the condition of isotropic turbulence w e r e made t o these two results
i n accordance w i t h t h e analysis given i n t h e appendix. The adjusted
frequency-response functions obtained are shown i n f i g u r e 9. Two e f f e c t s
may be noticed from t h i s adjustment. F i r s t , t h e two frequency-response
functions are now i n much b e t t e r agreement, and second, t h e adjustment
has raised t h e two curves by from 10 t o 20 percent over most of the fre-
quency range i n accordance with the span averaging functions of figure 10.
The b a s i s f o r t h e underestimation and d i s t o r t i o n i n the apparent frequency-
response function of figure 8 stems from the f a c t t h a t t h e point gust input
used i s t o o high. The e f f e c t i v e gust input i s the average gust v e l o c i t y
across the span which, as indicated i n reference 5 , tends t o f a l l below
the point input a t a l l frequencies. A s a consequence, t h e airplane
acceleration response i s lower than w h a t would be expected f o r uniform
10

turbulence across t h e span i n accordance w i t h t h e point measurements.

These r e s u l t s t h u s suggest t h a t the e f f e c t s of spanwise v a r i a t i o n s i n
turbulence are s i g n i f i c a n t f o r such airplanes as t h e B-47A. I n addition,
since a n a l y t i c r e s u l t s i n d i c a t e t h a t the e f f e c t s of spanwise v a r i a t i o n s
i n turbulence are primarily a function of span, t h e s e e f f e c t s may he
even more important f o r a i r p l a n e s with l a r g e r span than t h e present
airplane.

CONCLUDING REMARKS

It has been indicated t h a t , f o r the B-29 airplane, t h e e f f e c t s of the

first and second symmetrical bending m o d e s y i e l d moderate s t r a i n ampli-
f i c a t i o n s i n rough a i r a l l along t h e airplane span. Calculations involving
one o r preferably two s t r u c t u r a l modes appear t o y i e l d r e l i a b l e estimates
of f l e x i b i l i t y on t h e s t r a i n s .

For swept-wing airplanes, t h e e f f e c t s of f l e x i b i l i t y a r e complicated

by the importance of s t a t i c a e r o e l a s t i c i t y e f f e c t s i n addition t o t h e
dynamic response. I n the case of t h e B-47A airplane, t h e dynamic a n r p l i f i -
cations appear t o be quite large, p a r t i c u l a r l y i n the midspan region.
These amplifications are a t least p a r t i a l l y balanced by l a r g e and favor-
able s t a t i c a e r o e l a s t i c e f f e c t s associated with sweep. Finally, some
i n d i r e c t r e s u l t s suggest that spanwise v a r i a t i o n s i n tukbulence have a
s i g n i f i c a n t e f f e c t on t h e responses of the B-47A airplane.
 11

APPENDIX

EFFECTS OF SPANWISE VARIATIONS I N TURB-CE ON

MEASURFD FREQUENCY-RESPONSE FUNCTIONS

I n t h e main body of t h i s paper, it was indicated t h a t , f o r uniform

turbulence across t h e airplane span, t h e amplitude of t h e frequency-
response function f o r gust disturbances I
H(f) I
can be obtained experi-
mentally by two methods; the s p e c t r a l method which i s based on t h e
relation

and t h e cross-spectral method which i s based on t h e r e l a t i o n

The subscripts s and c a r e used t o d i f f e r e n t i a t e between t h e e s t i -

mates obtained by t h e two methods. I n t h i s appendix, t h e e f f e c t s on
these estimates of spanwise variations of turbulence w i l l be outlined.
The material t o be presented i s based on t h e analysis of t h e e f f e c t s of
spanwise gust variations given i n reference 5 .

The response of an airplane z(t) t o two-dimensional turbulence,

t h a t is, turbulence varying along t h e f l i g h t path and along t h e spanwise
direction, may be expressed as

where

b wing span

x(t,y) v e r t i c a l gust velocity impinging a t wing leading edge a t

time t and span location y

h(t,y) response a t time t due t o u n i t impulse gust impinging

a t wing between s t a t i o n y and y + dy a t time zero
1;: .
The power spectrum of z(t) as derived i n reference 5 i s given by

where

4Z(f> power spectrum of response z(t)

4x (f , y2-y1) cross-spectrum between gust v e l o c i t i e s impinging a t

wing leading edge a t s t a t i o n s y2 and yl

H(f,Y) influence function describing response of airplane

t o u n i t sinusoidal gusts impinging a t wing leading
edge a t s t a t i o n y and i s given by

"he a s t e r i s k designates t h e complex conjugate.

I n these terms, t h e frequency-response function obtained by t h e s p e c t r a l

method can be expressed as

where 4(f,O) i s the power spectrum of the gust input a t span s t a t i o n 0

(such as measured w i t h the vane).

The expression f o r 1 H s ( f )I given by equation (A5) may be viewed

as an average frequency-response function. Further i n s i g h t i n t o t h i s
result can be obtained by considering the s p e c i a l case

where
 13

For t h i s case

2
1
Equation (A7) indicates t h a t t h e quantity Hs ( f ) I determined from
measurements represents t h e "frequency part" of t h e frequency-response
function and i s . m u l t i p l i e d by the term i n brackets, which m i g h t be con-
sidered a span averaging function. For the case of uniform turbulence
across t h e span, equation (A7) reduces t o

If t h e turbulence i s isotropic w i t h a known spectrum, t h e averaging

function given by t h e term i n brackets i n equation (A7) can be evaluated.
Evaluations of the numerator of t h e term i n brackets a r e given i n r e f e r -
ences 5 and 6 f o r several assumed power spectra of turbulence and f o r
several assumed span d i s t r i b u t i o n s y. By using these results, t h e func-
t i o n 7 ( f ) was evaluated f o r a p a r t i c u l a r case and the results obtained
1
a r e shown i n figure 10. The r e s u l t shown i n figure 10 i s f o r t h e case of
a uniform span d i s t r i b u t i o n (7(y) = l/b) and a spectrum of v e r t i c a l gust
v e l o c i t y given by

where Q = -
25rf
v and V i s t h e airplane forward speed. I n equation (A8),
crw i s the root-mean-square gust v e l o c i t y and L i s t h e s c a l e of turbu-
lence. A value of L = 1,000 f e e t appears t o be representative of con-
d i t i o n s f o r atmospheric turbulence. For present purposes, a value of
b / L = 0.1 was assumed.

A similar analysis may be applied t o t h e cross-spectrum case and

y i e l d s the following results:
.. .. , I , > . _ I . I __-

14

where t h e weighting function given by the term i n brackets may a l s o b e

viewed as an average span weighting function y2( f ) .The function T2( f )
can a l s o be evaluated from figure 4 of reference 5 f o r t h e case of mi-
form span loading and t h e gust spectrum of equation ( A 8 ) . This func-
t i o n Y2(f) i s a l s o shown i n f i g u r e 10.

If t h e weighting functions Yl(f) and T 2 ( f ) given i n equations (A7)

and (Ag) are divided i n t o the measured values of I H s ( f ) Iand l H c ( f ) 1,
respectively, of f i g u r e 8, estimates of I H ( f ) l may be obtained which
a r e compatible. The r e s u l t s obtained i n t h i s manner a r e shown i n
figure 9.
 REFERENCES

1. Mickleboro, Harry C., and Shufflebarger, C . C.: F l i g h t Investigation

of t h e Effects of Trpnsient Wing Response on Wing S t r a i n s of a
Twin-Engine Transport Airplane i n Rough A i r . NACA TN 2424, 1951.

2. Murrow, Harold N., and Payne, Chester B.: F l i g h t Investigation of

t h e Effect of Transient Wing Response on Wing Strains of a Four-
Engine Bomber Airplane i n Rough A i r . NACA TN 2951, 1953.

3. Houbolt, John C., and Kordes, Eldon E.: S t r u c t u r a l Response t o

Discrete and Continuous Gusts of an Airplane Having Wing-Bending
F l e x i b i l i t y and a Correlation of Calculated and F l i g h t Results.
NACA Rep. 1181, 1954. (Supersedes NACA TN 3006; a l s o contains
e s s e n t i a l material from TN 2763 and TN 2897.)

4. Crane, Harold L., and Chilton, Robert G.: Measurements of Atmospheric

Turbulence Over a Wide Range of Wavelength for One Meteorological
Condition. NACA TN 3702, 1956.

5. Diederich, Franklin W.: The Response of an Airplane t o Random Atmos-

pheric Turbulence. NACA TN 3910, 1957.

6. Diederich., Franklin W., and Drischler, Joseph A.: Effect of Spanwise

Variations i n Gust I n t e n s i t y on t h e L i f t Due t o Atmospheric Turbu-
lence. NACA TN 3920, 1957.

7. Liepmann, H. W.: Extension of t h e Statistical Approach t o Buffeting

and Gust Response of Wings of F i n i t e Span. Jour. Aero. Sci.,
vol. 22, no. 3, Mar. 1955, pp. 197-200.
 TEST AIRPLANES

Figure 1

MEASURED 6-29 STRAIN

FREQUENCY RESPONSE FUNCTION

IH(f)l,
STRAIN
GUST VELOCITY

1 I I I I I
0 I 2 3 4 5
f , CPS
..
'E
17

CALCULATED B -29 STRAIN

FREQUENCY RESPONSE FUNCTION

Y/+= 0.15
A
IH(f
STRAIN
,I)
GUST VELOCITY

0 I 2 3 4 5
f,CPS

Figure 3

B -29 STRAIN A M P L l FIC A T I O N

%, FLEX
1.1
v€, REF
--_ -----_------
--*_
I .o I I I I I I I I 1 1
0'
o EXPERIMENTAL
CALCULATED
-----VERT. MOTION + 1st WING BENDING
\-VERT. MOTION + 1st AND 2d WING BENMNG

€FLEX
€REF

0 .2 .4 .6 .8 1.0
y/+

Figure 4
18

MEASURED B - 4 7 A STRAIN FREQUENCY-

RESPONSE F U N C T l O N

6
/\
I4,
STRAIN
4
GUST VELOCITY

0 I 2 3
f ,CPS

Figure 5 ( a )

MEASURED B-47A STRAIN FREQUENCY-RESPONSE FUNCTION

PHASE ANGLE

PHASE ANGLE I
DEG

I I I I I I I
0 I 2 3
f, CPS
227
Figure 5(b)
 B-47A STRAIN AMPLIFICATION

: %,
ai,FLEX
REF :
0
- o----
3 o--DYNAMIC AMPLIFICATION
STATIC ALLEVIATION +

2
'FLEX
'REF
I -----
-

0 .2 .4 .6 .8 I.o
y/*

Figure 6

VARIATION OF STRAIN PER g IN PULL-UP

WITH DYNAMlC PRESSURE

---- -\ I

STRAIN
INDICATION .4-
PER g GUST
CONDITION
TESTA
-2 - I
I
I I I I I,

Figure 7
2c

B-47A NORM4L-ACCELERATION FREQUENCY-

I?ESPONSE FUNCTION

ACCELERAT I ON
G U S T VELOCITY

0 .5 I.o L5 20 2.5
f ,CPS

Figure 8

B-47A NORMAL-ACCELERATION FREQUENCY-

RESPONSE FUNCTION
ADJUSTED FOR SPANWISE GUST VARIATIONS

IH(f)lADJ 3
ACCE L E R ATlON
GUST VELOCITY

0 .5 I.o 15 2.0 2.5

f,CPS .
-0
&&9
Figure 9
 21

SPAN AVERAGING FUNCTIONS 7 ( f ) AND % ( f ) FOR b/L= 0.1

.7 -

Figure 10

230
 LOADS lMPLICATI0NS OF GUST-ALLEVIATION SYSTESIS

By W i l l i a m H. P h i l l i p s

Langley Aeronautical Laboratory

SUMMARY
I
A review i s presented of the f a c t o r s affecting gust loads and the
methods o r devices which reduce these loads. Aerodynamic devices which
reduce the l i f t - c u r v e slope include spoiler-deflector controls, f o r
which some data a r e presented i n the Mach number range from 0.4 t o 1.1.
Systems a r e a l s o considered i n which a sensing device i s used t o operate
gust-alleviation controls. Two b a s i c a l l y d i f f e r e n t types of sensing
devices a r e possible, the load-sensing type and the angle-of -attack--
sensing type. These devices are compared and t h e i r limitations discussed.
Some preliminary f l i g h t measurements of wing-root bending moment due t o
turbulence a r e presented f o r a gust-alleviation system installed i n a
small twin-engine transport airplane. This system increased the wing-
root bending moments as compared with those of the basic airplane. This
increase r e s u l t e d from the f a c t that the system as tested w a s adjusted
t o reduce acceleration and, as a r e s u l t , overcompensated f o r the wing-root
bending moments due t o gusts. Some f l i g h t measurements of the e f f e c t s of
a yaw damper on the t a i l loads of a bomber airplane are a l s o presented.

INTRODUCTION

G u s t a l l e v i a t i o n has been of continued i n t e r e s t t o almost every

group i n aviation since i t s inception, but it has not been incorporated
i n production airplanes. Apparently the reason f o r the lack of use of
gust a l l e v i a t i o n is that detailed analyses of promising devices e i t h e r
pose problems insoluble at a given stage i n aircraft development o r
result i n p r a c t i c a l disadvantages t h a t seem t o outweigh the p o t e n t i a l
benefits. Systems have been studied by various organizations with the
objectives of providing improved riding comfort, increased s a f e t y due t o
load reductions, reduced s t r u c t u r a l weight, and more stable gun platforms.

Inasmuch as the various systems are perennially proposed as m e a n s of

improving a i r c r a f t , a need f o r a summary of the methods available f o r gust
a l l e v i a t i o n and the problems associated with these methods i s apparent.
The present r e p o r t considers the loads implications of gust-alleviating
methods.
2 '3

SYMBOLS

C wing chord

l i f t - c u r v e slope
cLa
9 dynamic pressure

U gust velocity

V t r u e airspeed

ACD drag-coefficient increment

M Mach number

a angle of a t t a c k

VS velocity a t s t a l l

vC
cruising speed

VMAX maximum speed

DISCUSSION

The f a c t o r s a f f e c t i n g gust loads a r e shown i n t a b l e 1. The

f i r s t f a c t o r i s the d i r e c t load due t o the gust. As indicated by Vne
formula, this load i s proportional t o the l i f t - c u r v e slope CL, the
dynamic pressure q, and the change i n angle of a t t a c k due t o the gust,
U/V, where U i s the gust velocity and V i s the t r u e airspeed. This
load may therefore be reduced by reducing o r by reducing q. The
cL,.
second f a c t o r i s the airplane motion due t o gusts o r due t o controls.
The airplane motion i s dependent on the basic airplane s t a b i l i t y , and
may a l s o be influenced by the operation of controls manually o r by an
autopilot. The t h i r d f a c t o r t o be considered i s the action of s p e c i a l
controls t o o f f s e t the gust load d i r e c t l y . This category would include
the relieving e f f e c t s due t o wing bending, the use of hinged surfaces or
w i n g s , and f i n a l l y the use of special gust-alleviating 'controls, such as
w i n g f l a p s , operated by a sfrvome2l-jpgsm.
t 3 ,CI
 3

E f f e c t of Spoiler-Deflector Control

Aerodynamic devices which reduce t h e l i f t - c u r v e slope include t h e

use of sweep o r reduced aspect r a t i o , t h e e f f e c t s of which a r e w e l l known,
and t h e use of chordwise s l o t s which, i n effect, reduce the aspect r a t i o .
Another device f o r reducing the l i f t - c u r v e slope i s the spoiler-deflector
control ( r e f . 1). The e f f e c t s of this device as a function of Mach num-
ber on a swept wing are shown i n f i g u r e 1. This f i g u r e shows the percent
of the basic wing load produced by the wing with a spoiler-deflector con-
t r o l covering 18 percent of the span. The increment i n drag c o e f f i c i e n t
i s a l s o shown. These data a r e taken from reference 2. The s p o i l e r height
above the w i n g w a s 0.02’3 and t h e d e f l e c t o r projection below the wing w a s
0 . 1 5 ~ . I n t h i s case, the reason f o r t h e s h o r t span of the s p o i l e r -
deflector control w a s t o locate it inboard of t h e a i l e r o n and outboard of
the horizontal t a i l . T h i s device provides a l a r g e increase i n drag as
well as a reduction i n l i f t - c u r v e slope. For t h i s reason, this c o n t r o l
might be useful f o r slowing an airplane down when rough a i r i s encountered
but it would not be desirable f o r continuous use i n high-speed f l i g h t .
Tests on s p e c i f i c configurations have shown t h a t this c o n t r o l may be
located so as t o minimize longitudinal t r i m changes. Location of t h e
spoiler-deflector control ahead of a n aileron, however, has been found t o
reduce g r e a t l y the a i l e r o n effectiveness, as might be expected. Possibly,
the spoiler-deflector control could be operated i n conjunction with t h e
a i l e r o n t o overcome t h i s d i f f i c u l t y .

E f f e c t of Sensor and Servo System Operating Special Controls

I n systems which use a sensing device t o d e t e c t t h e gusts and t o

operate gust-alleviation controls, two b a s i c a l l y d i f f e r e n t types of
sensing devices are possible: one, the load-sensing type such as s t r a i n
gages or an acceleroneter, and the other, t h e angle-of -attack-sensing
type. The e f f e c t s of these devices d i f f e r i n several important respects.
F i r s t , as shown i n figure 2, these sensing devices exhibit d i f f e r e n t
trends of effectiveness as a function of airspeed. The gust envelope
f o r a t y p i c a l transport airplane i s a l s o shown i n this f i g u r e . With the
load-sensing type of gust a l l e v i a t i o n t h e percent a l l e v i a t i o n increases
with increasing speed, whereas with the angle-of-attack-sensing type t h e
percent a l l e v i a t i o n tends t o remain constant. Thus, i f the two systems
are designed t o have t h e same effectiveness a t a given speed, the load-
sensing type w i l l show g r e a t e r effectiveness a t higher speeds. These
results apply only i f the system gain i s held constant as might occur
with t h e use of some simple types of aerodynamically operated gust a l l e v i -
a t o r s . If a servo system i s used, of course, it i s possible t o vary t h e
gain of the system as a function of speed and thereby change the e f f e c t s
of speed from those sham.
233
 The second difference between the two systems i s concerned with
the d i f f e r e n t nature of the e f f e c t s which l i m i t the maximum a l l e v i a t i o n
obtainable. The usual l i m i t a t i o n i n the case of the load-sensing system
i s the occurrence of a high-frequency i n s t a b i l i t y . Figure 3 shows the
percent.load experienced with a load-sensing system as a function of
the r e l a t i v e gain. A r e l a t i v e gain of 1 on this scale represents a
condition i n which, f o r example, a load increment corresponding t o 1 g
on the sensor w i l l operate the controls t o produce a load increment of
- l g on the airplane. Very high gains a r e required t o obtain a l a r g e
percent of a l l e v i a t i o n . A load-sensing system, however, is a t y p i c a l
closed-loop system, f o r which high gains a r e l i k e l y t o result i n
i n s t a b i l i t y . Analog-computer studies f o r c e r t a i n t y p i c a l cases have
shown t h a t reduction i n load below 50 percent of the unalleviated case
r e s u l t e d i n an o s c i l l a t o r y response and that these o s c i l l a t i o n s became
unstable a t a higher gain as shown.

With the angle-of-attack-sensing system, t h i s type of i n s t a b i l i t y

i s much less l i k e l y t o be encountered because this arrangement i s much
more nearly an open-loop system, that is, operation of the angle-of-
a t t a c k sensor causes deflection of the a l l e v i a t i o n controls but opera-
t i o n of the a l l e v i a t i o n controls has only a minor e f f e c t on the indica-
tions of the angle-of-attack sensor. For t h i s reason, the l i m i t a t i o n
i s less on t h e amount of gain ,which may be employed, and systems are
designed usually so t h a t the e f f e c t of a uniform gust i s completely
counteracted by the a l l e v i a t i o n controls. With t h i s type of system the
limits on the load reductions obtainable r e s u l t primarily from the f a c t
that sensing t h e gust a t one point does not give a representative indi-
cation of the average angle of attack across the wing span. Unpublished
t h e o r e t i c a l studies have shown t h a t the e f f e c t of nonuniform gust veloc-
i t y across t h e span f o r the angle-of-attack-sensing system i s a function
of the r a t i o of the w i n g span t o t h e scale of turbulence. Because of
the large s c a l e of atmospheric turbulence, values of a l l e v i a t i o n as high
as 80 percent may be obtained with a single sensor located ahead of the
nose which operates the controls with no lag. The addition of a s u i t a b l e
f i l t e r t o the output of the sensor which reduces the response t o high-fre-
quency gusts f u r t h e r improves the a l l e v i a t i o n t h e o r e t i c a l l y attainable.
Such a filter may a l s o be desirable i n order t o reduce the e f f e c t s of
s t r u c t u r a l feedback, which might cause the system t o reinforce s t r u c t u r a l
modes of o s c i l l a t i o n i f the system response were not attenuated a t high
frequencies.

F l i g h t Tests of Gust-Alleviation System

I n s t a l l e d i n Airplane

A f l i g h t investigation has been made of a gust-alleviation system

i n s t a l l e d i n a small twin-engine transport airplane. Some preliminary 4

, of t h i s study have been reported previously (refs. 3 and 4).

I w a s designed improvement of passenger
 5

camfort. The system uses an angle-of-attack vane t o operate wing f l a p s

through a servo system. The reduction i n acceleration obtained with
t h i s system i s shown i n f i g u r e 4. The r e l a t i v e values of normal accel-
e r a t i o n as a function of frequency, obtained w i t h t h e basic airplane and
the gust-alleviated airplane f o r comparable conditions of turbulence, and
the e f f e c t of the system on pitching velocity are shown. The r e l a t i v e
values p l o t t e d i n t h i s f i g u r e are proportional t o the square r o o t of the
power s p e c t r a l density of the response and show t h e c o r r e c t r e l a t i v e
values as well as t h e v a r i a t i o n of the response with frequency. The nor-
mal acceleration f o r the a l l e v i a t e d airplane w a s reduced t o . 3 0 o r 40 per-
cent of t h a t f o r t h e basic airplane i n the frequency range from 0 t o
2 cyles per second. The pitching velocity, which w a s small f o r t h e basic
airplane, w a s further reduced f o r the a l l e v i a t e d case.

Extensive strain-gage measurements have been made t o determine the

e f f e c t of this system on the s t r u c t u r a l loads. These data have not been
completely evaluated a t this t i m e . The e f f e c t of the system on wing-root
bending moment i s s h a m i n f i g u r e 5 . The wing-root bending moment is
a c t u a l l y increased by the gust-alleviation system. The explanation of
this increase i s i n d i c a t e d b y t h e i n s e r t on t h e f i g u r e which shows t h e
change i n span load d i s t r i b u t i o n f o r b a s i c and a l l e v i a t e d airplanes due
t o a small p o s i t i v e increment of angle of attack. I n the a l l e v i a t e d case,
the flaps on the wing are deflected up on the outboard sections and dawn
near the r o o t . This arrangement provides downwash conditions a t the
t a i l which minimize pitching moments due t o the gusts. The r e s u l t a n t
l i f t due t o this combination i s about zero, b u t because the t i p s e c t i o n s
are much more e f f e c t i v e i n producing bending moment, the result i s a
negative bending moment due t o an up gust. The magnitude of this nega-
t i v e bending moment i s a c t u a l l y greater f o r a given gust than the posi-
t i v e bending moment on the b a s i c airplane.

These r e s u l t s apply only s o long as the system i s operating i n i t s

lineax range. A t a gust v e l o c i t y of about 10 f e e t per second, the f l a p s
reach t h e i r stops. For greater up-gust v e l o c i t i e s , t h e bending moments
would increase i n the p o s i t i v e d i r e c t i o n as on the b a s i c airplane. Thus,
f o r a gust velocity of 20 f e e t per second, the bending moment would be
expected t o come back t o about zero, and f o r higher gust v e l o c i t i e s would
again become positive. This sytem i s therefore one which serves t o
improve passenger comfort i n the frequently encountered small gust veloc-
i t i e s , but which reduces t h e s t r u c t u r a l loads due t o severe gusts. No
f l i g h t data are available, however, t o show the c h a r a c t e r i s t i c s of t h e
system i n severe turbulence. The system increased the magnitude and
frequency of t a i l loads as well as the stresses i n minor s t r u c t u r a l com-
ponents such as the rear spar, t h e wing f l a p s , and s o f o r t h . This result
indicates that f a t i g u e loads would be a more serious problem f o r t h e gust-
a l l e v i a t e d airplane.
6

Effect of Yaw Damper on Vertical-Tail Loads

Some measurements have been made t o determine the e f f e c t of a yaw

damper on the v e r t i c a l - t a i l loads experienced by a bomber airplane i n
rough air a t various a l t i t u d e s . These r e s u l t s a r e shcwn i n figure 6,
which presents the probability of exceeding a given value of v e r t i c a l -
t a i l spar s t r a i n with the yaw damper on and off a t two a l t i t u d e s ,
35,000 f e e t and 5,000 f e e t . The yaw damper reduces the magnitude of
loads considerably i n the high-altitude case. The damping of the Dutch
r o l l motion of 'the airplane under these conditions i s l o w so that a
large resonance a t the Dutch roll frequency i s obtained. The e f f e c t of
the yaw damper i s primarily t o reduce t h i s amplification of load due t o
the Dutch roll motion. A t l o w a l t i t u d e where the damping of the air-
plane i s b e t t e r , the gains due t o the yaw damper are small.

Sane studies have been made t o determine the f e a s i b i l i t y of reducing

the loads on the wings by use of the normal elevator control. The r e s u l t s
are similar t o those obtained i n the l a t e r a l case; that is, i f the air-
plane has very l o w damping i n p i t c h the loads may be reduced through
elimination of the resonant peak of the short-period mode ( r e f . 5 ) . How-
ever, any attempt t o reduce the d i r e c t e f f e c t of the gust on the l i f t of
the surfaces by heading the airplane i n t o the gusts requires large
pitching motions of the airplane.

CONCLUDING REMAFKS

A b r i e f review has been given of the basic methods of gust a l l e v i a -

t i o n , and some r e s u l t s obtained i n f l i g h t t e s t s of a gust-alleviation
system have been presented. A system designed f o r improvement of pas-
senger comfort did not reduce s t r u c t u r a l s t r e s s e s while operating i n i t s
l i n e a r range. The system would be expected t o reduce the wing loads due
t o severe gusts, b u t loads i n the t a i l and other s t r u c t u r a l components
were increased.
 7
*
REFERENCES

1. Croom, Delwin R., Shufflebarger, C. C., and Huffman, Jarrett K.: An

Investigation of Forward-Located Fixed Spoilers and Deflectors as
Gust Alleviators on an Unswept-Wing Model. NACA TN 3705, 1956.

2. Croom, Iklwin R., and Huff'man, Jarrett K.: Investigation at Transonic

Speeds of Deflectors and Spoilers as Gust Alleviators on a 35' Swept
Wing - Transonic-Bump Method. (Prospective NACA paper. )

3. Kraft, Christopher C., Jr.: Initial Results of a Flight Investigation

of a Gust-Alleviation System. NACA TN 3612, 1956.

4. Cooney, T. V., and Schott, Russell L.: Initial Results of a Flight

Investigation of the Wing and Tail Loads on an Airplane Equipped
With a Vane-Controlled Gust-Alleviation System. NACA TN 3746, 1956.
5 . Vitale,A. James, Press, H., and Shufflebarger, C. C.: An Investiga-
tion of the Use of Rocket-Powered Models for Gust-Load Studies With
an Application to a Tailless Swept-Wing Model at Transonic Speeds.
NACA TN 3161, 1954.

237
8

TABLE I

FACTORS AFFECTING GUST LOADS

A LOAD = SUM OF
I-DIRECT LOAD DUE TO GUST-

(a CL,
oq
2-AIRPLANE MOTION DUE TO GUST OR CONTROLS
(A) BASIC AIRPLANE STABILITY
(B OPERATION OF CONTROLS MANUALLY OR BY AUTOPILOT
3-ACTION OF SPECIAL CONTROL TO OFFSET LOAD
(AI WING FLEXIBILITY
(B) HINGED SURFACES
(c] SENSOR AND SERVO SYSTEM OPERATING SPECIAL CONTROLS
 9

LOAD ALLEVIATION AND DRAG OF A SPOILER-DEFLECTOR

0 .4 .6 .8 I .o 1.2
M

Figure 1

EFFECT OF SENSING SYSTEM ON

VARIATION OF LOAD WITH AIRSPEED

LOAD, g UNITS

0 .5 I.o 4.5
"PC

Figure 2 233
lo

LOAD REDUCTION WITH LOAD- SENSING SYSTEM

80
loo R /SCILLATIONS

I I I I I I I I I
0 1 2 3 4 5 6 7 8
RELATIVE GAIN

Figure 3

EFFECT OF GUST- ALLEVIATION SYSTEM ON AIRPLANE MOTIONS

BASIC AIRPLANE
ACCELERATION
AMPLITUDE

0. \

PITCHING-
VELOCITY
AMPLITUDE

0
1-

I
------ 2 3
FREOUENCY, CPS

Figure 4
 11

EFFECT OF GUST- ALLEVIATION SYSTEM

ON ROOT BENDING MOMENTS

SPAN LOAD DISTRIBUTIONS

BENDING- GUST-ALLEVIATED AIRPLANE

MOMENT BASIC AIRPLANE
AMPLITUDE

I I I I I I I
0 I 2 3
FREQUENCY, CPS

Figure 5

EFFECT OF YAW DAMPER ON VERT1 CAL- TAIL LOADS

DAMPER OFF
--- DAMPER ON

M = 0.65 M = 0.35

.I I- ’\ \ t \\

\ ’
.001
0 0
VERTICAL-TAIL SPAR STRAIN

Figure 6
 RECENT DATA ON TI33 FRICTION DURING LANDING

By Sidney A. Batterson

Langley Aeronautical Laboratory

SUMMARY

An investigation w a s made a t the Langley landing-loads track t o

obtain data on the c o e f f i c i e n t of f r i c t i o n during wheel spin-up. A
landing gear w a s t e s t e d a t horizontal v e l o c i t i e s ranging from 0 t o
180 f e e t per second together with v e r t i c a l v e l o c i t i e s of 7.0 and 9.3 f e e t
per second. The r e s u l t s indicate the e f f e c t of forward speed and t i r e
i n f l a t i o n pressure on the c o e f f i c i e n t of f r i c t i o n .

INTRODUCTION

The National Advisory Committee f o r Aeronautics has been engaged for

some time i n an experimental study of the wheel spin-up phenomenon during
landing. A s a r e s u l t of these investigations, i t has been possible t o
separate the e f f e c t s of a number of parameters on the coefficient of f r i c -
tion. Until recently, however, data could not be obtained under controlled
conditions f o r e i t h e r the large landing gears or the high forward speeds
which are required f o r modern airplanes. I n order t o obtain such data the
NACA has put i n t o operation a new research f a c i l i t y called the landing-
loads track. This paper presents the f i r s t data obtained a t the t r a c k and
indicates the e f f e c t of forward speed and t i r e i n f l a t i o n pressure on the
coefficient of f r i c t i o n . Although the trends are c l e a r it i s not y e t pos-
s i b l e t o define the variations accurately because of the limited number of
available t e s t r e s u l t s .

TEST CONDITIONS

The t o t a l dropping weight w a s 20,000 pounds. The t e s t covered a

horizontal velocity range from 0 t o 180 f e e t per second and v e r t i c a l
velocities' of 7 and 3 . 3 f e e t per second. The landing gear w a s equipped
with a 44 x 13, type V I I , 26-ply-rating t i r e . The normal i n f l a t i o n
pressure f o r the 20,000-pound weight i s 140 pounds per square inch; how-
ever, the t i r e pressure w a s varied i n one s e r i e s of t e s t s . All tests
were made with the s t r u t inclined a t am angle of 15O. A l i f t force of
20,000 pounds w a s applied t o the dropping mass throughout the impact,
2

but the airplane f l e x i b i l i t y c h a r a c t e r i s t i c s were not simulated. Com-

parisons made of the landing surface a t the landing-loads track and an
active runway a t Langley Field, Va., shared the roughness of both sur-
faces t o be about the same.

RESULTS AND DISCUSSION

Effect of Horizontal Velocity

Figure 1 shows the e f f e c t of forward speed on the v e r t i c a l and drag

reactions a t the ground and on the Coefficient of f r i c t i o n . Time his-
t o r i e s a r e shown of three tests made a t horizontal v e l o c i t i e s vh of 110,
130, and 160 f e e t per second. The sinking speed Vv was 9 . 3 f e e t per
second f o r all three tests. I n reference 1 it was shown that the c o e f f i -
c i e n t of f r i c t i o n i s affected by the v e r t i c a l load. However, during the
three t e s t s shown i n figure 1, the maximum coefficient of f r i c t i o n
occurred a t approximately t h e same value of v e r t i c a l load. This i s
apparent from the upper s e t of curves where the c i r c l e s appearing on the
curves indicate the value of the v e r t i c a l load a t the i n s t a n t of maximum
c o e f f i c i e n t of f r i c t i o n f o r each t e s t . These data, therefore, i s o l a t e the
e f f e c t of forward speed on the maximum coefficient of f r i c t i o n , and show
that increases i n forward speed reduce the maximum coefficient of f r i c t i o n .
A t the present time i n s u f f i c i e n t data are available a t constant v e r t i c a l
load t o define t h i s v a r i a t i o n accurately.

I n figure l i t can be seen that the maximum drag reaction as well as

the maximum coefficient of f r i c t i o n changes over a limited range of forward
speed. Figure 2 shows the variation of maximum d r a g reaction over a much
larger range of horizontal v e l o c i t i e s . The v e r t i c a l velocity i s 9 . 3 f e e t
per second. This v e r t i c a l velocity i s the same as f o r the data i n f i g -
ure 1. Each point represents a separate t e s t , and the maximum drag reac-
t i o n i s p l o t t e d against the horizontal velocity of the t e s t . The curve
reaches a maximum a t a forward speed of about 110 f e e t per second. It
can be seen i n figure 1 that t h i s i s the forward speed where wheel spin-
up occurs j u s t as the v e r t i c a l load roughly l e v e l s o f f . A t forward
speeds l e s s than 110 f e e t per second spin-up occurs sooner and a t a
lower value of v e r t i c a l load and r e s u l t s i n smaller drag loads. A t
the higher forward speeds, although the v e r t i c a l loads remain approxi-
mately constant, the decrease i n maximum coefficients of f r i c t i o n
causes smaller drag loads.

Effect of Tire Pressure

Figure 3 shows the e f f e c t of variation i n t i r e pressure on the coef-

f i c i e n t of f r i c t i o n . Four t e s t s were made a t a horizontal velocity of
 .
3

160 f e e t per second and a v e r t i c a l velocity cf 7 f e e t per second. The

t i r e i n f l a t i o n pressures p covered a range from 35 t o 210 pounds p e r
square inch. I n figure 3 the coefficient of f r i c t i o n i s p l o t t e d against
the instantaneous skidding velocity. Skidding velocity i s defined as
the difference i n velocity between the peripheral speed of the t i r e and
the runway. Since the skidding velocity i s maximum a t the i n s t a n t of
touchdown, spin-up occurs from r i g h t t o l e f t i n figure 3 . I n the region
where the skidding velocity is large, the curves form two groups. Although
the difference between the two curves which form each group i s small, the
difference between the two groups i s large. Furthermore, the curves
obtained a t the higher t i r e pressures exhibit lower coefficients of f r i c -
tion. It is believed that t h i s i s caused primarily by differences i n t e m -
perature i n the t i r e f o o t p r i n t region. Since as the t i r e i n f l a t i o n pres-
sure i s increased the t i r e footprint areas decrease, the heat generated
during skidding i s d i s t r i b u t e d over a smll area, and the rubber i s hot
and primarily molten and exhibits a low coefficient of f r i c t i o n . However,
a t the low i n f l a t i o n pressures, where the t i r e f o o t p r i n t areas a r e l a r g e r ,
the heat i s d i s t r i b u t e d over a greater area; the rubber i s cooler and p r i -
marily i n the s o l i d s t a t e and exhibits a higher f r i c t i o n coefficient.
Evidently, the t r a n s i t i o n region from s o l i d t o molten rubber occurs some-
where i n the region of t i r e pressures between 70 and 140 pounds per square
inch. All the curves reach approximately the same maximum value a t the low
skidding v e l o c i t i e s . This indicates that the t i r e has turned s u f f i c i e n t l y
s o that cool rubber predominates i n the f o o t p r i n t region when spin-up
occurs.

Figure 3 shows that the maximm coefficient of f r i c t i o n obtained a t

a t i r e i n f l a t i o n pressure of 35 pounds per square inch occurs a t an
appreciably higher skidding velocity than those obtained a t the higher
t i r e pressures. This i s a t t r i b u t e d primarily t o the method used i n
measuring the skidding velocity. I n order t o obtain the peripheral speed
of the t i r e , measurenents a r e made of the r o t a t i o n a l velocity of the
wheel, and the r o t a t i o n a l motion of the t i r e with respect t o the wheel i s
neglected. As the t i r e pressure i s decreased the response of the wheel
t o the motion of the t i r e i s lowered. This i s not too c r i t i c a l i n t h e
e a r l y stages of the impact where the increase of the drag load is rela-
t i v e l y slow. However, appreciable e r r o r s a r e introduced as spin-up i s
approached since the wheel i s unable t o follow the rapid changes of t h e
t i r e motion caused by the rapidly changing drag load.

Effect of Braking and Wet Landing Surface

Figure 4 shows time h i s t o r i e s of drag reactions obtained f o r tests

made a t a horizontal velocity of 160 f e e t per second and a v e r t i c a l veloc-
i t y of 7 f e e t per second. Wheel spin-up i n all cases occurred subsequent
t o the leveling off of the vertical-load curve. The curve indicated as
"braked" w a s obtained during a landing made with p a r t i a l brake pressure
4

applied t o the wheel. The curve labeled "free" w a s obtained during a

normal landing with the wheel f r e e t o r o t a t e . Braking caused reduced
spin-up accelerations and, t o a c e r t a i n degree, simulated the e f f e c t of
increasing t h e moment of i n e r t i a of the wheel. It can be seen that, even
though the braked wheel skidded f o r a longer time, the maximum drag load
.
i n both cases w a s the same.

The curve labeled "wet" i n f i g u r e 4 w a s obtained from a landing made

on a wet runway with the wheel f r e e t o turn. The other curye w a s obtained
from a landing made with the wheel locked, ant! the rubber i n the f o o t p r i n t
area w a s undoubtedly molten. For both curves the skidding v e l o c i t i e s a r e
high, up t o 0.10 second after impact, and it can be seen that the drag
loads are approximately the same during t h i s time. This indicates t h a t
the l u b r i c a t i n g e f f e c t s of w a t e r and molten rubber a r e about the same.

I n the case of the landings on wet concrete where the lubricant is

spread along the runway, an appreciable reduction i s seen i n the m a x i m u m
drag load when compared with the maximum drag loads occurring i n e i t h e r
the f r e e o r t h e braked landing. This 'indicates t h a t runway lubrication,
i n addition t o reducing the coefficient of f r i c t i o n a t high skidding
v e l o c i t i e s , a l s o r e s u l t s i n a reduction i n the coefficient of f r i c t i o n a t
the very l o w skidding v e l o c i t i e s where the f r i c t i o n i s of the s t a t i c ,or F

interlocking type.

S W Y OF RESULTS

The p r i n c i p a l results indicated by these tests are as follows:

1. The m a x i m u m c o e f f i c i e n t of f r i c t i o n developed during wheel spin-

up appeared t o decrease as the horizontal velocity w a s increased.

2. Landings made with varying t i r e i n f l a t i o n pressures indicated

that a t the high skidding v e l o c i t i e s the coefficient of f r i c t i o n w a s
lower f o r t h e higher t i r e . i n f l a t i o n pressures. However, a t the low
skidding v e l o c i t i e s , the value of the maximum coefficient of f r i c t i o n
appeared t o be s u b s t a n t i a l l y independent of the t i r e i n f l a t i o n pressure.

3. The maximum drag load obtained during a landing made with p a r t i a l

brake application w a s the same as that obtained during a landing with the
same i n i t i a l conditions and with the wheel f r e e t o r o t a t e .

4. A t high skidding v e l o c i t i e s the lubricating e f f e c t of water on

the runway i s approximately the same as that of molten rubber.
 5

REFERENCE

1. Milwitzky, Benjamin, Lindquist, Dean C., and P o t t e r , Dexter M.: An

Experimental Study of Applied Ground Loads i n Landing. NACA
Rep. 1248, 1955. (Supersedes NACA TN 3246.)
€

GROUND REACTIONS AND COEFFICIENT OF FRICTION

5 0 r x IO3

GROUND
,REACTIONS,
LB

4
1
4
I I

OF
COEFF.

FRICTION
’ ;\
!\
I I :I I I I I
0 .I .2 .3 .4
TIME AFTER IMPACT, SEC

Figure 1

M A X I M U M DRAG REACTIONS
VERTICAL VELOCITY, 9.3 F P S

20-

MAXIMUM
DRAG 15-
REACT1ON,
LB

0 50 IO0 150 200

HORIZONTAL VELOCITY , V h ,FPS
 .a. a a . L . a 1 , 1- , ---
F

'IN.

.2-

.I -

I I I I

Figure 3

DRAG REACTIONS
HORIZONTAL VELOCITY, 160 FPS; VERTICAL VELOCITY, 7 FPS

I I J
0 .IO .20 .30
T I M E AFTER IMPACT,SEC

1.
 A SUMMARY OF GROUND-MADS STATISTICS - 4

By John R. Westfall, Benjamin Milwitzky, %

s
fb 2 . b

Norman S. Silsby, and Robert C . Dreher

Langley Aeronautical Laboratory

SUMMARY

This paper b r i e f l y summarizes t h e more important s t a t i s t i c a l data

obtained by t h e NACA on t h e subject of ground loads. The information
presented r e l a t e s primarily t o landing-impact and t a x i i n g loads; however,
some limited data are a l s o presented on one' phase of ground-handling
loads, namely, braking f r i c t i o n . A number of experimental and t h e o r e t i -
c a l papers dealing w i t h various aspects of t h e subject a r e l i s t e d i n t h e
bibliography.

IXLBODUCTION

The NACA has f o r some time been concerned with t h e study of ground
loads on a i r c r a f t ; these include landing-impact loads, t a x i i n g loads, and
loads due t o ground handling. A number of experimental and t h e o r e t i c a l
reports t h a t have been published on various aspects of the subject a r e
l i s t e d i n t h e bibliography. The present paper summarizes t h e more impor-
t a n t s t a t i s t i c a l data on ground loads obtained by the NACA, which may
serve as a basis f o r predicting t h e ground-loads experience of a i r c r a f t .

The two major phases of t h e ground-loads problem considered herein

a r e landing impact and t a x i i n g loads; however, some limited data a r e a l s o
presented on one phase of ground-handling loads, namely, braking f r i c t i o n .

The section on landing impact considers f i r s t t h e i n i t i a l contact

conditions, then landing-gear reactions and airplane response. !The con-
t a c t conditions are i n t h e form of s t a t i s t i c a l data on f i v e parameters:
v e r t i c a l velocity, horizontal velocity, bank angle, r o l l i n g velocity,
and wing l i f t . Although t h e r e are other parameters, t h e ones considered
appear t o be t h e most important. The discussion of ground reactions and
airplane response i s primarily concerned with t h e a n a l y t i c a l prediction
of v e r t i c a l and drag loads from the known i n i t i a l conditions.

On t h e subject of t a x i i n g loads, d a t a are available i n t h e form of

runway p r o f i l e s and acceleration measurements from VGH records i n taxiing;
t h e l a t t e r are presented i n s t a t i s t i c a l form. Some considerations on air-
plane s t r u c t u r a l response i n taxiing a r e also given.
2

With regard t o ground handling, some limited data are presented on

t h e c o e f f i c i e n t of f r i c t i o n as measured i n braking tests with an airplane
and with a s p e c i a l l y designed tow c a r t .

SYMBOLS

Fv v e r t i c a l ground force, lb

f frequency, cps

g acceleration due t o gravity, f t / s e c 2

An incremental center-of-gravity acceleration, Q units

Pti r e t i r e pressure, l b / s q in.

VV v e r t i c a l v e l o c i t y at contact, f p s

W airplane weight, l b

P c o e f f i c i e n t of f r i c t i o n

Subscript :

max maximum

FKEEXJLTS AND DISCUSSION

Landing Impact

S t a t i s t i c s . - The available s t a t i s t i c s on t h e i n i t i a l contact condi-

t i o n s a r e summarized i n figures 1t o 4. Figure 1 shows p r o b a b i l i t y curves
f o r t h e v e r t i c a l v e l o c i t y a t contact. The ordinate i s t h e p r o b a b i l i t y of
equaling or exceeding a given value of v e r t i c a l velocity; t h e abscissa i s
t h e v e r t i c a l velocity. Because t h e r e appeared t o be d e f i n i t e differences
i n the s t a t i s t i c s for c i v i l and military (land-based) airplanes, the prob-
a b i l i t y curves f o r these two categories are shown separately. Although
t h e reasons f o r t h e differences between t h e two categories are not com-
p l e t e l y understood, it i s probably s i g n i f i c a n t t h a t , on t h e whole, two
d i f f e r e n t generations of airplanes are involved - t h e c i v i l airplanes
were a l l p i s t o n powered, whereas t h e m i l i t a r y airplanes w e r e mostly jets.
Furthermore, some of t h e m i l i t a r y landings were t r a n s i t i o n a l - t r a i n i n g
f l i g h t s . The more severe character of t h e m i l i t a r y landings i s indicated
e
 ... 0 a . I , .. ..
3

by t h e generally higher values of v e r t i c a l velocity a t a given p r o b a b i l i t y

level. For example, 1 i n 10,000 landings of military airplanes would be
expected t o equal o r exceed a v e r t i c a l velocity of about 8$ f e e t per sec-
ond; t h e corresponding value f o r t h e c i v i l airplanes i s about 5 feet per
second.

Figure 2 shows similar probability curves f o r t h e airspeed a t con-

t a c t . The ordinate i s t h e probability of equaling o r exceeding a given
value of airspeed; t h e abscissa i s t h e airspeed a t contact, expressed as
percent above the s t a l l i n g speed. Again, t h e m i l i t a r y operations appear
t o give higher v e l o c i t i e s a t a given probability l e v e l .

With regard t o bank angle and r o l l i n g velocity a t contact, which are

of concern f o r unsymmetrical landings, wing span appeared t o be a better
c r i t e r i o n f o r separating t h e data than types of operations. Accordingly,
w i t h t h e number of engines taken as a crude measure of wing span, t h e
airplanes with four and more engines were put i n t o one category and t h e
single- and twin-engine airplanes i n t o another. Figure 3 presents t h e
probability curves f o r bank angle. A s might be expected from geometric
considerations, t h e l a r g e r span airplanes have somewhat smaller angles
of bank. The probability curves for r o l l i n g velocity a r e shown i n f i g -
ure 4; here t h e differences are more pronounced than f o r bank angle. It
may be t h a t these differences r e f l e c t not only t h e geometrical span
e f f e c t , but a l s o the g r e a t e r moments of i n e r t i a and damping i n roll of
t h e l a r g e r airplanes.

A study was made t o determine whether there was any s t a t i s t i c a l

correlation between bank angle and r o l l i n g velocity, on t h e one hand,
and v e r t i c a l v e l o c i t y or horizontal velocity, on t h e other hand. It was
found t h a t no such correlation existed. There was a l s o no c o r r e l a t i o n
between bank angle and r o l l i n g velocity. It therefore appears t h a t a l l
of these parameters may be t r e a t e d as independent quantities i n calcu-
l a t i n g unsymmetrical landing loads.

With regard t o t h e wing lift a t contact, both past and more recent
studies have shownthat, f o r c i v i l transport airplanes, t h e most probable
value of wing lift a t contact is l g a n d t h a t i n 95 percent of t h e landings
t h e wing lift was between O.9g and 1.1g.

The s t a t i s t i c a l analysis indicated no significant correlations

between wing l i f t and v e r t i c a l velocity.

Although a f a i r l y s u b s t a n t i a l amount of s t a t i s t i c a l data has been

obtained on these i n i t i a l contact conditions, t h e s i z e s of t h e samples
a r e s t i l l i n s u f f i c i e n t t o permit t h e f i n a l resolution of t h e e f f e c t s of
such variables as airplane type, size, wing loading, and so f o r t h .
4

Reactions and response.- On t h e subject of landing-gear reactions

and airplane response, t h e f i r s t problem t o be discussed i s t h e deter-
mination of t h e v e r t i c a l loads when t h e i n i t i a l contact conditions are
given.

I n recent years a n a l y t i c a l methods have been developed f o r predicting

t h e behavior of t h e landing gear during impact and t h e loads applied t o
t h e airplane. Good agreement has been obtained between t h e t h e o r e t i c a l
results and d a t a from landing-gear drop tests. The a n a l y t i c a l method a l s o
appears t o be reasonably s a t i s f a c t o r y when compared with f l i g h t - t e s t data.

Figure 5 shows a comparison of t h e maximum v e r t i c a l loads obtained

from f l i g h t t e s t s of t h e SNJ airplane and calculated maximum loads based
on t h e assumption t h a t t h e airplane i s a r i g i d body. The ordinate i s t h e
maximum center-of -gravity acceleration obtained f o r each impact, and t h e
abscissa i s t h e v e r t i c a l velocity a t contact. There i s considerable scat-
t e r i n t h e f l i g h t data, about 30 percent above 4 feet per second, r e f l e c -
t i n g t h e e f f e c t s of such f a c t o r s as shock-strut binding due t o drag loads,
unsymmetrical landings, s i d e force due t o yawed landings, and so f o r t h ,
a l l of which can modify t h e v e r t i c a l loads developed. A t t h e lower values
of v e r t i c a l v e l o c i t y t h e experimental data a r e somewhat higher than t h e
calculated curve, l a r g e l y because of t h e e f f e c t s of shock-strut preloading
and breakout f r i c t i o n which were not considered i n these calculations.
Above a v e r t i c a l v e l o c i t y of about 4 feet per second, t h e calculated curve
seems t o l i e f a i r l y close t o t h e mean of the experimental data.

Figure 6 shows a similar comparison of measured and calculated maxi-

mum v e r t i c a l loads f o r t h e l a r g e r and more f l e x i b l e B-29 airplane. The
ordinate i s t h e maximum landing-gear vertical-load f a c t o r and t h e abscissa
i s t h e v e r t i c a l velocity. The dashed l i n e represents t h e calculated
values. The break i n t h e curve r e s u l t s from t h e e f f e c t s of shock-strut
breakout, which were included i n these calculations. Although t h e r e i s
a s c a r c i t y of t e s t points a t t h e higher values of v e r t i c a l velocity, it
does appear t h a t t h e calculated r e s u l t s are i n f a i r l y good agreement with
t h e f l i g h t - t e s t data.

With regard t o t h e prediction of spin-up drag loads, t h e most impor-

t a n t unknowns have been t h e magnitude and v a r i a t i o n of t h e coefficient of
f r i c t i o n between t h e t i r e and runway during t h e impact. Figure 7 sum-
marizes t h e available experimental information on t h e v a r i a t i o n of coef-
f i c i e n t of f r i c t i o n during t h e wheel spin-up process. The ordinate i s
t h e coefficient of f r i c t i o n , and t h e abscissa i s t h e skidding velocity,
which i s defined as t h e instantaneous difference between t h e forward
speed of t h e airplane and t h e peripheral speed of t h e t i r e . The shaded
bands are envelopes of t h e data obtained i n several d i f f e r e n t t e s t pro-
grams. The two r e l a t i v e l y narrow bands represent tests with an SNJ landing
gear i n the Langley impact basin. The shorter of these two bands i s f o r
true forward-speed impacts; t h e longer one represents impacts with forward
".
 5

speed i n combination w i t h reverse wheel prerotation ( t o simulate higher

horizontal v e l o c i t i e s beyond t h e range of t h e impact-basin equipment)
The wider dotted band shows r e s u l t s obtained with t h e B-29 airplane i n
f l i g h t t e s t s . The lowest band represents data recently obtained with a
B-57 landirig gear a t the new landing loads track.
A l l these bands illustrate the f a i r l y well-known f a c t t h a t the coef-
f i c i e n t of f r i c t i o n decreases with increasing skidding velocity. The
w i d t h of t h e bands represents t h e e f f e c t s of other f a c t o r s which influ-
ence t h e coefficient of f r i c t i o n , such as the! s l i p r a t i o , t h e v e r t i c a l
load, variable-heating and contamination e f f e c t s , and variations i n t h e
runway surface conditions, plus, of course, experimental e r r o r s . It
w i l l be noted t h a t t h e S N J and B-29 data, which a r e f o r r e l a t i v e l y low
t i r e pressures, are generally i n good agreement w i t h one another, whereas
t h e results f o r t h e B-57 landing gear, which has a much higher t i r e pres-
sure, indicate appreciably lower coefficients of f r i c t i o n . These r e s u l t s
appear t o v e r i f y a previously suspected trend, namely, t h a t the coeffi-
cient of f r i c t i o n decreases at higher t i r e pressures.

Although the shapes of t h e bands are f a i r l y consistent f o r t h e

various tests, it i s not as yet possible t o take i n t o account quantita-
t i v e l y a l l t h e f a c t o r s t h a t contribute t o t h e spread of t h e values of
t h e coefficient of f r i c t i o n . Therefore, it seems t h a t the most p r a c t i -
c a l approach a t present i s t o s e l e c t a shape f o r the curve, and then
normalize it on t h e basis of t h e maximum value. From inspection of t h e
data it appears that a reasonable range of the maximums might be from
about 0.4 t o 0.9, t h e value t o be used depending on the conditions
involved.

Taxiing Loads

On t4e subject of t a x i i n g loads, e s s e n t i a l l y two types of data a r e

available a t present; t h a t i s , measurements of runway p r o f i l e s and center-
of-gravity acceleration data from t h e impact portion of VGH records.

Two possible ways of u t i l i z i n g runway p r o f i l e data are: (1)as

specific displacement inputs in analog computations, o r ( 2 ) as statisti-
c a l power spectra i n generalized harmonic analysis. Both of these
approaches present c e r t a i n p r a c t i c a l d i f f i c u l t i e s . I n the case of the
analog method, one of the major problems i s t o define a representative
runway. A s t o t h e use of generalized harmonic analysis, t h e main stumb-
l i n g blocks a r e as yet unsolved problems of nonlinear landing-gear
response and t h e determination of a unique t r a n s f e r function.

With regard t o t h e VGH acceleration data, several general r e l a t i o n -

ships have been found which appear t o be applicable t o a number of
6

d i f f e r e n t types of airplanes and which provide a common b a s i s f o r e s t i -

mating t h e loads experience i n taxiing. These a r e shown i n t h e following
table :

AIRPLANE

TAXIING T I M E APPROX. 5 MIN

C.G. RESPONSE I TO 2 CPS 5

SPEED AT A n M A X

SCATTER AT CUM.
FREQ. = 1 0 4
<I
< 20 MPH

1.5 I 4

F i r s t , an analysis of t h e time spent i n taxiing, for t h r e e transport

airplane types, showed that t h i s time was remarkably constant, averaging
within a few seconds of 5 minutes.

Second, it was found that the predominant frequency i n t h e center-

of-gravity accelerations - t h a t is, t h e frequency a t which t h e airplanes
respond most t o t h e runway roughness inputs - i s a l s o remarkably constant
f o r f i v e , d i f f e r e n t types of airplanes ranging f r o m t h e 13-36 bomber t o a
small j e t f i g h t e r . This predominant frequency varied only from about 1
t o 2 cps, with an average value of about 1-73 cps, despite a very wide
range of wing bending frequencies f o r the airplanes involved. This fre-
quency of about 1.77 cps appears t o be associated with t h e n a t u r a l vibra-
t i o n of t h e airplane on i t s t i r e s .

Third, it was found that, f o r four transport t y p e s , i n several hundred

taxi runs, t h e maximum t a x i i n g loads occurred a t speeds usually below
20 q h .

Fourth, d e t a i l e d analysis of a l l the loads experienced i n t a x i i n g

f o r four airplane types shared that t h e probability d i s t r i b u t i o n s w e r e
very similar, and t h e s c a t t e r i n t h e load l e v e l a t a given probability
was small, being about 1.5 t o 1, or less than t h e s c a t t e r f o r such param-
e t e r s as v e r t i c a l velocity o r horizontal velocity. Consequently, it may
be assumed t h a t a combined d i s t r i b u t i o n of t a x i i n g loads can be used t o
represent a l l t a x i i n g operations.

Figure 8 shows d i s t r i b u t i o n s of center-of-gravity accelerations i n

taxiing, per 1,000 f l i g h t s . The ordinate i s t h e number of times a given
value of center-of-gravity acceleration i s equaled o r exceeded i n
1,000 flights; t h e abscissa i s the center-of-gravity incremental accel-
eration. The ordinate scale can be ratioed i n d i r e c t proportion t o t h e -b

254 L
 7

number of f l i g h t s considered. I n calculating these curves, t h e approach

used was f i r s t t o construct a cumulative frequency d i s t r i b u t i o n of a l l
t h e individual incremental load peaks, both positive and negative, based
on a very d e t a i l e d analysis of a , r e l a t i v e l y few VGH records; these results
were ratioed up t o represent 1,000 flights, and a r e shown by t h e upper
curve. Then, on the basis of data f o r 7,000 f l i g h t s , a cumulative fre-
quency d i s t r i b u t i o n was obtained f o r t h e maximum t a x i i n g load i n each
f l i g h t . This d i s t r i b u t i o n of maximum loads was a l s o r a t i o e d t o represent
1,000 f l i g h t s and i s shown by t h e lower curve. I n the l i m i t , the two
curves should coincide a t a cumulative frequency of 1. I n view of the
nature of t h e assumptions involved and the fact t h a t the two d i s t r i b u -
t i o n s were obtained independently, t h e convergence appears t o be quite
good.

If the predominant frequency of t h e input t o t h e airplane i s assumed

t o be 1.75 cps, one can make a straightforward response calculation f o r
steady-state forced vibration and, thus, determine t h e v a r i a t i o n of t h e
acceleration along t h e span, which can then be compared w i t h t h e accelera-
t i o n a t t h e center of gravity. The r a t i o s r e s u l t i n g from such a compari-
son would permit converting t h e abscissa scale i n f i g u r e 8 from center-
of-gravity acceleration t o acceleration a t any point along t h e span.

As an example, such a response calculation has been made f o r t h e

B-29 airplane; two f l e x i b l e symmetrical bending modes were considered
and a steady-state sinusoidal input force a t 1.75 cps was assumed t o a c t
a t t h e landing-gear attachment point. Figure 9 shows a p l o t of the accel-
e r a t i o n response of t h e B-29 as a function of position along t h e span.
The dashed l i n e i s t h e rigid-body response; the s o l i d l i n e i s the t o t a l
dynamic response. For the input frequency of 1.75 cps, the r a t i o of
incremental wing-tip acceleration t o incremental center-of -gravity accel-
e r a t i o n i s about 7.9. It i s of i n t e r e s t t o note t h a t data obtained i n
t a x i i n g t e s t s of a B-29 a t t h e W r i g h t A i r Development Center showed
acceleration r a t i o s of the same order. However, there are s t i l l some
aspects of t h e problem which a r e not f u l l y understood; f o r example, even
though the B-29 records showed frequencies of about 1.5 t o 2.0 cps a t
t h e center of gravity, the frequencies a t t h e wing t i p were, i n sane
cases, considerably higher.

Braking Loads

The last topic t o be discussed i s t h e subject of braking f r i c t i o n .

Recent NACA t e s t s w i t h a s p e c i a l tow c a r t and with a C-123B airplane
have provided the data shown i n figure 10. Here the maximum coefficient
of f r i c t i o n i n braking on dry concrete i s p l o t t e d against the horizontal
velocity. The c i r c l e s are data obtained with t h e C-l23B airplane; t h e
squares are results obtained with t h e tow c a r t a t lower speeds. A l t h o u g h
t h i s type of information can be useful i n t h e calculation of maxhm

2 ,:5
 '

braking loads, no s t a t i s t i c a l data are as yet available regarding p i l o t

operating p r a c t i c e i n applying t h e brakes, t h a t is, t h e magnitude and
frequency d i s t r i b u t i o n s of braking loads.

CONCLUDING REMARKS

The foregoing material swmnarizes t h e more important NACA data

applicable t o t h e calculation of repeated ground loads. Additional work
w i l l be necessary t o extend t h e s i z e and scope of t h e s t a t i s t i c a l samples
arid t o f i l l i n t h e several a n a l y t i c a l gaps which s t i l l e x i s t .

256
 BIBLJ:OGRAPHY

S t a t i s t i c a l Data

Silsby, Norman S., Rind, Emanuel, and Morris, Garland J.: Some Measure-
ments of Landing Contact Conditions of Transport Airplanes i n Routine
Operations. NACA RM L33E05a, 1953.

Silsby, Norman S. : S t a t i s t i c a l Measurements of Contact Conditions of

478 Transport-Airplane Landings During Routine Daytime Operations.
NACA Rep. 1214, 1953. (Supersedes NACA TN 3194.)

Silsby, Norman S., and Harrin, Eziaslav N.: S t a t i s t i c a l Measurements

of Landing-Contact Conditions of a Heavy Bomber. NACA RM L55EO3, 1 9 5 .

Silsby, Norman S., and Harrin, Eziaslav N.: Ianding Conditions f o r Large
Airplanes i n Routine Operations. NACA RM LfS>E18c, 1955.

Kolnick, Joseph J., and Morris, Garland J.: S t a t i s t i c a l Measurements of

Landing Contact Conditions of t h e Boeing B-47 Airplane. NACA RM L55H24,
1955
Lindquist, Dean C.: A S t a t i s t i c a l Study of Wing L i f t at Ground Contact
f o r Four Transport Airplanes. NACA TN 3435, 1975.

Silsby, Norman S. : S t a t i s t i c a l Measurements of Landing Contact Conditions

of Five M i l i t a r y Airplanes During Routine Daytime Operations. NACA
RM L56F21a, 1956.

Harrin, Eziaslav N. : Comparison of Landing-Impact Velocities of F i r s t

and Second Wheel To Contact From S t a t i s t i c a l Measurements of Transport
Airplane Landings. NACA TN 3610, 1956.

Dreher, Robert: A Method f o r Obtaining S t a t i s t i c a l Data on Airplane

V e r t i c a l Velocity a t Ground Contact From Measurements of Center-of -
Gravity Accelerations . NACA TN 3541, 1956 .
Loads and Response

Milwitzky, Benjamin, and Cook, Francis E. : Analysis of Landing-Gear

Behavior. NACA Rep. 22-54, 1953. (Supersedes NACA TN 2735.)

Cook, Francis E., and Milwitzky, Benjamin: Effect of I n t e r a c t i o n on

Landing-Gear Behavior and Dynamic Loads i n a F l e x i b l e Airplane
Structure. NACA Rep. 1278, 1956. (Supersedes NACA TN 3467.)
* . * ..n * =c m. w 0 mm' a- ,1*

10

Milwitzky, Benjamin, Lindquist, Dean C., and Potter, Dexter M.: An

Experimental Study of Applied Ground Loads i n Landing. NACA Rep. 1248,
1955. (Supersedes NACA TN 3246.)
Sawyer, Richard H., H a l l , Albert W., and McKay, James M.: Vertical and
Drag Ground-Reaction Forces Developed i n Landing Impacts of a Large
Airplane. NACA RM L55E12c, 1955.

Milwitzky, Benjamin, and Lindquist, Dean C.: Evaluation of t h e Reduced-

Mass Method of Representing Wing-Lift Effects i n Free-Fall Drop Tests
of Landing Gears. NACA TN 2400, 1951.

Yntema, Robert To, and Milwitzky, Benjamin: An Impulse-Momentum Method

f o r Calculating Landing-Gear Contact Conditions i n Eccentric Landings.
NACA TN 2596, 1952.

Walls, James H., Houbolt, John C., and Press, Harry: Some Measurements
and Power Spectra of Runway Roughness. NACA TN 3303, 1954.

Houbolt, John C., Walls, James H., and Smiley, Robert F.: On Spectral
Analysis of Runway Roughness and Loads Developed During Taxiing. NACA
TN 3484, 1953.
 VERTICAL VELOCITY FOR CIVIL AND MILITARY AIRPLANES

PROBABILITY ,0-2 -

0 I 2 3 4 5 6 7 8 9 IO
vv, FPS

Figure 1

AIRSPEED AT CONTACT FOR CIVIL AND MILITARY AIRPLANES

MILITARY
I, 347 LANDINGS
PROBABILITY

10-4 -
I I I I I I I I I I
0 20 40 60 80 IO0
AIRSPEED AT CONTACT, % ABOVE STALL

Figure 2
 BANK ANGLES FOR SMALL AND LARGE AIRPLANES

PROBABILITY

257 LANDINGS

10-4 -
I I I I I I 1 I

0 2 4 6 8
BANK ANGLE, DEG

Figure 3

ROLLING VELOCITY FOR SMALL AND LARGE AIRPLANES

I
4-AND MORE ENGINE
CIVIL AND MILITARY
153 LANDINGS

I-AND 2-ENGINE CIVIL

AND MILITARY
686 LANDINGS

I I I I I I I I
0 4 8 12
ROLLING VELOCITY, DEG/SEC ,TOWARD
FIRST WHEEL TO CONTACT

Figure 4
 --. . e l -

CALCULATED AND EXPERIMENTAL MAXIMUM VERTICAL

LOADS IN FLIGHT TESTS OF S N J AIRPLANE

0 I 2 3 4 5 6 7 8 9 10
V, , FPS

Figure 5

CALCULATED AND EXPERIMENTAL MAXIMUM VERTICAL LOADS

I N FLIGHT TESTS OF 6-29 AIRPLANE
1.8
/
I’

.a’
. /’/
/
0

I .2
CALCULATED #’
5.6

I I I I 1 I I J
0 1 2 3 4 ’ 5 6 7 8 9
V , ,FPS

Figure 6
3 1 e ?em em e* e c e mea =e r 7 .

14

COEFFICIENT OF FRICTION DURING I-UP

PTI RE *
LB/SQ IN.

~~ 1 IMPACT-BASIN TESTS ,SNJ ---------32

FLIGHT TESTS, B -29 ------------- 7 5

I ,
I.Or

I
0 40 80 I20 160 200 240 280
SKIDDING VELOCITY, FPS

Figure 7

u
DISTRIBUTIONS OF TAXIING LOADS PER 1,000 FLIGHTS

‘“n
lo5

lo4
TOTAL ACCELERATIONS

CUMULATIVE lo3
FREQUENCY

IO2

IO
MAXIMUMS

Figure 8
 RESPONSE OF 8-29 AIRPLANE
f = 1.75 C PS

ACCELERATION -
RESPONSE FACTOR
5 r

1 I I II I I I I I I
0 I70 340 510 680 850
SPANWISE DISTANCE,IN.

BRAKING TESTS ON DRY CONCRETE

0 TOW CART
0 C-1238 AIRPLANE
1.0 -
0

.E -

.6 -
MAX. COEFF. 0
OF FRICTION,
PMAX.4 -

.2 -

I I I I I I J
40 60 80 100 120 140 160
HORIZONTAL VELOCITY, FPS

Figure 10 263
. c m.l I r . ' , 1 1.1 . -1

STRUCTURES
 .
75399 - -
AERODYNAMIC HEAT ?IRANSFEB I;o WING SURFACES

AND WING LEADING EDGES

By Aleck C. Bond, W i l l i a m V. Feller,

and W i l l i a m M. Bland, Jr.

Langley Aeronautical Laboratory

A canpilation i s presented of sane recent r e s u l t s obtained a t various

f a c i l i t i e s of the Langley Aeronautical Laboratory on the heat t r a n s f e r t o
wing surfaces and wing leading edges. Data obtained f’ran hypersonic
tunnel t e s t s , rocket-powered-model f l i g h t t e s t s , and high-stagnation-
temperature j e t t e s t s are included and canpared with applicable theories.
Measured heat t r a n s f e r t o wing surfaces exposed t o high-heat fluxes, i n
general, showed good agreement w i t h theory. Theory w a s a l s o shown t o be
adequate i n predicting the heat t r a n s f e r t o wing surfaces Over a range
of l o c a l Reynolds numbers fran 1 x lo6 t o 20 X 106. Heat-transfer meas-
urements on a blunted s l a b wing a t a Mach number of 6.86 showed t h a t
surface heating a t zero angle of attack was e s s e n t i a l l y t h e sane f o r
sweep angles of Oo, 40°, and 600; also, increasing t h e angle of a t t a c k
of the slab wing increased the heating of t h e windward surface and
decreased t h e heating of t h e leeward wing surface. Measurements a t high
Reynolds numbers showed t h a t r a t e s of heat t r a n s f e r that a r e much higher
than laminar r a t e s can be experienced on leading edges in the region of
t h e wing-body juncture. Transpiration-cooling t e s t s on a wedge surface
showed t h a t t h e o r e t i c a l predictions on t h e effectiveness of nitrogen as
a t r a n s p i r a t i o n coolant apply equally as well for conditions of e i t h e r
high or low heating potentials. Further cooling t e s t s showed t h a t helium
i s about f i v e times more e f f e c t i v e as a t r a n s p i r a t i o n coolant than
nitrogen.

INTRODUCTION

Considerable emphasis has recently been placed on the experimental

study of t h e aerodynamic heat t r a n s f e r t o wing surfaces and wing leading
edges. The wing leading edge has received a large amount of considera-
t i o n and has been t r e a t e d both t h e o r e t i c a l l y and experimentally i n r e f e r -
ences 1 t o 4. A q u a l i t a t i v e type of investigation on wing leading edges
a t conditions of high stagnation temperature has been reported i n r e f e r -
ence 5. Investigations a r e reported i n references 6 t o 8 of the
c o
F

r
e
.

i
ee I eo oo . *-. -- ..I

aerodynamic heating of wing surfaces a t angles of a t t a c k a t low super-

sonic speeds; however, data on the e f f e c t s of leading-edge bluntness,
sweep, and angle of attack, p a r t i c u l a r l y a t hypersonic speeds, as well
as the e f f e c t s of high heating p o t e n t i a l s and high Reynolds numbers on
wing-surface heating a r e indeed lacking i n t h e current l i t e r a t u r e .

Results of experiments conducted a t various f a c i l i t i e s of t h e

Langley Aeronautical Laboratory have recently becane available which
provide some extension of the s t a t e of t h e art regarding t h e heat
t r a n s f e r t o wing leading edges and wing surfaces. Rocket-powered-model
f l i g h t t e s t s have provided some large-scale measurements of the heat
t r a n s f e r t o wing surfaces as well as of the heating of blunt leading
edges i n the v i c i n i t y of t h e wing-body juncture. Hypersonic wind-tunnel
t e s t s have provided information on both the e f f e c t s of sweep md m g l e
of a t t a c k on the heat t r a n s f e r t o a wing having large leading-edge blunt-
ness. Tests i n a j e t having a high stagnation temperature have a l s o
provided r e s u l t s on the heating of wings under conditions of high r a t e s
of heat transfer. Furthermore, data on t h e effectiveness of transpira-
t i o n cooling of f l a t surfaces, under conditions of high stagnation tem-
perature, have a l s o recently become available *can tests i n a hot-jet
f a c i l i t y . It is, therefore, the purpose of t h i s paper t o present a
canpilation of these recent r e s u l t s which a r e applicable t o the high-
speed wing-heat ing problem.

SYMBOLS

F injection parmeter,
P 2V2

M Mach number

Stanton number based on l o c a l conditions, h

Nst
Cp2PZVZ

h
Stanton number based on free-stream conditions,
%t,m Cp,PmVm

R Reynolds number, -
PVS
IJ

Rd Reynolds number based on diameter, -

PmVCQd
IJW

T temperature, OF or ?R
v velocity, f t / s e c
Btu/slug
specific heat of a i r a t constant pressure,
OF
d diameter of c y l i n d r i c a l leading edge, f t
BtU
h l o c a l aerodynamic heat-transfer coefficient,
(sq f t )(set> (9)

9 dynamic pressure, lb/sq f t

S length from stagnation point t o measurement s t a t i o n , as indicated

when used, f t

A sweep angle, deg

U angle of attack, aeg

P v i s c o s i t y of a i r , lb-sec/sq f t

P density, slugs/cu f t

Subscripts :

AW adiabatic w a l l

C coolant

EXP experimental

2 conditions outside boundary layer

0 zero coolant flow r a t e

TH theoretical

t t o t a l conditions

W conditions pertaining t o skin of model

03 f r e e -stream conditions

267
1 -0 0. 0. 0 6 .*. .
I m .

RESULTS AND DISCUSSION

Heat Transfer a t H i g h Stagnation Temperature

In order t o study heating problems associated w i t h high-heat f l u e s ,

one of the f r e e jets of t h e Langley p r e f l i g h t j e t of t h e P i l o t l e s s
Aircraft Research Station a t Wallops Island, Va., has been modified t o
allow t e s t i n g of r e l a t i v e l y large t e s t specimens a t stagnation tempera-
t u r e s up t o about 3,fsoOO R which corresponds t o a Mach number of about
6.5 a t a l t i t u d e . The t e s t Mach number is maintained the same (M = 2.0);
however, t h e stream temperature is increased by burning ethylene gas
upstream of the e x i t nozzle. This mode of t e s t i n g provides heating rates
corresponding t o Mach numbers much higher than the stream Mach number
i n a t e s t medium not much d i f f e r e n t fram a i r . Recent t e s t s w i t h wings
subjected t o high heating r a t e s have yielded heat-transfer data which
show good agreement w i t h theory. Results f o r a wing t e s t e d a t a stag-
nation temperature of 3,476' R, a stream Mach number of 2.0, and a stream
Reynolds number of 2.4 X lo6 per foot a r e shown i n figures 1 and 2. The
wing w a s a tapered unswept wing of hexagonal cross section, w a s constructed
of magnesium, and had a span of 11 inches. The leading edge which was
blunted t o a 1,/16-inc% radius was covered with a 1/32-inch Inconel cap
t o increase i t s endurance a t t h e t e s t temperature. Sweep of t h e leading
edge was 17O. A t the 63-percent spanwise s t a t i o n (coinciding with t h e cen-
t e r l i n e of t h e j e t ) , temperature measurements were made a t f o u r chordwise
locations: t h e stagnation-point and t h e 1-, 2-, and 3.77-inch s t a t i o n s .
Temperature time h i s t o r i e s measured on the wing a r e shown i n f i g u r e 1.
The temperature a t t h e stagnation point rises rapidly and i n approximately
2.1 seconds (time of f a i l u r e of the leading-edge cap) has reached 2,731' R,
j u s t about 22g0 R below the melting temperature of Inconel. The tempera-
t u r e s a t the t h r e e reaxward s t a t i o n s show a maximum rise of about one-half
the stagnation-point value. It should be noted, however, t h a t t h e temper-
a t u r e of the 2-inch s t a t i o n , which is immediately behind the end of t h e
leading-edge cap, i s generally about X)Oo higher than t h e temperatures
of the other two s t a t i o n s during most of the t e s t .

In figure 2 these wing heating data are reduced t o t h e nondimensional

Stanton number evaluated a t l o c a l conditions and a r e a l s o plotted as a
function of time. In t h e p l o t a t the upper l e f t , t h e stagnation-point
d a t a a r e cmpmed w i t h t h e average laminar stagnation-point theory of
Goodwin, Creager, and Winkler ( r e f . 4). C a p a r i s o n w i t h t h e average
t h e o r e t i c a l stagnation-point values r a t h e r than w i t h l o c a l values is made
since it w a s f e l t that, because of t h e small physical s i z e of t h e leading
edge, t h e measurements a t t h i s point more nearly represented average values
r a t h e r than l o c a l point values. The f i r s t data point shows good agree-
ment w i t h the theory; however, w i t h increasing time, the data show
increasing deviation fran t h e theory. This disagreement i s believed t o
be due mainly t o l.arge l a t e r a l conduction losses fram t h e stagnation
F

point as t h e leading edge increases i n temperature, since it is known

that t h e temperature gradient across t h e j e t i s n o t uniform b u t tends
t o peak i n t h e v i c i n i t y of t h e j e t center l i n e . The data f o r t h e other
t h r e e s t a t i o n s are compared w i t h the f l a t - p l a t e t h e o r i e s of Van Driest
( r e f s . 9, 10, and 11) and, i n general, show good agreement. The theo-
r e t i c a l curves were determined by using calculated l o c a l condi!!xLons and
a Reynolds number length equal t o the streamwise distance f r c x n t h e stag-
nation point t o t h e measurement s t a t i o n . These data show that the flow
was i n i t i a l l y laminar a t t h e 1-inch s t a t i o n and t h a t t r a n s i t i o n took
place between t h e 1- and 2-inch s t a t i o n s and then gradually moved forward
w i t h time. I n i t i a l l y , t r a n s i t i o n could have been caused by t h e abrupt
discontinuity a t t h e end of the leading-edge cap, and the forward move-
ment w i t h time may have been due t o t h e increased temperature (i.e.,
temperature r a t i o ) of t h e surfaces w i t h time. It i s obvious t h a t , had
t h e t r a n s i t i o n point not moved forward, a serious hot spot would have
developed a t or n e a r . t h e 2-inch measurement s t a t i o n . The ccxnparison of
t h e data f o r t h e three downstream s t a t i o n s w i - those of t h e stagnation
point show t h a t t h e turbulent l e v e l of heating of t h e downstream s t a t i o n s
w a s of t h e order of one-half t h a t of t h e stagnation point.

It might be mentioned a t t h i s time t h a t , i n t h e reduction of t h e

heating d a t a f o r t h e 1-inch s t a t i o n t o l o c a l Stanton number, t h e l o c a l
conditions were determined by considering t h e l o s s e s through t h e normal
shock caused by t h e blunted leading edge. This procedure gave much
b e t t e r agreement w i t h the theory than w a s obtained when t h e l o s s e s w e r e
neglected. For t h e two s t a t i o n s downstream, t h e agreement between theory
and experiment was not enhanced when taking these l o s s e s i n t o account;
thus, t h e e f f e c t s of t h i s small amount of bluntness did not propagate
very far downstream i n t h i s case.

Large-Scale Heat-Transfer Measurements

In recent rocket-powered-model f l i g h t t e s t s , sane r e l a t i v e l y large-

s c a l e heat-transfer measurements have been obtained on two d i f f e r e n t
wing configurations a t free-stream Reynolds numbers up t o approximately
27 X lo6 based on t h e wing chord. This Reynolds number is equivalent
t o that of a wing of 120-inch chord f l y i n g a t an a l t i t u d e of 50,000 f e e t
and a Mach number of 2.5. I n order t o show t h e general agreement which
w a s obtained w i t h theory over t h i s l a r g e Reynolds number range, repre-
s e n t a t i v e data from these t e s t s a r e shown i n figure 3 as t h e r a t i o of
experimental t o t h e o r e t i c a l Stanton number as a function of t h e l o c a l
Reynolds number. It m i g h t be mentioned a t t h i s time that t h e r e w a s no
indication of laminar heating on e i t h e r of t h e wings throughout the
usable portion of t h e f l i g h t tests; however, a t t h e beginning of t h e
t e s t s there was some indication of t r a n s i t i o n a l flow a t t h e most forward
measurement s t a t i o n s . The t h e o r e t i c a l Stanton numbers were, therefore,
evaluated from t h e turbulent f l a t - p l a t e theory of Van Driest (refs. 10
6

and 11)by using calculated l o c a l conditions and a value of 0.6 f o r t h e

r a t i o of Stanton number t o skin-friction coefficient.

One of t h e t e s t wings was unswept, untapered, of approximately

20-inch chord and span, employed a ?-percent c i r c u l a r a i r f o i l section,
and had f o r a l l p r a c t i c a l purposes a sharp leading edge. The other
wing, t h e heating d a t a of which a r e reported i n reference 12, had a
leading-edge sweep of 30°, a span of 25 inches, and employed a hexagonal
a i r f o i l section w i t h a leading-edge radius of approximately 1/8 inch.
The data f o r t h e unswept wing a r e shown by t h e c i r c u l a r symbols and
represent measurements a t seven chordwise s t a t i o n s (35.7-percent. span)
and a free-stream Mach number range f r m 1.75 t o 2.66. The data f o r t h e
swept wing, indicated by t h e square symbols, were obtained a t f i v e chord-
wise s t a t i o n s (42-percent span) and a r e f o r a free-stream Mach number
range from 2.0 t o 3.64. Local Reynolds numbers up t o about 20 X 106 were
obtained on t h e unswept wing; however, on t h e swept wing the maximum
l o c a l Reynolds nunber iSLonly- of t h e order of 10 X 106 . Even though t h e
stream Reynolds numbers of t h e t w o wings were of t h e same order of magni-
tude, t h e l o c a l Reynolds numbers f o r t h e swept wing were lower as a r e s u l t
of i t s blunted leading edge.

In t h e lower Reynolds number range t h e agreement w i t h theory i s

generally within k l 5 percent, with t h e majority of t h e points showing
agreement within k10 percent. A t t h e higher Reynolds numbers t h e d a t a
show agreement within about 20 percent of t h e theory. The decreasing
trenii w i t h increasing Reynolds number which is exhibited by t h e data may
be due t o t h e reduction i n t h e r a t i o of Stanton number t o skin-friction
c o e f f i c i e n t w i t h increasing Reynolds number which w a s observed by Seiff
i n reference 13. Since t h e present data were obtained i n the presence
of temperature, pressure, and Mach number variations, it i s f e l t t h a t no
d e f i n i t e conclusions can be drawn on t h i s point. The important f e a t u r e
of these d a t a i s t h a t reasonable agreement w i t h theory is shown f o r two
t o t a l l y d i f f e r e n t wing configurations over a rather wide variation of
Reynolds number.

Heat Transfer a t Angle of Attack and Sweep

Theory and experiment have shown t h a t both leading-edge sweep and

bluntness have t h e general e f f e c t of reducing t h e l o c a l r a t e s of heat
t r a n s f e r t o t h e leading edge. In order t o study t h e e f f e c t s of sweep
as well as t h e e f f e c t s of angle of a t t a c k on the heating of t h e surfaces
of a wing w i t h a blunt leading edge, t e s t s have been conducted i n the
Langley 11-inch hypersonic tunnel a t a Mach number of 6086 on a s l a b wing
w i t h a s a i c i r c u l a r leading edge. In figure 4 a r e presented some of t h e
r e s u l t s of these t e s t s which show t h e e f f e c t of sweep on t h e wing-surface
heat t r a n s f e r a t zero angle of attack. The data a r e f o r a stream Reynolds

i
 7

number of 0.151 x 106 based on the 3/4-inch diameter of t h e leading edge

and correspond t o an a c t u a l f l i g h t a l t i t u d e of about 76,000 f e e t at t h e
t e s t Mach number. The data a r e plotted as Stanton number based on f r e e -
stream conditions as a function of s/d, where s i s t h e surface length,
measured i n a streamwise plane, froan t h e stagnation point t o the measure-
ment s t a t i o n , and d is t h e leading-edge diameter. Data a r e presented
f o r sweep angles of Oo, 40°, and 60'. In order t o show the r e l a t i v e
magnitude between the leading edge and surface heating, theory curves f o r
t h e cylinder a r e presented f o r each of the sweep angles; however, since
heat t r a n s f e r t o swept cylinders has been presented i n the p a s t ( r e f s . 3
and 4 ) , a t t e n t i o n i s c a l l e d t o the f l a t - p l a t e portion of t h e wing. The
measured data f o r a l l sweep angles l i e e s s e n t i a l l y on one l i n e and show
no e f f e c t of sweep angle on t h e surface heating. The theory curve
included in figure 4 w a s ccaqputed by using the Van Driest theory f o r
laminar boundary layer (ref. 9 ) w i t h l o c a l conditions determined by taking
i n t o account the losses through the normal shock and by assuming t h e
l o c a l s t a t i c pressure on t h e surface equal t o stream s t a t i c pressure. In
t h e v i c i n i t y of t h e leading edge (s/d = 1 t o 2 ) t h e data a r e considerably
higher than t h i s t h e o r e t i c a l calculation, mainly as a r e s u l t of t h e
higher-than-stream pressures which a r e known t o e x i s t there; but as t h e
pressure f a l l s off w i t h distance downstream, t h e data show b e t t e r agree-
ment w i t h t h e theory. Measured pressure d i s t r i b u t i o n s on the wing
showed t h a t , j u s t behind t h e leading edge, the surface pressures were
several times higher than stream s t a t i c pressure and dropped off t o
about twice the stream s t a t i c pressure a t values of s/d of about 7.5.
The theory curve shown i n the f i g u r e was a c t u a l l y computed f o r zero
sweep, but it i s Tnteresting t o note t h a t calculations f o r the 40° and
60' sweep angles showed only small variation from t h e '0 sweep condition.

Measured surface heat t r a n s f e r a t angle of a t t a c k f o r t h e 40° swept

s l a b wing is presented in figure 5 as being generally t y p i c a l of t h e
e f f e c t of angle of a t t a c k on t h e wing surface heating f o r a l l t h e sweep
angles. Here, again the data a r e p l o t t e d as Stanton number based on
free-stream conditions as a function of streamwise s/d. Measurements
a r e presented of surface heating of t h e lower or windward surface w i t h
t h e wing a t angles of a t t a c k of 5' and 7.5' and of t h e upper or lee-
w a r d surface with t h e wing a t angles of a t t a c k of 5 O and loo. The curve
of zero-angle-of-attack data, of course, applies t o both upper and lower
surfaces of the wing. Increasing the angle of a t t a c k increases t h e
heating of the lower surface and decreases t h e heating of t h e upper sur-
face. The v a r i a t i o n of heating with angle of a t t a c k i s e s s e n t i a l l y
l i n e a r f o r both t h e upper and lower surfaces; however, t h e r a t e of
increase of heating on t h e lower surface i s about twice the r a t e of
decrease of heating on t h e upper surface.
 1
.
6
e .e.
.e

a
Heating of Wing Leading Edge in Vicinity

of Wing-Body Juncture

The heat t r a n s f e r to swept cylinders i n an undisturbed flow f i e l d

has been studied over a wide range of conditions and, in general, t h e
experimental d a t a agree well w i t h existing theories. A rocket-powered-
model f l i g h t t e s t was recently conducted f o r t h e purpose of measuring
t h e heat t r a n s f e r t o leading edges i n the v i c i n i t y of t h e wing-body
juncture a t r e l a t i v e l y high Reynolds numbers. A sketch i s shown i n f i g -
ure 6 of t h e rocket-powered model which carried small stubs representing
only t h e leading-edge portions of Oo and 750 swept wings. The stubs had
c y l i n d r i c a l leading edges of 3/4-inch diameter which became tangent t o
f l a t surfaces inclined a t 4 . 3 O t o t h e chord plane. Thermocouple measure-
ments were made i n a d i r e c t i o n perpendicular t o the leading edge of each
stub a t locations shown in the cross-section sketch. Four measurement
p o i n t s were obtained on t h e Oo swept stub; however, on t h e 75O swept
stub, t h e most rearward thermocouple f a i l e d t o operate properly and,
hence, measurements a t only t h e three forward locations were obtained.

Measured heating d a t a f o r the two leading edges a t a stream Mach

number of 3.12 and a stream Reynolds number of 18.7 X lo6 per f o o t a r e
presented i n f i g u r e 7 along w i t h appropriate t h e o r e t i c a l curves f o r c m -
parison. The data a r e presented as t h e nondimensional Stanton number
evaluated a t free-stream conditions as a function of s/d where s i s
t h e surface length, measured i n a plane normal t o t h e leading edge, f r m
t h e stagnation point t o the measurement s t a t i o n and d is the leading-
edge diameter. The measured data on t h e c y l i n d r i c a l portion of the
leading edge f o r both t h e 0' and 75' swept stubs a r e considerably higher
than t h a t predicted by t h e laminar leading-edge theory i n reference 4.
On t h e Oo swept leading edge, the heating a t the stagnation point i s of
the order of 2 times t h a t which i s predicted by laminar theory, and on
t h e 7 5 O swept leading edge, t h e heating i s of t h e order of 4.5 times a s
great. Since t h e r e i s such a large difference between t h e measured
values and t h e laminar theory f o r t h e c y l i n d r i c a l portion of the leading
edges, it was f e l t t h a t t h e f l o w over t h i s portion must be turbulent.
This w a s v e r i f i e d f o r the 0' swept leading edge by integrating the meas-
ured values over t h e c y l i n d r i c a l portion t o obtain the average heat
,* = 34.96 >< 10-4) was found t o agree
t r a n s f e r . This average value (Nst
very well w i t h t h e t h e o r e t i c a l average (NSt
,* = 34.85 x lo4) cmputed
from t h e turbulent theory i n reference 3. Comparison of t h e measured
d a t a on t h e f l a t portion following t h e cylinder w i t h t h e Van Driest
turbulent f l a t - p l a t e theory (refs.. 10 and 11) a l s o shows good agreement
and tends t o e s t a b l i s h f u r t h e r t h a t t h e flow on t h e Oo swept leading
edge w a s turbulent. The analysis f o r turbulent heating on the 75O swept
leading edge w a s not c a r r i e d out because of t h e absence of information
on the l o c a l flow conditions a t t h i s sweep angle and Mach number.
 9

The reason f o r turbulent f l o w over t h e Oo swept leading edge cannot

be deduced f r a n the measurements made on t h e model. Since the measure-
ments were made a t a distance of only 11 inches fra the body, t h e heating
4
t o t h e leading edge could have been influenced by conditions e x i s t i n g i n
t h e body boundary layer. The Reynolds number f o r t h e body a t t h e leading-
edge body juncture w a s approximately 66 x lo6 and t h e body boundasy-layer
thickness was estimated t o be of t h e order of 1/2 inch. With these con-
d i t i o n s prevailing, it is possible t h a t interaction between t h e bow shock
ahead of t h e leading edge and the t h i c k turbulent boundary layer of t h e
body could have increased t h e heating of t h e leading edge t o t h e turbu-
l e n t level. Although exact simulation of t h e flow f i e l d a t t h e wing-body
juncture may not have been provided by t h e short leading-edge stubs,
these d a t a indicate that r a t e s of heat t r a n s f e r much higher than l a m i n a r
r a t e s can be experienced on leading edges i n the region of t h e wing-body
juncture, and t h a t , further, more canplete investigations a r e needed t o
understand t h i s phenomenon.

Transpiration Cooling a t H i g h Stagnation Temperatures

A t lower f l i g h t speeds t h e problem of convective heating of an a i r -

c r a f t surface can generally be handled by designing the surface as a
heat sink and allowing t h e skin t o a t t a i n some allowable equilibrium
temperature. A s f l i g h t speeds increase, t h e idea of sane type of cooling
f o r t h e h o t t e r surfaces appears more a t t r a c t i v e . One type of cooling
which shows promise i s transpiration cooling, i n which the coolant passes
from t h e i n t e r i o r of t h e a i r c r a f t through a porous skin i n t o t h e hot
boundary layer. Experimental data on t r a n s p i r a t i o n cooling a t r e l a t i v e l y
low heating p o t e n t i a l s have been reported in references 14, 15, and 16,
and have shown appreciable reduction i n the convective heating w i t h t r a n s -
p i r a t i o n cooling. In order t o investigate the effectiveness of trans-
p i r a t i o n cooling a t higher heating p o t e n t i a l s , exploratory t e s t s have
been conducted a t high stagnation temperatures i n t h e Langley p r e f l i g h t
j e t of t h e P i l o t l e s s Aircraft Research Station a t Wallops Island, Va.
The model employed f o r t h e tests was a blunted wedge of 200 half-angle
which had a porous s t a i n l e s s - s t e e l segment inserted i n one surface through
which coolant was ejected. Tests were conclucted a t a ncminal free-stream
dymmic pressure of 5,000 pounds per square foot and a free-stream Mach
number of 2.0; however, t h e local-surface Mach number on t h e wedge was
of the order of 1.2. Results of these t e s t s showing the e f f e c t of trans-
p i r a t i o n cooling on t h e wedge-surface temperature f o r both nitrogen and
helium coolants a r e presented i n figure 8. The ordinate is t h e non-
dimensional wall-temperature pwameter
- TW Tc
where Q i s t h e
-
TAW,O Tc
porous-wall temperature, T, i s the coolant temperature, and TAW,O i s
t h e boundary-layer recovery temperature f o r zero coolant flow r a t e . The

273
abscissa i s t h e coolant flow r a t e i n pounds per square f o o t per minute.
Such a presentation of cooling data shows d i r e c t l y t h e reduction i n
w a l l temperature which can be achieved f o r a given coolant flow rate.
Data a r e presented f o r stagnation temperatures i n t h e range from 2,35'j0 R
t o 3,370° R f o r nitrogen coolant and frcm 1,755O R t o 3,195O R f o r helium
coolant. For t h e various tests, r a t i o s of w a l l temperature t o l o c a l
temperature ranged from 0.2 t o 0.5 and t h e l o c a l Reynolds numbers ranged
from 0.6 X lo6 t o 8.2 X 106. It m i g h t a l s o be mentioned t h a t t h e present
data were obtained f o r average operating temperatures of t h e porous w a l l
i n t h e range frm about 20O0 F t o 1,300° F. Comparison of t h e nitrogen
and helium data i n f i g u r e 8 shows t h a t the h e l i m performs as a much
more e f f e c t i v e coolant than nitrogen, as would be expected because of
the higher s p e c i f i c heat of t h e helium. For example, i n order t o main-
t a i n t h e skin a t a t a p e r a t u r e of about 0.3 of t h e uncooled value,
approximately 10 pounds of nitrogen per square f o o t per minute would be
required as ccmpared w i t h 2 pounds per square f o o t per minute for helium.
It m i g h t be added t h a t t h i s r a t i o of 5 t o 1 i n required coolant flow
rates i s roughly t h e same as t h e r a t i o of t h e s p e c i f i c heat of helium
t o t h a t of nitrogen.

In order t o show how tZlese high-temperature data canpare w i t h theory

and other low-temperature data, as w e l l as t o show t h e e f f e c t of cooling
on t h e heat t r a n s f e r , t h e data were reduced t o heat-transfer c o e f f i c i e n t s
and are presented i n dimensionless form i n f i g u r e 9. The ordinate i s t h e
r a t i o of t h e Stanton number with cooling t o t h e Stanton number f o r zero
cooling, and t h e abscissa is t h e i n j e c t i o n parameter F (the r a t i o of
coolant weight flow t o l o c a l stream w e i g h t flow) divided by t h e Stanton
number f o r zero cooling. In the reduction of t h e measured data t o
Stanton nmber, t h e recovery temperature w i t h coolant flow w a s evaluated
by using recovery-factor values ccanputed from t h e theory of reference 17,
which gives t h e v a r i a t i o n of recovery f a c t o r w i t h Mach number, coolant
flow r a t e , and Reynolds number. Included f o r comparison are the theo-
r e t i c a l curve applicable t o nitrogen from t h e theory i n reference 18
f o r M = 1.0 and a r a t i o of w a l l temperature t o l o c a l temperature of
0.2, and t h e experimental data i n reference 14 f o r t h e t r a n s p i r a t i o n
cooling of an 8' cone w i t h both nitrogen and helium coolants a t R stag-
nation temperature of 1,060' R. The experimental wedge data f o r nitrogen
coolant form E band which has t h e same trend as t h e t h e o r j and shows
s l i g h t l y greater cooling effectiveness. The present data f o r t h e nitrogen
a l s o show good agreement w i t h t h e trend established by t h e 8' cone data.
The present he1i.w data do nut show as good agreement w i t h t n e p r i o r
cone data w i t h heliuni coolant as i s observed i n t h e case w i t h nitrogen.
Since both s e t s of data f o r t h e helium coolant a r e f o r s m a l l flow rates,
t h e discrepancy between t h e two s e t s of data may be due t o t h e l e s s e r
accuracy i n determining t h e flow r a t e s i n the lower range of flow rates.

With t h i s type of correlation, t h e canparison of t h e effectiveness

of helium with nitrogen as t r a n s p i r a t i o n coolants i s about i n t h e same

2;4
 11

r e l a t i o n as noted i n figure 8. The theory which i s s t r i c t l y derived f o r

t r a n s p i r a t i o n of a i r t o a i r but i s applicable a l s o t o nitrogen because
of t h e s i m i l a r i t y of t h e physical c h a r a c t e r i s t i c s may be said t o apply
equally as well f o r e i t h e r high or low heating potentials.

CONCLUSIONS

The pertinent conclusions which may be drawn f'ran the foregoing

canpilation of wing-heating data may be summarized b r i e f l y as follows:

1. Measured heat t r a n s f e r t c wing surfaces exposed t o high-heat

fluxes have, i n general, shown good agreement w i t h theory. Theory has
a l s o been shown t o be adeqmte i n predicting t h e heat t r a n s f e r t o wing
surfaces over a wide range of Reynolds number.

2. Heat-transfer measurements on a blunted s l a b wing e=ta Mach

number of 6.86 showed t h a t the surface heating a t zero angle of a t t a c k
was e s s e n t i a l l y t h e same for sweep angles of Oo, 40°, and 60°. Increasing
t h e angle of attack of t h e s l a b wing increased the heating of t h e wind-
ward surface and decreased t h e heating of t h e leeward surface.

3 . Measurements a t high stream Reynolds nmbers have shown t h a t

r a t e s 02 heat t r a n s f e r much higher than laminar r a t e s can be experienced
on leading edges i n t h e region of the wing-body juncture.

4. Theoretical predictions of t h e effectiveness of nitrogen as a

t r a n s p i r a t i o n coolant have been shown t o apply equally as w e l l f o r con-
d i t i o n s of e i t h e r high or low heating potentials. Measurements have
shown t h a t helium i s about f i v e times more e f f e c t i v e as a t r a n s p i r a t i o n
coolant than nitrogen.
12

1. Reshotko, E l i , and Cohen, Clarence B.: Heat Transfer a t the Forward

Stagnation Point of Blunt Bodies. NACA TN 3513, 1955.

2. Beckwith, Ivan E.: Theoretical Investigation of Laminar Heat Transfer

on Yawed I n f i n i t e Cylinders i n Supersonic Flow and a Comparison W i t h
Experimental Data. NACA RM L55FO9, 1955.

3. Beckwith, Ivan E., and Gallagher, James J.: Experimental Investigation

of t h e Effect of Boundary-Layer Transition on the Average Heat
Transfer t o a Yawed Cylinder i n Supersonic Flow. NACA RM L56E09,
1956
4. Goodwin, Glen, Creager, Marcus O., and Winkler, Ernest L.: Investiga-
t i o n of Local Heat-Transfer and Pressure Drag Characteristics of a
Yawed Circular Cylinder a t Supersonic Speeds. NACA RM A55H.31, 1956.

5. Bland, W i l l i a m M.,
Jr., and Bressette, Walter E.: Some Effects of
Heat Transfer a t Mach Number 2.0 a t Stagnation Temperatures Between
2,310' and 3,50O0 R on a Magnesium Fin With Several Leading-Edge
Modifications. NACA RM L57C14, 1957. (Prospective NACA Paper. )

6. Carter, Howard S.: Heat Transfer on t h e L i f t i n g Surfaces of a

60° Delta Wing a t Angle of Attack f o r Mach Number 1.98. NACA
m ~ 5 6 ~ 2 319%.
,
7. B a r t l e t t ,
G. E., Hilton, J. H., Vidal, R. J., and Woolard H. W.:
Experimental Investigations of Heat Transfer on a 10-Percent Double-
Wedge A i r f o i l a t Mach No. 2.0. Rep. No. CAL/CM-832 (Contract
NOrd-14>23), Cornell Aero. Lab., Inc., Jan. 1955.

8. Sterbutzel, Gerald A., and Kajencki, Stephen S.: Experimental Inves-

t i g a t i o n of Heat Transfer Frcxa Aerodynmic Bodies i n Supersonic
Flow. Rep. No. AF-473-A-9 (Contract 03-038-ac-16701), Cornell
Aero. Lab., Inc., Apr. 1950.

9. Van Driest, E. R.: Investigation of Laminar Boundary Layer i n

Compressible Fluids Using t h e Crocco Method. NACA TN 2597, 1952.

10. Van Driest, E. R.: The Turbulent Boundary Layer f o r Compressible

Fluids on a F l a t P l a t e With Heat Transfer. Rep. No. AL-997, North
American Aviation, Inc., Jan. 27, 19%.

11. V a n Driest, E. R.: The Turbulent Boundary Layer With Variable

Prandtl Number. Rep. No. AL-1914, North American Aviation, Inc.,
Apr. 2, 19%.

Akii?
 13

12. Swanson, Andrew G., and Rumsey, Charles B.: Aerodynamic Heating of a
Wing A s Determined From a Free-Flight Rocket-Model Test t o Mach
Number 3.64. NACA RM LfS6Flla, 1956.

13. S e i f f , Alvin:
Examination of t h e Existing Data on t h e H e a t Transfer
of Turbulent Boundary Layers a t Supersonic Speeds From t h e Point
of View of Reynolds Analogy. NACA TN 3284, 1954.

14. Chauvin, Leo T., and Carter, Howard S. : Ekploratory Tests of

Transpiration Cooling on a Porous 8 O Cone a t M = 2-03 Using
Nitrogen Gas, Helium Gas, and Water as t h e Coolants. NACA
RM L55C29, 1955-
15. Leadon, B. M., and Scott, C. J.: Measurement of Recovery Factors
and H e a t Transfer Coefficients W i t i n Transpiration Cooling i n a
Turbulent Boundary Layer a t M = 3 Using A i r and Helium as
Coolants. Res. Rep. No. 126, Univ. of Minnesota Inst. Tech., Dept.
Aero. Eng. (Contract AJ? 18(6oo)-1226), Feb. 1956.

16. Rubesin, Morris W., Pappas, Constantine C., and Okuno, Arthur F.:
The Effect of Fluid Injection on t h e Compressible Turbulent
Boundary Layer -
Preliminary Tests on Transpiration Cooling of
a F l a t Plate a t M = 2.7 With A i r as t h e Injected G a s . NACA
RM A551193 1955-

17. Rubesin, Morris W.: An Analytical Estimation of the Effect of

Transpiration Cooling on the Heat-Transfer and Skin-Friction
Characteristics of a Compressible, Turbulent Boundary Layer.
NACA TN 3341, 1954.

18. Dorrance, W i l l i a m H.,

and Dore, Frank J.: The Effect of Mass Transfer
on t h e Cmpressible Turbulent Boundary Lsyer Skin F r i c t i o n and
Heat Transfer. Rep. ZA-7-013, Convair, Aug. 5, 1954.
14

TIME HISTORIES OF WING TEMPERATURES

Mm.2; Tt= 3,476O R

STATION I" 3.77"

1
3,OOOr

2.500 1 1/32" INCONEL CAP

STAGNATION POlNT

0
I I
.5
I
I.o 15
I I
2.0
I
2.5
TIME, SEC

Figure 1

WING-SURFACE HEAT TRANSFER

Moo=2 ; T+ = 3,476O R ; R, = 2.4 X IO' PER FT
STAGNATION POINT
80 4!914LTHEORY
------
F
I" STATION
oooooo
60 0 rTURBULENT THEORY
0
n
POINT LAMINAR THE~RYO
20

0
1
80 X I O - ~
60 0
2" STATION
0
1 3.77" STATION

0 0
INCREASING TIME-

Figure 2
P
 COMPARISON OF EXPERIMENT AND THEORY FOR
LARGE-SCALE HEAT-TRANSFER TESTS

1.2r

I.o

.8

NSt, EXP .6 - 5"/0CIRCULAR ARC HEXAGONAL

NSt, TH
.4 -

.2 -
-
M,=1.75
....... 19.9"
To 2.66 Maz2.00 TO 3.64
I I I I I I

EFFECT OF SWEEP ON HEAT TRANSFER TO A BLUNT L.E. WING

a=OoiM.6.86; Rd=0.151~106

A ,DEG
0 0
n 40
A 60

- I
On
I I I
THEORY, L A M INAR
FLAT PLATE
A d
I I
Am
I
0 I 2 3 4 5 6 7 8
STREAMWISE s/d

Figure 4
\
16

EFFECT OF ANGLE OF ATTACK ON HEAT TRANSFER

TO A B L U N T L.E. WING

-4

JUNCTURE

30

0 2 4 6 0 IO
STREAMWISE S/d

Figure 5

LEADING-EDGE HEAT-TRANSFER MODEL

44
1-4
2
'.

MEASUREMENT STATION

LEADING-EDGE CROSS SECTION

Figure 6
 HEAT TRANSFER T O LEADING EDGE IN
VICINITY OF WING-BODY JUNCTURE
M OD =3.12; R
OD
=18.7X106 PERFT; Rd= 1.17x106

A =Oo A = 75'
50 6 x 1 0 - 4 c

40 - -
0

Nstp
0
--LAMINAR L.E.
I I
0 .4 .8 1.2 1.6 0 .4 .8 1.2
s/d s/d

Figure 7

EFFECT OF TRANSPIRATION COOLING ON

WEDGE-SURFACE TEMPERATURE
LB
q,SJ 5,000 SQFT i M a = 2
T+1 .OR

":;
-
NITROGEN HELIUM
0 3,370 0 3,195
' .8
*Or 0 2,921 b 2,591
0 2,355 A 1,.755

.2 0
"%' O.0 0
NITROGEN
HELIUM 043
98 0 0
0 A OOcQJ 4,
L I I I I
I I I I 1
0 2 4 6 8 IO 12 14 16 I8
LB/SQ FT
COOLANT FLOW RATE, MIN

Figure 8
281
 I

18

EFFECT OF TRANSPIRATION COOLING ON HEAT TRANSFER

T+ , O R

I.OK
-8kA
200 WEDGE

8" CONE
c NIT.ROGEN HELIUM
o
0
0

6
3,370
2,921
2,355
1,060
a 3,195
b 2,591
A 1,755
1,060

NITROGEN

DORRANCE AND DORE

HELIUM

0 .5 1.0 1.5 2.0 2.5 3.0 3.5

F/Nst,o

.
 .a
a.

TOTAL BEAT TRANSFER To BLUNT-NOSE SHAPES W I T H LAMINAR

BOUNDARY LAYERS AT HIGH SUPERSONIC SPEEDS

By John 0. Reller, Jr.

Arnes Aeronautical Laboratory

SUMMARY

A method of designing b l u n t shapes has been devised which proposes

t o reduce the heat t r a n s f e r t o a body by v i r t u e of low-velocity flow
over t h e nose and low-density f l o w over the afterbody. A t y p i c a l body
consists of a f l a t nose and a highly curved afterbody surface defined
by a modified Newtonian theory.

Tests w e r e conducted i n the Ames 10- by &inch supersonic wind

tunnel at Mach numbers from 3.0 t o 6.3 and it w a s found t h a t t o t a l heat-
t r a n s f e r rates f o r these flat-nose shapes i s less than t h a t of a cone
of about t h e same drag. Comparison of experimental r e s u l t s with theory
indicates higher than average heat-transfer rates near t h e shoulder of
a t y p i c a l shape and r e l a t i v e l y low values over most of t h e afterbody.

INTRODUCTION

It i s a well-known fact t h a t blunting can reduce t h e rate of heat

transfer t o the nose of a body i n supersonic f l o w . The problem is,
however, t o determine a type of blunting which tends t o minimize t h e
heat-transfer rate. The purpose of t h i s paper i s t o describe an inves-
t i g a t i o n of t h i s problem i n which it was undertaken first t o devise a
method of designing blunt shapes and then t o check by experiment the
effectiveness of these shapes i n reducing heat t r a n s f e r .

"he b a s i c heat-transfer equation, i n t h e form of Reynolds analogy,

i s shown at t h e top of figure 1. Although t h i s analogy i s not s t r i c t l y
applicable f o r blunt shapes, i n general t h e convective heat-transfer
c o e f f i c i e n t i s proportional t o the product of t h e l o c a l density, l o c a l
velocity, s p e c i f i c heat, and l o c a l skin-friction coefficient. The spe-
c i f i c heat i s a f a c t o r over which r e l a t i v e l y l i t t l e control can be
exerted. Likewise, c o n t r o l over the skin-friction coefficient i s limited
2

primarily by t h e extent t o which laminar boundary-layer f l o w can be

preserved. Attention i s therefore focused on the product of the l o c a l
density and t h e l o c a l v e l o c i t y . This product, and hence the l o c a l heat-
t r a n s f e r coefficient, m a y be kept low over a shape designed t o have low
v e l o c i t i e s at the nose and low d e n s i t i e s over t h e afterbody. Such a
shape would, i n i t s simplest form, be a truncated cone, inasmuch as t h e
f l a t nose minimizes l o c a l v e l o c i t i e s , while t h e highly inclined s i d e s
minimize l o c a l pressures and hence densities. The f a c t is, however,
t h a t the sharp corner a t the i n t e r s e c t i o n of the face and afterbody of
a truncated cone tends t o cause l o c a l separation and reattachment of
t h e flow with an attendant shock wave and unfavorable pressure gradient.
These conditions tend t o promote t r a n s i t i o n t o a turbulent boundary
layer, which increases t h e l o c a l heating r a t e . Therefore, the surface
i n t h i s shoulder region should be curved t o avoid l o c a l separation and
possible tripping of t h e boundary layer. I n order t o promote laminar
flow over t h e e n t i r e afterbody it i s desirable t o have a contour which
generates a continuously favorable pressure gradient. Such a contour
i s easily determined w i t h the modified Newtonian theory of Eggers,
Resnikoff, and Dennis ( r e f . l), providing the square of hypersonic s i m i -
l a r i t y parameter M i s large compared w i t h unity. "his condition i s
ID
satisfied by blunt bodies a t t h e high supersonic Mach numbers of t h i s
investigation. The r e s u l t a n t expression defining t h e shape of t h e a f t e r -
body surface i s shown as t h e second equation i n figure 1. Note t h a t t h e
coordinates iT and axe t h e l o c a l x and y dimensions, resp.ectively,
divided by t h e radius of the f l a t nose. The parameter K i n t h i s expres-
sion f i x e s t h e l e v e l and gradient of pressure on t h e afterbody.

Two families of blunt-nose bodies of revolution were designed

according t o t h i s equation. The f i r p t family i s shown a t the l e f t i n
f i g u r e 1 and consisted of eight bodies of the same fineness r a t i o , but
of varying diameters of t h e f l a t nose. For t h e second f a m i l y of bodies,
shown a t t h e r i g h t i n figure 1, the nose and base diameters were held
constant and the fineness r a t i o was varied f r m about 0.3 t o 1.6. The
corresponding values of K are shown on the superimposed sketches of
these bodies. The bodies of t h i s second family have approximately the
same pressure drag and for t h i s reason two reference bodies of similar
pressure drag w e r e included i n t h e t e s t program. These were t h e sharp-
pointed and hemispherical-tipped 60' cones shown i n figure 1. A f u l l
hemisphere of the same base diameter was also tested.

Tests were conducted i n the Ames 10- by &inch supersonic wind

tunnel at Mach numbers fram 3 t o 6 . 3 . As shown i n figure 2, models were
s t i n g supported fram the r e a r and were i n e f f e c t insulated t o prevent
 ., . ,.
b
r.

'
..
0 -
4
l: :
3

heat loss tm the support system by a guard heater which w a s used t o

equalize t h e temperature i n the support s h e l l . H e a t input was through
a resistance heater of Inconel wire wound on a copper spool. From the
known resistance and measured voltage t h e overall heat-transfer rate
could be determined f o r equilibrium conditions. This apparatus did not
permit t h e measurement of l o c a l heat-transfer rates. Models were m a d e
of aluminum and provided e s s e n t i a l l y constant-temperature heat-transfer
surfaces because of t h e i r high conductivity. Thermocouples were i n s t a l l e d
within t h e models t o measure temperatures near the outer surface. No
correction w a s made t o the measured data t o account f o r t h e heat flow
through t h e exposed portion of the m o d e l bases. This heat f l o w w a s esti-
mated t o be a small f r a c t i o n of t h e t o t a l and hence has been neglected
since t h e primary i n t e r e s t of t h i s investigation i s the r e l a t i v e e f f e c t s
of shape on heat t r a n s f e r .

EFFECT OF BLUNTNESS ON HEXT TRPLNSFER

The e f f e c t on heat t r a n s f e r of varying nose diameter w h i l e holding

fineness r a t i o constant i s shown i n f i g u r e 3 . Representative data a t zero
angle of attack and free-stream Mach number M .of 4.24 a r e presented i n
the form of nondFmensiona1 heat-transfer coefficients, t h a t is, Stanton
numbers, based on free-stream properties. A t t h e top of the figure,
Stanton numbers a r e referenced t o model base area and as such are a
d i r e c t comparison of the total heat-transfer rates. As nose diameter
i s increased frm zero, Stanton number first decreases s l i g h t l y , then
remains nearly constant up t o a diameter r a t i o of one-half, and there-
after increases substantially up t o a diameter r a t i o of 1. The Stanton
numbers presented i n the lower p a r t of figure 3 a r e referenced t o wetted
surface area, exclusive of t h e base, and hence are indicative of t h e
average heating r a t e per u n i t surface area. They decrease s i g n i f i c a n t l y
with increasing diameter r a t i o up t o about one-half and then remain
e s s e n t i a l l y constant. These r e s u l t s suggest t h a t a body w i t h a nose-
to-base diameter r a t i o of about one-half i s a good comprmise f o r low
values of both t o t a l and average heat t r a n s f e r . Furthermore, t h i s amount
of bluntness y i e l d s drags which are i n t h e range of p r a c t i c a l i n t e r e s t
f o r b a l l i s t i c m i s s i l e shapes. It was f o r t h i s reason that a diameter
r a t i o of 0.32 was used i n designing t h e second f a m i l y of bodies.

EFFECT OF FINENESS RATIO ON HEAT TRANSFEZ

To BLUNT 'BODIES

Total heat t r a n s f e r t o t h e bodies of d i f f e r e n t fineness r a t i o at

zero angle of attack i s shown i n figure 4 as a function of free-stream
4

Mach number. Camparative d a t a f o r a hemisphere are also presented.

Stanton numbers a r e referenced t o model base area. The upsweep of the
d a t a with increasing Mach number r e s u l t s frm the decrease of free-
stream Reynolds number R t y p i c a l of a wind tunnel operating w i t h a
fixed supply pressure. As determined frm shadowgraph pictures the
boundary layer w a s laminar over the flat-nose bodies at a l l t e s t condi-
t i o n s . This i s i n contrast t o the reference cones on which t r a n s i t i o n
occurred a t low Mach numbers and i s an i l l u s t r a t i o n of t h e e f f e c t of
blunting on the length of laminar run. In f a c t , a t M = 3, the length
of laminar run w a s increased by as much as a factor of 4. T o t a l Stanton
numbers f o r t h e flat-nose shapes were l e s s than corresponding cone values
a t a l l Mach numbers. Thus, by blunting it was possible t o increase sur-
face area and volume by a f a c t o r of 3 with no heat-transfer penalty.

When the Stanton numbers are referenced t o wetted surface area, a

pronounced e f f e c t of fineness r a t i o i s apparent as shown i n figure 5.
Note, f o r example, t h a t values of average Stanton number f o r the
K = 6 shape a r e 35 percent lower than those f o r the cone with all-laminar
boundary lqyer (M > 4). The maximum reduction i n average Stanton number
w i t h all-laminar f l o w i s 70 percent f o r t h e K = 9 shape a t M = 4.24.
These d a t a are, as are a l l t h e d a t a discussed previously i n t h i s report,
f o r zero angle of attack. It should be pointed out, however, t h a t no
measurable change i n Stanton-numbers has been observed f o r the flat-nose
shapes at angles of a t t a c k up t o 3'.

The f a c t should not be overlooked that the r a t e of aerodynamic'

heating i s actually proportional t o the product of Stanton number and
tpmperature recovery f a c t o r . Thus, it i s a t a c i t assumption of the
discussion i n previous sections t h a t recovery f a c t o r i s e s s e n t i a l l y con-
s t a n t . The v a l i d i t y of t h i s assumption i s i l l u s t r a t e d i n figure 6 where
average temperature recovery f a c t o r s based on free-stream conditions a r e
shown as a function of free-stream Mach number. It i s apparent t h a t
shape has l i t t l e e f f e c t on recovery factor; hence, t h i s f a c t o r plays no
s i g n i f i c a n t r o l e i n t h i s discussion of the e f f e c t of shape on heating.

DISTRIBUTION OF IxlCAL HEAT-TRANSFER COEFFICIENTS *

The measurement of o v e r a l l r a t e s of heat t r a n s f e r described i n this

paper can, at b e s t , give only a q u a l i t a t i v e idea of l o c a l heating r a t e s .
It was undertaken t o determine a t h e o r e t i c a l d i s t r i b u t i o n of heat-transfer
c o e f f i c i e n t s around one t y p i c a l shape and t o make a limited experimental
comparison. The method of Stine and Wanlass ( r e f . 2) w a s used t o calcu-
l a t e heat-transfer coefficients. This calculation requires a knowledge
of the l o c a l f l o w properties j u s t outside t h e boundary layer. These proper-
t i e s were derived from experimental pressure d i s t r i b u t i o n s such as those
shown i n f i g u r e 7. Also sham i n figure 7 a r e the predicted pressures
 .. D* .* .
5

of the modified Newtonian theory which was used t o design the blunt-nose
bodies. These predicted pressures a r e i n good agreement w i t h experiment,
t h e differences being most pronounced i n the region of the shoulder of
the body. The calculated variation of l o c a l heat-transfer c o e f f i c i e n t
w i t h &istance along the body surface i s shown i n figure 8 where t h e
r e s u l t s of the Stine-Wanlass method, which includes the e f f e c t of pres-
sure gradient, a r e compared at free-stream Mach numbers 3 and 5 w i t h
f l a t 7 p l a t e values taken from t h e laminar boundary-layer theory of
Van Driest ( r e f . 3 ) . All heat-transfer coefficients are based on l o c a l
flow properties j u s t outside t h e boundary layer. Reference values near
the stagnation point were computed by the method of Sibulkin ( r e f . 4 ) .
Predicted heat-transfer coefficients remain e s s e n t i a l l y constant over
t h e first half of t h e nose f l a t but then increase t o 2 t o 3 times t h i s
value near the shoulder. (A portion of t h e curve i n t h i s region has been
shown as a dashed l i n e because spacing of the pressure taps d i d not per-
m i t an accurate determination of the maximum.) Subsequently, l o c a l
coefficients decrease sharply t o l e s s than one-half the i n i t i a l value
and continue i n a gradual decline t o the base. The notable f e a t u r e of
t h i s prediction i s the pronounced increase of heat-transfer coefficients
over t h e forepart of t h e body as a r e s u l t of three-dimensional and
pressure-gradient e f f e c t s . These r e l a t i v e l y large l o c a l heating rates
indicate t h a t perhaps t h e design of these bodies has been somewhat over-
simplified. Very qualitatively, it appears that a s l i g h t l y convex nose
would tend t o reduce t h e heat-transfer peak at t h e shoulder by reducing
the l o c a l density, although a t the expense of a s l i g h t increase i n heat
t r a n s f e r t o the nose.

COMPARISON OF TBEIORY WITH EXPERlMENT

Integrated values of the t h e o r e t i c a l heat-transfer coefficients f o r

portions of the body surface are compared w i t h experimental measurements
i n figure 9. The flat-nose a d nose-shoulder data were obtained w i t h
camposite models t h a t isolated t h e heat-transfer surfaces fram t h e r e s t
of the body. Agreement is reasonably good except for t h e f l a t nose a t
free-stream Mach numbers above 3 . Both camputed and measured Stanton
numbers show t h a t roughly 55 percent of t h e t o t a l heat t r a n s f e r was
concentrated i n t h e nose-shoulder region which has l e s s than X) percent
of the surface area. Thus, on an average basis, heat-transfer r a t e s i n
t h i s region a r e greater by a factor of 5 than those f o r the rest of t h e
body.

These r e s u l t s apply only when boundary-layer flaw i s e n t i r e l y

laminar. A q u a l i t a t i v e indication of the e f f e c t of t r a n s i t i o n on t o t a l
heat t r a n s f e r w a s obtained, i n one case, by a r t i f i c i a l l y tripping t h e
boundary layer i n t h e shoulder region. The r e s u l t a n t Stanton number i s
shown as the s o l i d point i n figure 9. It i s indicated that the o v e r a l l
6

heating rate i s not s u b s t a n t i a l l y increased by the presence of a turbu-

l e n t boundary layer i n t h i s region of r e l a t i v e l y low density flow.

CONCLUDING REMARKS

The r e s u l t s of t h i s investigation c m be summarized as follows:

1. A method of designing blunt shapes has been devised which reduces

t h e heat t r a n s f e r t o a body by v i r t u e of low-velocity flow o%er t h e nose
and low-density f l o w over the afterbody.

2. An afterbody curvature has been found which augments the favor-

able e f f e c t of a f l a t nose i n promoting long runs of laminar boundary
layer.

3. Total heat t r a n s f e r t o these shapes i s the same o r l e s s than t o

a cone of about t h e same drag, although heat-transfer rates per u n i t of
surface area a r e considerably lower.

4. Camparison of theory and experiment indicates higher than average

.
heat-transfer rates near the shoulder of a t y p i c a l flat-nose shape and
r e l a t i v e l y low values over most of t h e afterbody.

1. Eggers, A. J., Jr., Resnikoff, Meyer M., and Dennis, David IS.: Bodies
of Revolution Having Minimum Drag a t U g h Supersonic Airspeeds.
NACA TN 3666, 1956. (Supersedes NACA RM's A51K27 by Eggers, Dennis,
and Resnikoff and A52D24 by Resnikoff ) .
2. Stine, Howard A . , and Wanlass, Kent: Theoretical and Experimental
Investigation of Aerodynamic-Heating and Isothermal Heat-Transfer
Parameters on a Hemispherical Nose With Laminar Boundary Layer a t
Supersonic Mach Numbers. NACA TN 3344, 19%.

3. Van D r i e s t , E. R.: The Laminar Boundary Layer With Vasiable Fluid

Properties. Rep. No. AL-1866, North American Aviation, Inc.,
Jan. 19, 1 9 9 .
4. Sibulkin, M. : Heat Transfer Near the Forward Stagnation Point of a
Body of Revolution. Jour. Aero. Sci. (Readers' Forum), vol. 19,
no. 8, Aug. 1952, pp. 570-571.
 7

BODY SHAPES USED IN HEAT-TRANSFER INVESTIGATION

REYNOLDS ANALOGY h a pucpcf

BODIES FOR LOW HEAT TRANSFER

FAMILY # 2

BODIES FOR LOW HEAT TRANSFER

FAMILY # I REFERENCE BODIES

Figure 1

HEAT TRANSFER APPARATUS

.010" CLEARANCE ALL AROUND

Figure 2
8

EFFECT OF BLUNTNESS ON HEAT TRANSFER AT M54.24

.0045r

.0030L I
nl I I I I I I
.0026{
L I D = -767
e
Z
Z
aa
$wm
5
$
az=
a
W
B
NOSE DIAMETER /BASE DIAMETER, d/D

Figure 3

TOTAL HEAT TRANSFER TO BODIES

.OlO -
R FOR
-
M MAX DIA
3.02 2.05 X IO6
4.24 1.80 x IO6
E .008 - 5.05 .90x IO6
m 6.30 .38 x IO6
fz
.006 -
0
I-
z B.L. PARTLY
a
& .004- TURBULENT
1
s
g .002 -
Q A

I I I I 1
0-
3 4 5 6 7
MACH NUMBER
 AVERAGE HEAT TRANSFER TO BODIES

E
g.004 - /
A
I /
3
Z /
5.003 - /'
I-
Z
a
/
m ~ a
B.L. PARTLY
I-
(J7.002- TURBULENT
v 03
w
a B 0 a4
a
(L
9 .OOl - m A
a A a8
4 n aK=g
- I I I I

Figure 5

AVERAGE TEMPERATURE-RECOVERY FACTORS OF BODIES

1.00 -
oz
.80 -
i?
>-
5.60 -
s>
k! .40-
w
(3
U
a -
w.20
B
-
3 4 5 6 7
MACH NUMBER

Figure 6
10

PRESSURE DISTRIBUTION ABOUT A BLUNT BODY

pq
2.0 - o M.3.02
o M.5.05

& 1.6 -
W '
2
LL
LL
w 1.2-
0
V

W MODIFIED NEWTONIAN THEORY

a .e-
2
v) c
W
U

DISTANCE ALONG SURFACE, FT

J.O. 5229 RELLER MARCH 4-6, 1957
A-223157CONF: AIRCRAFT LOADS, FLUTTER 8 STRUCTURES,LANGLEY FIELD, VA.

Figure 7
.

THEORETICAL LOCAL HEAT-TRANSFER COEFFICIENTS

FOR A BLUNT BODY; LAMINAR BOUNDARY LAYER
.04 r I I I

STIN E- WANLASS

_--- VAN DRIEST (FLAT PLATE)

0 .08 .I6 .24 .32

DISTANCE ALONG SURFACE, FT
(a) M.3.02

Figure 8
 11

HEATTRANSFERTOABLUNTBODY

EXPERIMENT
THEORY (STINE-WANLASS)
5.008 -
m
if
z
z ,006- K=6
0
I- TURBULENT BOUNDARY LAYER
z ON AFTERBODY
a
$ ,004 -
-J
F
e .002-
LI I I I
-I
0 3 4 5 6
MACH NUMBER

Figure 9
 .'

LOCAL HEAT TRANSFER TO BLUNT NOSES AT HIGH SUPERSONIC SPEEDS

By W i l l i a m E. Stoney, Jr.

Langley Aeronautical Lab or a tory

A b r i e f summary i s presented of :he recent t h e o r e t i c a l and experi-

mental work on l o c a l heat-transfer rates on blunt-nose bodies. Compari-
sons of t h e o r e t i c a l and measured heating r a t e s indicate t h e following
conclusions: The calculation of l o c a l conditions over noses of high
radius of curvature needs more study and t e s t i n g at t h e present time.
If l o c a l flow conditions a r e known, t h e laminar heating r a t e s over t h e
whole nose shape can be calculated h i t h i n most engineering accuracy
requirements f o r t h e whole range of Mach numbers up t o 13.8. For t h e
case of turbulent flow, turbulent f l a t - p l a t e theory based on l o c a l flow
properties can provide good estimates of t h e heating r a t e s possible.
The prediction of t r a n s i t i o n remains a t present t h e biggest unknown.

INTRODUCTION

The importance of blunt-nose shapes has increased i n recent years

because such shapes have many advantages i n comparison with t h e more
common sharp-nose shapes f o r missiles having high heating rates. High
drag is sometimes desirable f o r b a l l i s t i c missiles and does not penalize
t h e t o t a l e f f i c i e n c y of these missiles since they operate a t e s s e n t i a l l y
drag-free a l t i t u d e s f o r t h e major p a r t of t h e i r f l i g h t . As shown i n r e f -
erence 1, extremely blunt noses develop lower t o t a l heating r a t e s than
do sharp noses of about t h e same s i z e and d r a g . Finally, and perhaps
most importantly, t h e r e are many indications that extremely blunt shapes
f o s t e r longer runs of laminar flow than do sharp shapes.

For these reasons t h e calculation of t h e l o c a l heating r a t e s on such

shapes has become extremely important. The present paper gives a b r i e f
summary of present-day techniques and t h e i r effectiveness. T h i s summary
together w i t h t h e attached l i s t of references should be helpful t o those
orienting themselves i n t h i s fast-changing f i e l d .
 I

CP pressure coefficient

2 surface.distance measured from stagnation point

M Mach number

P pressure

9 heating rate, Btu/( sec) (sq f t )

k,d free-stream Reynolds number based on body diameter

rb base radius

rh nose radius o r radius of curvature

S surface distance f r o m stagnation point t o junction of hemisphere

and cylinder; on f l a t face, distance from stagnation point t o
edge, t h a t is, cylinder radius

T temperature, OF

X distance along center l i n e measured from apex of nose

e angle between f r e e streaan and noma3 t o surface of nose

CI- v i s c o s i t y of a i r

P density of a i r

Subscripts :

HEMIS hemispherical nose

2 local

W based on temperature of w a l l surface

0 a t stagnation point

m f r e e stream
 I.

DISCUSSION

Calculation of Local Flow Conditions

Before any attempt can be made t o calculate heat-transfer r a t e s ,

the l o c a l flow conditions must be calculated. Blunt noses can be c l a s s i -
f i e d roughly i n t o two groups: those f o r which t h e Newtonian flow concept
(Cp,z/Cp,o = cos20) i s applicable and thus f o r which t h e l o c a l conditions
can be e a s i l y calculated, and those f o r which t h e Newtonian flow concept
i s not applicable and f o r which no simple solutions e x i s t .

A rough boundary can be fixed between these two groups of nose shapes
by consideration of a s e r i e s of noses having constant r a d i i of curvature.
Such a s e r i e s i s i l l u s t r a t e d i n figure 1; t h i s s e r i e s progresses from t h e
hemispherical nose on t h e l e f t t o t h e f l a t nose ( i n f i n i t e radius) on t h e
r i g h t . The intermediate case f o r t h i s s e r i e s would be t h e nose shown i n
t h e center - the nose f o r which, according t o Newtonian theory, t h e flow
j u s t before t h e corner i s a t a Mach number of 1. Until s t r i c t e r c r i t e r i a
are provided, Newtonian calculations should be checked with experiment
o r with a more comprehensive theory f o r noses having a radius of curva-
t u r e a t t h e stagnation point greater than Recently, s e v e r a l theo-
r e t i c a l approaches have been presented t o t h e blunt-nose problem. (See
r e f s . 2 t o 5; t h e o r i e s presented i n r e f s . 4 and 5 may a l s o be found i n
appendix D of r e f . 6. ) Some of these appear t o be promising, but more
experience w i t h them i s needed before t h e i r general usefulness can be
determined.

This paper considers only shapes whose l o c a l conditions a r e known.

Experimental d a t a and t h e o r e t i c a l r e s u l t s a r e applied t o t h e hemispherical
and t h e f l a t noses t o b r i n g out t h e following points:

(1)If t h e r a t e of change of velocity a t t h e stagnation point i s

known, t h e heat-transfer r a t e a t t h i s point can be calculated accurately
enough f o r most engineering purposes.

(2) With t h e proper pressure d i s t r i b u t i o n t h e laminar heating r a t e s

over t h e e n t i r e nose can be calculated s a t i s f a c t o r i l y , and reasonably
good estimates of t h e possible turbulent heating rates can be made.

Prediction of Stagnation-Point Heating Rates

The prediction of stagnation-point heating rates i s well proved

f o r t h e lower Mach number range and needs demonstration only at t h e
higher Mach numbers where r e a l gas e f f e c t s come i n t o play.
 I n f i g u r e 2 t h e r a t i o s of measured stagnation-point heating r a t e s
t o t h e o r e t i c a l calculations a r e presented as functions of free-stream
Mach number. The t e s t r e s u l t s were obtained from a f l i g h t of t h e Lockheed
X-17 rocket having a nose defined by t h e equation r / r b = (X/*b) 1/3
( r e f . 6 ) , a f l a t - f a c e cylinder i n f r e e f l i g h t ( r e f . 7), and shock-tube
experiments performed by Alexander P. Sabol of t h e Langley Aero-Physics
Section f o r 1-inch-diameter spheres. The l i n e f o r t h e Lockheed data repre-
s e n t s t h e Mach number t h e h i s t o r y f o r t h e f l i g h t , the lower portion being
t h a t f o r t h e accelerating p a r t of t h e f l i g h t and t h e upper portion repre-
senting t h e decelerating Mach numbers. The difference between accelerating
and decelerating values a t any given Mach number i s probably a measure of
t h e t e s t accuracy r a t h e r than a r e a l e f f e c t due t o t h e higher Reynolds
numbers of t h e deceleration portion of t h e f l i g h t path. The NACA rocket
d a t a should be considered as preliminary data.

The t h e o r e t i c a l heating rates f o r both rockets were calculated by

t h e stagnation-point theory of Fay and Riddell, c i t e d i n appendix B of
reference 6, f o r t h e equilibrium boundary layer w i t h a Prandtl number
of 0.71 and a Lewis number of 1.4. The stagnation-point theory of
Lees ( r e f . 8) w a s compared w i t h t h e shock-tube experiments. The t h e o r i e s
a r e d i f f e r e n t because Lees assumes pp = p w h = Constant through t h e
boundary l a y e r while Fay and Riddell use an approximately correct varia-
t i o n . Probstein ( r e f . 9) notes that, f o r hypersonic f l i g h t conditions,
t h e heat-transfer r a t e f o r a r e a l i s t i c v a r i a t i o n of pp across t h e bound-
a r y l a y e r i s s i g n i f i c a n t l y lower than t h a t calculated from t h e approxi-
mation pp = p+w. If F'robstein's conclusions a r e assumed t o apply
d i r e c t l y t o Lees' calculations, t h e l e v e l of t h e r a t i o of t h e shock-tube
experimental values t o Lees ' calculated values shown i n f i g u r e 1 would be
r a i s e d t o somewhere between 1.1 and 1.4. This would not change t h e general
conclusion derived from t h i s f i g u r e t h a t , although t h e s c a t t e r shown indi-
cates t h a t many d e t a i l s of t h e heat-transfer processes under these high-
temperature conditions a r e s t i l l unknown, t h e agreement of these d a t a
i s close enough so t h a t it can be f e l t that t h e i r general nature is
understood .
Calculation of Heating Rates Over E n t i r e Nose

I n t h e calculation of heating r a t e s over t h e e n t i r e nose, t h e

stagnation-point heat-transfer r a t e s of the hemispherical noses a r e used
as datum points, and a l l t h e d a t a and calculations are presented as r a t i o s
of these values.

Figure 3 presents the r a t i o s of l o c a l laminar heating r a t e s on a

hemispherical nose t o stagnation heating r a t e s p l o t t e d as functions of
p o s i t i o n along t h e surface of t h e hemisphere. The t e s t points are f o r
 5

a variety of conditions from Moo = 2 t o 6.8 and free-stream Reynolds

numbers based on diameter of 1.0 x 106 t o 14.3 x 106. (Data f o r M = 2
and M = 3.9 are presented i n r e f . 10 and d a t a f o r M = 6.8 i n r e f . 11.
The data f o r M = 2.5 were obtained from a f r e e - f l i g h t investigation
conducted by James J. Buglia a t t h e Langley P i l o t l e s s Aircraft Research
Station a t Wallops Island, Va.) Good agreement with the t h e o r e t i c a l dis-
t r i b u t i o n s , which were calculated by t h e method of Lees ( r e f . €9, can be
seen immediately. Corrections by the Probstein theory t o Lees' stagnation-
point theory mentioned previously would probably not a f f e c t the r a t i o s gre-
sented i n figure 3 . It is a l s o important t o notice t h a t t h e d i s t r i b u t i o n
of heating r a t e s does not vary markedly w i t h Mach number.
Another i n t e r e s t i n g condition hriicated by the data i n figure 3 i s t h e
extremely high Reynolds number (14.3 x 106) f o r which laminar flow was
obtained i n one f l i g h t t e s t . This model, however, was a s p e c i a l case
for which extreme pains were taken t o obtain a m i r r o r f i n i s h on t h e nose
of the order of 2 microinches - not an i d e a l process f o r assembly-line
fabrication.

S i m i l a r agreement w i t h t e s t r e s u l t s can be shown f o r t h e other

extreme i n t h e s e r i e s of nose shapes - t h e p e r f e c t l y f l a t face. I n f i g -
ure 4 the l o c a l heating r a t e s were not divided by t h e f l a t - f a c e stagna-
t i o n heating r a t e s but by the appropriate hemispherical stagnation heating
r a t e s . T h i s allows a d i r e c t comparison of t h e f l a t and hemispherical
values. The s o l i d l i n e gives the t h e o r e t i c a l laminar r e s u l t s for the
f l a t face by the method of Lees ( r e f . 8); and the dashed l i n e above it,
the same r a t i o calculated by the method of S t i n e and Wanlass ( r e f . 1 2 ) .
The l a t t e r calculations d i f f e r from Lees' because they include a cor-
rection f o r t h e e f f e c t of pressure gradient. (For t h e r e l a t i v e l y l o w
pressure gradients of t h e hemisphere, t h e two methods agree closely.)
A s can be seen from the figure t h e agreement w i t h e i t h e r theory i s rea-
sonably good and the uncertainties i n t h e measurements do not permit a
choice between them.

Since t h e f l a t nose i s a shape f o r which Newtonian f l o w concepts

do not work a t all, a pressure-ratio d i s t r i b u t i o n pz/po obtained experi-
mentally a t a Mach number of 2 (see ref'. 13 and t h e o r e t i c a l work of
r e f . 14) was assumed t o be constant f o r a l l higher Mach numbers and was
used i n t h e t h e o r e t i c a l calculations. The agreement of the data, which
cover Mach numbers from 2 t o 13.8, indicates that t h i s approximation was
adequate f o r t h i s case. (Data f o r M = 2 are presented i n r e f . 13; data
f o r M = 13.8 are presented i n r e f . 7; and data f o r M = 5 were deter-
mined i n an investigation of 2-inch f l a t - f a c e cylinders conducted by
Morton Cooper in t h e Mach number 5 axisymmetric blowdown j e t a t t h e
Langley Gas Dynamics Branch.)
6

What i s gained by the f l a t face in lower l o c a l heating r a t e s i n

t h e center i s l o s t a t i t s edge. However, it must be remembered t h a t high
l o c a l heating r a t e s are not necessarily the c r i t i c a l points i n a p a r t i -
cular design, as can be shown by the time h i s t o r y of temperature p r o f i l e s
..
on the f r o n t and s i d e of the f l a t - f a c e rocket model on which t h e heating
r a t e s a t a Mach number of 13.8 were obtained. These temperature p r o f i l e s
and a sketch of t h e nose are presented i n f i g u r e 5. The nose was made of
copper and was three-sixteenths of an inch t h i c k on t h e f r o n t and one-
eighth of an inch thick at t h e sides. The datum points represent t h e
temperatures as measured along t h e f r o n t surface and down t h e side f o r
t h r e e d i f f e r e n t times during t h e high Mach number p a r t of the f l i g h t .
A t the e a r l i e s t time not much hea%ing has occurred and t h e temperature
p r o f i l e i s r e l a t i v e l y f l a t . A t peak Mach number the e f f e c t s of conduc-
t i o n i n the skin can be seen since the highest temperature is reached
a t 0.82/s i n s p i t e of t h e peak heating which occurs a t t h e corner. The
reason t h a t the sides are able t o a c t as good heat sinks l i e s i n t h e i r
extremely l o w heating r a t e s . Measurements of these r a t e s f o r t h e f i r s t
two s i d e thermocouples are shown in figure 4. Seven seconds l a t e r , t h e
model has slowed down t o a Mach number of 7, and t h e l a t e r a l flow of
heat i n the skin w a s s o large t h a t the maximum temperature of t h e f l i g h t
occurred i n t h e center of t h e model - not a t the corners. T h i s is, of
course, only an example but it does show t h a t a f a i r l y d e t a i l e d study
of the p a r t i c u l a r nose m u s t be made i f i t s effectiveness i s t o be evalu-
ated correctly.

There are some indications ( f o r example, r e f . 1) t h a t f l a t noses

o r closely a l l i e d shapes may have some advantages in retaining l a m i n a r
flow, and, as i s shown, on a hemispherical nose t h i s retention can be
very important since turbulent heating r a t e s can be very high. I n f i g -
ure 6 t h e r a t i o s of l o c a l heating r a t e s t o stagnation heating r a t e s a r e
presented as functions of surface location. The t e s t points were deter-
mined f r o m an investigation conducted by Ivan E. Beckwith and James J.
Gallagher i n a blowdown j e t a t t h e Langley G a s Dynamics Branch. For
t h i s investigation a Mach number of 2 and free-stream Reynolds num-
bers (based on body diameter) of 2.7 x 106 and 3.4 x 106 were used.
Transition obviously took place forward on the hemisphere f o r both t e s t s
and heating r a t e s nearly 2-1 times the stagnation r a t e were reached. The
2
s o l i d l i n e represents f l a t - p l a t e turbulent values based on the l o c a l
Reynolds numbers around the body. T h i s comparison and similar compasi-
sons f r o m other t e s t s indicates that even t h i s r e l a t i v e l y crude t h e o r e t i -
c a l approach may have considerable value in estimating t h e turbulent
heating r a t e s . (See a l s o r e f s . 15 and 16. The t h e o r e t i c a l approach of
r e f . 16 may a l s o be found i n appendix C of r e f . 6.)
F

CONCLUDING REMARKS

The calculation of t h e l o c a l heating conditions over noses of high

radius of curvature needs more investigation a t t h e present time, although
data obtained i n low Mach number t e s t s may be adequate a t much higher Mach
numbers f o r use i n these calculations. However, i f the l o c a l conditions
are known, t h e laminar heating r a t e s over the whole nose shape can be
calculated within most engineering accuracy requirements f o r t h e whole
range of Mach numbers up t o a t l e a s t 13.8.

For turbulent f l o w , f l a t - p l a t e theory may provide good estimates

of t h e heating r a t e s possible. O f course, t h e prediction of t r a n s i t i o n
s t i l l remains and i s the biggest unknown a t the present time.
 8

1. Reller, John O., Jr.: T o t a l Heat Transfer t o Blunt-Nose Shapes With

Laminar Boundary Layers a t High Supersonic Speeds. (Prospective
NACA paper.)

2. Maslen, Stephen H., and Moeckel, W. E.: Inviscid Hypersonic Flow Past
Blunt Bodies. Preprint No. 665, S.M.F. Fund Preprint, Inst. Aero.
Sei., Jan. 1957.

3. Probstein, Ronald F.: Inviscid Flow i n t h e Stagnation Point Region

of Very Blunt-Nosed Bodies a t Hypersonic Flight Speeds. WADC
TN 56-39? (Contract No. AF 33(616)-2798), Wright A i r Dev. Center,
U. S. Air Force, Sept. 1956. (Also available from ASTIA as Doc.
No* AD97273.1
4. Hayes, Wallace D.: Some Aspects of Hy-personic Flow. The Ramo-Wooldridge
Corp., Jan. 4, 1955.

5. Hayes, Wallace D.: Hypersonic Flow Fields a t S m a l l Density Ratio.

Doc. No. 34, The Ramo-Wooldridge Corp., May 12, 1955.

6. Anon. : X-17 Re-Entry Test Vehicle - R-3 Fin& Flight Report. Rep.
No. MSD-313 (Contract No. AF 04 (645)-7), Lockheed A i r c r a f t Corp.,
Oct. 31, 1956.

7. Stoney, W i l l i a m E., Jr., and Swanson, Andrew

G.: H e a t Transfer on
a Flat-Face Cylinder Measured in Free Flight a t Mach Numbers Up
t o 13.8. (Prospective NACA paper.)

8. Lees, Lester: Laminar Heat Transfer Over Blunt-Nosed Bodies a t Hyper-

sonic F l i g h t Speeds. J e t Propulsion, vol. 26, no. 4, Apr. 1956,
PP. 259-269.
9. Probstein, Ronald F.: The Effect of Variable Fluid Properties on t h e
Equilibrium Laminar Boundary Layer Surface Heat Transfer Rate a t
Hypersonic Flight Speeds. WADC TN 56-2 (Contract No. AF 33(616)-
2798), Wright Air Dev. Center, U. S. Air Force, Dec. 1955.

10. Garland, Ben jamine J., and Chauvin, Leo T. : Measurements of Heat
T r a n s f e r and Boundary-Layer Transition on an 8-Inch-Diameter Hemi-
sphere Cylinder in Free Flight f o r a Mach Number Range of 2.0 t o 3.88.
(Prospective NACA paper. )

U. Crawford, D a v i s H., and McCauley, W i l l i a m D. : Investigation of t h e

Laminar Aerodynamic Heat-Transfer Characteristics of a Hemisphere-
Cylinder in t h e Langley 11-Inch Hypersonic Tunnel a t a Mach Number
of 6.8. NACA TN 370

f -
t - ._
30%
 9

12. Stine, Howasd A., and Wanlass, Kent: Theoretical and Ekper b e n t al
Investigation of Aerodynamic-Heating and Isothermal Heat-Transf e r
Parameters on a Hemispherical Nose With Laminar Boundary Layer a t
Supersonic Mach Numbers. NACA TN 3344, 1954.

13. Maskley, J. T., and Stoney, W i l l i a m E., Jr.:

Heat-Transfer and Pres-
sure Measurements on a Flat-Face Cylinder a t a Mach Number of 2.
(Prospective NACA paper. )

14. Maccoll, J. W., and Codd, J.: Theoretical Investigations of t h e Flow

Around Various Bodies i n t h e Sonic Region of Velocities. B r i t i s h
Theoretical Res. Rep. No. 17/45, B.A.R.C. 45/19, Ministry of Supply,
Armament Res. Dept., 1945.

15. Van Driest, E. R.: The Problem of Aerodynamic Heating. Aero. Eng.
Rev., vol. 15, no. 10, Oct. 1956, pp. 26-41.

16. Denison, M. Richard: Turbulent Boundary Layer on Blunt Bodies of

Revolution a t Hy-personic Speeds. Res. Memo., Lockheed Aircraft
Corp., Missiles Systems Div., Apr. 13, 1956.

f
 BLUNT-NOSE FLOW FIELDS

r S O N l C LINE

Figure 1

COMPARISON OF MEASURED AND CALCULATED

STAGNATION HEAT-TRANSFER RATES

F \

QcALc -6 -
-
.4 -
- --X-17 ROCKET, LOCKHEED
.2- SHOCK TUBE
0
}
- ~//l//////l
ROCKET
NACA
1 1 1 1 1 1 1 1 1 1 1 1 1 ~
LAMINAR HEATING RATES ON HEMISPHERICAL NOSES

Ma, R@,d TEST

0 2.0 2.7X106 ROCKET
o 2.5 14.3 ROCKET
A 3.9 5.1 ROCKET
v 6.8 1.0 TUNNEL

1.2
1.0
.8
.6
Ma,
=*ILAMINAR
MaJ = 4 THEORY
.4
.2

0 .2 .4 .6 .8 1.0
1/s

Figure 3

LAMINAR HEATING RATES ON A FLAT-FACE CYLINDER

M a Ra,,d TEST
A 2 4.7x1O6 TUNNEL
0 5 4.6 TUNNEL

I I I I I I l n j
0 .2 .4 .6 .8 1.0 1.2 1.4 1.6
z/s

Figure 4 304
52

1EMPERATURE HISTORY ON FACE AND SIDES

OF ROCKET MODEL

1,800
1,600
1,400
1,200
1,000
T,OF
800
600
400
200

0 .2 .4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6

1/5

Figure 5

TURBULENT HEATING RATES ON A HEMISPHERICAL NOSE

TUNNEL TESTS AT M a = 2

Rqd
0 2.7X106
0 3.4

LOCAL TURBULENT
FLAT-PLATE THEORY

.4
-
.2 -
I I
\

I
\ -,p
I 1
LAMINAR THEORY

0 .2 .4 .6 .8 1.0
1 /s

t
i
acs Figure 6
 HEAT TRANSFER TO BODIES AT A N G U S OF ATTACK

By W i l l i a m V. F e l l e r

Langley Aeronautical Laboratory

SUMMARY

Heat-transfer rates were measured on a modified Ka.rman nose shape

a t an angle of a t t a c k of Oo a t Mach llumbers of 6.86 and 3.69, and a t
angles of a t t a c k up t o 2 5 O a t a Mach number of 3.69. Data are presented
f o r a smooth model, which showed n a t u r a l t r a n s i t i o n , and f o r the model
with roughness s t r i p s , which caused f u l l y turbulent flow.

INTROGUCTION

The heat t r a n s f e r t o bodies of revolution a t zero angle of a t t a c k

with laminar flow has been extensively studied, both t h e o r e t i c a l l y and
experimentally, so t h a t i n general calculations can be made f o r a given
configuration with confidence. For turbulent boundary layers a t zero
angle of attack, the solutions f o r cones and f l a t p l a t e s have been used
with t h e l o c a l flow conditions along t h e body t o obtain good approxi-
mations. Very l i t t l e study has been made of t h e heat t r a n s f e r t o bodies
a t angles of attack. A t very lasge angles of attack, it might be expected
t h a t t h e r e s u l t s found f o r swept i n f i n i t e cylinders would be applicable,
but there i s a decided lack of information on heat t r a n s f e r a t i n t e r -
mediate angles of attack.

This paper presents some of the r e s u l t s from an experimental study

of t h e heat-transfer rates t o a p a r t i c u l a r body of revolution a t angles
of a t t a c k up t o 25O, i n order t o give some idea of t h e d i s t r i b u t i o n of
heat-transfer rates t h a t can be expected a t supersonic speeds on an air-
plane fuselage.

SYMBOLS

NSt Stanton number based on free-stream a i r properties

*D Reynolds number based on body maximum diameter and free-stream

air properties
 I

X a x i a l distance measured from nose

D body maximum diameter

A sweep angle, deg

a angle of attack, deg

M Mach number

APPARATUS

The body t e s t e d i s shown'at t h e top of figure 1. It i s a Karman

nose shape of fineness r a t i o 5 , modified by t h e addition of a tangent
1
0' half-angle cone i n f r o n t , and a length of c y l i n d r i c a l section behind.
Tests w e r e made a t zero angle of a t t a c k i n the Langley 11-inch hypersonic
tunnel a t a Mach number of 6.86, and a t angles of a t t a c k from Oo t o 2 5 O
i n t h e Langley Unitary Plan wind tunnel at a Mach number of 3.69.

RESULTS

I n figure 1 are shown t h e r e s u l t s of the tests a t t h e two Mach num-

bers a t zero angle of a t t a c k f o r laminar flow. The laminar correlating
parameter NStK i s p l o t t e d on a logarithmic scale against t h e axial
s t a t i o n along t h e m o d e l i n diameters, X/D. The c i r c l e s are the data
from t h e t e s t s a t M = 6.86 and t h e squares a t M = 3.69. The dashed
curves, calculated by t h e theory of S t i n e and Wanlass ( r e f . 1) f o r t h e
two Mach numbers, f i t t h e data w e l l over t h e body, except t h a t , a t axial
s t a t i o n s behind X/D = 4.2 a t M = 3.69, t h e r i s e i n t h e data indicates
t r a n s i t i o n t o turbulent flow. Also shown i n figure 1 are curves calcu-
l a t e d f o r M = 6.86, by t h e f l a t - p l a t e theory of Van D r i e s t (ref. 2)
f o r t h e cone and f l a t p l a t e . The calculated curves f o r t h e cone and
f l a t p l a t e agree fairly w e l l with t h e experimental data on t h e p a r t s of
t h e body t h a t are nearly conical and cylindrical, respectively.

I n figure 2 are shown data f o r t h e same model a t M = 3.69 and

zero angle of a t t a c k but with turbulent flow tripped by roughness applied
i n a r i n g near t h e nose as indicated i n t h e sketch a t t h e t o p of t h e f i g -
ure. The same parameter N s ~ F D i s presented i n order t o f a c i l i t a t e
comparison with t h e laminar values i n figure 1. The data near t h e nose
are not q u i t e up t o t h e values calculated i n turbulent cone theory
( r e f . 3) but show t h e same t r e n d along t h e length as did t h e laminar
data, and agree w e l l with t h e turbulent f l a t - p l a t e theory ( r e f . 4 )
toward t h e rear of t h e body.
 C'
1F ..
7

Figure 3 shows the variation of the Stanton number based on the

free-stream conditions along the lower or windward meridian l i n e of t h e
body a t several angles of attack. On t h e l e f t a r e shown t h e data f o r
the smooth model. The curve a t zero angle of a t t a c k i s t h e same as w a s
shown i n figure 1, with t r a n s i t i o n s t a r t i n g a t X/D = 4.2. As the
angle of a t t a c k increases, the heat-transfer r a t e s increase, as would
be expected, and the point of t r a n s i t i o n moves forward, up t o 14' angle
of a t t a c k . The curves a t angles of a t t a c k of 21' and 25' show a some-
w h a t d i f f e r e n t trend.

I n order t o obtain turbulent heat-transfer data, a s t r i p of rough-

ness was applied i n a band along the lower meridian l i n e as well as i n
a r i n g ne= the nose. The r e s u l t s a r e sham on the r i g h t i n f i g u r e 3.
As i n the case of laminar f l o w , the Stanton numbers increase w i t h angle
of a t t a c k from Oo t o 140. !&e curve f o r a = 21° does not follow t h e
trend with angle of attack, and i s i n f a c t lower than t h a t f o r a = 14O.
The curve f o r a = 25O i s very i r r e g u l a r and w i l l be discussed l a t e r .
The behavior of the Stanton number d i s t r i b u t i o n curves suggests that
the flow around the body changes t o a mainly crosswise flow a t some
angle between 1 4' and 21'.
I n figure 4 the d i s t r i b u t i o n of Stanton number along t h r e e meridian
l i n e s i s shown f o r two angles of attack, 7' on t h e l e f t , and 25' on t h e
right. Again, roughness was applied t o produce turbulant flow. A t an
angle of a t t a c k of 7' t h e Stanton number drops from i t s value a t the
windward meridian t o about half a t 900, and about l/3 on the leeward
side (180°). On t h e leeward side, behind X/D = 3, the Stanton number
i s independent of X/D, so that separated flow i s indicated.

A t an angle of a t t a c k of 25O, t h e values on t h e windward meridian

a r e very e r r a t i c , ranging from values near those found f o r laminar flow
up t o values i n good agreement with t h e turbulent swept-cylinder theory.
This behavior i s due t o the f a c t t h a t the l o c a l Reynolds numbers over
t h e roughness s t r i p a r e very low, so t h a t t r a n s i t i o n i s not f u l l y com-
pleted a t a l l s t a t i o n s . The range of values of t h e data represents var-
ious stages i n t h e t r a n s i t i o n t o fully turbulent flows. The curves a t
meridian s t a t i o n s goo and 180° a r e smooth. Values of Stanton number a t
the 90' s t a t i o n f o r a = 25' a r e not very d i f f e r e n t from those found
f o r a = 7 O , but a t the 180' s t a t i o n a r e somewhat lower than a t a = 7 O .

Figure 5 shows a polar p l o t of Stanton number around t h e circum-

ference of the body a t X/D = 5.12, where t h e body i s cylindrical, f o r
several angles of attack. On the r i g h t side a r e t h e d i s t r i b u t i o n s f w
M l y turbulent flows, produced by a roughness s t r i p along the windward
meridian. On the l e f t side of t h i s figure a r e values obtained on the
smooth model. As the angle of a t t a c k increases, t h e fully turbulent
Stanton numbers increase on the windward side and decrease on t h e lee-
ward side up t o 14O. The curve f o r a = 21' nearly coincides with that
4

f o r a = 14O, and the curve f o r a = 2 3 O shows a f u r t h e r increase. The

long-short-dash curve f o r a = 250 i s calculated by a semiempirical
method developed f o r swept cylinders. The basis of t h e calculatian i s
t h e theory f o r t h e average heat-transfer c o e f f i c i e n t s over the f r o n t half
of swept c i r c u l a r cylinders with fully turbulent boundary layers presented
by Beckwith and Gallagher i n reference 5 . This theory w a s conibined w i t h
an unpublished experimentally measured d i s t r i b u t i o n of l o c a l heat-transfer
c o e f f i c i e n t s around a swept cylinder obtained a t the Langley Laboratory
by Beckwith and Gallagher, on the assumption t h a t the d i s t r i b u t i o n curve
i s independent of sweep angle and Mach number ( 8 s has been shown f o r
laminar flow). The agreement with experiment i s very good.

On t h e l e f t side of t h e figure a r e shown the d i s t r i b u t i o n s of

Stanton number obtained on the smooth model a t a = Oo, 70, and 250.
The values a t a, = Oo, w i t h n a t u r a l t r a n s i t i o n , a r e i n fair agreement
with the a s t i f i c i a l t r a n s i t i o n values on the r i g h t . A t an angle of
a t t a c k of 7 O , t h e 'boundary layer i s laminar near the windward meridian
and low heat-transfer values a r e obtained. Transition starts a t a
meridian angle about 45O. A t a = 25O, t r a n s i t i o n starts e a r l i e r ,
about 20° from t h e lower meridian l i n e , a f t e r which the Stanton numbers
increase and f a i r i n t o t h i s curve f o r f u l l y turbulent flow a t a = 25'
which w a s transposed f r o m t h e data on the r i g h t .

In figure 5 the d a t a a t a = 2 5 O agree well with crossflow theory.

The behavior of the curves a t a = 14O and 21° indicates t h a t the ty-pe
of flow may be changing i n t h e i n t e r v a l between. To show the trend more
clearly, t h e Stanton numbers on t h e windward meridian a r e p l o t t e d i n
figure 6 against angle of a t t a c k f o r two s t a t i o n s on the model;
X/D = 2.8 and 5.1. The curve labeled "longitudinal-flow theory" was
calculated by f l a t - p l a t e theory, by use of t h e l o c a l flow conditions
outside t h e boundary layer f o r the equivalent f r e e stream. The curve
labeled "crossflow theory" w a s calculated by use of the semiempirical
method described e a r l i e r f o r swept cylinders, t h e body being considered
t o be made up of a s e r i e s of cylinders of varying diameter and sweep
angle. The experimental values increase uniformly from a = 0' t o 14'
but the trend does not agree with that predicted by t h e modified f l a t -
p l a t e (longitudinal-flow) theory. A t a = 2 1' and 25O, t h e data agree
w e l l w i t h the crossflow theory curve.

CONCLUDING REMARKS

The experimental r e s u l t s presented i n t h i s paper have shown that

a t zero angle of attack, t h e available t h e o r e t i c a l methods give a good
estimate of the heat-transfer coefficients on a body of revolution. A t
l a r g e angles of attack, t h e methods based on the cross flow over swept
 5

cylinders are applicable. For intermediate angles of attack, however,

there is as yet no theoretical approach, and reliance w i l l have to be
placed on extrapolation from experimental studies. It is believed that
the results presented in this paper can be considered typical of simple
fuselages of moderate fineness ratio at supersonic Mach numbers for which
the air may still be considered an ideal gas.

REFERENCES

1. Stine, Howard A., and Wanlass, Kent. : Theoretical and Experimental

Investigation of Aerodynamic-Heating and Isothermal Heat-Transfer
Parameters on a Hemispherical Nose With Laminar Boundary Layer at
Supersonic Mach Numbers. NACA TN 3344, 19%.

2. Van Driest, E. R.: The L d n a r Boundary Layer With Variable Fluid

Properties. Rep. No. AL-1866,North American Avaiation, Inc , .
Jan. 1.9, 19%.

3 . Van Driest, E. R.: Turbulent Boundary Layer on a Cone in a Super-

sonic Flow at Zero Angle of Attack. Jour. Aero. Sci., vol. 19,
no. 1, Jan. 1952, pp. 55-57, 72.

4. Van Driest, E. R.: Turbulent Boundary Layer in Compressible Fluids.

Jour. Aero. Sci., vol. 18, no. 3, Mar. 1951, pp. 145-160, 216.
5. Beckwith, Ivan E., and Gallagher, James J.: Experimental Investiga-
tion of the Effect of Boundary-Layer Transition on the Average
Heat Transfer to a Yawed Cylinder in Supersonic Flow. NACA
m ~56~09,
1956.
c

HEAT TRANSFER TO KARMAN NOSE AT ZERO ANGLE OF ATTACK

LAMINAR FLOW

oM.6.86
OM= 3.69

CONE, M.6.86

.I -
I I I I I I I 1 I
0 I 2 3 4 5 6 7 8
4
D

Figure 1

HEAT TRANSFER TO KARMAN NOSE AT ZERO ANGLE OF ATTACK

TURBULENT FLOW
IO

0 M13.69; RD= 1.25 X IO6

NSt 6

I I I I I I I I
0 I 2 3 4 5 6 7
-
X
D

Figure 2
i 31%
 HEAT TRANSFER TO LOWER MERIDIAN
KARMAN NOSE j M.3.69 j R D = 1.2 x I06

SMOOTH MODEL ARTIFICIAL TRANSITION

a,DEG
0 0
0 7
0 14
A 21
A 25

I I I I I
0 1 2 3 4 5 0 1 2 3 4 5
X/D X/D

Figure 3

HEAT TRANSFER TO KARMAN NOSE

M.3.69 j R D = 1 . 2 ~ 1 j0ARTIFICIAL
~ TRANSlTlON

Q=7O Q =25O
MERIDIAN
ANGLE, DEG

'1
6 o
0
0180
0 (WINDWARD)
90
00 . -
TURBULENT
CROSS FLOW
THEORY
-- ---E

NSt

0 1 2 3 4 5 0 1 2 3 4 5
X/D X/D

Figure 4
8

CIRCUMFERENTIAL DISTRIBUTION O F HEAT TRANSFER

K A R M A N NOSE;$ ~5.12; Rj-j=1.2XI06
M~3.69;

Figure 5

HEAT TRANSFER TO LOWER MERIDIAN LINE AT ANGLES OF ATTACK

M.3.69; R ~ ' i . 2 3 x I o ~ ;TURBULENT FLOW

X
-
D ~2.8
X
-
D = 5.1
8 10-3 8X10-3

CROSSFLOW THEORY CROSSFLOW THEORY

6 6-

NSt
4
mM = 3.69 4-

2 2- 0

FLOW THEORY C FLOW THEORY

I I I I I I I
C 10 20 30 40 0 10 20 30
a, DEG a, DEG

Figure 6
313
 HEAT TRANSFER IN REGIONS OF SEPARATED AND REATTACHED FLOWS

By Davis H. Crawford and Charles B. Rumsey

Langley Aeronautical Laboratory

SUMMARY

Past experimental work has indicated that separated flow can greatly
increase the heat transfer to a surface; whereas, some theoretical studies
have indicated a possible decrease. Recent investigations have helped
to clarify the effects of separation on heat transfer and have indicated
a method of reducing separation. This paper considers the results of
some of these investigations and shows the heat transfer in regions of
separation and reattachment for a few specific shapes. These results
have shown that the heat transfer in a separated region is strongly
affected by the extent of separation, the location of the reattachment
point, and the location of transition along the separated boundary.

INTRODUCTION

Recent experimental and theoretical investigations have shown that

the separation of the boundary layer from a surface can have a great
effect on the heat flow to the surface. At high altitudes and high Mach
numbers, the presence of separation is more likely, and the regions of
separation are more extensive because of the thicker boundary layers
characteristic of this type of flight. Past experimental work has indi-
cated that separated flow can greatly increase the heat transfer to a
surface (ref. l), whereas some theoretical studies have indicated a pos-
sible decrease (ref. 2). Recent investigations have helped to clarify
the effects of separation on heat transfer and have indicated a method
of reducing separation. This paper considers the results of some of these
investigations and shows the heat transfer in regions of separation and
reattachment for a few specific shapes. For one of these shapes, boundary-
layer bleed was used to delay separation and to alter the heat transfer
in the region of possible separation.

SYMBOLS

b distance along surface from zero station to sphere-cylinder

juncture
CD drag coefficient

d diameter of vehicle a t c h a r a c t e r i s t i c s t a t i o n , as indicated i n

figures

2 length

M Mach number

NSt Stanton number

9 heat t r a n s f e r per u n i t area per u n i t time

%,ns heat t r a n s f e r per u n i t area per u n i t time a t zero s t a t i o n with

no spike

R Reynolds number based on c h a r a c t e r i s t i c length along surface

Rd Reynolds number based on d

S distance along surface from zero s t a t i o n

X distance along axis of t e s t vehicle, as indicated i n figures

DISCUSSION

Flare Skirt

One i n t e r e s t i n g case of separation i s t h e flow about an ogive-

cylinder body with a 30' t a i l f l a r e ( r e f . 3 ) . The t a i l f l a r e i s a pos-
s i b l e s t a b i l i z i n g device as well as a drag brake f o r many missile shapes.
The lower curve of f i g u r e 1 shows t h e position of t h e start of separation
as a function of Reynolds number. Figure 2 shows the approximate shape
of the flow p a t t e r n f o r three Reynolds numbers. A t low Reynolds numbers
t h e s t a r t of separation moves forward s l i g h t l y with increasing Reynolds
number as i s expected from laminar theory ( r e f . 4 ) . The shape of t h e
flow p a t t e r n when t h e separation i s a maximum i s shown i n t h e top diagram
of figure 2. A s the Reynolds number increases, t h e boundary l a y e r near
t h e start of separation tends t o become t r a n s i t i o n a l which delays t h e
separation. Then t h e position of t h e start of separation moves back along
t h e model u n t i l a t a Reynolds number of 0.5 x 106 t h e f l o w p a t t e r n
appears as i s shown i n t h e center diagram of f i g u r e 2. A t high Reynolds
numbers, t h e boundary layer i s f u l l y turbulent ahead of t h e separation,
as i s shown i n t h e lower diagram of figure 2, and t h i s causes the flow t o

315
24

EFFECT OF WING ON TAIL LOADS

M=2.46; azoo; (L= 20"
1.5

-.5
-I
\
\
\

Figure 19
 23

EFFECT OF TAIL ROLL CONTROL

ON VERTICAL-TAIL LOAD

-.3
.4
' I 1
.6
I
.0
I I I
I.o
MACH NUMBER

Figure 17

EFFECT OF WING ON TAILLOADS

M=2.46; 8=0";
a=6O

1.0

.5
-
Z
C
0

-.5

-1.0
-I 0 I
1-5or -
CN
CN&W

Figure 18
 FLIGHT MEASUREMENTS OF AIRPLANE STRUCTURAL TEMPERATURES

AT SUPERSONIC SPEEDS

By Richard D. Banner

NACA High-speed F l i g h t Station

SUMMARY

Skin and s t r u c t u r a l temperatures have been obtained on t h e X-lB

and X-IE research airplanes under t r a n s i e n t aerodynamic heating condi-
t i o n s a t speeds up t o Mach numbers near 2.0. Extensive temperature
measurements w e r e obtained throughout t h e X-u3 airplane, and temperature
d i s t r i b u t i o n s are shown on t h e nose cone, t h e wing, and t h e v e r t i c a l
t a i l . Temperatures f o r t h e X-1E wing leading edge and i n t e r n a l wing
s t r u c t u r e w e r e compared w i t h similar data f o r the X-lB.
P

No c r i t i c a l skin and s t r u c t u r a l temperatures were obtained on t h e

two airplanes over t h e range of these tests.

Simplified calculations of t h e skin temperatures i n t h e laminar-

flow regions of t h e nose cone and the leading edges agreed favorably
w i t h t h e general trends i n t h e measured data. The f l a t - p l a t e skin-
temperature calculations i n t h e turbulent-flow regions agreed favorably
w i t h t h e measured data on t h e nose cone and a t t h e midsemispan s t a t i o n
of t h e wing but overestimated the v e r t i c a l - t a i l skin temperatures and
a l s o t h e upper wing skin temperature near t h e wing t i p . The r e l a t i v e l y
low values of t h e upper skin temperatures t h a t w e r e measured a t t h e wing
t i p were believed t o be caused by separated-flow e f f e c t s i n t h i s region.

INTRODUCTION

I n t h e design of supersonic a i r c r d t , aerodynamic heating i s

becoming increasingly important. Analytical studies and controlled
t e s t i n g represent t h e basic methods u t i l i z e d ' i n t h e design of complex
s t r u c t u r e s t o withstand t h e e f f e c t s of aerodynamic heating.

Concurrent with t h e basic research studies, t h e National Advisory

Camittee f o r Aeronautics i s conducting a program a t the NACA High-speed
F l i g h t S t a t i o n a t Edwards, C a l i f . , t o investigate t h e skin and s t r u c t u r a l
temperatures a c t u a l l y experienced by airplanes during f l i g h t a t super-
sonic speeds. The purpose of t h i s paper is t o summarize t h e results of
 2

t h i s program t o t h e present time. The r e s u l t s of simplified calculations

of the skin temperatures a r e compared w i t h t h e measured data i n t h e
regions of the fuselage nose cone and the wing and v e r t i c a l - t a i l skins
and leading edges. .
SYMBOLS

b/2 wing semispan

C chord length

hav average heat-transfer coefficient, Btu/sq f t -hr-%

5 pressure a l t i t u d e , f t

M Mach number
P l o c a l surface pressure, lb/sq f t

Pm free-stream s t a t i c pressure, l b / s q f t

free-stream dynamic pressure, lb/sq f t

r recovery f a c t o r

T skin temperature, ?F

Taw adiabatic w a l l temperature, ??

TT free-stream stagnation temperature, T,(I + 0.a2), OF

Tm free-stream ambient a i r temperature, %

t time, sec

a angle of attack, deg

r thickness, in.

TESTS
.-.)
-. -c- .c *, -.:
-.
'.,. .2".
..-.,,pp---' Data have recently been obtained on the X - U airplane a t Mach num-
bers up t o 2.10 and on the X-lB a t Mach numbers up t o 1.94. (See f i g . 1.)

329
 3

These speeds do not necessarily represent t h e maximum speed capabilit5es

of t h e airplanes. Both airplanes are constructed primarily of aluminum.
Temperature measurements were made a t approximately 60 locations on the
X-lE, and approximately 300 temperatures were measured on t h e X-lB.

Figure 2 shows t h e f l i g h t conditions f o r both airplanes under which

t h e temperatures were obtained. The Mach number, pressure a l t i t u d e ,
ambient a i r temperature, and angle of a t t a c k are shown as time h i s t o r i e s .
As can be seen, t h e two f l i g h t s a r e generally similar.

Transient heating conditions were experienced by both airplanes

during t h e f l i g h t s . The maximum heating r a t e was experienced on t h e
thinner skinned X-IB and was on t h e order of 3' per second. This heating
r a t e i s r e l a t i v e l y low i n comparison w i t h t h e rates being obtained on
missiles and rocket models and i n controlled wind-tunnel tests; however,
it i s believed t o be representative of t h e heating r a t e s which a r e being
experienced or which will be experienced i n t h e near f u t u r e by f i g h t e r
and interceptor a i r c r a f t .

For laminar flow i n t h e stagnation regions, approximate heat-transfer

c o e f f i c i e n t s were calculated from expressions r e l a t i n g t h e Nusselt number
and the Reynolds number given by Stine and Wanlass. (See ref. 1.) I n
t h e regions of f l a t - p l a t e turbulent flow, approximate heat-transfer coef-
f i c i e n t s w e r e calculated by using Colburn's expression. These calcula-
t i o n s w e r e g r e a t l y simglified by t h e use of free-stream conditions.
This procedure was considered j u s t i f i e d i n t h a t only t h e o v e r a l l e f f e c t s
were desired. A d e t a i l e d t h e o r e t i c a l analysis i s not considered t o be
within t h e scope of t h i s report.

The results of t h e calculations j u s t described indicated r e l a t i v e l y

small v a r i a t i o n s i n t h e heat-transf er c o e f f i c i e n t s with t i m e ; theref ore,
average values were determined and used i n calculating t h e skin
temperatures.

Newton's l a w of heat flow t o t h e skin, which considers t h e heat capac-

i t y of t h e material and neglects t h e e f f e c t s of conduction, was assumed.

RESULTS AND DISCUSSION

The X-IB i n s t a l l a t i o n provides a r a t h e r d e t a i l e d case h i s t o r y of the

temperatures that exist throughout an actual airplane; and, i n view of
t h i s f a c t , more at-f;entioni s given t o t h e X-IB data i n t h e present paper.
c
 4

Th6-X-lE data are used t o point out any differences t h a t e x i s t i n t h e

measured temperatures due t o either t h e configuration o r t h e construction
o r both.
r
Some of the maximum temperatures t h a t w e r e measured on the X-IB and
an indication of t h e higher temperature areas on t h e airplane are shown
i n figure 3 . The nose of t h e airplane and t h e leading edges of t h e wing
and t a i l surfaces are shaded i n figure 3 t o indicate the higher tempera-
ture areas. The approximate thermocouple locations a r e indicated by t h e
dark points. The maximum stagnation temperature f o r t h e f l i g h t was 220' F.

A maximum skin temperature of 185' F was measured a t t h e forward point

of t h e nose cone. On the w i n g leading edge t h e maximum temperature w a s
168' F. A maximum temperature on t h i s order was a l s o measured on t h e
leading edge of t h e v e r t i c a l t a i l . The maximum temperatures i n other
general areas are a l s o shown. Some of t h e temperatures shown a t these
locations are appreciably a f f e c t e d by i n t e r n a l heat sources and heat
sinks; f o r instance, t h e fuselage skin temperatures adjacent t o t h e
l i q u i d oxygen tank a r e r e l a t i v e l y low (130 F); whereas, j u s t ahead of
that area on the fuselage skin a maximum of 122'-F w a s measured. Other
areas of i n t e r e s t a r e t h e windshield and canopy and t h e rearward p a r t
of the fuselage i n t h e region of t h e rocket engine.

It should be pointed out t h a t t h e maximum temperatures shown i n f i g -

ure 3 did not a l l occur a t the same t i m e . I n areas where i n t e r n a l con-
duction i s negligible, skin thickness, boundary-layer temperature, and
t h e heat-transfer coefficient are the primary f a c t o r s a f f e c t i n g t h e skin
temperature rise .
Figure 3 gives an o v e r a l l p i c t u r e of t h e measured temperatures.
I n subsequent f i g u r e s , the temperature d i s t r i b u t i o n s i n t h e areas of the
nose cone, t h e wing skins, t h e leading edges and t h e v e r t i c a l t a i l are
considered i n greater detail; and comparisons are made with calculated
r e s u l t s . Attention i s given only t o t h e supersonic portion of the f l i g h t s ,
since a t t h e subsonic speeds the combined e f f e c t s of increasing Mach num-
ber and decreasing ambient a i r temperature produced o n l y small changes i n
the skin temperatures.

Figure 4 shows the nose of t h e X-IB, where both skin-temperature and

pressure measurements w e r e made a t i n t e r v a l s along t h e nose between sta-
t i o n 0 and s t a t i o n 55.0. The pressure data are p l o t t e d as pressure coef-
f i c i e n t s a t t h e bottom of t h e figure. The measured skin temperatures a r e
shown i n t h e upper p a r t of the figure as the open symbols. The skin t e m -
perature decreases over the forward p a r t of the nose cone, the laminar
f l o w calculations showing f a i r l y good agreement. This v a r i a t i o n i s f o l -
. lowed by a region of no change i n the skin temperature and then an increase
. L'
-- t-oward t h e rear of the nose cone where t h e calculated turbulent f l a t - p l a t e
I . . and cone temperatures a r e approached. It i s about here t h a t t h e nose-cone
F 5

shape asymptotically approaches t h e cylindrical shape of t h e fuselage,

and t h e pressure drops t o near t h e free-stream value. The o v e r a l l t e m -
perature variations that a r e shown suggest t h a t t r a n s i t i o n from laminar
t o turbulent flow takes place along the nose a t about 25 t o 30 percent
of t h e nose-cone length, f u l l y turbulent flow developing - a t the r e a r o f -
the nose cone. A summary of various wind-tunnel data on cones and bodies
of revolution indicated t h i s general t r a n s i t i o n a l area.

The skin here on t h e nose cone was r e l a t i v e l y t h i n i n comparison

w i t h t h a t on other areas on t h e airplane. The e f f e c t s of varying skin
thickness on the skin temperatures can be seen i n f i g u r e 5. Figure 5
shows t h e spanwise d i s t r i b u t i o n s of maximum temperatures on t h e upper
wing skin a t the 66-percent-chord l i n e and a t t h e leading edge. The
spanwise variation i n t h e skin thickness i s shown below. The temperature
variations indicate an inverse relationship between t h e skin temperature
and t h e skin thickness a t both chordwise positions. The trend which
would give r i s e t o thermoelastic considerations a t higher temperature
l e v e l s i s c l e a r l y seen i n t h e data here.

The calculated temperatures, shown by the dashed l i n e s i n f i g u r e 5 ,

were based on a constant spanwise heating input, laminar f o r the leading
edge and turbulent f o r the 66-percent-chord l i n e ; and the same v a r i a t i o n s
due t o thickness a r e seen i n the calculated temperatures.

The d&ta i n figure 5 a l s o i l l u s t r a t e t h e differences i n t h e skin

temperature w i t h chordwise position. Both top and bottom skin tempera-
t u r e s were measured a t several chordwise positions a t about midsemispan
and near t h e t i p of the wing. These data a r e presented i n figure 6.
I n t h i s f i g u r e t h e calculated temperatures were estimated on t h e basis
of zero angle of attack. No detailed consideration i s given t o t h e
e f f e c t s of angle of a t t a c k on t h e measured temperatures; however, from
an overall standpoint it should be recalled from f i g u r e 2 t h a t t h e angles
of a t t a c k were positive during t h e f l i g h t and were on t h e order of 2'
t o loo.

The data of figure 6 i l l u s t r a t e several i n t e r e s t i n g trends. F i r s t ,

notice t h e chordwise temperature gradients that a r e shown a t t h e mid-
semispan s t a t i o n . The higher temperatures a r e experienced a t t h e leading
edge and t h e t r a i l i n g edge (which a t t h i s location i s t h e outboard t i p of
t h e f l a p ) , and t h e lower temperatures are experienced on the t h i c k e r
skinned wing box section.

Secondly, note t h e differences i n temperature between t h e top and

bottom skins a t t h e two span stations, the bottom skin temperature being
higher i n both cases. A t t h e t i p station, t h e f a i r l y large differences
seen between t h e top and bottom skin suggest that the flow might be p a r t l y
separated i n t h i s Etrea.
6

The approximate calculations, which were based on the assumption

of laminar flow over the leading-edge section and turbulent f l a w over
t h e remainder of the chord, give a f a i r l y good o v e r a l l estimate of t h e
skin temperatures a t t h e midsemispan stations. A t the t i p s t a t i o n t h e
calculations agree f a i r l y well with t h e bottom skin temperature; however,
t h e y indicate an overestimate of t h e t o p s k i n temperature, probably
because of the flow e f f e c t s previously mentioned.

An example of t h e i n t e r n a l temperatures that were measured through

t h e wing i s seen i n f i g u r e 7. Shown i n the figure a r e the f r o n t wing
spars a t about midsemispan on the X-U3 and t h e X-lE. The temperatures
w e r e measured a t the locations shown by the black dots on t h e structures,
and t h e values a r e given on the right-hand side of the figure. The higher
temperatures shown f o r t h e X - l B were measured on the s k i n a s l i g h t d i s -
tance from the spar.

I n order t o give an indication of t h e temperature r i s e t h a t has taken

place i n the i n t e r n a l structure, it i s worthwhile t o mention that the
ambient a i r temperatures were between -TO0 F and -90' F and t h a t t h e
assumed turbulent adiabatic w a l l temperatures were on the order of 200' F
f o r t h e highest Mach numbers shown here. The measured i n t e r n a l tempera-
tures are r e l a t i v e l y low and show only slight temperature gradients across
t h e thickness of t h e X-lB wing, t h e lower temperature being obtained on
the spar center l i n e . E s s e n t i a l l y no differences a r e seen on t h e t h i c k
spar construction of the X-IE wing, t h e heavier type of construction of
the X-IE having a temperature-neutralizing tendency due t o t h e higher
heat capacity. The thermal l a g e f f e c t i s a l s o seen i n f i g u r e 6 by t h e
increase i n t h e measured temperatures as t h e Mach number decreases i n
t h e later portions of the flights.

Effects of material thickness differences are a l s o seen i n t h e meas-

ured leading-edge temperatures. These data are shown i n f i g u r e 8 i n
time-history form together w i t h time h i s t o r i e s of t h e assumed laminar
adiabatic w a l l temperatures. The locations a t which the temperatures
were measured a r e shown by t h e dark points on t h e leading-edge sketches,
and the material thicknesses are given a t these locations. Values of
the average heat-transfer coefficients u t i l i z e d f o r the calculated t e m -
peratures are shown below t h e sketches. The calculated temperatures a r e
seen t o agree very w e l l with the measured data.

M a x i m u m temperatures on t h e same order of magnitude were measured

on t h e wing and v e r t i c a l - t a i l leading edges of t h e X-lB. For comparison,
the temperatures t h a t were measured a t t h e rear of the s o l i d leading edge
o f ' t h e X-IE are seen i n t h e middle of the figure. It w i l l be noticed
here t h a t t h e maximum measured temperature was on t h e order of 20' F.
!$e high heat capacity of the s o l i d leading edge i s the contributing
f a c t o r t o t h e small rise i n t h e temperature measured a t t h i s location.
The calculated s k i n temperature i s shown t o agree .very w e l l w i t h t h e

323
 7

measured data; however, t h i s r e s u l t i s not considered s i g n i f i c a n t because

the measured temperatures a r e r e l a t i v e l y low. (For example, a 50-percent
reduction i n t h e assumed heat-transfer coefficient a t t h i s point would
produce a decrease i n t h e calculated maximum temperature of 7 O F.)

The chordwise variation i n the v e r t i c a l - t a i l temperatures i s shown

i n figure 9 f o r a time near maximum Mach number a t near midspan. The
temperatures were measured on t h e skin and spar center l i n e s at t h e loca-
t i o n s shown by t h e black dots on t h e sketch. No appreciable gradients
are seen i n t h e chordwise variation of t h e measured skin temperature.

Transition from laminar t o turbulent flow w a s assumed t o take place

a t the point where t h e leading-edge section attaches t o t h e f r o n t spar
because inspection revealed a r e l a t i v e l y large discontinuity i n t h e skin
a t t h i s point. Skin temperatures calculated from the average heat-
t r a n s f e r coefficients shown and based on these assumptions agree f a i r l y
well with t h e measured trends i n t h e leading-edge region but deviate
somewhat over t h e remainder of t h e chord and give a conservative o v e r a l l
estimate i n t h i s region.

CONCUTDING REMARKS

Skin and s t r u c t u r a l temperatures have been obtained on t h e X-IB

and X-IE research airplanes under t r a n s i e n t aerodynamic heating condi-
t i o n s a t speeds up t o Mach numbers near 2.0. Extensive temperature
measurements were obtained throughout t h e X-IB airplane, and temperature
d i s t r i b u t i o n s are shown on t h e nose cone, t h e wing, and t h e v e r t i c a l
t a i l . Temperatures f o r t h e X-1E wing leading edge and i n t e r n a l wing
structure were compared with similar data far the X-U3.

No c r i t i c a l skin and s t r u c t u r a l temperatures w e r e obtained on t h e

two airplanes over the range of these tests.

Simplified calculations of t h e skin temperatures i n t h e laminar-

flow regions of t h e nose cone and t h e leading edges agreed favorably
w i t h the general trends i n t h e measured data. The f l a t - p l a t e skin-
temperature calculations i n t h e turbulent-flow regions agreed favorably
w i t h the measured data on t h e nose cone and a t the midsemispan s t a t i o n
of t h e wing but overestimated t h e v e r t i c a l - t a i l skin temperatures and
a l s o t h e upper wing skin temperature near t h e wing t i p . The r e l a t i v e l y
low values of the upper skin temperatures t h a t were measured a t the wing
t i p were believed t o be caus6d by separated-flow e f f e c t s i n t h i s region.

324
 REFERENCE

1. Stine, Howard A., and Wanlass, Kent: Theoretical and Experimental

Investigation of Aerodynamic-Heating and Isothermal Heat-Transfer
Parameters on a Hemispherical Nose With Laminar Boundary Layer at
Supersonic Mach Numbers. NACA TN 3344, 1954.
 9

RESEARCH AIRPLANES

X-IE X-IB
Mz2.10 Mz1.94
60 TEMP. GAGES 300 THERMOCOUPLES

d d
Figure 1

FLIGHT CONDITIONS

X-l E X-I6

M
0.5

" l i : 7 1-fl
-
hp, FT
25

Ta°F "
-100
7
L-'
12
a, DEG
0 100 200 300 4000 100 200 300 400
TIME, SEC TIME, SEC

Figure 2
 MAXIMUM MEASURED TEMPERATURES, X- I 6

THERMOCOUPLE LOCATIONS

122OF

Figure 3

NOSE CONE TEMPERATURES AND PRESSURES, X- I B

M = 1.94, t = 270 SEC

"I0 --
MEAS.
1
0

----- LAMINAR
-TURBULENT CALC.
STA 55.0

TEMP, CONE
OF o
- . o o o 0 0 0

0 , , , o , u 0 T F :

PRESSURE ORIFICES

'0 IO 20 30 40 50 60
DISTANCE, INCHES

Figure 4
MAXIMUM SPANWISE SKIN TEMPERATURES, X-IB WING
L.E., M = 1.90, t = 290 SEC 66O/oc, Mz1.32, t = 3 2 5 SEC

.F
66 % c

CHORDWISE SKIN TEMPERATURES, X-IB WING

M a = 1.90, TT'212 OF

-+-BOTTOM SKIN I
-0- TOP SKIN \
200r CALC (CY=O)
r

---- ------------
~

\TOP AND BOTTOM

BOTTOM
0 %C 100 0 %C IO0

Figure 6
MEASURED INTERNAL WING TEMPERATURES
X-18

0 40 80 120
TP F

X-IE
M
2.04
//1.30
.40 TYF!

L
0 40 80 120
T: F

LEADING-EDGE TEMPERATURES
0 MEAS ---- CALC - Taw (r=0.85)
X-IB WING X-IE WING X-IB VERTICALTAIL
54% b/2 63.3% b/2 M ID -SPAN

150 250 350 150 230 310 150 250 350

t, SEC t, SEC t, SEC
1-

Figure 8
IF

CHORDWISE TEMPERATURES-X- IB VERTICAL TAIL

M=1.90. t = 290 SEG

ha,,
-
BTU/SQ FT HR-"F
3Lk-L-
0 20 40
To CHORD
60
TURBULENT

80 100

Figure 9
 1 b.. -

TWO FACTORS INFLUENCING TEMPERATURE DISTRIBUTION&

AND THERMAL STRESSES IN STRUCTURES
*g N?
By W i l l i a m A. Brooks, Jr., George E, G r i f f i t h ,
and H. K u r t Strass

Langley Aeronautical Laboratory

The influence of j o i n t conductivity and i n t e r n a l r a d i a t i o n on t e m -

perature d i s t r i b u t i o n and thermal s t r e s s e s has been discussed. J o i n t s
of poor conductivity can occur i n normal f a b r i c a t i o n procedure and
g r e a t l y a l t e r temperature d i s t r i b u t i o n s and increase thermal s t r e s s e s .
On the other hand, i n t e r n a l radiation tends t o make t h e temperature
d i s t r i b u t i o n s more uniform and thereby r e l i e v e s thermal s t r e s s .

INTRODUCTION

Thermal s t r e s s e s arre unquestionably an important consideration i n

t h e design of supersonic a i r c r a f t . Such s t r e s s e s a r e usually produced
by nonuniform temperature d i s t r i b u t i o n s - the greater the tanperature
varriation within t h e s t r u c t u r e , the larger t h e thermal stresses. The
present paper deals b r i e f l y w i t h t w o f a c t o r s which may a f f e c t the tem-
perature d i s t r i b u t i o n , and thus t h e thermal s t r e s s e s , within a structure;
theae two f a c t o r s a r e j o i n t conductivity and i n t e r n a l radiation. In
order t o indicate some of t h e e f f e c t s of these two f a c t o r s without
unnecessary s t r u c t u r a l ccanplications, t h e basic s t r u c t u r e considered
consists of a length of skin w i t h an i n t e g r a l or an attached web. In
addition t o t h e o r e t i c a l r e s u l t s , experimental data were obtained by
heating t h e s t r u c t u r e e i t h e r aerodynamically or by laboratory radiant-
heat sources.

SYMBOLS

U c o e f f i c i e n t of thermal expansion

C spec if i c heat

E modulus of e l a s t i c i t y
33.8
 - .A
2

E s-missi v i t y

h heat-transfer c o e f f i c i e n t

hj j o i n t conductivity c o e f f i c i e n t

k thermal conductivity

M Mach number

heating r a t e

t skin thickness

T temperature

TO i n i t i a l temperature

T1 maximum skin temperature

T2 minimum web temperature

Taw equilibrium temperature

7 time

W spec i f i c weight

CT thermal stress

Subscripts :

RC
.
r a d i a t i o n and conduction

C conduction

MAX maximum

DISCUSSION

Simgle I n t e g r a l Structure

Consider f i r s t t h e simplest possible type of skin-web combinstion,

one without joint; and without i n t e r n a l r a d i a t i o n e f f e c t s . Figure 1
shows such h s t r u c t u r e , a symmetrical i n t e g r a l H-section w i t h a flange
( o r skin) of thickness t symmetrically heated a t a constant r a t e q; .-- .1
 3

within t h i s s t r u c t u r e heat t r a n s f e r takes place by conduction only. The

actual temperature T minus the i n i t i a l temperature To, multiplied by
the material thermal conductivity k and divided by the heating r a t e q
and skin thickness t is p l o t t e d against a time parameter i n which T
i s the elapsed time, c t h e specific heat, a,nd w t h e s p e c i f i c weight.
The choice of these dimensionless parameters makes it possible t o present
on t h e same p l o t t h e r e s u l t s f o r a given cross section of different
materials subjected t o d i f f e r e n t heating r a t e s . The calculated results,
shown by t h e s o l i d l i n e s , a r e obtained by a simple an-sis assuming
constant material thermal properties and no heat losses. Point 1, which
i s f a r t h e s t f r o m t h e heat sink afforded by the web, w i l l be t h e h o t t e s t
point i n t h e structure, whereas point 2, which i s f a r t h e s t frm t h e heat
source, w i l l be t h e coolest. A s a consequence of assuming no heat l o s s e s
and a constant heat input, t h e average temperature of the s t r u c t u r e i s
proportional t o t h e mass of t h e section and varies l i n e a r l y w i t h time.
In order t o check t h e theory, t h r e e tests were made involving constant
heating r a t e s of 5 , 16, and 41 Btu/ft2-sec. Each t e s t is represented
by a d i f f e r e n t symbol, and t h e experimental temperatures a r e p l o t t e d f o r
various times until t h e maximum skin temperature reached 4500 F. The
parameters a r e such t h a t the data f o r t h e highest heating r a t e , shown by
the diamonds, appear a t t h e extreme lower end of t h e curves, whereas the
data f o r the lowest heating r a t e extend t h e f u l l length of t h e curves.
The t e s t r e s u l t s a r e i n f a i r l y good agreement with t h e theory; any dis-
crepancy i s due largely t o the assumption of constant thermal properties
used i n t h e theory.

A t any point i n t h e structure, t h e thermal stress i s proportional

t o t h e difference between the average temperature and t h e temperature
of t h e point i n question. When multiplied by t h e appropriate material
properties, these differences can be converted i n t o thermal s t r e s s e s ,
such as shown i n f i g u r e 2.

In t h i s figure thermal-stress d i s t r i b u t i o n s are shown f o r both t h e

l o w and high heating r a t e s of 5 and 4 1 Bu/ft2-sec f o r the times when
t h e maximum skin temperature reached 450° F. The skin stresses appear
i n region A and the web s t r e s s e s , i n region B. The s o l i d l i n e s represent
thermal s t r e s s e s calculated from measured temperatures s h m i n figure 1,
and the symbols represent experimental s t r e s s e s obtained from strain-gage
readings corrected f o r d i f f e r e n t i a l expansion between the gage and speci-
nen. The agreement between theory and experiment i s quite good. As
would be expected, the higher heating r a t e allows less time f o r heat t o
be conducted i n t o t h e web, thus provides M g e r tem-prature differences
i n the struct-we, and consequently y i e l d s larger thermal s t r e s s e s . For
the heating r a t e of 41 Btu/ft2-sec the maximurn skin s t r e s s , shown a t the
extreme right, and t h e maximum web s t r e s s , a t t h e extreme l e f t , are more
than twice t h e values shown for the r a t e of 5 Btu/ft2-sec, despite t h e
f a c t t h a t t h e skin temperature r i s e was t h e same.

333
- 4

Another s i g n i f i c a n t point i s t h a t t h e maximum s t r e s s e s a r e not

r e l a t e d i n the same proportion as t h e heating r a t e s , which have an 8 t o
1 r a t i o . The s t r e s s e s shown here a r e those which existed when t h e s k i n
temperature Tl w a s 4 p o F. I f , however, the skin temperature were not
l i m i t e d t o 430° F and a ccanparison made a t t h e same time, then, indeed,
t h e s t r e s s e s would be proportional t o t h e heating rate.

It is a l s o possible t o present t h e s t r e s s e s i n a dimensionless form

as was done f o r the temperatures. Figure 3 shows t h e dimensionless
thermal s t r e s s e s at various times during t h e progress of t h e three t e s t s
a t heating r a t e s of 5 , 16, and 4 1 13tu/ft2-sec. It i s again evident t h a t
t h e high-heating-rate data a r e grouped at the lower end of the t h e o r e t i c a l
curves which are represented by t h e s o l i d l i n e s , whereas t h e data f o r t h e
low heating r a t e extend t h e f u l l length of t h e curves. The t h e o r e t i c a l
s t r e s s e s a r e obtained by using t h e t h e o r e t i c a l temperature of f i g u r e 1.
The agreement between theory and experiment i s about t h e same as t h a t
i n f i g u r e 1, and any discrepancy i s due t o the same reason.

The r e s u l t s of these t h r e e figures indicate t h a t f a i r l y accurate

temperatures and s t r e s s e s can be predicted i n a simple i n t e g r a l s t r u c t u r e
where only conductive heat t r a n s f e r need be considered and, f u r t h e r , t h a t
strain-gage readings, i f t r e a t e d properly, can give reasonable s t r e s s
results. However, if t h i s s t r u c t u r e had contained any j o i n t s , greater
temperature differences would have resulted, and accurate calculations
would have became smewhat more d i f f i c u l t .

Structure With Attached Webs

I n order t o show sane of the e f f e c t s of j o i n t s , t h e temperatures i n

one skin-web canbination of a multiweb wing, a more r e a l i s t i c a i r c r a f t -
type s t r u c t u r e which contains riveted j o i n t s , w i l l be examined. When
s i x 20-inch-chord, m u l t i w e b wings i d e n t i c a l i n cross section -that is,
w i t h t h e same o v e r a l l size, skin thickness, material, and so f o r t h -
were
t e s t e d a t sea-level conditions a t a Mach number of 2 i n a blowdown j e t ,
t h e r e s u l t s were as shown i n f i g u r e 4. This f i g u r e shows t h e temperature
difference T1 - T2 between the maximum temperature i n t h e skin (point 1)
and t h e minimum web temperature (point 2) divided by the maximum possible
s k i n temperature r i s e Taw - TO plotted against time. The upper dot-
dash curve gives t h e t h e o r e t i c a l temperature-ratio drop if no heat i s
conducted i n t o t h e s t i f f e n e r , that is, i f the j o i n t conductivity h j i s
equal t o zero, and t h e lower dot-dash curve gives the t h e o r e t i c a l r e s u l t s
f o r p e r f e c t j o i n t conductivity. The experimental data f o r f i v e of t h e
wings, indicated by the shaded area, l i e close t o t h e curve for p e r f e c t
j o i n t conductivity. On t h e other hand, the data f o r the s i x t h wing show
t h a t t h e increased temperature drop is sanewhat closer t o t h e curve f o r
no j o i n t conductivity. Since t h e temperature d i s t r i b u t i o n depends mainly

a.
e

s34
 I

upon heat conduction within the structure, t h i s increased temperature

difference can be a t t r i b u t e d almost s o l e l y t o a lower j o i n t conductivity.
The much poorer thermal conductivity of t h e j o i n t s of t h e one wing, con-
pared w i t h t h e other f i v e wings, was not obtained by design but rather
r e s u l t s f r m the small variations i n otherwise i d e n t i c a l s t r u c t u r e s
expected frcpn normal f a b r i c a t i o n techniques. The 50-percent increase
i n temperature drop of t h e one wing over t h e average of t h e other f i v e
wings a l s o resulted i n a substantial increase i n thermal s t r e s s . In
f a c t , the wing w i t h the l a r g e s t temperature drop f l u t t e r e d and then
suffered a dynamic f a i l u r e ; whereas, t h e orie wing among the other f i v e ,
which w a s i d e n t i c a l i n every d e t a i l , survived i t s t e s t without damage.
Construction d e t a i l s , such as the inclusion of ribs, invalidate further
comparison w i t h t h e other models. However, it seems obvious that j o i n t
conductivity can have an important, and sanetimes adverse, e f f e c t on the
temperature v a r i a t i o n i n a structure and hence on t h e thermal s t r e s s e s
and, as a r e s u l t , on t h e s t r u c t u r a l i n t e g r i t y .

In f i g u r e 5 the maximum nondimensional temperature drops have been

plotted against a dimensionless j o i n t conductivity parameter h j t / k
where h j i s t h e thermal conductivity of the j o i n t . Calculated r e s u l t s
a r e shown by the s o l i d l i n e s . I n the calculations the value of 0.013 f o r
ht/k, the B i o t number (an index of the r a t e of external t o i n t e r n a l heat
t r a n s f e r ) , w a s determined by the experimental aerodynamic heat-transfer
c o e f f i c i e n t h. The symbols indicate data obtained from t h e tests and
show a 6 t o 1 variation i n which apparently can be expected even
hj
from good shop p r a c t i c e . Data obtained by Barzelay, Tong, and Hollmay
( r e f s . 1 and 2 ) cover an even wider range i n h j . Although a consider-
able s c a t t e r i n h j is shown by the f i v e lowest t e s t points, these
points a r e i n good agreement w i t h the theory since they l i e i n a region
where a fairly large change i n j o i n t conductivity has l i t t l e e f f e c t on
the temperature drop. Thus, t h e b a d f o r t h e f i v e wings shown in f i g -
ure 4 masks a considerable variation i n h j . The s i x t h point, which
does not show such good agreement, is i n a region where a small change
i n h j can a l t e r t h e temperature drop appreciably. The spread of h j
within t h e band shown i n f i g u r e 4 i s of the same order of magnitude as
the v a r i a t i o n frm the top of t h e band t o t h e curve f o r the one wing.

If the maximum skin or s t i f f e n e r s t r e s s were p l o t t e d instead of t h e

temperature drop, a similar variation w i t h h j t / k could be observed -
the
lower t h e value of h j , t h e higher t h e thermal stress, as evidenced by
t h e f a i l u r e of t h e one wing.

A c t u a l f l i g h t - t e s t r e s u l t s a r e shown i n f i g u r e 6 for a duplicate of

one of t h e f i v e ‘wings which had very good j o i n t conductivity. The Mach
number and a l t i t u d e f o r t h e f l i g h t a r e shown a t t h e top of t h e figure.
The temperature difference TI - T2 frm t h e skin t o t h e s t i f f e n e r
 I

center l i n e is p l o t t e d against time. When t h i s wing w a s attached t o a

rocket model, t h e r e s u l t s were as shown by t h e s o l i d l i n e , which l i e s
much closer t o t h e calculated curve f o r h = 300 Btu/ft2-hr-* than
J
t o t h e curve f o r p e r f e c t j o i n t conductivity. Since a value of h j = 300
would l i e s l i g h t l y t o t h e l e f t of t h e t e s t point f o r t h e one wing w i t h
t h e l a r g e s t temperature difference shown i n figure 5 , t h e f j o i n t s of t h i s
wing would be considered poor from the standpoint of thermal conductivity.
I n t h i s case f a i l u r e of t h e wing was prevented by t h e inclusion of a
chordwise r i b .

I n t e r n a l Radiation i n an Integral Structure

As w a s mentioned previously, t h e e f f e c t of j o i n t s is t o increase

t h e magnitude of t h e thermal s t r e s s e s . However, i f the temperatures
involved a r e s u f f i c i e n t l y high, it i s possible t h a t i n t e r n a l r a d i a t i o n
i n a s t r u c t u r e may bring about some thermal-stress r e l i e f . The h o t t e r
p a r t s of t h e s t r u c t u r e t r a n s f e r heat t o the cooler p a r t s by r a d i a t i o n
and thereby lessen t h e temperature differences. Although t h e r e i s an
increased i n t e r e s t i n t h i s phenomena, very l i t t l e information can be
found on t h i s subject.

Therefore, i n order t o evaluate some of t h e e f f e c t s of i n t e r n a l

r a d i a t i o n on t h e temperature d i s t r i b u t i o n within a m u l t i c e l l beam, such
as shown i n f i g u r e 7, an analog solution w a s made t o determine the
temperature d i s t r i b u t i o n of an i n f i n i t e l y long box whose i n t e r n a l surfaces
have an emissivity of 1. The procedure w a s t o assume a constant r a t e of
heat input u n t i l t h e maximum skin temperature reached a prescribed l i m i t
and then t o maintain t h i s skin temperature while t h e web temperature
approached t h e s k i n temperature.

Heat w a s put i n t o both skins a t t h e rate of 5 Btu/Ft*-sec u n t i l

t h e maximum skin temperature TI reached 1,200’ F, a t which time t h e
skin temperature w a s kept constant. The maximum skin temperature Ti
and t h e minimum web temperature T2 a r e p l o t t e d against time i n seconds.
The dashed l i n e s are t h e solution when i n t e r n a l r a d i a t i o n i s neglected;
t h e s o l i d l i n e s represent t h e case when radiation i s included. In t h i s
s i t u a t i o n it can be seen that, when r a d i a t i o n is included, steady s t a t e
is reached much more rapidly than when conduction only i s considered. It
can a l s o r e a d i l y be seen that t h e maximum temperature difference between
t h e skin and t h e web, when conduction and r a d i a t i o n are considered, occurs
a t approximately 75 seconds and i s about two-thirds of that when conduc-
t i o n only is considered.

In order t o make a preliminary survey, other heating r a t e s and skin-

temperature l e v e l s were a l s o investigated as shown i n figure 8. Here,
P
F
7

i n order t o show one of the e f f e c t s of i n t e r n a l radiation, t h e r a t i o of

t h e maximum difference between skin temperature and web t a p e r a t u r e f o r
both radiation and conduction t o t h e maximum difference f o r conduction
only is p l o t t e d against t h e s k i n temperature T1. For a given heating
r a t e , t h e r a t i o of t h e temperature differences w i l l approach a minimum
value as t h e skin temperature increases. In the case of q = 1 Btu/ft2-sec,
t h e l i m i t i s nearly reached a t a skin temperature of 1,200° F. The curve
f o r t h e highest heating rate w i l l eventually reach t h e lowest l i m i t but a t
a temperature far i n excess of 1,200° F. For t h e conditions i n the figure,
r a d i a t i o n e f f e c t s a r e most s i g n i f i c a n t when heating rates are l o w and tea-
peratures a r e high.

Figure 9 shows some preliminary experimental results. The specimen

w a s made of
x -inch Inconel w i t h the cross section as shown and w a s
12 inches long. The emissivity of the i n t e r n a l surfaces w a s about 0.8.
The heating r a t e s were approximately 5 and 2 5 Btu/ft2-sec. Although t h e r e
i s a measurable reduction i n the maximum temperature difference, the reduc-
t i o n i s not as large as indicated by the idealized t h e o r e t i c a l r e s u l t s .

CONCLUDING REMARKS

In summary, two f a c t o r s , j o i n t conductivity and i n t e r n a l radiation,

which influence tanperature d i s t r i b u t i o n and therefore thermal stresses,
have been discussed. Even normal f a b r i c a t i o n techniques can produce
j o i n t s of such poor conductivity as t o cause t h e temperature differences
t o increase markedly over Those of an i n t e g r a l structure. On t h e other
hand, a t t h e higher temperatures i n t e r n a l r a d i a t i o n has t h e e f f e c t of
making t h e temperature differences l e s s severe and t h e thermal stresses
smaller. A t t h e higher temperatures these two e f f e c t s tend t o cancel
one another.

REFERENCES

1. Barzelay, Martin E., Tong, Kin Nee, and Holloway, George F.: Effect
of Pressure on Thermal Conductance of Contact J o i n t s . NACA TN 3295,
1955
2. Barzelay, Martin E., Tong, Kin Nee, and Holloway, George F.: Thermal
Conductance of Contacts i n Aircraft Joints. NACA TN 3167, 1954.

337
8 d

TEMPERATURES IN RADIANTLY HEATED H-SECTION

250

- CALCULATED
k
(T-To) q

0 250

Figure 1

THERMAL STRESSES IN RADIANTLY HEATED H-SECTION

*
5}q,
4'
BTU
FT*-SEC F
T
I-B
rrrrc
t t t t t
20

c r,SEC
-\
35.0
-e- 3.3
I
I
-20 I

t--B+A+
 9

THERMAL STRESSES IN RADIANTLY HEATED H-SECTION

q

40

20

- 20
-40 L
I I I I I
0 50 100 150 200 250
-
kr
2
cwt

Figure 3

TEMPERATURE DIFFERENCES IN 6 MULTIWEB WINGS

Tl-T2 .5
TAW -TO

0 6 12
TIME, SEC

Figure 4
339
VARIATION OF TEMPERATURE DIFFERENCE WITH JOINT CONDUCTIVITY

I\ - CALCULATED

0 .2
y.I
k

Figure 5

’4
TEMPERATURE DIFFERENCE FOR MULTIWEB WING IN FLIGHT
6,000

M ALT., FT
ALTITUDE
0 0
160 -

T-T2,OF 80 -

a 16
TIME, SEC

Figure 6
 11

TEMPERATURE HISTORIES FOR BOX BEAM WITH INTERNAL RADIATION

e- 1.0
BTU
q=5-
FT~-SEC

0 100 200
TIME, SEC

Figure 7

INTERNAL. RADIATION EFFECTS FOR VARIOUS HEATING RATES

BTU
q*
FT*-SEC

Figure 8
EXPERIMENTAL RESULTS FOR INTERNAL RADIATION

4
5

Figure 9
 -I

EFFECT OF TRANSIENT E A T I N G ON VIBRATION FREQUERCIES

OF SOME SIMPLF, WING STRUCTURES

By Louis F. Vosteen, Robert R. McWithey,

and Robert G. Thomson

Langley Aeronautical Laboratory

SUMMARY

Thermal stresses caused by nonuniform temperature d i s t r i b u t i o n s

associated with transient heating can cause changes i n the e f f e c t i v e
s t i f f n e s s of wing structures. Some of the e f f e c t s of this change i n
s t i f f n e s s w e r e investigated experimentally by t e s t i n g three types of
simple wing structures under various radiant-heating conditions. The
s t r u c t u r e s t e s t e d w e r e a uniform p l a t e , a s o l i d double-wedge section,
and a circular-arc multiweb-wing section. Changes i n s t i f f n e s s were
measured by measuring t h e changes i n natural frequency of vibration
during t r a n s i e n t heating. I n order t o measure changes i n frequency, a
resonance-following system w a s developed which keeps the model v i b r a t -
ing a t i t s n a t u r a l frequency. Some of the data a r e compared with theo-
r e t i c a l calculations and show that, a t temperature differences near
those required f o r thermal buckling, d i s t o r t i o n s have a marked e f f e c t on
the frequency. The conclusion w a s made that, i n order t o p r e d i c t t h e
e f f e c t s of aerodynamic heating on the s t i f f n e s s of wing structures by
means of laboratory radiant-heating t e s t s , care must be taken t o simu-
l a t e closely the temperature d i s t r i b u t i o n s produced aerodynamically.

INTRODUCTION

One of the e f f e c t s of the nonuniform temperature d i s t r i b u t i o n s pro-

duced by aerodynamic heating i s a change i n the e f f e c t i v e s t r u c t u r a l
s t i f f n e s s caused by thermal atresses. Changes i n s t i f f n e s s have been
observed i n laboratory t e s t s of several types of simple wing structures,
and, i n reference 1, thermal stresses were c i t e d as the cause of failure
of wing models subjected t o aerodynamic heating. I n order t o obtain
more information on t h i s thermoelastic phenomenon, laboratory tests
were conducted on several types of simplified wing structures f o r which
changes i n s t i f f n e s s were measured during rapid heating. The results of
these t e s t s a r e presented and some comparisons with t h e o r e t i c a l calcula-
t i o n s a r e made.

313
2

The s t r u c t u r e s considered and the manner i n which they were heated

are shown i n f i g u r e 1. The s t r u c t u r e s had a square plan form and were
mounted as cantilevers. The f i r s t structure i s a . p l a t e of uniform thick-
ness which was heated r a d i a n t l y along the longitudinal edges. The second
i s a s o l i d double-wedge section which w a s subjected t o a constant heat
input over the top and bottom surfaces. The t h i r d i s a symmetrical
circular-arc multiweb a i r f o i l . The heat input varied along the chord
as indicated i n the figure.

TEST PROCEDURE AND NSULTS

Resonance-Following System

Inasmuch as the natural frequency of vibration i s a function of

the model s t i f f n e s s , changes i n s t i f f n e s s can be measured by measuring
the changes i n n a t u r a l frequency during t r a n s i e n t heating. The mecha-
nism used t o detect the changes i n frequency i s shown schematically i n
f i g u r e 2. The mechanism consists b a s i c a l l y of a forcing system, the
vibrating body, a vibration pickup, and the servo detecting system. The
system i s operated by s e t t i n g the frequency of the signal generator s o
t h a t the model i s forced i n one of i t s natural modes of vibration. The
phase detector i s then s e t t o maintain the phase r e l a t i o n that e x i s t s
between the applied force, as indicated by %he signal generator, and
the response of the model, as indicated by the vibration pickup. If the
natural frequency of the model changes, the phase r e l a t i o n between the
force and the response w i l l change. The servo system detects t h i s
change, and, through a servomotor, mechanically drives the s i g n a l gen-
e r a t o r until the resonant phase r e l a t i o n i s recovered. In this manner
the model i s continuously forced a t i t s resonant frequency.

Uniform P l a t e

The p l a t e t e s t e d w a s of aluminum a l l o y 20 inches square and

1/4 inch thick. A s the p l a t e w a s heated, the temperature a t the edges
rose rapidly, but the center section remained r e l a t i v e l y cool. The ther-
m a l s t r e s s e s which r e s u l t e d caused a change i n the natural frequencies
of the plate. I n reference 2 results were presented f o r changes i n fre-
quency of t h e f i r s t two modes. These t e s t s w e r e repeated by using the
resonance-following system; i n addition, the next three higher modes
were investigated. The frequency h i s t o r i e s which were obtained during
t r a n s i e n t heating a r e shown i n figure 3.

The v a r i a t i o n w i t h time of the temperature difference AT, between

the edge and the midchord i n degrees Fahrenheit, i s shown a t the top of
f i g u e 3 . Heat w a s applied t o the p l a t e f o r about 20 seconds a t which
M

ji 344
?
3

time the temperature difference between the edge and the midchord w a s
about 190° F. I n the lower portion of the figure the r a t i o of the fre-
quency measured during heating f t o that of the unheated p l a t e f o i s
plotted as a function of time i n seconds f o r the f i r s t f i v e n a t u r a l modes.
The mode shape and corresponding i n i t i a l frequency are s h a m f o r each
curve. O f these f i v e modes, the f i r s t t o r s i o n mode underwent the l a r g e s t
change i n frequency, a reduction of about 35 percent. The f i r s t chordwise
bending mode w a s l e a s t affected by t h i s type of heating. Since the p l a t e
w a s not i n i t i a l l y f l a t , the heating a l s o caused the p l a t e t o deform.
These deformations were primarily t o r s i o n a l and, therefore, an analysis
t h a t considered only t o r s i o n a l deflections w a s made f o r the e f f e c t of
heating on the t o r s i o n a l mode of vibration. Such an analysis i s given
i n reference 3 . Some results of t h i s analysis are shown i n f i g u r e 4.

I n f i g u r e 4 the r a t i o of the frequency of a heated p l a t e t o t h a t of

a f l a t unheated p l a t e i s shown as a function of the r a t i o of the tempera-
t u r e difference between the edge and the midchord AT t o the t h e o r e t i c a l
temperature difference required t o produce t h e m 1 buckling of t h e f l a t
p l a t e AT,,. Curves a r e shown f o r various values of i n i t i a l d i s t o r t i o n
which have been expressed as a t w i s t parameter .8, For the p e r f e c t
p l a t e , 8, = 0, the frequency r a t i o decreases u n t i l the buckling tempera-
t u r e of the p l a t e i s reached. A t this time the p l a t e buckles; and, since
the s t i f f n e s s of the buckled p l a t e i s g r e a t e r than t h a t of the f l a t p l a t e ,
the t o r s i o n a l frequency begins t o increase. For a p l a t e which has an
i n i t i a l twist, the heating has two e f f e c t s . The thermal s t r e s s e s lower
the frequency, but, since the p l a t e i s a l s o deforming, i t s s t i f f n e s s due
t o t w i s t i s increasing. The frequency therefore reaches a minimum below
the buckling temperature and any f u r t h e r heating causes the frequency t o
increase. The p l a t e t e s t e d had a value of 8, of 0.06 which corresponded
t o a t i p t w i s t of about one-third degree. As the frequency h i s t o r y indi-
cates, the s t i f f n e s s a c t u a l l y began t o increase w h i l e heat w a s s t i l l being
applied.

Figure 5 shows a comparison between measured and calculated frequency

h i s t o r i e s . The r a t i o of t h e frequency during heating t o the frequency of
the unheated p l a t e i s shown as a function of time i n seconds. A small-
deflection theory which neglects the e f f e c t s of d i s t o r t i o n on the s t i f f n e s s
overestimates the frequency change. When a large-deflection theory which
includes the e f f e c t s of d i s t o r t i o n i s used, the agreement i s s u b s t a n t i a l l y
improved.

Double-Wedge Section

The double-vedge section t e s t e d w a s made of s t a i n l e s s s t e e l 1 inch

thick and 20 inches square. Both the upper and lower surfaces were

3.35
4

subjected t o a constant heat input by means of radiant heating. The

temperature d i s t r i b u t i o n i n the chordwise d i r e c t i o n i s then primarily
a function of the mass d i s t r i b u t i o n . Again the changes i n s t i f f n e s s
were measured by measuring the changes i n n a t u r a l frequency. Some of
the t e s t r e s u l t s a r e shown i n figure 6.

The v a r i a t i o n of temperature r i s e T with time f o r an edge, the

surface a t the midchord, and a point a t the centroid of the section i s
shown a t the top of f i g u r e 6. The changes i n frequency r a t i o with time
are shown f o r t h e first f i v e modes i n the lower p a r t of the figure. The
mode shapes and t h e i r corresponding s t a r t i n g frequencies are shown f o r
each curve.

The reductions i n frequency which occurred during heating are sur-

p r i s i n g l y small i n comparison with the reductions calculated by Budiansky
and Mayers ( r e f . 4) f o r the aerodynamic heating of similar wings instan-
taneously accelerated t o moderately high Mach numbers. The discrepancies
a r i s e because of the marked difference between aerodynamic heating and
this type of r a d i a n t heating. I n aerodynamic heating, the temperature
a t the leading and t r a i l i n g edges i s limited by the boundary-layer tem-
perature. I n the radiant heating, the temperature r i g h t a t the leading
and t r a i l i n g edges becomes prohibitively high before a temperature d i s -
t r i b u t i o n i s obtained which r e s u l t s i n l m g e s t i f f n e s s changes. A theo-
r e t i c a l calculation was made f o r the f i r s t t o r s i o n mode by using the
same method as t h a t used by Budiansky and Mayers but based on the experi-
mental temperature d i s t r i b u t i o n obtained from the radiant-heating t e s t .
These calculations gave a frequency reduction about the same as t h a t
obtained i n the t e s t . Therefore, l a r g e r reductions i n frequency probably
would have been obtained i f the t e s t could have been continued f o r a
longer time o r i f the temperature d i s t r i b u t i o n had more closely simulated
that generated aerodynamically.

Multiweb-Wing Section

The t h i r d section t e s t e d w a s a 5-percent-thick multiweb-wing s e c t i o n

of aluminum-alloy construction with a 20-inch chord and span. This speci-
men i s the same as one which w a s t e s t e d i n the p r e f l i g h t j e t of the
Langley P i l o t l e s s A i r c r a f t Research S t a t i o n a t Wallops Island, Va., a t a
Mach number of 2. A description of the t e s t of this model, designated
MW-2, i s given i n reference 1. The radiant-heating t e s t did not simulate
the aerodynamic heating c o r r e c t l y although an e f f o r t was made t o reproduce
the average heat input. The t e s t r e s u l t s a r e shown i n figure 7.

The temperature i n degrees Fahrenheit i s p l o t t e d as a function of

time i n seconds f o r a point on the skin and a t the center of a spar a t
the top of f i g u r e 7. The changes i n the frequency r a t i o with time have
been shown f o r f i v e modes of vibration a t the bottom of the f i g u r e . The
 5

l a r g e s t reduction i n frequency, as shown by the lowest curve, occurred

f o r a mode which involved chordwise bending. The smallest reduction i n
frequency occurred f o r the f i r s t t o r s i o n mode.

The radiant heating has not simulated the aerodynamic heating very
well; therefore, these results cannot be applied d i r e c t l y t o aerodynamic
t e s t s . However, the type of frequency reductions obtained here are s i g -
n i f i c a n t . The similar model (described i n r e f . 1) t e s t e d a t a Mach
number of 2 f l u t t e r e d i n a mode which involved a large amount of chord-
w i s e bending as d i d the one which had the l a r g e s t frequency reduction.

A t the present t i m e , no t h e o r e t i c a l method i s available f o r pre-

dicting the e f f e c t s of t r a n s i e n t heating on the frequencies of the more
complicated types of modes which a r e t h e important ones f o r this
structure .
CONCLUDING REMARKS

Three types of simple wing s t r u c t u r e s have been t e s t e d under various

radiant-heating conditions and changes i n s t i f f n e s s as indicated by
changes i n natural frequency of vibration were measured. I n order t o
measure the changes i n frequency during t r a n s i e n t heating, a resonance-
following system w a s developed which keeps the specimen vibrating a t i t s
natural frequency.

For temperature differences near those required f o r thermal buckling,

i n i t i a l d i s t o r t i o n s have a marked e f f e c t on the frequency. Calculations
which include the e f f e c t s of these d i s t o r t i o n s are found t o be i n good
agreement with the measured frequency.

The r e s u l t s obtained from these tests indicate that the magnitude of

the e f f e c t s of thermal s t r e s s depends strongly on the manner i n which the
s t r u c t u r e i s heated. Therefore, i n order t o determine the e f f e c t s of
aerodynamic heating on the s t i f f n e s s of wing s t r u c t u r e s by means of lab-
oratory radiant-heating t e s t s , care m u s t be taken t o simulate c l o s e l y the
temperature d i s t r i b u t i o n s produced aerodynamically.
 REFERENCES

1. Heldenfels, Richard R., and Rosecrans, Richard: Preliminary Results

of Supersonic-Jet Tests of Simplified Wing Structures. NACA
RM L53E26a, 1953.

2. Vosteen, Louis F., and Fuller, Kenneth E.: Behavior of a Cantilever

Plate Under Rapid-Heating Conditions. NACA RM L55E20cY1955.

3. Heldenfels, Richard R., and Vosteen, Louis F.: An Approximate Analysis

of the Effects of Large Deflections and I n i t i a l Twist on the Tor-
s i o n a l S t i f f n e s s of a Cantilever P l a t e Subjected t o Thermal S t r e s s ,
(Prospective NACA paper. )

4. Budiansky, Bernard, and Mayers, J.: Influence of Aerodynamic Heating

on the Effective Torsional S t i f f n e s s of Thin W i n g s . Jour. Aero.
Sci., vol. 23, no. 12, %c. 1956, pp. 1081-1093, 1108.
 7

MODEL CONFIGURATIONS

PLATE WEDGE MULTIWEB

Figure 1

RESONANCE-FOLLOWING SYSTEM

SIGNAL MOTOR PHASE

GENERATOR DETECTOR

Figure 2
8

FREQUENCY H ISTO R IES

CANTILEVER PLATE

0 5 IO 15 20
T I M E ,SEC

Figure 3

THEORETICAL FREQUENCY CHANGES

FIRST TORSION MODE
1.5

1.0

-
f
f0

.5

i
0 .5 I.o 1.5 2.0
AT/ATcr
P

Figure 4
 9

COMPARISON OF FREQUENCY HISTORIES

FIRST TORSION MODE
1.0

f
-
f0
.5 -

SMALL- DEFLECTION

0 4 8 12 16 20 24
TIME, SEC

Figure 5

FREQUENCY HISTORIES
DOUBLE WEDGE

1,200

T,OF 600

0
I .o

J- .95
f0

.9
I I I I

0 2 4 6 8
TIME ,SEC

Figure 6
351
10

FREQUENCY H ISTOR IES

MULTI W EB

400r

T ,OF 2 0 0 ~

I I , I I I
0 fo ,CPS

-
f .95 -
f0

0 I 2 3 4 5 6
TIME ,SEC
Figure 7
 I

EFFECTS OF RAPkD IlEAlcING ON STRENGTH

OF AIRFRAME COMPONENTS
By Richard A. Pride, John B. Hall, Jr.,
and Melvin S. Anderson

Langley Aeronautical Laboratory

SUMMARY

Results of several experimental investigations axe presented which

indicate the effects of rapid heating on the bending strength of multiweb
beam and ring-stiffened cylinders. It is shown that thermal stresses
reduce the bending load carried at buckling by both beams and cylinders.
The influence of thermal stress on maximum load is found to depend
largely on the mode of buckling. For beams that buckle locally, no
apparent effect of thermal Stress on the maximum load has been found.
A reduction in maximum load has been observed for beams that buckle in
the wrinkling mode and for cylinders.

INTRODUCTION

Aerodynamic heating rates currently contemplated in airplanes and

missiles may impose severe thermal stresses on primary structures. This
paper considers the influence of such thermal stresses on the bending
strength of multiweb beams and ring-stiffened cylinders.

Observation of the behavior of various types of structures under

combinations of loads and thermal stresses has indicated that the effect
of thermal stress on buckling is to reduce the bending load which may
be carried prior to buckling and that the effect of thermal stress on
maximum load may be correlated with the type of stress-shortening
diagram of the structure.

Some typical stress-shortening diagrams an2 associated buckling

modes are presented in figure 1. Local buckling has a stress-shortening
curve that permits considerable increase in both stress and shortening
to occur after buckling and prior to failure. Within this region of
increasing stress and shortening, it is possible that a redistribution
of stress would alleviate the detrimental effect of thermal stress on
m a x i m load.
 In contrast, the stress-shortening curve for a structure such as
a ring-stiffened cylinder usually drops abruptly after buckling (fig. 1).
Hence, a reduction in buckling stress produced by thermal stress should
be reflected as a corresponding loss in maximum strength. A stress-
shortening diagram in which considerable shortening occurs for negli-
gible increase in stress exists for structures such as multiweb beams
with formed-channel webs which buckle in a mode known as wrinkling.
This mode is characterizedby buckles extending across the chord of the
beam. The considerable shortening that can occur after buckling indi-
cates that some influence of thermal stress on failure may be anticipated;
however, this influence m y not be as great as for the cylinder.

Tests to explore themnal-stress effects on the bending strength of

these three types of structures have been made and some of the results
are presented in preliminary form in this paper.

SYMBOLS

A cross-sectional area, sq in.

b plate width, in.

C distance from neutral axis to center of skin, in.

EB secant modulus for bending stress in skin, ksi

ES secant modulus for skin, ksi

Ew secant modulus for web, ksi

I moment of inertia, in.4

k plate buckling coefficient

2 cylinder ring spacing, in.

M bending moment, in-kips

r cylinder radius, in.

S peripheral distance, in.

 3

T temperature, OF
T average temperature, 9
?? temperature rate with respect t o time, %/see

t p l a t e thickness, in.

a coefficient of thermal expansion, per ?F

? p l a s t i c i t y reduction f a c t o r

(r stress, ksi
-
(r average s t r e s s , k s i

P Poisson's r a t i o

Subscripts:

B bending

cr critical

CY 0.2-percent-offset compressive y i e l d

f failure

S skin

w web

BEAM S!EWVG'.I!K

The t e s t setup f o r applying both bending loads and rapid heating

on beams is shown i n figure 2. A one-cell beam i s shown supported on
columns about one-third t h e distance from each end and loaded i n bending
by t h e hydraulic jacks a t the ends. Rapid heating i s applied t o t h e
t o p and bottom skins of the beam by means of high-intensity quartz-tube
lamp r a d i a t o r s located behind the r e f l e c t o r s . S t r a i n gages and deflec-
t i o n pickups mounted on t h e cold webs give an indication of o v e r a l l
beam behavior during t h e t e s t .
4

F a i l u r e Associated With Local Buckling

A l a r g e number of tests have been made on one-cell beams t h a t

buckle i n t h e l o c a l mode. (Some of t h e results have been reported i n
ref. 1.) A representative temperature d i s t r i b u t i o n around t h e cross
s e c t i o n is shown i n figure 3 f o r a beam 7.35 inches square with a w a l l
thickness of 0.152 inch. The v a r i a t i o n of temperature with peripheral
distance f o r one quadrant of t h e beam i s shown. For t h e maximum heating
time i n these t e s t s , t h e temperature of t h e center of t h e skin rises t o
about 600° F, while t h e temperature of most of t h e webs rises t o only
about 100' F. Temperature d i s t r i b u t i o n s such as t h i s w i l l produce l a r g e
thermal s t r e s s e s i n these beams.

For these aluminum-alloy beams, skin temperatures i n t h e range

from 400° t o 600° F a l s o produce d e t e r i o r a t i o n i n t h e material properties.
I n order t o separate t h e loss i n beam s t r e n g t h caused by d e t e r i o r a t i o n
of material properties from t h e l o s s i n beam s t r e n g t h caused by thermal
s t r e s s , a d i f f e r e n t type of t e s t w a s made i n which al1,sides of t h e beam
were uniformly heated a t a high r a t e . The t e s t r e s u l t s f o r l o c a l buckling
of these one-cell beams are shown i n f i g u r e 4. The bending load s t r e s s
i n t h e s k i n of t h e beam i s p l o t t e d against t h e average skin temperature.
For t h e uniform heating case t h e skin temperature i s t h e average temper-
a t u r e around t h e e n t i r e beam cross s e c t i o n j whereas, t h e average s k i n
temperature i s indicated f o r t h e nonuniform heating case i n which t h e
skins are heated a t about 1 00' F/sec. For t h e nonuniform heating tests
t h e existence of a difference i n temperature between skins and webs i s
implied as w a s shown i n f i g u r e 3 .

Results a r e given i n figure 4 f o r beams having two d i f f e r e n t values

of width-thickness r a t i o f o r t h e skin; f o r one beam proportion buckling
occurred i n t h e e l a s t i c - s t r e s s range so t h a t a l a r g e increase i n stress
a f t e r buckling was possible, and f o r t h e other beam proportion buckling
occurred i n t h e p l a s t i c range s o t h a t only a s l i g h t increase i n stress
beyond buckling was obtained. The upper curve f o r b / t = 48 gives t h e
calculated buckling stress f o r uniformly heated beams and shows t h e
e f f e c t of material d e t e r i o r a t i o n on t h e e l a s t i c buckling stress. The
buckling stress was determined from t h e following equation:

The s l i g h t reduction i n buckling s t r e s s shown i s t h e ' r e s u l t of changes

i n e l a s t i c modulus with temperature. The open square t e s t points i n
f i g u r e 4 are t h e confirming t e s t s f o r t h i s curve. - ,
 5

The lower curve f o r b / t = 48 determined from t h e following

equation i s the calculated amount of bending load s t r e s s required t o
produce buckling when thermal s t r e s s i s present along w i t h material
deterioration :

Equation ( 2 ) follows d i r e c t l y from equations (1)and ( 2 ) of reference 2.

The region between the t w o lower curves i n f i g u r e 4 i s a measure of t h e
thermal s t r e s s as calculated by the l a s t term i n equation ( 2 ) . The
s o l i d square t e s t points indicate confirming t e s t s f o r t h i s e l a s t i c
thermal-stress buckling condition.

For b / t = 30 the buckling s t r e s s i s i n t h e p l a s t i c range f o r t h i s

material. The upper curve i n f i g u r e 4 gives the calculated e f f e c t of
material deterioration on buckling s t r e s s f o r uniform heating. The
buckling s t r e s s was again determined from equation (1)by using the
p l a s t i c i t y reduction coefficient 7 = Es/E. The lower curve gives t h e
calculated amount of bending load s t r e s s required t o produce buckling
when both material d e t e r i o r a t i o n and thermal s t r e s s are considered
(eq. ( 2 ) ). Again the region between these t w o cwves represents t h e
amount of thermal s t r e s s present a t buckling. The t e s t s axe i n agree-
men% w i t h t h e calculations. It should be noted t h a t about one-half of
t h e t e s t r e s u l t s presented i n f i g u r e 4 were obtained f o r beams that were
loaded f i r s t and then heated, with t h e remaining beams heated f i r s t and
then loaded. No discernable e f f e c t s of sequence could be found.

Thus, it appears that l o c a l buckling of beams w i t h i n t e g r a l webs

can be predicted a t elevated temperatures e i t h e r with uniform heating
o r with temperature gradients, and i n both t h e e l a s t i c and t h e p l a s t i c
s t r e s s ranges.

The maximum strength of these same beams is given i n f i g u r e 5.

The curves show the maximum strength of the beams f o r uniform heating
calculated from the following equation:

- -- -
Of
Mf“
(3)
I
6

where

Equation (4) i s taken from reference 1 and i s dependent only on beam

dimensions and material properties a t elevated temperature. The calcu-
l a t e d curves of f i g u r e 5 correspond t o the same uniform heating case
as t h e upper buckling curves f o r each value of b / t i n f i g u r e 4. A
considerable difference e x i s t s between the buckling stress and failure
f o r b / t = 48. A t b / t = 30 t h i s difference i n s t r e s s i s p r a c t i c a l l y
gone as a r e s u l t of p l a s t i c buckling; however, a considerable amount
of beam d e f l e c t i o n occurred between buckling and f a i l u r e , even though
t h e s t r e s s increased only s l i g h t l y . The open t e s t points corresponding
t o uniform heating indicate t h a t t h e f a i l u r e calculation i s v a l i d f o r
both types of beams. The s o l i d points represent t e s t conditions a t
maximum load f o r the beams which had developed large thermal s t r e s s e s
a t buckling. As indicated by t h e intermingling of t h e open and s o l i d
t e s t points, no e f f e c t of thermal s t r e s s i s evident on t h e maximum
load. This r e s u l t i s equally t r u e f o r beams t h a t buckled e i t h e r
elastically or plastically.

F a i l u r e Associated With Wrinkling

"he preliminary r e s u l t s obtained f o r a s e r i e s of multiweb beams

designed t o a i r c r a f t proportions and fabricated from 17-7 PH s t a i n l e s s
s t e e l are given i n f i g u r e 6. The i n t e r i o r s t r u c t u r e consists of t h i n
formed-channel webs r i v e t e d t o t h e skins. I n t h i s p a r t i c u l a r configu-
r a t i o n , no post-buckling strength w a s found i n the room-temperature
t e s t . Therefore, t h e upper curve calculated from equation (1)gives
the buckling s t r e n g t h and failure strength f o r uniform heating based
on material d e t e r i o r a t i o n with temperature. The lower curve gives t h e
calculated bending load s t r e s s t h a t produces buckling under the
nonuniform-heating t e s t condition (eq. ( 2 ) ) . The region between t h e
curves then represents t h e amount of thermal s t r e s s t h a t w a s calculated
t o be present i n t h e t e s t s . The c i r c l e s give t h e failure-test r e s u l t s
obtained from t h r e e beams. The r e s u l t s from two of the t e s t s f o r non-
uniform heating indicate t h a t thermal s t r e s s i n t h e beam reduced t h e
maximum load. The maximum possible reduction was not large i n t h e
p a r t i c u l a r beam proportions tested, and it remains t o be determined
whether o r not corresponding r e s u l t s w i l l be obtained i n beam proportions
where t h e induced thermal s t r e s s e s a r e a l a r g e r percentage of t h e beam
buckling s t r e s s .
 7

CYLINDE3 STRENGTH

Tests have been completed on 2024-T3 aluminum-alloy ring-stiffened

c i r c u l a r cylinders ( t a b l e I ) , and a program has been s t a r t e d w i t h 17-7PH
s t a i n l e s s s t e e l cylinders. The t e s t setup i s shown i n figure 7. The cyl-
inder i s loaded i n bending by t h e large weight cage a c t i n g through a l i n k -
age system. A quartz-tube lamp r a d i a t o r constructed mound t h e outside
of t h e cylinder t e s t section permits rapid heating of t h e cylinder skin
with heat flowing i n t o t h e rings primarily by conduction from t h e skin.

The e f f e c t of heating a t d i f f e r e n t temperature r a t e s on cylinder

strength i s shown i n f i g u r e 8. The s o l i d curve i s t h e calculated bending
strength of t h e ring-stiffened cylinder f o r uniform heating of both skin
and rings and i s based on t h e room-temperature t e s t r e s u l t s of reference 3
The decrease i n strength shown by the curve i s caused by t h e change i n
e l a s t i c modulus a t high temperatures.

Three nominally i d e n t i c a l cylinders were loaded with two-thirds of

t h e room-temperature f a i l i n g load i n bending and then heated t o f a i l u r e
a t temperature rates of '1 F/sec, 20' F/sec, and PoF/sec. With t h e
slowest temperature rate which corresponded e s s e n t i a l l y t o a uniform
heating t e s t , t h e r i n g temperature followed t h e skin temperature very
closely. For t h i s case it would be expected t h a t t h e buckling failure
would be influenced mainly by t h e change i n e l a s t i c modulus with tempera-
t u r e . The close agreement of the t e s t a t lo F/sec with t h e curve f o r
uniform heating checks t h e v a l i d i t y of such a calculation. For the two
higher heating rates, r i n g temperatures lagged the s k i n temperature by
a large amount. The lover curve i n f i g u r e 8 i s f a i r e d f o r these rapid
heating t e s t s and t h e gap between t h e curves indicates t h e e f f e c t of
restrained expansion. Variations i n cylinder diameter caused by r e s t r a i n t
of thermal expansion of t h e skin by t h e rings were observed i n t h e t e s t s .
The bulging between rings and circumferential compression s t r e s s produced
by t h i s r e s t r a i n t a r e t h e most probable causes f o r t h e reduction of
strength of t h e two cylinders subjected t o t h e high heating rates. For
these p a r t i c u l a r cylinder proportions, temperature rates s u b s t a n t i a l l y
l e s s than t h e 20° F/sec would be required t o avoid t h i s reduction i n
strength.

The e f f e c t of changing the r i n g spacing i s shown i n f i g u r e 9 f o r a

constant temperature r a t e of 9 0' F/sec. A t room temperature t h e same
value of failure stress i s obtained f o r 2/r equal t o 1/2 o r greater.
A t a uniform elevated temperature a single calculated curve represents
t h e strength of these cylinders. For Z/r equal t o 1/4, t h e small r i n g
spacing r e s u l t s i n an increased bending strength f o r uniform heating con-
d i t i o n s as shown by t h e top s o l i d curve. The dashed curves a r e drawn t o
t h e i r respective t e s t points f o r nonuniform heating a t the high r a t e .
The gap between t h e dashed curves and t h e corresponding s o l i d curves i s

359
8

a measure of the loss in bending strength caused by the restrained skin

expansion. As the ring spacing increases, the loss in bending strength
also increases.

It is believed to be significant that the mode of failure of these

tests was the usual diamond buckle pattern and not a circumferential
bulge such as would correspond to the eccentricity of the restrained
skin. This behavior and the greater reduction in failure strength with
increasing ring spacing suggest that the circumferential compression due
to restrained skin expansion interacts more with the bending stress than
does the eccentricity. This interaction becomes more serious as ring
spacing is increased, inasmuch as the ratio of the actual to the critical
circumferential compression stress increases.

CONCLUDIIVG REMAHKS

It may be concluded fromthese tests that thermal stresses will

interact with bending load stresses to reduce the buckling load carried
by a structure. Depending upon the mode of buckling, the thermal stresses
may or may not affect the magnitude of the maximum lokd. In tests of
structures where increases in load are required to produce deformation
in the post-buckling range, no apparent effect of thermal stress on the
failure load has been found; in structures which tend to deform with loss
of load beyond buckling, tests indicate that a loss of strength due to
thermal stress can be expected.

1. Pride, Richard A.: Transient Heating Effects on the Bending Strength

of Box Beams. M. S. Thesis, Va. Polytechnic Inst., 1956.
2. Pride, Richard A,: An Investigation of the Effects of Rapid Skin
Heating on Box Beams Loaded in Bending. NACA RM L55Bo3, 1955.

3. Peterson, James P.: Bending Tests of Ring-Stiffened Circular Cylinders.

NACA TN 3735, 1956.
F 9

%
3r-z
k d
ald

c l m
+I
0
;I
d o

co f c o
Y ? ! ? ?
0

k
d
c
d
rl
x
u
STRESS-SHORTENING CURVES FOR VARIOUS STRUCTURAL COMPONENTS

STRESS

SHORTENING

Figure 1

TEST SETUP FOR RAPID HEATING AND BENDING OF BEAMS

L-91930.1
Figure 2
365
 TEMPERATURE DISTRIBUTION IN BEAM CROSS SECTION
2014-T6 ALUMINUM ALLOY; tmlOO°F/SEC
600 -
I
I

400 -

200 -

I
I I I I I I I I
0 2 4 6 8
s, IN.
PERIPHERAL DISTANCE, S,

Figure 3

RAPID -HEATING EFFECTS ON BEAM BUCKLING

2014-T6 ALUMlNUM ALLOY;? u 100DF/SEC

60 -

40 -
BENDM
LOAD -
STRESS,

\T
KSI U

20 -
bt
-=48
-

I I I I I I I

Figure 4 3G3
I "
*i
..
0 e. .e
r
a.
< .
e
a
.. . -...- - - 0. ..-
RAPID-HEATING EFFECTS ON.BEAM FAILURE
2014 -T6 ALUMINUM ALLOY 100' F/SEC

60 - b
-=30

40 -
BENDING
LOAD -
STRESS,
KSI
20 - \
-

0 200 400 600 800

SKIN TEMPERATURE, 7

Figure 5

RAPID -HEATING EFFECTS ON BEAMS (WRINKLING MODE)

17-7 PH STAINLESS STEEL; IOO'F/SEC
140
BUCKLING AND FAILURE,
UNIFORM HEATING

BUCKLING WITH
IO0 THERMAL STRESS

BENDING 80 -
LOAD
STRESS, -
bS 25
KSI 60- tS

40 -
.025 tw
- I 0.32
20 - tS

I I I I I I I I
TEST SETUP FOR RAPID HEATING AND BENDING OF CYLINDERS

Figure 7 L-90701.1

RAPID -HEATING-RATE EFFECT ON RING-STIFFENED

CYLINDER STRENGTH
2024-T3ALUMINUM ALLOY;*=+; +=300

20 r
UNIFORM HEATING

NONUNIFORM HEATING
BENDING
IO -
STRESS,
KSI

0
t
I I I
200
I
400
I I I
600
SKIN TEMPERATURE, OF

Figure 8 365
14 '.

RING SPACING EFFECT ON STRENGTH OF RAPIDLY HEATED

RING -STIFFENED CYLINDERS
2024-T3 ALUMINUM ALLOY j f = 3 0 0 ; T = S O OF /SEC

25 -

20 -

BENDING l5
-
LOAD
STRESS,
KSI 10 -

UNIFORM HEATING
5-
--- NONUNIFORM HEATING
I I I I I I I

Figure 9
 THE COMBINATIONS OF THE3MAL AND LOAD STRESSES FOR
THE ONSET OF P m BUCKLING IN PLATES
By George W. Zender and Richard A. Pride

Langley Aeronautical Laboratory

A simple and practical method for evaluating the onset of permanent

buckling in plates in the presence of combined thermal and load stresses
is outlined. A particular application of the method shows reasonable
agreement with tests of 17-7PH stainless-steel square tubes. The
results indicate that the load stress which the plate can support at the
onset of permanent buckling is substantially reduced as the temperature
difference of the plate and adjoining members increases.

INTRODUCTION

The use of high-density materials in supersonic aircraft causes con-

cern with respect to buckling of the thin skin surfaces, particularly
where permanent buckles may develop in the structure. Methods of analy-
sis are needed which include the effects of thermal stresses in combina-
tion with the usual load stresses. In order to provide such methods,
the behavior of plates due to combined thermal and load stresses is being
studied; this paper presents initial results obtained for predicting the
onset of permanent buckling and compares the results with experiment.

SYMBOLS

AS cross-sectional area of skin, sq in.

Aw cross-sectional area of webs or stringers restraining thermal

expansion of skin, sq in.

b width of skin, in.

e unit shortening of skin

2

ES modulus of e l a s t i c i t y of skin materialrat average skin tem-

perature, k s i

Ew modulus of e l a s t i c i t y of web material a t average web tempera-

ture, ks i

t thickness of skin, i n .

m average temperature of skin minus average temperature of web,

?F

U l i n e a r coefficient of thermal expansion, per ?F

%r s t r a i n a t which buckling i n i t i a t e s

E e last ic l i m i t s t r a i n
e2

Q average s t r e s s over cross-sectional area of skin, k s i

average load s t r e s s over cross-sectional area of skin, k s i

aL

average thermal s t r e s s over cross-sectional area of skin, k s i

OT

METHOD OF ANALYSIS

A simple approximation f o r t h e beginning of permanent buckling i n

a p l a t e subjected t o compressive load i s suggested by t h e experimental
observation t h a t permanent buckling begins when t h e unit shortening of
t h e p l a t e i s about t h e same value as the e l a s t i c - l i m i t s t r a i n of t h e
material. This concept appears useful f o r approximating t h e compressive
load required f o r t h e onset of permanent buckling i n a p l a t e i n t h e pres-
ence of thermal stresses. Consider a p l a t e which has been shortened
beyond t h e value required f o r buckling. (See f i g . 1.) I n t h e usual
sense, t h e shortening of t h e p l a t e comes from compressive loads; however,
i n e f f e c t , shortening a l s o occurs when the p l a t e i s heated but t h e t h e r -
mal expansion i s restrained, such as would occur i f t h e edges of t h e
p l a t e were bounded by s t r i n g e r s or shear w e b s a t a lower temperature.
This e f f e c t i v e shortening due t o restrained thermal expansion i s t h e
difference i n length between the restrained and the unrestrained p l a t e
when heated. The lower curve of figure 2 shows t h e manner i n which
t h e e f f e c t i v e u n i t shortening due t o restrained thermal expansion
increases w i t h the difference i n average temperature of t h e p l a t e and
F 3

t h e adjoining member. The dashed l i n e i n f i g u r e 2 indicates t h e c r i t i c a l

s t r a i n o r t h e u n i t shortening f o r buckling of t h e p l a t e . The upper s o l i d
curve indicates t h e e l a s t i c - l i m i t s t r a i n which decreases somewhat with
increasing temperature. For a given temperature difference, then, t h e
p l a t e i s i n a state of shortening due t o r e s t r a i n e d thermal expansion;
additional shortening by compressive loading causes t h e p l a t e t o buckle;
and f u r t h e r compressive-load shortening causes t h e buckles t o deepen but
t h e buckles are not permanent u n t i l t h e l i m i t given by t h e upper curve
i s exceeded. The region between t h e s o l i d l i n e s , then, defines the per-
missible amount of compressive-load shortening which may be applied i n
conjunction w i t h t h e e f f e c t i v e shortening from r e s t r a i n e d thermal expan-
sion without causing permanent buckling.

SPECIMENS AMI =OD OF TESTING

I n order t o t e s t t h e v a l i d i t y of t h e foregoing approximate analysis,

1
t e s t s were performed on --inch-thick 17-7 PH s t a i n l e s s - s t e e l p l a t e s fab-
16
r i c a t e d i n t o square tubes by welding the corners. The tubes were 32 inches
long and had b / t r a t i o s of 40, 60, or 80. Two opposite walls of t h e
tubes ( s k i n s ) were exposed t o heat supplied by two banks of quartz-tube
r a d i a t o r s as shown i n figure 3 . The other two w a l l s (webs) were shielded
from t h e r a d i a t o r s by aluminum p l a t e s which ran lengthwise of the tube
and projected diagonally outward from each corner of t h e tube. The shields
have been removed i n figure 3 i n order t o show t h e tube. Temperature d i s -
t r i b u t i o n s i n t h e tubes were obtained w i t h thermocouples, and records of
t h e extension near t h e corners, along 15 inches of t h e length, were
obtained w i t h four d i f f e r e n t i a l transformers. Figure 3 shows t h e setup
when t h e square tubes were subjected t o heat without end load. The same
t e s t setup w a s placed i n a t e s t i n g machine and compressive loads were
applied i n combination w i t h heat. I n addition, compression t e s t s of
tubes a t room temperature were performed. The square tubes were unloaded
and/or cooled t o room temperature a f t e r loading and/or heating, and pro-
f i l e s of t h e amplitude of t h e permanent buckles were obtained with a
pantograph apparatus which multiplied the amplitude by a f a c t o r of 11.
For each run, t h e average of t h e buckle depths (measured from c r e s t t o
valley or twice maximum amplitude) i n t h e skins w a s obtained. Figure 4
shows an example of t h e permanent-buckle information obtained f o r t h e
square tubes subjected t o heat. The p a r t i c u l a r data shown a r e f o r t h e
square tube subjected t o heat without compressive load and w i t h b / t = 60.
The ordinate shown i n f i g u r e 4 is t h e difference of t h e average tempera-
tures of the skin and web of t h e tube. Before t h e tube w a s subjected t o
heat, t h e average value of buckle depth w a s obtained from t h e pantograph
measurements and i s indicated i n f i g u r e 4 ( i n i t i a l imperfection). After
t h e tube had been subjected t o an elevated temperature and cooled t o room
temperature, t h e permanent-buckle depth w a s s l i g h t l y l a r g e r than t h e
4

i n i t i a l imperfection. Subsequent t e s t s on t h e same tube but with pro-

gressively higher values of temperature r e s u l t e d i n progressively l a r g e r
values of permanent-buckle depth. The temperature difference when t h e
permanent-buckle depth began t o exceed t h e i n i t i a l imperfection w a s
defined as t h e s t a r t of permanent buckling, as i s indicated i n f i g u r e 4.
Similar information was obtained f o r a l l tubes t e s t e d except that load
s t r e s s w a s p l o t t e d i n place of temperature difference f o r t h e square
tubes subjected t o compressive load but not t o heat. The b a d s t r e s s
i n t h e skin at the start of permanent buckling f o r t h e square tubes sub-
jected t o both heat and compressive load w a s evaluated from t h e extension
measurements and t h e t o t a l load on t h e tube since t h e compressive s t r e s s
i n t h e skin f o r t h i s case d i f f e r s considerably f r o m t h a t i n the webs.

RESULTS AND DISCUSSION

Figure 5 presents a comparisoii of t h e experimental r e s u l t s with

curves obtained from t h e approximate analysis previously given. The
ordinate shown i n f i g u r e 5 i s t h e average compressive load s t r e s s f o r
the onset of permanent buckling i n the p l a t e , and t h e abscissa is t h e
difference i n average temperatures of t h e skin and web. The d e t a i l s
of evaluation of t h e curves a r e given i n t h e appendix. It i s apparent
that the load s t r e s s which t h e p l a t e can support a t t h e onset of perma-
nent buckling is s u b s t a n t i a l l y reduced as t h e temperature difference
increases.

I n general, reasonable agreement of t h e approximate analysis and

t h e experiment e x i s t s . Somewhat b e t t e r agreement e x i s t s when load s t r e s s
only i s applied than when large thermal s t r e s s e s a r e present. The poorer
agreement i s l a r g e l y a t t r i b u t e d t o the assumption that the properties of
t h e heated skin a r e represented by properties f o r t h e average value of
skin temperature. This representation i s somewhat inaccurate a t high
values of temperature difference when r a t h e r large-temperature gradients
exist.

CONCLUDING R E m S

A simple and p r a c t i c a l method for evaluating t h e onset of permanent

buckling i n p l a t e s i n t h e presence of thermal s t r e s s e s i s given. A par-
t i c u l a r application of t h e method shows reasonable agreement w i t h t e s t s
of 17-7 PH s t a i n l e s s - s t e e l square tubes. The r e s u l t s indicate that the
load s t r e s s which t h e p l a t e can support a t t h e onset of permanent buckling
i s s u b s t a n t i a l l y reduced as t h e temperature difference of t h e p l a t e and
adjoining members increases.
 5

APPENDIX

EVALUATION OF LOAD S!?2RESS

Compressive stress-strain curve"s of 17-7 PH stainless-steel material

for a wide range of temperatures are presented in reference 1 and are
assumed to represent the material of the square tubes. Beyond kuckling,
the average stress in the buckled skin is approximated by the expression
(ref. 2):

which is in substantial agreement with the experimental measurements

obtained on the square tubes subjected to compressive loads without ther-
m a l stresses (room-temperature tests). The combined thermal and load
stress for the onset of permanent buckling is then

T +o',=E Sb/ Ee2Ecr

O'

The thermal stress as given by equation (2) of reference 3 is

OT = aEs (3)
1+-
EWAW
Equation ( 3 ) was modified for thermal stresses exceeding the buckling
stress by employing assumptions consistant with equation (1), in order
to allow for the effect of the reduced longitudinal stiffness of the
buckled skin. After evaluating the thermal stress, the load stress uL
was obtained from equation (2) and the results are shown by the curves
of figure 5 .

E
6

REFERENCES

1. Stein, Bland A . : Compressive Stress-Strain Properties of 17-7 PH

and AM 350 Stainless S t e e l Sheet a t Elevated Temperatures. (Pro-
spective NACA paper. )

2. Von K d d n , Theodor, Sechler, Ernest E . , and Donnell, L . H.: The

Strength of Thin Plates i n Compression. A.S.M.E. Trans., APM-54-5,
vol. 54, no. 2, Jan. 30, 1932, pp. 53-57.

3 . Pride, Richard A.: An Investigation of t h e Effects of Rapid Skin

Heating on Box Beams Loaded i n Bending. NACA RM L55BO3, 1955.
 BUCKLED PLATE

Figure 1

METHOD O F ANALYSIS

UNIT
SHORTENING
w ELASTIC LIMIT

-----

RESTRAINED THERMAL EXPANSION

TEMPERATURE DIFFERENCE

Figure 2
.
8 : . ' "

TEST SETUP FOR THERMAL PERMANENT BUCKLING

OF SQUARE TUBE

Figure 3 147-64 1

EX PE R I MENTA L MEASUREMENTS OF PERMANENT - B U C K L E DEPTH

I 7 -7 PH STAINLESS S T E E L , b / t = 60

OF

':-n.l
0
IN ITlAL IMPERFECTION

.02 .04 .06

I

OB
I

IO
I

PERMANENT -BUCKLE DEPTH ,IN.

,

'12

374 Figure 4
 9

COMBINATIONS OF LOAD STRESS AND TEMPERATURE

DIFFERENCE FOR THE START OF PERMANENT BUCKLING

17-7 PH STAINLESS STEEL; AS’ AW

EXPERIMENT
b/t
0 40
A 60
0 80

0
0 100 200 300 400 500 600 700
TEMPERATURE DIFFERENCE, OF

Figure 5
F

SOME E X P E R m S WITH INSULATED STRUCTURES

By Richasd Rosecrans, Aldie E. Johnson, Jr.,

and W i l l i a m M. Bland, Jr.

Langley Aeronautical Laboratory

Two methods of insulating structures, one f o r short-time use and

t h e other f o r somewhat longer protection, were studied by radiant-heating
t e s t s and supersonic-wind-tunnel t e s t s . Results of these t e s t s indicated
t h a t e f f e c t i v e temporary protection can be obtained w i t h f a i r l y l i g h t -
weight structures; however, it is emphasized that care should be given
t o design d e t a i l s t o prevent f l u t t e r of t h e insulation cover i n some
cases.

INTRODUCTION

One proposed method of coping with t h e problem of aerodynamic heating

i s t o provide an insulating cover or radiation shield t o reduce t h e amount
of heat entering the s t r u c t u r a l airframe. In order t o be most effective,
such a covering must have high insulating value and as l i t t l e weight and
thickness as possible. If it i s necessary t o protect t h e basic s t r u c t u r e
f o r only a short time, a simple, thin, lightweight system can be employed.
For longer protection, a more elaborate method is necessary. Even t h e
more extensive insulating systems permit heat t o get through eventually;
therefore, f o r very long time exposure t o high temperatures, cooling must
be provided. This paper discusses the f i r s t two classes -
those which
do not require cooling.

TESTS

Experiments have been conducted upon several devices which appeared

l i k e l y t o afford s i g n i f i c a n t protection.

Solid Protective Covering

One short-time method involved a combination of insulation and heat

storage. I f t h e basic s t r u c t u r e is of lightweight material incapable of
2

withstanding high temperatures, it can be protected f o r a short period

by covering it with one o r more layers of heavier material which can
t o l e r a t e high temperature. Much of t h e heat which enters the s t r u c t u r e
during t h e f i r s t few seconds i s stored in t h e covering material. Heat c

f l o w t o t h e inner structure is delayed p a r t l y by t h e conductivity of t h e

material and p a r t l y by thermal resistance of t h e j o i n t s . I f t h e need f o r
protection ends by the time a damaging amount of heat enters t h e basic
structure, as it might i n the case of a short-range missile, t h e objective
i s accomplished. Such a method was used t o protect some tapered f i n s .
With no protection, the leading edge of a magnesium f i n melted i n l e s s
than 2 seconds. With the leading edge covered with one layer of 0.008-inch-
t h i c k glass c l o t h and one layer of 0.031 Inconel, the temperature of the
magnesium had r i s e n t o l e s s than l25O F i n 2 seconds. A heavier covering,
i n which one l a y e r covered the e n t i r e f i n and a second layer covered only
t h e leading edge, was even more effective. Figure 1 shows t h a t , while
t h e temperature of the outer layer of Inconel rose rapidly and reached
1,750° F in 5 seconds, t h e temperature of t h e inner magnesium s t r u c t u r e
had r i s e n t o only 90' F i n 2 seconds, when t h e unprotected f i n had f a i l e d ,
and t o l e s s than 300' F i n 5 seconds. During the f i r s t 2 seconds, l e s s
than 5 percent of the heat which entered t h e leading edge penetrated t o
t h e load-carrying structure.

Motion pictures of t h e t e s t showed t h a t , soon a f t e r t h e f i n entered

t h e j e t stream, t h e cover becage red near midspan, where the temperature
of t h e gases f r o m t h e j e t were highest. Because the thickness of the
covering varied between t h e leading and t r a i l i n g sections, the tempera-
t u r e of the cover along the chord was f a i r l y uniform. After 8.4 seconds,
t h e j e t pierced t h e protective cover and the f i n f a i l e d almost immediately.

Double-Wall Construction

Two insulating configurations ( f i g . 2), designed t o protect t h e a i r -

frame over a longer period of heating, were investigated. One was a
single-faced, corrugated-core sandwich of Inconel X; the other was a
s t a i n l e s s - s t e e l honeycomb sandwich. The exposed surface w a s supported
by t h e longitudinal corrugations i n one case and by t h e core and inner
face of the sandwich i n the other; i n both cases t h e sandwich was
separated from t h e load-carrying structure by e i t h e r bulk insulation o r
an a i r gap. Corrugated panels were constructed with and without bulk
insulation; a l l honeycomb panels were made with only an a i r gap. Because
of t h e very l i g h t weight of these panels, t h e p o s s i b i l i t y of f l u t t e r and
f a i l u r e existed; consequently, t h e i r s t r u c t u r a l i n t e g r i t y i n a supersonic
airstream, as well as t h e i r insulating q u a l i t i e s , was determined.

Corrugated panels.- The corrugated panels were designed and fabricated

by t h e B e l l A i r c r a f t Corporation. Both the surface and corrugations were
 3

of 0.005 Inconel X. A one-quarter-inch s t e e l p l a t e c o n s t i t u t e d t h e base

o r load-carrying s t r u c t u r e . Retaining s t r a p s held t h e panel against t h e
supports, and provision w a s made f o r expansion i n both directions. Thermo-
couples on t h e under s i d e of t h e outer face and a t s e v e r a l locations in
t h e i n t e r i o r measured t h e temperature during a t e s t .

I n order t o determine t h e insulating capacity of this type of con-

struction, t h e outer surface was heated, i n a s t a t i c t e s t , a t approxi-
mately 1/2 Btu/sq ft/sec u n t i l it reached l , 5 0 O o F, a f t e r which t h e skin
was maintained a t t h a t temperature f o r about 50 seconds. Temperature
h i s t o r i e s are shown i n f i g u r e 3 . The temperature of t h e exposed surface
rose a t a constant r a t e f o r about 70 seconds and then w a s held steady
u n t i l 2 minutes had elapsed f r o m t h e start of t h e t e s t . The temperature
of t h e inner surface of t h e corrugation lagged behind that of t h e exposed
surface and f i n a l l y reached an equilibrium temperature about 150° F lower
than t h a t of t h e outer face. The combination of t h e r a d i a t i o n shield and
b u l k i n s u l a t i o n protected the load-carrying s t r u c t u r e so t h a t it experi-
enced only a negligible temperature r i s e .

Additional t e s t s were made i n a blowdown j e t a t a Mach number of 1.4,

sea-level s t a t i c pressure, and a stagnation temperature of 6000 F. In
some of t h e t e s t s , added heat w a s supplied by a quartz-lamp r a d i a t o r
which faced the panel from j u s t outside t h e j e t stream and r a i s e d t h e
surface temperature t o nearly l,OOOo F i n some cases. Temperature his-
t o r i e s f o r a t e s t on a corrugated panel without b u l k i n s u l a t i o n a r e shown
i n figure 4. Surface temperatures of about 800° F were reached. The
temperature of t h e inner p a r t of t h e corrugation lagged s e v e r a l hundred
degrees behind t h a t of t h e exposed surface and t h e load-carrying s t r u c -
t u r e experienced only a s m a l l temperature r i s e . The data terminate after
10 seconds because of the f l u t t e r and f a i l u r e of t h e corrugated cover.
Aerodynamic heating was augmented by t h e radiant heater and t h e combined
e f f e c t of t h e heating and t h e a i r loads caused f l u t t e r t o begin about
10 seconds a f t e r t h e t e s t began. The panel w a s mounted as an extension
of t h e tunnel w a l l and w a s divided i n t o three bays by expansion j o i n t s .
From motion pictures, f l u t t e r w a s f i r s t observed i n t h e downstream bay.
It spread quickly t o t h e other bays and the panel was destroyed after
about 1 2 seconds. High-speed p i c t u r e s taken a t 640 frames per second
show t h e very high frequency of t h e f l u t t e r and indicate that t h e mode
shape was much l i k e t h a t of corrugated m e t a l . The corrugations were i n
t h e same d i r e c t i o n as those of t h e supporting corrugations of t h e panel
but appeared t o have a longer wave length by a f a c t o r of perhaps two o r
three. Not a l l t h e t e s t s of corrugated panels ended i n t h i s manner. A
similar configuration, but with an improved design o f . t h e edges, survived
t h e same t e s t . However, t h e r e s u l t emphasizes t h e need f o r c a r e f u l
experimental checking of such designs, a t l e a s t u n t i l more adequate theo-
r e t i c a l methods a r e available.

3'78
4

Honeycomb panels.- The honeycomb panels were fabricated of 0.005-inch-

t h i c k facing sheets brazed t o both sides of a honeycomb core. All sandwich
material w a s 17-7PH s t a i n l e s s s t e e l . The base or load-carrying s t r u c t u r e
w a s 1/8-inch-thick aluminum p l a t e . The same type of s t a t i c heating t e s t
w a s made on t h e honeycomb panels as on the corrugated panels.. Tempera-
t u r e h i s t o r i e s are shown i n f i g u r e 5 f o r the outer and inner faces of t h e
sandwich and f o r t h e base s t r u c t u r e . The heating cycle is evident from
t h e surface temperature and w a s very nearly t h e same as t h a t f o r the cor-
rugated panels. The inner-face temperature lagged behind and approached
1,250' F toward t h e end of t h e t e s t . The aluminum base w a s heated t o
about 300' F, which i s a higher temperature than t h a t of t h e base p l a t e
behind the corrugated panel with bulk insulation, even when allowance
i s made f o r t h e difference i n heat capacities of t h e two base s t r u c t u r e s .
I n t h e wind-tunnel t e s t s , however, t h e bending s t i f f n e s s of t h e honey-
comb panels was s u f f i c i e n t l y high t o prevent f l u t t e r , and t h e only unde-
s i r a b l e e f f e c t of t h e t e s t w a s t h e pock-marked appearance of t h e outer
surface. Thermal expansion of t h e heated skin caused t h e facing sheet
t o buckle over each c e l l of t h e honeycomb i n t e r i o r . This buckling reduced
t h e smoothness of t h e aerodynamic surface but had t h e advantage of reducing
t h e o v e r a l l expansion of t h e panel.

DISCUSSI O N

L i t t l e difference i n insulating value w a s found between t h e corrugated

and honeycomb panels, s o long as bulk insulation was not used between t h e
sandwich and t h e base structure. The corrugated panels with bulk insula-
t i o n were considerably b e t t e r from an insulation standpoint than e i t h e r
type without it but, when bulk insulation w a s used, the protective covering
w a s t h r e e times as thick as when it w a s omitted. The corrugated panels
have c e r t a i n advantages because they employ only one face; they can be
fabricated by seam welding instead of brazing, which makes them e a s i e r
t o manufacture and, also, they can e a s i l y be bent around curved surfaces
s o long as t h e curvature i s i n only one direction. Similarly, t h e honey-
comb panels have advantages because of t h e i r g r e a t e r strength; they are
considerably l e s s susceptible t o f l u t t e r and should require l e s s sup-
porting structure.

CONCLUDING REMARKS

These t e s t s represent only a s m a l l e f f o r t i n t h e f i e l d of insulated

structures; many additional variations can be studied. Results so far
indicate t h a t e f f e c t i v e short-time insulation can be achieved with f a i r l y
lightweight structures; however, it i s emphasized t h a t care should be
given t o design d e t a i l s i n order t o insure s t r u c t u r a l i n t e g r i t y .
 *
5

ETHYLENE HOT -EXHAUST JET TEST

-
INCONEL PROTECTED MAGNESIUM F I N

TEMP.,
2*ooor
.031 INCONEL
OF

1,000,

0 2 4 6
TIME, SEC

Figure 1

I NSU L AT I NG PAN E LS
CORRUGATED PANEL
RETAINER

CORRI
BULK I NSUL AT ION
BASE
HONEYCOMB PANEL

OUTER FACE
INNER FACE
BASE PLATE

SUPPORT PIN
4
 LABORATORY RADIANT - H E A T I N G T E S T
CORRUGATED P A N E L

1,500

1,000
TEMP.,
OF

500

0 60 I20
TIME,SEC

Figure 3

BLOWDOWN-JET TEST
CORRUGATED PANEL

800 -

TEMP,
OF

400 -

I
0 5 IO
T I M E ,SEC

Figure 4
381
 L A B O R A T O R Y R A D I A N T - H E A T I N G TEST
HONEYCOMB P A N E L
1,500

1,000

TE M P.,
OF

500

0 60 I20
T I M E , SEC

Figure 5

382
 , ., -
F

SOME RESEARCH RESULTS ON SANDWICH STRUCTURES

By Melvin S. Anderson and Richard G. Updegraff

Langley Aeronautical Laboratory

SUMMARY

The r e s u l t s of compressive-buckling t e s t s of s t e e l sandwich p l a t e s

are given, and the s i g n i f i c a n t parameters which a f f e c t the strength of
t h e p l a t e s a r e discussed. The various types of sandwich construction
a r e shown t o be comparable on a weight-strength basis with conventional
high-s trength aluminum- alloy construct ion.

INTRODUCTION

The use of high-density, heat-resistant materials in modern a i r c r a f t

has served t o reemphasize the need f o r lightweight methods of construction.
One such method receiving wide attention i s sandwich construction which
permits almost f u l l u t i l i z a t i o n of the strength of t h i n gages of materials.
Success of t h i s approach i s t o a large extent dependent upon advances i n
production techniques and p r a c t i c a l engineering experience. For these
reasons a number of sandwich Configurations have evolved, some of which
a r e shown in figure 1. The honeycomb sandwich has been produced by adhe-
s i v e bonding, brazing, and welding techniques. In high-temperature m t e -
rials there i s considerable i n t e r e s t i n t h e welding approach and a t t e n t i o n
has, therefore, been directed toward other configurations which a r e more
amenable t o welding. Representative of these configurations are t h e t r u s s -
core, tube-core, and web-core sandwiches a l s o shown in f i g u r e 1. These
configurations have t h e common c h a r a c t e r i s t i c that t h e core elements can
be joined t o t h e faces by welding along p a r a l l e l longitudinal l i n e s .
Certain obvious differences e x i s t between these configurations and t h e
honeycomb; f o r example, they are more d i r e c t i o n a l i n t h e i r s t i f k e s s and
strength properties. They require a heavier core t o achieve comparable
panel thicknesses, but the core c a r r i e s d i r e c t loads and provides a high
shear strength. The core a l s o provides n a t u r a l passages f o r the circula-
t i o n of coolants i n applications where t h i s may be desirable. The behavior
of the truss-core sandwich is t y p i c a l of t h i s group and i s considered along
w i t h the honeycomb sandwich in t h i s paper.
2

b p l a t e width

D p l a t e f l e x u a l s t i f f n e s s per u n i t width

DQ p l a t e shear s t i f f n e s s per u n i t width

E Young's modulus of e l a s t i c i t y

h o v e r a l l height of sandwich

k plate-buckling c o e f f i c i e n t

2 panel length

pi compressive load per u n i t width

t p l a t e- element thickness

?-I p l a s t i c i t y reduction f a c t o r

e angle between face sheet and core element

Poisson's r a t i o

stress

Qcr buckling s t r e s s

Subscripts :

C core

f face sheet

1 upper face

2 lower face
 3

EXPERIMENT

Honeycomb construction e f f e c t i v e l y increases t h e thickness of t h e

sandwich p l a t e which r e s u l t s i n a high o v e r a l l bending s t i f f n e s s as com-
pared with an equivalent-weight p l a t e of s o l i d material. A t t h e same
time, however, t h e l i g h t core tends t o make a sandwich s e n s i t i v e t o con-
centrated loads and causes shearing deformations t o play an important
r o l e i n determining t h e stress t h a t a sandwich p l a t e can carry. Because
of t h i s e f f e c t , sandwich development has required considerable experi-
mental work using s p e c i a l t e s t techniques and fixtures.

Test Technique

A f i x t u r e t h a t w a s found t o be s u i t a b l e f o r a simple t e s t of t h e
strength of a sandwich panel i n compression i s i l l u s t r a t e d - i n f i g u r e 2.
The panel i s loaded on i t s ends by a t e s t i n g machine, and t h e panel edges
a r e alined by a f i x t u r e designed t o give simple support. A cross-sectional
view of t h e panel and f i x t u r e i s shown i n f i g u r e 3. The I-beam and knife
edges prevent l a t e r a l d e f l e c t i o n of t h e panel edges but permit rotation.
They can be adjusted t o accommodate panels of d i f f e r e n t thickness and
width. Clearance between t h e f i x t u r e and t h e t e s t i n g machine permits
shortening of t h e panel without loading t h e f i x t u r e .

The sandwich p l a t e shown i n f i g u r e 2 has a honeycomb core. For t h i s

type of sandwich it w a s found necessary t o reinforce t h e panel a t t h e
loaded ends t o prevent end f a i l u r e s . The l i g h t areas of t h e t e s t speci-
men are externally applied doubler p l a t e s which a r e adhesive-bonded t o
t h e panel. It i s considered s i g n i f i c a n t t h a t in t e s t s of truss-core
sandwich no s p e c i a l end reinforcement was required. This i s a t t r i b u t e d
t o t h e higher shear strength of t h e core of t h i s construction.

T e s t Results

Panels.- Test results f o r some honeycomb panels a r e compased i n

f i g u r m t h t h e buckling theory of reference 1. These panels varied
i n thickness from 1/4 inch t o 3/4 inch with face-sheet thicknesses of
0.017 inch t o 0.064 inch. The compressive buckling-stress c o e f f i c i e n t
i s p l o t t e d as a function of a s h e a r - f l e x i b i l i t y parameter. Theory pre-
d i c t s a l a r g e l o s s i n panel buckling; strength as core shear f l e x i b i l i t y
increases. If t h e shear s t i f f n e s s DQ of the panel i s large, t h e
buckling-stress coefficient approaches t h e value of 4 which i s associated
wfth s o l i d p l a t e s . Theory i s capable of predicting t h e influence of t h e
geometrical q u a n t i t i e s which make up t h e s h e a r - f l e x i b i l i t y parameter as
i s i l l u s t r a t e d by t h e open t e s t points f o r both brazed and adhesive-bonded
s t e e l honevcomb -Danels. The core density associated with these points w a s
4

f$lb/cu f t (l/k-inch c e l l with 0.002-inch f o i l ) and I 2 lb/cu f t (l/k-inch

c e l l with 0.003-inch f o i l ) . However, when panels having a core density
of 624 lb/cu f t (l/k-inch c e l l with 0.0015-inch f o i l ) were tested, pre-
dicted strengths were not consistently obtained as indicated by the dark-
ened t e s t points. These l o w points a r e believed t o be caused by the low
shear strength associated with t h e l i g h t e s t core. The influence of core
shear strength i s not included i n buckling theory. The l o s s i n buckling
strength f o r these panels i s much greater than the reduction i n weight
over a panel which had a heavier core but sustained t h e predicted load.
Hence, on a weight-efficiency basis as well as from the standpoint of
obtaining consistent and r e l i a b l e r e s u l t s , these t e s t s indicate t h a t
adhesive-bonded cores should have a density greater than 6-3 lb/cu f t .
4
In t h e sandwiches with t h e heavier cores, the s t r e s s e s i n t h e face sheets
varied up t o 200,000 p s i , t h e compressive y i e l d s t r e s s of t h e material
tested.

Beams.- Since the matter of core density i s an important f a c t o r in

determining t h e weight of honeycomb panels, f u r t h e r t e s t s have been made
i n which sandwich panels were used as t h e compression covers of box beams
i n bending. Figure 5 i s a photograph of one of the beams a f t e r a com-
pression f a i l u r e of the sandwich cover. Of t h e three beams t e s t e d t o date,
only the one having a sandwich with a core density of 1 2 lb/cu f t approached
t h e load predicted by theory (14 percent l e s s than theory). The remaining
t w o beams had core d e n s i t i e s of & and d lb/cu f t and f a i l e d a t loads
2 4
considerably l e s s than the predicted ones. These r e s u l t s suggest that,
t o achieve adequate core shear strength, heavier cores may be required
f o r a p r a c t i c a l s t r u c t u r e than f o r simple compression t e s t s under more
i d e a l conditions.

LOCAL BUCKLING OF TRUSS-CORE SANDWICHES

Plate-buckling theory has been applied t o truss-core sandwich p l a t e s

and t h e available t e s t r e s u l t s indicate t h a t t h e theory i s again adequate.
With the truss-core sandwich no problems wirth core shear s t i f f n e s s or
strength were encountered because of t h e higher core d e n s i t i e s required
t o prevent l o c a l buckling of t h e individual p l a t e elements of t h e sand-
wich configuration. The l o c a l buckling s t r e s s can be calculated with t h e
a i d of a diagram such as t h a t shown i n figure 6. By knowing the l o c a l
buckling s t r e s s as well as the o v e r a l l p l a t e i n s t a b i l i t y s t r e s s as a
function of sandwich dimensions the proportions can be varied t o obtain
t h e most e f f i c i e n t sandwich f o r any given loading condition. In figure 6
t h e buckling c o e f f i c i e n t f o r l o c a l i n s t a b i l i t y of t h e sandwich configura-
t i o n i s p l o t t e d against the r a t i o of core-elemen$ thickness t o face-sheet
thickness over a range of values found t o give e f f i c i e n t proportions.
It should be noted t h a t t h e core elements are of t h e same order of thick-
ness as t h e face sheets i n contrast t o t h e f o i l in honeycomb cores which
may be only a f r a c t i o n of t h e thickness of t h e face sheets. Inasmuch as
t h e sandwich i s orthotropic, t h e buckling-stress c o e f f i c i e n t i s given f o r
compression i n both t h e longitudinal d i r e c t i o n indicated by t h e upper
curves and t h e transverse d i r e c t i o n indicated by t h e lower curves. A
lower strength i s obtained f o r t h e transverse loading inasmuch as t h e
face p l a t e s are subject t o column f a i l u r e between truss-panel p o i n t s -
I n both cases, the buckling-stress c o e f f i c i e n t i s r a i s e d because of i n t e r -
ference r e s t r a i n t s caused by t h e t r i a n g u l a r arrangement of t h e members.
The values indicated i n f i g u r e 6 have been substantiated by crippling
t e s t s on s m a l l specimens such as t h a t shown i n f i g u r e 7. This p a r t i c u l a r
specimen i s of welded construction, 17-7 PH s t a i n l e s s s t e e l , and sustained
a s t r e s s of 185,000 p s i a t f a i l u r e .

By using f i g u r e 6, p l a t e proportions can be adjusted s o t h a t l o c a l

buckling of t h e sandwich elements i s equal t o o r g r e a t e r than t h e o v e r a l l
plate-buckling s t r e s s . For example, i f -it i s desired t o achieve a
longitudinal-compression s t r e s s of 180,000 p s i ( a t y p i c a l value f o r t h e
y i e l d stress of high-strength s t e e l ) , t h e proportions given on t h e l e f t -
hand portion of f i g u r e 6 meet t h e requirements; a l s o given i s t h e c r i t i c a l
compressive s t r e s s i n t h e transverse d i r e c t i o n which i s almost two-thirds
t h e value f o r t h e longitudinal d i r e c t i o n . The o v e r a l l height of t h e sand-
wich, indicated by t h e r a t i o h/tf = 15, i s such that a very favorable
weight efficiency i s obtained f o r panels of t h i s sandwich loaded in edge
compression.

EFFICIENCY OF SANDWICH CONSTRUCTION

The weight of unstiffened plates, sandwiches, and s t i f f e n e d panels

subjected t o longitudinal-compressive s t r e s s i s p l o t t e d as a function of
t h e appropriate s t r u c t u r a l index i n f i g u r e 8. The weight of these struc-
t u r e s can be compared a t i d e n t i c a l values of t h e s t r u c t u r a l index inasmuch
as t h e p l a t e width b and panel length 2 a r e simply t h e support spacing
i n an a c t u a l structure. For t h i s p a r t i c u l a r p l o t , t h e u n i t s a r e such t h a t ,
i f t h e ordinate i s multiplied by t h e support spacing, t h e weight i s given
d i r e c t l y i n pounds per square foot of surface. For example, at an ordi-
nate value o f 0.2, a p l a t e 10 inches wide would weigh 0.2 times 10 (or
2 lb/sq f t ) .

The weight e f f i c i e n c y of t h e honeycomb-sandwich construction taken

from reference 2 has been calculated by assuming t h a t a core density of
10 lb/cu f t was necessary t o obtain t h e s t r e s s e s indicated by theory.
6

The most e f f i c i e n t proportions f o r t h e truss-core sandwich involve core

d e n s i t i e s varying from 30 t o 50 lb/cu f t . Despite the g r e a t e r core density
f o r t h e truss-core sandwich, there i s l i t t l e difference i n weight between
t h e t w o types of sandwich construction; w i t h t h e honeycomb sandwich more
e f f i c i e n t a t low values of the s t r u c t u r a l index and because of t h e load-
carrying core, t h e truss-core sandwich i s more e f f i c i e n t a t the higher
values&-% %.l.so shown in figure 8 a r e the weight-efficiency curves f o r
high-gtrength aluminum-alloy p l a t e s that would occur in multispar con-
s t r u c t i o n ' a n d f o r conventional stiffened-panel construction. It appears
t h a t under compressive loadings sandwich construction in s t e e l i s com-
parable i n weight t o e f f i c i e n t conventional construction i n aluminum
alloy.

THE3MAL STRESSES IN SANDWICH PLATES

Aconsideration of the response of sandwich p l a t e s t o t r a n s i e n t

heat- indicates t h a t c e r t a i n adjustments t o sandwich proportions may
be desirable t o minimize t h e e f f e c t of thermal s t r e s s e s . For example,
i n f i g u r e 9 are shown the r e s u l t s of thermal-stress calculations f o r a
t y p i c a l sandwich which is heated on one face t o &IOo F a t a r a t e of 80 F
per second. Heat i s transferred t o the other face by conduction and
radiation. The sandwich is assumed t o be constrained t o remain f l a t ,
and t h e r e s u l t i n g maximum thermal s t r e s s i n t h e t w o faces i s p l o t t e d
against the thickness r a t i o of the two faces. T h i s r a t i o was varied
while holding t h e t o t a l weight of t h e sandwich constant.

For equal-thickness faces, t h e tension s t r e s s in the cooler face i s

equal t o the compression s t r e s s i n the heated face: As t h e thickness of
t h e cooler face i s decreased, r e l a t i v e l y l i t t l e change occurs in t h e ten-
sion s t r e s s while the compression s t r e s s i n t h e heated face i s reduced
considerably. This favorable a l t e r a t i o n in t h e thermal s t r e s s e s i s due
p a r t i a l l y t o a reduction i n t h e maximum temperature difference between
t h e t w o faces and p a r t i a l l y t o t h e change i n t h e r e l a t i v e areas of t h e
faces. The decrease i n compressive thermal s t r e s s permits an increase
i n load s t r e s s t o be cazried before buckling of t h e sandwich. I n addi-
tion, the more even d i s t r i b u t i o n of temperature through t h e thickness
of t h e sandwich permits t h e absorption of a greater quantity of heat
before t h e h o t t e r face exceeds i t s allowable temperature.

CONCLUDING REMARKS

A few of the f a c t o r s which a f f e c t the design of any p a r t i c u l a r sand-

wich configuration have been presented. For honeycomb construction, t h e
core should be of s u f f i c i e n t s t i f f n e s s t o give a u g h buckling coefficient
 7

in order t o obtain a minimum-weight sandwich. In addition, t h e core

should be of adequate strength t o prevent premature core failures and
insure r e l i a b l e results. For sandwiches such as t h e truss core, tube
core, o r web core, shear strength or stiffness i s generally no problem;
but t h e main consideration i s proportioning t h e sandwich s o that o v e r a l l
p l a t e i n s t a b i l i t y i s not preceded by l o c a l buckling or crippling. Sand-
wiches of t h i s type can be proportioned so that they compare favorably
with honeycomb construction on a weight-strength basis. For elevated-
temperature applications, a sandwich with FL t h i c k e r outer face appears
b e t t e r able t o cope with t h e e f f e c t s of heating and r e s t r a i n e d expansion.
.. .....
.._

REFERENCES

1. Seide, Paul, and Stowell, Elbridge Z.: E l a s t i c and P l a s t i c Buckling

of Simply Supported Solid-Core Sandwich P l a t e s in Compression.
NACA Rep. 967, 1930. (Supersedes NACA TN 1822. )

2. Johnson, Aldie E., Jr., and Semonian, Joseph W.: A Study of t h e

Efficiency of High-Strength, Steel, Cellular-Core Sandwich P l a t e s
i n Compression. NACA TI? 3751, 1956.
 SANDWICH CONFIGURATIONS

HONEYCOMB CORE TRUSS CORE

TUBE CORE WEB CORE

Figure 1

TEST SETUP
IF

GROSS SECTION OF FIXTURE AND SPECIMEN

HONEYCOMB

Figure 3

BUCKLING COEFFICIENTS FOR SANDWICH P L A T E S

k r27)D
Pi = COR E
b2 DENSITY
LBKUFT
675 8.50,
0 '6 BRAZED
m n ADHESIVE
BONDED

't
I
0
I
.2
I I
4
I I
.6
I I
.8
I

SHEAR FLEXIBILITY PARAMETER,

82D
-
b2 DQ

Figure 4 391
 SANDWICH BOX BEAM AFTER FAILURE

Figure 5 L-57-650.1

BUCKLING OF TRUSS-CORE SANDWICH ELEMENTS

2~
I

Figure 6
 CRIPPLING SPECIMEN

Figure 7 L-57-95 1

WEIGHT-STRENGTH CURVES

.7-UNSTIFFENED PLATE ,7075-T6

.6 -

CONVENTIONAL
STIFFENED PANEL,
7075 -T6
TRUSS-CORE SANDWICH, 17-7PH

I I I I I I I I
0 I 2 3 4
Pi Pi
STRUCTURAL INDEX,-
b OR ~ KSI

Figure 8 393
EFFECT OF SANDWICH DIMENSIONS ON THERMAL STRESSES

50r
c-2,TENS ION

THERMAL
STRESS,
KS I
40 -

30 -

20 -
2
/ Lt1
7-
=k
10-
/
P- CT I ,COMPRESS ION Lt2

I I I I I I I

Figure 9
 RECENT RESEARCH ON THE CREEF’ OF AIRFRAME’ COMPONENTS
By Eldon E. Mathauser, Avraham Berkovits,
and Bland A. Stein

Langley Aeronautical Laboratory

The results of recent research of the National Advisory Committee

for Aeronautics on the creep of airframe components at elevated tempera-
tures are summarized. Experimental lifetime data from creep tests of
stainless-steel plates and aluminum-alloy unstiffened circular cylinders
are presented and compared with results predicted from isochronous stress-
strain curves. The results of a study to determine the magnitude of
creep strains that produce significant structural deformations are
included. A comparison of structural weight determined from assumed
strength and creep criteria is made to establish temperature ranges in
which creep is expected to influence structural design for various
materials.

INTRODUCTION

Many studies have been made at the National Advisory Committee for
Aeronautics during the past few years to obtain basic knowledge of the
creep behavior of structural elements at elevated temperatures. These
studies have ranged from analytical and experimental investigations of
simple structural elements such as columns and plates (refs. 1 to 3) to
the development of a variational theorem (ref. 4) suitable for applica-
tion in many structural creep problems. Studies have a l s o been made to
establish approximate methods (for example, ref. 3) for predicting creep
collapse of structural components. This paper presents comparisons
between experimental and predicted lifetime results for stainless-steel
plates and for aluminum-alloy unstiffened circular cylinders; The results
of an analysis to determine the magnitude of creep strains that produce
significant structural deformations are given. Temperature ranges in
which creep is expected to influence aircraft structural design are
indicated for various materials.
 T C

? e
-
-
2

b width, in.

ES secant modulus, ksi

Et tangent modulus, ksi

g acceleration due to gravity, ft/sec 2

k constant
r radius, in.

t thickness, in.

%r critical (buckling) stress, ksi

uCY 0.2-percent-offset compressive yield stress, ksi

-Of average failure stress, ksi

DE'l!T2RMINATION OF CREEP C0I;LPSSE

The approximate methods that have been investigated for predicting

creep collapse of structural elements are based on the use of isochronous
stress-strain curves in conjunction with methods established for pre-
dicting maximum strength. An example of isochronous stress-strain curves
is given in figure 1. The dashed line indicates the material compressive
stress-strain curve for 17-7 PH stainless-steel sheet (condition TH 1,050)
at 800' F. The solid lines, designated as isochronous stress-strain
curves, indicate the strain produced on application of a given stress
plus the creep strain obtained at that stress for the various times.
Curves such as these can be obtained by cross plotting compressive creep
curves to give stress as a function of strain for different times. The
isochronous stress-strain curves shown in figure 1 were obtained from
compressive creep tests of the stafnless-steel sheet at 800° F. The
tick marks indicate the 0.2-percent-offset compressive yield stresses.
,

Plates

The use of isochronous stress-strain curves for the prediction of

creep lifetime of plates will be considered first. A comparison between
predicted and experimental lifetimes is shown in figure 2. Applied stress .b
is plotted against lifetime defined as collapse time for 17-7 PH
stainless-steel plates (condition TK 1,050) at 800' F. m e symbols
indicate compressive-creep-test results from V-groove edge-supported
plates for width-thickness ratios ranging from 15 to 60. The solid
lines indicate plate life predicted fromthe following equation:

-af = 1.60&acy t

where 5f is the average applied stress to produce creep collapse of

the plate, E, is the secant modulus associated with acy is the
compressive yield stress, and t/b is the plate thickness-width ratio.
Equation (1)gives maximum strength of V-groove edge-supported plates
at elevated temperatures (ref. 3), if the material parameter /=
is evaluated from the material compressive stress-strain curve. Evalua-
tion of the material parameter from isochronous compressive stress-strain
curves, in general, gives a satisfactory approximation for plate lifetime.
This equation has been used to predict creep lifetime for both 2024-T3
and 7075-6 aluminum-alloy plates. Similar agreement between experimental
and predicted results was obtained.

Unstiffened Circular Cylinders

Prediction of creep lifetime of unstiffened circular cylinders using

isochronous stress-strain curves will now be considered. In figure 3,
bending moment is plotted against lifetime defined as collapse time for
5052-0 aluminum-alloy cylinders at 500° F. The symbols indicate experi-
mental results obtained from reference 5 for radius-thickness ratios
ranging from 125 to 250. Predicted lifetimes indicated by the curves
were obtained by substituting materials data from isochronous stress-
strain curves into the following relation:

where Ocr is the critical or buckling stress, k is a constant assumed

to be 0.6, Es and Et are secant and tangent moduli, respectively, and
t/r is the cylinder thickness-radius ratio. !The predicted buckling
stresses were then converted to bending moments by using elementary beam
theory. !€!he isochronous stress-strain curves required for this study
were obtained from compressive creep tests of 5052-0 aluminum-alloy sheet
at 500° F. The experimental results shown in figure 3 are the only data
available on the lifetime of cylinders subjected to bending. Although

397
the predictions are i n good agreement w i t h the experimental data, addi-
t i o n a l studies w i l l be needed t o determine whether equation (2) w i l l
predict lifetime s a t i s f a c t o r i l y f o r cylinders of other materials and t o
establish the appropriate value of k.

CREXP DEFLECTIONS

The results of the s t u d i e s of creep of p l a t e s and cylinders and of

other s t r u c t u r a l components investigated previously indicate t h a t l i f e -
time defined by collapse can be estimated i n general by s u b s t i t u t i n g
material data obtained from isochronous compressive s t r e s s - s t r a i n curves
i n t o appropriate r e l a t i o n s that define maximum strength of the struc-
t u r a l elements. For some types of structures, it i s realized t h a t l a r g e
creep deformations can be obtained i n a f r a c t i o n of the a c t u a l collapse
t i m e . Such deformations i n many cases may determine the useful l i f e of
the s t r u c t u r e . A study w a s made accordingly t o determine the range of
values f o r creep s t r a i n t h a t would be expected t o govern design of
structures where deformation r a t h e r than collapse would be of primary
interest.

The s t r u c t u r e considered f o r this study w a s a constant-stress wing

i n which the s t r e s s e s are assumed t o be of the same magnitude a t a l l
s t a t i o n s along the wing. The deflections of a constant-stress wing are
determined from the product of wing configuration and the s t r a i n s pro-
duced by the applied stresses. The deflections t h a t would be produced
by creep over a complete range of stress f o r stainless-steel wings are
shown i n f i g u r e 4. I n t h i s figure, stress i s plotted against the r a t i o
of wing deflection produced by creep t o wing deflection produced by load
f o r 17-7 PH s t a i n l e s s s t e e l a t 800° F. Load deflection i s defined as
the s t a t i c deflection of the wing obtained immediately upon application
of any stress. A t a stress of 50 k s i , f o r example, the creep deflection
of t h e wing i n 1 hour i s equal t o 0.1 of the s t a t i c or load deflection
obtained immediately upon application of the stress. Ilf t h i s stress i s
applied f o r approximately 300 hours, the creep deflection increases t o a
value equal t o t h e s t a t i c deflection. A shaded area i s indicated f o r
the range of creep t o load deflection r a t i o s from 0 . 1 t o 1.0. This area
i s assumed t o define the region of i n t e r e s t f o r s t r u c t u r e s such as air-
c r a f t wings. Creep deflections t o the l e f t of the shaded area would be
p r a c t i c a l l y negligible; whereas, t o the right, t h e creep deflections
would undoubtedly be considered excessive f o r most s t r u c t u r a l applications.

Creep s t r a i n s that are associated w i t h t h e range of deflection r a t i o s

shown i n f i g u r e 4 are indicated i n figure 5. The s o l i d curves have been
reproduced from f i g u r e 4. The dashed curves indicate creep s t r a i n s pro-
duced i n the specified t i m e s f o r the range of stresses shown. Ncte t h a t
creep s t r a i n s of approximately 0.0002 t o 0.002 are associated w i t h r a t i o s
i
r
4F '5

of creep t o load deflections ranging from 0 . 1 t o 1.0. These results

apply t o bending deflections of any constant-stress wing regardless of
s t r u c t u r a l configuration because the r a t i o of creep deflection t o load
deflection i s determined from the r a t i o of creep s t r a i n t o load s t r a i n .

This method of analysis which is used t o determine the range of

creep s t r a i n s t h a t a r e of i n t e r e s t f o r structures subjected t o bending
was applied t o t w o other materials: 2024-T3 aluminum alloy a t 400° F
and Inconel X a t 1,350° F. The r e s u l t s f o r t h e 2024-T3 aluminum alloy
are shown i n f i g u r e 6, and t h e r e s u l t s f o r Inconel X are sham i n f i g -
ure 7. For 'both materials approximately the same range of values of
creep s t r a i n from 0.0002 t o 0.002 was obtained f o r r a t i o s of creep t o
load deflections from 0 . 1 t o 1.0.

'IIEMPERATURE RANGES FOR CREEP

Consideration w i l l now be given t o the determination of temperature

ranges i n which creep w i l l be expected t o influence s t r u c t u r a l design
f o r various materials. These temperature ranges a r e determined by com-
paring s t r u c t u r a l weight required f o r strength w i t h the weight required
f o r creep at d i f f e r e n t temperatures. The r e s u l t s obtained from t h i s
analysis f o r s t a i n l e s s s t e e l a r e presented i n f i g u r e 8. The required
weight of a t e n s i l e member i n a r b i t r a r y units i s plotted against tempera-
t u r e f o r 17-7PH s t a i n l e s s s t e e l . The s o l i d l i n e indicates the w e i g h t
required f o r strength based on ultimate load after 1,000 hours exposure
t o temperature. Ultimate load i s assumed t o be 3.75 times the 1 g load.
The dashed curves indicate weight required f o r t h e three d i f f e r e n t creep
c r i t e r i a f o r 1,000 hours a t 1 g load; namely, creep s t r a i n s of 0.0002
and 0.002 and creep rupture.

The dotted curve f o r 0.0002 creep s t r a i n i n f i g u r e 8 indicates that

the t e n s i l e member can be designed on the basis of strength up t o 6500 F.
Above 6500 F, the weight of the t e n s i l e member would increase very
rapidly i n order not t o exceed 0.0002 creep s t r a i n . The c r i t e r i o n of
0.002 creep s t r a i n would govern the design above 8250 F. Above 825O F,
s i g n i f i c a n t deflections would be expected t o occur i n s t r u c t u r e s sub-
jected t o bending as discussed previously. The region between t h e curve
f o r 0.002 creep s t r a i n and the curve f o r rupture defines t h e temperature
range where creep would be a very i m p o r t a t f a c t o r i n s t r u c t u r a l design
f o r t h i s material. The region between the two creep-strain l i n e s may be
considered t o be t h e temperature range i n which creep s t r a i n s become
perceptible and gradually increase t o a magnitude t h a t produces s i g n i f -
i c a n t s t r u c t u r a l deformations. It i s of i n t e r e s t t o note t h a t design
on the b a s i s of a given creep s t r a i n requires a very large increase i n
weight f o r small increases i n temperature. It appears that it w i l l be
more f e a s i b l e t o convert t o a higher strength material than t o add weight
6: C”

in order not to exceed a given value of creep strain. If other creep

and strength criteria are considered, the position of the creep lines
is shifted relative to the strength curve; however, the temperature
ranges defined by the distance between the various creep lines remain
essentially constant.

This method of analysis of weight required for strength and creep

to determine temperature ranges in which creep will be a design consid-
eration was applied to two other materials which are shown in figure 9.
The materials considered are 2024-T3 aluminum alloy, 17-7PH stainless
steel and Inconel X. The solid lines indicate the weight required for
strength based on ultimate load. The shaded regions define the tempera-
ture range for each material where creep may be an important factor in
structural design. The width of the shaded regions for each material
was established by determining required weight for creep on the basis of
several different creep criteria. These results indicate that creep
problems will be restricted to a rather narrow range of temperatures for
each material and that creep does not become a design consideration until
temperatures are reached where the strength of the material deteriorates
rapidly. Note that the weight required for creep increases very rapidly
with small increases in temperature for all materials considered. It
appears that, whenever a temperature is reached where creep is a design
consideration, less structural weight will be required by converting to
a higher strength material than by designing for creep with the original
material. It is realized that conversion to a higher strength material
introduces many new problems such as changes in production methods and
consideration of availability and strategic importance of the higher
strength material. Because of such factors, it is anticipated that the
weight of some structural components will be increased to account for
material creep rather than converting to a higher strength material.

CONCLUDING REMARKS

The results presented indicate that material creep w i l l influence

structural design over a rather narrow range of temperatures for each
striictural material. In this temperature range, lifetime or collapse
time can be estimated satisfactorily for structural elements by using
isochronous stress-strain curves in conjunction with established methods
for predicting maximum strength. If useful life is detedned by deflec-
tions rather than collapse, the simplified analysis of creep deflection
of constant-stress wings indicates that creep strains ranging from
approximately 0.0002 to 0.002 define the region of interest for struc-
tures subjected to bending. . L

p c:,
4L J
 7

REFERl3NCES

1. Libove, Charles: Creep-Buckling Analysis of Rectangular-Section

Columns. NACA TN 2956, 1953.

2. Mathauser, Eldon E., and Brooks, William A., Jr.: An Investigation

of the Creep Lifetime of 75s-T6 Aluminum-Alloy Columns. NACA
TN 3204, 19%.
3. Mathauser, Eldon E., and Deveikis, William D.: Investigation of the
Compressive Strength and Creep Lifetime of 2024-Tf,Aluminum-Alloy
Plates at Elevated Temperatures. NACA TN 3552, 1956. (Supersedes
NACA RM L55Ellb.)
4. Sanders, J. well, Jr., McComb, Harvey G., Jr., and Schlechte,
Floyd R.: A Variational Theorem f o r Creep with Applications to
Plates and Columns. NACA TN 4003, 1957. (Prospective NACA paper.)

5. Hoff, N. J., Erickson, Burton, et al.: Creep Bending and Buckling

of Thin Circular Cylindrical Shells. PIBAL Rep. No. 355 (Contract
Nos. Naw-6392 and PJaw-6447), Polytechnic Inst. Brooklyn, July 1956.
3 f

COMPRESSIVE STRESS-STRAIN CURVES

17-7 PH ST41NLESS STEEL; 800" F

/v----
0 TIME, HR
I20
I4O[ /
I
100 IO
100
1,000
80
STRESS,
KSI
60

---MATERIAL
40
-ISOCHRONOUS
20

0 .002 .004 .006 .008 .010

STRAIN

Figure 1

CREEP LIFETIME OF PLATES

17-7 PH STAINLESS STEEL; 800" F
120 -

100

80 0
STRESS, -0
KSI
-30

6ot .
4 A.,
40

I I I I I I l l 1 I I l l
.OL .&I I " ; 5 IO 50
LIFETIME, HR

4
 9

LIFETIME OF UNSTIFFENED CYLINDERS

5052-0 ALUMINUM ALLOY; 500" F
120- c c r = k G { -
PREDICTED
000.0 EXPERIMENT

r
100 -

80 -
BENDING
MOMENT,
IN-KIPS

20
O L
.I .5 I 5 IO
LIFETIME, HR

RATIO OF CREEP TO LOAD DEFLECTIONS FOR WINGS

17-7 PH STAINLESS STEEL;800" F
TIME, HR

''""I251
.^^

75.
STRESS,
KS I
50.

25

0
CREEP DEFLECTION
LOAD DEFLECTION

Figure 4 403
 CREEP STRAIN IN WINGS
17-7 PH STAINLESS STEEL; 800" F
TIME, HR
I25

Ioc

75
STRESS,
KS I
5c

25

0 I I I I I 1 1 1 1 I I l l
.05 .I .5 I 5
CREEP DEFLECTION
LOAD DEFLECTION

Figure 5

CREEP STRAIN IN WINGS

2024-T3 ALUMINUM ALLOY; 400"F
TIME, HR
CREEP STRAlN=O.OI .

404 Figure 6
 CREEP STRAIN IN WINGS
INCONEL X;l,35Oo F

70-
CREEP STRAIN = 0.01

50
STRESS,
KS I
40 -
\ '. 1,000

20 -
I II I1111 I I l l
OL .d5' ' I '!I .5 I 5
CREEP DEFLECTION
LOAD DEFLECTION

Figure 7

WEIGHT FOR CONSTANT TENSILE LOAD

17-7 PH STAINLESS STEEL

TENSILE STRENGTH, ULTIMATE LOAD,

1,000 HR EXPOSURE
1 I I I I 1
0 200 400 600 800 1,000
TEMP, OF

Figure 8 ' -405

 TEMPERATURE RANGES FOR CREEP

2.001

1.75,

1.50.
WEIGHT
REQUlRED
1.25.

1.00,

, ".,
.75- '-- -.- 'I' CREEP
I I I I
0 500 1.000 1,500
TEMP., O F

Figure 9
 A RELATION BETWEEN STRESS, STRAIN RATE, TIME, AND

TEMPEF3lTURE FOR METALS AT ELEVATED TEMPEBA!lYJFU3S

By Elbridge Z. Stowell and George J. Heimerl

Langley Aeronautical Laboratory

A r e l a t i o n between s t r e s s , s t r a i n r a t e , time, and temperature i s

suggested t o account f o r the behavior of polycrystalline metals a t e l e -
vated temperatures. The metal i s assumed t o have t h e basic properties
of e l a s t i c i t y , thermal expansion, and viscosity. An appropriate com-
bination of these three properties has been shown t o account f a i r l y well
f o r t h e steady creep r a t e , time t o creep rupture, s t r e s s - s t r a i n curves,
and rapid-heating curves f o r 7073-6 aluminum a l l o y . Similar calcula-
t i o n s made f o r low-carbon steels show good correlation between ultimate
t e n s i l e strength, time f o r creep rupture, and the steady creep r a t e s .
In t h e case of a complex a l l o y l i k e Inconel X, fair correlation i s a l s o
obtained provided one of t h e constants i n t h e r e l a t i o n i s permitted a
slow v a r i a t i o n with temperature.

INTRODUCTION

The need f o r knowledge of t h e behavior of metals a t elevated tem-

peratures i s constantly becoming more and more important as f l i g h t speeds
increase and as t h e e f f e c t s of aerodynamic heating become more pronounced.
Information i s needed concerning -
(1)The creep of metals a t these elevated temperatures

( 2 ) The time a t which rupture can be expected i f the creep i s allowed

t o continue

( 3 ) The s t r e s s - s t r a i n c h a r a c t e r i s t i c s a t some p a r t i c u l a r temperature

and s t r a i n r a t e

(4) The behaxior of t h e metal when heated rapidly up t o some high

temperature

Evidently t h e job of providing t h i s information would c a l l f o r a great

d e a l of t e s t i n g , and it would be a useful achievement i f some method of
2

calculation could be found which would simplify t h e t a s k . Such a method

follows from t h e r e l a t i o n between stress, s t r a i n r a t e , and temperature
proposed i n t h i s paper.

SYMBOLS

strain

s t r a i n r a t e , per hr

t time, hr

U stress, k s i

E Young ’s modulus, ks i

U l i n e a r expansion coefficient, per 4c

T temperature, ?K unless otherwise indicated

S constant, per hr per OK

AH a c t i v a t i o n energy f o r creep, c a l per mole

R gas constant, taken as 2 c a l per mole per ?K

00 constant, k s i

I- t i m e t o rupture, hr

kl’k2 constants

STATEMENT OF THE RELATION

The s t r a i n rate $$ at the absolute temperature T and t h e stress

Ub

u may be stated as follows:

a dT -AH/RT
,=dt(E)
+ a -
dt
+ 2sTe sinh

The f i r s t term t o t h e r i g h t of t h e equal sign gives t h e contribution t o

t h e s t r a i n r a t e due t o a change i n stress a; .E i s Young’s modulus,
 _.
3

which i s considered t o be known as a function of temperature. The mid-

d l e term gives t h e contribution due t o a possible thermal expansion o r
contraction; a, t h e l i n e a r expansion coefficient, i s assumed known as
a function of t h e temperature. The l a s t term on t h e r i g h t gives the
contribution due t o viscous flow. The exponential term, which describes
t h e usual e f f e c t s of temperature on viscous materials, contains AE, t h e
constant a c t i v a t i o n energy f o r creep of t h e material, and R, t h e gas
constant. The remaining constants i n t h e viscous term are s and bo.
The r e l a t i o n (eq. (1))holds only f o r temperatures high enough f o r t h i s
viscous term t o have an appreciable value, that i s , f o r temperatures
higher than about one-half t h e melting point.

Equation (1)m y be integrated i n various ways t o provide informa-

t i o n on such properties as steady creep, rupture t i m e , s t r e s s - s t r a i n
r e l a t i o n s , and t h e e f f e c t s of rapid heating. For example, a t constant
stress and temperature, - da = -dT = 0, and t h e r e l a t i o n (eq. (1))becomes
at at
simply t h e equation f o r steady creep. If t h e creep i s allowed t o con-
tinue without interruption, t h e material w i l l rupture i n a time T given
d€ = Constant. The s t r e s s - s t r a i n curve a t any con-
by t h e r e l a t i o n T -
at
s t a n t temperature T i s obtained by s e t t i n g t h e s t r a i n r a t e equal t o
the known constant t e s t i n g r a t e and then solving t h e r e s u l t i n g equation
for a (with as a function of t h e time. And f i n a l l y , i f t h e
material i s held under a constant stress and heated rapidly, t h e
r e s u l t i n g s t r a i n i s found by integrating t h e expression with
( at
with respect t o time.

The d e t a i l s of such applications t o one material, 7075-T6 aluminum

alloy, are given i n reference 1.

APPLICATIONS

Creep and Rupture Parameters

I n recent years, various empirical and semiempirical temperature

parameters have been proposed f o r the creep of metals. Two of these
parameters may be derived from t h e r e l a t i o n given by equation (1)when
t h e stress a i s large enough t o permit replacement of t h e hyperbolic
s i n e by t h e exponential. Under t h i s condition,

a = sol$ + log, i
) + log, kg
4

where kl i s a constant. Thus, i f t h e s t r e s s i s p l o t t e d against t h e

parameter @ + log, 6 , a single s t r a i g h t l i n e with a slope equal t o bo
RT
should be obtained.

I f , i n addition, use i s made of t h e r e l a t i o n between creep r a t e k

and rupture time T

6~ = Constant
as found by Monkman and Grant ( r e f . 2), then equation ( 2 ) may a l s o be
written as

e (3)

where k2 i s another constant. Thus, i f t h e s t r e s s i s p l o t t e d against

t h e parameter -AH
RT - loge T, a single s t r a i g h t l i n e with a slope equal
to o0 should be obtained.

I n figure 1, which presents a correlation of creep and rupture d a t a

f o r 7075-T6 clad aluminum alloy, t h e parameters
RT
+ log
e
6 and
--
AH loge T a r e p l o t t e d on t h e same abscissa scale against s t r e s s a s
RT
ordinate. The data points, taken from reference 3 , f a l l along two
s t r a i g h t l i n e s having t h e same slope. Since t h e temperature of 212' F
is known t o be t o o low f o r equation (1)t o hold, t h e data points
taken a t that temperature do not f a l l on t h e l i n e s but were included
f o r comparison.

Steady Creep Rate

Figure 2 shows t h e steady creep r a t e per hour as a function of

stress f o r two clad aluminum alloys: 7075-6 on t h e l e f t and 2024-T3
on t h e r i g h t . The solid-line curves represent a c t u a l data from r e f e r -
ence 3 and it is seen t h a t f o r t h e 7075-T6 a l l o y t h e solid-line curves
are nearly p a r a l l e l s t r a i g h t l i n e s . The steady-creep l a w given by equa-
t i o n (2) requires that such curves be p a r a l l e l s t r a i g h t l i n e s ; t h e
dashed-line curves were computed from equation ( 2 ) with t h e use of t h e
constants i n t a b l e I. The data f o r t h e 2024-T3 a l l o y do not form par-
a l l e l s t r a i g h t l i n e s and f o r t h i s material t h e r e l a t i o n does not apply.
 Stress-Strain Curves

Figure 3 shows t h e s t r e s s - s t r a i n curves f o r t h e 7075-6 aluminum

a l l o y a t three temperatures and f o r a s t r a i n r a t e of 0.002 per minute.
The solid-line curves represent data taken from reference 4. The dashed-
l i n e curves were computed by using the constants i n t a b l e I. The agree-
ment i s good a t t h e two higher temperatures and i s .poor a t t h e lowest
temperature, which i s below t h e temperatures a t which t h e theory can be
expected t o hold.

Rapid-Heating Curves

Figure 4 shows t h e s t r a i n s obtained when specimens of 7075-6 alu-

minum alloy, held under t h r e e constant s t r e s s e s as indicated, are heated
rapidly a t widely d i f f e r e n t temperature rates. The s l a n t portions of t h e
curves give mainly t h e e f f e c t of thermal expansion. The sharp upward
t u r n at t h e r i g h t of each curve gives t h e e f f e c t of viscosity. The s o l i d -
l i n e curves a r e data taken from reference 4, and t h e dashed-line curves
a r e computed by using t h e constants i n t a b l e I. Agreement i s good a t t h e
higher temperatures.

The constants l i s t e d i n t a b l e I f o r t h e d i f f e r e n t applications t o

7075-6 aluminum a l l o y are not q u i t e t h e same. For a discussion of these
small differences, see reference 1.

Low-Carbon S t e e l s

I n order t o show how t h e r e l a t i o n works f o r materials other than

aluminum alloys, an attempt w a s made t o check t h e available data on low-
carbon s t e e l s ( r e f . 5 ) . The s o l i d l i n e s of f i g u r e 5 show t h e v a r i a t i o n
with temperature of t h e ultimate t e n s i l e strength, t h e stress t o give
creep rupture i n 100 hours, and t h e stress t o give a creep rate of LOW6
per hour. The dashed-line curves give t h e corresponding q u a n t i t i e s com-
puted by using t h e constants i n t a b l e I. The agreement i s good, espe-
c i a l l y above about 8oOo F.

Inconel X

The s o l i d - l i n e curves of f i g u r e 6 represent t h e same q u a n t i t i e s as

i n f i g u r e 5 but are f o r t h e a l l o y Inconel X; t h e d a t a are taken from
reference 6. The corresponding dashed-line curves were computed by using
t h e constants i n t a b l e I; note that f o r t h i s complex a l l o y t h e values
of a0 vary slowly with t h e temperature. It is believed that t h i s var-
i a t i o n r e f l e c t s t h e microstructural changes which occur i n Inconel X a t
elevated temperatures. The agreement with t h e data i s f a i r and seems t o
6

indicate that t h e proposed r e l a t i o n w i l l be s a t i s f a c t o r y even f o r a

complex alloy, provided allowance i s made f o r some i n s t a b i l i t y with
temperature.

CONCLUDING REMARKS

A r e l a t i o n between s t r e s s , strain r a t e , time, and temperature has

been suggested f o r metals a t elevated temperatures. The r e l a t i o n has
been applied successfully t o steady creep, creep rupture, s t r e s s - s t r a i n
r e l a t i o n s , rapid heating, and t e n s i l e strength f o r a f e w materials. The
r e l a t i o n makes use of t h e e l a s t i c i t y , thermal expansion, and v i s c o s i t y
of t h e metal, and i s an equation of s t a t e i n r a t e form. The r e l a t i o n is
inapplicable t o metals which a r e markedly unstable o r which undergo aging
o r changes i n s t r u c t u r e . The metal must be such that i t s steady creep
i s describable by the r e l a t i o n i n order that t h e constants may be deter-
mined. With such a metal, however, t h e r e l a t i o n may be useful i n the
prediction of other properties from steady-creep data.

RFFERENCES

1. Stowell, Elbridge Z.: A Phenomenological Relation f o r Metals at

Elevated Temperatures. NACA TN 4000, 1957. (Prospective NACA paper.)

2. Monkman, Forest C., and Grant, Nicholas J.: An Empirical Relationship

Between Rupture L i f e and Minimum Creep Rate i n Creep-Rupture Tests.
Paper NO. 72, A.S.T.M. (Presented a t Fifty-Ninth Annual Meeting of
A.S.T.M., June 17-22, 1956.)

3. Anon.: Strength of Metal A i r c r a f t Elements. ANC-5 B u l l . , Rev. ed.,

Depts. of A i r Force, Navy, and Commerce, Mar. 1955.

4. Heimerl, George J., and Inge, John E.: Tensile Properties of 7075-T6
and 2024-T3 Aluminum-Alloy Sheet Heated at Uniform Temperature Rates
Under Constant Load. NACA TN 3462, 1955.

5. Simmons, Ward F., and Cross, Howard C.: Elevated-Temperature Properties

of Carbon S t e e l s . Special Tech. Pub. No. 180, A.S.T.M., 1955.

6. Anon. : Inconel "X" - A High Strength, High Temperature Alloy, Data

and Information. The International Nickel Co., Inc., Dev. and Res.
Div., Jan. 1949.
 7
c

Metal Application as, 6,

‘7
cal/mole ‘0, ai at ‘
9 9 per per OK
7075-T6 clad aluminum alloy Steady creep 35,800 4.3 All 1.62 X lo6 0.0077
7075-T6 aluminum alloy Stress-strainrelations 34,700 4.3 All 1.50
7075-T6 aluminum alloy Rapid heating 34,700 4.3 All 3.60
Low-carbon steels All 72,000 2.0 All 9-15X lo8 0.025
Inconel X All 4.9 1,ooO 3,000 lo8 0.0095

I
X
4.9 1,200
4.6 1,350
4.1 1,425
3.85 1,450
3.50 1,475
3.a 1,500
2.70 1,550 -
2.4 1,600
 CORRELATION OF C R E E P AND R U P T U R E DATA e.
FOR 7075-T6 CLAD ALUMINUM ALLOY

0 0
60- 0

50 -
40- -RUPTURE
STRESS,
KS' 30-

20 - A 375OF
0 300°F
IO - 0 212OF

I I I I I
0 25 30 35 45 50
40
4
RT
+ log, k ffy
OR -- loge r

Figure 1

STEADY CREEP O F 7075-T6 AND

2024-T3 C L A D ALUMINUM ALLOYS

50 -7 0 7 5 - T 6 /

40 -

30 -
STRESS,
KSI
20 -
EXPERIMENT

10 - -----C A L C U LATE D
 STRESS - S T R A I N C U R V E S FOR
7 0 7 5 - T 6 ALUMINUM ALLOY

EXPERIMENT
---- C ALC U LAT ED

,/--

,0' 2OOOF

60
STRESS,
KSI
--- 4 O O O F

0 .002 .004 .006 .008 .OlO

STRAIN

RAPID - H E A T I N G CURVES FOR 7 0 7 5 - T 6

ALUMINUM ALLOY AT T H R E E STRESSES

TEMP. RATE ,OF/SEC

.020 - 2 96

STRAIN

I I I I
0 200 400 600 800
TEMPERATURE, OF

Figure 4
415
 STRENGTH OF LOW-CARBON STEELS

-E X P E R I M E N T A L
CALCULATED

40 -
STRES
KS I
30 -

CREEP RATE

I
0 500 1,000 1,500 2,000
TEMPERATURE, OF

Figure 5

STRENGTH O F I N C O N E L X
I40 F

I20

100

80 -RUPTURE
STRESS,
KS' 60 -
40 -
20
I I
0 1,000 1,200 1,400 1,600 1,800
TEMPERATURE, OF

Figure 6
 REACTIONS OF MATERIALS I N HIGH-TEMPERATURE AIR FLOWS

By Joseph G. Thibodaux, Jr., and Joseph N. Kotanchik

Langley Aeronautical Laboratory

SUMMARY

Ceramic-heated air jets and high-intensity arc-powered jets which

are capable of duplicating some of the environmental conditions of hyper-
sonic flight up to reentry velocities have been developed and are being
operated. Initial tests of structural materials in these facilities at
stagnation temperatures of 4,000° F and above indicate that these mate-
rials will be ignited, melted, or be severely oxidized. Surface heat
inputs due to oxidation can be great enough to cause melting in high-
velocity airstreams with stagnation temperatures several hundred degrees
below the melting point of the materials. Laboratory experiments indicate
that recombination of dissociated atoms on catalytic material surfaces
can cause severe heating and greatly increased oxidation rates.

INTRODUCTION

In order to meet the current demand for high-temperature data, the

National Advisory Cormnittee for Aeronautics is now operating, or is
putting into operation, many new high-temperature facilities. In these
facilities, attempts are being made to duplicate environmental conditions
which might be encountered by hypersonic airplanes and missiles in flying
through and in reentering the earth's atmosphere. Current research pro-
grams are aimed, first, at furnishing sufficient data to insure the suc-
cessful design and operation of hypersonic aircraft and, second, at
furthering the understanding of high-temperature phenomena. Of particular
interest at this time are reactions between structural materials and high-
temperature environments. This paper w i l l briefly describe two types of
high-temperature air-jet facilities and discuss results of some materials
tests conducted in them.

CERAMTC-HEATED AIR JET

Experience with rocket and combustion jets has sham that results
were influenced $y jet composition, that is, concentration of oxygen and
steam, both of which oxidize most of the conventional aircraft structural
materials. I n order t o avoid t h i s limitation, ceramic-heated air j e t s
have been designed and put i n t o operation. The general arrangement of
t h i s type of f a c i l i t y i s sham i n f i g u r e 1. It consists of a c y l i n d r i c a l
s t e e l pressure vessel which i s lined w i t h various types of r e f r a c t o r i e s
and interlocking ceramic refractory bricks. The c e n t r a l hole i s f i l l e d
with 3/8-inch-diameter lime-stabilized zirconia pebbles. A burner a t
the top supplies hot combustion gases which are drawn downward through
the bed. When the required bed temperature and temperature d i s t r i b u t i o n
a r e reached, compressed air i s blown back through the bed and expanded
by a water-cooled nozzle.

Operating c h a r a c t e r i s t i c s of two completed f a c i l i t i e s are as

follows :

Operating conditions Laboratory m o d e l P i l o t model

......
J e t - e x i t diameter, i n . 0.75 t o 1.00 4.0
..
Stagnation pressure, lb/sq i n . To 100 To 1,600
Stagnation temperature, ?F .... To 4,000 To 4,000
Jet velocity, f t / s e c....... 3,500 t o 5,200 4,300 t o 6,100
Mass flow, lb/sec ......... 0.2 t o 0.5 6 t o 10
G a s composition ..........
The maximum operating temperature i s now limited t o 4,000° F because the
pebble bed begins t o soften and lose strength a t slightly higher tempera-
tures. Tests a r e now i n progress which indicate t h a t the maximum temper-
ature may soon be increased t o 5,000° F by t h e use of t h o r i a bricks and
pebbles.

Up t o the present time, materials tests were conducted t o determine

i f the material reacted w i t h the high temperature, high-velocity airstream,
and, i f so, t o w h a t extent the reaction contributed t o t h e melting or t h e
destruction of the material. These tests consisted of putting m o d e l s of
various materials i n t o the j e t and observing t h e i r behavior by means of
high-speed color motion pictures. mese pictures were carefully studied
t o determine i f oxidation or i g n i t i o n occurred and t o determine the nature
and severity of t h e reaction. J e t stagnation temperature i s usually
higher than the melting point of the materials, and, i n t h i s event,
melting would occur even i n the absence of any reaction.

Solid models of various shapes, having a maximum diameter of 1/2 inch,

were used i n a l l tests. Typical shapes were cones with d i f f e r e n t angles
and nose radii, hemisphere cylinders, and f l a t - f a c e cylinders. I n i t i a l

.. n
428
 3

heating rates were greatly influenced by shape as w e l l as by airstream

properties. 'Ilests of i d e n t i c a l models of various materials were used
t o compare t h e r e l a t i v e behavior of the materials. Tests of i d e n t i c a l -
materials of various shapes were used t o compare the e f f e c t of shape and
heating rate.

Results of tests of t y p i c a l s t r u c t u r a l materials i n the ceramic-

heated a i r j e t w i l l now be discussed.' Operating conditions f o r the
t e s t s are as follows:

Stagnation temperature, OF .................. 4,000

..............
Stagnation pressure, lb/sq i n . abs 105
.....................
J e t velocity, f t / s e c 5,200
Jet Mach nmber ....................... 2
....................
J e t - e x i t diameter, in. 0.8

Magnesium

The magnesium model w a s a s o l i d 20° cone with a 1/32-inch nose radius.

The material being t e s t e d formed only the f i r s t 3/8 inch of t h e cone. B e
remainder of the cone w a s a s t e e l model support which w a s separated from
t h e t e s t specimen by a b a k e l i t e i n s u l a t o r . I n t h e motion p i c t u r e s , t h i s
i n s u l a t o r w i l l be observed t o i g n i t e and burn before the model m e l t s or
i g n i t e s . This burning of t h e i n s u l a t o r plays no p a r t i n the behavior of
the model and should be disregarded. This cone-shape model had a maximum
i n i t i a l heating r a t e of 2,000 Btu per square f o o t per second.

The magnesium began melting 0.2 second after entering t h e j e t with

no noticeable oxidation or i g n i t i o n . Melting continued and the m e t a l
flowed back f r e e l y and f r o z e on cooler portions of t h e model support.
Intermittent i g n i t i o n and combustion appeared i n 0.6 second. The model
w a s completely melted i n 0.9 second. Downstream i n the wake, out of t h e
camera f i e l d , magnesium w a s vaporized and burned vigorously.

Aluminum

The aluminum m o d e l was of t h e sa.he design as t h e magnesium m o d e l .

The aluminum began melting 0.4 second after entering t h e j e t . Melting
w a s not accompanied by noticeable oxidation or i g n i t i o n . Aluminum flowed
f r e e l y and f r o z e on the cooler portions of the model support. The model
was completely melted i n 1.5 seconds. Downstream i n the wake, out of the
view of the camera, aluminum ignited and burned i n locations where the
model support acted as a flame holder.
A' notion-picture f i l m supplement (L217) has been prepwed and i s
available on loan from NACA Headquarters, Washington, D. C.
 Steel

The s t e e l model w a s t h e same design as the magnesium model. S t e e l

began oxidizing immediately after entering t h e j e t as evidenced by d i s -
coloration a t the t i p . The m o d e l heated t o incandescence. Melting and
simultaneous i g n i t i o n were noted 0.7 second a f t e r entering t h e j e t . The
r e a c t i o n rate increased rapidly and completely engulfed t h e model and
model-support surface i n flame. The model w a s completely destroyed i n
3.5 seconds.

Stainless Steel

The stainless-steel modelwas t h e same design as the magnesium model.

S t a i n l e s s s t e e l began heating up w i t h the formation of oxides as evidenced
by t i p discoloration. The oxide began t o build up and formed a hard,
crusty, adherent deposit a t the t i p . This sxide afforded some protection
from f u r t h e r oxidation u n t i l subsurface melting began 1.1 seconds after
the model entered t h e j e t . The oxide w a s blown away, new surface w a s
exposed, and t h e oxidation process w a s then repeated. Soon, a s t h e model
heated up, t h e r e a c t i o n became vigorous enough t o cause continuous melting.

Copper

The copper m o d e l was the same design as t h e magnesium model. Copper

began t o form a black cupric oxide 0.5 second after entering t h e j e t .
This oxide continued t o form a t a r e l a t i v e l y slow rate. It w a s very
adherent and appeared t o p r o t e c t t h e copper from rapid oxidation. I n
2.9 seconds, the cupric oxide began t o decompose t o t h e red cuprous oxide
a t t h e t i p , and i n 3.0 seconds melting began. The model w a s completely
destroyed i n 5 seconds.

Titanium

The titanium m o d e l w a s a 1/2-inch-diameter hemisphere cylinder.

I n i t i a l heating r a t e w a s approximately 500 Btu per square f o o t per second.
The c y l i n d r i c a l model required a much longer time' t o heat than the cone-
shape models because of the lower heating rate and t h e l a r g e r surface-to-
mass r a t i o . The m o d e l began oxidizing 2 seconds after entering the j e t .
This oxide formed u n t i l a stable coating protected t h e model from further
oxidation. In approximately 13 seconds, a t a temperature approaching the
melting point, titanium suddenly burst i n t o a v i o l e n t reaction which
completely destroyed the m o d e l i n approximately 1 second.
 5

Molybdenum

The molybdenum model w a s the same design as the magnesium m o d e l

except t h a t the rear conical support w a s graphite and no bakelite sepa-
r a t o r w a s used. The melting point of molybdenum w a s approximately 600° F
above t h e j e t stagnation temperature. The model entered t h e j e t and, i n
0.2 second, a black oxide, probably the sesquioxide, w a s formed. I n
1.3 seconds, a gray oxide, probably a mixture of the sesquioxide and
trioxide, formed. I n 2.5 seconds, a yellow oxide, molybdenum t r i o x i d e ,
formed. All these oxides were v o l a t i l e and condensed out on portions of
the model and support which were cooler than t h e sublimation temperature
of t h e oxide. A s t h e temperature increased, they a l l vaporized. I n
3 . 1 seconds, a vigorous reaction began a t the t i p . I n 3.2 seconds, i g n i -
t i o n occurred and melting began. Vigorous oxidation and burning a t the
surface furnished s u f f i c i e n t a d d i t i o n a l heat input t o cause and s u s t a i n
melting.

Tungst e n

The tungsten model w a s t h e same design as the molybdenum m o d e l . The

melting p o i n t of tungsten was approximately 2,000' F above the stream
stagnation temperature. Tungsten began t o oxidize 0.5 second a f t e r
entering the j e t with the formation of a black oxide, probably t h e
dioxide. I n 2.5 seconds, a yellow oxide, probably t h e t r i o x i d e , w a s
seen forming and condensing out on the cooler portions of the model sup-
port. This oxidation, although not vigorous, w a s steady. The tungsten
oxidized with reasonable l o s s of mass of t h e m o d e l ; however, oxidation
w a s not s u f f i c i e n t t o cause melting or i g n i t i o n . 'Phe t r i o x i d e vaporized
or reacted with the graphite model holder as t h e temperature increased.
!The graphite m o d e l holder w a s oxidized t o a much greater e x t e n t than it
would have been i f it were tested alone; t h i s f a c t indicates t h a t tungsten
or one of i t s oxides catalyzed the oxidation of graphite.

Graphite Coated With Flame-Sprayed Zirconia

!The zirconia-coated graphite m o d e l w a s a 30° cone with a diameter

of 3/4 inch and a nose radius of 1/8 inch. The model survived 10 seconds
of t e s t with no damage. The t e s t was terminated by f a i l u r e of t h e support
system of the m o d e l .

HIGH-INTENSITY ARC-POWERED JETS

Ceramic-heated jets can produce temperatures which m e l t or severely

oxidize most conventional s t r u c t u r a l materials. These temperatures are

4ZI
6

considerably less than environmental temperatures which are associated

with reentry. It is theoretically impossible to obtain reentry tempera-
tures by chemical means, either directly or indirectly. For production
of these high temperatures, the Langley Laboratory has been developing
jets in which the energy is supplied by a high-intensity electric arc.

Initial operation of these arc-powered jets was conducted with

water and liquid oxygen as the jet medium. More recently, they have
been successfully operated with liquid and gaseous air as the jet medium.
Operation with air is now being emphasized inasmuch as this medium more
nearly reproduces the atmospheric environment at reentry velocities.

An arc-powered jet is shown schematically in figure 2. It consists

of a pressure chamber that houses the arc and the jet medium which is
to be heated. The anode enters the chamber through a hole in the bottom.
The top forms the cathode and a hole in the top forms the nozzle. A
high-intensity arc is struck and maintained between the anode and the
cathode. Air enters the bottom of the chamber and flaws into the arc
chamber through an annular opening around the anode. It is then heated
and expanded to high velocity through the nozzle. Power is supplied as
indicated. This diagram shows a direct-current arc. 'Ilhree-phase
alternating-current arcs have also been operated successfully.

Characteristics of typical arc-powered jets which have been operated

at this laboratory are as follows:

.................
Jet-exit diameter, in. 0.25 to 1.00
Power, kw
Stagnation temperature, +
....... ...............
...............
.............
Stagnation pressure, lb/sq in.
60to1,000
>15,000
3 to 30
. . . . . . . . . . . . . . . . . . Air
Jet composition (dissociated)
......
Heating rate (hemisphere nose), Btu/ft2-sec 2,200
Temperatures and pressures indicated represent what has been achieved up
to the present time and do not represent the maximum capabilities of this
type of facility.
Materials tests similar to those conducted in the ceramic-heated air
jets have also been run in the high-intensity arc-powered jets. For the
purpose of comparing the relative severity of test conditions and behavior
of materials in these two jets, results of tests of titanium, molybdenum,
and graphite models will be presented. Jet conditions for these tests
are as follows:
7F
7

. . . . . . . . . . . . . . . . . >13,000
Stagnation temperature, OF
...............
Stagnation pressure, lb/sq i n . 10 -

J e t velocity ........................ Subsonic

J e t medium ......................... Air
Jetdia.meter,in. ..................... 0,375
. . . . . . . . . . . 1,000 t o 2,200
I n i t i a l heating r a t e , Btu/ftZ-sec

Titanium

The titanium model i s a l/;l-inch-diameter f l a t - f a c e cylinder. A

hemisphere-cylinder model with one-half the i n i t i a l heating r a t e with-
stood 13 seconds i n the ceramic-heated a i r j e t before i g n i t i o n occurred.
The model i n the arc-powered a i r j e t i g n i t e d i n approximately 0.1 second.
The reaction was violent. Heating rates were extremely high. Examina-
t i o n of t h e model afterward showed l i t t l e penetration of heat i n t o t h e
model and indicates t h a t most of the material was l o s t i n a manner
resembling ablation.

Molybdenum

The molybdenum model was the same shape as the titanium model. The
molybdenum model having similar i n i t i a l heating r a t e s required 3.1 seconds
t o melt i n the ceramic-heated air j e t . The model i n the arc-powered air
j e t began t o r e a c t violently i n 2 seconds. Examination of the model a f t e r
the t e s t a l s o indicated t h a t material w a s l o s t i n a manner resembling
ablation.

Graphite

The graphite model was a 1/2-inch-diameter hemisphere cylinder.

Graphite w a s r e l a t i v e l y undamaged by a 20-second exposure i n the ceramic-
heated air j e t . I n the present case, t h e graphite model suffered l i t t l e
damage due t o a 5-second exposure t o the arc-powered a i r j e t . Some
material was l o s t because of oxidation o r sublimation, although t h e model
was rapidly heated t o incandescence.

The t e s t s i n both j e t s were contrived t o produce rapid model heating

so t h a t materials reactions, i f present, would begin soon a f t e r t h e models
were introduced i n t o the j e t s . Oxidation and i g n i t i o n are complex
phenomena which depend on many f a c t o r s and conditions. Some of these
are the temperature of both reactants, the thermal properties of react-
ants and products, the thermal and xass transport properties of react-
ants and products, the k i n e t i c s of reaction, the physical properties of
reactants and products, and the surface-to-mass r a t i o of t h e model.
Changes i n the s i z e and shape of models may well a f f e c t some of these
423
properties and conditions. Until a suitable explanation for the inter-
relation of these properties and conditions is found, results of these
tests should be regarded as qualitative, and should not be applied
directly to models of other sizes and shapes.

Both types of jets are being operated as free jets exhausting into
the atmosphere. This type of operation allows reasonable duplication of
stagnation condi%ions, and to some extent, velocity. It does not allow
duplication of static conditions. Exact duplication of both static and
stagnation conditions will require the use of vacuum chambers, diffusers,
or auxiliary ejectors.

Both jets may be operated with inert gases, pure gases, and gas
mixtures. Use of inert gases would allow determination of heat transfer
in the absence of oxidation. Pure and mixed gases may be used to study
reactions in controlled environments, and to study the aerothermochemical
phenomena.

One such phenomenon, that of enhanced oxidation by oxygen atoms at

elevated temperatures, has been studied and reported by Fryburg in refer-
ence 1. Platinum was chosen for this investigation because of its known
catalytic activity in promoting recombination of oxygen atoms. Platinum
strips were heated and their temperatures were maintained at temperatures
of l,OOOo C and above solely by recombination of oxygen atoms at the
surface. The rates of oxidation of platinum were measured and found to
be directly proportional to the number of atoms striking the surface.
Oxygen atoms were 400 times more effective in oxidizing platinum than
oxygen molecules, thereby indicating that extrapolation of low-temperature-
oxidation data obtained in conventional environments to reentry conditions
could cause serious errors.

CONCLUDING REMARKS

High-temperature air-jet facilities which can duplicate some environ-

mental condition of hypersonic flight up to reentry velocities are now
being operated.

Tests of materials in high-velocity airstreams at temperatures of

4,000° F and greater indicate that conventional structural materials
undergo melting, ignition, .or severe oxidation.

Oxidation and ignition characteristics of materials may be limiting

properties rather than strength, creep, or fatigue.

Under certain conditions, recombination of dissociated atoms on

catalytic surfaces can cause severe heating and oxidation.
 9

The f a c t t h a t materials undergo oxidation and ignition, and t h a t

these reactions can give rise t o large surface heat inputs, suggests t h e
use of caution i n extrapolating law-temgerature heat-transfer d a t a t o
reentry conditions.

1. Fryburg, George C.: Enhanced Oxidation of Platinum i n Activated

Oxygen. Jour. Chem. Phys., vol. 24, no. 2, Feb. 1956, pp. 175-180.
 CERAMIC- HEATED J E T
WATER -COOLED N O Z Z L E
WATER -COOLED CHAMBER
c

BURNER
ZIRCONIUM OXIDE SPHERES OIL INLET

BURNER AIR I N L E T
INSULATING ZIRCONIA BRICK
INSULATING FIREBRICK

DENSE ZIRCONIA BRICK

CAST INSULATING BRICK

DIRECTION OF COMBUST10
PRODUCTS DURING FIRING

DIRECTION OF AIR FLOW

DURING BLOWDOWN
4

Figure 1

ELECTRIC-ARC AIR JET

I ANODE

426 Figure 2
 4

FATIGUE-CRACK PROPAGATION AND RESIDUAL STATIC S'I!RENGTH

OF BUILT-UP STRETURES

By Herbert F. W d r a t h and Richard E. Whaley

Langley Aeronautical Laboratory

Fatigue tests were conducted on box beams and tension panels i n

order t o study sane of t h e f a c t o r s affecting fatigue-crack propagation.
The box beams had e s s e n t i a l l y the same configuration except f o r t h e mode
of connecting s t r i n g e r s t o t h e tension cover. The beams w i t h bonded
s t r i n g e r s had t h e lowest r a t e of crack growth, and beams w i t h riveted
and i n t e g r a l s t i f f e n e r s had successively higher rates of crack growth.
Crack growth w a s slower i n beams w i t h close r i v e t spacing than i n beams
w i t h greater r i v e t spacing. The tension panels were a l l of the same
general configuration except t h a t t h e proportions.of cross-sectional
areas of skin, s t r i n g e r s , and flanges were varied. Panels w i t h heavy
s t r i n g e r s and thin s k i n had lower r a t e s of crack growth than did panels
w i t h heavy skin and l i g h t s t r i n g e r s .

S t a t i c t e s t s were performed on box beams, on tension panels, and

on two types of wings, a l l of which contained f a t i g u e cracks. The com-
parison of results with predictions made by a simple theory indicates
that t e s t r e s u l t s were affected by a r e d i s t r i b u t i o n of loads among t h e
various r m a i n i n g elements and by whether cracks terminated a t r i v e t
holes.

INTRODUCTION

During the p a s t several years t h e idea of "fail-safe" design has

become a very popular t o p i c f o r discussion among a i r c r a f t s t r u c t u r a l
designers. Although t h e conditions f o r c a l l i n g a given design f a i l - s a f e
have not been c l e a r l y defined, almost a l l engineers concerned agree on
several general conditions. F i r s t , the progress of a f a t i g u e crack through
a s t r u c t u r e must be reasonably slow, preferably i n a readily inspectable
location. Second, t h e s t r u c t u r e containing a crack must r e t a i n enough
s t a t i c strength t o withstand s m e specified load.

The m e a n s f o r accmplishing these ends involve such f a c t o r s as the

selection of materials w agation and s t a t i c
I
427
2

notch-strength properties, t h e arrangement of material t o i n h i b i t crack

growth, the provision of multiple load paths, and others. The purpose
of t h e present paper is t o review current research which deals w i t h t h e
systematic study of sane of these f a c t o r s and t h e i r application t o the ?

design of s t r u c t u r e s used i n wings.

CRACK PROPAGATION

Box Beams

One phase of t h e study of crack propagation involves f a t i g u e t e s t s

of box beams such as those shown i n f i g u r e 1. Two general configurations
were t e s t e d . For t h e first configuration t h e tension cover had i n t e g r a l
s t i f f e n e r s and was machined from a plate. For t h e other configuration
t h e s t r i n g e r s and skin were e i t h e r bonded or riveted. The webs and cam-
pression covers on a l l beams were of i d e n t i c a l built-up construction f o r
simplicity i n construction and analysis. A l l beams were 20 inches w i d e
and 8 f e e t long. I d e n t i c a l beams were constructed in each of t h e aluminum
alloys, 2024 and "Om. The beams were loaded as shown i n f i g u r e 1 t o
produce t e n s i l e s t r e s s e s of 13 f 6.5 k s i i n t h e carry-through bay. A n
oblong hole was made i n t h e center of t h e carry-through bay t o i n i t i a t e
t h e crack a t t h a t point. Cracks generally g r e w symmetrically across t h e
chord. Although sane of t h e d a t a have been published previously (ref. 1) ,
representative r e s u l t s a r e given i n figure 2.

In figure 2 t h e percentage of t h e tension area l o s t by fatigue

cracking i s p l o t t e d as a function of t h e number of cycles of load applied
a f t e r crack i n i t i a t i o n . The curves a r e f o r beams w i t h integral, riveted,
and bonded covers. The material i n each case w a s 7075 aluminum alloy.
The r e s u l t s f o r 2024 aluminum-alloy beams are not s h o q , but t h e same
general observations apply except t h a t crack growth i s appreciably slower
i n beams made of 2024 aluminum alloy.

A s indicated by t h e curves, t h e cracks g r e w l e a s t rapidly i n beams

w i t h bonded covers. The crack growth w a s confined t o t h e s k i n and pro-
gressed a t a reasonable r a t e , probably controlled by support Tram t h e
s t r i n g e r s . Measurements of s t r i n g e r s t r e s s e s taken a t i n t e r v a l s during
t h e t e s t indicated that these s t r e s s e s increased more slowly than s t r e s s e s
i n beams w i t h other types of connections, probably because of the f a c t
t h a t t h e bond between the skin and s t r i n g e r s peeled back as t h e crack grew
across t h e beam. Since there were no r i v e t holes i n t h e s t r i n g e r s , no
s t r e s s raisers w e r e present, and no s t r i n g e r s failed before the crack had
grown completely through t h e skin.
In riveted covers t h e s t r i n g e r s were sanewhat more vulnerable t o
f a i l u r e because s t r i n g e r s t r e s s e s increased more rapidly and s t r e s s
 3

r a i s e r s due t o r i v e t holes were present. The l o s s of s t r i n g e r s i n t h e

case of riveted beams contributes t o t h e more rapid l o s s of tension
material. In t h e case of the i n t e g r a l l y s t i f f e n e d covers the cracks
grew a t t h e f a s t e s t r a t e because no natural b a r r i e r s t o crack growth
were present,

These observations indicate t h a t crack propagation is very much a

function of t h e effectiveness of connections between t h e sheet and
s t r i n g e r s . An extension of t h i s work t o beams i n which t h e r i v e t p i t c h
was varied was therefore undertaken, and some of t h e r e s u l t s of t h i s
work are shown in f i g u r e 3.

I n f i g u r e 3 t h e curves from f i v e 2 a r e replotted as s o l i d l i n e s .

New curves f o r beams w i t h r i v e t pitches of one-half and of twice t h e
r i v e t p i t c h used i n t h e previous beam a r e shown by dotted l i n e s . The
symbols represent t h e stage of t h e t e s t when one o r more s t r i n g e r s had
completely failed. Crack growth was slower i n beams w i t h a r i v e t p i t c h of
3/4 inch than i n those w i t h a r i v e t p i t c h of 1
; inches and was much slower
than i n beams w i t h a r i v e t p i t c h of 3 inches. The s t r e s s measurements
previously mentioned indicated t h a t s t r e s s e s i n s t r i n g e r s straddling
cracks increased more rapidly i n beams with a r i v e t p i t c h of 3/4 inch
than i n other beams w i t h riveted s t r i n g e r s . This increase i n s t r e s s e s
had two e f f e c t s : a slow r a t e of crack propagation i n t h e sheet as indi-
cated by t h e low i n i t i a l slope of t h i s curve, and increased probability
of s t r i n g e r f a i l u r e as indicated by t h e f a c t t h a t i n t h i s beam two
s t r i n g e r s f a i l e d after only about 8 percent of t h e s t r u c t u r e had been
l o s t by skin cracking. On t h e other hand, i n beams w i t h a r i v e t p i t c h
of 3 inches, t h e s t r i n g e r s t r e s s increased more slowly and indicated
t h a t l e s s support w a s given t o the sheet; consequently, crack growth
was very rapid i n t h e skin. The r e s u l t s of these t e s t s show t h a t the
closer t h e r i v e t pitch, t h e b e t t e r t h e resistance t o crack propagation.
Fabrication limitations prevent decreasing t h e r i v e t p i t c h further.
I n t e g r a l construction, which m i g h t appear t h e same as r i v e t s w i t h a zero
pitch, displays rapid crack growth. The reason appears t o be t h a t i n
i n t e g r a l construction only one crack needs t o be s t a r t e d , and then t h a t
crack grows caupletely through t h e panel.

Tension Panels

Another phase of the investigation of crack propagation involves

tension tests of stiffened panels. Sme of t h e results a r e shown i n
figure 4. The panels t e s t e d were 30 inches wide and were composed of
a skin, f o u r s t r i n g e r s , and two flanges. The parameters varied were t h e
percentages of t h e areas of skin, s t r i n g e r s , and flanges as indicated by
values listed i n t h e figure. Two of t h e configurations had 40 percent
of t h e area i n t h e skin, and one configuration had 80 percent of t h e
4 +*

area i n the skiin. Many current a i r c r a f t have proportions within this

range. Repeated tension loads were applied t o produce nominal s t r e s s e s
of 14 & 4.7 ksi. Fatigue cracks were i n i t i a t e d a t cutouts w i t h t h e
shapes indicated i n f i g u r e 4. Although only a few results a r e available,
preliminary discussions a r e of i n t e r e s t .

The heavier s t r i n g e r s i n panels w i t h 40 percent of t h e area i n t h e

skin appear t o have controlled the r a t e of crack growth more e f f e c t i v e l y
than i n t h e panels with 80 percent of the area in the skin. The result
i s an appreciably lower r a t e of crack propagation i n panels with lighter
gage skin. One configuration had a 3-inch-square cutout which w a s t h e
width of one skin panel. This cutout removed approximately 7 percent of
t h e o r i g i n a l net section, and t h e beginning of t h e curve f o r t h i s specimen
i s p l o t t e d a t t h a t value. In t h i s case t h e growth of f a t i g u e cracks w a s
very slow i n t h e i n i t i a l stages of t h e t e s t and i l l u s t r a t e d t h e b e n e f i c i a l
e f f e c t of framing members adjacent t o cutouts.

RESIDUAL STATIC STRENG!TH

The r e s t of t h i s paper deals w i t h t h e study of r e s i d u a l s t a t i c

strength i n the specimens j u s t described. The r e s u l t s which a r e pre-
liminary i n nature a r e discussed i n order t o indicate trends. These
r e s u l t s a r e then compared with r e s u l t s of a simple analysis. S t a t i c
t e s t s of Convair 240 (designated herein as CV-240) and c-46 wings con-
t a i n i n g fatigue cracks a r e a l s o discussed.

Box Beams
In figure 5 t h e ordinate represents the ultimate load producing
s t a t i c f a i l u r e of beams w i t h cracks expressed as a percent of t h e load
calculated t o produce f a i l u r e of the tension covers i n uncracked beams.
The abscissa represents t h e length of the f a t i g u e crack i n t h e cover skin.
The symbols represent the r e s u l t s of s t a t i c t e s t s of box beams which were
made of 7075 aluminum a l l o y and which were i d e n t i c a l except f o r t h e
fastenings between t h e skin and s t r i n g e r s . The symbols represent e i t h e r
bonded covers or riveted covers as shown i n f i g u r e 5. The dotted l i n e
represents t h e strength of t h e beam having a skin crack only, w i t h no
allowance made f o r s t r e s s concentration K due t o t h e crack. The s o l i d
l i n e s represent predictions made by calculations of t h e s t a t i c strength
of a specimen containing a fatigue crack i n t h e skin only. The basis of
t h i s method i s t o compute a stress-concentration f a c t o r , K for the
sheet by t h e method outlined i n reference 2. The residual s t a t i c strength
of a sheet containing a crack was added t o the s t a t i c strength of s t r u c -
turalmembers such as s t r i n g e r s and flanges t o produce t h e values f o r t h e
upper curve. Each of t h e other curves was ccanputed i n a similar w a y
F

except t h a t one or more s t r i n g e r s were assumed f a i l e d . Each t e s t point

i s connected by a v e r t i c a l l i n e t o the curve appropriate f o r t h e number
of s t r i n g e r s failed i n t h e specimen represented.

When t h e crudeness of t h e method used is considered, t h e agreement

between predicted strength and a c t u a l strength of r i v e t e d beams i s good.
The beams w i t h a r i v e t p i t c h of 3/4 inch f a l l farthest below t h e predicted
curves, probably because of t h e higher s t r e s s e s c a r r i e d by s t r i n g e r s .
Evidently t h e d i s t r i b u t i o n of loads among the remaining members i n the
s t r u c t u r e m u s t be taken i n t o account i n order t o improve t h e predictions.
The strengths of beams w i t h bonded covers are higher than t h e respective
predicted strengths while other strengths are lower than predicted. No
reason f o r t h i s behavior has been found. One point f e l l above t h e dotted
l i n e which represents loss of strength equal t o loss of area. This high
strength w a s caused by t h e f a c t t h a t t h e a c t u a l material strength w a s
higher than t h e specification values used i n computations. Adjustment
of the computation f o r a c t u a l material proportions would a f f e c t a l l t h e
results.

Tension Panels

Figure 6 gives results similar t o those i n f i g u r e 3 f o r s t a t i c t e s t s

of t h e tension panels w i t h 80 percent of the area i n the skin. I n t h i s
case t h e data were somewhat higher than predictions (shown by s o l i d l i n e s
i n t h e f i g u r e ) made by t h e method previously used. The reason appears
t o be t h a t i n these panels t h e fatigue cracks ended a t r i v e t holes.
The computation of t h e stress-concentration f a c t o r should, therefore,
allow f o r a radius of curvature p equal t o t h e radius of t h e r i v e t
instead of t h e e f f e c t i v e radius a t t h e r o o t of a crack. Calculations
based on t h i s assumption are indicated by t h e dashed l i n e s , and t h e
t e s t r e s u l t s f a l l below t h e predictions as before.

CV-240 Wings

Same CV-240 outer wing panels were subjected t o repeated loading i n

order t h a t crack growth m i g h t be studied, and then s t a t i c tests were
performed i n order t h a t the r e s i d u a l s t a t i c strength might be determined.
Figure 7 presents t h e r e s u l t s of three s t a t i c t e s t s of CV-240 wings con-
t a i n i n g f a t i g u e cracks. The p l o t of f i g u r e 7 i s similar t o t h e p l o t s of
figures 5 and 6 except t h a t t h e abscissa i s t h e tension cross-sectional
area failed expressed as a percentage of the t o t a l tension cross-sectional
area. The dotted l i n e represents reduction i n s t a t i c strength i n t h e same
proportion as the reduction i n area. These wings were constructed of
7On aluminum alloy, and t h e s t r u c t u r e w a s somewhat more cmplex than t h a t
of t h e box beams discussed previously. A s before, t h e stress concentra-
t i o n was computed f o r t h e skin only. Also, since cracks terminated a t
6

r i v e t holes o r g r e w t o r i v e t holes a t an e a r l y stage of t h e s t a t i c t e s t ,

t h e r i v e t radius w a s used i n t h e calculations. The cracks i n these winns
originated i n t h e r e a r spar caps and then g r e w across t h e chord. "he c a l -
culation w a s , therefore, f o r a sheet w i t h a notch on one side only. The
t e s t r e s u l t s f e l l s l i g h t l y below t h e prediction as before.

C-46 wings
The f a t i g u e t e s t s of C-46 wings have been discussed i n references 3
and 4. Thirteen wings were s t a t i c tested a f t e r various amounts of t h e
tension material were f a i l e d i n f a t i g u e tests. The r e s u l t s of the s t a t i c
t e s t s have been presented i n reference 4.

The results a r e a l s o shown i n figure 8 i n the same ty-pe of p l o t as

w a s used i n f i g u r e 7. The predicted curve w a s computed on the a s s m p t i o n
t h a t t h e skin w a s continuous across t h e chord. The r i v e t radius w a s used
i n t h e calculation f o r t h e same reason as before. I n s p i t e of t h e f a c t
t h a t a w i d e v a r i e t y of s t r u c t u r a l members failed during t h e f a t i g u e tests
on these specimens, t h e predicted strengths f e l l i n a very narrow band
which i s represented i n t h e figure as a s i n g l e l i n e . The discrepancy
between results and predictions w a s somewhat greater than i n t h e previous
cases. This discrepancy was t o be expected as a result of t h e much more
complicated s t r u c t u r e i n t h e C-46 wings. Obviously, t h e r e d i s t r i b u t i o n
of loads i n t h i s s t r u c t u r e w i l l have t o be considered before more accurate
predictions can be made.

The apparent l a r g e l o s s of strength w i t h small cracks applies only

t o t h e tension surface. In compression-critical wings, such as the C-46
and CV-240, t h e l o s s i n wing strength is very s m a l l u n t i l large cracks
a r e present.

CONCLUDING REMARKS

Crack-propagation and s t a t i c - s t r e n g t h t e s t s i n several types of

built-up specimens and f u l l - s c a l e wings have been reviewed. The r e s u l t s ,
t o date, i n d i c a t e t h a t the rate of crack propagation i s influenced strongly
by t h e mode of connecting t h e skin t o s t r i n g e r s and by the proportions of
areas of the s k i n and s t r i n g e r s . The analysis of r e s i d u a l s t a t i c strength
of complex s t r u c t u r e s indicates the f e a s i b i l i t y of applying simple methods,
'Gut the results are subject t o questions regarding t h e r e d i s t r i b u t i o n of
loads, i n t e r a c t i o n s between various members, and such seemingly t r i v i a l
considerations as whether or not a crack terminates a t a r i v e t . Much work
remains t o be done on these problems. Other configurations designed t o
improve both the rate of crack propagation and r e s i d u a l s t a t i c strength
should be investigated.
 ’
7

l3ERBENCES

1. Hardrath, Herbert F., Leybold, Herbert A., Landers, Charles B., and
Hauschild, Louis W: : Fatigue-Crack Propagation i n Aluminum-kloy
Box Beams. NACA TN 3856, 1956.

2. McEvily, Arthur J., Jr., I l l g , Walter, and Hardrath, Herbert F.:

S t a t i c Strength of Aluminum-Allay Specimens Containing Fatigue
Cracks. NACA TN 3816, 1956.

3. McGuigan, M. J., Jr., Bryan, D. F., and Whaley, R. E.: Fatigue

Investigation of Full-scale Transport-Airplane Wings - Surmnary
of Constant-Amplitude Tests Through 1953. NACA ”IV 3190, 1934.

4. Whaley, Richard E., McGuigan, M. J., Jr., and Bryan, D. F.: Fatigue-
Crack-Propagation and Residual-Static-Strength Results on Full-scale
Transport-Airplane Wings. NACA TN 3847, 1956.

433
5

CONFIGURATION OF BOX BEAMS

INTEGRAL RIVETED OR BONDED

t
STRESS 3 13.0 f 6.5 KSI

Figure 1

CRACK GROWTH IN BOX BEAMS

7075 ALUMINUM ALLOY

AREA f/-RIVETED
LOST,
%
15

0 20 40 60 80 100x103
CYCLES AFTER INITIATION OF CRACK

434
 s

CRACK GROWTH IN BOX BEAMS

7075 ALUMINUM ALLOY
FAILURE
?I 0 2 STRINGERS
I 0 I STRINGER
I RIVET
&INTEGRAL PITCH,
INCHES
30

AREA
LOST,
"/.
15

0 20 40 60 80 1OOX, 1 0 3
CYCLES AFTER INITIATION OF CRACK

Figure 3

CRACK PROPAGATION IN TENSION PANELS

7075 ALUMINUM ALLOY
CUTOUT

SHEET AREA,%
0

............ 80
0

40
040
STRINGERS,% .............I O 30 30
FLANGES, %.................IO 30 30
50

-
AREA
LOST, -
Sb

Figure 4 ,
435
 STATIC STRENGTH OF BOX BEAMS
707'5 ALUMINUM ALLOY
100 t

80

60
ULTlMATE

1
LOAD,
OF ORIGINAL
-3
40
1 0 BONDED h
2o h $INPITCH]
o I i I N . PITCH RIVETED
* 3 IN. PITCH
I I I I 1
0 5 10 15 20
CRACK LENGTH, INCHES

Figure 5

STATIC STRENGTH OF TENSION PANELS

7075 ALUMINUM ALLOY

100

ULTIMATE
LOAD,% 50
OF ORIGINAL

0 10 20 30
CRACK LENGTH, INCHES

Figure 6
436
 11

STATIC STRENGTH OF CV-240 WINGS

7075 ALUMINUM ALLOY

ULTIMATE 4
"/, OF ORIGINAL

0 20 40 60 80 100
AREA LOST,%

Figure 7

STATIC STRENGTH OF C-46 WINGS

2024 CLAD ALUMINUM ALLOY

100

80

60
ULTIMATE LOAD,
"/o OF ORIGINAL
40

20

Figure 8
437
F

SOME ASPECTS OF FAIL-SAFE DESIGN OF PRESSURIZED FUSELAGES

By Paul K u h n and Roger W. Peters

Langley Aeronautical Laboratory

Separate investigations have dealt with the critical crack length

of flat sheets or of unstiffened cyliniiers and with the type of rupture
experienced by stiffened cylinders. These investigations are correlated,
supplemented by new tests, and combined into a uniform scheme for pre-
dicting critical crack length and type of rupture in stiffened pressur-
ized cylinders.

INTRODUCTION

The fail-safe design of pressurized fuselages is a problem that has

attracted much attention in the past few years. This paper is a progress
report on work in this field and an attempt to correlate several lines of
investigation.

ANET net area, sq ini

AR ring area (cross sectional), sq in.

KU stress-concentration factor for ultimate load

K stress-concentration factor for cylinder

u,C-YL
1 ring spacing, in.

P load, ksi

r radius of cylinder, in.

tS skin thickness, in.

438
s length of o r i g i n a l s l i t , i n .

U stress, ksi

U maximum s t r e s s , k s i
MAX

UU ultimate t e n s i l e s t r e s s , k s i

DEFINITIONS

The problem i s defined i n a general way by t w o questions: If i n i -

t i a l damage, such a s a f a t i g u e crack, i s i n f l i c t e d on a pressurized
s t i f f e n e d s h e l l , how large can the damage be before t h e pressure causes
a rupture of the s h e l l , and w h a t i s the nature of the rupture? The
only type of i n i t i a l damage which i s considered here i s a longitudinal
crack o r s l i t of length 6 as shown i n figure 1. (For aluminum a l l o y s ,
t o which t h i s discussion is confined the difference between a fatigue
crack and a f i n e s l i t is negligible. For any specified pressure o r
hoop tension, a c r i t i c a l length of crack e x i s t s a t which the pressure
w i l l rupture t h e skin. The term "confined rupture" i n t h i s paper w i l l
r e f e r t o a rupture which stops a t t h e nearest rings, as indicated on
t h e sketch a t t h e l e f t . The term ''unconfined rupture" w i l l r e f e r t o a
rupture which extends i n t o adjacent bays, as indicated on the sketch
a t the r i g h t . An unconfined rupture of t h e skin often r e s u l t s i n f a i l -
ure of t h e rings and sometimes of t h e s t r i n g e r s .

'SURVEY OF PRFVIOUS INVESTIGATIONS

C r i t i c a l Crack Length

Figure 2 shows schematically t h e f i r s t two steps i n t h e investiga-

t i o n of c r i t i c a l crack length. A t the top i s shown a f l a t sheet under
tension with a c e n t r a l crack of length 6. When t h e maximum s t r e s s uMAx
i n the sheet a t the two ends of the crack becomes equal t o t h e t e n s i l e
strength of t h e material, t h e sheet w i l l t e a r a p a r t . A t t h i s i n s t a n t ,
t h e maximum s t r e s s is equal t o t h e product of t h e net-section s t r e s s
P / A ~and ~ a stress-concentration f a c t o r G.

A method f o r calculating t h e f a c t o r K, has been published i n

reference 1. The calculation involves the s t r e s s - s t r a i n curve and a
size-effect constant which i s determined from a tension t e s t on a speci-
men with a sharp notch of known radius. The presence of t h i s s i z e e f f e c t
i
 3

i n problems involving cracks or sharp notches invalidates t h e mechanical

l a w of s i m i l a r i t y that geometrically similar s t r u c t u r e s f a i l at' the same
s t r e s s and makes it impossible t o draw generalized curves based on dimen-
sionless parameters.

The lower sketch i n f i g u r e 2 shows a pressurized unstiffened cylinder

w i t h a longitudinal crack. The stress-concentration f a c t o r f o r such a
cylinder i s calculated by t h e formula shown. The quantity Ku i s t h e
stress-concentration f a c t o r calculated f o r t h e configuration obtained by
unwrapping t h e cylinder i n t o a plane. The term i n parentheses is t h e
"curvature correction," which w a s found empirically and applies t o 2024-T3
ahminun a l l o y as well as 7075-T6 aluminum a l l o y . The experimental basis
f o r t h e formula may be found i n reference 2.

The importance of t h e curvature correction i s shown i n figure 3.

For t h e two aluminum alloys, t h e c r i t i c a l crack length is shown as func-
t i o n of t h e t e n s i l e stress f o r f l a t sheet and f o r cylinders w i t h a radius
of 15 inches, which i s a widely used size representing roughly a 1/4-
s c a l e model of a fuselage. For a l l but very short cracks, t h e drop i n
s t r e n g t h due t o curvature e f f e c t is obviously s u b s t a n t i a l . Curves of
c r i t i c a l crack lengths of unstiffened cylinders corresponding t o t h e
curves shown i n f i g u r e 3 w i l l be used later as a yardstick o r reference
basis f o r s t i f f e n e d cylinders.

me of Rupture

The i n i t i a l investigation of t h e problem of type of rupture w a s con-

ducted on cylinders w i t h r a d i i of 15 o r 24 inches, s t i f f e n e d by s t r i n g e r s
and riveted-on rings. Subjecting these cylinders t o a constant i n t e r n a l
pressure and t o repeated t o r s i o n loads resulted i n f a t i g u e cracks a t 45'
t o t h e cylinder a x i s .

Figure 4 shows the results of t h e i n i t i a l investigation on 2024-T3

cylinders. Hoop stress i s p l o t t e d as t h e ordinate and ring-reinforcement
r a t i o , t h e r a t i o of t h e cross-sectional a r e a of a r i n g AR t o t h e area
of t h e associated skin 2 t s , i s p l o t t e d as t h e abscissa. The dashed
l i n e , labeled " t h e o r e t i c a l c r i t e r i o n , 'I is based on elemetary considera-
t i o n s and gives t h e a r e a which t h e rings must have t o carry t h e hoop
load if t h e skin i t s e l f cannot carry it because it is cracked or cut.
The c i r c l e s denote confined ruptures, t h e x-marks denote unconfined rup-
t u r e s i n t h e tests. The s o l i d l i n e , labeled "empirical c r i t e r i o n , " i s
approximately t h e upper boundary of t h e confined ruptures. The ring-
s i z e c r i t e r i o n can thus be used as a b a s i s f o r predicting t h e type of
rupture. The data are taken from reference 3, except that the curve
showing t h e empirical c r i t e r i o n is drawn here i n a somewhat more con-
servative fashion than i n the reference.
4

A few t e s t s of the same nature have been made on cylindeis of 7075-T6

material. The number of t e s t s was t o o small t o e s t a b l i s h an empirical
c r i t e r i o n f o r r i n g size, and the method of t e s t i n g , pressure combined
w i t h cyclic torsion, has been replaced by pressure cycling. t

NEW INVESTIGATIONS

Presented herewith a r e evaluations of more recent data. The pres-

entation is aimed a t answering two questions:

(1)What correlation e x i s t s between t h e c r i t i c a l crack length f o r

unstiffened cylinders and that f o r s t i f f e n e d cylinders?

(2) How r e l i a b l e i s t h e ring-size c r i t e r i o n f o r predicting the nature

of the rupture of s t i f f e n e d cylinders when t h e cracks a r e longitudinal
r a t h e r than at 45O, a s in t h e o r i g i n a l investigation?

The majority of t h e tests discussed were a c t u a l l y made t o obtain

some preliminary information on t h e e.ffect of parameters not considered
previously, f o r instance, type of rings used. A s a r e s u l t , t h e t e s t s
available a t t h i s time a r e inadequate f o r giving d e f i n i t e answers t o t h e
two questions posed, but they do indicate trends.

The t e s t s f a l l i n t o t h r e e groups according t o s i z e of cylinder:

small cylinders with a radius of 3.6 inches; medium-size ones with a
radius of 13 inches; and f u l l - s c a l e ones w i t h a radius of about 70 inches.
A l l were s t i f f e n e d by s t r i n g e r s and rings or hoops.

The small cylinders had precut s l i t s and were brought t o rupture by

increasing t h e pressure steadily. The crack-length r e s u l t s a r e shown i n
figure 5 . The curve, labeled b,gives t h e c r i t i c a l crack length of
unstiffened cylinders having the same radius. The c i r c l e s denote again
confined ruptures; t h e x-marks, unconfined ruptures. Regardless of t h e
type of rupture, a l l t h e c r i t i c a l crack lengths p l o t f a i r l y close t o t h e
reference curve, which means that the c r i t i c a l crack lengths of t h e stiff-
ened cylinders i n t h i s group of t e s t s a r e equal t o those of corresponding
unstiffened cylinders.

Figure 6 shows t h e ring-size c r i t e r i o n p l o t f o r t h e same.group of

t e s t s . Each t e s t point represents t h e same specimen as t h e point a t t h e
same s t r e s s l e v e l i n figure 5 . The empirical c r i t e r i o n l i n e i s taken
from figure 4. A l l ruptures observed a r e i n agreement with t h e c r i t e r i o n :
unconfined i f above t h e curve, confined if below t h e curve. The cylinders
i n t h i s t e s t group had e i t h e r r a t h e r l i g h t rings o r e l s e r a t h e r heavy
rings; as a r e s u l t t h i s test group does not give a s e n s i t i v e check on t h e

i 441
 5

accuracy of t h e ring-size c r i t e r i o n . On t h e other hand, t h e f a c t that

\
r a t h e r extreme r i s i z e s were used tends t o increase t h e weight of t h e
evidence regarding c ck length shown i n f i g u r e 5 .

The next group of t e s t s i s on medium-size cylinders (15-inch radius)

of 2024-T3 material. On these cylinders, t h e i n t e r n a l pressure w a s cycled
t o grow a f a t i g u e crack, s t a r t i n g from an i n i t i a l s l i t . Figure 7 shows
the crack-length p l o t . The c r i t i c a l crack lengths f o r t h i s group of t e s t s
a r e consistently longer than indicated by t h e reference curve, t h e excess
length varying roughly from 40 percent t o 100 percent. The result thus
d i f f e r s markedly from that obtained on t h e small cylinders.

Figure 8 i s t h e ring-size p l o t f o r t h e medium-size cylinders. The

types of f a i l u r e a r e i n agreement with t h e prediction, i f t h e prediction
f o r t h e border-line case on t h e l i n e i s made conservatively. This p l o t
shows only 3 points, and thus no counterparts a r e shown f o r many of
the points shown i n figure 7; t h e missing points a r e discussed i n t h e
following paragraph.

On t h e small cylinders, and on t h e t h r e e medium-size cylinders d i s -

cussed i n t h e previous paragraph, t h e rings were riveted continuously
t o the skin. On the other medium-size cylinders, t h e rings were e i t h e r
floating, that is, not touching t h e skin a t a l l , o r they were touching
the skin, but riveted t o it only a t t h e intersections with t h e s t r i n g e r s .
Rings of e i t h e r type are believed t o have l i t t l e i f any power t o confine
rupture; therefore, t h e assumption is made t h a t t h e presence of e i t h e r
type of r i n g j u s t i f i e s a prediction of unconfined rupture, and conse-
quently, it i s unnecessary t o p l o t the point on t h e ring-criterion p l o t
i n order t o a r r i v e a t a prediction of t h e type of rupture. So far, t h e r e
appears t o be no evidence that t h e assumption i s unduly conservative.

Figure 9 shows t h e crack-length r e s u l t s f o r medium-size cylinders

of 7075-6. A l l t h e points p l o t close t o t h e reference curve. Since
the corresponding p l o t f o r 2024-T3 material ( f i g . 7) showed excess crack
lengths ranging from 40 percent t o 100 percent, it may be s a i d t h a t
2024-T3 has a "hidden margin of safety," under some conditions, which
appears t o be lacking i n 7075-T6 material.

Figure 10 i s t h e ring-size p l o t f o r t h e 7075-6 cylinders. Only t h e

t h e o r e t i c a l c r i t e r i o n i s shown, since no empirical c r i t e r i o n is estab-
lished, as mentioned previously. The type of rupture i s as expected i n
a l l cases, but t h e location of t h e points i s such that t h e empirical c r i -
t e r i o n s t i l l remains undefined.

Figure 1 1 w i t h data taken from references 4 and 5 shows two t e s t s

made by two a i r c r a f t manufacturers on f u l l - s c a l e models of fuselages with
a radius of about 70 inches. The c r i t i c a l crack length i s j u s t above t h e
6

reference value f o r specimen A and about 60 percent l a r g e r f o r specimen B.

This s c a t t e r suggests a s c a t t e r band somewhat similar t o that obtained on
t h e medium-size cylinders. 8'

I n specimen A, a s a w s l i t had been made through t h e skin as w e l l as

through t h e r i n g underneath it. With one r i n g thus out of action, t h e
ring-reinforcement r a t i o w a s marginal, which would lead t o t h e prediction
that t h e rupture w i l l probably be unconfined. For specimen B, an uncon-
fined rupture would be predicted because t h e rings were f l o a t i n g . The
symbols indicate that both specimens had unconfined ruptures. It should
be noted that both specimens served as s t a r t i n g points i n t h e Gevelopment
of f i n a l designs.

CONCLUDING REMARKS

The crack-length c r i t e r i o n and t h e ring-size c r i t e r i o n appear t o

o f f e r some promise as t o o l s f o r putting some aspects of f a i l - s a f e design
on a quantitative basis. The crack-length c r i t e r i o n i n i t s present form
states t h a t t h e c r i t i c a l crack length of a s t i f f e n e d cylinder is a t least
equal t o that of an unstiffened cylinder of t h e same radius. The ring-
s i z e c r i t e r i o n , which is reasonably w e l l established f o r 2024-T3 aluminum
a l l o y but not f o r 7075-6 aluminum alloy, appears t o permit a prediction
whether the rupture w i l l be confined o r unconfined when the rings are
continuously r i v e t e d t o t h e skin. When the rings a r e not s o riveted ( o r
otherwise fastened), t h e rupture w i l l probably be unconfined.

Much work remains t o be done, however. The conditions under which

t h e c r i t i c a l crack length of t h e stiffened cylinder can be greater than
that of t h e unstiffened cylinder should be established more f u l l y . Only
s t r a y b i t s of information a r e available a t present on a number of factors,
such as e f f e c t of s t r i n g e r s , of load-carrying members bridging a crack,
and of producing t h e i n i t i a l damage very rapidly.
 -7
I

1. McEvily, Arthur J., Jr., Illg, Walter, and Hardrath, Herbert F.:
Static Strength of Aluminum-Alloy Specimens Containing Fatigue
Cracks. NACA TN 3816, 1956.

2. Peters, Roger W., and Kuhn, Paul: Bursting Strength of Unstiffened

Pressure Cylinders With Slits. NACA TN 3993, 1957.

3. Peters, Roger W., and Dow, Norris F.: Failure Characteristics of

Pressurized Stiffened Cylinders. NACA TN 3851, 1956.

4. Sorensen, Arne: Some Design Considerations for Tear-Resistant Air-

plane Structures. Preprint No. 618, S.M.F. Fund Preprint, Inst.
.
Aero. Sci , Jan. 1956.

5. Spaulding, E. H.: Observations on the Design of Fatigue-Resistant

and 'Fail Safe' Aircraft Structures. Session 8. Paper 2. Presented
at International Conference on Fatigue of Metals sponsored by British
Inst. Mech. Eng. and A.S.M.E., Sept. 10-14 (London) & Nov. 28-30
(New York), 1956.
8

TYPES OF RUPTURE

d8t

CONFINED RUPTURE UNCONFINED RUPTURE

Figure 1

STRENGTH OF CRACKED SPECIMENS

K,,CYL = Ku (I + 46#)
FOR 2024-T3 AND 7075-T6
ALUMINUM ALLOY

Figure 2
3F 9

CRITICAL CRACK LENGTHS

UNSTIFFENED CYLINDERS

HOOP
STRESS,
KSI

2o Lt
\I
r45
I I I I I
2024 -T 3
7075-T 6

2024 -T3
L7075-T6
I

Figure 3

RING-SIZE CRITERION FOR 2 0 2 4 - T 3

AND 24 IN.; PRESSURE AND TORSION
r=15

EMPIRICAL CRITERION

HOOP 2 0 1
STRESS,
t /d
o<
’
0

0
KSI THEORETICAL CRITERION

0 CONFINED RUPTURE
X UNCONFINED RUPTURE

0 -2 .4 .6
RING-REINFORCEMENTRATIO, AR/ltS

Figure 4
426
10

CRACK LENGTHS FOR 2024-T3

r=3.6 IN.; PRESSURE RISING

RUPTURE
0 CONFINED
x UNCONFINED
4

0 I 2
8, INCHES

Figure 5

RING-SIZE CRITERION FOR 2024-T3

r = 3.6 IN.; PRESSURE RISING
30 -
X
RUPTURE
0 CONFINED
X UNCONFINED

0 0

CRITERION

I I I
0 .2 .4 .6
AR/2k

Figure 6
I
 CRACK LENGTHS FOR 2024-T3
r=15IN.; PRESSURE CYCLING
30-
RUPTURE
0 CONFINED
X UNCONFINED

-
20.

UNSTIFFENED
HOOP
STRESS, .
' CYLINDER
KSI

I O .- X
4

tI I I I I I
0 2 4 6 8 IO
8, IN.(INCLUDING INITIAL DAMAGE)

Figure 7

RING-SIZE CRITERION FOR 2024-T3

r=15IN.; PRESSURE CYCLING

/
/
- /
/ I I I

Figure 8
 CRACK LENGTHS FOR 7075-T6
r = 15 IN.; PRESSURE CYCLING
301-

\
RUPTURE
0 CONFINED
X UNCONFINED

20-

HOOP
STRESS, -
KSI

IO -
UNSTIFFENED
CYLINDER

0 2 4 6 8 10
6, IN. (INCLUDING INITIAL: DAMAGE)

Figure 9

RING-SIZE CRITERION FOR 7075 -T6

r = 15 IN.; PRESSURE CYCLING
/
/
/

20
"/
STRESS,
KSI
t xx

/
/ "
/

I I , I
0 .2 .4 .6
AR/%

449 Figure 10
 CRACK LENGTHS FOR 2024-T3
r=70IN.; PRESSURE CONSTANT
30-

20 -
HOOP
STRESS,
KSI
- UNSTIFFENED
CYLINDER

0 4 8 12 16 20
8, INCHES

Figure 11
 By John B. Gamin

Langley Aeronautical Laboratory

The importance of t h e "fail-safe" design concept with regard t o

fatigue i n modern airplane structures i s evidenced by the great a t t e n t i o n
which has recently been directed t o t h i s subject. Involved i n t h i s con-
cept are a need t o know something of t h e behavior of a fatigue crack as
it s t a r t s and grows i n an airplane component and, i n addition, a need
f o r howledge of a b i l i t y t o estimate t h e fatigue l i f e of t h e a i r c r a f t
structure.

This paper presents r e s u l t s from an experimental investigation of

full-scale-airplane fatigue. Crack behavior and fatigue l i f e were studied
during the application of random o r variable-amplitude loads t o t h e wings
of a C-46 transport-type airplane. These w i n g s were the same types of
structures which were used i n constant-amplitude fatigue tests reported
i n references 1 and 2. The fatigue t e s t s of these wings provide t h e
f i r s t opportunity t o compare r e s u l t s from f u l l - s c a l e constant-amplitude
loadings with those simulating a c t u a l f l i g h t .

The loading chosen f o r the variable-amplitude t e s t s was the gust

spectrum reported by Rhode and Donely f o r a wide sampling of transport
operations i n t h e United States. (See ref. 3.) The gust spectrum w a s
selected because t h e loads due t o f l i g h t maneuvers and ground operations
were not considered t o play a s i g n i f i c a n t p a r t i n t h e fatigue l i f e of
t h i s airplane.

Figure 1presents the incremental loading spectrum f o r the C-46.

The ranges of incremental load f a c t o r f o r each loading l e v e l are shown
plotted against the number of loads applied a t each level. Loads were
applied a t t h e mean of each of t h e 16 load ranges designated by t h e
c i r c l e d points. Thus, t h e lowest value of incremental load w a s 0.225g,
representing the range from the threshold value of O.l5g t o 0.30g. The
numbers of cycles were established on t h e b a s i s of a representative
f l i g h t distance of 10 million m i l e s . The maximum loading associated
with a f l i g h t operation of t h i s duration would., as i s indicated by t h e
position of t h e dashed l i n e i n t h e figure, be somewhat greater than l i m i t
load f o r t h i s airplane.

In order t o apply t h e loadings i n such a way as t o avoid putting on

a l l one s i z e load p r i o r t o any other, t h e spectrum of figure l w a s divided
by 100 which would correspond t o a distance of 100,000 miles or about
500 flying hours under t h e assumed t y p i c a l conditions of f l i g h t . The
loads which appeared less than 100 times i n t h e o r i g i n a l spectrum would
appear only occasionally i n t h e reduced spectra.

For t h e purpose of introducing randomness i n t o t h i s method of

loading, the a c t u a l order of application of the loading s t e p s w a s estab-
lished by reference t o a table of random numbers. Such an arrangement has
I1
been called a "Sequence. The random arrangements of t h e loadings i n the
reduced spectra f o r t h e first three sequences a r e i l l u s t r a t e d i n t a b l e I.
The f i r s t incremental loading i n the first sequence i s the 0.525g load
applied 3,510 times; t h i s i s followed by the 0.675g load applied 4 0 6 7
times. The load a t 2.475g and a l l other loads i n t h i s sequence which
appear with a zero value f o r the number of cycles applied are those
loads which occur l e s s than 100 times i n the o r i g i n a l LO-million-mile
spectrum and thus only appear occasionally i n t h e 100 sequences. It
can be readily seen t h a t t h e orders of application of t h e loads i n t h e
3 sequences shown are different, and it m a y be of i n t e r e s t t o note t h a t
none of the 100 sequences were i d e n t i c a l with regard t o order of load
application t o t h e specimen. Loading sequences greater than 100 repeated
the o r i g i n a l order; thus, t h e 101 sequence was t h e same as the first 1

sequence.
L
The f a t i g u e machine used f o r t h i s investigation i s seen i n f i g u r e 2.
This machine w a s b a s i c a l l y a concentrated-eccentric-mass type of shaker
capable of being operated at any frequency up t o 4 cycles per second.
The wing attached t o a c e n t r a l portion of t h e a s e l a g e can be seen inverted
and mounted between two large supporting s t r u c t u r e s near t h e center of
t h e picture. The eccentric weights can be seen under t h e wing t i p ; t h e
gearbox and drive shafting are under t h e wing; and t h e necessasy controls
and cycle-measuring equipment were located c e n t r a l l y near the main support
structures. Loads were applied dynamically up to incremental values of
t1 g . Those loads above t h i s l e v e l were applied s t a t i c a l l y by means of
t h e hydraulic ram located d i r e c t l y over the eccentric weights. A l l loads
applied during t h e t e s t s were monitored by e i t h e r i n t e r n a l s t r u c t u r a l
s t r a i n gages or external load-measuring dynamameters which are located
on top of t h e loading rams.

This program has, up t o t h e present time, performed variable-amplitude

t e s t s on three complete wings, each camposed of one center section and two
outer panels. A t o t a l of 46 cracks have been observed, of which one crack
on each of t h e s i x outer panels grew t o such s i z e as t o indicate imminent
f a i l u r e of the wing. Actual f i n a l lifetimes accumulated by these wings
ranged between la and 219 sequences of load corresponding t o assumed
f l i g h t times of from 62,000 t o 114,000 f l i g h t hours.

Three subjects appear of i n t e r e s t from t h e results of these v a r i a b l e

amplitude t e s t s : the area of fatigue crack i n i t i a t i o n , t h e r a t e of growth
o r propagation of t h e crac ion of f a t i g u e damage i n t h e e.
s t r u c t u r e during variable
r
4s2
 I

With regard t o crack i n i t i a t i o n , it was noted t h a t fatigue cracks

which grew t o f i n a l f a i l u r e i n t h e variable-amplitude t e s t s i n i t i a t e d
i n areas where cracks i n i t i a t e d only during t h e high-level constant-
amplitude loadings. I n other words, the serious cracks found under t h e
random t e s t conditions did not occur a t t h e loadings which might be
considered representative of that for g r e a t e s t damage i n the constant-
l e v e l tests. Therefore, a constant-level fatigue t e s t at t h e l e v e l of
greatest fatigue damage might not reveal the proper area of crack i n i t i -
ation which would be of i n t e r e s t t o operational inspection crews.

Crack-growth information i s shown i n figure 3 . The abscissa scale

i s the percentage of f i n a l l i f e f o r the specimen. This scale was chosen
f o r convenience i n displaying the r e s u l t s f o r specimens which were sub-
jected t o d i f f e r e n t values of constant-amplitude loading as w e l l as those
subjected t o random loadings. This w i l l be c l e a r when it i s r e a l i z e d
that the specimen t e s t e d a t a constant alternating load l e v e l of f l g
sustained only 100,000 cycles of load, while the variable-amplitude ’
specimens accumulated about 10,000,000 cycles of load before f i n a l f a i l -
ure. The dashed curves show crack growth f o r two representative constant-
amplitude specimens t e s t e d a t a low and a high value of a l t e r n a t i n g load,
and the s o l i d curves show r e s u l t s f o r two representative variable=
amplitude specimens. All curves appear t o exhibit a c h a r a c t e r i s t i c shape
denoting slow growth of t h e crack from i n i t i a t i o n a t about one-third
t o one-half of the f i n a l l i f e u n t i l about 95 percent of t h e l i f e had
been spent, a f t e r which there w a s a very rapid growth of t h e crack u n t i l
the e n t i r e specimen f a i l e d . This rapid growth i n i t i a t e d a t a point where
t h e crack had approached a value between 5 and 10 percent of t h e t o t a l
tension material. The appearance of the s t e p shown i n the two upper
curves at about 70 percent of the l i f e i s associated with t h e f a i l u r e
of a f a i r l y heavy element of the structure. This f a i l u r e c h a r a c t e r i s t i c
appears t o be associated w i t h a p a r t i c u l a r configuration and might not
be generally applicable t o other structures. It would seem, however,
from these r e s u l t s , t h a t crack growth under variable-amplitude loading
conditions is, i n general, l i k e crack growth under t h e constant-amplitude
loadings.

An assessment of t h e amount of fatigue damage which an airplane wing

might accumulate during the variable-amplitude loadings of f l i g h t may be
determined i n several ways. The cumulative-damage c r i t e r i o n , usually
a t t r i b u t e d t o Miner, has been selected f o r a f i r s t comparison. (See
r e f . 4.) Since t h i s damage c r i t e r i o n i s based on an endurance curve
from constant l e v e l t e s t i n g and since such a curve was available from
previous t e s t s on an i d e n t i c a l structure, having the same load d i s t r i b u -
t i o n and corresponding load ranges, it w a s f e l t t h a t t h i s damage c r i t e r i o n
might give an excellent prediction of l i f e f o r both crack i n i t i a t i o n and
f o r f i n a l f a i l u r e of t h e w i n g .
4

Figures 4 and 5 show t h e agreement obtained f o r t h e six outer wing

panels t e s t e d under variable-amplitude loadings. A bar graph i s used
t o show t h e r e l a t i v e l i f e of these panels. The standard of camparison
i s t h e l i f e calculated from t h e r e s u l t s of the constant-level tests.
Figure 4 shows t h e r e l a t i v e l i f e u n t i l i n i t i a t i o n of the crack which grew
t o f a i l each specimen, and f i g u r e 5 shows t h e r e l a t i v e l i f e u n t i l f i n a l
f a i l u r e . The calculated l i f e f o r crack i n i t i a t i o n w a s 17.4 sequences
o r about 8,700 f l i g h t hours while t h e calculated l i f e t o f i n a l f a i l u r e
w a s 37 sequences o r about 18,500 hours. It can be seen i n figure 4 t h a t
t h e average a c t u a l l i f e t o crack i n i t i a t i o n i s about 35 times t h e pre-
dicted l i f e w i t h a spread of about 5 t o 1. Similar comparisons f o r a l l
t h e 46 cracks which i n i t i a t e d i n these wings would show that t h e average
l i f e would have been about 4& times the predicted l i f e , while t h e o v e r a l l
spread would be about 33 t o 1. The e a r l i e s t crack which occurred i n any
of t h e wings was found t o occur at about one-third of t h e l i f e predicted
by t h e cumulative-damage c r i t e r i o n .

Figure 5 shows t h a t t h e actual f i n a l l i f e of these specimens was

about four times t h a t predicted by t h i s damage c r i t e r i o n and t h a t the
spread i n t h e f i n a l failure values i s slightly less than 2 t o 1. It
would appear from t h i s that an estimation of a f i n i t e l i f e e i t h e r to
crack i n i t i a t i o n o r f i n a l f a i l u r e by t h e cumulative-damage method based
on constant-level fatigue-test r e s u l t s would be uncertain.

Another method of analyzing fatigue l i f e w a s performed on these d a t a

f o r camparison purposes. This method, h o w n as t h e "intersect" method,
makes use of t h e summation curve f o r t h e spectrum loading i n condunction
w i t h t h e appropriate endurance curves from the constant-level tests.

These r e s u l t s f o r t h e same specimens a r e shown i n figures 6 and 7.

I n figure 6, the difference between predicted l i f e and measured l i f e has
been reduced t o a value of about 2 while t h e spread, as would be expected,
remains t h e same as f o r t h e cumulative-damage method. I n the case f o r
f i n a l f a i l u r e shown i n figure 7, differences of about 2 t o 1 are seen as
were noted i n t h e case f o r crack i n i t i a t i o n by t h i s method; here again
the spread i s t h e same as f o r t h e cumulative-damage method. Although
t h e i n t e r s e c t method indicates closer values of l i f e prediction than t h e
cumulative-damage method., t h e r e appears t o be l i t t l e advantage of one
method over t h e other because of t h e influence of spread i n t h e data,
p a r t i c u l a r l y i n the crack-initiation case.

The significance of any p a r t i c u l a r value of estimated l i f e by either

method. would appear t o be questionable; however, t h i s would be the case
i n any attempt t o use the r e s u l t s from one o r several constant-level
fatigue tests t o p r e d i c t t h e endurance of a similar component subjected
t o t h e variable loads of a c t u a l f l i g h t .
 I

The apparent increase i n l i f e of t h e c-46 wings as indicated by

the absolute values of both methods might be explained by t h e application
of the higher loads of t h e spectrum. It is noted t h a t these l a r g e r loads
were of such amagnitude as t o cause a considerable increase i n Life as
indicated by r e s u l t s from constant-level t e s t s on Meteor 4 t a i l p l a n e s
reported by Raithby and Longson i n reference.5. Whether o r not t h i s
e f f e c t fully accounts f o r the increase i n l i f e noted here i s not known.

I n summary, it appears t h a t a constant-amplitude fatigue t e s t might

not reveal t h e area wherein a crack would i n i t i a t e t o cause failure of
a s t r u c t u r e subjected t o t h e variable loadings of f l i g h t ; crack propa-
gation behavior based on the f i n a l l i f e of t h e specimen appears t o be
s i m i l a r under both constant-- and variable-amplitude loadings j and, f i n a l l y ,
the assessment of f a t i g u e damage by e i t h e r t h e cumulative damage method
or the i n t e r s e c t method appears questionable when based only on constant-
l e v e l f atigue-test r e s u l t s .

REFERENCES

1. McGuigan, M. J., Jr., Bryan, D. F., and Whaley, R. E.: Fatigue Inves-
t i g a t i o n of Full-scale Transport-Airplane Wings -
Summary of
Constant-Amplitude Tests Through 1953. NACA TN 3190, 19%.

2. Mhaley, Richard E., McGuigan, M. J., Jr., and Bryan, D. F.: Fatigue-
Crack-Propagation and Residual-Static-Strength Results on Full-scale
Transport-Airplane Wings. NACA TN 3847, 1956.

3. mode, Richard V., and Donely, Phillip: Frequency of Occurrence of

Atmospheric Gusts and of Related Loads on Airplane Structures.
NACA WR L-121, 1944. (Formerly NACA ARR LkI21. )

4. Miner, Milton A.: Cumulative Damage i n Fatigue. Jour. Appl. Mech.,

vol. 12, no. 3, Sept. 1945, pp. A-159 - A-164.

5. Raithby, K. D., and Longson, Jennifer: Some Fatigue Characteristics

of a Two Spar Light Alloy Structure (Meteor 4 Tailplane). Rep.
No. Structures 195, B r i t i s h R.A.E., Jan. 1956.
5

TABLE I. - ORDER OF LOADING

F i r s t sequence Second sequence Third sequence

Load, An Sycles Load, An Cycles bad, Lh Cycles

0.525 3,510 1.725 0 0.225 39,3=

2 475 0 2.175 0 1.125 21
1.875 0 .825 236 1.425 1
2,325 0 1.425 1 2.029 0
675 1,067 1.275 5 2.475 0
2.025 0 1.575 1 975 73
975 73 1.125 20 1.875 1
1.575 1 2 475 0 525
9 3,510
1.725 0 375 15,444 1.575 1
1.275 6 .225 39,312 2.325 0
2.175 0 675 1,067 .825 235
1.125 20 525 3,510 675 1,067
1.425 1 975 73 1.725 0
.825 235 2.025 0 1.275 5
,225 59,312 2.325 0 2.175 0
375 15,444 1.875 1 375 15,444
 ‘7
I

C-46 LOAD SPECTRUM

1O7MlLES OF FLIGHT

2.7
2.4
LIMIT LOAD
2. I __------
I .a
INCREMENTAL
LOAD FACTOR, LOAD VALUES
All 1.2
.9
.6
.3
THRESI+OLD
0
100 101 102 103 104 105 106 to7
CYCLES

Figure 1

C- 46 VARIABLE-AMPLITUDE FATIGUE MACHINE

Figure 2
L-57-479.1
457
8

FATIGUE -CRACK PROPAGATION

VARIABLE AMPLITUDE
40 - CONSTANT AMPLITUDE
------ An = 0.35
--- An = 1.00
TENSION 30
AREA
FAILED,
% 20

IO

0 20 40 60 80 IO0
LIFE TO FINAL FAILURE, %

Figure 3

DAMAGE EVALUATION
CUMULATIVE DAMAGE METHOD ; CRACK INITIATION

I
I
5 I
I
~I
SPECIMEN
NUMBER
3 v-
I
AVERAGE

0 I 2 3 4 5 6 7

4
A&. Figure 4
 s

DAMAGE EVALUATION
CUMULATIVE DAMAGE METHOD ; FINAL FAILURE

4 AVERAGE LIFE
SPECIMEN
NUMBER

I
I f I I I f I I I I l l f I 1 1
0 I 2 3 4 5 6 7
ACTUAL
RELATIVE LIFE,
CALCULATED

Figure 5

DAMAGE EVALUATION
INTERSECT METHOD j CRACK INITIATION

5 I

SPECIMEN 4
NUMBER
PAVERAGE
I
I

Figure 6 \-.
 I

DAMAGE EVALUATION
INTERSECT METHOD ; FINAL FAILURE

AVERAGE LIFE

ACTUAL
RELATIVE LIFE,
CALCULATED

Figure 7
 STUDIES OF STRUCTURAL FAILURE

ACOUSTIC LOADING
- -
By Robert W. Hess, Robert W. Fralich,
and Harvey H. Hubbard

Langley Aeronautical Laboratory

Some discussion of t h e acoustic f a t i g u e problem of a i r c r a f t s t r u c -

t u r e s i s given along with data pertaining t o t h e acoustic inputs from
some powerplants i n common use. Comparisons a r e given f o r r e s u l t s of
some f a t i g u e tests of f l a t panels and cantilever beams exposed t o both
random- and discrete-type inputs. I n t h i s regard it appears that both
t h e s t r e s s l e v e l of t h e t e s t and t h e type of model a r e significant;
hence, no generalization can be m d e a t t h i s time. With regard t o
increasing t h e fatigue l i f e , it w a s noted that increased s t i f f e n i n g of
a panel due t o curvature and pressure d i f f e r e n t i a l i s p a r t i c u l a r l y
beneficial.

INTRODUCTION

It i s well-known that f a t i g u e damage can occur t o a i r c r a f t s t r u c -

t u r e s which a r e exposed t o intense acoustic pressure loads. Damage
usually occurs i n t h e secondary s t r u c t u r e of t h e a i r c r a f t as a result
of a large number of r e l a t i v e l y small loads applied a t the r a t e of sev-
eral hundred loading cycles per second. This paper presents information
pertaining p a r t i c u l a r l y t o t h e problem of exposure t o random noise such
as that encountered from turbojets, ram jets, rocket engines, and aero-
dynamic boundary layers.

Some of t h e phenomena involved i n t h i s problem can be discussed with

t h e a i d of t h e block diagram of t a b l e I. L e t us first d i r e c t our a t t e n -
t i o n t o t h e blocks themselves. The acoustic inputs are i n t h e form of
f l u c t u a t i n g pressures on t h e exposed surface of t h e s t r u c t u r e . They
impose loads that tend t o vibrate t h e surface. Depending on i t s s t r u c -
t u r a l c h a r a c t e r i s t i c s , such as geometry and method of construction, t h e
surface w i l l have a c e r t a i n dynamic response. This dynamic response
influences t h e s t r e s s patterns i n t h e s t r u c t u r e which, i n turn, deter-
mine t h e f a t i g u e l i f e .
 Analyses have been published (refs. 1 and 2) wherein the phenomena
denoted as A i n t a b l e I were made use of t o calculate stresses on a
simple panel. A knowledge of t h e noise spectra a t an arbitrary point
and t h e dynamic response t o a uniform load have made it possible t o c a l -
culate, by s p e c t r a l techniques, t h e s t r e s s e s a t an arbitrary point on
t h e panel. For more complex s t r u c t u r e s o r inputs o r both, t h i s approach
may not be s u f f i c i e n t and it may thus be necessary t o use the q u a n t i t i e s
denoted as B i n table I f o r t h e stress analyses. For instance, a
knowledge of t h e correlation functions of the acoustic input and t h e
complex dynamic influence coefficients of a panel surface might make it
possible t o calculate t h e s t r e s s d i s t r i b u t i o n over t h e surface.

Thus, it i s seen that t h e blocks i n t a b l e I each represent rather

complex phenomena. Detailed analyses of a l l p a r t s of t h i s problem would
involve considerable e f f o r t and a r e beyond t h e scope of t h i s paper. The
rest of t h i s paper w i l l d e a l mainly w i t h the f a t i g u e l i f e of s t r u c t u r e s
exposed t o noise and, i n p a r t i c u l a r , w i l l compare results of random and
d i s c r e t e frequency fatigue t e s t i n g . The other p a r t s of t h e problem are
discussed only b r i e f l y .

SYMBOLS

nozzle diameter

a x i a l distance measured from nozzle e x i t plane

power s p e c t r a l density of noise input

mechanical impedance of panel

power spectrum of stress i n panel

ACOUSTIC INPUTS

The ingredient of t h i s problem which d i f f e r e n t i a t e s it from other

f a t i g u e problems i s t h e nature of t h e input function. This can be
described b r i e f l y w i t h t h e aid of figure 1. Shown here are the acoustic
inputs f o r various powerplants i n common use. The data shown are t h e
noise pressure loads on a surface p a r a l l e l t o t h e t h r u s t a x i s of t h e
engine and four e x i t diameters d i s t a n t from it. (Free space measure-
ments were increased by 3 decibels t o adjust them t o conditions a t t h e
surface of t h e w a l l . ) Both the noise pressure l e v e l s i n decibels and
t h e equivalent noise pressures i n pounds per square foot are given on
t h e v e r t i c a l scale as a function of distance f r o m t h e e x i t nozzle of

t
t h e engine. Surface pressure data a r e given f o r two t u r b o j e t engines
and one rocket engine. It should be noted t h a t t h e numbers associated
w i t h t h e coded legend are t h e pounds of t h r u s t developed by t h e engine
per square foot of nozzle e x i t area. The short s o l i d t u r b o j e t curve
applies t o t h e J34-WE-22 t u r b o j e t engine whereas t h e curve of long
dashes applies t o t h e 557-P-3 engine f o r which free-space data are
given i n reference 3 . These two curves i l l u s t r a t e t h e growing s e v e r i t y
of t h e problem as the engine performance increases because of t h e
increased engine pressure r a t i o . The curve of long-short dashes applies
t o a World War I1 rocket engine. It i s seen that these pressures are of
t h e order of 10 decibels higher than those f o r t h e 557 t u r b o j e t engine.
For rocket engines operating a t higher pressure r a t i o s , there i s some
evidence t h a t t h e acoustic pressure would a l s o tend t o be higher. A l l
these engines generate intense noise i n t h e range of frequencies that
i s detrimental t o a i r c r a f t s t r u c t u r e s . For t h e type of random spectra
generated by these engines, f a t i g u e damage can occur at o v e r a l l l e v e l s
of t h e order of 140 decibels or higher. The amount of damage incurred
a t any given level, of course, i s a function of t h e (1)d e t a i l design
of t h e s t r u c t u r e , ( 2 ) t h e length of exposure t o t h e noise, and, a l s o ,
( 3 ) t h e spectrum of t h e noise.

DYNAMTC RESPONSE

The importance of t h e noise spectrum with r e l a t i o n t o t h e dynamic

response of a given s t r u c t u r e i s i l l u s t r a t e d by t h e diagrams of f i g u r e 2.
Conditions a r e defined f o r t e s t s of f l a t panels exposed t o d i s c r e t e f r e -
quency noise from a s i r e n and t o random noise from a j e t .

The input spectra of f i g u r e 2 are r e l a t e d t o t h e s t r e s s spectra by

t h e panel admittance. It can be seen that the noise from t h e s i r e n w a s
contained l a r g e l y i n t h e fundamental frequency, t h e harmonics being f e w
i n number and r e l a t i v e l y weak. For these t e s t s the s i r e n was tuned
u n t i l t h e fundamental frequency coincided w i t h t h e f i r s t n a t u r a l mode
of t h e panel. As a result nearly a l l t h e energy accepted by t h e panel
i s i n i t s f i r s t mode. A panel being excited by t h e broad-band spectrum
of the j e t , on t h e other hand, responds i n some degree t o a l l frequencies.

It can be seen from an inspection of t h e f i g u r e that a panel w i l l

accept e s s e n t i a l l y a l l t h e energy available from t h e s i r e n whereas a
l a r g e p a r t of t h e energy available from t h e j e t i s not accepted. For
these p a r t i c u l a r tests, the stress developed by a discrete-type e x c i t a -
t i o n would be expected t o be greater a t a given o v e r a l l noise pressure.
4

FATIGUE LIFE

Effects of Overall Noise Level .

Some f a t i g u e r e s u l t s obtained f o r panels exposed t o both of these
types of input are given i n figure 3 . Fatigue l i f e as a function of
t h e o v e r a l l noise l e v e l i s shown f o r an O.032-inch gage f l a t panel
11 inches by 1.3 inches mounted over a rectangular cutout i n a r i g i d
frame. The panel w a s attached by small round head b o l t s tightened t o
a predetermined torque. This configuration w a s chosen f o r t h e reason
that it f a c i l i t a t e d assembly and disassembly of models while stress con-
centrations similar t o those i n a riveted s t r u c t u r e were retained. For
the s o l i d points which represent f a t i g u e data obtained w i t h t h e siren,
t h e curve has been sketched i n through t h e available points t o indicate
a general trend of t h e data. Fatigue l i f e i s very strongly dependent on
t h e l e v e l of noise excitation, since it v a r i e s from under a minute t o
several hours i n t h e noise-level range of t h e t e s t s .
Attention is called p a r t i c u l a r l y t o the open points which are data
obtained w i t h t h e j e t . These data f a l l generally t o t h e r i g h t of the
curve i n f i g u r e 3; thus, a longer fatigue l i f e i s indicated. This d i f -
ference i n f a t i g u e l i f e i s due i n p a r t t o the f a c t that t h e panel has
a higher root-mean-square s t r e s s l e v e l when excited by the s i r e n f o r a
given overall noise l e v e l than when excited by t h e a i r j e t .

Effects of Method of Mounting

During t h e d i s c r e t e frequency t e s t s w i t h these simple panels, t h e

opportunity w a s taken t o change t h e manner of mounting t o evaluate pos-
s i b l e e f f e c t s on f a t i g u e l i f e . These mounting configurations are shown
schematically i n f i g u r e 4 along with some of the t e s t results i n bar
graph form. For a l l t h e mountings, t h e gage and s i z e of panel were con-
s t a n t and t h e input noise l e v e l s were a l s o constant. The basic configu-
r a t i o n A i s t h e same as that f o r which d a t a w e r e presented i n f i g u r e 3.
Failures i n t h e skin panel occurred first near t h e b o l t heads and t h e
average fatigue l i f e $or t h i s configuration i s used as a reference i n
t h e figure.

Configuration B i s t h e same as configuration A except that a l a y e r

of bonding material is placed between t h e panel and t h e r i g i d frame.
During testing, t h e panel first peeled away from the bonding and then
f a i l u r e i n t h e skin occurred near t h e b o l t heads. This configuration
l a s t e d on the average about 50 percent longer than configuration A.
An attempt w a s made t o eliminate peeling by bonding both s i d e s and
clamping t h e panel between two r i g i d surfaces as i n configuration C.
I n t h i s case failure occurred a t t h e edge of t h e frame and t h e average-
model l a s t e d twice as lon

4e:.4
 . . . ..
0 , .& b
..m
0 .
. I
. -.
0
I
)
I.

I n order t o study t h e e f f e c t s of curvature some panel models were

r o l l e d t o an 8-foot radius and were fastened f o r t e s t i n g t o a curved
r i g i d frame as indicated i n configuration D. F a i l u r e s were i n i t i a t e d
near t h e b o l t heads as i n configurations A and B. This condition r e s u l t e d
i n an average l i f e about 15 times as long as f o r configuration A even
though t h e frequency of t h e stress cycles increased by about 50 percept.
A f u r t h e r increase of f a t i g u e l i f e was obtained by leaving t h e edges of
t h e frame sharp instead of rounding as i n t h e figure. T h i s l a t t e r scheme
caused t h e failures t o occur a t t h e edge of the frame instead of a t t h e
b o l t head; by so doing t h e f a t i g u e l i f e was doubled as indicated by t h e
bar of dashed l i n e s . Configuration E i s t h e same as configuration D
except that t e s t s were made w i t h a pressure d i f f e r e n t i a l of 6 pounds per
square inch across t h e panel. This high i n t e r n a l pressure caused t h e f i r s t
panel frequency t o nearly treble and the f a t i g u e l i f e w a s g r e a t l y increased
as shown, i n s p i t e of t h e much faster rate of application of stress cycles.

A s a matter of i n t e r e s t a 0.064-inch-gage panel was t e s t e d i n a con-

f i g u r a t i o n similar t o configuration A f o r comparison. It w a s found that
doubling t h e gage thickness of t h e panel increased i t s f a t i g u e l i f e t o
about twenty times that of configuration A. This finding w a s confirmed
i n both the j e t and s i r e n t e s t s .

A Limited number of other tests have been made on l a r g e r and more

complex panels. I n a l l cases failures came first i n t h e s t i f f e n e r ele-
ments; thus, t h e importance of d e t a i l design of t h e panel supporting
s t r u c t u r e is emphasized. It w a s a l s o noted that crack growth w a s mark-
edly more rapid i n bonded s t r u c t u r e s than i n r i v e t e d s t r u c t u r e s .

Comparison of Random and Discrete Frequency Tests

F l a t panels.- The rest of t h e paper w i l l d e a l w i t h comparisons

of f a t i g u e l i f e under d i s c r e t e and random loading a t t h e same root-mean-
square (RMS) s t r e s s l e v e l s . The r e s u l t s of f l a t - p a n e l tests are given
i n figure 5. Time t o failure i n hours is shown f o r various root-mean-
square stress l e v e l s f o r an 0.032-inch-gage panel. The s o l i d points
were obtained by means of d i s c r e t e frequency e x c i t a t i o n from a siren
whereas t h e open points were obtained with random e x c i t a t i o n from an
a i r j e t . The curve is a least-square curve through t h e s o l i d points.
It can be seen that a t a given root-mean-square stress l e v e l , failure
occurs i n a shorter time when t h e panels are excited by t h e random j e t
noise. The f a c t that t h e random noise of t h e j e t i s more d e s t r u c t i v e
may r e s u l t f r o m t h e f a c t that some o f t h e peak stress responses are sev-
eral times as great f o r a given root-mean-square v a l u e ' a s they are f o r
t h e constant-level s i r e n tests. This phenomenon i s p a r t i c u l a r l y notice-
able a t t h e lower stress l e v e l s where the differences i n f a t i g u e l i f e
tend t o be t h e g r e a t e s t and t h e panel damping i s r e l a t i v e l y low. A t
6
a # "<$

higher stress l e v e l s t h e panel damping i s g r e a t e r and t h e differences

i n f a t i g u e l i f e tend t o be smaller.

Cantilever beams.- Similar fatigue t e s t s have been made f o r notched

cantilever beams a t various stress levels f o r both random- and d i s c r e t e -
type inputs. A schematic diagram of t h e model along with some of t h e
t e s t r e s u l t s are i l l u s t r a t e d i n f i g u r e 6.

Here again t h e time t o failure i n hours is shown a t various root-

mean-square stress l e v e l s f o r t h e beams t e s t e d i n bending. The speci-
mens were 3 inches long, 1 inch wide, and 1/4 inch deep with 3/16-inch
notches located 1/2 inch from t h e root. The open points were obtained
by means of an amplified tape recording of j e t noise f e d i n t o a shaker
attached t o t h e t i p . The s o l i d points were obtained by applying a sinus-
o i d a l load a t t h e t i p i n a Sonntag bending f a t i g u e machine. There is a
tendency i n these t e s t s a l s o f o r the random load t o be r e l a t i v e l y less
destructive a t t h e higher s t r e s s l e v e l s and more destructive a t t h e
lower s t r e s s l e v e l s than the sinusoidal load. For t h e data of figures 5
and 6 t h e strain-gage locations were a r b i t r a r y and hence t h e root-mean-
square s t r e s s e s shown i n figures 5 and 6 a r e not necessarily comparable.

A t h e o r e t i c a l prediction of t h e time t o failure f o r t h e random

loading is given by t h e s o l i d curve. This theory i s e s s e n t i a l l y one
given by Miles ( r e f . 1) and i s based on Miner's rule of l i n e a r accumur
l a t i o n of damage. It can be seen that t h i s t h e o r e t i c a l curve f i t s t h e
data f a i r l y well a t low stress l e v e l s but i s very conservative a t t h e
higher stress l e v e l s .

CONCLUDING REMARKS

The problem of acoustic f a t i g u e of a i r c r a f t structures has been

discussed with p a r t i c u l a r emphasis on a comparison of t h e fatigue l i f e
due t o d i s c r e t e - and random-type loadings. I n t h i s regard it appears
that both t h e stress l e v e l of t h e t e s t and t h e type of model are s i g n i f -
icant; hence, no generalizations can be made a t t h i s time. With regard
t o increasing t h e f a t i g u e l i f e , it w a s noted that increased s t i f f e n i n g
of a panel due t o curvature and pressure d i f f e r e n t i a l i s p a r t i c u l a r l y
beneficial.

4
 7

REFEBENCES

1. Miles, John W.: On S t r u c t u r a l Fatigue Under Random Loading. Jour.

Aero. Sci., vol. 21, no. 11, Nov. 1954, pp. 753-762.

Hess, Robert W., and Hubbard, FIarvey H.:

2. Lassiter, L e s l i e W., An Exper-
imental Study of t h e Response of Simple Panels t o Intense Acoustic
Loading. Jour. Aero. Sci., vol. 24, no. 1, Jan. 1957, pp. 19-24, 80.

3 . Howes, Walton L., and Mull, Harold R.: Near Noise F i e l d of a Jet-
Engine Exhaust. I - Sound Pressures. NACA TN 3763, 1956.
 .
I.
-1

1..
..
8

TABLE I

ACOUSTIC FATIGUE PROBLEM

ACOUSTIC A. NOISE SPECTRUM AT ARBITRARY POINT

INPUTS B. CORRELATIONS

STRUCTURAL A. DYNAMIC RESPONSE TO UNIFORM LOAD

CHARACTER1STlCS B. DY NAM IC INFWENCE COE FF IC IE NTS (COMPLEX )

STRESS A. STRESS AT ARBITRARY POINT

8. STRESS DISTRIBUTION OVER SURFACE

-..
 9

ACOUSTIC INPUTS

AIRCRAFT SURFACE-’ T THRUST,

I c-x 41d LB/SQ FT
--- c -
TURBOJET
2,600
430 -- 3,800
ROCKET
,/’\,,-\-/
/ ---25,000
I60 - I 40
OVERALL
NOISE c
j NOISE
PRESSURE,
LEVEL,
i
/---

DECIBELS 150 ,
/ /
I
/
/

- \
I3
LB/SQ FT

f
I
4:-0
4
1
-4 0 4 8 12 16 20
AXIAL DISTANCE, x/d

Figure 1

INPUT AND RESPONSE CHARACTERISTICS

SlREN
SPECTRUM
INPUT @N~u)l 7 AIR JET

SPECTRUM @$d

FREQUENCY FREQUENCY
10

156 -Ym 0 8 o O o
GAGE
&*& \
:
:.e'
*...**'

148

OVERALL NOISE
LEVEL, DffiIBEW
-
140 -
- \

am ..'C
'-.,....
'x x" ? B B a O

.-h0
a&&l %"" o+
M

- 0 LABORATORY AIR JET

, .--SIREN

Figure 3

RELATIVE FATIGUE LIFE

PANEL GAGE AND SIZE CONSTANT
0.032"

----_---------
3___________130

DIFFERENTIAL
PRESSURE=6 P S I
RELATIVE TIME TO FAILURE
 11

FATIGUE OF FLAT PANELS

0 AIR JET

12,000
0 - SIREN

8,000
RMS STRESS,
PSI t

4,000

0 .01 .I I IO
TIME TO FAILURE, HOURS

Figure 5

FATIGUE OF NOTCHED CANTILEVER BEAMS

FUNDAMENTAL FREQUENCY, 119 CPS

0 JET NOSE LOAD

0 SINUSOIDAL LOAD
-THEORY
20,000
Q
RMS STRESS, 0
PSI
I0,OOO

LI I I I I
0 .01 .I I IO 100
TIME TO FAILURE, HOURS

Figure 6
FLUTTER AND
BUFFETING
 *e

'IIIE USE OF WIND TUNNELS TO PREDICT FLIGHT BUFFET LOADS

By Don D. Davis, Jr., and Wilbur B. Huston

Langley Aeronautic a1 Laboratory

Methods and techniques developed f o r the study of buffeting by use

of wind-tunnel models are described. Requirements on model damping and
frequency c h a r a c t e r i s t i c s , the model support system, and the e f f e c t s of
wind-tunnel turbulence on buffet tests a r e outlined. The results a r e
sham of a recent comparison of b u f f e t loads measured on a research air-
plane w i t h t h i n unswept wings and on two d i f f e r e n t models of the airplane
i n two d i f f e r e n t wind tunnels.

INTRODUCTION

One objective of b u f f e t research has been t o l e a r n how t o predict

f l i g h t buffet loads from wind-tunnelmeasurements. A method f o r meeting
t h i s objective i s described i n reference 1. Bending-moment measurements
were made w i t h e l e c t r i c a l s t r a i n gages mounted near the wing root. By
combining knowledge from two d i f f e r e n t f i e l d s , those of s t r u c t u r a l beam
theory and s t a t i s t i c a l analysis, an equation w a s derived f o r predicting
f l i g h t bending moments from the buffet measurements on the model. Results
were presented f o r two s p e c i f i c cases, and they were encouraging.

Additional experience has been gained, since the publication of

reference 1, with regard t o t e s t techniques and modelrequirements. The
primary purpose of the present paper i s t o discuss these f a c t o r s .

SYMBOLS

average chord

first-mode generalized l i f t - c u r v e slope f o r damping

component of aerodynamic force due t o wing vibration
Mach number
473
velocity
2

a angle of a t t a c k #7?
P air density

UM root-mean-square value of bending moment due t o buffet

aerodynamic input parameter; parer s p e c t r a l density, a t

%,1 first-mode resonant frequency, of generalized normal-
force c o e f f i c i e n t f o r first-mode vibration

03 c i r c u l a r frequency

Subscript :

MAX maximum

DISCUSSION

Model Damping Requirement

One requirement f o r a b u f f e t model concerns the damping of the

buffet vibration. The question that arises i s whether the damping i s
aerodynamic, s t r u c t u r a l , or some combination of the two. mis question
can be answered, i n a given case, by studying the variation of buffet
bending moment w i t h air density as i s done i n figure 1. I n order t o
f a c i l i t a t e comparison between d i f f e r e n t tests, the density values f o r
a p a r t i c u l a r t e s t have been normalized i n figure 1 b y dividing by t h e
maximum density f o r that t e s t . Similarly, the bending moments f o r 'a
p a r t i c u l a r t e s t have been normalized by dividing by t h e bending moment
a t the maximum density. The bending moment i s herein defined as t h e
root-mean-square value of the vibratory p a r t of the wing-root bending
moment; the s t a t i c bending moment has been removed from t h e data.

If the damping i s s t r u c t u r a l , the damping c o e f f i c i e n t w i l l be con-

s t a n t and the bending moment w i l l be d i r e c t l y proportional t o the e x c i t i n g
force and thus t o the a i r density. This variation i s shown by the dashed
l i n e i n figure 1.

If the damping i s aerodynamic, on the other hand, t h e damping coef-

f i c i e n t and the exciting force w i l l both be d i r e c t l y proportional t o the
density. By the methods of power s p e c t r a l analysis, the root-mean-square
bending momer;t i s found t o be proportional t o the square r o o t of the
density. (See ref. 2.) This v a r i a t i o n i s shown by the s o l i d l i n e i n
f i g u r e 1.

F l i g h t - t e s t d a t a indicate that the damping of the buffet vibration

i s primarily aerodynamic. ( S ) I n order t o i l l u s t r a t e
t h i s point, some f l i g h t r e s u l t h c i r c u l a r symbols i n
 ,.
). .. .
.a 0 0
o m .
- -
*.
-

figure 1. Such r e s u l t s l e d t o the assumption of aerodynamic damping i n

developing the equation of reference 1 f o r scaling buffet bending moments
from wind-tunnel models t o f u l l - s c a l e airplanes. The equation i s applic-
able, therefore, only when the damping i s primarily aerodynamic f o r the
model as w e l l as f o r t h e airplane.

In order t o determine whether the damping of models a c t u a l l y i s

aerodynamic, bending-moment d a t a f o r three ordinary f o r c e - t e s t models
are p l o t t e d i n figure 1. Model 1 i s s u i t a b l e f o r buffet loads tests
because the damping i s primarily aerodynamic. For models 2 and 3 , on
the other hand, the damping i s largely s t r u c t u r a l ; thus, t h e s c a l i n g
equations of reference 1 d o not apply t o these m o d e l s . The model d a t a
shown i n f i g u r e 1 represent two extremes, and anything between these
extremes i s possible.

If the model i s t e s t e d i n a variable-density wind tunnel, a p l o t

l i k e figure 1provides a means f o r learning, a f t e r the t e s t i s finished,
whether the model w a s suitable f o r buffet loads tests. It would be
better, of course, t o know t h i s before making the t e s t . As a guide f o r
t h a t purpose, the following procedure i s suggested. With tabulated
f l u t t e r coefficients, make a rough estimate of the aerodynamic damping.
In order t o approximate t h e s t r u c t u r a l damping, measure t h e damping of
the wing i n a s t i l l - a i r vibration test. If the measured s t r u c t u r a l
damping i s as low as 1/10 of the estimated aerodynamic damping, the
model w i l l probably be s a t i s f a c t o r y . Experience i n transonic tunnels
operating a t atmospheric stagnation pressure has shown t h a t s o l i d m e t a l
wings are generally s a t i s f a c t o r y i f t h e wing-fuselage j o i n t s me t i g h t l y
clamped. D i f f i c u l t y has been experienced when i n s u f f i c i e n t j o i n t f i x i t y
r e s u l t e d i n high s t r u c t u r a l damping and when tests w e r e run a t such l o w
dynamic pressures that t h e aerodynamic damping was t o o low.

Model Frequency Requirements

Another requirement f o r the b u f f e t m o d e l i s that the vibratory mode

shape and reduced resonant frequency be t h e same as those f o r the airplane.
O f these two quantities, simulation of the resonant frequency probably i s
t h e more important. The power s p e c t r a l density of the wing-root bending
moment i s plotted as a function of the reduced frequency i n f i g u r e 2 f o r
an airplane and f o r a 0.075-scale model of the airplane a t the same Mach
nmiber. The vibration of the airplane wing i s concentrated i n t h e first
symnetrical bending mode. This i s a l s o t h e case f o r t h e model; f u r t h e r -
more, the reduced resonant frequency f o r the m c d e l i s about the same as
t h a t f o r the airplane. This comparison i s i n t e r e s t i n g because the model
has a standard solid-metal force-test wing. I n t h e design of t h i s model,
there was no consideration whatever given t o the resonant frequencies of
the wing and y e t frequency simulation of the f i r s t symmetrical bending
mode w a s obtained. This seems t o be a normal c h a r a c t e r i s t i c of solid-
n e t a l models of f i g h t e r - t e a s t i n t h e absence of

475
4

e x t e r n a l s t o r e s ) ; thus, t h e simulation of wing resonant frequency f o r

such airplanes presents no serious model design problems.

I n the case of t h e t a i l , however, t h e s i t u a t i o n i s very d i f f e r e n t .

The power s p e c t r a l density of t h e bending moment a t t h e root of a hori-
zontal s t a b i l i z e r i s p l o t t e d as a function of the reduced frequency i n
f i g u r e 3 f o r an airplane and f o r a 0.25-scale model of the airplane.
Again there w a s no attempt i n the model design t o simulate the frequency
c h a r a c t e r i s t i c s of t h e airplane. For the airplane, t h e most prominent
mode i s associated with fuselage torsion b u t t h i s mode i s hardly v i s i b l e
on the model, presumably because t h e model fuselage i s more r i g i d than
t h a t of the airplane. Note a l s o t h a t , f o r the model, the frequencies
of t h e various vibration modes are very d i f f e r e n t from those f o r t h e
airplane. B u f f e t bending moments measured on t h i s model s t a b i l i z e r
would probably bear l i t t l e r e l a t i o n t o those measured on the airplane.
It appears then t h a t , although force-test models a r e usually s a t i s f a c t o r y
f o r t h e study of wing b u f f e t loads, the scaling of t a i l b u f f e t loads w i l l
require models s p e c i a l l y designed f o r t h a t purpose.

Support System

The vibration modes of t h e model may be influenced by the support

system. Figure 4 shows power spectrums obtained with two d i f f e r e n t types
of model supports. The spectrum a t the l e f t i s f o r a floor-mounted
semispan model t h a t w a s t e s t e d i n t h e Ames 12-foot pressure tunnel. The
predominant mode i s first symmetrical bending as it w a s i n t h e case of
t h e airplane wing i n figure 2. The spectrum on the r i g h t i s f o r a sting-
mounted full-span model. Two new modes are evident i n t h i s spectrum.
The antisymmetrical bending mode i s actually a wing mode but the other
new mode i s e s s e n t i a l l y a rigid-body r o l l i n g vibration due t o the t o r -
s i o n a l f l e x i b i l i t y of t h e s t i n g support. This mode has no counterpart
i n f l i g h t and therefore should be eliminated before an attempt i s made
t o calculate f l i g h t bending moments.

A t present, there are three methods of eliminating the mode due t o

t h e s t i n g support. The f i r s t method i s t o determine the power spectrum
and then subtract the response i n t h i s mode. The second method, which
i s the same i n principle but less costly i n practice, i s t o remove the
s t i n g mode with an e l e c t r i c a l f i l t e r e i t h e r a t t h e t i m e t h e d a t a are
taken or l a t e r . For some combinations of model and s t i n g supports, t h e
n a t u r a l frequency of t h e s t i n g mode i s so close t o t h a t of t h e symmetrical
bending mode t h a t these two methods a r e not applicable. When these
methods cannot be used, t h e t h i r d method can be used. The s t i n g t o r s i o n
response can be canceled by placing bending-moment gages on both wings
and summing t h e outputs e l e c t r i c a l l y . The dashed l i n e i n the p l o t a t
t h e r i g h t of figure 4 shows a spectrum t h a t w a s obtained by t h i s method.
 5

Note, however, t h a t not only the s t i n g torsion mode b u t a l s o the a n t i -

symmetrical wing mode i s eliminated. The elimination of t h i s antisym-
metrical mode would be undesirable i n a case where t h i s mode contributed
a large p a r t of the t o t a l wing-root bending moment. Fortunately, however,
the contribution of modes higher than the first-symmetrical bending mode
i s small f o r the f i g h t e r airplanes f o r which f l i g h t spectra a r e
available.

Wind-Tunnel Turbulence

Experience has shown t h a t wind-tunnel turbulence sometimes compli-

cates the r e s u l t s of buffet t e s t s . An example i s i l l u s t r a t e d i n f i g -
ure 5. The ordinate f o r the curves i n t h i s figure i s a parameter t h a t
i s proportional t o the root-mean-square value of the wing-root bending
moment. The numerator i s associated w i t h the input force t h a t causes
the wing t o b u f f e t and the denominator w i t h the damping force t h a t
l i m i t s t h e amplitude of the buffet vibration. This parameter i s plotted
as a function of the angle of a t t a c k f o r three d i f f e r e n t Mach numbers.
A t a Mach number of 0.4, the f l a t portion of the curve a t low angles of
attack i s the response due t o wind-tunnel turbulence. The point where
the curve r i s e s steeply i s considered t o be the b u f f e t boundary. A t
Mach numbers of 0.7 and 0.85, the response t o turbulence i s much higher.
This increase has two e f f e c t s . First, the turbulence tends t o obsc&e
the b u f f e t boundary, as i s the case a t a Mach number of 0.7. In addi-
t i o n , the bending moments are higher because of t h e contribution of the
wind-tunnel turbulence. The manner i n which the turbulence and buffet
inputs add t o give the t o t a l i s not yet completely understood. Any
correction of the measured t o t a l f o r the e f f e c t s of turbulence depends
upon the frequency d i s t r i b u t i o n of t h i s input and on t h e degree of corre-
l a t i o n between the t w o inputs, f a c t o r s t h a t have not y e t been determined.
A subtraction process i s indicated schematically on t h e curve f o r a Mach
number of 0.85 where the b u f f e t response i s considered t o be only t h e
p a r t labeled "buffeting." Note t h a t the r e s u l t s i n f i g u r e 5 t h a t were
chosen for purposes of i l l u s t r a t i o n are f o r a p a r t i c u l a r l y bad case.
The e f f e c t s of turbulence are not always s o severe as shown here. On
t h e other hand, it i s important t o r e a l i z e that because the wing i s a
resonant system with very l i t t l e damping, a small exciting force due t o
turbulence may cause a sizable response.

For b u f f e t t e s t s , t h e important quantity is not the o v e r a l l t u r b u -

lence l e v e l i n the wind tunnel but r a t h e r the power s p e c t r a l density of
the turbulence a t the resonant frequency of the wing. There i s l i t t l e
information available on t h e turbulence spectra of various wind tunnels
i n the frequency range of i n t e r e s t here. Wind-tunnel turbulence has
been investigated r a t h e r extensively i n connection with studies of
boundary-layer t r a n s i t i o n . The frequencies of i n t e r e s t f o r boundary-
layer t r a n s i t i o n , however, are much higher than t h e b u f f e t frequencies.
. .. *-
7 . 1
-c
a
s
i c
- .-
- ... . . .
*- ? . 0. .*

At the present time, therefore, the best way to learn whether a given
tunnel is suitable for buffet tests probably is to make such a test.

The effect of turbulence on buffet measurements is not a problem

for the wind tunnels only. On occasion, turbulence has interfered with
buffet tests of both rocket-propelled free-flight models and airplanes.
In these cases, the solution to the problem is to avoid atmospheric
turbulence when making buffet tests.

Instrumentation

With regard to instrumentation, tape recordings of the wing strain-

gage output have been particularly useful. Power spectrums are obtained
by recording the strain-gage output on magnetic tape and later analyzing
the tape records (ref. 4). Root-mean-square bending-moment values can
be obtained from the tape record or they can be obtained at the time of
the test by means of a thermocouple meter, which is sensitive to the
mean-square value of a random electrical current. The Ames Aeronautical
Laboratory has developed an instrument that measures the peak buffet
loads in successive 10-second intervals over a period of several minutes.
The instrument is basically a condenser charged through a diode that
conducts current only during a peak that is higher than any previous
peak. Thus, the charge on the condenser is a measure of the highest
peak. Either peak or root-mean-square measurements can be used to study
buffeting. The measured peak values in the tests at the Ames laboratory
were two to four times higher than the root-mean-square values and thus
were about what would be expected for a Gaussian random process.

Additional Results Bearing on Validity of !Theory

The experimental technique that is discussed in this paper is based

on the assumption that buffeting is essentially a Gaussian random process
and that the response of the wing can be treated as the linear response
of a lightly damped single-degree-of-freedom elastic system. A recent
analysis of flight-test data (ref. 5) has shown that wing buffet loads
do indeed exhibit the characteristics of a Gaussian random process. For
a representative stall maneuver, the loads are normally distributed, and
the probability that a load peak will exceed a given level is in agree-
ment with the theoretical results obtained in reference 6. For a repre-
sentative pull-up into buffeting in the shock regime, the buffet intensity
appears to vary linearly with penetration beyond the buffet boundary.
The loads under maneuvering conditions are therefore not stationary and
are thus non-Gaussian. By means of a simple linear transformation, how-
ever, the buffet loads in maneuvering flight can be treated as a
Gaussian process.
 ..
I 8.
... .. e m .
.*
8
.*
.
I .
-,
w
*
r.

Another r e s u l t of t h e o r e t i c a l i n t e r e s t has been established by an

analysis (unpublished) of parer spectrums of wing bending moments
obtained i n a wind tunnel. It w a s found t h a t the relationship of the
maximum value of the spectrum, t h e band width of the spectrum a t the
1/2-parer points, and the integrated mean-square bending moment i s the
same f o r the wing during buffeting as f o r a lightly damped l i n e a r single-
degree-of-freedom system. This result i s important not only as v e r i f i -
cation of a basic assumption but a l s o because it implies that t h e magni-
tude of the wing damping forces can be evaluated, a t l e a s t i n principle,
from the parer spectrum.

Comparison of Flight and Wind-Tunnel Results

Wing-root bending moments during buffet have been measured i n

f l i g h t on a thin, unswept wing, research airplane (X-1E). They have
a l s o been measured on two models of d i f f e r e n t s i z e s i n two d i f f e r e n t
wind tunnels. The wind-tunnel results have been scaled up t o f l i g h t
conditions by using the equation presented i n reference 1. The values
obtained are compared with the f l i g h t bending moments i n f i g u r e 6. When
the f a c t t h a t buffeting i s inherently a random process i s considered,
the agreement between f l i g h t and wind-tunnel results shown i n t h i s f i g -
ure i s regarded as s a t i s f a c t o r y . Apparently, f l i g h t buffet loads can
be estimated from wind-tunnel r e s u l t s , a t least f o r simple wings. It
follows from t h i s r e s u l t t h a t wind tunnels can a l s o be used t o study
the e f f e c t of airplane modifications on the buffet loads.

CONCLUDING REMARKS

A description has been given of recent developments i n the appli-

cation of a method f o r predicting f l t t buffet loads from wind-tunnel
!?
modelmeasurements. Model requiremen s have been outlined f o r proper
scaling of the damping c h a r a c t e r i s t i c s and vibration frequencies. These
requirements appear t o be easily met f o r wing loads, but s p e c i a l care
w i l l be needed i n the construction of models f o r the study of t a i l loads.

‘The influence of t h e m o d e l support system and of the turbulence

l e v e l i n t h e wind tunnel on t e s t results has been discussed, and the
instrumentation f o r wind-tunnel studies has been described. A comparison
has been made of t h e wing buffet loads on a research airplane with t h i n
unswept wings and the loads that were measured on two models of the air-
plane b u i l t t o d i f f e r e n t scales and tested i n d i f f e r e n t wind tunnels.
The comparison indicates t h a t f l i g h t wing loads can be estimated from
wind-tunnel results; thus, wind tunnels could a l s o be used t o study the
e f f e c t s of airplane modifications on b u f f e t loads.
e
e
e .. P
o r
,
.
- ...
*.a

REFERENCES

1. Huston, Wilber B., Rainey, A. Gerald, and Baker, !Thomas F.: A Study
of t h e Correlation Between F l i g h t and Wind-Tunnel Buffeting Loads.
NACA RM L55El6b, 1955.

2. Liepmann, E. W.: Parameters f o r U s e i n Buffeting F l i g h t Tests.

Rep. No. SM-14631, Douglas Aircraft Co., Inc., Jan. 3, 1953.

3. Huston, Wilber B., and Skopinski, T. H. : Measurement and Analysis

of Wing and T a i l Buffeting Loads on a Fighter Airplane. NACA
Rep. 1219, 1955. (Supersedes NACA TN 3080.)

4. Smith, Francis B.: Analog Equipment f o r Processing Randomly Fluc-

tuating D a t a . Aero. Eng. Rev., vol. 14, no. 5, May 1955,
pp. 113-119.

5. Huston, Wilber B., and Skopinski, T. H. : Probability and Frequency

Characteristics of Some F l i g h t B u f f e t Loads. NACA TN 3733, 1956.

6. Rice, S. 0.: Mathematical Analysis of Random Noise. P t s . I and 11.

B e l l Syst. Tech. Jour., vol. X X I I I , no. 3, July 1944, pp. 282-332;
P t s . I11 and IV, vol. XXIV, no. 1, Jan. 1945, pp. 46-136.
 9

DAMRNG
AERODYNAMIC OR STRUCTURAL

I
AERODYNAMIC

.5

aM (P) 0 AIRPLANE
ffM (PMAX) 0 MODEL I
0 MODEL 2
/STRUCTURAL A MODEL 3
A DAMPING
(rMaP)
.I
Ll I I I
0 .I .2 .5 I
PPMAX

Figure 1

MODEL DESIGN
SI MULATlON OF WING NATURAL FREQUENCY

AIRPLANE

SYMMETRICAL
BENDING

POWER
SPECTRAL
DENSITY , 0.075-SCALE MODEL

BENDING

0 .2 4 .6 .0
REDUCED FREQUENCY,-

Figure 2

5 - . .,..'
481
10

MODEL DESIGN
DISSIMILARITY OF TAIL NATURAL FREQUENCIES
AIRPLANE
I

TORSION
STABILIZER
BENDING
BENDING

POWER I J
SPECTRAL 0
DENSITY

Ir STABILIZER
BENDING
FUSELAGE

SUPPORT SYSTEM
EFFECT ON MODEL BUFFET MODES

-
---- RIGHT WING
SUMMATION
FLOOR- MOUNTED STING-MOUNTED
SEMISPAN MODEL FULL-SPAN MODEL

FIRST
BENDING

SECOND

FREQUENCY FREQUENCY

4s2 Figure 4 h
-4
J

TURBULENCE COMPLICATES BUFFET TESTS

12
r XIO-3
M = 0.4

4
I

V
(%)F
M = 0.7 M = 0.85
BUFFETING

4 TURBULENCE

0 4 8 12 16 0 4 8 12 16
a,cm . Q, DEG

Figure 5

BUFFET LOAD
FLIGHT AND WIND-TUNNEL COMPARISON FOR X-IE

c1
FLIGHT
30 xi03 00.25- SCALE MODEL

M -0.7
-
0 0.075 SCALE MODEL

25

20
U

.
BUFFET 0
BENDING 15 - 0

YE? 0.0, 0
M mO.8 U

10 - 00
0 0 00

.
0

0
5 -
0
0
I -
I- I I

a , DEG

Figure 6
483
 .
. ..
... ... . . ...
I .
.b
8

EFFECTS OF WING AND FUSELAGE MODIFICATIONS ON BUFFETING

By Fred B. Sutton and J. Walter Lautenberger, Jr.

Ames Aeronautical Laboratory

INTRODUCTION

The performance of many high-speed a i r c r a f t has been adversely

affected by shock-induced separation encountered a t moderate l i f t coef-
f i c i e n t s a t high subsonic speeds. This phenomenon i s usually evidenced
by severe decreases i n longitudinal s t a b i l i t y , large increases i n drag,
and heavy buffeting. Much research has been directed toward the a l l e -
v i a t i o n of the s t a b i l i t y and drag d i f f i c u l t i e s , but l i t t l e i s known
about the e f f e c t s on buffeting of such s t a b i l i t y and d r a g "fixes" as
wing fences, wing leading- and trailing-edge modifications, and fuselage
indentations. This paper presents some r e s u l t s of investigations i n
the Langley 8-foot transonic pressure tunnel and i n the Ames 12-foot
pressure tunnel t o obtain an indication of the e f f e c t s of these devices
on buffeting .
DISCUSSION

The t e s t s i n t h e 8-foot tunnel were conducted with a model of a

modified delta-wing airplane and some t y p i c a l r e s u l t s of t h e t e s t s i n
t h i s tunnel a r e shown i n figure 1. Fluctuating wing bending moments
measured a t t h e wing root of the model were reduced t o show the r e l a -
t i v e buffeting input due t o t h e aerodynamics of several model configu-
rations. The apparent buffeting input a t low l i f t coefficients i s
believed t o be due t o wind-tunnel turbulence. Test data a r e shown
f o r a Mach number of 0.97, which corresponds t o a f l i g h t speed a t which
buffeting i s c r i t i c a l .

Figure 1 shows one-half of t h e sting-mounted full-span model used

f o r the investigation. The heavy s o l i d l i n e i s t h e outline of the basic
configuration. The buffet inputs of the wing as a function of t h e lift
coefficient a r e shown i n t h i s figure. When conical camber was added t o
t h e leading edge of the wing, t h e buffet i n t e n s i t y at the higher l i f t
coefficients was reduced considerably. By f u r t h e r adding a swept t r a i l i n g
edge t o t h e wing and a bump t o t h e body i n order t o improve the a r e a dis-
t r i b u t i o n of t h e airplane, the buffet i n t e n s i t y was s t i l l f u r t h e r reduced.
These modifications, which were designed t o improve t h e airplane perform-
ance, thus were quite e f f e c t i v e a l s o i n reducing t h e buffet loads a t
transonic speeds.
484
 e..
..e
4 P

The investigation in the 12-foot tunnel was made with a wing-

fuselage-tail combination having a relatively thick sweptback wing with
a high aspect ratio. This combination was typical of transport and
bomber airplanes designed for long-range flight at high subsonic speeds.
Reference 1 describes the semispan-model technique and the instrumenta-
tion eqloyed in the investigation of buffeting in the 12-foot tunnel.
Figures 2 and 3 illustrate the semispan-model setup and the model con-
figurations used in this wind tunnel, respectively. Included in the
investigation were studies of the effect of wing fences, a leading-edge
extension, and a fuselage indentation of the Kuchemann type. The wing
used in this investigation was twisted and cambered and had NACA 6 4 ~ -
series thickness distributions. A strain-gage bridge was mounted at the
wing root to measure fluctuations of wing bending moment due to buf-
feting. The bridge output was fed into an electronic recorder-analyzer.
This apparatus permitted the convenient recording and rapid analysis of
data samples corresponding to several thousand fluctuations of bending
moment and provided the maximum peak intensities as well as the usually
measured root-mean-square intensities (refs. 2 and 3 ) of the fluctuating
wing bending moments.

It has not been possible to deduce from the buffet outputs meas-
ured in the 12-foot tunnel the buffet inputs due to the model aerody-
namics as was the case with the results obtained in the 8-foot tunnel.
Instead, the measured fluctuating bending moments have been converted
to approximate values of fluctuating normal-force coefficient mN.
Because the dynamic characteristics of the model were essentially con-
stant for the various configurations tested, these results do indicate
relative buffeting.
The effects of wing fences and a leading-edge extension on the buf-
feting of the wing-fuselage-tail combination are shown in figures 4 and
5. Figure 4 shows the variation of X N , with ~ lift coefficient at
a Mach number of 0.86, and figure 5 shows boundaries of lift coefficient
and Mach number for heavy buffeting.

The intensity chosen for heavy buffeting, A C N , = ~ W.08, is

strictly arbitrary and is intended only to indicate buffeting of rela-
tively heavy degree. The boundaries shown were determined from maximum
buffet intensities; however, boundaries from comparable root-mean-square
intensities were very similar. The test results indiEated that, at most
Mach numbers, the wing fences generally reduced the erratic variation of
buffet intensity with increasing lift coefficient and reduced buffet
intensities at moderate and high lift coefficients. The wing leading-
edge extension (figs. 4 and 5 ) was relatively ineffective as a means of
reducing buffeting and, for some test conditions, increased buffeting.
Figure 5 also shows that the fences greatly reduced the adverse effects
of increasing Mach number on heavy buffeting. The combination was also
 3

tested with a fuselage indentation which was designed to reduce the

wing-fuselage interference effects. These results are not shown but
they indicate that the fuselage indentation reduced buffeting for some
test conditions; however, these effects were small compared with the
improvements derived from the use of wing fences.

The lift coefficients for drag divergence and for pitching-moment-

curve inflection or pitch-up are considered important design parameters
in analyzing the static longitudinal characteristics of airplanes. These
parameters are compared in figure 6 with boundaries of Mach number and
lift coefficient for light buffeting and the previously described bound-
aries for heavy buffeting. The intensity selected for light buffeting,
~ C0.02, approximates the boundary for buffet onset. At the
A C N , =
higher Mach numbers, the lift coefficients for drag divergence are close
to the boundaries for buffet onset for both the model with and without
wing fences. However, heavy buffeting was indicated at lift coefficients
considerably lower than those for pitch-up. This is significant, inas-
much as the occurrence of heavy buffeting at these comparatively low
lift coefficients indicates that the usable range of lift coefficients
is probably much less than the lift-coefficient range for stability.
Wing fences did much to lessen this difference, but heavy buffeting was
still indicated at lift coefficients which were appreciably lower than
those for pitch-up.

SuMMeRY OF RESULTS

The results of these investigations may be swmnarized as follows:

1. Many modifications intended to improve drag and stability char-

acteristics are also effective buffet alleviators.

2. For a model of a modified delta-wing airplane, a conically cam-

bered leading edge, a swept trailing edge, and body bumps to improve
area distribution all reduced buffeting.

3 . For sweptback wings having high aspect ratios, fences were a

very effective means of reducing buffeting. To a much lesser degree,
a Kuchemann fuselage indentation also reduced buffeting. A wing leading-
edge extension was ineffective as a means of alleviating buffeting and,
for some test conditions, increased the buffeting.

4. At high subsonic speeds for sweptback wings having high aspect

ratios, the lift-coefficient boundaries for buffeting onset are close
to the lift-coefficient boundaries for drag divergence; however, heavy
buffeting at these speeds usually occurred at lift coefficients which
were considerably lower than the lift coefficients for pitch-up.
 I

REFERENCES

1. Davis, Don D., Jr., and Huston, Wilber B.: The Use of Wind Tunnels
To Predict Buff e t Loads. (Prospective NACA paper. )

2 . Polentz, Perry P., Page, W i l l i a m A., and Levy, Lionel L., Jr.: The
Unsteady Normal-Force Characteristics of Selected NACA P r o f i l e s at
High Subsonic Mach Numbers. NACA RM A33CO2, 1933.

3 . Huston, Wilber B., Rainey,A. Gerald, and Baker, Thomas F. : A Study

of the Correlation Between Flight and Wind-Tunnel Buffeting Loads.
NACA RM LZE16b, 1953.
 EFFECTS OF MODIFICATIONS
M-0.95

o BASIC MODEL
v;
0 L.E. CAMBER
0 L.E. CAMBER + SWEPT T.E. t FUSELAGE BUMP

BUFFET
INPUT
0
A m
0 “ 0

”
-2 0 .2 .4 .6 .8
GL

Figure 1

GENERAL ARRANGEMENT

BASIC FUSELAGE
----- INDENTED FUSELAGE
54.61‘I

TUNNEL
FLOOR

Figure 2

* ” 488
 WING VARIABLES
A= 40'
A = 7.0

TYPICAL FENCE

CATIANC

FENCES

ACN, MAX

0 .2 .4 .6 .8 I .o
CL

Figure 4
3G
7

EFFECTS OF WING MODIFICATIONS

ACN,MAX 20.08

1.0 -
.8-
J- +FENCES

.6 -
CL

.4- \

*2 t
L L
0 .6
I
.7
I
.8
M
I
.9
I
1.0

Figure 5

COMPARISON OF BUFFET BOUNDARIES WITH

STATIC LONGlTUDlNAL PARAMETERS

1.0 '

.8

.6
GL (LIGHT BUFFET)
.4 -

.2 -
1 FENCES OFF '\ FENCES ON -\

LI I I I 1 I I I I I
0 .6 .7 .8 .9 I.o .6 .7 .8 .9 1.0
M M
 0.

TH!IORETICAL AND EXPERIMENTAL INVESTIGATIONS

OF DELTA-WING VIBRATIONS
By Edwin T. Kruszewski, Eldon .E.Kordes,
and Deene J. Weidman

Langley Aeronautical Laboratory

SUMMARY

The results of some theoretical and experimental investigations of

delta-wing vibrations are discussed.

Nodal-line patterns and frequencies of a 45' built-up thin-skin

delta-wing specimen obtained experimentally are compared with those cal-
culated by two analytical methods - the idealized-structure type of
method as described by Levy and EL limited-deformation type of method
proposed by Stein and Sanders. It is shown that when the effects of
transverse shear are included into the Levy approach the agreement
between calculated and experimental frequencies is exceptionally good.

Experimental nodal-line patterns and frequencies for a 60° thick-

skin delta wing are also shown.

INTRODUCTION

The prediction of the vibrational characteristics of aircraft struc-

tures is a problem of importance to the designer of high-speed aircraft.
As was discuksed in reference 1, methods of predicting modes and fre-
quencies for the large-apect-ratio box-beam type of structure are very
successful. However, for low-aspect-ratio or delta wings, the problem
is still of particular concern because of the analytical difficulties
involved in predicting their stiffness characteristics.
An experimental investigation of the modes and frequencies of two
large-scale built-up delta-wing specimens has recently been completed.
The results of this experimental investigation are being used to evalu-
ate theoretical methods of deflectional analysis. This paper will deal
with the findings of this evaluation.
491
... .

DISCUSSION

One of t h e delta-wing specimens used i n t h e experimental investiga-

t i o n i s shown i n figure 1. It i s a built-up large-scale 45' d e l t a with
a span of 18 f e e t 8 inches, a midchord of 8 f e e t , and a uniform carry-
through section of 2 f e e t 8 inches. The wing i s uniform i n depth i n t h e
chordwise d i r e c t i o n but varies i n depth i n t h e spanwise d i r e c t i o n from
5-i-
1 inches a t t h e carry-through section t o 12 inches a t t h e t i p . The
4
covers are made up of a t h i n sheet s t i f f e n e d by spanwise s t r i n g e r s . I n
order t o f a c i l i t a t e construction, the s t r i n g e r s w e r e placed on t h e out-
s i d e of t h e covers. The i n t e r n a l construction consisted of four spanwise
spars and a bent leading-edge spar with l i g h t streamwise r i b s spaced a t
close i n t e r v a l s . A detailed description of t h e s t i f f n e s s properties and
weight d i s t r i b u t i o n of t h e specimen i s given i n reference 2.

A general view of t h e vibration t e s t setup i s shown i n f i g u r e 2.

The d e l t a wing i s hung v e r t i c a l l y by f l e x i b l e a i r c r a f t cables from t h e
gallows shown i n t h e f i g u r e . The specimen w a s vibrated i n the horizon-
t a l direction s o t h a t t h e wing could be considered t o be e s s e n t i a l l y
free-free. The vibrations were produced by four electromagnetic shakers
( t h r e e of which can be seen i n f i g . 2 ) . The mode shapes were obtained
from pickups mounted at t h e i n t e r s e c t i o n of t h e spars and r i b s while t h e
node-line locations were determined with a portable probe pickup. A
t o t a l of 10 n a t u r a l modes of vibration was established, f i v e symmetrical
and f i v e antisymmetrical, with frequencies ranging from 43 t o 216 cycles
per second. Detailed descriptions of t h e vibration t e s t s a r e given i n
reference 2.

I n recent years a v a r i e t y of methods of load d e f l e c t i o n a l analysis

f o r low-aspect-ratio and d e l t a wings has been proposed. (See, f o r exam-
ple, r e f s . 3 t o 6.) Two of these methods have been used f o r calculating
t h e modes and frequencies of t h i s specimen. One w a s described by Levy
i n reference 3 and the other was developed by S t e i n and Sanders i n ref-
erence 4. Although both of these methods were discussed i n d e t a i l i n
references 3 and 4, a recapitulation i s thought t o be worthwhile. Both
of t h e methods are an influence-coefficient type of procedure; that i s ,
t h e v i b r a t i o n a l problem i s set up through t h e use of s t a t i c influence
coefficients and t h e i n e r t i a l c h a r a c t e r i s t i c s of t h e s t r u c t u r e . Further-
more, both methods neglect t h e e f f e c t s of transverse shear.

The e s s e n t i a l features of t h e two methods a r e shown schematically

i n f i g u r e 3 . The Levy method deals with a simplified s t r u c t u r e but
allows a r b i t r a r y deflections of that s t r u c t u r e . On t h e other hand, t h e
Stein-Sanders method analyzes t h e a c t u a l s t r u c t u r e but r e s t r i c t s t h e
allowable deflection shape.
c 63-
hd4
 3

Consider first the Levy method. A s shown i n f i g u r e 3 the actual

wing is idealized i n t o a system of interconnecting beams and torque
boxes. A l l t h e spanwise normal-stress-carrying material of t h e covers,
both sheet and stringers, i s concentrated i n t o t h e spars of t h e i d e a l -
ized wing while a l l t h e chordwise normal-stress-carrying material i s
concentrated i n t o t h e ribs. The shear-carrying capacity of t h e cover
sheets i s accounted f o r by torsion boxes i n t h e spar-rib cells.

The s t i f f n e s s influence coefficients f o r each component p a r t of t h e

idealized structure, t h a t i s , t h e spars, r i b s , and t o r s i o n boxes, are
superimposed t o yield t h e s t i f f n e s s influence coefficients of t h e com-
p l e t e structure. The f l e x i b i l i t y influence coefficients are then obtained
d i r e c t l y by inversion. Note that because of t h e nature of t h e idealiza-
t i o n used, t h e Levy method i s intended t o be applied primarily t o t h i n -
skin structures, such as t h e p a r t i c u l a r 4 5 O delta-wing specimen used i n
t h e experimental investigation.

The Stein-Sanders method deals w i t h t h e deflections of t h e n e u t r a l

surface of t h e a c t u a l w i n g . It i s assumed that t h e chordwise v a r i a t i o n
of t h e deflection w is parabolic and can be expressed i n t h e form shown
i n figure 3 . I n t h i s case Id, gl, and g2
give the spanwise v a r i a t i o n
of t h e deflection, slope, and chordwise curvature a t t h e t r a i l i n g edge.
Since t h e e f f e c t s of transverse shear a r e neglected, t h e d i s t o r t i o n s of
the various elements i n t h e wing can be expressed .in terms of t h e 6 ' s .
Further by taking a number of s t a t i o n s along t h e wing and by going through
a straightforward calculation, t h e v a h e s of t h e 6 ' s a t t h e various
s t a t i o n s can be expressed i n terms of t h e loading a t those s t a t i o n s .

Since the Stein-Sanders method deals w i t h t h e a c t u a l structure, it

should handle a thick-skin wing as well as a thin-skin w i n g . However,
t h e r e s t r i c t i o n t o parabolic deflections i n t h e chordwise d i r e c t i o n may
lead t o serious e r r o r s unless t h e center section i s very s t i f f against
chordwise bending.

The comparisons of t h e r e s u l t s of these two t h e o r e t i c a l solutions

and those obtained from t h e experimental investigation a r e shown i n
f i g u r e 4. I n t h i s figure t h e nodal-line p a t t e r n s and frequencies a r e
shown f o r t h e f i r s t f i v e symmetrical modes.

A s can be seen, t h e nodal-line patterns from t h e Stein-Sanders

method agree quite w e l l with the ones obtained experimentally. The
frequency agreement, however, is poor; t h e e r r o r s ranging from 7 per-
cent i n t h e f i r s t mode t o 38 percent i n the f i f t h mode. On t h e other
hand, t h e frequency agreement i n t h e Levy method i s much b e t t e r . The
l a r g e s t e r r o r i n t h e f i r s t four modes (which occurs i n t h e t h i r d mode)
i s only 7 1 percent while t h e e r r o r i n t h e f i f t h mode i s 20 percent.
2
4

Some of the errors in the Stein-Sanders method are undoubtedly due

to the assumption of parabolic chordwise deformation. This particular
specimen had no extra chordwise stiffening in the center of the wing,
such as would be furnished by a fuselage, for example. On the other
hand, as previously noted, the Levy approach should be applicable to
this specimen because of its relatively thin covers. Therefore, it is
not too surprising that the Levy method gives better results.

Although the Levy method does give better results, the results are
still somewhat unsatisfactory, especially for the fifth mode. For this
reason, an investigation of the influence of transverse shear, which was
neglected in the preceding calculations, was undertaken. This investi-
gation dealt solely with the Levy method for which an approximate cor-
rection for transverse shear could be made with little additional labor.
The effects of the shear deformations of the webs were simply included
in the stiffness-coefficient calculations of the individual spars and
ribs, but the torsion-box coefficients were left unchanged.
The results from this recalculation are presented in table I which
summarizes the frequencies obtained by the different methods for the
first five symmetrical modes. The frequencies as obtained experimentally
are tabulated in the first row. The corresponding frequencies as calcu-
lated by the Stein-Sanders method and by the Levy method without shear
are shown in the second and third rows, respectively. Shown in the
fourth row are the results from the Levy method with the effects of trans-
verse shear included. The frequencies shown in the last row will be dis-
cussed a little later in the paper. A s can be seen, the frequencies cal-
culated by the Levy method with shear are in excellent agreement with the
experimental frequencies. The largest error, which occurred in the fourth
mode, being siightly less than 4 percent. Furthermore, a comparison of
the frequencies in the third and fourth rows shows that the effects of
transverse shear can be appreciable; the largest effect being in the fifth
mode where the inclusion of transverse shear causes an 18-percent reduc-
tion in the calculated frequency. The effects of transverse shear on the
calculated nodal-line patterns were slight. The changes that did occur,
however, tended to improve the comparison between theory and experiment.

Although only the symmetrical modes of vibration have been discussed,

the antisymmetrical results indicate the same conclusions; that is, the
Levy method predicts the modes and frequencies of thin wings with thin
skins such as the 45' delta wing tested with more accuracy than the Stein-
Sanders method and, secondly, the effects of transverse shear, which can
be appreciable, are easily and accurately incorporated into the Levy
method.

Now to digress a little, consider the frequencies shown in the last

row of table I, which were calculated from influence coefficients deter-
mined experimentally. These resulss are of interest because a popular
e
I
i
 .
e

method of obtaining frequencies i s t o measure influence c o e f f i c i e n t s

e i t h e r on a model o r a f u l l - s c a l e s t r u c t u r e and then t o use these coef-
f i c i e n t s i n a vibrational analysis. For t h e delta-wing specimen, t h i s
p a r t i c u l a r method yields r e s u l t s which are a l s o very good. However,
it must be remembered that the influence coefficients were measured
on the a c t u a l wing; therefore, no e r r o r s due t o modeling a r e involved.
The advantage of t h e purely a n a l y t i c a l method i n t h e design stage i s
obvious.

An investigation of t h e vibrational c h a r a c t e r i s t i c s of another

delta-wing specimen of e n t i r e l y d i f f e r e n t type of construction i s being
carried out. Although the calculations a r e not complete, t h e eqerimen-
t a l information obtained should be of i n t e r e s t . A sketch of t h i s speci-
men i s shown i n f i g u r e 5. It i s a built-up 60° d e l t a wing with a span
of 8 f e e t and a midchord of 7 f e e t 4 inches. A chordwise cross section
of t h e wing i s shown i n t h e figure. The covers a r e p l a t e s with i n t e g r a l
waffle-like s t i f f e n i n g and are fastened t o r e l a t i v e l y l i g h t spars and
ribs.

The sketches of t h e experimental nodal l i n e s of t h e 60° d e l t a wing

f o r t h e f i r s t s i x modes ( 3 symmetrical and 3 antisymmetrical) are shown
i n f i g u r e 6. The corresponding natural frequencies are noted a t t h e
bottom of each sketch. The frequencies range from 82.5 cycles per sec-
ond i n t h e first mode ( f i r s t antisymmetrical) t o 207.9 cycles per sec-
ond i n t h e s i x t h mode ( t h i r d antisymmetrical). Only s i x modes of vibra-
t i o n were obtained due t o excessive panel vibrations of t h e covers. It
i s a l s o believed that these panel vibrations could be p a r t l y responsible
f o r t h e asymmetry of t h e nodal-line patterns.

The specimen i s representative of thick-skin construction and thus

w i l l furnish an acid t e s t f o r t h e Levy method. Although calculations
have not been made, it i s believed that the Levy method w i t h correction
f o r transverse shear w i l l not predict the frequencies of t h e 60° d e l t a
wing as w e l l as it predicted the frequencies of t h e 45' d e l t a wing. The
p r i n c i p a l source of e r r o r i n t h e Levy method i s t h e neglect of the i n f l u -
ence of t h e Poisson's r a t i o e f f e c t on the i n t e r a c t i o n of t h e chordwise
and spanwise s t r e s s e s . This e f f e c t , which i s important i n thick-skin
wings, cannot be r e a d i l y included i n t o t h e Levy approach.

CONCLUSIONS

Comparisons of experimental and calculated modes and frequencies

of a delta-wing specimen have shown that t h e idealized-structure type
of method as proposed by Levy gives excellent r e s u l t s f o r thin-skin wings
provided that corrections a r e made f o r the e f f e c t s of transverse shear.
The l i m i t i e d deformation type of met-hod proposed by S t e i n and Sanders i s

495
6

apparently inapplicable t o low-aspect-ratio wings when t h e center sec-

t i o n is not very s t i f f against chordwise bending.

REFERENCES

1. Hedgepeth, John M.: Summary of Recent Theoretical and Experimental

Work On Box-Beam Vibrations. NACA RM L55EOga, 1955.

2. Kordes, Eldon E., Kruszewski, Edwin T., and Weidman, Deene J.: Experi-
mental Influence Coefficients and Vibration Modes of a Built-Up
45O Delta-Wing Specimen. NACA TN 3999, 1957. (Prospective NACA paper.)
3 . Levy, Samuel: S t r u c t u r a l Analysis and Influence Coefficients f o r
Delta W i n g s . Jour. Aero. Sci., vol. 20, no. 7, July 1953,
pp. 449-454.

4. Stein, Manuel, and Sanders, J. Lyell, Jr.: A Method f o r Deflection *

Analysis of Thin Low-Aspect-Ratio Wings. NACA TN 3640, 1956.

5 . Schuerch, H. U., and Freelin, J. R . : S t r u c t u r a l Analysis of a D e l t a

Wing Structure by E l a s t i c Coefficients. Rep. No. zs-182, Convair,
May 5 , 1953.
6. Williams, D.: Recent Developments i n t h e S t r u c t u r a l Approach t o Aero-
e l a s t i c Problems. Jour. R.A.S., vol. 58, June 1954, pp. 403-428.
 7

TABLE I

COMPARISON OF EXPERIMENTAL AND CALCULATED FREQUENCIES

Frequency, cps, for -

Frequency determined by - 1 s t I 2d
mode mode
I I I
3d
mode
4th
mode
5th
mode

Experiment ........... .. . 43.3 88.8 122.8 164.2 179 7

.........
Stein-Sanders method 46.4105.3 150.0 202.0 248.0
.....
Levy method (without shear) 44.6 94.7 132.0 172.0 216.0
.......
Levy method (with shear) k2.8 88.9 120.1158.0
..
Experimental influence coefficient
-
43.1 83.0 118.0 146.0 172.0
 45" DELTA - W I N G SPECIMEN

RIBS
-{SPARS
STRINGERS

Figure 1

VIBRATION-TEST SETUP OF DELTA-WING SPECIMEN

Figure 2
L-88071
498
 9

T H E O R E T I C A L APPROACHES CONSIDERED

LEVY STEIN - SANDERS

NODE LINES AND FREQUENCIES FOR 45" DELTA WING

1st MODE z d MODE

A
-EXPERIMENTAL

----LEVY (44.6)
---STEIN -
(43.3)

SANDERS (46.4)
3 d MODE
-EXPER I MENTAL (88.8)
---- LEVY (94.7)
---STEIN - SANDERS (105.3)
--____

.\. _
-_.
I

-EXPERIMENTAL (122.8)

A
4th
_---LEVY (132) 5 th MODE
---STEIN-SANDERS (150)

I ;, _... '\'
-EXPERIMENTAL(I64.2) -EX PER I M ENTAL ( I 79.7)
--__ LEVY (172) ----LEVY (216)
STEIN -SANDERS (202) ---STEIN -SANDERS ( 2 4 8 )
10

60" DELTA - W I N G SPECIMEN

t=i

6"

I 8' 0" I
NODE LINES AND FREQUENCIES OF 60" DELTA WING

1st MODE

2 d MODE
A 90.6 CPS

3 d MODE

A A 102.3 CPS

(a) SYMMETRICAL.
171.2 CPS

1st MODE

A 82.5 CPS

&.,A143.1 CPS
?id MODE

207.9 CPS

(b) ANTISYMMETRIGAL.

Figure 6
50%
 OSCILLATING AIR FORCES AND A PRESENTATION OF

SOME FLU'ITER CALCULATIONS

By Charles E. Watkins, Donald S. Woolston,

and Herbert J. Cunningham

Langley Aeronautical Laboratory

INTRODUCTION

This paper is made up in two parts: the first part being a summary
of the present status with regard to the calculation of aerodynamic forces
necessary in aeroelastic problems, such as flutter; the second part being
a presentation of some recent results of flutter calculations that are
compared with experimentally determined results.

SYMBOLS

A aspect ratio

XtY;S,rl Cartesian coordinates

cp velocity potential

f,Fn functions of 5 and q

an constant coefficients

P density

M Mach number
u) circular frequency

9JU2 circular frequency of first and second vibration modes,

respectively

v velocity

root chord
2

k frequency parameter, -
2v

amplitude of downwash velocity

lift distribution

kernel function

region of integration

amplitude of lift coefficient associated with translatory

oscillations

amplitude of lift coefficient associated with pitching

oscillations

amplitude of lift coefficient associated with parabolic

chordwise camber oscillations

amplitude of lift coefficient associated with parabolic

bending oscillations

amplitude of moment coefficient associated with transla-

tory oscillations
amplitude of moment coefficient associated with pitching
oscillations

phase angle associated with lift due to translatory

oscillations

phase angle associated with lift due to pitching

oscillations

phase angle associated with lift due to parabolic chord-

wise camber oscillations

phase angle associated with lift due to parabolic bending

oscillations

phase angle associated with moment due to translatory

oscillations

phase angle associated with moment due to pitching

os cillations
5;:3
 *I
.*a
3
*
- a- * *
- .*
L
1 -

a.

s t i f f n e s s parameter
I

A sweep

trailing-edge sweep

AERODYNAMIC FORCES

An a e r o e l a s t i c problem involves an e l a s t i c structure, such as an

airplane wing or control surface, on which aerodynamic forces a c t . A
solution t o such a problem might imply a determination of t h e deformation
that the s t r u c t u r e undergoes because of aerodynamic loading or it might
imply a determination of conditions under which t h e s t r u c t u r e can v i b r a t e
or f l u t t e r i n an undamped mode of o s c i l l a t i o n .

The aerodynamic forces referred t o herein a r e not t h e forces required

t o sustain f l i g h t but are superimposed forces generated by deformations
o r o s c i l l a t i o n s that the s t r u c t u r e undergoes. This means t h a t a i r forces
and s t r u c t u r a l deformations a r e mutually dependent; that i s , t h e air
forces depend on t h e s t r u c t u r a l deformations and the s t r u c t u r a l deforma-
t i o n s , i n turn, depend on t h e a i r forces. Hence, i n order t o solve an
a e r o e l a s t i c problem, it must f i r s t be put i n some form whereby t h e i n t e r -
action between deformations and a i r forces can be accounted f o r . This
can be done by various considerations, depending, of course, on t h e prob-
lem a t hand.

With f l u t t e r problems t h e most comon and perhaps most p r a c t i c a l way

of r e l a t i n g s t r u c t u r a l deformations and air forces is by r e s o r t i n g t o a
so-called modal analysis. This analysis involves an assumption that f l u t -
t e r modes of vibration can be s a t i s f a c t o r i l y represented by a superposi-
t i o n of a few known supplementary modes, f o r example, modes of t h e type
i l l u s t r a t e d i n figure 1. The mode shapes used f o r t h i s i l l u s t r a t i o n a r e
t h e first three natural modes of a tapered configuration that can be seen
i n outline i n f i g u r e 1. They were determined experimentially by a photo-
graphic process and can be seen t o involve considerable camber o r curva-
ture i n both t h e chordwise and spanwise directions.

Before the aerodynamic forces necessary i n t h i s modal approach f o r

solving f l u t t e r problems a r e discussed, perhaps a few words regarding
i t s a p p l i c a b i l i t y a r e i n order. This approach has proved t o be very
s a t i s f a c t o r y f o r wings o r configurations that are s o constructed that
they behave as simple beams. However, t h e behavior of t h i n , low-aspect-
r a t i o wings of many current and future airplanes i s more l i k e that of a
p l a t e than of a beam. With configurations t h a t behave as p l a t e s , a modal
504
4

approach f o r solving f l u t t e r problems becomes somewhat precarious because

of t h e d i f f i c u l t y involved i n choosing mode shapes t h a t w i l l unite satis-
f a c t o r i l y t o y i e l d a f l u t t e r mode. The most l o g i c a l modes t o choose a r e
t h e f r e e vibration modes, if available, of t h e configuration under con-
sideration. But even with these modes, it appears necessary t o e q l o y
considerably more modes f o r p l a t e - l i k e s t r u c t u r e s than f o r beam-like
structures.

The aerodynamic forces necessary i n a modal approach a r e t h e forces

associated with each chosen supplementary mode. It i s of i n t e r e s t t o
note that t h e l i t e r a t u r e on unsteady aerodynamics deals mainly with t h i s
form of forces, although f o r t h e most p a r t t h i s l i t e r a t u r e pertains t o
special plan forms and very simple mode shapes. I n recent years, how-
ever, considerable advancement has been made i n calculating aerodynamic
forces f o r a r b i t r a r y plan forms and mode shapes f o r both subsonic and
supersonic speeds. These advancements w i l l be discussed momentarily but
f i r s t it i s of i n t e r e s t t o review b r i e f l y t h e pertinent plan forms and
give some indication as t o methods by which the aerodynamics f o r unsteady
wings has developed. The plan forms f o r which o s c i l l a t o r y aerodynamic
forces a r e known, roughly i n order of development, are as follows:

I. Wings of i n f i n i t e aspect r a t i o
(a) Velocity p o t e n t i a l
( b ) I n d i c i a l functions

11. Wings of vanishingly small aspect r a t i o

(a) Velocity p o t e n t i a l

111. Circular p l a t e i n incompressible flow

( a ) Pressure p o t e n t i a l

I V . F i n i t e wings for which flow normal t o a l l

edges i s supersonic
( a ) Velocity p o t e n t i a l
( b ) I n d i c i a l functions
( c ) Simple piston theory /k = -
1W-c r <<
(z
M 2V

V. Rectangular and arrowhead plan forms i n

supersonic flow
( a ) Velocity p o t e n t i a l expanded i n
powers of frequency

V I . Wings of any plan form

(a) L i f t d i s t r i b u t i o n
(1)Box method
( 2 ) Kernel function method
 5

These w i l l be d e a l t with very b r i e f l y i n t h i s paper. A good liography

and accounting of developments can be found i n reference 1.

The f i r s t item i s t h e wing of i n f i n i t e aspect r a t i o . The velocity

potential, and hence t h e forces and moments, f o r t h i s wing undergoing
harmonic t r a n s l a t i o n and pitching o s c i l l a t i o n s is now known f o r the com-
p l e t e Mach number range. I n the past t h e t h e o r e t i c a l aerodynamics of
f l u t t e r has been based mainly on t h i s two-dimensional theory. This theory
has been employed i n various strip-analysis methods and has been found t o
be very useful, p a r t i c u l a r l y f o r high-aspect-ratio wings with beam-like
behavior. I n addition t o the velocity p o t e n t i a l f o r t h e i n f i n i t e wing,
t h e i n d i c i a l response functions associated with sudden v e r t i c a l and
pitching motions are numerically known. These functions have applica-
t i o n s i n c e r t a i n types of analog studies of f l u t t e r problems. They a l s o
have useful applications i n such problems as determination of an air-
plane's response t o gust.

The second plan form l i s t e d is a wing of vanishingly small aspect

r a t i o . Results f o r t h i s case a r e based on t h e assumption that a l l t h e
forces a r i s e from crossflows and that t h e t r a i l i n g edge has no influence
on the nature of the aerodynamic forces that a c t upon the wing. These
r e s u l t s a r e of academic i n t e r e s t i n that they correspond t o the one
extreme, A = 0; whereas t h e two-dimensional r e s u l t s correspond t o t h e
other extreme, A = m .

Results f o r the c i r c u l a r p l a t e a r e based on a consideration of a

pressure o r acceleration p o t e n t i a l . Aerodynamic coefficients associated
with a few simple modes f o r t h i s configuration have been tabulated. Cor-
responding r e s u l t s f o r other modes can, without very'great d i f f i c u l t y , be
obtained. It i s not proposed that such r e s u l t s would ever be very use-
f u l i n a c t u a l f l u t t e r studies, but they can be very useful i n assessing
other approximate procedures f o r obtaining aerodynamic forces.

The f o u r t h plan form considered i s a wing f o r which flow normal t o

a l l edges is supersonic. This case stands somewhat alone, because the
velocity p o t e n t i a l i s known, a t l e a s t i n t h e form of a surface integral,
f o r any given mode shape. In addition t o t h i s , i n d i c i a l functions cor-
responding t o instantaneous v e r t i c a l and pitching displacements f o r such
plan forms are e x p l i c i t l y known. I n order t o obtain t h e d i s t r i b u t i o n of
forces associated with t h e i n t e g r a l representation of t h e velocity poten-
t i a l , it i s generally necessary t o r e s o r t t o approximate procedures f o r
handling surface i n t e g r a l s . I n t h i s regard an approximation that applies
when t h e Mach number i s large and t h e reduced frequency i s low has come
t o be known as "piston theory" and has created eonsiderable i n t e r e s t i n
t h e p a s t f e w months.

Large values of Mach number M and low values of frequency parame-

ter k imply small values of t h e r a t i o k/M. When t h i s r a t i o i s small
6

enough, the velocity p o t e n t i a l f o r t h i s case can be e x p l i c i t l y expressed

i n a simple limiting form rather than as a surface i n t e g r a l . The forces
obtained a t a given point from such a limiting form depend s o l e l y on t h e
motion of t h e s t r u c t u r e a t that point. Hence, each point of the surface
can be imagined t o behave as a s m a l l independent piston; thus, t h e name
"piston theory." The word "simple" i s used here t o distinguish r e s u l t s ,
which a r e s t r i c t l y l i n e a r , that a r e obtained from t h i s l i m i t i n g process
from more elaborate second-order r e s u l t s that are a l s o known as piston
theory. For example, see t h e results of Hayes and L i g h t h i l l ( r e f . 1).
The main concern here i s t h e subsonic and low supersonic speed range
wherein there i s no b a s i s f o r piston theory.

Item V i n t h e l i s t p e r t a i n s t o rectangular- and arrow-type plan

forms. The velocity p o t e n t i a l corresponding t o f a i r l y simple mode shapes
f o r these plan forms can be derived i n t h e form of a power s e r i e s i n
terms of t h e frequency of o s c i l l a t i o n . This procedure and c e r t a i n appli-
cations thereof have been published f o r several years now. It i s men-
tioned a t t h i s time f o r two reasons: F i r s t , because of some s i g n i f i c a n t
derivations and applications that have recently been made and a r e d i s -
cussed l a t e r i n t h i s paper and second, because r e s u l t s of the expansion
procedure are used as a yardstick by which other l e s s r e s t r i c t e d approxi-
mate procedures t h a t a r e about t o be discussed can be evaluated.

I t e m V I i n t h e l i s t p e r t a i n s t o a r b i t r a r y plan forms. It i s f o r
such plan forms t h a t aerodynamic information has been c r i t i c a l l y lacking
f o r a long time. It is a l s o f o r these plan forms that s i g n i f i c a n t
advancements have recently been made. The advancements referred t o
are not based on new p r i n c i p l e s . The general methods on which they are
based have been known f o r some time, but t h e r e has been a drawback t o
t h e i r application because of t h e enormous amount of calculation involved.
With the high-capacity computing equipment t h a t i s now available, how-
ever, t h i s phase of t h e work becomes simply a t a s k of systemization.
Such systemizations a r e being accomplished, and it appears that a point
w i l l soon be reached where t h e aerodynamics of linearized theory can be
completely exploited.

The ''box method" as used herein pertains only t o wings with a l l

supersonic edges. Basically, it is a method f o r numerically handling
c e r t a i n surface i n t e g r a l s t o which t h e forces corresponding t o a given
mode can be reduced. The form of these i n t e g r a l s i s shown i n t h e f o l -
lowing express ion :

The function f i n t h e i n t e g r a l i s r e l a t e d t o a mode of o s c i l l a t i o n and

i s known. The function Cp i s t h e p o t e n t i a l f o r a source and is, of
 7

course, known. The region of integration S ' i s indicated by shaded

areas i n t h e sketches i n figure 2. This method w a s named by Pines and
others a t Republic Aviation Company, who proposed a procedure termed
"box method" f o r obtaining generalized forces on f l e x i b l e wings. A s
o r i g i n a l l y proposed, t h e wing plan form i s replaced by a rectangular
g r i d as indicated i n t h e sketch a t the l e f t i n f i g u r e 2. The leading
edge i s made jagged as shown. The integrations, i n t h i s o r i g i n a l t r e a t -
ment, were effected by expanding integrands and approximating a term-
by-term integration over t h e portion of each l i t t l e box or rectangle
that can a f f e c t t h e force a t a given point x,y. When high-capacity
computing equipment, such as an IBM 650 o r higher capacity machine, i s
used very l i t t l e i s a c t u a l l y gained i n computing time by approximating
the plan form o r by approximating the integrands. Therefore, t h i s pro-
cedure has been programed f o r calculating forces f o r wings with super-
sonic edges without expanding t h e integrands and i n such a way that t h e
plan form is completely accounted for, as indicated i n t h e sketch a t
the r i g h t i n figure 2. The leading edge i s not broken as i n t h e sketch
a t t h e l e f t . The present program i s f o r an IBM 650 machine and has been
t e s t e d by making a comparison with r e s u l t s f o r simple mode shapes cal-
culated by the frequency expansion method. The two s e t s of results
agree very s a t i s f a c t o r i l y . The program w i l l probably be converted t o
an IBM 704 machine i n t h e near future.

The kernel-function approach w i l l now be considered. This method

applies t o e i t h e r subsonic o r supersonic speeds and i s p e r f e c t l y general
with regard t o plan form and frequency. It involves approximating solu-
t i o n s t o an i n t e g r a l equation, namely,

I n t h i s equation w(x,y) i s t h e known amplitude of downwash o r t h e nor-

m a l velocity of points (x,y) of a wing; p i s the f l u i d density; L(6,q)
i s t h e unknown l i f t d i s t r i b u t i o n acting on t h e wing; and K(x,y;(,v;M,cu)
i s t h e known kernel function of t h e i n t e g r a l equation. The region of
integration S t extends over t h e portion of t h e wing that can a f f e c t
the induced normal velocity o r downwash a t a given point (x,y). For
subsonic speeds S ' extends over the e n t i r e wing surface but f o r super-
sonic speeds it depends on Mach number. The region of integration f o r
a t r i a n g u l a r plan form and one p a r t i c u l a r supersonic speed and f o r a l l
subsonic speeds is i l l u s t r a t e d i n f i g u r e 3 .

The procedure f o r approximating solutions t o t h e i n t e g r a l equation

i s t o assume that t h e lift can be represented by a s e r i e s of functions
that a r e known except f o r c e r t a i n constant coefficients. A general form
of expression f o r t h e l i f t and t h e r e s u l t i n g downwash equation a r e indi-
cated i n t h e following equation:
a

I n t h i s equation represents unknown constants and t h e functions Fn

are known expressions chosen t o be combined s o as t o represent t h e l i f t
.
d i s t r i b u t i o n L( g, q) Hence, when t h e coefficients are determined,
an approximate expression f o r t h e l i f t i s known. I n making a choice of
the functions Fn, one must be guided by c e r t a i n conditions that must be
s a t i s f i e d along the wing edges and t h e r e s u l t s f o r t h e few cases f o r
which t h e l i f t d i s t r i b u t i o n i s known.

I n order t o determine t h e coefficients an, a system of l i n e a r alge-

b r a i c equations i n which these q u a n t i t i e s are t h e only unknowns can be
obtained by choosing N + 1 points a t which t o s a t i s f y t h e known down-
wash. The coefficients of % i n t h i s system of equations however
involve complicated surface i n t e g r a l s . These integrals, which must be
evaluated t o make t h e equations e x p l i c i t , have t o be handled by numeri-
c a l methods and, because of strong s i n g u l a r i t i e s involved i n t h e kernel
function, t h e evaluations require s p e c i a l procedures. This i s espe-
c i a l l y t r u e i n the neighborhood of t h e s t r i p , shown darkened i n f i g -
ure 4, extending ahead of t h e point a t which t h e downwash i s being
s a t i s f i e d . Furthermore, t h e tremendous t a s k of computations involved
i n evaluating t h e i n t e g r a l s make t h e t a s k f e a s i b l e only with t h e use
of high-capacity computing equipment.

What makes t h e number of computations s o great i s t h e kernel func-

t i o n . It has t o be evaluated many times i n t h e process of determining
t h e l i f t f o r one s e t of conditions and it i s of such a nature that it
can be evaluated quickly only by automatic high-speed equipment. I n an
e f f o r t t o accelerate development and uses of t h i s kernel function approach
f o r t r e a t i n g aerodynamic problems, t h e A i r Research and Development
Command i s sponsoring a tabulation of t h e kernel function f o r a f e w Mach
numbers ( M = 0.6, 0.8, and 1.0). This tabulation, which i s now about
complete, i s being made a t the Harvard University Computation Laboratory.
These t a b l e s w i l l have many uses but t h e r e now e x i s t computing machines
that a r e capable of handling t h e complete aerodynamic problem, including
generation of t h e kernel function as it i s needed, s o t h a t searching
f o r these values i n voluminous t a b l e s i s not necessary.

With regard t o evaluating t h e surface i n t e g r a l s involved i n t h e

kernel-function approach, several procedures have been considered.
Runyan and Woolston ( r e f . 2) have developed a method of t r e a t i n g t h e
case f o r subsonic speeds by a procedure similar t o that of Falkner
( r e f . 3) f o r t r e a t i n g steady f i n i t e wings. Although t h i s procedure
 9

appears t o give s a t i s f a c t o r y r e s u l t s , it i s not r e a d i l y adaptable f o r

systematically programing f o r machine calculation. Furthermore, t h e r e
i s no b a s i s f o r i t s application when the speed i s supersonic.

Nelson and Diederich ( r e f . 4) have developed a method f o r subsoiiic

speeds that involves dividing t h e wing i n t o many small elements i n each
of which t h e downwash i s s a t i s f i e d a t t h e center of area. The lift i n
t h e elements bounded by t h e wing edges i s represented by functions that
have t h e proper behavior as t h e edge i s approached. I n a l l other e l e -
ments the l i f t i s represented by simple polynomials. Integrations a r e
effected over each individual element. (See r e f . 4.)

Recently, a method somewhat l i k e that of Multhopp ( r e f . 5 ) f o r

t r e a t i n g steady wings has been developed and programed f o r machine cal-
culation a t t h e Langley laboratory. It applies i n p r i n c i p l e t o both
subsonic and supersonic speeds but, since the kernel function d i f f e r s
f o r subsonic and supersonic speeds, a separate program i s required f o r
t h e two speed regimes. I n order t o describe t h i s procedure b r i e f l y ,
t h e machine i s instructed t o calculate t h e spanwise distance over which
t h e integrations extend and then t o divide t h i s distance i n t o a specified
number of i n t e r v a l s . Numerical integrating formulas a r e then applied
along chordwise sections a t each subdivision. Results of t h e chordwise
integrations are then employed i n t h e spanwise integrating formulas t o
complete t h e integrations. These integrations a r e simple and s t r a i g h t -
forward except i n t h e QeighborhGod of s i n g u l a r i t i e s . I n these neighbor-
hoods, use of limiting forms and s p e c i a l formulas make it very easy t o
overcome t h e d i f f i c u l t i e s .

This setup f o r subsonic speeds has j u s t recently been programed f o r

an IBM 704 machine. Although the program i s j u s t being tested, t h i s
machine requires less than ll minutes per control point t o calculate t h e
3
forces associated with f i v e vibration modes f o r a given frequency value
and Mach number. This means that, f o r 9 control points, which should be
s u f f i c i e n t i n most cases, t h e IBM 704 machine w i l l require l e s s than
15 minutes running time per frequency value and Mach number t o calculate
t h e forces associated with f i v e d i f f e r e n t vibration modes.

The supersonic case has been recently programed f o r an IBM 650 machine
and t h i s machine requires about s i x t y t i m e s as much time as t h e IBM 704
t o make t h e corresponding calculations. That i s , it takes the IBM 650
machine about 15-hours running t i m e per frequency value and Machnumber
t o calculate the forces based on 9 control points f o r f i v e d i f f e r e n t
vibration m o d e s . The program f o r t h e supersonic case w i l l be converted
t o t h e IBM 704 machine i n t h e near f u t u r e and it i s estimated t h a t it
w i l l require less time on t h i s machine than t h e subsonic case does.
 I n order t o give some indication as t o w h a t might be expected with
regard t o accuracy of r e s u l t s calculated by this kernel function proce-
dure, some comparisons a r e made f o r supersonic speeds by t h i s method
w i t h r e s u l t s calculated by t h e frequency-expansion procedure f o r a tri-
angular wing. I n reference 6 t h e expansion procedure i s applied t o m c h
w i n g s deforming according t o a general quadratic equation. The deforma-
t i o n s considered f o r comparison are shown i n figure 4. They include
v e r t i c a l t r a n s l a t i o n and pitching of t h e wing considered as r i g i d , i l l u s -
trated by s e c t i o n AA; a parabolic chordwise camber mode, i l l u s t r a t e d by
section AA; and a parabolic bending mode, i l l u s t r a t e d by section BB. The
p a r t i c u l a r plan form considered i s a 6 0' d e l t a wing. The reduced fre-
quency parameter i s 3/8 o r 0.375 and t h e Mach number is 2. "his i s t h e
Mach number condition f o r which t h e flow normal t o edges swept 60' i s
sonic.

Amplitudes of t h e forces and associated phase angles have been c a l -

culated by t h e two methods. R e s u l t s f o r the r i g i d modes a r e shown i n
f i g u r e s 5 and 6. The magnitudes of lift and moment as well as phase angle
a r e seen t o be i n very good agreement. Results f o r t h e camber and bending
modes a r e shown i n f i g u r e 7. The amplitudes a r e seen t o be i n almost per-
f e c t agreement, and t h e phase angles a r e i n f a i r agreement. There i s thus
a good indication that t h e kernel-function procedure as programed i s sound.
Furthermore, t h e r e seems t o be no reason why it would not give as good
r e s u l t s f o r subsonic speed as these results i n d i c a t e it does f o r super-
sonic speed. It therefore appears t h a t t h e r e i s a sound and f e a s i b l e way
of determining t h e l i n e a r i z e d aerodynamic forces f o r any given plan form,
mode shape, Mach number, and frequency. It i s further pointed out that,
by considering t h e mode shape i n t h e form of a matrix of d e f l e c t i o n func-
t i o n s , t h e kernel-function procedure can be handled s o as t o y i e l d aero-
dynamic influence c o e f f i c i e n t s useful i n a d i r e c t a t t a c k on the eigen-
value problem f o r t h e f l u t t e r mode.

SOME EXAMPUS OF FLUTTEB CALCULATIONS

Consideration i s now given t o a few f l u t t e r calculations. The con-

f i g u r a t i o n f o r which results are f i r s t presented is a model of a 4 5 O delta
wing. Results f o r t h i s model a r e shown i n f i g u r e 8. "he wing plan form
i s shown i n sketch a t the r i g h t of t h e figure. The modal l i n e s corre-
sponding t o t h e first four natural modes of t h e model are shown. The nec-
essary information regarding mode shapes f o r t h i s p a r t i c u l a r model were
obtained by f i x i n g a s m a l l mirror a t 24 d i f f e r e n t points on t h e surface
and recording the t r a c e of a r e f l e c t e d l i g h t G O U T C ~ as t h e wing w a s forced
t o v i b r a t e a t a n a t u r a l frequency. Unfortunately, only four modes were
measured. The model w a s f l u t t e r t e s t e d i n the Langley 8-foot transonic
pressure tunnel and experienced f l u t t e r a t a Mach number of 0.85 and a
density corresponding t o an a l t i t u d e of about 33,000 f e e t .
I 51%
 11

Aerodynamic forces appropriate f o r t h i s plan form and i t s mode

shapes were determined by t h e kernel-function method and f l u t t e r c a l -
culations have been made f o r several a l t i t u d e conditions employing f i r s t
t h r e e and then four modes. Results of t h e three-mode analyses are indi-
cated by squares and r e s u l t s f o r t h e four-mode analyses, by c i r c l e s .
The four-mode r e s u l t s may be noted t o be within about 10 percent of t h e
measured f l u t t e r speed. Although t h e r e s u l t s appear t o have almost con-
verged, it would be desirable t o include a d d i t i o n a l modes t o see t h e
e f f e c t that they might have on t h e calculated r e s u l t s .

The next r e s u l t s are f o r t h e all-supersonic edge case. Cunningham

( r e f . 7) has recently employed t h e velocity-potential expansion proce-
dure t o develop aerodynamic forces associated with r o l l i n g and s y m e t r i c
flapping modes f o r arrowhead wings with all-supersonic edges. The moti-
vation f o r these derivations w a s that t h e r e s u l t s were needed f o r ana-
lyzing t h e f l u t t e r of some all-movable control surface models t e s t e d i n
t h e Langley 9-inch by 18-inch supersonic f l u t t e r tunnel.

From observation and measurement of t h e node l i n e s of t h e n a t u r a l

modes, it w a s noted that t h e models had e s s e n t i a l l y a l l t h e i r f l e x i b i l i t y
concentrated i n t h e supporting shaft. Two coupled modes were used i n
t h e analyses, and each coupled mode involved components of v e r t i c a l t r a n s -
l a t i o n , pitching, and flapping. The amount of each component could be
determined by t h e l o c a t i o n of t h e node l i n e s because t h e control surface
was essentially rigid.

A sample result of t h e analysis i s shown i n f i g u r e 9. This is f o r

an all-movable 45' d e l t a plan form with i t s supporting a x i s a t 55.4 per-
cent of t h e r o o t chord. The Mach number i s 1.6. The v e r t i c a l coordinate
f o r t h e f i g u r e i s a s t i f f n e s s parameter cpl/2V where cr i s r o o t
chord, V is t h e airspeed, and 'ol i s t h e frequency of t h e f i r s t natu-
r a l mode. The horizontal coordinate i s t h e r a t i o of t h e frequencies of
t h e f i r s t and second n a t u r a l modes. The calculated f l u t t e r boundary i s
a smooth curve with a l a r g e peak near a frequency r a t i o of 1. The f l u t -
t e r region is below t h e curve and t h e peak indicates that more and more
s t i f f n e s s i s required t o prevent f l u t t e r as t h e f i r s t and second fre-
quencies approach one another.

Two experimental points are shown. The open point i s a n o - f l u t t e r

point and t h e dark point denotes that f l u t t e r occurred a f t e r t h e shaft
had been weakened by a c e r t a i n amount. The c r i t i c a l f l u t t e r speed should
then f a l l between t h e two points. The theory i s on t h e conservative s i d e
f o r both points and agrees very w e l l with experiment. Similar r e s u l t s
were obtained f o r several other arrowhead wings f o r M = 1.6 and t h e s e
are shown i n f i g u r e 10. The plan forms are i d e n t i f i e d by leading-edge
sweep and trailing-edge sweep. Sweepback i s p o s i t i v e . Location of t h e
supporting a x i s is given i n percent root chord. I n t h e upper left-hand
12

corner of f i g u r e 10 are results f o r t h e same configuration as t h a t j u s t

discussed i n f i g u r e 9 but with t h e p i t c h a x i s moved rearward t o t h e
63.3 percent root chord. I n t h e upper right-hand corner a r e r e s u l t s
f o r a wing with t h e leading edge swept back 450 and t h e t r a i l i n g edge
swept forward 13O. I n t h e lower left-hand corner a r e r e s u l t s f o r a
50' d e l t a wing and i n t h e lower right-hand corner a r e results f o r t h i s
50' wing with t h e t r a i l i n g edge swept back 15O. The theory i s seen t o
be consistently conservative and t o give results that a r e i n reasonable
agreement with t h e experimental data.

CONCLUDING RFMARKS

Several approaches t o t h e determination of t h e aerodynamic forces

on o s c i l l a t i n g wings have been b r i e f l y considered. These approaches
take i n t o account the a c t u a l plan form under consideration and t h e spe-
c i f i c modes of o s c i l l a t i o n . The few f l u t t e r calculations t h a t have
been discussed indicate that, when s u f f i c i e n t s t r u c t u r a l modes along
with appropriate aerodynamic forces are taken i n t o account, a reason-
able correlation between theory and experiment can be obtained. It is
emphasized t h a t the kernel-function approach discussed is a highly ver-
s a t i l e procedure and i s currently being systematically programed f o r a
high-speed calculator f o r both subsonic and supersonic speeds. Results
obtained thus far indicate that, a t l e a s t where l i n e a r theory i s appli-
cable, it should be a very powerful a i d t o t h e f l u t t e r analyst.
 REFERENCES

1. Garrick, I. E.: Nonsteady Wing Characteristics. Aerodynamic Compo-

nents of Aircraft at High Speeds. Vol. VI1 of High Speed Aerody-
namics and Jet Propulsion, sec. F, A. F. Donovan and H. R. Lawrence,
eds., Princeton Univ. Press, 1957, pp. 658-793.

2. Rmyan, Harry L., and Woolston, Donald S.: Method for Calculating the
Aerodynamic Loading on an Oscillating Finite Wing in Subsonic and
Sonic Flow. NACA TN 3694, 1956.

3. Falkner, V. M.: The Calculation of Aerodynamic Loading on Surfaces

of Any Shape. R. & M. No. 1910, British A.R.C., Aug. 1943.

4. Nelson, H. C., and Diederich, F. W.: A Numerical Integration Method

for Calculating Pressure Distributions on Wings of Arbitrary Plan
Form Deforming Harmonically in Subsonic Flow. (Prospective NACA
paper. )
5. Multhopp, H.: Methods for Calculating the Lift Distribution of Wings
(Subsonic Lifting Surface .Theory). Rep. No. Aero. 2353, British
R.A.E., Jan. 1950.

6. Watkins, Charles E., and Berman, Julian H. : Velocity Potential and

Air Forces Associated With a Triangular Wing in Supersonic Flow,
With Subsonic Leading Edges, and Deforming Harmonically According
to a General Quadratic Equation. NACA TN 3009, 1953.

7. Cunningham, H. J.: Lift and Moment on Thin Arrowhead Wings With

Supersonic Edges Oscillating in Symmetric Flapping and Roll and
Application to the Flutter of All-Movable Control Surfaces. (Pro-
spective NACA paper. )

514
14

.
MODES OF VIBRATION FOR A FLEXIBLE WING

Figure 1

SCHEME OF CALCULATION B Y BOX M E T H O D

SUPERSONIC E D G E S

Figure 2
515
 SCHEME OF C A L C U L A T I O N B Y K E R N E L
FUNCTION METHOD

SUPERSONIC S U B S O N IC

Figure 3

SOME DEFLECTION MODES TREATED BY FREQUENCY

EXPANSION METHOD

DEFLECTION VERTICAL PITCHING PARABOLIC

MODES: TRANS
X2 Y2

Figure 4
 SECTION LIFT AND PHASE ANGLES FOR TRANSLATION
AND PITCH ABOUT MIDCHORD
M=2; b0.375

-*b
1.2 r - EXPANSlON
o KERNEL

‘cz’h‘ .4

0 .5 I
0 .5 I

ICl,al
DEG

-I 0
0 .5 1 0 .5 I
SPAN SPAN

Figure 5

SECTION MOMENTS AND PHASE ANGLES FOR TRANSLATION

AND PITCH ABOUT MIDCHORD
M=2; k.0.375
- EXPANSION
o KERNEL
50

, % . d . 2 5 ~ +g::;L

IClqal .8

0
.”.-
‘.2p7)
0

5 I
-50
0 b I
SPAN SPAN

Figure 6
SECTION LIFT AND PHASE ANGLES FOR PARABOLIC MODES
M = 2; k = 0.375
-EXPANSION
0 KERNEL

DEG , 80 L
u
0 .5 I

.5
r
0 I
loo

0 .5
SPAN
I 0
- .5
SPAN
I

Figure 7

CALCULATED AND EXPERIMENTAL FLUTTER SPEED

OF A DELTA WING
M=0.85

1,000 -

FLUTTER 6oo
SPEED,
800
- \
-
MODE
-2

Fps 400-
EXPEWMENT

0 4 8 12 16~10~
DENSITY, S L U W FT

Figure 8
D3 .
0
. .
. . C ' c
. ,
..C
. . e
.. ' >*
1
.

18

45" DELTA ALL-MOVABLE CONTROL

PITCH AXIS AT 55.4% ROOT CHORD; M.1.6
.
v 7 ~ r

EXPERIMENT
NEAREST NO-FLUTTER
FLUTTER

0 I
FREQUENCY RATIO, 01/02

FLUTTER RESULTS FOR SOME ARROWHEAD

ALL - MOVABLE CONTROL SURFACES
M = 1.6
A = 45" A = 45"
e 7 0 [ - ATE = 00 A [ATE=-!~~
PITCH AXIS AT
65.3% ROOT CHOR 55.8% ROOT CHOR
.35

STIFFNESS
PARAMETER, "

PITCH AXIS AT
69.4% ROOT CH

0 I O I
FREQUENCY RATIO, W l / W 2
519
Figure 10
 .. b
.- .
4
4
I ..
d .
0

OG

FLUTTER AT VERY HIGH SPEEDS

By Harry L. Runyan and Homer G. Morgan
Langley Aeronautical Laboratory

SUMMARY

This paper is concerned with a discussion of some of the problems

of flutter and aeroelasticity that are o r may be important at high speeds.
Various theoretical procedures for treating high Mach number flutter are
reviewed. Application of two of these methods, namely, the Van Dyke
method and piston-theory method, is made to a specific example and com-
pared with linear two- and three-dimensional results. It is shown that
the effects of thickness and airfoil shape are destabilizing as compared
with linear theory at high Mach number. In order to demonstrate the
validity of these large predicted effects, experimental flutter results
are shown f o r two rectangular wings at Mach numbers of 6.86 and 3 . The
results of nonlinear piston-theory calculations were in good agreement
with experiment, whereas the results of using two- and three-dimensional
linear theory were not.

In addition, some results demonstrating the importance of including

camber modes in a flutter analysis are shown, as well as a discussion
of one case of flutter due to aerodynamic heating.

INTRODUCTION

This paper is concerned with some problems of flutter and aero-

elasticity at very high flight speeds. For this purpose high speeds
will be defined as starting in the Mach number range of 2.5 to 3 .

Some of the problems which are or may be important at high speeds

-
are discussed according to the forces in the aeroelastic problem aero-
dynamic, structural, and inertial. Under the aerodynamic part are:
(a) Nonlinear effect of airfoil shape, thickness, and angle of
attack: There appears to be a very large effect of these factors on
the flutter speed, which is discussed subsequently.

(b) Effect of shocks: Information is lacking and constitutes an

area requiring research effort.
52.3
 ,
c
. 0
- I

( c ) Boundary l a y e r and viscous e f f e c t s : Here again information i s

lacking, (of course, even i n the low speed case),but, with t h e t h i c k
boundary l a y e r s encountered a t high-speed flight, t h e dynamics of the
boundary layer could become important, p a r t i c u l a r l y , i n t e r a c t i o n s of
t h e boundary layer w i t h shocks.

( d ) Plan form: Some of t h e new plan forms having sweep angles of

t h e order of 7 5 O w i l l pose s p e c i a l problems with respect t o unsteady
a e r o w a m i c s , and there arises t h e d i f f i c u l t problem of studying and
developing t h e o r i e s t h a t w i l l take i n t o account t h e e f f e c t of both air-
f o i l shape and aspect r a t i o .

(e) Controls: Controls have always been a source of trouble f o r

t h e f l u t t e r analysts. I n t h e Mach number range of 10 t o 20, t h e type
of control t h a t w i l l prove t o be s a t i s f a c t o r y i s not known. But from
p a s t experience, whatever type of aerodynamic control, i f any, i s found
t o be s a t i s f a c t o r y , it w i l l probably c o n s t i t u t e a f l u t t e r problem.

Structures required f o r high-speed f l i g h t present another. a r e a of

d i f f i c u l t y . Some of t h e problems are:

( a ) Aerodynamic heating: An example of aerodynamic heating r e l a t i n g

t o f l u t t e r i s b r i e f l y discussed.

( b ) Panels and heat shields: For t h e flat-bottomed highly heated

a i r c r a f t now envisioned f o r high-speed f l i g h t , it appears t h a t t h e
f l u t t e r of panels and heat shields w i l l be a very real problem. Under
high-temperature conditions, buckling w i l l probably occur and w i l l
require nonlinear treatment.

( c ) Plan form: For wings of high aspect r a t i o , t h e d i s t o r t i o n s of

t h e w i n g involved mainly a t w i s t i n g and bending of t h e wing, s o t h a t the
elementary concepts of "beamology" could be used. However, the low-
a s p e c t - r a t i o wings now being considered behave more l i k e plates and
involve a large amount of chordwise d e f l e c t i o n .

I n e r t i a l force i s t h e t h i r d type of f o r c e i n t h e a e r o e l a s t i c prob-

lem. The a i r c r a f t s t r u c t u r a l weight i s decreasing i n comparison with
the weight of the f u e l , p a r t i c u l a r l y with regard t o missiles. Conse-
quently, such nonlinear problems as f u e l sloshing and swirl a r e becoming
exceedingly important.

Three of these problems w i l l be discussed: t h e e f f e c t of a i r f o i l

shape, s t r u c t u r a l plan form, and aerodynamic heating.
 ...
.*. <
( a
*
m
m
a
a
~ a , * - a _ _ -- --* *
,. a
a
.
3

A aspect r a t i o

a speed of sound

b half chord, i n .

CP pressure coefficient, P - Po
9

IU moment of i n e r t i a about e l a s t i c a x i s

M Mach number

m mass of wing per u n i t of span length, lb- sec

in. 2
P pressure a t point x,y

PO pressure i n undisturbed stream

9 dynamic pressure

rU radius of gyration, ru2 = Ia

mb2
t thickness r a t i o

W v e r t i c a l induced v e l o c i t y o r downwash

v stream velocity, f t /sec

X, Y Cartesian coordinates

U angle of a t t a c k

Y s p e c i f i c heat r a t i o

9-l bending frequency, radians/sec

“h; t o r s i o n a l frequency, radians /sec

f i r s t bending frequency, radians/sec

Fr mass r a t i o , -m
npb2
4

P f l u i d density

cp velocity p o t e n t i a l

Subscripts:

L linear

NL nonlinear

ANALYTICAL METHODS

This section i s concerned with a b r i e f description of t h e t h e o r e t i c a l

methods available f o r high Mach number studies. The complete nonlinear
p a r t i a l d i f f e r e n t i a l equations f o r t h e p o t e n t i a l and the pressure coeffi-
c i e n t a r e shown by

‘pL+‘pNL=o

and can be broken down i n t o a l i n e a r p a r t plus a nonlinear p a r t .

The l i n e a r s o l u t i o n now i n general use i s obtained from equations (1)
and ( 2 ) by s e t t i n g the nonlinear p a r t equal t o zero, as given by

cq=o (3)

cp = Cp,L

and then solving the equations with s u i t a b l e boundary conditions.

Several approximate methods a r e available for obtaining nonlinear

solutions. The first of these i s the solution of Van ‘Dyke ( r e f . 1) H e .
f i r s t eliminated t h e third-order terms from these two nonlinear equations
( t h i s procedure, i n e f f e c t , eliminates the e f f e c t of f i n i t e shocks) and
then inserted the solution f o r t h e l i n e a r equation ( 3 ) i n the nonlinear
p a r t of equation (1). This procedure resulted i n a l i n e a r p a r t i a l d i f f e r -
e n t i a l equation plus a known function as

523
 .*.
B . 8
. . m
m i )
1 ) .
~

8
, ,’
r
.
. $
,
.-
8
L.

By a laborious technique, Van Dyke then solved these equations f o r t h e

pressure on s p e c i f i c a i r f o i l s .

Another nonlinear method i s t h e so-called p i s t o n theory. T h i s pro-

cedure w a s o r i g i n a l l y suggested by Hayes ( r e f . 2 ) , w a s used by L i g h t h i l l
( r e f . 3 ) t o check the r e s u l t s of Van Dyke a t high Mach number, and l a t e r
was elaborated on and applied t o t h e f l u t t e r problem by Ashley and
Zartarian ( r e f . 4) . The advantage of piston theory i s i t s u t t e r s i m -
p l i c i t y as compared with other theories. The pressure c o e f f i c i e n t i s
e a s i l y derived on the basis of a piston moving i n a one-dimensional
channel. The expression f o r t h e pressure c o e f f i c i e n t i s given i n

cp = 2kk) + ($2 + y+l

12 M($ + .. 1
where w i s t h e instantaneous v e r t i c a l velocity of a point on t h e wing,
and V i s t h e stream velocity. Note that, as M i s increased, the
f i r s t term would become l e s s important, and the higher order nonlinear
terms would begin t o take an added importance.

Another method i s use of the Newtonian concept. I n t h i s procedure

it i s assumed t h a t t h e flow s t r i k i n g the exposed surface i s compressed
t o a very t h i n boundary l a y e r and t h e force exerted on t h e a i r f o i l i s
due t o t h e component of momentum perpendicular t o t h e surface. The
r e s u l t i n g pressure c o e f f i c i e n t

1 2 w*,4 + ..
cP = 2/(!)2 - 3 (7,
was obtained by expanding cos2 M. For a curved surface o r an o s c i l l a t i n g
v
surface, additional terms due t o centrifugal force could be added. Note
t h a t the first t e r m i s missing a s compared with p i s t o n theory and t h e
coefficient of the squared t e r m has a f a c t o r of 1 as compared with 0.6
f o r 7 = 1.4. Later, use w i l l be made of t h e Van Dyke and piston-theory
solution.

APPLICATION TO SPECIFIC MAMPLES

Some applications and comparisons of t h e various t h e o r i e s are given.

I n figure 1 are shown t h e r e s u l t s of calculating t h e f l u t t e r of a rectan-
gular wing of panel aspect r a t i o 1.5 throughout t h e Mach number range
of 1 . 3 t o 10. The a i r f o i l section w a s 63 series, tapering from 4 percent
a t the r o o t t o 3 percent a t t h e t i p . Four theories have been used:
l i n e a r two-dimensional theory, l i n e a r three-dimensional theory, nonlinear

52%
6

p i s t o n theory, and Van Dyke theory. The r e s u l t s are p l o t t e d a g a i n s t t h e

s t i f f n e s s - a l t i t u d e parameter, 6. The f l u t t e r region i s below t h e
curves. Constant-altitude l i n e s are horizontal and constant-dynamic-
pressure l i n e s are radial l i n e s emanating from the o r i g i n . There are
four points of i n t e r e s t i n t h i s p l o t . F i r s t , the l a r g e difference between
the l i n e a r t h e o r i e s and t h e two t h e o r i e s which include the e f f e c t of
thickness a t t h e higher Mach numbers. This e f f e c t i s primarily due t o
a forward s h i f t i n the center of pressure due t o a i r f o i l shape whereas
the center of pressure f o r t h e two-dimensional l i n e a r theory i s f i x e d a t
the 70-percent chord and t h e forward s h i f t of the center of pressure i n
t h e three-dimensional t i p does not p r e d i c t as much forward movement as the
nonlinear t h e o r i e s . Another point i s the agreement of t h e more compli-
cated Van Dyke theory with t h e simpler and more r e a d i l y used p i s t o n
theory a t t h e higher Mach number. A t h i r d point of i n t e r e s t i s the cross
over of t h e two- and three-dimensional theory a t M = 1.6. It has usually
been assumed that inclusion of t h e three-dimensional e f f e c t s i s a
r e l i e v i n g e f f e c t when compared w i t h t h e two-dimensional theory. This
i s usually t r u e a t the lower Mach numbers but it i s not necessarily t r u e
a t the high Mach numbers. A f o u r t h point i s that the e f f e c t of a i r f o i l
shape and thickness i s d e s t a b i l i z i n g . For instance, a t an a l t i t u d e cor-
responding t o a value of t h e ordinate of 3.3, nonlinear theory i n d i c a t e s
t h a t f l u t t e r would be experienced a t a Mach number of 5 , whereas the
l i n e a r theory would p r e d i c t t h e a i r f o i l t o be f l u t t e r f r e e .

I n f i g u r e 2 i s shown t h e calculated e f f e c t of thickness and a i r f o i l

shape on f l u t t e r a t M = 10 obtained by using nonlinear p i s t o n theory.
The s t i f f n e s s - a l t i t u d e parameter i s shown p l o t t e d against the r a t i o of
bending t o t o r s i o n frequency. The f l u t t e r region i s below t h e curve.
Curves are presented f o r a f l a t p l a t e , a 4-percent wedge, and a 4-percent
biconvex a i r f o i l . L e t us focus our a t t e n t i o n on the curves f o r t h e f l a t
p l a t e or zero-thickness a i r f o i l and t h e biconvex a i r f o i l . For low-
frequency r a t i o , t h e zero-thickness a i r f o i l gives no f l u t t e r s o l u t i o n
whereas t h e biconvex a i r f o i l shows a d e f i n i t e f l u t t e r solution. A s t h e
frequency r a t i o i s increased, however, the curves tend t o approach each
other and a t - oh = 1 . 2 they a c t u a l l y cross. That is, the e f f e c t of
%
thickness i s d e s t a b i l i z i n g f o r low values of the frequency r a t i o and
s t a b i l i z i n g for high values, a t least f o r t h i s case. Note t h a t t h e wedge
has a shape similar t o the f l a t p l a t e except it i s s l i g h t l y d e s t a b i l i z i n g .

I n figures 1 and 2 are shown some r a t h e r l a r g e and disturbing e f f e c t s

of thickness and a i r f o i l shape i n reducing t h e f l u t t e r speed. The ques-
t i o n i s then "are these l a r g e e f f e c t s , i n f a c t , true." I n an attempt
t o answer t h i s question two wings have been f l u t t e r e d a t high speed. The
frequency r a t i o selected f o r these wings w a s d e l i b e r a t e l y chosen so that
as w i d e a spread as possible between t h e zero-thickness and t h e thickness
s o l u t i o n could be obtained. For these cases t h e frequency r a t i o was
,I
i 52.5
 7

approximately 0.35. Although t h e parameters f o r figure 2 are not t h e

same as those f o r t h e experiment, t h e t r e n d s are t h e same and a t a fre-
quency r a t i o of about 0.35 t h e r e i s quite a difference between t h e zero-
thickness and the thickness cases.

COMPARISON OF MPERIMENTAL AND T€BORECICAL RESULTS

F l u t t e r a t Mach Numbers of 6.86 and 3.0

Results f o r two rectangular wings are shown i n figures 3 and 4,

each having a panel aspect r a t i o of 0.8. One i s an 11-percent double-
wedge s e c t i o n and t h e other i s a 4-percent f l a t wing. The p r o p e r t i e s
of these wings a r e given i n t h e following t a b l e :

11-percent wedge 4-percent p l a t e

b . . ....... 2 -55 2 -57

m . . . . . . . . . 0.0001276 O.OOOl.27
ra2 ........ 0.251 o .269
xo ........ 0.467 0.46
........ 0 -0545 0.0745
........ 110 .g 106
........
'"tL 314 322
VatM=3.0 . . . 2,110 2,120
VatM=6.86.. . 3,290 . 3,255

(The t o r s i o n mode f o r both wings w a s taken as u n i t y across t h e span.

The bending mode f o r t h e 11-percent wing w a s taken as f h = 0.23 + 0 . 1 9 2 5 ~
and f o r t h e 4-percent w i n g as f h = 0.335 + 0 . 1 8 6 ~ where x v a r i e s
fYom 0 t o 4 inch&.) The wings were very r i g i d and were mounted on
f l e x i b l e s h a f t s so t h a t , i n e f f e c t , they corresponded t o all-movable
controls. The results are again p l o t t e d as t h e s t i f f n e s s - a l t i t u d e
parameter against Mach number. The experimental results are shown as
s o l i d points and were obtained a t Mach numbers of 6.86 and 3 i n t h e
Langley 11-inch hypersonic tunnel and t h e Langley 9- by 18-inch super-
sonic f l u t t e r tunnel, respectively. Let us examine f i r s t t h e double
wedge. The s o l i d l i n e i s the r e s u l t of using nonlinear p i s t o n theory and
fairly good agreement i s indicated with t h e experiment. The two-
dimensional, zero-thickness method gave no solution. The three-
dimensional l i n e a r case indicated a f l u t t e r - f r e e wing a t a Mach number
of 6.86 b u t gave a solution a t a lower Mach number as indicated. For
t h e &-percent p l a t e , simi&r -@ston theory and
526
experiment was obtained. Again, two-dimensional zero thickness gave no
solution and indicated t h e wing t o be f l u t t e r f r e e ; whereas, inclusion of
t h e three-dimensional t i p e f f e c t gave a solution as indicated. The value
of reduced frequency k f o r the M = 6.86 t e s t was - 1 and a f i r s t - o r d e r
70
theory i n frequency such as the p i s t o n theory should be s a t i s f a c t o r y .

Thus, it appears that the detrimental e f f e c t of thickness on f l u t t e r

as predicted by p i s t o n theory i s i n f a c t t r u e and that nonlinear theories
must be used a t the high f l i g h t speeds. One i n t e r e s t i n g f a c t i s that
l i n e s drawn through t h e experimental points i n t e r s e c t t h e origin; thus
a constant "q" f l u t t e r v a r i a t i o n i s indicated.

F l u t t e r of Delta Wings

Now l e t us t u r n our a t t e n t i o n t o some f l u t t e r calculations of t w o

low-aspect-ratio cantilever wings. I n figure 5 t h e s t i f f n e s s - a l t i t u d e
coefficient has been p l o t t e d against Mach number f o r 45' and 60' d e l t a
wings. The wings were f l a t p l a t e s with beveled leading and t r a i l i n g edges.
The c i r c u l a r points a r e the experimental r e s u l t s from reference 5
and the s o l i d and dashed l i n e s a r e a n a l y t i c a l result-s. Piston theory
w a s used f o r t h e aerodynamic input. A modal type of analyses which was
based on experimentally measured mode shapes was used. Since these
modes had a l a r g e amount of deflection i n t h e chord direction, it did
not seem.that t h e d e f l e c t i o n curves could be approximated by t h e usual
procedure of bending and twisting of a straight l i n e . Hence, a n a l y t i c a l
curves were f i t t e d t o the experimental deflection curves a t each of
10 spanwise s t a t i o n s f o r use i n t h e analysis. The r e s u l t s a r e shown by
s o l i d l i n e s and show f a i r l y good agreement with experiment. I n order
t o assess the e f f e c t of chordwise deflection, t h e camber was a r b i t r a r i l y
eliminated from each mode and then recalculated. The results a r e shown
by the dashed l i n e s . For t h e 600 wing, the curve w a s s h i f t e d over t o
t h e nonconservative side, whereas f o r the 4 5 O wing a very wide divergence
i s found. Thus the importance of including the camber deflection i n the
analysis of a low-aspect-ratio w i n g i s demonstrated.

F l u t t e r Due t o Aerodynamic Heating

/
Another well-known problem of high speed i s the e f f e c t of aero-
dynamic heating. With regard t o f l u t t e r , t h e main e f f e c t of aerodynamic
heating i s t o cause a loss i n t o r s i o n a l s t i f f n e s s , p a r t i c u l a r l y during
t r a n s i e n t conditions. A s o l i d duralumin wing has been t e s t e d a t a Mach
number of 2 i n t h e p r e f l i g h t j e t of the Langley P i l o t l e s s A i r c r a f t
Research S t a t i o n a t Wallops Island, Va. Two runs were made, a cold run
during which the.wing d i d not f l u t t e r and a hot run during which t h e wing
f l u t t e r e d . This phenomena can be explained with the a i d of figure 1. These
i
L 527 .
LG
9

calculations apply t o t h i s heated wing. I n the cold condition t h e value

of t h e s t i f f n e s s - a l t i t u d e parameter at a Mach number of 2 i s 3.07 and i s
well i n t h e s t a b l e region. During t h e fast start of t h e tunnel, t h e
leading and t r a i l i n g edges heated up much more r a p i d l y t h a n t h e t h i c k e r
center section; t h i s condition causes a momentary l o s s i n t o r s i o n a l s t i f f -
ness. Thus t h e t o r s i o n a l frequency was reduced; t h e s t i f f n e s s - a l t i t u d e
parameter i s correspondingly reduced and would follow a v e r t i c a l l i n e t o
an i n t e r s e c t i o n of t h e f l u t t e r curve. Calculations of t h e l o s s i n t o r -
s i o n a l s t i f f n e s s have been made and show a reduction i n t h e t o r s i o n a l fre-
quency by 50 percent which i s s u f f i c i e n t t o i n t e r s e c t the f l u t t e r region.
Thus, f l u t t e r which has been induced by aerodynamic heating, a t least f o r
a simple s o l i d wing, can be calculated.

CONCLUDPNG REMARKS

New f l u t t e r and a e r o e l a s t i c problems w i l l appear a, high f l i g h t

speeds. Configurations dictated by high-speed requirements will probably
a l s o e x h i b i t new problems i n t h e lower speed ranges. An e s s e n t i a l fea-
t u r e of many of these problems i s t h e i r inherent nonlinearity. For
accurate f l u t t e r prediction, inclusion of these nonlinearities, such as
the e f f e c t of a i r f o i l thickness and shape, i s a necessity.

REFEREXCES

1. Van Dyke, Milton D.: Supersonic Flow Past Oscillating A i r f o i l s

Including Nonlinear Thickness Effects. NACA Rep. 1183, 19%.
(Supersedes NACA TN 2982.)

2. Hayes, Wallace D.: On Hypersonic Similitude. Quarterly Appl. Math.,

vol. V, no. 1, Apr. 1947; pp. 105-106.

3. L i g h t h i l l , M. J.:
Oscillating A i r f o i l s a t High Mach Number. Jour.
Aero. Sci., vol. 20, no. 6, June 1953, pp. 402-406.

4 . Ashley, Holt, and Zartarian, Garabed: -

Piston Theory A New Aero-
dynamic Tool f o r t h e Aeroelastician. Jour. Aero. Sci., vol. 23,
no. 12, Dec. 1956, pp. 1109-1118.

5 . Tuovila, W. J .
, and McCarty, John Locke: m e r i m e n t a l F l u t t e r
Results f o r Cantilever-Wing Models a t Mach Numbers up t o 3.0.
NACA RM L55El1, 1955.

F
Y
10

FLUTTER FOR VARIOUS AERODYNAMIC THEORIES

4-

bw,&
a 3.

2.

I-
FLUTTER REGION
v
I I I I 1
0 2 4 6 8 10
M

Figure 1

CALCULATED EFFECT OF AIRFOIL SHAPE

-_----4 Oo
/ BfCONVEX
I I I I I
0 .4 .0 1.2 I .6 2 .o
Wh/ w a

529
 FLUTTER OF A DOUBLE WEDGE WING AT HIGH SPEED
E.A.=46.7%; c.G.=49.4% i A z 0 . 8
--=zzm-
t= 11%
6-

-
4-
PISTON THEORY ( t = 1 1 % )

a
-
2-

-
EXPERIMENT -? [-;DIMEN. ( t = 0)
--_ ---_
2-DIMEN. ( t = 0 )
I I t r- --,I I

FLUTTER OF A THIN WING AT HIGH SPEED

E.A. = 46 Oi'o ; C.G. - 5 0 % ;A = 0.8

t = 4 O/O
6

EXPERIMENT
-7

I I I I I I I / I
0 I 2 3 4 5 6 7
M

Figure 4

'. " .
 FLUTTER OF DELTA WINGS
2.c

60°

I .5

I .c
bwlfi
a

A
-
0 EXPERIMENT 0 /'
.5
-WITH CAMBER /
'
- --- WITHOUT CAMBER
':FLUTTER
REGION
0 I 2 3 f 3
M

.... ..-anarr
.
... .
 FWTTER OF WINGS WITH AND WITHOUT EXTEXNAL STORE3

AT TRANSONIC AND SUPERSONIC SPEEDS

By Laurence K. Iaftin, Jr., and William T. Lauten, Jr.
Langley Aeronautical Laboratory

INTRODUCTION

Much of the available experimental information on wing flutter at

transonic and supersonic speeds deals with plain wings of moderate aspect
ratio and taper ratio. In the present paper, some aspects of the flutter
behavior at transonic and supersonic speeds of delta wings and highly
tapered swept wings, and of wings with mass balance and external stores
w i l l be discussed. The Mach number range to be considered is from 0.8
to 3.0. The majority of the material to be presented was obtained
recently, is rather fragmentary, and should be regarded as suggestive of
possible leads or hints rather than as a basis for forming broad conclu-
sions or design rules. Trends for the plain wings will be discussed
first and then some results for wings with mass balance and stores will
be presented.

A aspect ratio
a speed of sound, ft/sec
b streamwise root semichord, ft

C streamwise chord, ft

M Mach number

9 dynamic pressure, lb/sq ft

dynamic pressure at flutter of wing without'stores or

90 bodies, lb/sq ft

s, dynamic pressure at flutter of wing with stores or bodies,

lb/sq ft
 , .* ,
*.
..I L
.e
e.

A angle of sweep, deg

A taper ratio

v nondimensional density parameter (ratio of mass of exposed

semispan wing to mass of air contained in the frustum of
a cone whose upper and lower base diameters are equal to
strewise tip and root chords)

% frequency of predominantly torsion mode, radians/sec

cue uncoupled store pitching frequency, radians/sec

RESULTS AND DISCUSSION

Plain-Wing Results

The flutter trends to be discussed for wings without stores or

balance weights are shown in figure 1. This is a plot of the stiffness-
altitude parameter baufi Gainst Mach number. The flutter region is
a
the region below the curve. Constant altitude is indicated by a hori-
zontal line, constant dynamic pressure by a straight line through the
origin. An increase in altitude for a gi-renwing corresponds to an
increase in the stiffness-altitude parameter. Wends are presented for
a 60° delta wing, a 45’ swept wing with a taper ratio of 0.2 and, for
comparison, a 4 5 O swept wing with a taper ratio of 0.6. The plan forms
of the wings are indicated by the sketches on the left. The curves shown
are based on data obtained in the 26-inch Langley transonic blowdown
tunnel at Mach numbers from 0.8 to 1.4 and in the Langley 9- by 18-inch
supersonic flutter tunnel from 1.3 to 3.0. The trends shown in this
figure are thought to be particularly interesting in that the configura-
tions investigated at transonic and supersonic Mach numbers are nearly
the same and an acceptable basis for evaluating the relative severity
of the flutter problem at transonic and supersonic Mach numbers is thus
provided, at least for the plan forms considered. &e trends show the
existence of a critical flutter region mound Mach number 1.0, after
which the three wings are flutter-free at altitudes equal to or greater
than that necessary to avoid flutter at transonic speeds, until the Mach
number reaches values of the order of 2.0. No especial significance
should necessarily be attached to the Mach number around 2.0, at which
value the stiffness-altitude parameter becomes equal to the critical
value around Mach number 1.0, except insofar as the particular configura-
tions considered in the figure are concerned. Investigations at both
transonic and supersonic Mach numbers indicate that the values of the
stiffness-altitude parameter corresponding to flutter can be al+,eredby

I
ji 533
 3.
1 . 3
0. ..
3

changes in wing plan form, center-of-gravity position, and perhaps other

parameters. The trends of this figure, however, emphasize the necessity
for careful consideration of the flutter boundaries at both transonic
and supersonic speeds on any new high-speed aircraft.
An interesting detailed feature of the results shown in figure 1
is the behavior of the highly tapered wings at Mach nurdbers around 1.0.
In contrast to the results for the wing with taper ratio of 0.6, the
value of the stiffness-altitude parameter for the highly tapered wings
is seen to drop suddenly to a larer value aad then increase .slowly as
the Mach number is increased. Thus, a large stable region for these two
wings is indicated at Mach nuuibers somewhat in excess of 1.0. The sudden
decrease in stiffness-altitude parameter has been found to accompany an
increase in flutter frequency by a factor of approximately 2, with the
wing fluttering at about the natural torsional frequency. This change
in flutter mode is felt to be of interest because the nature of a ''fix"
required to suppress flutter may be different for the different flutter
modes. Tne combination of parameters responsible for the favorable
decrease in the stiffness-altitude parameter with the accompanying
increase in the flutter frequency is not understood; therefore, the range
of application of the results cannot be assessed at this time.

Effects of Mass Balance

The results of some recent studies at transonic speeds of the effects

of mass balance will now be discussed. The wing plan form employed was
the same 4
5' swept wing with taper ratio of 0.2 discussed in figure 1.
The configuratims investigated are shown at the top of figure 2. The
shaded areas shown at the leading edge represent the two positions of
balance mass for which experiments were made. Balance masses of 6 and
12 percent of the wing mass were attached at each position and in each
configuration the mass extended over 25 percent of the span. The results
obtained from the four mass-balance configurations are sham in the bottom
part of the figure. These results are plotted in the form of dynamic pres-
sure required for flutter against Mach nurdber. The area above the curve
is the flutter region. B e solid line represents data taken from figure 1
for the wing without balance. The dashed line is a fairing of data taken
for all balance conditions since no significant difference existed between
the data for the four configurations. The results show that in the crit-
ical flutter range between Mach numbers of 0.9 and 1.05, where the very
pronounced peaks in stiffness-altitude parameter were evident in figure 1,
the leading-edge balance mass increased the dy&c pressure required
for flutter by as much as 100 percent. The hatched area labeled "no
flutter" indicates that flutter could not be obtained beyond Mach num-
ber 1.1for azzy of the mass-balance configurations up to the maximum
dynamic pressure employed in the investigation. Obviously, substantial
improvements in the transonic flutter chasacteristics of this wing can

534
4

be obtained by the use of mass balance. In t h i s connection, it is per-

haps of some i n t e r e s t t o point out t h a t an increase of approximately
100 percent i n the plain-wing t o r s i o n a l s t i f f n e s s would be required i n
order t o achieve the e f f e c t produced by t h e addition of 6 percent of
t h e wing weight as a mass balance. The f a c t t h a t decreasing t h e amount
of balance mass by a f a c t o r of 2, t h a t is, f r o m 12 percent t o 6 percent
wing mass, had no e f f e c t on the improvement shown leads t o the question
of j u s t how l i t t l e mass balance could be used i n order t o obtain a cer-
t a i n desired increase i n dynamic pressure f o r f l u t t e r . There i s no
general c r i t e r i o n f o r the optimum amount and position of the,balance mass
and each case w i l l probably have t o be a subject f o r individual study.

Wings With External Stores

I n contrast t o the work j u s t discussed on d i s t r i b u t e d mass, some

exploratory studies have been made a t Mach numbers from 0.8 t o 1.3 on
the e f f e c t on f l u t t e r of concentrated masses i n t h e form of external
stores. This work was aimed a t gaining some i n s i g h t i n t o the nature of
the problems encountered, and, if possible, w h a t properties of the s t o r e
might be s i g n i f i c a n t t o t h e f l u t t e r problem. Three configurations
studied a t transonic speeds are shown i n f i g u r e 3. !be wing was swept
back 45O, had no taper, and had an aspect r a t i o of 4.0. The three s t o r e s
employed were of the underslung type mounted on pylons, and had an aero-
dynamic shape consistent with current transonic design practice. The
s t o r e s are labeled as A, B, Elnd C. Stores A and B were nearly identical,
except f o r a change i n center-of-gravity position, and had a weight of
about 70 percent of the wing weight-. Store C w a s 40 percent heavier
(approximately equal t o the wing weight). The center of gravity of
s t o r e A was located a t the 38-percent-chord point a t midsemispan and
t h e center of gravity of s t o r e B was at the 12-percent-chord point, a l s o
a t the midsemispan. The center of gravity of s t o r e C w a s 25 percent
ahead of the leading edge a t about 80 percent of t h e semispan.

The experimental r e s u l t s f o r the various s t o r e configurations are

s h a m i n figure 4 i n which the r a t i o of the dynamic pressure a t f l u t t e r
of the wing w i t h s t o r e s t o t h a t of the wing without s t o r e s i s plotted
against Mach number. No s i g n i f i c a n t difference was found i n the r e s u l t s
f o r s t o r e s A and B, and t h e s o l i d l i n e i n this f i g u r e represents a f a i r i n g
of both s e t s of data. It may be seen t h a t the dynamic pressure f o r f l u t t e r
of the wing w i t h stores A and B l i e s between 70 and 80 percent of t h a t f o r
t h e p l a i n wing i n the range of Mach number from 0.8 t o about 1.3. The
dashed l i n e indicates the r e s u l t s obtained from the t e s t s of configura-
t i o n C. As can be seen, the dynamic pressure f o r f l u t t e r a t M = 0.95
has been increased from about 75 percent t o about 90 percent of t h a t f o r
t h e wing without stores as a result of t h i s change i n t h e weight p o s i t i o n j
outside the Mach number range of 0.88 t o 1.05, the. dynamic pressure
required f o r f l u t t e r i s l a r g e r f o r the wing w i t h t h e s t o r e than f o r the
wing without the s t o r e . +.
!G 5

The f a c t that forward movement of the s t o r e center of gravity had

no e f f e c t on the results obtained f o r s t o r e s A and B i s perhaps not too
surprising when consideration i s given t o the nddal p a t t e r n s f o r t h e
two s t o r e configurations. It may be seen in figure 5 that the s h i f t i n
the center of gravity makes only a small change i n the location of the
nodal p a t t e r n and t h e uncoupled frequencies would not be expected t o be
much d i f f e r e n t . In regard t o configuration C, it w a s assumed that a
large forward movement of the predominantly t o r s i o n node l i n e might, as
previous investigators have suggested, be i n d i c a t i v e of a favorable
coupling e f f e c t w i t h a corresponding increase i n t h e dynamic pressure of
f l u t t e r . Consequently, a vibration survey w a s made w i t h the s t o r e located
i n numerous spanwise and chordwise positions. Configuration C w a s found
t o give the type of node l i n e desired and yielded an increase i n t h e
f l u t t e r dynamic pressure. Any differences i n t h e e f f e c t s of t h e aero-
dynamic forces and moments of the s t o r e on the f l u t t e r condition as the
s t o r e w a s moved from the location of A and B t o C were assumed t o be of
second-order importance. The results shown i n t h i s f i g u r e i n d i c a t e t h a t ,
j u s t as i s the case a t lower speeds, the addition of s t o r e s should not
necessarily cause a serious f l u t t e r problem a t transonic speeds and may
perhaps be employed as a m e a n s f o r a l l e v i a t i n g the problem.

A s h o r t exploratory study has a l s o been made a t transonic speeds of

the e f f e c t s of the introduction of f l e x i b i l i t y i n t o the s t o r e mount s o
as t o give a pitching degree of freedom. lPle configuration and some of
t h e results obtained are shown i n f i g u r e 6. The same untapered wing used
i n connection with f i g u r e s 4 and 5 w a s employed i n these studies. A
schematic diagram of the s t o r e mount i s shown a t t h e right of the figure.
The s t o r e w a s mounted close t o the'under surface of t h e wing and w a s
connected t o the wing by a spring i n such a manner t h a t t h e s t o r e w a s
e s s e n t i a l l y r i g i d i n yaw and s i d e bending b u t had f l e x i b i l i t y i n p i t c h .
The spanwise position of the s t o r e w a s chosen t o correspond approximately
t o t h e plain-wing second bending node l i n e which w a s a t about 75 percent
of the semispan. 'phe s t o r e mass w a s about the same as that of t h e p l a i n
semispan wing. The center of gravity w a s located a t 3 l p e r c e n t of t h e
wing chord. The d a t a shown are i n the form of the r a t i o of the dynamic
pressure required f o r f l u t t e r f o r t h e wing w i t h s t o r e t o t h a t f o r the
wing without store, p l o t t e d against the r a t i o of the uncoupled s t o r e
pitching frequency t o the uncoupled t o r s i o n frequency % of the
p l a i n wing. The highest value of t h i s r a t i o shown here represents about
as near a r i g i d condition as could be obtained. The c i r c l e d points
represent values of the frequency r a t i o f o r which d a t a w e r e obtained.
The results a r e f o r a Mach number of approximately 0.85. The trend of
the d a t a indicates a d e f i n i t e optimum value of the p i t c h f l e x i b i l i t y .
For the p a r t i c u l a r configuration studied, the optimum value of the r a t i o
of pitching frequency t o t o r s i o n frequency was about 0.6, w i t h reductions
i n t h e dynamic pressure required f o r f l u t t e r accompanying e i t h e r increases
or decreases i n the f l e x i b i l i t y of t h e s t o r e mount. The results sham i n
t h i s figure i n d i c a t e the p o s s i b i l i t y of lessening the f l u t t e r problem w i t h
s t o r e s a t transonic speeds b i l i t y i n the s t o r e mount.
5%
.. .
a .
m
-e
e
-. .
0
e
2
e
, 0
- 1
.
0 ..
e.

Studies of the e f f e c t s of external bodies on f l u t t e r a t supersonic

speeds are being- carried out i n the Langley 9- by 18-inch supersonic
f l u t t e r tunnel. The plan forms f o r w h i c h these s t u d i e s are being made
include those shown i n figure 1. Some of the r e s u l t s f o r the 600 d e l t a
wing and the 4 5 O swept wing with taper r a t i o of 0.2 w i l l be presented.
The configurations investigated on the d e l t a wing are shown i n figure 7.
The symbols indicate the I 2 positions of the body center of gravity f o r
which tests were made. The positions were -25, 0, 25, and 50 percent of
the wing chord. The spanwise positions were 1/4, 1/2, and 3/4 of the
semispan. The s i z e and shape of the bodies are indicated by the outline
drawings a t the 50-percent-semispan s t a t i o n . The mass of the body w a s
slightly more than one-half of the wing semispan &ass. In all cases the
body was equal i n length t o the wing chord at the 50-percent-semispan
position and had a fineness r a t i o of 10. The shape of t h e body was r a t h e r
crude; however, the slope of the l i f t and moment curves on a body of such
high fineness r a t i o i s r e l a t i v e l y small and i s not affected t o any great
extent by d e t a i l e d changes i n the shape. It w a s therefore assumed t h a t
t h e shape of the bodies was r e l a t i v e l y unimportant i n determining t h e
f l u t t e r c h a r a c t e r i s t i c s . A f e w t e s t s with bodies of cleaner aerodynamic
shape have indicated that t h i s assumption i s valid.

The e f f e c t of t h e center-of-gravity location on t h e f l u t t e r character-

i s t i c s of the 600 d e l t a wings i s shown f o r Mach numbers 1.3 and 2.0 i n
figures 8 and 9, respectively. The ordinate i s again the r a t i o of dynamic
pressure necessary f o r f l u t t e r with body attached t o t h a t necessary f o r
f l u t t e r w i t h the p l a i n wing. The abscissa is the f r a c t i o n of t h e semispan.
The c i r c l e s indicate the center o f - g r a v i t y ahead of the leading edge, the
squares indicate the leading-edge center-of-gravity position, the diamonds
indicate the 25-percent-chord position, and the t r i a n g l e s indicate the
50-percent-chord position. The open points indicate that no f l u t t e r was
obtained f o r these body locations up t o the maximum dynamic pressure
possible i n t h e tunnel. Points above the s o l i d l i n e a t a value of qs/Q
of 1 indicate an increase i n the dynamic pressure required f o r f l u t t e r
r e s u l t i n g from i n s t a l l a t i o n of the external body. Large increases i n the
dynamic pressure required f o r f l u t t e r are seen t o accompany many body
locations a t Mach numbers of both 1.3 and 2.0. Although the r e s u l t s are
not completely consistent, the far-forward, 3/4-semispan position of the
center of gravity appears t o be the most favorable.

The configurations investigated on the 45' swept wing with the taper
r a t i o of 0.2 a r e shown i n figure 10. The chordwise center-of-gravity
locations were the same as f o r the 60° delta wing but t h e spanwise posi-
t i o n s were a t l/3, 2/3, and f u l l semispan. 'he e f f e c t of the i n s t a l l a t i o n
of the bodies on the f l u t t e r c h a r a c t e r i s t i c s of t h e 45O sweptback wing
i s shown i n f i g u r e s 11 and 12 f o r Mach numbers of 1.3 and 2.0, respec-
t i v e l y . %e ordinate and abscissa a r e the same as f o r t h e d e l t a wing.
As the case of the 60° d e l t a wing, large increases i n the dynamic
pressure required f o r f l u t t e r are seen t o be associated with most of the

537
 +.
,* .* o

body i n s t a l l a t i o n s . The far-forward position of the center of gravity

seems t o cause the greatest increase i n t h e dynamic pressure f o r f l u t t e r .
The most favorable spanwise position would appear t o be near the
2/3-semispan position, p a r t i c u l a r l y a t the higher Mach number. A t f i r s t
glance, t h i s may appear t o contradict the r e s u l t s obtained on the d e l t a
wing but on the d e l t a wing the outermost position w a s only 3/4 semispan.

The improvements shown i n figures 8, 9, 11, and 1 2 t o accompany t h e

i n s t a l l a t i o n of external bodies on the d e l t a and sweptback wings indicate
t h a t the increase i n the s t i f f n e s s - a l t i t u d e parameter notedlin figure 1
f o r the plain wings a t the higher Mach numbers may be a l l e v i a t e d by the
considered location of an external body.

CONCLUDING REMARKS

The trends which have been presented emphasize the necessity f o r

careful consideration of wing f l u t t e r boundaries a t both transonic and
supersonic speeds on any new high-speed a i r c r a f t . The use of mass
balance would seem t o hold promise as a means f o r lessening the wing
f l u t t e r problem. Large external bodies need not cause severe penalties
from the f l u t t e r point of view and, i n f a c t , hold promise as a possible
m e a n s f o r a l l e v i a t i n g the f l u t t e r problem a t both transonic and supersonic
Mach numbers.

538
8

EFFECT OF MACH NUMBER

-L NO FLUTTER

-60°
A A
A
x
WING

-\
n
-
-- 45O
---- 45O
4

4
.2
.6

FLUTTER

I I I
0 I 2 3
M

Figure 1

EFFECT OF MASS BALANCE

BALANCE MASS 6 AND 12 "Lo OF WING MASS

q,LB/SQ FT

i*.

539
 STORE CONFIGURATIONS

A = 4 5 O ; A=4.0 ; Xz1.0

1 AIRSTREAM

Figure 3

EFFECT OF STORES
A=45': Az4.0; X=i.O

P- A AND B

1 I I I I I
0 .8 I,o 12 1.4
M

Figure 4
10

WING TORSIONAL NODE LINES

A=45O; A=4.0; X=I.O

FAIRSTREAM

Figure 5

EFFECT OF STORE-MOUNT FLEXIBILITY

M = 0.85

I I I I
0 4 .0 I .2 I.6
FREQUENCY RATIO, -
we
W a

Figure 6 4

I 5432
 11

60° DELTA P L A N FORM SHOWING

C.G. LOCATION OF EXTERNAL BODIES

-25% C 0 O/O c

~ 25% C
50% C

Figure 7

E F F E C T O F C.G. S H I F T OF E X T E R N A L BODIES
M.1.3 i A = 6 O o

I
C.G., To C
A ;4 / NO FLUTTER

.5
I I I I
0 ‘/4 ‘/2 3/4 I
FRACTION OF SEMISPAN

Figure 8
EFFECT OF C.G. SHIFT OF EXTERNAL BODIES
M=2.0$A = 6 0 °
C.G.,%C
0 -25
m o
+ 25
/
1 0
/ \NO FLUTTER
/

.5 L
I i I I I
0 I /4 112 314 I
FRACTION OF SEMISPAN

45O SWEPT PLAN FORM SHOWING

C.G. LOCATION OF EXTERNAL BODIES

Figure 10
 e .e 3
e

13G

EFFECT OF C.G. SHIFT OF EXTERNAL BODIES

M.1.3 j A=45O
C.G. ,% C
-25
mo
+ 25
2.0 A. 5 0

1.0 '

.5L
I 1 I J
0 '/3 5 I
FRACTION OF SEMISPAN

Figure ll

EFFECT OF C.G. SHIFT OF EXTERNAL BODIES

M z 2 . 0 ; A=45O

I I I 1
0 113 2/3 I
FRACTION OF SEMISPAN

Figure 12
 ..
0.
A
0
0
00 8
0

AERODYNAMICS OF OSCILLATING CONTROL SURFACES

- AT TRANSONIC SPEEDS
By Robert F. Thompson and Sherman A. Clevenson

Langley Aeronautical Laboratory

SUMMARY

Oscillating flap-type and all-movable controls a r e discussed w i t h

p a r t i c u l a r emphasis on the aerodynamic forces and moments a t transonic
speeds. Hinge-moment r e s u l t s from recent wind-tunnel and rocket-powered-
model t e s t s are summarized f o r trailing-edge flap-type controls t o i l l u s -
t r a t e the e f f e c t s of control hinge-line position and p r o f i l e shape on
one-degree-of-freedom f l u t t e r of t h i s type of control. From a wind-
tunnel investigation of a model considered representative of an a l l -
movable control, t h e aerodynamic e f f e c t s due t o rigid-body-oscillation
modes i n r o l l and i n p i t c h a r e presented.

The general magnitudes of t h e unstable aerodynamic damping moments

f o r the flap-type controls t e s t e d a r e pre,sented. No significant benefits
toward a l l e v i a t i n g t h e unstable aerodynamic damping i n the control r o t a -
tionalmode at transonic speeds were obtained f o r a wide range of hinge-
l i n e positions t e s t e d . O f three control p r o f i l e modifications t e s t e d ,
only a "wedge" modification t o a 35-percent-overhang balanced control
gave s i g n i f i c a n t improvements i n "buzz" s t a b i l i t y . This wedge control
gave s t a b l e aerodynamic damping i n t h e control r o t a t i o n a l mode up t o t h e
maximum speed t e s t e d . However, t h e s t a b l e damping and improvement i n
f l u t t e r c h a r a c t e r i s t i c s were limited t o o s c i l l a t i o n amplitudes l e s s than
about 3 O .

For an all-movable control, t h e o s c i l l a t i n g aerodynamic derivatives

which define t h e separate e f f e c t s of rigid-body p i t c h and rigid-body r o l l
have been presented through t h e transonic speed range. The aerodynamic
w i n g f o r t h i s low-aspect-ratio all-movable control was s t a b l e a t a l l
conditions t e s t e d .

INTRODUCTION

O s c i l l a t i n g aerodynamic forces and moments a r e needed i n analyzing

t h e dynamic o r f l u t t e r c h a r a c t e r i s t i c s of any airplane component. "hese
aerodynamic values cannot be accurately computed f o r t h e mixed-flow
conditions a t transonic speeds, and t h e r e i s a current need f o r experi-
mental data. The purpose of t h i s paper is t o summarize some recent
t e s t s on o s c i l l a t i n g flap-type and all-movable controls, wherein t h e
aerodynamic e f f e c t s a t transonic speeds have been measured. Flap-type
controls are discussed first, and results p e r t a i n primarily t o one-
degree-of-freedon f l u t t e r o r "buzz." Then, t h e aerodynamic e f f e c t s due
t o rigid-body modes i n p i t c h and i n r o l l are presented f o r an all-movable
control.

SYMBOLS

flap-type-control chord, distance from hinge l i n e t o

t r a i l i n g edge of control ( s e e f i g . 2), f t

flap-type-control balance chord, distance from hinge

l i n e forward t o leading eQe of control (see f i g . 2 ) ,
ft

r o o t chord of all-movable control, f t

t o t a l control chord a t midspan of control, cb + ca, f t

ka reduced frequency of all-movable-control p i t c h o s c i l l a -

tion,
2v

reduced frequency of flap-type-control o s c i l l a t i o n ,

2v

reduced frequency of all-movable-control r o l l o s c i l l a -

4 tion, W- r
2v

M free -stream Mach number

M' area moment of flap-type-control area rearward of and

about t h e hinge l i n e , cu f t

9 free-stream dynamic pressure, l b / s q f't

S semispan area of all-movable control, sq f t

V free-stream velocity, f t / s e c

c
 3

‘h flap-type-control hinge-moment coefficient,

Hinge moment
2M‘q

cL lift coefficient, Model lift

(4s
rolling-moment coefficient (fig. 7 ) ,
Model rolling moment about roll axis

qs -
Cr
2

‘m pitching-moment coefficient (fig 7),.

Model pitching moment about pitch axis

U amplitude of all-movable-control pitch oscillation,

radians except as noted

6 amplitude of flap-type-control oscillation, measured

in plane perpendicular to hinge line, radians except
as noted

# amplitude of all-movable-control roll oscillation,

radians except as noted

0 angular frequency of oscillation, radians/sec

derivative with respect to 6, a, or # as noted

derivative with respect to -

6C,

2v

derivative with respect to -

&Cr

2v
derivative with respect to &
2v

Subscript:

u) derivative obtained from an oscillation

4

DISCUSSION OF RESULTS

Flap-Type Controls

The aerodynamic hinge moment existing on t h e o s c i l l a t i n g flap-type

controls discussed herein i s represented i n complex notation by the
relationship
Resultant hinge moment - C + ik C
a'q6 h6,U %;Lo

where ch
89 0
represents an aerodynamic spring -moment derivative pro -
portional t o t h e component of the t o t a l aerodynamic moment i n phase with
control position. The product k C i s an aerodynamic damping param-
%,a
e t e r , proportional t o t h e component of t h e t o t a l aerodynamic moment i n
phase with control r o t a t i o n a l velocity, and contributes t h e damping.
The p a r t represents an aerodynamic viscous-damping derivative.
%,0

Theoretical considerations.- Theoretical values f o r control aerody-

namic damping a r e shown i n f i g u r e 1. These two-dimensional-flow r e s u l t s
were obtained from references 1 t o 3 and a r e presented t o provide a frame-
work f o r evaluating $he experimental flap-type-control data. Theoretical
damping derivatives a r e p l o t t e d against Mach number f o r a 30-percent-
chord flap-type control hinged at the leading edge. The dotted portion
of t h e l i n e s i s an a r b i t r a r y f a i r i n g between the subsonic and supersonic
theories, and t h e Mach number v a r i a t i o n indicated i s i n general agree-
ment with experiment. Theory shows t h e aerodynamic damping t o be unstable
f o r some values of reduced frequency at Mach numbers from about 0.9 t o
1.5. I n this region, control-surface f l u t t e r can occur unless s u f f i c i e n t
nonaerodynamic damping i s present i n t h e control system t o provide damping
moments g r e a t e r than t h e unstable aerodynamic moments. The subsonic and
supersonic damping derivatives are f a i r l y independent of Mach number and
reduced frequency. However, large e f f e c t s a r e indicated a t transonic
speeds, and t h e idealized theory indicates s t a b l e damping f o r reduced
frequencies representing very high o s c i l l a t i n g frequencies. Experimental
damping r e s u l t s t o date (refs. 4 t o ll) have a l l been i n t h e frequency
range where theory indicates i n s t a b i l i t y , and these experimental data
show unstable damping i n t h e control r o t a t i o n a l mode a t transonic speeds.
The magnitudes indicated by t e s t and theory are often i n poor agreement,
but this is not surprising considering t h e mixed-flow conditions which
e x i s t a t transonic speeds and t h e possible influence from nonpotential
sources. Based on these r e s u l t s , a fundamental approach i n a l l e v i a t i n g
control-surface "buzz" would be t o provide enough nonaerodynamic damping
i n t h e control system t o overcome t h e unstable aerodynamic e f f e c t s . This
approach generally necessitates the addition of some type of a r t i f i c i a l

548
damping t o t h e control system, f o r example, the type provided by a
hydraulic damper. This addition can often lead t o mechanical complexi-
t i e s , especially when such factors as control f r e e piay a r e considered;
and it would be desirable t o s t a b i l i z e t h e control aerodynamic danrping
by some change i n geometric c h a r a c t e r i s t i c s provided overall control
efficiency can be maintained.

Control hinge position and p r o f i l e shape a r e known t o have l a r g e

e f f e c t s on s t a t i c hinge moments, and some of t h e i r e f f e c t s on dynamic
hinge moments have recently been determined i n the hope that s t a b l e con-
t r o l aerodynamic damping a t transonic speeds can be achieved.

Effects of hinge position.- Studies of the e f f e c t s of control hinge

position and p r o f i l e shape were made i n t h e Langley high-speed 7- by
10-foot tunnel with t h e wing-control configuration shown i n figure 2
( r e f s . 7 and 8). These t e s t s were at a Reynolds number of about 2 x 106
based on the wing mean aerodynamic chord. Model plan form w a s held con-
s t a n t and the t o t a l control chord w a s 30 percent of t h e wing chord. The
range of test conditions covered i s indicated i n f i g u r e 2. From f i g u r e 1
it can be seen t h a t t h e test frequencies a r e i n a range where two-
dimensional-flow theory indicates unstable aerodynamic damping at tran-
sonic speeds f o r a 30-percent-chord control hinged a t t h e leading edge.
This paper first presents t h e e f f e c t s of hinge-line position, as indicated
by r e s u l t s f o r t h e three conventional control p r o f i l e s shown' i n f i g u r e 2,
and then summarizes t h e e f f e c t s of the p r o f i l e modifications indicated.
The hinge l i n e i s located by t h e r a t i o of t h e balance chord cb t o t h e
chord of t h e control rearward of t h e hinge l i n e ca, and 20-, 35-, and
100-percent-overhang balanced controls were t e s t e d .

I n figure 3 the e f f e c t s of hinge position on t h e control aerodynamic

damping a r e surmnarized. These data a r e presented f o r a reduced frequency
of 0.10 and an angle of a t t a c k of ; '
0 i n general, the variations i n angle
of a t t a c k and reduced frequency covered i n these t e s t s had small e f f e c t s
on t h e control hinge-moment r e s u l t s . A f r e e - o s c i l l a t i o n t e s t technique
was used and t h e damping derivative i s p l o t t e d against o s c i l l a t i o n anrpli-
tude f o r representative t e s t Mach numbers i n figure 3(a) with t h e complete
Mach number variation shown i n figure 3(b) f o r l o w o s c i l l a t i o n amplitudes.
Positive values of C indicate unstable aerodynamic damping. The
%,a
f l u t t e r o r buzz associated with t h e unstable damping shown f o r t h i s model
w a s a self-excited o s c i l l a t i o n and b u i l t up i n amplitude u n t i l t h e aero-
dynamic energy fed i n t o the control system over a complete cycle was
belanced by t h e energy dissipated due t o s t r u c t u r a l and f r i c t i o n a l damping.
Steady-state f l u t t e r points f o r the t e s t control system a r e indicated by
t h e c i r c u l a r symbols i n figure 3(a). The damping derivative f o r t h e 20-
and 35-percent-overhang balanced controls at subsonic speed was stable
and f a i r l y constant t o o s c i l l a t i o n amplitudes of about loo. Damping f o r
t h e LOO-percent-overhang balanced control was nonlinear with amplitude
6

a t t h i s speed and unstable a t the higher o s c i l l a t i o n amplitudes. I n

order t o i n i t i a t e t h e model f l u t t e r shown f o r t h i s subsonic Mach number,
it was necessary t o displace i n i t i a l l y t h e control t o some intermediate
amplitude and suddenly release it. This i n s t a b i l i t y a t a low Mach num-
b e r i s believed t o be closely r e l a t e d t o a s t a l l - f l u t t e r type of phe-
nomenon. Increasing t h e Mach number i n t o t h e transonic speed range
caused t h e control aerodynamic damping t o become unstable f o r all hinge
positions t e s t e d . For t h e 100-percent-overhang balanced control there
w a s f i r s t a l a r g e s t a b l e increase i n damping with increasing Mach number
before t h e damping became unstable. Model f l u t t e r a t these transonic
speeds w a s i n i t i a t e d by random tunnel disturbances, and t h e f l u t t e r
amplitude was markedly decreased a t sonic speed w i t h t h e hinge located
at t h e control midchord. The damping a t transonic speeds was nonlinear
with amplitude f o r all hinge positions and indicated t h e influence of
nonpotential e f f e c t s which r e s t r i c t the application of any p o t e n t i a l
theory.

Shown i n figure 4 a r e recent control aerodynamic damping r e s u l t s

obtained from a f r e e - f l i g h t rocket-powered-model test by the Langley
P i l o t l e s s A i r c r a f t Research Division. These data were obtained by a
f r e e - o s c i l l a t i o n t e s t technique similar t o that described i n reference 10.
T e s t Reynolds number, based on t h e wing mean aerodynamic chord, varied
.
from 2 x 106 t o 13 x 106 The delta-wing configuration i s i l l u s t r a t e d
and t h e trailing-edge flap-type control w a s hinged s l i g h t l y rearward of
t h e control midchord. The experimental damping derivative evaluated f o r
an o s c i l l a t i o n amplitude of 1.8' is shown f o r Mach numbers from about
0.5 t o 1.9, and t h e test reduced frequency varied as indicated through
t h e Mach number range. For this midchord hinge position, there was a
l a r g e increase i n stable damping a t high subsonic speeds and unstable
aerodynamic damping i n t h e transonic bhch nuuiber region, with t h e damping
again becoming s t a b l e a t t h e higher t e s t supersonic speeds. These aero-
dynamic damping trends w i t h Mach number, measured i n f r e e f l i g h t , are i n
agreement w i t h t h e t h e o r e t i c a l and wind-tunnel results presented. Based
on t h e data shown i n f i g u r e s 3 and 4, reasonable hinge positions do not
o f f e r much promise i n a l l e v i a t i n g t h e unstable damping i n the control
r o t a t i o n a l mode at transonic speeds.

The e f f e c t s of hinge position and Mach number on t h e control i n

phase o r s t i f f n e s s derivative f o r t h e wind-tunnel model a r e presented
i n f i g u r e 5 . Derivatives obtained from dynamic and s t a t i c t e s t s a r e
compared t o i l l u s t r a t e t h e e f f e c t s of o s c i l l a t i n g t h e control. Positive
values of t h i s derivative indicate overbalanced o r s t a t i c a l l y unstable
hinge moments. The balancing e f f e c t of s h i f t i n g t h e hinge l i n e rearward
i s shown, and t h e midchord hinge position overbalances t h e control through-
out t h e speed range t e s t e d . The o s c i l l a t i n g spring moments a r e approxi-
mately equal t o t h e s t a t i c values f o r the 20- and 35-percent-overhang
balanced controls, and t h e differences shown f o r t h e 100-percent-overhang
4G 7

balanced control can be attributed to the larger deflection range over

which it was necessary to evaluate the stiffness derivative for the
dynamic tests. This agreement indicates that, for a wide range of hinge
positions, static hinge moments can be used to predict accurately the
frequency of control buzz.

Effects of profile shape.- The profile modif2cations studied in the

wind-tunnel investigation are illustrated in figure 2. For the 20-percent-
overhang balanced control the thickness at the trailing edge was made
equal to the hinge-line thickness, and the portion of the control for-
ward of the hinge was not altered. Results for the oscillating hinge
moments for this control were similar to the results for the original
profile with the aerodynamic damping still unstable at transonic speeds.
This modification caused a slight beneficial shift in the level of unsta-
ble damping and an increase in the aerodynamic spring stiffness; however,
the flutter characteristics of the model were not appreciably improved.
(See ref. 7.)

On the 35-percent-overhang balanced control two profile changes

were made. For the lower modification shown, the rear half of the con-
ventional control chord was replaced by a thin "splitter" plate over
essentially the full span of the control. This control was similar to
some controls tested in a flight investigation by North American Aviation,
Inc. (For example, see ref. 12.) In the present tunnel tests, the
oscillating hinge-moment and flutter results measured for this splitter-
plate control did not show any significant differences relative to the
original profile, and the aerodynamic damping indicated about the same
level of instability at Mach numbers from about 0.92 to 1.01, the maxi-
mum speed tested. In reference 12 a trailing-edge splitter plate com-
bined with a slight thickening of the forward portion of the control has
given qualitative indication of improved buzz stability. Direct com-
parison of the model and flight results is not feasible since the aero-
dynamic damping was not measured in flight, and there are appreciable
geometric differences between the configurations.

The "wedge" modification to the 35-percent-overhang balanced con-

trol did give some beneficial effects on control damping. The trailing-
edge thickness was a little more than twice the thickness tested on the
modified 20-percent-overhang balanced control. The leading and trailing
edges were connected by a straight line which resulted in an increase in
the hinge-line thickness relative to the original profile. Damping
results for the wedge control are shown in figure 6 and are compared
, with the damping of the conventional profile. The aerodynamic damping
of the wedge control was stable at low oscillation amplitudes for the
complete-test speed range. However, at transonic speeds, stability is
confined to oscillation amplitudes less than about.'3 Damping for the
wedge control is unstable for amplitudes greater than about 3'; if the
8

model wedge control is manually displaced t o these unstable anrplitudes

and released, it would f l u t t e r w i t h c h a r a c t e r i s t i c s very similar t o those
i n t h e response of t h e conventional control.

All-Movable Control

The model used t o measure some o s c i l l a t i n g aerodynamic derivatives

at transonic speeds f o r conditions of p a r t i c u l a r application t o a l l -
movable cdntrols is shown i n figure 7. These tests were made i n t h e
Langley 2- by 2-foot transonic f l u t t e r research tunnel a t Reynolds numbers,
based on t h e control mean aerodynamic chord, from 1.4 x 106 t o 3.6 x 106.
The rigid-body degrees of freedom i n p i t c h and r o l l t o be considered a r e
i l l u s t r a t e d . The p i t c h axis passes through the midchord a t t h e r o o t
and t h e leading edge a t t h e t i p . The r o l l a x i s was a r b i t r a r i l y chosen
inboard of t h e model root, and the p i t c h and r o l l motion together with
t h e angles describing t h e modes are i l l u s t r a t e d . Nomenclature similar
t o that used i n discussing t h e flap-type controls i s a l s o used f o r t h i s
all-movable control with the r e s u l t a n t aerodynamic forces and moments
reduced t o derivatives that a r e e i t h e r i n o r out of phase with t h e
motion. The aerodynamic forces and moments e x i s t i n g on the o s c i l l a t i n g
model a r e represented i n complex notation f o r pitching motion as

Resultant lift -
- cL,,(I,+ +c%,u

Resultant pitching moment

= k,, + ikcm=,os
.
qs -
2
a

and f o r r o l l i n g motion as
Resultant r o l l i n g moment

L i f t and pitching-moment derivatives a r e presented f o r t h e pitching

motion, and t h e damping due t o r o l l i s shown f o r t h e r o l l i n g o s c i l l a t i o n .
These experimental data have not y e t been compared with e x i s t i n g tbree-
dimensional-flow theory ( r e f s . 13 and 14).

L i f t components due t o p i t c h o s c i l l a t i o n of the model a r e shown i n '

figure 8, and these data were evaluated f o r a forced-oscillation a n r p l i -
tude of about 3' t o each side of Oo angle of a t t a c k . The l i f t i n phase
with t h e motion i s shown on t h e l e f t and t h e out-of-phase component i s
shown on t h e r i g h t . D a t a were measured f o r a Mach nuztiber range from 0.6
t o 1.2, and t h e curves represent values f o r various reduced frequencies.
 9

For a constant reduced frequency, the in-phase derivative increases with

Mach numbers from 0.6 t o 1.0 and decreases w i t h Mach numbers from 1.0
t o 1.2. The out-of-phase derivative decreases with increasing Mach num-
bers from 0.6 t o 1.0 and changes sign from positive t o negative near a
Mach number of 1.0. Increasing t h e reduced frequency decreases t h e
in-phase component of the t o t a l lift a t a l l t e s t speeds and causes a
positive increment i n t h e magnitude of t h e out-of-phase component.
Although it i s not presented herein, experimental data have shown t h a t
the r o l l i n g moments due t o p i t c h o s c i l l a t i o n follow t h e same trends as
those shown herein f o r t h e l i f t .

I n f i g u r e 9 t h e pitch-damping parameter due t o p i t c h o s c i l l a t i o n

of the control i s plotted as a function of reduced frequency and Mach
number. These aerodynamic data were evaluated from t h e f r e e - o s c i l l a t i o n
response of t h e control following removal of t h e forcing function a t an
o s c i l l a t i o n amplitude of about 3 O . For additional information on t h e
t e s t technique used, see references 15 and 16. The aerodynamic viscous-
w i n g derivative was e s s e n t i a l l y l i n e a r f o r amplitudes from'0 t o 3'.
The damping moments were s t a b l e throughout t h e complete Mach number range
from 0.6 t o 1.2, and there was a tendency f o r t h e damping parameter t o
become more s t a b l e up t o M = 1.0 and then decrease. The damping
increased as t h e reduced frequency was increased. For conditions roughly
approximating those of t h e present t e s t s , two-dimensional-flow theory
( r e f s . 2 and 3 ) indicates unstable values f o r t h i s parameter a t t r a n -
sonic speeds. However, three-dimensional-flow theory ( r e f s . 13 and 14)
indicates a r a t h e r large s t a b i l i z i n g e f f e c t a t these speeds due t o aspect
r a t i o ; therefore, these experimental r e s u l t s showing s t a b l e damping
appear t o be i n reasonable agreement w i t h existing theory.

Figure 10 shows t h e roll-damping parameter due t o r o l l o s c i l l a t i o n

of t h e control as a function of reduced frequency and Mach number. These
data were evaluated from t h e f r e e - o s c i l l a t i o n response of t h e control
following r o l l o s c i l l a t i o n s up t o amplitudes $ of about 3'. The r o l l -
damping moments were s t a b l e throughout t h e speed range tested, and t h e
damping tends t o become more s t a b l e as t h e t e s t Mach number i s increased.
Increasing t h e reduced frequency a l s o increases t h e s t a b l e damping-in-
r o l l parameter.

CONCLUDING REMARKS

The aerodynamic forces and moments acting on o s c i l l a t i n g flap-type

and all-movable control surfaces at transonic speeds have been summarized.
The discussion on flap-type controls has b r i e f l y reviewed w h a t could be
considered t h e fundamental approach i n a l l e v i a t i n g one-degree-of-freedom
f l u t t e r of t h i s type of control, namely, incorporating s u f f i c i e n t non-
aerodynamic damping i n t h e control system t o overcome t h e unstable
10

aerodynamic moments. Experimental results were presented which estab-

l i s h t h e general magnitude of t h e unstable aerodynamic moments f o r t h e
t e s t models. However, it i s desirable t o have a control configuration
with inherent aerodynamic s t a b i l i t y ; therefore, the aerodynamic e f f e c t s
of control hinge-line position and some p r o f i l e modifications were
studied. Based on these r e s u l t s , no s i g n i f i c a n t benefits i n control
aerodynamic damping were obtained f o r a wide range of hinge positions.
O f three control p r o f i l e modifications t e s t e d , only a "wedge" modifica-
t i o n t o a 33-percent-overhang balanced control gave s i g n i f i c a n t improve-
ments i n "buzz" s t a b i l i t y . This wedge control gave s t a b l e aerodynamic
damping i n the control r o t a t i o n a l mode up t o t h e maximum speed t e s t e d .
However, t h e s t a b l e damping and improvement i n f l u t t e r c h a r a c t e r i s t i c s
were limited t o o s c i l l a t i o n amplitudes l e s s than about 3'.

For an all-movable control, t h e o s c i l l a t i n g aerodynamic derivatives

which define t h e separate e f f e c t s of rigid-body p i t c h and rigid-body
r o l l have been presented through t h e transonic speed range. The aero-
dynamic damping f o r t h i s all-movable control w a s s t a b l e a t a l l condi-
tions tested.
 11

REFERENCES

1. An n.: -
Tables of Aerodynamic Coefficients f an os illating Wing-
Flap System in a Subsonic Compressible Flow. Rep. F.151, Mationaal
Luchtvaartlaboratorium, Amsterdam, May 1954.

2. Nelson, Herbert C., and Berman, Julian H.: Calculations on the Forces
and Moments for an Oscillating Wing-Aileron Combination in Two-
Dimensional Potential Flow at Sonic Speed. NACA Rep. 1128, 1953.
(Supersedes NACA TN 2590.)

3. Garrick, 1. E., and Rubinow, S. I.: Flutter and Oscillating Air-Force

Calculations for an Airfoil in a Two-Dimensional Supersonic Flow.
NACA Rep. 846, 1946. (Supersedes NACA TN 1158.)

4. Reese, David E., Jr.: An Experimental Investigation at Subsonic and

Supersonic Speeds of the Torsional Damping Characteristics of a
Constant-Chord Control Surface of an Aspect Ratio 2 Triangular Wing.
NACA RM A53D27, 1953.

5. Martin, Dennis J., Thompson, Robert F., and Martz, C. William: Explor-
atory Investigation of the Moments on Oscillating Control Surfaces
at Transonic Speeds. NACA RM L55E31b, 1955.

6 . Reese, David E., Jr., and Carlson, William C. A.: An Experimental

Investigation of the Hinge-Moment Characteristics of a Constant-
Chord Control Surface Oscillating at High Frequency. NACA RM A55J24,
1955*
7. Thompson, Robert F., and Moseley, William C., Jr.: Oscillating Hinge
Moments and Flutter Characteristics of a Flap-Type Control Surface
on a 4-Percent-Thick Unswept Wing With Low Aspect Ratio at Trarnsonic
Speeds. NACA RM L55KJ-7, 1956.

a. Thompson, Robert F., and Moseley, William C., Jr.: Effect of Hinge-
Line Position on the Oscillating Hinge Moments and Flutter Charac-
teristics of a Flap-Type Control at Transonic Speeds. NACA RM L57C11,
1957. (NACA prospective paper. )
9. Tuovila, W. J., and Hess, Robert W.: Aerodynamic Damping at Mach
Numbers of 1.3 and 1.6 of a Control Surface on a Two-Dimensional
Wing by the Free-Oscillation Method. NACA RM L56A26a, 1956.

10. Martz, C. William: Experhental Hinge Moments on Freely Oscillating

Flap-Type Control Surfaces. NACA RM ~56~20,1956.
e. 0. a , ,
a - -

12

11. Clevenson, Sherman A.: Some Wind-Tunnel Experiments on Single-

Degree-of-Freedom F l u t t e r of Ailerons i n t h e High Subsonic Speed
Range. NACA TN 3687, 1956. (Supersedes NACA RM L ~ E Q ~ . )

12. Anon.: F l i g h t T e s t Progress Report No. 23 f o r Period Ending

7 October 1953 f o r Model FJ-4 Airplanes. Rep. No. NA 54H-374-23
(Contract NOa( s )54-323), North American Aviation, Inc , Oct 20, . .
1955
13. Nelson, Herbert C., Rainey, Ruby A . , and Watkins, Charles E.: Lift
and Moment Coefficients Expanded t o the Seventh Power of Frequency
f o r O s c i l l a t i n g Rectangular Wings i n Supersonic Flow and Applied
t o a Specific F l u t t e r Problem. NACA “I 3076, 1954.

14. Runyan, Harry L., and Woolston, Donald S.: Method f o r Calculating
t h e Aerodynamic Loading on an O s c i l l a t i n g F i n i t e Wing i n Subsonic
and Sonic Flow. NACA TN 3694, 1956.

15. Clevenson, Sherman A., and Widmayer, Edward, Jr.: Experimental

Measurements of Forces and Moments on a Two-Dimensional O s c i l l a t i n g
Wing A t Subsonic Speeds. NACA TN 3686, 1956. (Supersedes NACA
RM LgK28a.)

16. WiWyer, Edward, Jr., Clevenson, Sherman A . , and Leadbetter, Sumner A . :

Some Measurements of Aerodynamic Forces and Moments a t Subsonic
Speeds on a Rectangular Wing of Aspect Ratio 2 O s c i l l a t i n g About
t h e Midchord. NACA RM L53F19, 1953.
 THEORETICAL CONTROL DAMPING
TWO -DIMENSIONAL THEORY

20

15

10
C
hi,,
5

-5 I I 1 I
0 -5 1.0 1.5 2.o
M

Figure 1

FLAP -TYPE CONTROL MODEL

WIND -TUNNEL TESTS
MODEL PARAMETERS
ASPECT RATIO 1.8
TAPER RATIO 0.74
NACA 6 4 A 0 0 4
SCOPE O F TESTS

0.6TO 1.01
ANGLE O F ATTACK 0°AND6'
REDUCED FREQUENCY 0.06 TO 0.13

Figure 2
537
14 *

EFFECT OF HINGE POSITION ON CONTROL DAMPING

k=O.lOi a = O 0

I I
0- 0-
10 I

0
C
h8 ,w
-5
0 STEADY- STATE
-10 FLUTTER
AMPLITUDE

EFFECT OF HINGE POSITION ON CONTROL DAMPING

kZO.10; Q = O O ; 8=O0
10

-Cb/ca
5
I
D(, 0.20
0
I

0.35
\ I
I

1.00
-15

I
-25 Lk I I I I
.7 .8 .9 1.0 1.1
M
 EFFECT OF MACH NUMBER ON CONTROL DAMPING
ROCKET-MODEL TEST;8=+1.8"; a!=O"

k = Oilo 96 94
5

chbyw-/
-10 I
DLi
M

Figure 4

EFFECT OF HINGE POSITION ON CONTROL STIFFNESS

k.0.10 ) Q = O o

6
-DYNAMlG
0 STATIC
4

C cb/ca
hs,, 2
-D- I 1.00
I
0 I

1".c; -
A
v -0- .35
UNDERBALAN C ED I(
-2 .1, I - n c ' 20
.a .9 1.0 I
M
 EFFECT OF WEDGE PROFILE ON CONTROL DAMPING
k=O.IO ;a=O"

-I
a
I

0
chs,w
-5
o STEADY STATE
FLUTTER AM PLlTUDE
-I 0
0 5 IO 15 0 5 10 15
28 ,DEG +S,DEG

Figure 6

A L L -MOVABLE CONTROL SURFACE

RIGID-BODY DEGREES OF FREEDOM

PANEL ASPECT RATIO = 1.25

TAPER RATIO =0.283
NACA 6 5 A 0 0 5 a
T

it 560
Figure 7
 .
EFFECT OF MACH NUMBER ON LIFT COMPONENTS DUE
TO PITCH OSCILLATION
k

3-
CLap
2-

I-

M M

Figure 8

EFFECT OF MACH NUMBER ON PITCH-DAMPING

COMPONENT DUE TO PITCH OSCILLATION

.3
k
0 0.10
.II5
.I 5
k Cmb,w -3 .I75

-.6

-:9
L I I I I I I I I I
0 .4 .'5 -6 .7 .8 .9 1.0 1.1 1.2
M

Figure 9
EFFECT OF MACH NUMBER ON ROLL-DAMPING
COMPONENT DUE TO ROLL OSCILLATION

0.5

0
k
-0.5-0.10
.I I5

.*O \
-2.0-

-2.5-
I
04 .6 .f .k -6 I
1.0
I
1.1 1.2
M

Figure 10
 STATUS OF FLUTTER OF FLAT AND CURVED PANELS

By Robert W. Leonard and John M. Hedgepeth

Langley Aeronautical Laboratory

SUMMARY

Representative r e s u l t s are presented t o show t h e current status

of the panel f l u t t e r problem. The discussion includes f l a t panels with
and without midplane s t r e s s e s , buckled panels, and both unstiffened and
s t i f f e n e d i n f i n i t e l y long c i r c u l a r cylinders.

INllRODUCTION

The f l u t t e r of the skin panels of supersonic airplanes and missiles

has aroused considerable i n t e r e s t i n recent years. Numerous investiga-
tions ( r e f s . 1 t o 27) have determined that panel f l u t t e r may govern t h e
design of a t l e a s t very t h i n panels such as those used i n f a i r i n g s and
i n radiation shielding. Representative r e s u l t s of some of these inves-
t i g a t i o n s , which tend t o i l l u s t r a t e the current status of t h e problem,
a r e presented i n t h i s report. 'Ihe discussion i s divided roughly i n t o
three p a r t s : F i r s t , the f l u t t e r of f l a t panels; second, the f l u t t e r of
panels stressed beyond t h e bucMing load; and t h i r d , t h e f l u t t e r of
thin-walled c i r c u l a r cylinders.

SYMBOLS

a panel dimension i n chordwise direction

b panel dimension i n spanwise direction

D panel s t i f f n e s s ,
Et3
** ab e 1 , , - -
e -

E Young's modulus

i,j,m,p integers

M Mach number

n integer, number of circumferential waves around the cylinder

px chordwise midplane load per u n i t spanwise length

pY spanwise midplane load per u n i t chordwise length

% dynamic pressure
P
R = J
PX
cylinder radius

thickness

panel deflection

chordwise coordinate

spanwise coordinate

dynamic-pressure parameter, zqa3

&?y-D
Poisson's r a t i o (taken equal t o l / 3 throughout)

mth mode of uniform clamped-clamped beam of u n i t length

5 dimensionless coordinate

4
V4=a7+2 a4 + -a4
bX ax2ay* ay4
 FLAT PANELS

1 I t o the present time, the l a r g e s t research e , , x t has been devoted

t o t h e o r e t i c a l study of f l a t panels. Typical a n a l y t i c a l r e s u l t s for f l a t
panels are shown i n figures 1 and 2.

The r e s u l t s i n figure 1 apply t o panels of i n f i n i t e aspect r a t i o ,

t h a t i s , panels which extend t o i n f i n i t y i n the spanwise direction and,
hence, behave s t r u c t u r a l l y l i k e a beam. Figure 2 contains r e s u l t s for
a panel with f i n i t e aspect r a t i o . Both p l o t s give t h e panel thickness
r a t i o t/a 'necessary t o prevent f l u t t e r , a t various Mach numbers, of
s t e e l panels a t sea l e v e l . The r e s u l t s a r e r e s t r i c t e d t o panels which
are unstressed; t h a t i s , they have no midplane tension or compression.

I n both figure 1 and f i g u r e 2, the dashed curves a t low supersonic

Mach numbers a r e two-mode r e s u l t s based on the use of unsteady linearized
a i r forces. (See r e f s . 8, 10, and 20.) The s o l i d curves a t high Mach
numbers a r e two-mode r e s u l t s obtained by using much simpler " s t a t i c " air
forces i n which no account i s taken of unsteady e f f e c t s . (See refs. 22
and 26.) Experience has shown t h a t , f o r the high Mach numbers, the
s t a t i c approximation y i e l d s r e s u l t s i n close agreement with those yielded
by more refined theories. The smooth merging of the dashed and s o l i d
l i n e s f o r each configuration i l l u s t r a t e s t h i s close agreement.

I n figure 1, curves a r e shown for three d i f f e r e n t configurations:

an array of panels continuous over equally spaced supports, a single
panel with pinned ends, and a single panel with clamped ends. The general
e f f e c t of panel boundary conditions may be seen from the r e l a t i v e posi-
tions of t h e curves a t high Mach numbers. The most c r i t i c a l configwa-
tion, requiring thicker panels t o prevent f l u t t e r , i s t h e array of panels
on equally spaced supports.

The e f f e c t of aspect r a t i o on required panel thickness i s shown i n

figure 2. I n t h i s case, both curves a r e f o r a single panel w i t h pinned
edges. The upper curve applies t o an i n f i n i t e span panel and i s t h e
same as the middle curve of figure 1. The lower curve applies t o a
square panel. As would be expected, reducing t h e aspect r a t i o from
i n f i n i t y t o one reduces the required thickness.

It should be pointed out t h a t the moderate reduction i n thickness

r a t i o a t high Mach numbers i s due primarily t o the s t r u c t u r a l e f f e c t s of
aspect r a t i o . (See r e f . 26.) De aerodynamic e f f e c t i s l e s s than 2 per-
cent throughout the range of the s o l i d curves. Thus, the use of aero-
dynamic s t r i p theory would give accurate boundaries f o r the square p l a t e
f o r these high Mach numbers.
 The situation is apparently different for the low Mach numbers.
In addition to the moderate structural effect, there is now a large
aerodynamic effect of aspect ratio and a resulting large reduction in
thickness in changing from the infinite-aspect-ratio panel to the square
panel. It is interesting to note that, although the low supersonic Mach
number range appears to be critical for infinite-aspect-ratio panels,
this may not be the case for the simply supported square panel for which
the results show a steady increase in thickness with increase of Mach
number.

It should be pointed out that the dashed curves at the low Mach
numbers in figures 1 and 2 are based on a rather limited number of cal-
culations and this region has not yet been carefully explored. It is
known, however, that structural damping is fairly effective in reducing
the large thicknesses required for infinite-aspect-ratio panels. (See
ref. 10.)

It is also worthwhile to note that a small amount of experimental

data has been reported at Mach numbers of 1.3 and 1.56 for clamped panels
which behave like the infinite-aspect-ratio panel. (See refs. 6 and 13.)
The data are not shown in figure lbecause they apply to a different
altitude and material. When compared with calculations made on the same
basis, experiment and theory agree at Mach number 1.56. At Mach num-
ber 1 . 3 , the deta show an increase of thickness but not the large increase
predicted by the theory. The difference may be due in part to structural
damping.

So far, only panels with no midplane stress have been considered.

However, the effect of midplane tension or compression stresses on the
required thickness of flat panels has been investigated and is found to
be important (ref. 26). This effect is illustrated by the results shown
in figure 3 .

Figure 3 is a plot of a modified-thiclmess-ratio parameter

({Cl q'3t 9 containing both panel and
D
air properties, against a
chordwise compression parameter sX .The coefficient of
(Buckling Px)pv=o
J
the thickness ratio contains the Mach number My Young's modulus E,
and the dynamic pressure q and allows the application of the plotted
results to all Mach numbers above about 1.5. The compression parameter,
which specifies stress in'the chordwise direction, is the ratio of
chordwise compression Px to the buckling value of chordwise stress
which corresponds to zero stress in the spanwise direction. Positive
values of this parameter indicate chordwise compression and negative
values imply tension. Curves are shown for two pinned-edge panels with
aspect ratios of 0.5 and 1.
I

566
 5

It m u s t be remembered that the calculations apply t o f l a t panels

only. Hence, the curves i n figure 3 are v a l i d only up t o t h e chordwise
load which results i n buckling. If the midplane stress i n t h e spanwise
d i r e c t i o n i s zero, the curves are therefore valid, by d e f i n i t i o n , only
up t o t h e compression p a s m e t e r equal t o one. If tension i s applied t o
t h e panel i n the spanwise direction, t h e chordwise load necessary t o
cause buckling is raised; the curves would then be v a l i d t o a higher ,
value of t h e compression parameter. Spanwise compression would have
the opposite e f f e c t . It should be noted that this i s the only influence
of spanwise stress on the f l u t t e r boundaries of f l a t panels. (See
r e f . 26.)

The influence of chordwise s t r e s s i s sham by the v a r i a t i o n of t h e

curves themselves. Note especially that, f o r each aspect r a t i o , there
i s a c r i t i c a l value of chordwise compression f o r which t h e theory requires
very large thicknesses t o prevent f l u t t e r . In r e a l panels, t h e required
thickness i n this range i s dependent on t h e amount of s t r u c t u r a l damping.
However, it i s apparent that these c r i t i c a l combinations of aspect r a t i o
and chordwise compression should be avoided i n design.

The mode shapes of f l u t t e r of f l a t panels have an i n t e r e s t i n g f e a t u r e

t h a t i s i l l u s t r a t e d i n figure 4. Calculated chordwise variations of ty-p-
i c a l flutter-mode shapes a r e shown f o r panels with both pinned and clamped
edges. The air flow i s from l e f t t o r i g h t . Note t h a t the motion i s con-
centrated toward the rear of the panel f o r both panels. This " t a i l -
wagging" type of motion i s c h a r a c t e r i s t i c of t h e observed motion i n a c t u a l
instances of panel f l u t t e r .

Consideration w i l l now be given t o the f l u t t e r of panels which have

been buckled by heating or by the application of midplane loads. Buckled
panels have been t r e a t e d i n a number of t h e o r e t i c a l investigations, most
of which a r e confined t o panels of i n f i n i t e aspect r a t i o . (See r e f s . 1,
2, 3, 9, 11, 23, and 26 .) I n addition, a f e w experimental s t u d i e s have
been made. (See refs. 13, 16, and 23.)

The e f f e c t of buckling may be seen from t h e results i n f i g u r e 5.

The required thickness r a t i o t/a i s p l o t t e d against Mach number f o r
s t e e l panels of i n f i n i t e span i n a i r a t sea l e v e l . Curves a r e s h m a t
high Mach nmbers f o r both pinned and clamped panels. The upper p a i r of
curves ( f i g . 5 ) a r e calculated boundaries which apply t o buckled panels
of i n f i n i t e span without regard t o buckle depth. They result frcan closed-
form solutions based on t h e so-called " t r a n s t a b i l i t y " concept. (See
ref. 1.) The lower pair of curves ( f i g . 5 ) are the corresponding results
f o r f l a t panels. Note that t h e buckled panels require much l a r g e r thick-
nesses t o prevent f l u t t e r than do t h e unbuckled panels.

567
\io,.
0 O0
C 00 0 0 0 , ---
c
0

Experimental r e s u l t s f o r clamped buckled panels are a l s o shown i n

f i g u r e 5. With the exception of the point a t M = 2.18, these r e s u l t s
were taken from reference 16. (They have been corrected t o apply t o
s t e e l a t sea l e v e l . ) The point a t M = 2.18, which was obtained inde-
pendently ( r e f . 23), appears t o confirm the other experimental r e s u l t s .
The dashed t h e o r e t i c a l curve f o r clamped buckled panels d i f f e r s on the
conservative side by about I 2 percent.

In references 13 and 16, some experimental r e s u l t s were presented

f o r a group of clamped rectangular panels buckled by heating. These
r e s u l t s a r e rep_eated i n figure 6.

Figure 6 i s constructed w i t h modified-thickness-ratio parameters

for both the ordinate and abscissa. One thickness r a t i o i s based on the
chord and t h e other, on t h e span. The d a t a a r e applicable t o a l l Mach
numbers greater than about 1 . 3 . The construction of the p l o t i s such
that data f o r panels of a given aspect r a t i o f a l l on a r a d i a l l i n e from
the origin. The aspect r a t i o s and orientation of the t e s t panels a r e
shown by the small figures a t the outer ends of the !radial l i n e s . I n
each case, node l i n e s a r e shown t o i l l u s t r a t e the observed buckle pat-
t e r n s . The distance outward from the origin of the p l o t i s a measure
of the panel thickness. Solid points designate panels t h a t f l u t t e r e d
and open points indicate t h a t no f l u t t e r was observed.

An attempt has been made t o calculate a t h e o r e t i c a l f l u t t e r boundary

f o r t h i s p a r t i c u l a r group of panels buckled by heating. Because the
midplane s t r e s s e s i n the t e s t panels had not been measured, it was neces-
sary t o estimate these s t r e s s e s from t h e observed buckle patterns. Some
d e t a i l s of t h i s calculation are given i n the appendix. The s o l i d curve
i s the r e s u l t i n g approximate boundary. The curve has been shown dashed
on the lower right because the calculation i s a two-mode approximation
which must be considered unreliable f o r the very low-aspect-ratio panels.

The agreement i n figure 6 between t h e t h e o r e t i c a l boundary and the

experimental data encourages the viewpoint t h a t successful analyses of
rectangular buckled panels are possible. O f p a r t i c u l a r i n t e r e s t i s the
finding t h a t , as i n the case of f l a t rectangular panels, there a r e c r i t -
i c a l combinations of aspect r a t i o and s t r e s s f o r which even t h i c k panels
can f l u t t e r . One such c r i t i c a l point occurs i n the calculated boundary
near an aspect r a t i o of 0.5. Perhaps t h i s explains why the t e s t panels
with an aspect r a t i o of 0.5 were more prone t o f l u t t e r than the other
t e s t panels were.

It must be pointed out t h a t the t h e o r e t i c a l boundary i s very sensi-

t i v e t o the type of buckling load, especially f o r aspect r a t i o s between
0.5 and 2. Consequently, t h i s boundary applies only t o the p a r t i c u l a r
s e r i e s of t e s t s shown. It should be emphasized t h a t this boundary i s not
intended t o be a universal. f l u t t e r boundary f o r f i n i t e - a s p e c t - r a t i o
buckled panels. In f a c t , no such universal boundary e x i s t s .
 7

CIRCULAR CYLINDERS

One other configuration which has received some preliminary a t t e n -

t i o n i s t h e i n f i n i t e l y long thin-walled c i r c u l a r cylinder with a x i a l
flow over i t s outer surface. (See r e f s . 14, 21, and 27.) Typical
r e s u l t s , f o r empty s t e e l cylinders, are shown i n f i g u r e 7.

The ordinate i n figure 7 i s t h e r a t i o of w a l l thickness t o cylinder

radius t/r and the abscissa i s the Mach number. The curves a r e theo-
r e t i c a l f l u t t e r boundaries which apply t o unstiffened i n f i n i t e l y long
cylinders. The s o l i d curves a r e boundaries t h a t r e s u l t when there i s
no static-pressure d i f f e r e n t i a l across t h e w a l l ; the dashed curves, on
the other hand, correspond t o the existence of enough i n t e r n a l pressure
t o cause a circwnferential tension near the ultimate.

It w i l l be noted t h a t the required w a l l thicknesses of unstiffened

i n f i n i t e l y long cylinders increase rapidly w i t h increase i n Mach number
and very quickly reach prohibitive values, even a t very high a l t i t u d e s .
I n t e r n a l pressure i s seen t o produce only small percentage reductions
i n these large thicknesses. Furthermore, it has been pointed out i n
reference 27 t h a t negligible benefit can be expected i f i n t e r n a l damping
i s taken i n t o account.

On the other hand, an e f f e c t t h a t shows promise of reducing t h e

thicknesses t o reasonable values i s t h e e f f e c t of f i n i t e length. For
the purpose of i l l u s t r a t i n g t h i s e f f e c t , an approximate calculation has
been m a d e f o r a s t e e l cylinder with r i g i d ring-stiffeners spaced a radius
apart. The calculation i s based on equation (42) of reference 2 1 w i t h
the a i r forces replaced, f o r simplicity, by t h e i r plane s t a t i c approxi-
mation. This simplification imposes the requirements t h a t the Mach
number M be large and t h a t M 2 >> 2 where n i s the nu-ber of f u l l
a 2n
waves around the circumference of the cylinder (corresponding t o the
maximum required thickness f o r the prevention of f l u t t e r ) and a i s the
distance between s t i f f e n e r s . The r e s u l t i n g thickness r a t i o required t o
prevent f l u t t e r a t M = 6 of the stiffened s t e e l cylinder a t an a l t i t u d e
of 35,000 f e e t with no i n t e r n a l pressure i s shown by t h e small c i r c l e i n
figure 7. This r e s u l t i s only approximate because the above conditions
are only approximately s a t i s f i e d (maximum t / R corresponds t o n = 26).
However, it has the correct order of magnitude. Furthermore, i n contrast
t o the extremely large thickness r a t i o s predicted a t M = 6 f o r the
unstiffened i n f i n i t e l y long cylinder, the r e s u l t f o r t h e s t i f f e n e d cyl-
inder has a reasonable order of magnitude. Thus, the large thicknesses
predicted by analyses of i n f i n i t e l y long unstiffened cylinders apparently
do not apply t o p r a c t i c a l configurations of f i n i t e length.
i e 1 e6 0. .. e. . , . ---

C O N C D I N G REMARKS

It may be stated t h a t much progress has been made i n the s o l u t i o n

and understanding of t h e panel f l u t t e r problem. Although, i n general,
panel f l u t t e r i s of concern only i n t h e design of very t h i n panels, there
appear t o be c r i t i c a l combinations of panel aspect r a t i o and midplane
stress f o r which even very t h i c k panels may f l u t t e r . It should be
emphasized t h a t there i s a need f o r more work on c e r t a i n phases of t h e
panel f l u t t e r problem. I n p a r t i c u l a r , there i s a need f o r f u r t h e r con-
s i d e r a t i o n of cylinders of f i n i t e length and f o r more experimental
results f o r assessing the v a l i d i t y of t h e o r e t i c a l analyses.
 .a *.

APPENDIX

TRANSTABILITY ANAIYSIS OF THE FLUTPER OF EUCKLFD

m C W G U L A R PANEW W I T R CULMPED EDGES

The application of the "transtability" concept (ref. 1) to rectan-

gulm buckled panels with pinned edges is discussed in reference 26.
The corresponding results, for clamped panels, are presented in this
appendix.

The transtability problem is governed by the plate equation

4
DV w + PXwn + PYwn +

where w is the panel deflection and the subscripts denote differentia-

tion, and by the proper boundary conditions. The assumption is made that
the panel deflection shape is adequately represented in the chordwise
direction by a linear combination of the first two modes of a uniform
clamped-clamped beam and in the spanwise direction by a single mode, the
pth mode. Then, by the Galerkin method, the condition for the existence
of a nontrivial solution is found to be

where
10

The function e',(k) is the nth mode for a beam of unit length. The
beam modes and the integrals I$ ) given conveniently in refer-
are
ences 28 and 29.

As pointed out in reference 26, the critical value of the dynamic-

pressure parameter A depends on the midplane loads Px and Py applied
to the panel at the boundaries or induced by heating a panel with fixed
boundaries. Let Py = RPx. Substituting % and % from equations (3)
into equation (2) yields a quadratic equation for the buckling loads Px
corresponding to given values of R and A. As A increases from zero,
the buckling loads approach each other until at the critical "transtability"
value of A they coalesce and disappear. This condition is given by the
vanishing of the discriminant of the quadratic and yields the result

Ill

Calculations have been made of the quantities:

and

for a series of test panels buckled by heating. me integer p, speci-

fying the number of half-waves in the spanwise direction, was taken as
the number of spanwise half-waves in the observed buckle patterns.
!
i
5";'Z
Furthermore, since the a c t u a l loads i n the t e s t panels had not been
measured, t h e r a t i o - w a s estimated from the buckle p a t t e r n s
R = py
PX
with t h e a i d of f i g u r e 5 of reference 30. Specifically, f o r each aspect
r a t i o , the average longitudinal buckling load corresponding t o the number
of observed spanwise or chordwise half-waves w a s used along with t h e
corresponding transverse load. For an aspect r a t i o of 1, it w a s assumed
t h a t R = 1 (Px = Py). The r e s u l t i n g calculated boundary i s shown i n
figure 6 along with the t e s t r e s u l t s .
0 . .
* .*- 0.
, e. , a
. ..
- - - - e
.
. 0. .*

12

REFERFNCES

1. Isaacs, R. P.: Transtability Flutter of Supersonic Aircraft Panels.

U. S. Air Force Project RAND P-101, The Rand Corp., July 1, 1949.
2. Hayes, W.: A Buckled Plate in a Supersonic Stream. Rep. No. AL-1029,
North American Aviation, Inc., May 10, 1950.

3. Miles, John W.: Dynamic Chordwise Stability at Supersonic Speeds.

Rep. No. AL-1140, North American Aviation, Inc., Oct. 18, 1950.
4. Shen, S. F.: Non-Stationary Aerodynamics of a Two-Dimensional Bump
in a Uniform Stream and Its Effect on the Vibration Characteristics
of an Elastic Panel. M.I.T. Tech. Rep. (Contract No. N5ori-07833,
NR 064~59)~ k y 1952.
5. Shen, S. F.: Flutter of a Two-Dimensional Simply-Supported Uniform
Panel in a Supersonic Stream. Contract No. N5oriLO7833, Office of
Naval Res., Dept. Aero. Eng., M.I.T., Aug. 6 , 1932.

6. Sylvester, Maurice A., and Baker, John E.: Some Experimental Studies
of Panel Flutter at Mach Number 1.3. NACA TN 3914, 1957. (Supersedes
NACA RM ~52116.)

7. Goland, Martin, and Luke, Yudell L.: An Exact Solution for Two-
Dimensional Linear Panel Flutter at Supersonic Speeds. Jour. Aero.
Sei. (Readers' Forum), vol. 21, no. 4, Apr. 1954, pp. 275-276.
8. Hedgepeth, John' M., Budiansky, Bernard, and Leonard, Robert W. :
Analysis of Flutter in Compressible Flow of a Panel on Many Supports.
Jour. Aero. Sci., vol. 21, no. 7, July 1954, pp. 475-486.

9. Fung, Y. C. : The Static Stability of a Two-Dimensional Curved Panel

in a Supersonic Flow, With An Application to Panel Flutter. Jour.
Aero. Sci., vol. 21;, no. 8, Aug. 199, pp. 556-565.
10. Nelson, Herbert C., and Cunningham, Herbert J.: llheoretical Investi-
gation of Flutter of Two-Dimensional Flat Panels With One Surface
Exposed to Supersonic Potential Flow. NACA Rep. 1280, 1956.
(Supersedes NACA TN 3465.)
11. Fung, Y. C.: The Flutter of a Buckled Plate in Supersonic Flow.
GAZCIT Rep. No. OSR-TN-55-237, July 1955.

12. Eisley, 5. G.: The Flutter of Simply Supported Rectangular Plates in

a Supersonic Flow. QSR-TN-55-236, GUCIT, July 1955.
7G 13

13. Sylvester, Maurice A., Nelson, Herbert C., and Cunningham, Herbert J.:
Experimental and Theoretical Studies of Panel Flutter at Mach
Numbers 1.2 to 3.0. NACA RM L5'jEl8b, 1955.

14. Miles, John W.: Supersonic Flutter of a Cylindrical Shell. The

Ramo-Wooldridge Corp., Guided Mssile Res. Div.
I. General Theory. Rep. No. AM 5-2, Aug. 19, 1955.
11. Pressurization and Internal Fluid Effects. Rep. No. AM 5-ll,
Nov. 14, 1955.
111. Aeolotropic Shell. Rep. No. AM 5-l2, Dec. 2, 1955.
IV. Effects of Non-Uniform Steady Flow. Rep. No. AM 5-16,
Dee. 7, 19559
15. Shen, S. F.: Remarks on "An Exact Solution for Two-Dimensional
Idnear Panel Flutter at Supersonic Speeds." Jour. Aero. Sci.
(Readers' Forum), vol. 22, no. 9, Sept. 1955, pp. 636-657.

16. Sylvester, Maurice A.: Experimental Studies of Flutter of Buckled

Rectangular Panels at Mach Numbers From 1.2 to 3.0 Including
Effects of Pressure Differential and of Panel Width-Length Ratio.
NACA RM ~55130,1957.
17. Luke, Yudell L., St. John, Andrew, and Goland, Martin: An Exact
Solution for Two-Dimensional Linear Panel Flutter at Supersonic
Speeds. WADC Tech. Note 56-460 (Contract No. AF 33( 616)-2897),
Midwest Res. Inst., Mar. 15, 1956.

18. Luke, Yudell L., St. John, A. D., and Gross, Betty: Panel Flutter
at Supersonic Speeds. Third Quarterly Progress Rep. (Contract
No. AF 33(616)-2897), Midwest Res. Inst., Appl. Phys. Div.,
w. 29, 1956.
19. Luke, Yudell L., St. John, A. D., and Gross, EWtty: Panel Flutter
at Supersonic Speeds. Fourth Quarterly Progress Rep. (Contract
No. AF 33(616)-2897), Midwest Res. Inst., Appl. Phys. Div.,
M ~ Y29, 1956.
20. Luke, Yudell L., and St. John, A. D.: Panel Flutter at Supersonic
Speeds. Fifth Quarterly Progress Rep. (Contract No. AF 33(616)-2897),
Midwest Res. Inst., Appl. Phys. Div., Oct. 31, 1956.

21. Ieonard, Robert W., and Hedgepeth, John M.: On Panel Flutter and
Divergence of Infinitely Long Unstiffened and Ring-Stiffened Thin-
Walled Circular Cylinders. NACA TN 3638, 1956.
22. Hedgepeth, John M.: On the Flutter of Panels at High Mach Numbers.
Jour. Aero. Sci. (Readers' Forum), vol. 23, no. 6, June 1956,
pp. 609-610.
14

23. Eisley, J. G.: The F l u t t e r of a Two-Dimensional Buckled P l a t e With

Clamped Edges i n a Supersonic Flow. OSR-m-56-296, GALCIT, July
1956.
24. Miles, John W.: On the Aerodynamic I n s t a b i l i t y of Thin Panels.
Jour. Aero. Sci., vol. 23, no. 8, Aug. 1956, pp. 771-780.

25. Ashley, Holt, and Zartarian, Garabed: P i s t o n Theory -

A New Aero-
dynamic Tool f o r t h e Aeroelastician. Jour. Aero. Sci., vol. 23,
no. 12, BC. 1956, pp. 1109-1-1.3.8.

26. Hedgepeth, John M.: F l u t t e r of Rectangular Simply Supported Panels

a t High Supersonic Speeds. Preprint No. 713, S.M.F. Fund Paper,
Inst. Aero. Sci., Inc., Jan. 28-31, 1957.

27. Miles, John W.: Supersonic F l u t t e r of a Cylindrical Shell. Jour.

Aero. Sci., vol. 24, no. 2, Feb. 1957, pp. 107-118.

28. Young, Dana, and Felgar, Robert P., Jr. : Tables of Characteristic
Functions Representing Normal Modes of Vibration of a Beam. Univ.
of Texas, Pub. No. 4913, Eng. R e s . Ser. No. 44, Bur. Eng. Res.,
July 1, 1949.

29. Felgar, Robert P., Jr.: Formulas f o r I n t e g r a l s Containing Character-

i s t i c Functions of a Vibrating Beam. C i r . No. 14, Bur. Eng. Res.,
Univ. of Texas, 1950.

30. Libove, Charles, and Stein, Manuel: Charts f o r C r i t i c a l Combinations

of Longitudinal and Transverse Direct S t r e s s f o r F l a t Rectangular
P l a t e s . NACA WR L-224, 1946. (Formerly NACA ARR L6AO5.)
 e

REQUIRED THICKNESS OF INFINITE -ASPECT-RATIO

UNSTRESSED STEEL PANELS AT SEA LEVEL

7 STATIC A I R FORCES
---- UNSTEADY AIR FORCES

I I I I
2 3 4 5
M

Figure 1

EFFECT OF ASPECT RATIO ON REQUIRED THICKNESS OF

UNSTRESSED SIMPLY -SUPPORTED STEEL PANELS
.015-
AT SEA LEVEL
I

n
I
I
I
I
1.7
I
I
I
.010- I
I
I
-
t I
I
a I
I
I
I
I
.005- I

c--
'--
I

-STATIC AIR FORCES

---- UNSTEADY AIR FORCES
I I I I
01 2 3 4 5
M
16

E F F E C T OF M I D P L A N E S T R E S S ON F L U T T E R
OF UNBUCKLED PANELS
M > 1.5

1.25-

1.00 -

.75 -
-
t
a
50 -

.25 -

I I I
0' -i 0 I 2
PX
(BUCKLING P X ) ~-o
Y-

TYPICAL FLUTTER - M O D E SHAPES

PINNED

CLAMPED

L E A D I N G EDGE T R A I L I N G EDGE

Figure 4
 EFFECT OF BUCKLING ON REQUIRED THICKNESS OF
I N F I N I T E - ASPECT-RATIO S T E E L PANELS AT SEA L E V E L

-OI5r

I - ----- PINNED EDGES

.OlO c O
CLAMPED EDGES
CLAMPED EDGES EXPERIMENT

00 0
0 ------------- BUCKL.ED
0
___________________ _____
I
}FLAT

REQUIRED THICKNESSES OF A SERIES OF

TEST PANELS BUCKLED BY HEATING

0 FLUTTER
0 NO FLUTTER
.8

-I .6
3 ) 'q +
.4

.2

0 .2 .4 .6 .8 1.0 1.2

Figure 6
579
 REQUIRED THICKNESS OF INFINITELY
LONG STEEL CYLINDERS
.020-

.015-

q
NO INTERNAL PRESSURE
MAX. INTERNAL PRESSURE

35,000 FT
NO INTERNAL PRESSURE
I 1
5 6
M

Figure 7
 \
L - ’
... b I .. .

FLUTTER AND DIVERGENCE OF RECTANGULAR

WINGS OF VERY LOW ASPECT RATIO

By Robert W. Fralich, John M. Hedgepeth,

and W. J. Tuovila

Langley Aeronautical Laboratory

SUMMARY

Slender-body aerodynamic theory is used in conjunction with thin-

plate theom in the flutter analysis of low-aspect-ratio rectangular
wings of constant thickness when chordwise variations of deflections
are considered. The spanwise variation of deflection is given by a
parabola, and the chordwise variation is allowed complete freedom. The
results show the variation of flutter speed and mode shape with aspect
ratio. Comparisons are made with results obtained by approximating the
chordwise deflection shape by the first few terms of a power series.
Comparisons with some preliminary experimental results are also included.

INTRODUCTION

The prediction of flutter of wing and tail surfaces of very low

aspect ratio is a problem of some concern to aircraft designers. The
difficulties in the flutter analysis of such surfaces are mainly con-
nected with the presence of large amounts of chordwise curvature in the
flutter mode. It is of interest to investigate the complexity of the
chordwise deflection shape at flutter and to determine to what degree
of accuracy the chordwise deflection must be represented in order to
obtain good results. This paper is concerned with the flutter behavior,
both theoretical and experimental, of the simple low-aspect-ratio con-
figuration shown in figure 1. The analysis is similar to that of ref-
erence 1, which treated the static divergence behavior of the same
configuration.

SYMBOLS

coordinate system (see fig. 1)

wing deflection, positive in z-direction
2

free-stream velocity

wing thickness
wing chord
wing semispan

time

chordwise deflection shape

-
dynanic pressure, pv2 .
2

E Young's modulus of elasticity

P free-stream density of fluid

density of material
period of oscillation

Mach number

perturbation-velocity potential

h flutter-speed parameter, z(1

4
- p2)$
s3
-
t3

K flutter-frequency parameter, u)

E mass-ratio parameter, S p s
2 ht

CL Poisson's ratio (taken as 1/3 in all computations)

D plate stifmess in bending, Et3/12(l - p2)

U) flutter frequency

THEDREZICAL APPROACH

The configuration analyzed in this paper consists of a rectangular

plate of constant thickness from a rigid wall. (See fig. 1.)
r
I 582
This plate may be thought of as representing one-half of a wing with a
chord c and a semispan s. The plate is located in a fluid flow with
a free-stream velocity V. The deflection shape of such a low-aspect-
ratio plate can be expected to vary in a much more complicated manner
in the chordwise direction than in the spanwise direction. For this
reason the deflection w is assumed to vary as w(x,y,'r) = y2F(x,T),
where the spanwise deflection is given by a simple parabola and the
chordwise variation F of the deflection is an arbitrary function of
the chordwise coordinate x and time 7 . The distortions of the plate
are found through the use of ordinary thin-plate theory. The aerodynamic
loadings are found most simply by using slender-body aerodynamic theory.
In this theory streamwise perturbations are neglected in comparison with
perturbations in the crossflow direction. The use of this approximate
aerodynamic theory simplifies the aeroelastic problem to the extent that
an exact solution is possible. A brief description of the analysis and
the resulting equations are given in the appendix.

RESULTS AND DISCUSSION

Some results are given by the boundaries shown in figure 2. The

ordinate is the dynamic-pressure parameter *, in which q is the
(*/SI3
dynamic pressure, E is Young's modulus of elasticity for the plate,
and t/s is the ratio of thickness to semispan. The abscissa is the
ratio of chord to semispan c/s. The variation of the dynamic-pressure
parameter for flutter with the ratio of chord to semispan is dependent upon
the mass-ratio parameter E in which is the ratio of air density
Rtlt Qm
to plate density. The flutter boundaries are given for two values of
mass-ratio parameter; the region above a particular boundary is unstable
while that below is stable. Also shown is the result for static diver-
gence obtained from reference 1. This result, which is, of course,
independent of the mass ratio, is indicated by a single curve. Note
that divergence is less critical than flutter for these particular mass
ratios. Note also that, for the higher mass ratio, the flutter boundary
consists of a series of loops approaching a constant value of dynamic-
pressure parameter. The lower curve also has these characteristics; how-
ever, the loops are so elongated in this case that only one can be seen
in this figure.

The kind of flutter mode shapes obtained fromthe analysis is

shown in figure 3 . The top set of curves shows the components of tip
deflection which are in phase and out of phase with the maximum leading-
edge deflection for a value of c/s corresponding to the tick mark on
the bottom of the first loop on the flutter boundary in figure 2. The

583
4

t i p d e f l e c t i o n i s given when the leading edge has i t s maximum amplitude

and a t one-quarter of a period later when the leading edge has zero
d e f l e c t i o n . The bottom set of curves gives t h e components of mode shape
f o r a value of c/s given by the t i c k mark on t h e second loop of t h e
f l u t t e r boundary of figure 2. The e f f e c t of increasing the chord i s t o
add more waves t o t h e mode shape.

I n the r e s u l t s discussed so far, t h e chordwise v a r i a t i o n s of deflec-

t i o n s were allowed complete freedom and an exact s o l u t i o n w a s possible.
I n a p r a c t i c a l case, an exact solution would not be f e a s i b l e , and some
s o r t of approximation of t h e chordwise d e f l e c t i o n shape would be nec-
essary. Some f l u t t e r boundaries obtained by approximating the chord-
w i s e d e f l e c t i o n s by t h e f i r s t few terms of a power series are shown i n
figure 4. The dashed curve gives t h e results f o r parabolic deformations,
and t h e long-and-short-dashed curve gives t h e results f o r cubic deforma-
t i o n s . The exact boundary i s a l s o shown f o r comparison. Both approxi-
mations y i e l d good results f o r the lower values of c / s . The cubic
approximation i s almost exact. For longer chords, however, both approxi-
mations y i e l d poor r e s u l t s . Apparently, i n order t o analyze t h e f l u t t e r
behavior of wings i n t h i s range, higher order terms i n t h e d e f l e c t i o n
shape must be used.

It can be seen from figure 2 t h a t f o r each value of mass r a t i o a

l i m i t i n g value of t h e dynamic-pressure parameter can be obtained by con-
sidering t h e values a t the bottom of t h e loops as t h e chord becomes
l a r g e . The v a r i a t i o n of t h i s l i m i t i n g value w i t h mass r a t i o i s shown
i n figure 5 . The r e s u l t s obtained s o far indicate t h a t t h i s curve gives
a conservative estimate of t h e f l u t t e r speed f o r t h i n rectangular plates
of very l o w aspect r a t i o . It should be noted that the f l u t t e r speed i s
less than the divergence speed and seems t o approach it asymptotically
f o r high mass r a t i o .

Several preliminary t e s t s of some low-aspect-ratio p l a t e s a t super-

sonic Mach numbers have been run i n t h e Langley 9- by 18-inch supersonic
f l u t t e r tunnel. The r e s u l t s of these t e s t s are shown i n f i g u r e 6. The
f l u t t e r speed of t h e various models i s shown as a r a t i o of experimental
f l u t t e r speed t o calculated f l u t t e r speed, where t h e calculated f l u t t e r
speed w a s obtained from the curve i n figure 5 . T h i s r a t i o i s p l o t t e d
a g a i n s t t h e r a t i o of chord t o semispan. The agreement between theory and
experiment i s f a i r l y good i n view of t h e approximations inherent i n t h e
theory and of the preliminary nature of the t e s t s .

During t h e t e s t s t h e f l u t t e r modes were observed, and it w a s noted

t h a t t h e specimens w i t h the l a r g e r r a t i o s of chord t o semispan had t h e
more complicated mode shapes as i s predicted by theory.

.
 5

CONCLUDING REMARKS

The f l u t t e r analysis of low-aspect-ratio rectangular p l a t e s indi-

cates t h a t the flutter-mode shape has an increasing number of waves i n
t h e chordwise d i r e c t i o n as t h e aspect r a t i o i s reduced; approximating
t h e chordwise d e f l e c t i o n shape by parabolic or cubic curves y i e l d s f l u t -
t e r speeds i n f a i r agreement with those of t h e more exact theory, pro-
vided t h a t t h e aspect r a t i o i s not too l o w . The cubic curve gives some-
what b e t t e r r e s u l t s . For lower aspect r a t i o s , higher order approximations
must be used. Experimental data indicate that the f l u t t e r speed and t h e
type of mode shape yielded by t h e theory are i n f a i r agreement with
experimental results.

585
 APPENDIX

AmLYSIS

In the application of slender-body aerodynamic theory to the present

unsteady-flow problem, terms containing time derivatives in the velocity-
potential equation for linearized flow axe neglected in addition to those
containing streamwise x-derivatives. The velocity-potential equation
thus reduces to Laplace's equation in the crossflow plane. The boundary
conditions on velocity and the pressure-potential relations are taken to
be the same as those ordinarily used in unsteady linearized aerodynamic
theory.

The potential cp and the aerodynamic loads resulting f'romthe

given deformation shape are calculated in a manner similar to that used
in reference 1 for the static-divergence problem.
The principle of minimum potential energy is used to derive the
differential equation of equilibrium for the function F(x,T) in a man-
ner analogous to that used in reference 1. Solution yields an eigen-
value equation which relates the flutter speed to the properties of the
plate and the surrounding air:

L
Dl cos + E1 cosh cosh +3
F1 sinh sinh 9)+
i(D2 sin + E5!
U sinh cosh E +2 cosh % sinh Ej
S P S

where

-I -.-..
(A3 + hB7) 12A3 - ( 47* + a2 + P2> AJ,] +

586
 00.
. I , a * 0 I , , 0. .. .*

F 1 = (A3 + AB3)[2u2P2A4 - (4y2 + u2 + P2)A3] +

u2p2(A4 - A) [+ - ( 4r2 + a2 + P2)A4]

D2 = -( A1B2 - A2B1)

+ 8a1 - 8A + - 2(r2 - h)
7

in which

B2
= 20 (1 -
3
1.1)A + (Y 2 - A) (2. - j"")
y 1 2

m o (1 -
-9 p)(l - 21.1) - 20 (1 - p)A + 10 (1 - p ) r 2 + h2 IF -
-- 5€h
A3 - 3 3 472 *
587
A4 = 10 (1 - p) + r2
3

- 10 CI.
B3 - 3 - Y2

I n equation (l), u and p can be e i t h e r r e a l or imaginary q u a n t i t i e s .

Solution by t r i a l yields t h e curves given i n f i g u r e 2.

The divergence boundary obtained from reference 1 can a l s o be

obtained from t h e present analysis by s e t t i n g t h e flutter-frequency
parameter K equal t o zero.

Mode Shapes

The flutter-mode shapes shown i n f i g u r e 3 can be obtained from t h e

following equation, whe t h e component i n phase and
588
edge deflection:

f(x) =
Cleiml
-
the imaginary part, the component out of phase with t h e max-

X
= + C2eim2 5
c1 + c2
+ C3ea 3
+ c3 + c4
5 + C4eim4 Q
9

leading-

where

-i(ml+m4 ) 4j
(5- %)(a2 - m42)%.3' +
 C
(5- ml)(a2 - %22 ) 13
~e -i(%+%)
5 +

(ml - (9 -
9) m32)A12e

i n which

ml = 7 + ij3

m;!=y-i$

m3 = -7 + ia

m4 = -7 - ia

Approximate Solutions

The approximahe analysis used p a r a l l e l s the analysis f o r divergence

presented i n reference 1. Cubic approximation of the chordwise varia-
t i o n of deflection r e s u l t s i n the following complex determinant:
G

6
N
v

0
II

&
f

+
4
&It-

*
cu
I I

M 0
v

81- L2L.I
+ II
4
+
4

v
N

No1
4
+
-

t
12

REFERENCE

1. Hedgepeth, John M., and Waner, Paul G., Jr.: Analysis of Static
Aeroelastic Behavior of Low-Aspect-Ratio Rectangular Wings. NACA
!JD3958, 1957.
CANTILEVER PLATE OF VERY LOW ASPECT RATIO

Figure 1

FLUTTER AND STATIC DIVERGENCE BOUNDARIES

‘r
DIVERGENCE

c/s

Figure 2
593
 FLUTTER MODE SHAPES

IN-PHASE (T"0)

I I I I I I I
0 I 2 3 4 5 6
x/s, DISTANCE IN SEMISPANS

Figure 3

FLUTTER BOUNDARIES OBTAINED FROM APPROXIMATE

DEFLECTION FUNCTIONS

1.0 -
ps .00636
t ,P =
.8 - \

.6

(t/s) 3 -
.4 EXACT
----- PARABOLIC OEFORMATION

.2
- --- CUBIC DEFORMATION

I I I I I
0 2 4 6 8 10
CIS

591 Figure 4
 ENVELOPE VALUES OF CRITICAL
DYNAMIC-PRESSURE PARAMETER

I I I I I I
.01 .02 .03 .04 .05 .06
-
PS
Pm+

Figure 5

1.2-
H
1.0 ' - H
F 0
(2r' 0

.0 -
0
0 .
VEXP
VCnLC .6

M
.4
0 1.64
} ALUMINUM
.2 l.3 } STEEL
H 1.64

I I I I I I I
0 I 2 3 4 5 6
c/s

Figure 6
595
 FLUTTER EXPlKtMENTS W I T H VARIOUS CONTROL CONFIGURATIONS

By Robert W. Boswinkle, Jr., and Homer G. Morgan

Langley Aeronautical Laboratory

SUMMARY

The f l u t t e r c h a r a c t e r i s t i c s of various control-surface configurations

have been under study a t the Langley Aeronautical Laboratory. Presented
herein i s a compilation of some of the more pertinent r e s u l t s . Most of
the models studied were dynamically and e l a s t i c a l l y scaled frm proposed
a i r c r a f t . The f l u t t e r investigations w e r e made i n various f a c i l i t i e s over
both the transonic and supersonic speed ranges. Configurations t e s t e d
included a wing with t i p ailerons, all-movable elevators, v e r t i c a l t a i l s
with trailing-edge rudders, and a T-tail.

INTRODUCTION

A large portion of the f l u t t e r problems which plague modern aircraft

i s associated with control systems. No r e a l l y r e l i a b l e a n a l y t i c a l method
i s available f o r designing controls i n the transonic zone. Such a v a r i e t y
of configurations arise that an overall trend study of c h a r a c t e r i s t i c s i s
impractical. The a l t e r n a t i v e i s t o investigate experimentally the various
s p e c i f i c configurations as they a r i s e . Much of the work on controls a t
the Langley Aeronautical Laboratory has been of this type.

The Langley Laboratory has worked with various organizations on

control-surface f l u t t e r problems of new a i r c r a f t . The work has included
transonic and supersonic f l u t t e r t e s t s of models. Most of the models used
were both dynamically and e l a s t i c a l l y scaled from the prototype a i r c r a f t .
Some of the more pertinent r e s u l t s of these investigations w i l l be pre-
sented. The controls t o be discussed a r e shown as configurations A t o H
i n figure 1. They include a wing with t i p ailerons, two all-movable
controls, four v e r t i c a l t a i l s with trailing-edge rudders, and a T - t a i l .

SYMBOLS

a speed of sound, f t / s e c 596

b r o o t half chord, f t
2

g s t r u c t u r a l damping coefficient

M Mach number

dynamic pressure, lb/sq f t

v free-stream velocity, f t / s e c

Mass of wing
P =
Mass of a i r contained i n truncated cone determined by wing

LUf f l u t t e r frequency, radians/sec

bending frequency, radians/sec

a
, t o r s i o n a l frequency , radians/sec

control-surface r o t a t i o n a l frequency, radianslsec

LUe p i t c h frequency of all-movable control, radians/sec

W 1 f i r s t coupled frequency, radians/sec

DISCUSSION

Configuration A - Wing With Tip Ailerons

Configuration A w a s a wing of arrowhead plan form, shown i n f i g -
ure 1, w i t h trailing-edge f l a p s and all-moving t i p ailerons. The leading
edge of the wing was swept back 55' and the t r a i l i n g edge w a s swept f o r -
ward 10'. The a i l e r o n w a s about 20 percent of the exposed wing area and
i t s hinge l i n e w a s swept back about 4.5'. The hinge l i n e w a s located a t
56 percent of the a i l e r o n r o o t chord (where the a i l e r o n root chord i s the
chord which contains the innermost parting l i n e ) . The a i r f o i l w a s 3 per-
cent thick. The wing w a s cantilever mounted a t the r o o t . The t r a i l i n g -
edge f l a p s had a fixed r o t a t i o n a l s t i f f n e s s . F l u t t e r points were obtained
f o r different values of simulated a i l e r o n actuator s t i f f n e s s .

The r e s u l t s are shown i n f i g u r e 2, p l o t t e d as the s t i f f n e s s - a l t i t u d e

parameter against Mach number. Increases i n a l t i t u d e correspond t o
increases i n the value of the s t i f f n e s s - a l t i t u d e parameter. I n t h i s type
figure, constant a l t i t u d e occurs as a horizontal l i n e , and constant-
dynamic-pressure curves are s t r a i g h t l i n e s through the origin. I n f i g -
ure 2, the frequency i n the s t i f f n e s s - a l t i t u d e parameter i s the t o r s i o n
frequency of t h e wing. The actuator s t i f f n e s s i s the a c t u a l value divided
by the o r i g i n a l value.

5
 Data f o r the o r i g i n a l value of actuator s t i f f n e s s a r e shown on the
l e f t s i d e of figure 2. The no-flutter points correspond t o about sea-
l e v e l conditions. The mode of f l u t t e r which occurred a t a Mach number
of about 0.9 w a s predominantly bending and torsion of the w i n g , whereas
the f l u t t e r mode a t supersonic Mach numbers w a s predominantly a i l e r o n
rotation.

These f l u t t e r points w e r e considered t o be too close t o the a i r -

plane f l i g h t regime and an improvement i n the conditions w a s sought by
increasing the actuator s t i f f n e s s t o three times the o r i g i n a l s t i f f n e s s .
This increased actuator s t i f f n e s s w a s s t i l l a feasible value. The data
obtained with the s t i f f e r actuator a r e shown by the squares on the r i g h t
side of f i g u r e 2. The no-flutter points a t supersonic speeds so obtained
were a t lower a l t i t u d e s than the f l u t t e r points previously obtained with
the lower actuator s t i f f n e s s . No-flutter points were a l s o obtained a t
supersonic speeds w i t h an i n f i n i t e actuator s t i f f n e s s . I n f i n i t e actuator
s t i f f n e s s w a s simulated by gluing the a i l e r o n t o the wing so no r e l a t i v e
motion could take place. It i s i n t e r e s t i n g t o note, however, that, even
with i n f i n i t e actuator s t i f f n e s s , bending-torsion-type f l u t t e r w a s s t i l l
obtained near a Mach number of 0.9 and occurred a t about the same a l t i -
tude as f o r the lowest value of actuator s t i f f n e s s .

Increasing the actuator s t i f f n e s s w a s one way t o increase the

a i l e r o n r o t a t i o n frequency; another way w a s t o reduce t h e moment of
i n e r t i a of th? aileron. Since there was considerable windup of the
ailerons, a large reduction i n e f f e c t i v e moment of i n e r t i a w a s possible
by cutting off the t i p s of the ailerons. Accordingly, the t i p s were cut
off along the dashed l i n e s shown i n figure 1. Removal of these t i p s
gave beneficial results a t supersonic speeds but again l e f t the subsonic
f l u t t e r region e s s e n t i a l l y unaffected.

Increasing the actuator s t i f f n e s s o r cutting off the a i l e r o n t i p s

reduced the a l t i t u d e a t which the aileron-rotation-type f l u t t e r occurred
a t supersonic speeds. These changes had no e f f e c t on the bending-torsion-
type f l u t t e r obtained a t subsonic speeds.

Configuration B - All-Movable Stabilizer

Configuration B, as s h a m i n figure 1, w a s an all-movable s t a b i l i z e r .

The model had an aspect r a t i o of 3.3, a taper r a t i o of 0.42, 35' sweepback
of the quarter-chord l i n e , a i r f o i l sections tapering from 6 percent t h i c k
a t the r o o t t o 4 percent a t the t i p , a rounded t i p , and a p i t c h a x i s a t
78.9 percent of the center-line chord. T e s t s were made a t a constant
value of s t a b i l i z e r t w i s t s t i f f n e s s and r a t i o of bending frequency t o t o r -
sion frequency. The pitching frequency of the s t a b i l i z e r on i t s spindle
w a s the main test variable.
 These t e s t r e s u l t s are shown i n figure 3 where the s t i f f n e s s - a l t i t u d e
parameter i s shown as a function of Mach number f o r various values of the
r a t i o of bending frequency t o p i t c h frequency. The frequency i n the
s t i f f n e s s - a l t i t u d e parameter i s the t o r s i o n frequency of the s t a b i l i z e r .
The lower curve i s f o r a frequency r a t i o of 0.50. A s the Mach number i s
increased, the f l u t t e r boundary shifts t o higher a l t i t u d e s . This Mach
number e f f e c t becomes more pronounced as the frequency r a t i o i s increased
toward 1.0, as shown by the slopes of the curves f o r frequency r a t i o s of
0.62, 0.77, and 0.94.

The e f f e c t of frequency r a t i o a t a Mach number of 0.8 i s shown i n

f i g u r e 4. The flutter-speed coefficient V/bcDa i s p l o t t e d as a function
of frequency r a t i o f o r a given value of mass-density r a t i o . It should be
noted t h a t on this p l o t the f l u t t e r side of the boundary is above the
curve. The experimental data a r e shown as a s o l i d curve, cross-plotted
from the previous figure. The other two curves are calculated results
using two-dimensional incompressible aerodynamic coefficients and t h r e e
uncoupled modes. The calculations were made a t a fixed value of the
r a t i o of bending frequency t o t o r s i o n frequency, %/u+, = 0.303. The
lower dashed-line curves assume zero s t r u c t u r a l damping coefficient and
the upper dashed-line curves use a s t r u c t u r a l damping coefficient i n each
of the three modes equal t o the measured values. This calculation again
shows the importance of damping on f l u t t e r speeds a t frequency r a t i o s
near 1. The trend a t frequency r a t i o s below 1 appears t o be predicted
and the calculations a r e on the conservative side of the boundary. The
calculations p r e d i c t a favorable jump i n f l u t t e r speed as the frequency
r a t i o i s increased above 1 although experimental data were not obtained
a t large enough frequency r a t i o s f o r v e r i f i c a t i o n . For frequency r a t i o s
above 1, these calculations indicate that the f l u t t e r speed increases
almost d i r e c t l y w i t h the t o r s i o n frequency. Changes i n r a t i o of bending
frequency t o p i t c h frequency would have p r a c t i c a l l y no e f f e c t on the
f l u t t e r speed. However, a t frequency r a t i o s between about 0.3 and 0.8,
increases i n t o r s i o n frequency may not result i n proportionate increases
i n f l u t t e r speed. The f l u t t e r speed i n this region varies i n a much more
complicated manner. For example, notice that an increase i n t o r s i o n fre-
quency wa w h i l e holding the r a t i o of bending frequency t o torsion f r e -
quency constant would require increases i n %, the bending frequency.
This, i n turn,,would increase the r a t i o of bending frequency t o p i t c h
frequency and lower the flutter-speed coefficient. A more complete dis-
cussion of t h i s behavior i s given i n reference 1.

Configuration C - All-Movable Stabilizer

Configuration C y shown i n figure 1, i s a s t a t i c a l l y balanced a l l -

movable s t a b i l i z e r . The control w a s spindle mounted with the r o t a t i o n
 5

axis a t 46 percent of the root chord. Its semispan aspect r a t i o w a s

1.56, leading-edge sweepback w a s 20°, t h e t r a i l i n g edge w a s swept f o r -
ward 24O, and the thickness r a t i o w a s 5 percent. Torsion frequency O f
the surface w a s very high, s o that the predominant control-surface modes
were p i t c h about the spindle axis and bending. Studies were made a t
various values of the r a t i o of bending frequency t o p i t c h frequency.
The r e s u l t s a r e shown i n figure 5 , with s t i f f n e s s - a l t i t u d e parameter
p l o t t e d as a function of Mach number. The frequency i n the s t i f f n e s s -
a l t i t u d e parameter i s the r o t a t i o n frequency of the s t a b i l i z e r . The
upper curve i s the f l u t t e r boundary established f o r an uncoupled bending-
p i t c h frequency r a t i o of 1.02. When the frequency r a t i o w a s increased t o
1.41 o r decreased t o 0.92, favorable shifts i n the f l u t t e r boundary were
obtained. For the higher frequency r a t i o s the f l u t t e r mode involved both
bending and torsion, whereas f o r the lowest frequency r a t i o t h e f l u t t e r
mode involved primarily bending. "hese tests demonstrate that a change
i n frequency r a t i o which moves the frequency r a t i o away from 1 i s bene-
f i c i a l and requires less s t i f f n e s s t o avoid f l u t t e r .

Mass balancing i s a common method f o r increasing the f l u t t e r speed

of controls. Such a method w a s studied f o r t h i s case. The results a r e
shown i n f i g u r e 6. The plain-wing data f o r a frequency r a t i o of 1.02 a r e
repeated from the previous figure. Also shown i s t h e f l u t t e r boundary
obtained when a heavy leading edge w a s added t o the same configuration on
the outboard 50 percent of the control, as shown by t h e sketch i n f i g u r e 6.
The added weight w a s about 7 percent of the plain-wing weight. The f r e -
quency r a t i o ,of the modified model dropped t o 0.95. Adding t h e weight i s
seen t o be b e n e f i c i a l s o that less s t i f f n e s s i s required t o prevent f l u t -
t e r . However, the increase i n f l u t t e r speed may be due t o the change i n
frequency r a t i o as w e l l as the center-of-gravity s h i f t .

Calculations have been made f o r comparison with the experimental

data obtained with the p l a i n wing a t frequency r a t i o s of 1.02 and 1.41.
The calculated and experimental r e s u l t s are compared i n figure 7.
Experimental f l u t t e r points are shaded symbols and experimental no-
f l u t t e r points are open symbols. The calculated results are shown as
the curves. These calculations used the f i r s t two measured coupled
modes of the system. Zero s t r u c t u r a l damping w a s assumed and the aero-
dynamic forces were obtained from piston theory ( r e f . 2 ) including the
e f f e c t of a i r f o i l thickness.

The calculated f l u t t e r boundary f o r a frequency r a t i o of 1.02 i s

seen t o be i n excellent agreement with t h e experimental data over t h e
Mach number range investigated - 1.5 t o 2.8. The agreement i s not s o
good f o r a frequency r a t i o of 1.41. For t h i s case, the calculated curve
passes through experimental no-flutter points a t Mach numbers of 1.6 and
2.0 and s o p r e d i c t more s t i f f n e s s than necessary. Not enough s t i f f n e s s
i s predicted a t a Mach number of 2.2. These two examples give some
6

i n d i c a t i o n of the r e l i a b i l i t y of l o w supersonic flutter-speed calcula-

t i o n s using piston theory. For one case, agreement between experiment
and calculations i s good and i n the other case, not s o good.

Configurations D and E - Vertical Tails With

Trailing-Edge Rudders

Configurations D and E, sham i n figure 1, were swept v e r t i c a l

t a i l s with unbalanced trailing-edge rudders. Studies were made i n two
d i f f e r e n t f a c i l i t i e s t o obtain the f l u t t e r boundary f o r each configura-
t i o n a t both transonic and supersonic speeds. The models were c a n t i -
lever mounted and had fairings t o keep them out of the tunnel boundary
layer. The models were f l a t p l a t e s that w e r e tapered i n thickness along
t h e span. Both models had 60° sweptback leading edges. Vertical t a i l D
had a panel aspect r a t i o of 0.72 and a t a p e r r a t i o of 0.43. Its rudder
hinge l i n e w a s swept back 43' and i t s rudder w a s 24 percent of the
exposed surface area. Vertical tail E had a panel aspect r a t i o of 0.77
and a taper r a t i o of 0.48, based on a plan form without a rounded t i p .
I t s rudder hinge l i n e had 50' of sweepback and i t s rudder w a s about
36 percent of the exposed surface.
The results f o r configuration D a r e presented i n figure 8, where
the s t i f f n e s s - a l t i t u d e parameter i s p l o t t e d as a function of Mach number.
The frequency appearing i n the s t i f f n e s s - a l t i t u d e parameter i s the rota-
t i o n frequency of the rudder. This model w a s cantilever mounted along
i t s e n t i r e r o o t chord. These controls had a r a t i o of first bending
frequency t o rudder-rotational frequency of about 0.4. Also shown i n
f i g u r e 8 is the f l u t t e r frequency r e f e r r e d t o the control r o t a t i o n
frequency. A t a Mach number of about 1.2, an abrupt increase i n s t i f f -
ness i s required t o prevent f l u t t e r . A t t h e same Mach number the f l u t t e r
frequency a l s o changes abruptly. A t lower Mach numbers, the f l u t t e r fre-
quency i s less than the rudder-rotation frequency, whereas a t higher Mach
numbers the f l u t t e r frequency i s greater than the rudder-rotation fre-
quency. This change i n f l u t t e r frequency is associated with a change i n
the f l u t t e r mode. Previously, f l u t t e r mode changes i n w i n g s without con-
t r o l s have been observed t o produce more stable configurations i n the
transonic region. I n the present case, however, the change i s toward a
less stable system.

Figure 9 shows the results f o r configuration E wherein the plan form

w a s cantilevered from a stub. In this case, the r a t i o of first-bending
frequency t o rudder-rotation frequency i s about 0.3. Again, the f i g u r e
shows both the s t i f f n e s s - a l t i t u d e parameter and r a t i o of f l u t t e r frequency
t o rudder-rotation frequency as functions of Mach number. The f l u t t e r
frequency again suggests a change i n the f l u t t e r mode p a s t a Mach number
of 1 but t h e change i s gradual as Mach number increases and n o t abrupt as
 7

f o r configuration D. Also, no destablizing jump o r discontinuity i n

the s t i f f n e s s - a l t i t u d e parameter occurs as Mach number increases. If
a flutter-mode change does occur, the shift i s gradual and produces no
detrimental e f f e c t s .

For v e r t i c a l t a i l D, a destablizing jump w a s found i n the f l u t t e r

boundary a t low supersonic speeds, whereas none w a s obtained f o r v e r t i -
c a l t a i l E. A very d i f f e r e n t type of f l u t t e r boundary w a s obtained f o r
somewhat similar v e r t i c a l tails with rudders. However, f o r both con-
t r o l s , i f the surface were designed with j u s t enough s t i f f n e s s t o avoid
f l u t t e r a t transonic Mach numbers, a new f l u t t e r region would be
encountered a t the same a l t i t u d e a t a Mach number of about 2.5. This
again shows the existence of two c r i t i c a l f l u t t e r regions -
transonic
and supersonic. A complete description and discussion of the tests on
these two v e r t i c a l tails w i l l be found i n reference 3.

Configurations F and G - Vertical T a i l s With

Trailing-Edge Rudders

Configurations F and G, shown i n f i g u r e 1, were swept v e r t i c a l t a i l s

with s t a t i c a l l y balanced trailing-edge rudders. S t a t i c balancing w a s
accomplished w i t h overhanging-balance weights located a t the top of the
rudders. Vertical t a i l F had an aspect r a t i o of 1.33, taper r a t i o of
0.44, and i t s rudder w a s about 24 percent of the exposed surface area.
Vertical t a i l G had an aspect r a t i o of 1.58, a taper r a t i o of 0.40, and a
rudder which w a s about 20 percent of the exposed surface area. Both con-
figurations had 45' sweep a t the quarter-chord line, 2g0 sweep on their
rudder hinge l i n e s and tapered i n thickness from about 6 percent a t the
root t o about 4 percent a t the t i p . The models w e r e cantilever mounted
a t the root.

The results f o r v e r t i c a l t a i l F are presented i n figure 10. The

frequency appearing i n the s t i f f n e s s - a l t i t u d e parameter is t h e t o r s i o n
frequency of the f i n . Fin properties remained e s s e n t i a l l y unchanged
during the t e s t , w h i l e the actuator and rudder twist s t i f f n e s s e s w e r e
varied. The actuator and twist s t i f f n e s s e s shown i n the f i g u r e have been
divided by the values of the s t i f f n e s s e s f o r the o r i g i n a l configuration.
The f l u t t e r points obtained w i t h the o r i g i n a l v e r t i c a l t a i l a r e indicated
by the c i r c l e s . This f l u t t e r boundary f e l l within the airplane f l i g h t
regime. I n an attempt t o study the nature of this f l u t t e r , a new system
w a s t r i e d wherein the rudder t w i s t s t i f f n e s s w a s increased t o 1 . 3 t i m e s
the o r i g i n a l value w h i l e actuator s t i f f n e s s w a s reduced t o 0.8 of i t s
value. The squares show where f l u t t e r w a s obtained with the changed
s t i f f n e s s e s . They show that the cbange w a s quite detrimental t o the
f l u t t e r boundary; that i s , with a s t i f f e r rudder and a weaker actuator
the f l u t t e r boundary s h i f t e d t o higher a l t i t u d e s . The coupled frequency
spectrum of the two models remained about the same: ‘up/uu = 0.33 and
qJwu = 0.22.

Some t e s t s were made with viscous dampers on the o r i g i n a l rudder

i n an attempt t o eliminate the f l u t t e r from the f l i g h t envelope of t h e
airplane. The dampers were mounted a t about midspan on the rudder.
F l u t t e r w a s s t i l l obtained when the damping produced by the dampers w a s
60 percent of c r i t i c a l . The f l u t t e r was eliminated when the damping w a s
120 percent of c r i t i c a l . The damping values quoted here a r e based on an
i n e l a s t i c rudder and back-up s t r u c t u r e . Since these assumptions y i e l d
high values of c r i t i c a l damping, the damping a c t u a l l y obtained w a s prob-
ably somewhat less than the s t a t e d value.

The c i r c l e s f o r the o r i g i n a l v e r t i c a l t a i l a r e repeated as c i r c l e s

again i n f i g u r e 11where dynamic pressure i s shown as a function of Mach
number. For this p l o t the f l u t t e r side of the boundary i s above the
curve. Also shown i n t h i s f i g u r e a r e three no-flutter points and a f l u t -
t e r point obtained w i t h v e r t i c a l t a i l G which had a longer and s t i f f e r f i n
than v e r t i c a l t a i l F. As indicated i n the figure, the rudder actuator and
t w i s t s t i f f n e s s f o r v e r t i c a l t a i l G were a l s o s o m e w h a t greater than f o r
v e r t i c a l t a i l F, although u”n/% and up/% were about the same. These
r e s u l t s indicate that the f l u t t e r boundary f o r v e r t i c a l t a i l G would be
a t s u b s t a n t i a l l y higher dynamic pressures than f o r v e r t i c a l t a i l F. Haw-
ever, the one f l u t t e r point f o r v e r t i c a l t a i l G w a s s t i l l considered t o be
too close t o the airplane f l i g h t regime. An adequate f l u t t e r boundary
appeared t o be obtained when viscous dampers were used. The dampers pro-
duced about 15 percent of c r i t i c a l damping.

An additional t e s t w a s made without dampers but w i t h the f i n t i p

c u t off along the dashed l i n e indicated i n figure 1. With the f i n t i p
c u t o f f , v e r t i c a l tail G had about %he same plan form as v e r t i c a l t a i l F
but the frequency r a t i o s w e r e reduced t o cop/ua = 0.25 and d%
= 0.18.
The one f l u t t e r point obtained w i t h the f i n t i p cut off, shown i n f i g -
ure 11, reveals a large increase i n dynamic pressure required t o f l u t t e r
this t a i l compared with v e r t i c a l t a i l F. This increase i n dynamic pres-
sure required t o f l u t t e r the clipped configuration G over that required
f o r configuration F can be a t t r i b u t e d t o increases i n the s t i f f n e s s of
t h e f i n , actuator, and rudder. The rudder t w i s t s t i f f n e s s w a s 2.8 t i m e s
that of the o r i g i n a l configuration and the actuator s t i f f n e s s w a s 1.5
times that of the o r i g i n a l configuration.

Studies of v e r t i c a l t a i l s F and G i n the transonic speed range

have shown that t h e i r f l u t t e r boundaries are s e n s i t i v e t o changes i n the
rudder twist s t i f f n e s s and rudder actuator s t i f f n e s s . F l u t t e r could be
eliminated by using viscous dampers, but the amount of damping required
varied g r e a t l y between the two configurations.
 s -
9

Configuration H - T-tail

Configuration H, shown i n figure 1, i s a T - t a i l which did not have

any movable control surfaces. The torsion and side f l e x i b i l i t i e s of the
fuselage were simulated i n the t a i l mounts. The f i n had an aspect r a t i o
of 1.0, a taper r a t i o of 0.56, and 45' sweepback of the quarter-chord
l i n e . The s t a b i l i z e r had an aspect r a t i o of 1.8, a taper r a t i o of 0.43,
and 40' sweepback of the quarter-chord l i n e . The r a t i o of cantilevered
f i n bending t o f i n t o r s i o n frequencies w a s about 0.55.

The r e s u l t s of the f l u t t e r investigation a r e shown i n figure 12.

S t i f f n e s s - a l t i t u d e parameter i s p l o t t e d as a function of Mach number and
the frequency appearing i n the parameter i s the t o r s i o n frequency of the
f i n . The o r i g i n a l s t a b i l i z e r had 1-5' of positive dihedral and the f l u t -
t e r points obtained with t h i s configuration a r e indicated by the c i r c l e s .
Models with zero s t a b i l i z e r dihedral were a l s o investigated and the
f l u t t e r points obtained w i t h this configuration are indicated by t h e
squares. The results indicate that l e s s f i n s t i f f n e s s i s required f o r
zero s t a b i l i z e r dihedral and so s t a b i l i z e r dihedral i s indicated t o be
detrimental .
The detrimental e f f e c t of s t a b i l i z e r dihedral can be explained
q u a l i t a t i v e l y by considering the s t a b i l i z e r as a yawing wing. When the
f i n t w i s t s , the s t a b i l i z e r yaws. As a result of dihedral this yaw motion
produces a r o l l i n g moment on the s t a b i l i z e r s o that f i n t w i s t r e s u l t s i n
f i n bending.

Simple Models of All-Movable Controls

I n addition t o the previously discussed scaled models, some simple

models of all-movable controls have been investigated a t low supersonic
speeds. As f i g u r e 13 shows, they had arrowhead plan forms - a 45' d e l t a
and one with a 45O sweptback leading edge and a 15' sweptforward t r a i l i n g
edge. The models w e r e f l a t p l a t e s with beveled edges and were supported
by a s h a f t mounted i n a bearing. The f l e x i b i l i t y w a s primarily i n t h e
shaft s o that the control surface experienced only rigid-body torsion,
flapping, and t r a n s l a t i o n . The r a t i o s of f i r s t coupled frequency t o
second coupled frequency varied from about 0.3 t o about 0.7.

The r e s u l t s are presented as the s t i f f n e s s - a l t i t u d e parameter p l o t t e d

against rotation-axis position i n percent of r o o t chord. The frequency i n
the s t i f f n e s s - a l t i t u d e parameter i s the f i r s t coupled frequency of the
system. The trend shown a t a Mach number of 1.2 for both surfaces pro-
duces a more s t a b l e configuration as the r o t a t i o n a x i s i s moved forward.
The two f l u t t e r points obtained a t a Mach number of 1.6 indicate that, f o r
rearward r o t a t i o n axes, more s t i f f n e s s i s required t o prevent f l u t t e r as
the Mach number increases. However, the s t i f f n e s s required a t a Mach
number of 1.6 rapidly decreases as the r o t a t i o n axis i s moved forward s o
t h a t , a t the a x i s location shown i n the figure, the controls were s t a b l e
a t a zero value of the s t i f f n e s s - a l t i t u d e parameter. This means the
controls were f l u t t e r f r e e when unrestrained i n r o t a t i o n and when f r e e
t o f l o a t . Thus, these r e s u l t s show that movement o f ’ t h e axis location
forward w a s b e n e f i c i a l on these plan forms a t low supersonic Mach numbers.

CONCLUDING REMARKS

A compilation of f l u t t e r experiments w i t - , various control conf-gura-

tions a t transonic and supersonic speeds has been presented. Some trends
on s p e c i f i c configurations a r e indicated and improvements i n f l u t t e r
c h a r a c t e r i s t i c s f o r various cases a r e shown. It would be risky t o d r a w
general conclusions from these r e s u l t s , but the discussion should be of
value.

1. Land, Norman S., and Abbott, Frank T., Jr.: Transonic F l u t t e r Inves-
t i g a t i o n of an All-Movable Horizontal T a i l f o r a Fighter Airplane.
NACA RM ~ 5 6 ~ 0 61957.,
2. Ashley, H o l t , and Zartarian, Garabed: Piston Theory - A New Aerody-
namic Tool f o r the Aeroelastician. Jour. Aero. Sci., vol. 2 3 ,
no. 12, ~ e c .1956, pp. ~ 0 9 - ~ 1 8 .

3 . Hanson, Perry W., and Rainey, A. Gerald: Experimental Investigation

of the Transonic and Supersonic F l u t t e r Characteristics of the Upper
and Lower Vertical Tails of an Air-to-Ground Missile. (Prospective
NACA paper. )
 TEST CONFIGURATIONS

A B /

C D E

F G H

Figure 1

TEST CONFIGURATION A
WING WITH TIP AILERONS
NO ACTUATOR
FLUTTER FLUTTER STIFFNESS
0 0 I .o
8 0 3.0
A A a0

6 6

bwa
a

-d4[2
- L
0 .6
8%
STABLE

ooP
I.o
M
FLUTTER

1.4
:j
-L
0 .6
f!!)
&
A
/A

I.o
M
o m &
A

1.4

Figure 2
 TEST CONFIGURATION B
ALL-MOVABLE STABILIZER

'r .50

A .94

0 .6 .7 .8 .9 1.0 1.1 1.2 1.3

M

TEST CONFIGURATION B
ALL-MOVABLE STABILIZER

-----
--- O }CALCULATED
gEXP
41- FLUTTER

0 .5 I .o I.5 2.0
Wh/

i;cr Figure 4 Y
 TEST CONFIGURATION C
ALL-MOVABLE STABILIZER

8 D 1.4I
A A 0.9 I

Q
D

0
2

I I t 1
0 LO 1.5 2.0 2.5 3.0
M

Figure 5

TEST CONFIGURATION C
ALL-MOVABLE STABILIZER

IO

("hP8
1.02

.95 HEAVY L.E.

'k l l ,
0 1.5 2.0 2.5 3.0
M

Figure 6
 TEST CONFIGURATION C
ALL-MOVABLE STABILIZER

FLUTTER FL?&ER wh/w8

1.02 A

CALCULATED
6-

a 0
4 -
8

2- e -
CALCULATED

0 LI I I I I I
.5 1.0 1.5 2.0 2.5 3.0
M

Figure 7

TEST CONFIGURATION D
VERTICAL TAIL WITH T.E. RUDDER

4.0 r

I .2
1.0 Wf

0 .5 1.0 I .5 2.0 2.5 3.0

M

Figure 8
 TEST CONFIGURATION E
T.E. RUDDER
VERTICAL TAIL WITH

FLUTTER

I I I I I I I
0 .5 I .o 1.5 2.0 2.5 3.0
M
C

TEST CONFIGURATION F
VERTICAL TAIL WITH TE. RUDDER

STIFFNESS

5-0 r ACTUATOR TWIST

I
I.o I.o

.8 1.3
-
bw 4
0
a 45

4.O
L I I I I I
0 1.0 1.1 1.2
M

Figure 10
 TEST CONFIGURATIONS F AND G
VERTICAL TAILS WITH T.E. RUDDER
STIFFNESS
NO
CONFIG. FLUTTER FLUTTER ACTUATOR TWIST
F 0 I .o I .o
G m 0 I .5 2.8

4#000 r FIN TIP REMOVED-

-
0

q,LB/SQ F T 3,000 -
0

Figure 11

TEST CONFIGURATION H
T-TAIL

DIHEDRAL
0 15O

I.o - 8 Oo

.8 -
E
!&
.a -
.6 -
FLUTTER
c

0 .6 .7 .8 .9 LO 1.1 1.2 1.3

All

1 GI1 Figure 1 2 c
 FLUTTER OF TWO ALL-MOVABLE CONTROLS

5 r .
+
NO
FLUTTER FLUTTER
o w

4-

3-
34
a
2 -

0 't, I
4
&
M = 1.6

ROTATION
.5
I I
.6
FLUTTER

AXIS, % ROOT CHORD

I I
.7

NACA - Langley Field, Va.

,_-. -.