You are on page 1of 30

LI br ARY

R O Y A L A I ~'','~,~ - -~-,',':' ':-f" L S F A B L I S ' - ' - ~ '~ i c-,''* 1~


. a
R. & M. No . 3 0 7 7
( 1 9 , 8 ~ 0 )
A.R.O. Technical Report
MI NI S T RY OF SUPPLY
AERONAUTI CAL RESEARCH COUNCI L
REPORTS AND MEMORANDA
Methods for Determining the Wave
Drag of Non-Lifting Wing-Body
Combinations
By
L. M. SHEPPARD
Crown Copyright z958.
L O N D O N :
HER MAJESTY'S STATIONERY OFFICE
1958
TE' N SHI LLI NGS NET
Methods for Determining the Wave Drag
Non-Lifting Wing-BodyCombinations
By
L. M. SH~.I'PARD*
of
COMMUNICATED BY THE DII~ECTOI~-GENEr.AL OF SCIENTIFIC RESEARCH "(ATR)
MINISTRY OF SUPPLY
Reports and Memoranda No. 3o77I
A p r i l 1957
Summa~y.~The area-rule, moment of area-rule and transfer-rule methods for estimating the wave drag Of wing-body
combinations are discussed. It is pointed out t hat the moment of area rule and the transfer rule are different forms
of the area rule, and that the transfer rule expresses the interference wave drag in a simple form. The existing methods
of wave-drag estimation are restricted t o combinations with bodies having continuous surface slope andher e an
extension to combinations with bodies having discontinuous surface slope is given. This paper is concerned with the
theoretical methods and their associated numerical techniques and no numerical results .or particular applications
are presented.
1. Introduction.--Important problems in the design of eransonic and supersonic aircraft are
the calculation of the wave drag of t he configuration and the design of the fuselage so t hat the
overall wave drag is a mi ni mum for 'certain specified conditions. Solutions of these two problems
can now be found by using the area rule. The justification of the area rule at transonic speeds
depends solely upon experiment.
Originally the area rule was applicable at sonic speed only and, in this form, is due t o
Oswatitsch, Whi t comb 1 and Lord. ~ Whi t comb I based his formulation of the sonic area rule
upon an experimentally verified equivalence between wings and bodies at sonic speed; it was
shown t hat there does exist a marked similarity between the shock wave pat t erns around a
wing-body combination and t hat body of revolution which has the same cross-sectional area
distribution. This result is clearly of a non-linearised nature since shock waves are considered.
A different form of the sonic area rule is due to Lord ~ who used the supersonic linearised t heory
to derive an expression for t he wave drag ; thus quant i t at i ve theoretical comparisons can be made
between the sonic wave drags of different configurations. The design of opt i mum wing-body
combinations with mi ni mum sonic wave drag has been examined by Lord s, 4 while the calculation
of t he sonic wave drag has been simplified remarkabl y by Emi nt on and Lord 4.
The extension of the area rule to supersonic speeds is due to Jones 5 and Whitcomb, who have
made the implicit assumption t hat the effect of the interference velocity potential on the wave
drag is negligible. Several aspects of this extension have been examined by Lord 6 and WarrenE
An i mport ant presentation of the mat hemat i cal basis of the area rule has been given by Ward s, 9,10,
who also suggested an alternative form, which is described as the transfer rule. A special case
of the supersonic area rule, called t he moment-of-area rule, was given by Baldwin and Dickey 11,
whose result has been obtained i ndependent l y in a different form. It is interesting to note t hat
.the supersonic area rule could be deduced directly from a wing drag result obtained by Hayes 1~
m 1947.
* Attached scientist from the Weapons Research Establishment, Salisbury, South Australia.
t R.A.E. Report Aero. 2590, received 12th June, 1957.
A comprehensive summary of American work on the wave drag of wing-body combinktions
has been given recently by Lomax and Heaslet ~3. This paper 1~ and the references cited therein
contain i mport ant theoretical analyses of the wave drag of configurations under conditions for
which the area rule is not applicable. The area rule applies to combinations of non-lifting thin
wings and slender bodies, which can be represented solely by surface source distributions with the
source strengths proportional to the local surface slopes in the streamwise direction. In certain
circumstances these source distributions may be discontinuous and it appears that, in so far as
the use of the area rule for the estimation of wave drag is concerned, discontinuities in source
distributions give rise to difficulties only when they occur along straight lines inclined such t hat
the component of free-stream velocity normal to the line is supersonic (or sonic), in the case of a
wing, or when t hey occur around the circumference of a cross-section of a body. Such discontinuous
source distributions are not discussed by Lomax and Heaslet ~ and so some consideration is given
to t hem in this report. The cases of lifting configurations, which require the introduction of
surface doublet distributions, and combinations with non-slender bodies, which can be t reat ed
by using axial distributions of multipoles as well as simple sources, are discussed in detail by
Lomax and Heaslet a~ and are not considered here. Furt her information on configurations
incorporating non-slender bodies is given in Ref. 14.
This present report is concerned only with theoretical met hods and their associated numerical
techniques and no numerical results or particular applications are given. In Section 2 a summary
is presented of the theoretical methods for estimating the wave drag of configurations to which
the area rule does apply, namel y combinations of wings and bodies with continuous source
representations, and the problem of body design for low total wave drag is also discussed; the
sonic area rule, supersonic area rule, moment of area rule and transfer rule are t reat ed separately.
In Section 3 extensions are given of the area and transfer rules to configurations incorporating
bodies represented by discontinuous source distributions. The appropriate numerical techniques
are described in Section 4. Since this report is concerned with idealised combinations of wings
and bodies, Section 5 contains some remarks of a more practical nature and it is indicated how
the methods developed for wing-body combinations may be applied to aircraft. Fi nal l y, the
main conclusions of the report are given in Section 6.
2. Combi~atons wi t h Bodies R@resent ed by Cont i nuous Source Di st ri but i ons. - - 2. 1. So~,ic Area
R~l e. - - Bef or e preceeding with a description of the area rule and its alternative forms a st at ement
of the sonic area rule will be given, since this is a simple special case of the general area rule and
thus provides a useful introduction to it.
2.1.1. Est i mat i on of wave drag. - - Consi der a non-lifting configuration of unit length and let
S(x), 0 <~ x <~ 1, be t he axial distribution of cross-sectional area. The area distribution S(x)
must be such t hat S' ( x) ( = dS/ dx) is continuous and S'(O) -- 0 ----- S'(1). Then the sonic wave
drag D is given by
D_ I l og
q 2~ 0 0 -- ' . . . . . . . .
where q is the kinetic pressure and S"(x) denotes d2S/dx 2. An alternative expression for the sonic
wave drag when S"(x) is continuous is given by Legendre 15 i n the form
q 2~ 0x, ( 1 - - x , ) ~ 0 0L x , - - x 2 A
This form has the advantage t hat it is non-singular, introduces only S' ( x) and shows t hat the
drag is positive. A mote general form of this result when S"(x) is not continuous is more complex,
although equation (1) remains valid as it stands.
The development of the area rule began after the work of Whi t comb ~ who obtained experimental
verification of the importance of the cross-sectional area distribution at transonic speeds. How-
ever, the physical interpretation of the sonic wave drag given by equation (1) requires further
elucidation and must depend upon experiment unless an adequate non-linear theory, which
2
is applicable to t he t r ansoni c range, can be found. Following t he suggestions of Lord s, t he sonic
wave drag of a wi ng- body combi nat i on ma y be i nt er pr et ed physi cal l y as a measure of t he sudden
t ransoni c drag rise of t he configuration, when such a drag rise can be defined.
I t is wor t hy of not e t h a t t he usefulness of equat i on (1) is gr eat l y enhanced by t he fact t ha t
t he l i neari sed t heor y predi ct s zero l i ft -dependent wave drag at M = 1. Thus equat i on (1) is
not rest ri ct ed to non-l i ft i ng configurations since i t gi ves t he t ot al sonic wave drag of a l i ft i ng
confi gurat i on (for a discussion of l i ft -dependent wave drag, see for example, Lomax and Heasletla).
2.12. Des i gn f o r l ow wave d r a g . - - T h e design of a wi ng-body combi nat i on wi t h low sonic wave
drag requi res t he det er mi nat i on of opt i mum area di st ri but i ons whose associated wave drag,
gi ven by equat i on (1), is a mi ni mum. Lord and Emi nt on have i nvest i gat ed t hi s probl em and
present ed opt i mum area di st ri but i ons for a number of specified conditions. Thus in order to
obt ai n low sonic wave drag i t is necessary to al t er t he body, and possi bl y t he wing also, so t h a t
t he cross-sectional area di st ri but i on becomes an opt i mum, or near opt i mum. A simple exampl e
of an opt i mum area di st ri but i on is t hat of a so-called Sears-Haaek body, whi ch i s poi nt ed at.
bot h ends and has mi ni mum wave drag for gi ven l engt h and volume.
2.2. Ar e a Ru l e . - - 2 . 2 . 1 . Es t i ma t i o n o f wave dr ag. - - - The area-rule met hod of est i mat i ng wave
drag is due to Jones 5 and Whi t comb. However, t he deri vat i on of t he area rule requires t he use
of an assumpt i on even t hough t he wi ng-body combi nat i on is represent ed by source di st ri but i ons
only. The assumpt i on is t hat t he effect of t he i nt erference vel oci t y pot ent i al on t he wave drag is
negligible, i . e. , t he per t ur bat i on vel oci t y pot ent i al of a wi ng-body combi nat i on is assumed to be
equal to t he per t ur bat i on vel oci t y pot ent i al of t he i sol at ed exposed wing t oget her wi t h t he
per t ur bat i on vel oci t y pot ent i al of t he i sol at ed body.
I t is conveni ent to define an el ement al area di st ri but i on. Consider t he f ami l y of paral l el Mach
pl anes i ncl i ned at t he Mach angle to t he body axis and such t hat t he ort hogonal pl ane cont ai ni ng
t he body axis is i ncl i ned at t he angle 0 to t he ' pl ane of t he wi ng '. A gi ven f ami l y of Mach planes
cuts t he wi ng- body combi nat i on obl i quel y and t hus defines an oblique sect i onal area di st ri but i on
s( x, O, M) , where M is t he Mach number under consideration. The cross-sectional area di st r i bu-
t i on S ( x , O, M) = s(x, O, M) sin if, where ff is t he Mach angle, is defined to be an el ement al area
di st ri but i on. I t s associated wave drag is gi ven by equat i on (1) as
q - - 2~ S' (x~,O,M) S" ( &, O, M) l og ] x ~ - - & [ d &d &; . . . . (2)
where S" ( x, O, M) denot es ~2S(x, O, M) / ~ x ~ and i t is assumed t hat S ' ( x , 0, M) is cont i nuous
everywhere. When M = 1 i t is seen t ha t all t he el ement al area di st ri but i ons are t he same, bei ng
equal to t he cross-sectional area di st ri but i on.
Hayes TM and Heasl et , Lomax and Sprei t er 16 obt ai ned a resul t whi ch applies to wings onl y but
i t s appl i cat i on to wi ng- body combi nat i ons follows i mmedi at el y from t he assumpt i on t hat t he
effect of t he i nt erference vel oci t y pot ent i al on t he wave drag is negligible*. The wave drag of
t he configurat!on t herefore can be expressed i n t he form
D = 2 7 j o D(O, M) dO , . . . . . . . . . . . . . . . . (3)
i . e. , t he wave drag of a wi ng- body combi nat i on at a Mach number M is t he mean of t he wave
drags associ at ed wi t h all t he el ement al area di st ri but i ons. This is a complete st at ement of t he
* I n a pr i vat e c ommuni c a t i on L. E. Fr a e nke l has s hown t h a t t hi s as s umpt i on is val i d f or smal l val ues of t he r at i o of
b o d y r adi us t o wi ng chor d. Thi s condi t i on ma y be der i ved b y consi der i ng a conf i gur at i on wi t h a ci r cul ar cyl i ndr i cal
b o d y and t h e n obs er vi ng t h a t t he i mp o r t a n t pa r a me t e r of an i nt er f er ence flow field is BR,/c wher e B = ~/ ( M 2 -- 1),
R, is a t ypi cal b o d y r adi us a nd c is t he wi ng- r oot chord. Thus t he i nt er f er ence vel oci t y pot ent i al can be negl ect ed
f or smal l val ues of BR,/c because i t s effect vani shes when R, = 0 a nd when B ----- 0 (i.e., M ---- 1). Wh e n t he a bove
condi t i on is not sat i sfi ed t he er r or i nt r oduc e d b y t he negl ect of t he i nt erfererfce vel oci t y pot ent i al is not known.
However , i t is possi bl e t h a t t he er r or will not be l arge because t i l e i nt er f er ence vel oci t y pot ent i al can be negl ect ed
al so whe n t he r at i o BR,/c is l arge, a n y i nt er f er ence effects vani s hi ng whe n c = 0 or B = oo.
3
(71666) A*
area rule originally given by Jones 5. Ward 8 has derived t he same result by using operational
methods, while Graham, Beane and Licher 17 have also given this result. When M = 1, equations
(2) and (3) reduce to the sonic area-rule result, equation (1).
It is i mport ant to note t hat the derivation of the area rule (equation (3)), required all the
elemental area distributions to be such t hat S' ( x , O, M) is continuous everywhere. Nevertheless,
provided t hat almost all the elemental area distributions satisfy this requi rement the wave drag
may still be given by equation (3). This condition is much less restrictive t han the original one*.
An elemental area distribution of a thin-wing-slender-body combination can be written in the
approximate form S( x, O, M) = S(x) + Sw(x, O, M) , where S(x) is t he cross-sectional area distri-
bution of t he body and Sw(x, O, M) is the contribution from the exposed wing. The wing is t aken
to lie in the plane z = 0 (see Fig. 1) and its thickness at the point. (x, y) is denoted by T( x, y) .
Consider any Mach plane cutting the axis at the point x = xl. Then this Mach plane cuts the
wing obliquely along an area which, when projected on to the plane x = xl, defines the area
Sw(x, O, M) . This projection may be regarded as a projection on to a plane normal to the wing
followed by a projection on to the plane x = xl. The former projection defines an area equal
to the wing thickness i nt egrat ed along the line in which t he Nach plane cuts t he wing, i . e. , -
x -- xl = y t an v where tan ~ = cot/~ cos 0. I t follows t hat
Sw(x, O, M) = f T ( x + Yl tan v, Yl) dyl .
T#O
Heaslet, Lomax and Spreiter 16 obtained this result in an equivalent form.
The above result applies to a wing lying in the plane z = 0. However, its extension to a more
general wing is not difficult and follows i mmedi at el y from the definition of an elemental area
distribution. The general result, which includes the previous result as a special case, can be
written in t he form
Sw(x, O, M ) = f T (f f ) dr ,
T#O
where an element of the wing plan-form is denot ed by dx dr, a general point in the co-ordinate
system of Fig.. 1 is denot ed by R = (x, r), R = (x + Be0 r, r), where B = cot #, ro is t he unit
vector (cos O, sin 0) and 0 is defined by y = r cos 0, z = r sin 0.
2.2.2. Desi gn f or low wave dr ag. - - The problem of reducing the supersonic wave drag of a
wing-body combination with a given wing was examined first by Jones 5. At sonic speed, a body
may be modified so t hat the equivalent body, which has the same cross-sectional area distribution
as the wing-body combination, is an opt i mum with the same total volume, i.e., the modified body
is formed by shaping the opt i mum equivalent body in the region of the wing. At higher speeds
t he wing may be influenced by, and may itself influence, a larger part of the body. Thus t he
corresponding opt i mum equivalent body shaping extends outside the region of the wing and
may be shown to be over t hat part of the body within t he Mach di amond of the wing. Details
of this drag-reduction procedure are given in section 2.4.2. where the transfer rale is discussed.
Anot her problem of interest is t hat of reducing the wave drag of a general configuration and
this has been discussed by Jones 5 in a qualitative form. All the elemental area distributions
should be such t hat their associated wave drags are as low as possible. The design of combinations
* Since this report was completed, Lock TM has shown that the wave drag of a rectangular wing can be found using
the area rule (equations (2), (3)). Thus the smoothness condition that S'(x, O, M) be continuous everywhere appears
to be unnecessary for the wing of a configuration.
satisfying this requi rement appears to be very difficult except in certain special cases ; for example,
at low supersonic speeds only a small number of elemental area distributions are required and
so an iteration met hod can be used to obtain a modified configuration with low wave drag.
In t he case of an isolated wing it is interesting to use the area rule to determine the opt i mum
thickness distribution for mi ni mum wave drag, t he plan-form and volume of t he wing being
given. A particular case of this problem was first examined by Jones 1 who showed t hat if t he
plan-form is elliptic t hen t he wing section must be parabolic-arc biconvex with thlckness/chord
ratio proportional to local chord. However, Graham, Beane and Licher 17 have used t he area
rule to obtain the same result and have shown t hat all t he elemental area distributions represent
Sears-Haack bodies (i.e., bodies of mi ni mum wave drag for given length and volume.) The
result for the wave-drag coefficient of the opt i mum wing of elliptic plan-form is
~A~B ~ ~A~B~/ ~
where CD is based upon the wing plan-form area
A is t he aspect ratio of the wing
*0 is t he thickness/chord ratio at the wing root.
Also , o = (32V)/(~Ac3), where-V is t he volume of t he wing and c is t he wing-root chord.
2.3. Mome nt of Area Rul e . - - 2. 3. 1. Es t i mat i on of wave d r a g . - - Th e moment of area-rule met hod
is due to Baldwin and Dickey ~, who expanded the wave drag in a series o.f powers of
B~- - - (M s - 1). The results presented later in this section are similar and were obtained
i ndependent l y of Baldwin and Dickey ~ whose results will be described first.
The first step towards a series solution in ascending powers of B ~ was made by Adams and
Sears ~ who developed a ' not -so-sl ender' wing theory by obtaining an expansion up to the first
power of B ~, t he t erm i ndependent of B ~ being the slender-wing solution. These results emphasise
t he fact t hat the series solution is actually an expansion in ascending powers of B2k ~ where k is
the slenderness paramet er introduced by Adams and Sears ~. For example, in some wing problems
k may be defined as t he aspect ratio.
Before proceeding with a description of the moment of area rule developed by Baldwin and
Dickey 11, it is necessary to obtain an alternative form of equation (1). Using t he substitution
x = (1 -- cos ~), 0 ~< x ~< 1, S' ( x) is expressed as a Fourier sine series in t he new variable ~ ;
namel y
S' ( x) = ~ A, ~s i nn~, 0 ~< ~ ~< ~.
Substituting this series in equation (1) shows t hat the wave drag is given by
D 7~ oo
- - ~. nA,,~ .
q
Therefore, using similar substitutions equation (3) can be written as
q s o ,~=1
where A, (O. M) is t he Fourier coefficient appropriate to the elemental area distribution S( x, 0, M) .
Baldwin and Dickey n expanded t he Fourier coefficient A,,(O, M) in a series of powers of
5
B ~ = (M ~ - 1) and i t is found t hat , wi t hi n t he usual order of approxi mat i on, t he wave drag
of a configuration wi t h t he wing in t he pl ane z = 0 (see Fig. 1) depends solely upon t he moment s
of area defined by
= s ( x ) + t
T(x,
Y)
dy
d T.~ 0
and
f f
M, (x) = | T(x, y) yi dy, i even, i >~
2 ,
d T#O
where S(x) is t he cross-sectional area di st ri but i on of t he body and, in t he not at i on of Fig. 1,
T( x, y) is t he wing t hi ckness at t he poi nt (x, y). Clearly, Mo(x) is equal to t he cross-sectional
area di st ri but i on of t he wi ng-body combi nat i on and M~(x) is equal t o t he di st ri but i on of t he
cross-sectional pol ar moment of i nert i a of t he wing.
The final result for t he wave drag may be wr i t t en in t he form D/q = ao + a=B ~ -/O(B~), where
a0 depends upon Mo(x) and a~ depends upon bot h Mo(x) and M~(x). The higher t erms in t hi s
series are more complex and so, as Bal dwi n and Di ckey *t have st at ed, t he met hod is simple t o
appl y onl y at low supersonic speeds. Nevertheless, wi t hi n t hi s speed range it does offer a simple
qual i t at i ve met hod for reduci ng wave drag; a useful requi rement is t ha t t he moment of area
di st ri but i ons be as smoot h and ' sl ender ' as possible. I t is suggested t hat , i n general, i t will be
sufficient to consider onl y Mo(x) and M~(x).
Begi nni ng from equat ons (2) and (3) it has been shown t hat , for a confi gurat i on of uni t length,
t he wave drag is given by
q - o o log
+ - ~ {~,~(&)} ~-~ {~,~(&)} log Ix~-- x~[ d &d & + O(B~), . . . . (4)
where %, ,(x), J i , a(x) are generalised moment s of area defined by
1 Ri+2(x, O) cos 40 dO
- i + o
1 ff~R~' ~(x, 0) sin 40 dO
0
where r = R(x, 0) is t he equat i on of t he cross-section at t he st at i on x. I t is assumed t ha t t he
series in equat i on (4) is convergent and t hat all necessary cont i nui t y and di fferent i abi l i t y con-
di t i ons are satisfied. Furt hermore, since t he body is slender i t is i mpor t ant t o not e t ha t t he
expressions for t he generalised moment s of area can be simplified. Wi t hi n t he usual order of
approxi mat i on, t he cont r i but i on of t he body can be neglected for all %,~, ~, ~( i # 0). Thus i n
equat i on (4) t he body cont ri but es t o t he wave drag onl y in so far as does S(x), t he cross-sectional
area di st ri but i on of t he wi ng-body combi nat i on. For a wing symmet ri cal about t he planes
0 = O, ~/2 (see Fig. 1) t he generalised moment s of area can be wri t t en in t he simpler f or ms
~'i,a(x) = 0, Z odd,
Y,,a(x) ---- O, all ,~,
and since ~, 0(x) is t he i t h moment of area, i t follows t ha t ~, 0(x) = M~(x)~
6
Thus equat i on (4) becomes
B 2
D = D{S} -J---4 [D{S + M( ' } - - D { S } - D{M('}] -}- O(B4), . . . . . . (5)
Wahere D{S}, D{M(' }, D{S + M( ' } denot e t he wave dr ag o f ' b o d i e s ' wi t h t he cross-sectional
r e d' di s t r i but i ons S(x), M( ' ( x) , S(x) + M(' (x) respect i vel y and i t has been assumed t ha t
M(H(x) is cont i nuous everywhere. At first si ght t hi s condi t i on seems to l i mi t ver y seri ousl y t he
appl i cabi l i t y of equat i on (5) but i t is not undul y rest ri ct i ve sincel for example, i t is satisfied b y
an unswept , t aper ed wi ng wi t h st r ai ght edges and bi convex section of const ant t hi ckness ratio.
Moreover, t he condi t i on is satisfied by t he opt i mum second-moment di st r i but i on gi ven in Section
2.3.2, al t hough i t is not satisfied by t he Jones elliptic wi ng discussed in Section 2.2.2. When
t he wi ng is defined numeri cal l y, equat i on (5) has t he serious di sadvant age t ha t t he wave drag
cannot be det er mi ned ver y accur at el y since i t is necessary to carry out t wo successive numer i cal
di fferent i at i ons in order t o cal cul at e M2"(x).
The i mpor t ance of equat i on (5) lies in t he fact t hat i t reduces t he cal cul at i on of t he wave drag
of a ' not -so-sl ender ' confi gurat i on to t he cal cul at i on of t he wave drag of t hree ' bodies '. Thus
numeri cal eval uat i on of t he wave drag is consi derabl y simplified and can be carried out usi ng
t he met hods of Section 4.2. However, i t shoul d be emphasi sed t ha t bot h S' (x) and M( "( x) must
be cont i nuous everywhere before equat i on (5) can be used. Therefore t he Fourier-series met hod
described earlier in t hi s Section ma y be used, t he subst i t ut i ons bei ng
x = (1-- cos g) , 0 ~<x ~< 1,
S' (x) = ~ A,, sin n9
n = l
and
M21;r(x) = ~ B , , sin n9
Equat i on (5) now becomes"
D
q 91: n = l V'T ~ = 1
I t will be not i ced t ha t onl y t he coefficients B,, for whi ch t he correspondi ng coefficient A,, is non-
zero appear in t he wave drag. Thus t hi s form of t he wave drag is useful when most of t he A,,
vani sh. I t is especi al l y useful when t he area di st ri but i on S(x) is an opt i mum di st r i but i on wi t h
a finite Fouri er series. For exampl e, t he Sears-Haack body has A,, = 0, n # 2.
2.3.2. Design f or low wave dr ag. - - I t has al r eady been emphasi sed t ha t t he mai n usefulness of
t he moment of area rule is as a qual i t at i ve design procedure for low supersonic speeds. Bal dwi n
and Di ckey 1~ have exami ned t he use of t he met hod i n designing configurations wi t h mi ni mum
wave drag under specified conditions. The procedure used i s a st ep-by-st ep opt i mi sat i on t echni que
wher eby successive t erms in t he series sol ut i on are opt i mi sed one aft er t he other. Thus a uni que
set of moment di st ri but i ons is found.
The opt i mum set of moment di st ri but i ons for mi ni mum wave drag, t he l engt hs and t he
' v o l u me s ' bei ng fixed, is gi ven by Bal dwi n and Di ckey 11. I n t hi s cont ext t he ' vol ume " is
t he p t h moment ' vol ume '* of t he confi gurat i on defined by
V~ = f' P M / x) dx ,
0
* The p t h mo me n t ' v o l u me ' does not ha ve a n y phys i c a l si gni f i cance unl es s 15 ~ 0 whe n i t is t he a c t u a l v o l u me of
t h e conf i gur at i on.
7
where lp is t he l engt h of t he pt h moment di st ri but i on, i.e., l0 ---- t, t he l engt h of t he combi nat i on
and, for p ~> 2, Ip = l~ , t he l engt h of t he proj ect i on of t he wing on t he body axis. The resul t is
t ha t t he opt i mum moment di st ri but i ons whi ch vani sh at x = 0, I p , are
[-
/
Z
z , / 2 ] J L
\
p e v e n , O~<x~<I p
and t ha t t he wave drag of t he opt i mum confi gurat i on is given by
_ _ F M ( I , 1 , _ _ / 2 ) _
D_ 4 1 1 s [ Mo ~ l l 2 ) - ] + 1 6 5 3 7 . 5 L l~ ~ j B ~ I + O ( B ~ ) .
q
I t will be not i ced t ha t no t er m in B ~ appears in t hi s result.
Since all t he t erms in t he above expression for t he wave drag are positive, i t has been poi nt ed
out by Bal dwi n and Di ckey 11 t ha t t he wave drag of a configuration, whi ch is opt i mum for t hei r
conditions, must increase monot oni cal l y wi t h i ncreasi ng Mach number and be ver y large when
t he second t er m is ver y much great er t han t he first t erm, i.e.,
( M o ( Z / 2 ) Z Z
B - - - 0 ]"
Thus t he Mach number i ndi cat ed by t hi s rel at i on can be used to find an obvious upper l i mi t to
t he range of appl i cabi l i t y of t he opt i mum t echni que. Thi s emphasises once agai n t he rest ri ct i on
of t he met hod to low supersonic speeds and t herefore t hat i t is necessary to consider onl y t he
first few t er ms of t he series solution. Bal dwi n and Di ckey 11 have poi nt ed out t ha t t he st ep-by-st ep
opt i mi sat i on t echni que fails at hi gher supersonic speeds because i t el i mi nat es all except posi t i ve
definite t erms in t he expression for t he wave drag. I t appears t ha t a more real i st i c opt i mi sat i on
procedure, based on equat i on (5), will be obt ai ned if t he t erm in B ~ is negat i ve i nst ead of zero
as i n t he above procedure. The fact t ha t t he t er m in B 2 can be negat i ve is shown r eadi l y b y
considering an unswept wing, such as t he Jones elliptic wi ng discussed in Section 2.2.2.
Exper i ment al confi rmat i on of t he usefulness of t he moment of area rule for wi ng-body com-
bi nat i ons has been gi ven by Bal dwi n and Di ckey 11. Fi r st l y, t he opt i mum area di st ri but i on was
found and t hen an opt i mum second moment di st ri but i on was obt ai ned by addi ng bodies of
revol ut i on near t he wi ng t i ps while modi f yi ng t he body to r et ai n t he opt i mum area di st ri but i on.
The t est s i nvol ved an unmodi fi ed configuration, one wi t h t he opt i mum area di st ri but i on as well
as one wi t h bot h t he opt i mum area di st ri but i ons and, i t will be not ed, an opt i mum, l engt hened
second moment di st ri but i on. By considering t he difference bet ween t he wave drag of t he wi ng-
body combi nat i on and t he wave drag of t he body alone t he agreement bet ween t heor y and
exper i ment was found to be good over t he Mach-number range M = 1. 0 to M = 1.4, for an
elliptic wi ng of aspect rat i o 2. At M = 1. 4 t he moment of area-rule model had a sl i ght l y lower
wave drag t ha n t he unmodi fi ed configuration. Thus t he sonic area rule is ext ended to low
supersonic speeds by usi ng t he moment of area rule. However, because 12 has been i ncreased
wi t hout changi ng M~(0), t hese exper i ment al resul t s d o n o t j ust i f y anyt hi ng more t han t he use
of smoot h, ' sl ender ' second moment di st ri but i ons.
I t will be not ed t ha t t he above t echni que requi res t he addi t i on of bodies to t he wing. Thi s is
in cont r ast to t he met hod described in Section 2.2.2, wher eby t he mai n body onl y was shaped
for mi ni mum wave drag. Thus i t would be desirable to exami ne combi nat i ons wi t h subsi di ary
bodies near t he wi ng t i ps by using t he met hods appl i cabl e at any supersonic speed, namel y,
t he area rule and t he t r ansf er rule. The probl em of shapi ng such subsi di ar y bodies is exami ned
i n Section 5.2.2.
2. 4.
War d s, 9, ~0, who has based his consi derat i ons upon t he drag of source di st ri but i ons.
obt ai ned by War d I is t ha t
D 1
q - - 2= ; f T"(R1)T"(R2)cosh-* ( [ &- &]
( **- . q
f f ' , > V c l , ex=
Transfer Rule. --2. 4. 1. Estimation of the wave drag. - - The t ransfer rule met hod is due t o
The result
where d~, d~
R I J ~2
PI ~ [' 2
T "
. . . . . . (6)
denot e elements of area of t he wing surface
denot e poi nt s on t he wing surface (i.e., using Fig. 1, R is t he vect or (x, y, z))
denot e t he vect or di st ance of t he elements d~,, d~,~ from t he body axis (i.e.,
using Fig. 1, r is t he vect or (y, z))
denotes a"T/ax", T bei ng t he wing thickness,
and t he i nt egrat i ons are t o be t aken over real values of t he i nt egrands. War d ~ has not pl aced
any rest ri ct i ons on t he val i di t y of equat i on (6) but i t is val i d onl y if t he effect of t he i nt er f er ence
vel oci t y pot ent i al on t he wave drag is negligible.
I t is 0t i nt erest to not e t ha t if t he t hree t erms of equat i on (6) are denot ed by D~, Dw~ and DB
respectively, t hen equat i on (6) can be wr i t t en in t he form D = D w + DwB q- DB, where Dw is
t he i sol at ed exposed wing wave drag, DwB is t he i nt erference wave drag and Dz is t he i sol at ed
body wave drag. Following War d 1 t he i nt erference wave drag Dw~ can be expressed in a simpler
form and a useful al t er nat i ve to equat i on (6) derived. Put
1 f dV , . . . . . . . . ( 7 )
A(x) = ~ V{B=r, =_ ( x - - &)2}, . . . .
where dV, = T(R1) d~, is an el ement of vol ume of t he wing and t he i nt egr at i on is over t ha t par t
of t he wing for whi ch I x - - x~l ~< B[ r , ] . Thi s shows t hat A(x) vani shes outside t he Mach
di amond whi ch encloses t he wing. Fur t her mor e, War d s shows t ha t equat i on (7) can be used for
wings carryi ng slender bodies (e.g., t i p t anks).
Using equat i on (7) and assuming t hat ~A/~x is cont i nuous for all values of x, equat i on (6) can
be wr i t t en as
1 ( I x 1 - x21
q
1 f f A"(xl)A"(x2)log ([& -- x2l) d& d&
+N
f f [ S ' t ( X l ) - { - A t t ( X l ) ] ~ . ~ t t ( X 2 ) @ A / t ( ~ 2 ) ] l o g ( [ . 5 ~ 1 - - X 2 ] ) ' d x l d X 2 ( S )
2~
where A"(x) denotes ~2A/~x2. Thi s result enables t he i nt erference wave drag t o be expressed in
t he si mpl er form
Dw. - - It X - -
( o )
q
Thus DwB depends upon Mach number since A (x) does.
9
The t ransfer rule is t he name suggest ed by War & for t he wave drag resul t gi ven by equat i on (8).
Therefore A(x) will be referred to as t he t ransferred area di st ri but i on of t he wing. I t follows
from equat i on (7) t hat t he t ransferred area di st ri but i on can be det er mi ned by spreadi ng each
el ement of vol ume of t he wi ng over t he x axi s accordi ng to t he law dV~/~/{B~r~ ~ -- (x -- xd~}.
Hence a vol ume el ement at t he poi nt (xx, r~) cont ri but es to t he t r ansf er r ed area di st ri but i on onl y
wi t hi n t he range x~ -- B I rl I < x < x~ + B I r~ l- I n addi t i on, t he vol ume of a body wi t h t he
cross-sectional area di st r i but i on A (x) is equal t o t he vol ume of t he wing.
The work of Jones 5 suggests an al t er nat i ve met hod for cal cul at i ng A(x). At t ent i on is con-
cent r at ed upon a poi nt on t he x-axis r at her t han a poifit on t he wing. Then i t is i mpl i ci t i n t he
work of Jones 5 t hat A(x) must be equal to t he mean wi ng el ement al area di st ri but i on. Thi s
resul t can be verified for a wi ng l yi ng t he tile (x, y) pl ane bu put t i ng dV1 = T(RI) dx~dy~ and
changi ng t he vari abl e of i nt egr at i on x~ in equat i on (7) from x~ to 0 = cos -~ { ( x~- x)/Byl},
0 ~< 0 ~< =. Thus t he double i nt egr at i on in equat i on (7) is replaced by an i nt egr at i on wi t h respect
to y, followed by an i nt egr at i on wi t h respect to 0 and, using Section 2.2.1, it can be shown t hat
1 ST(X, 0 M) dO A(x) = , ,
where Sw(x, O, M) defines a wing el ement al area di st ri but i on.
Now, r et ur ni ng to a consi derat i on of t he i nt erference wave drag Dw~, t he wave drag, gi ven by
equat i on (8), is wr i t t en i n t he form
D = Dw + D{S + A}- D{A}, . . . . . . . . . . ( 1 0 )
where D{S + A}, D(A} denot e t he wave drag, gi ven by equat i on (1), associated wi t h t he area
di st ri but i ons S(x) + A(x), A(x), and Dw is t he wave drag of t he exposed wing. Equat i on (10)
shows t hat t he i nt erference wave drag can be expressed as DwB = D{S + A} -- D{ S } - ,D{A}.
For a slender wing, Dw = D{A} to t he order of appr oxi mat i on considered and so D = D{S -1- A},
vchich is t he st andar d resul t of sl ender-body t heory.
However, equat i on (10) is not in a form sui t abl e for comput at i on since Dw is t he wave drag
of t he exposed wing, not t he wave drag of some sui t abl y chosen gross wing. Denot e a sui t abl y
chosen gross wi ng by W0 and t he port i on bl anket ed by t he body by A W = W0 -- W. Fur t her -
more, denot e t he body source r epr esent at i on by Q~, t he wing source represent at i on by Qw and
so on. Then, t he f undament al assumpt i on t hat t he i nt erference vel oci t y pot ent i al can be negl ect ed
shows t hat t he source di st ri but i on represent i ng t he wi ng-body combi nat i on is Q = Qw + Q~.
Therefore Q = Qwo + QB -- Q~w, where Q~w is t he source di st r i but i on represent i ng t he port i on of
t he gross wi ng bl anket ed by t he body. Thus Q~w does not necessari l y represent a slender body
al t hough A W is a l ow-aspect -rat i o wing. But , QAw can be i ncorporat ed wi t h t he sl ender-body
source di st r i but i on QB, i.e., for t he purposes of wave drag eval uat i on Q~w may be replaced by t he
equi val ent sl ender-body source di st ri but i on Q~A, where A A is t he (transferred) area di st r i but i on
of t he bl anket ed wi ng A W. The wave drag of t he confi gurat i on can now be simplified and
ec~uation (10) becomes
D = Dwo + D{S + A} -- D{A + AA} , . . . . . . . . (11)
i t bei ng assumed t hat AA' (x) is cont i nuous everywhere. Thi s resul t is advant ageous when
Dwo is known from exi st i ng resul t s in supersonic wi ng t heory.
An al t er nat i ve met hod of changi ng equat i on (10) i nt o a form sui t abl e for comput at i on is to
replace Dw by DWN, t he wave drag of t he net wing. To const ruct t he net wing, t he segment s of
t he exposed wi ng are pl aced t oget her to form a cont i nuous wing. Thi s procedure is not un-
reasonabl e because Dw = DWN at M -= 1, oo. The fact t ha t Dw ~ DWN at ot her Mach number s
seri ousl y l i mi t s t he accuracy of any resul t s obt ai ned by set t i ng Dw = Dw~v i n equat i on (10).
10
I t has been poi nt ed out t ha t t he moment -of-area rule is applicable at low supersonic speeds and,
since t hi s met hod is useful, i t is desirable to i nvest i gat e t he t ransfer rule at such speeds. Fr om
t he resul t s gi ven in t he previ ous discussion i t will be seen t hat a simpler expression for t he
t ransferred area di st r i but i on of t he wing syst em is required. Thi s is r eadi l y found at M = 1 since
t he t r ansf e/ r ed area di st ri but i on is t hen equal to t he cross-sectional area di st ri but i on of t he wing.
At low supersonic speeds t he t ransferred area di st ri but i on A (x) can be replaced by a di st ri but i on
deri ved as follows.
Consider a narrow st reamwi se st ri p of t he wing at a di st ance R from t he axis and wi t h t he
cross-sectional area di st r i but i on ~(x). I t has been shown t hat A(x) is equal to t he mean wi ng
el ement al area di st ri but i on and so, denot i ng t he cont r i but i on of t he st reamwi se st ri p to A (x) by
bA (x), i t follows t hat
dA (x) = ~ ~( x - - B R cos 0) dO
o
o r
2 ~a/~
aA(x) = ~J o , {~(x -- B R cos O) + ~( x + B R cos 0)} dO .
Thus, when B R = 0, dA (x) ----- V(x) and A (x) is equal to t he cross-sectional area di st r i but i on of
t he wing. -When B R is small t he i nt egr and is al most const ant and may be replaced by
{~(x -- B R cos ~/4) + N( x + B R cos ~/4)}, since 0 = ~/4 is a mean val ue of 0 in t he range of
i nt egrat i on, i.e.,
B R
and, if 7x"(x) is cont i nuous everywhere,
aA(x) _ +
(this l ast resul t may be deri ved by expandi ng A (x) in ascendi ng powers of B~R~).
Therefore, ~it low supersonic speeds where B R is small, t he simplified procedure for cal cul at i ng
A (x) is to t ransfer an el ement of wi ng cross-sectional area d ~, si t uat ed at t he poi nt (xl, rl) by
pl aci ng el ement s d~112 at t he poi nt s Ix -- xl[ = Bi t 1] X/2. Thi s procedure can be appl i ed easi l y
and qui ckl y.
2.4.2. Desi gn f or low wave d r a g . - - Th e probl em of det ermi ni ng t he opt i mum fuselage when t he
wing is gi ven has been exami ned by War d 9,10. Fr om equat i on (10), i t follows t ha t t he wave drag,
at a gi ven Mach number, will be a mi ni mum when t he wave drag, D{ S - / A } , of t he ' body ' wi t h
t he cross-sectional area di st ri but i on S(x) -k A( x) is a mi ni mum. Thus t he probl em has been
reduced to an equi val ent sl ender-body probl em whi ch can be t r eat ed usi ng t he met hods gi ven by
Lord and Emi nt on. For a given wi ng and Mach number, t he opt i mum body can be found but
i t will be not i ced t hat t he new exposed wi ng usual l y differs sl i ght l y from t he gi ven wing.
Fol l owi ng War d 1 t hi s difference is t aken i nt o account by combi ni ng i t wi t h t he body. However,
since t he effect on t he body shape is small, t hi s addi t i onal refi nement can be negl ect ed if i t is
desired to si mpl i fy t he analysis.
The appl i cat i on of t hi s procedure at low supersonic speeds can be compared wi t h t he moment -
of-area rule met hod, whi ch is described in Section 2.3.2. The area-rule and t ransfer-rul e met hods
consider confi gurat i ons wi t h fixed wings and derive a body shape which, i t will be noted, vari es
wi t h Mat h number. On t he ot her hand, t he moment -of-area rule considers var i at i ons i n wi ng
geomet ry, such as t he addi t i on of bodies near t he wi ng tips, and derives one body shape for t he
ent i re low supersonic Mach-number range.
A more general mi ni mum drag probl em arises if, for example, onl y t he wi ng pl an area, span
and vol ume are fixed as well as t he body l engt h and volume. Such problems have not been
exami ned in t he l i t er at ur e al t hough War d 9 has made some r el evant comment s, when t he Mach
11
di amond enclosing t he wing lies wi t hi n t he body l engt h. He shows t hat t he wave drag Of a
wi ng- body combi nat i on subj ect to these rest ri ct i ons cannot be reduced to t hat of t he Sears-Haack
body wi t h t he same t ot al vol ume and l engt h. Thi s resul t does not necessari l y i mpl y t ha t t he
wave drag cannot be reduced to a low value.
I t is suggest ed by War d 9 t h a t t he fuselage shoul d al ways be designed sufficiently long for t he
wi ng syst em to be i ncl uded wi t hi n i t s Mach di amond, i.e., S(x) is chosen so t hat i t is non-zero
wherever A (x) is non-zero.
An i mpor t ant class of opt i mi sat i on problems has been exami ned by Lomax and Heasl et 13, who
have i nvest i gat ed t he opt i mum fuselage design for a r ange of Mach numbers, t he wi ng bei ng
fixed (it has been shown earlier in t hi s Section t hat D{S + A} is t hen t he onl y t er m dependi ng
on t he fuselage). Lomax and Heasl eP 3 consider in effect t he mi ni mi sat i on of t he i nt egral
{s + A} d i n ,
D
where M~ ,<. M. ~< M,~ is t he range of Mach number bei ng exami ned and f(M) is a wei ght i ng
funct i on appropri at e to t he par t i cul ar desi gn problem. For ~xample, if t ransoni c acceleration is
i mpor t ant t he nf ( M) can be chosen to emphasi se t he t ransoni c cont r i but i on to t he above i nt egral .
Let D {S + A} denot e t he ' me a n ' val ue defined by
q
- - = f (M ) {S + A} dM . . ~' ~f ( M ) dM
q J M l / g/I,Z 1"
Then, usi ng t he double i nt egral of equat i on (1) i t can be shown t hat
D {s + A) - (A} = D {S + {X} .
Therefore, for a fixed wing (i.e., A (x) fixed), mi ni mi si ng (D/q){S + A} is equi val ent to mi ni mi si ng
(D/q) {S -~- z{}. Thi s l at t er probl em is of t he same t ype as t hat al r eady considered earlier in t hi s
section and so t he mi ni mi sat i on can be carried out di rect l y once t he ' me a n ' A(x) has been
cal cul at ed. The si mpl i ci t y of t hi s procedure is not ewor t hy.
Lomax and Heasl et *a have solved onl y t he case for whi ch t he opt i mum area di st ri but i on is a
cylinder, i.e., S(x)+ e g( x) = const ant . The present anal ysi s has generalised t hi s par t i cul ar
resul t to ar bi t r ar y opt i mum area di st ri but i ons, whi ch include t he const ant (or cylindrical) area
di st r i but i on as a special case. Fur t her mor e, as poi nt ed out by L omax and Heasl et la, t he present
anal ysi s can be appl i ed to mi ni mi si ng expressions .of t he form
i =1
whi ch consider t he wave drag at a discrete number of Ma c hnumbe r s only. Here A~s~ is t he
t ransferred area di st ri but i on appropri at e to t he Mach . number M~. The definitions of ' mean '
val ues follow exact l y as in t he above analysis.
The resul t for D {S -1-- A} obt ai ned above, namel y
q
{s + (S +
can be used to derive an expression for Cv, t he mean val ue of t he wave-drag coefficient of a
wi ng- body combi nat i on t hr oughout a Mach-number range. Fr om equat i on (10) i t follows t hat
CD = C/ ) w -t - Cv{s+a} - - CD/~} , where all t he wave-drag coefficients are based upon t he same
reference area.
12
3. Conf i gurat i ons Repr es ent ed by Di s c ont i nuous Source Di s t r i b u t i o n s . - - 3 . 1 . I sol at ed l / Vi n g s . ~
War d ~ has shown t hat t he wave drag associ at ed wi t h a di scont i nuous surface source di st ri but i on is
f f
D - - P c o s h - I d Q ( R I ) a s 2 d Q ( R ) . . . . ( 1 2 )
4 z ~ ~ I r , - r . I '
Where s0 is t he free-st ream densi t y
poQ(R) is t he source densi t y at t he poi nt R = (x, y, z)
r is t he posi t i on vect or in a pl ane x = const ant , i.e., r = (y, z)
d S dx denotes a general vol ume element,
t he co-ordinates (x, y , z) are chosen as in Fig. 1 and t he i nt egr at i on is over real values of the
i nt egrand. Fur t her mor e, since Q(R) is a di scont i nuous functiop, of x t he above i nt egral is a
double St i el t j es i nt egral whi ch must be wr i t t en in t hi s form. I t is possible to replace dQ(R) by
Q'(R) dx onl y when Q(R) is a cont i nuous funct i on of x.
I t follows di rect l y from equat i on (12) t hat t he wave drag of a wi ng is gi ven by
0 U 2 ( ( / I x l - x21
D- -
4 3 1 J ~ c s h - ~ \ B T r ( - - r o l ] d T ' ( R 1 ) _ , "" ""
(13)
where dx d r denot es an el ement of area of the wing, T(R) denot es t h e wi ng t hi ckness at t he
poi nt R, and t he i nt egr at i on is over those par t s of t he wi ng for whi ch t he i nt egr and is real. Thi s
resul t has been gi ven by War d% When t he wi ng' has rounded subsonic l eadi ng edges i t is modified
by t he addi t i on of a leading-edge ~ drag force. Such forces are exami ned in Section 5.
I t is Of i nt erest to not e t hat War d 1 has suggest ed t hat equat i ons (12) and (13) shoul d be
modified when t he t ot al source st r engt h is not zero, i.e., t he body is not closed. Such a modifica-
t i on is needed to allow for t he base drag term. Equat i on (13) gives t he ext er nal wave drag of
a wing, except for t he base drag, whet her t he t ot al source st r engt h is zero or not. Thi s resul t
can be proved most easi l y if t he sources are i ncl uded in t he equat i on of cont i nui t y and t he
aer odynami c forces are cal cul at ed by t he met hod of War d 21.
A par t i cul ar form of t he result, equat i on (13), has been gi ven by Li ght hi l P 2 for a pl ane wi ng
wi t h pol ygonal st reamwi se sections. The wave drag of a wi ng l yi ng in t he pl ane z = 0 (see Fig. 1)
is found to be
D _
where t he di scont i nui t i es i n T' occur along t he lines x = c,(y) and are of magni t ude 2/ Ua, , ( y)
( n = 1 , 2 , 3 , . . . ) .
Wa r d 8 has observed t ha t equat i on (13) is exact l y equi val ent to t he wave-drag resul t obt ai ned
. by i nt egr at i ng t he pressure di st ri but i on over t he wi ng plan-form. Fur t her mor e, in vi ew of t he
usefulness of t he sonic, area rule, t he resul t as M- - > 1 in equat i on (13) is of i nt erest . I n ma n y
cases t he presence of di scont i nui t i es in T ' implies t ha t t he l i neari sed t heor y yi el ds i nfi ni t e wave
drag at M = 1 but t here are cases for whi ch i t yi el ds finite wave drag at M = 1. Thi s poi nt
has been discussed by Heasl et , Lomax and Sprei t er 16 and by Lord, Ross and Emi nt on% I t
would appear t hat t he wave drag is finite at M = 1 when t he r at e of change of t he cross-sectional
area di st ri but i on is cont i nuous everywhere.
13
(71666) B
3.2. I s o l a t e d B o d i e s . - - T h e wave drag of a slender body whi ch can be represent ed by a discon-
t i nuous surface source di st ri but i on has been gi ven by Fr aenkel and Por t noy ~. The surface
slope ~ is di scont i nuous on a finite number of cont ours C~(i = 1, 2, . . . , ~) l yi ng in pl anes x = &
and separat ed by di st ances of magni t ude 0(1), t he body l engt h bei ng 0(1). Then, wi t hi n t he
usual order of approxi mat i on, t he wave drag of t hi s body is gi ven by
D 1 1
1 ( A S , ) ~ l o g 2
where ~,
( A , b , , ) ,
A~
A S I '
" "-]- ii=l ~ C i [(~,,)~=~,_ A w - - (A4~,)~ n,=~+] dr , . . . . . . (14)
i s~t hat port i on of t he slender body per t ur bat i on vel oci t y pot ent i al whi ch can
be det ermi ned from t he t wo-di mensi onal incompressible cross-flow
is t he increase in ~ at t he poi nt of di scont i nui t y x = &
denotes t he di scont i nui t y in t he surface slope at t he poi nt x = x~
denot es t he di scont i nui t y in S ' ( x ) at t he poi nt x = &
denot es t he finite par t of a double St i el t j es i nt egral t aken over all values of xl
and x2 from j ust upst r eam of t he body nose to j ust downst r eam of t he body
base
and tile i nt egr at i on in t he t hi r d t er m is, over t he cont our C~.
Fr aenkel and Por t noy 2~ ~ave shown t hat t he wave drag, gi ven by equat i on (14), remai ns
unchanged if t he st r eam di rect i on is reversed. Fur t her mor e, t he var i at i on of t he wave drag
wi t h Much number depends solely on t he di scont i nui t i es in S ' ( x ) and does not depend upon t he
cross-sectional shape. The case when t he body profile is di scont i nuous has been exami ned by
Fr aenkel and Por t noy ~-~but no appl i cat i ons of t he resul t to engine i nt akes have been given.
Fi nal l y, i t should be ment i oned t hat t he resul t obt ai ned by War d ~ differs from equat i on (14)
and is not correct for a slender body of a r bi t r a r y cross-sectional shape.
The effect of cross-sectional shape on t he wave drag of a slender body can be i nvest i gat ed using
equat i on (14). The presence of t he t hi r d t er m in equat i on (14) shows t h a t t he cross-sectional
shape at and wi t hi n t he nei ghbourhood of a di scont i nui t y in surface slope is i mpor t ant and
t hat t herefore t he wave drag of slender bodies does not depend sol el y on t he cross-sectional
area di st ri but i on. I t follows t ha t t he wave drag is i ndependent of t he cross-sectional shape
away from t he discontinuities. War d 1 has poi nt ed out t hat t he wave drag is a maxi mum for
t he body of revol ut i on. Bodies of elliptic cross-section have been exami ned by Fr aenkeP 5 who
has shown t ha t t he wave drag of t he equi val ent body of revol ut i on is great er by an amount
2g~ i : l
where e is t he eccent ri ci t y of t he el l i pt i c cross-section. I t was poi nt ed out by Fr aenkeP 5 t ha t if
t he cross-section is ' near l y ' circular, in t he sense t hat e --- 0(#/~), where t is t he t hi ckness rat i o
of t he body, t hen t he wave drag, whi ch is 0( # log t), will not change to t hi s order. Hence i t ma y
be assumed t hat most slender bodies of pr act i cal i mport ance have a wave drag i ndependent of
cross-sectional shape.
14
The form of equat i on (14) for a body of revol ut i on was gi ven first by Li ght hi l P 6 and, for a body
wi t h radi us R~ at t he poi nt x~, is found to be
1 1
1
- I - R ? ( d i ) l o g 2 . . . ( 1 5 )
~=1 B R ~ ' " . . . . . .
where t he i nt egrat i ons are Ri emann i nt egrat i dns i gnori ng t he poi nt s at whi ch S " ( x ) is undefi ned
and t he first t hree t erms are t he expanded form of t he finite par t of t he double St i el t j es i nt egr al
gi ven in equat i on (14). The l ast t er m in equat i on (15) gives t he resul t t hat t he wave drag of a
slender body wi t h di scont i nui t i es in S ' ( x ) is infinite at M = 1 according to t hi s t heory. Also, as
al r eady not ed, i t follows from equat i on (14) t hat equat i on (15) gives t he wave drag of a slender
body whi ch has ci rcul ar sections at and wi t hi n t he nei ghbourhood of t he di scont i nui t y cont ours C,:.
3.3. W i n g i B o d y C o m b i n a t i o n s . - - A n ext ensi on of t he area rule and t he t ransfer rule to con-
fi gurat i ons i ncorporat i ng bodies wi t h di scont i nui t i es in surface slope is given below. The cross-
sect i onal Shape of t he body is ar bi t r ai y.
Consider t he flow field on a circular cyl i ndri cal control surface at a large di st ance from t he
combi nat i on. Then t he wave drag is to be found by cal cul at i ng t h e rat e of flux of moment um
across t hi s surface. Heasl et , Lomax and Sprei t er 16, for example, have shown t hat t he per t ur bat i on
vel oci t y pot ent i al due to t he wi ng alone, at a gi ven angul ar posi t i on 0 on t he cont rol surface, is
t he same as t hat associated wi t h t he appropri at e wi ng . el ement al area di st ri but i on. Let t he
effect of any i nt erference vel oci t y pot ent i al on t he wave drag be neglected~. The flow field on
t he cont rol surface is t hen t aken as t he sum of t he per t ur bat i on vel oci t y pot ent i al s due to t h e
body alone and to an appropri at e wi ng el ement al area di st ri but i on so t hat , at a gi ven angul ar
posi t i on on t he cont rol surface, t he source di st ri but i ons represent i ng t h e body and t he appropri at e
el ement al area di st ri but i on are superimposed. The wave drag may be wr i t t en as
1
f l D{s s w ( x , o, M)}
D = - ~ . dO , . . . . . . . . . . (16)
where D { S + S w( x , O, M)} is i nt er pr et ed by exami ni ng t he flow field near t he axis, i.e.,
D { S + S w ( ) , O, M)} denot es t he wave drag of a ' combi ned' slender body. Thi s combi ned'
body is a st ream surface of t he di scont i l mous surface source di st r i but i on Q~, represent i ng t he
i sol at ed body, and t he cont i nuous source di st ri but i on defined by S w ' ( x , O, . M) . Since al most a l l
t he S w' ( X, O, M) are cont i nuous everywhere those di sconri nuous source di st ri but i ons associ at ed
wi t h t he di scont i nuous S w ' ( x , O, M ) are omi t t ed from t he i nt egral in equat i on (16).
Usi ng equat i on (12) t he wave drag of t he ' combi ned ' body can be expressed as
D{S + S.,(x, O, M ) } = D{ S} + D}Sw(x, O, M ) }
poU ( B I C I ' ) aSlag~(R.) Sw"(~~, o, M) d,x~.,
3 1 f c o s h - 1 ( 1 1 "
where D{ S } , t he i sol at ed body wave drag, is gi ven by equat i on (14) and D{ S w ( x , O, M)}, t he Wave
drag associ at ed wi t h S w( x , O, M ) , is gi ven by equat i on (1), wi t hi n t he usual order of approxi mat i on.
t This assumpt i on is not valid for slender wing-body combinations when t he source distribution representing the
body (and/or t he wing cross-sectional area distribution) is discontinuous wi t hi n t he region of t he wing because
equat i on (14), derived by Fraenkel and Port noy 24, cannot be obtained by neglecting the interference velocity potential.
The assum ?tion is valid for any other t ype of slender wing-body combination.
15
(71666) B 2
Following War d' s anal ysi s 1 for slender bodies of revol ut i on t he inverse cosh in t hi s result is
expanded in t erms of l ogari t hms. The unknown body source densi t y QB(R) is el i mi nat ed
by not i ng t hat t he t ot al source st r engt h in a pl ane x = const ant is equal t o US' ( x) , where S(x)
~s t he cross-sectional area di st ri but i on of t he body 1. Since S~v'(x, O, M) vani shes at its end
points, it can be shown qui t e si mpl y t hat t he wave drag of t he ' combi ned' body becomes, wi t hi n
t he usual order of approxi mat i on,
D { S -~- S w ( x , O, M) } = D { S } 4- D{Sw(x, O, M) }
p o V ~_ f c
2~ log Ix, - - x2 ] dS' (xl ) Sw"(x2, O, M) dx~ . . . . . (17)
Equat i ons (16) and (17) provi de a general i sat i on of t he area rule, equat i ons (2) and (3), to com-
bi nat i ons i ncorporat i ng bodies wi t h di scont i nui t i es in surface slope.
The t ransferred wing area di st ri but i on A (x) is equal to t he mean wing el ement al area distribu-
tion, i.e.,
1 (==
A (x) = ~J o Sw(x, o, M) do.
Thus, using equat i on (17), equat i on (16) becomes
= D { S } + Dw p o V f f l o g a S ' ( x , ) ( 1 8 ) D
2 ~ I ~ " "
where D{S} is given by equat i on (14) and Dw by equat i on (13), for example. Thi s result is t he
ext ended form of t he t ransfer rule, equat i ons (8) and (10).
Equat i on (18) implies t hat t he i nt erference wave drag Dw~ is give n by
whi ch is t he general i sat i on of equat i on (9). I t shoul d be not ed t ha t t he i nt erference wave drag
Dw~ ".is i ndependent of t he cross-sectional shape of t he body. The result for t he i nt erference wave
drag can be derived di rect l y from equat i on (12) wi t h Q(R) ---- QB(R) + Qw(R) ; Qw(R) = UT' ( R) .
I t is necessary to use War d' s i nt egral equat i on ~ for t he t ransferred area di st ri but i on A( x) .
I n Section 2.4.1 i t was shown t hat t he wi ng-body combi nat i on can be replaced by an equi val ent
combi nat i on made up of a sui t abl y chosen gross wing W0 and a modified body whi ch is denot ed
by B - - A A, where A A denotes t he ( t r ansf er r ed) ar ea di st ri but i on o f t he por t i on of t he gross
wing bl anket ed by t he body. I t is assumed t hat S' ( x) is cont i nuous where AA( x) is defined and
t ha t AA' ( x ) is cont i nuous everywhere as otherwise t he source di st ri but i on represent i ng t he
modified body is not equal (in t he wave-drag sense) to t hat represent i ng t he body less bl anket ed
wing, i.e., Q(B-~A)-/-~QB-~w. The generalised form of equat i on (11) follows from equat i on (18)
and can be shown to be
D ": Dwo + D{S + A} - - D{A + A A } , . . . . . . . . (19)
where D{S - / A } denotes t he wave drag of a ' combi ned' body, i.e.,
D{S + A} = D{S} + D{A} Pog~ ~ f " 2u J~ log Ix, -- x~[ dS'(x~) Sw (x~, O, M) dx~ .
4. Numeri cal Eval uat i on of Wave Dr a g . - - I n t he previous sections at t ent i on has been concen-
t r at ed upon giving t he t heoret i cal met hods r at her t hen t he pract i cal appl i cat i on of these met hods.
The present section is i nt ended to give a brief description of t he numeri cal met hods t hat have
been developed for eval uat i ng wave drag. I t will be seen t ha t these met hods have been devel oped
for i sol at ed wings or i sol at ed bodies but t hei r appl i cat i on t o wi ng-body combi nat i ons is not
difficult because t he results obt ai ned for t he wave drag of wi ng-body combi nat i ons require onl y
t he eval uat i on of wing and ' b o d y ' wave drags.
16
4 . 1 . Isolated Wi n g s . - - T h e r e has been a large number of i nvest i gat i ons in t he field of supersonic
wi ng t heor y and no at t empt will be made to summari se t hem here. However, i t is sufficient to
notice t hat t he resul t s for wave drag have been found by i nt egr at i ng pr essur e di st ri but i ons.
Also, t he wing pl an-forms are st rai ght -edged and t he Wing sections simple shapes. A s ummar y
of t he resul t s obt ai ned has been gi ven by Bi s hop and Cane ~7. Some resul t s for sonic speed have
been obt ai ned by Lord, Ross and Eminton. ~3 using t he sonic area rule.
Fr om t he resul t s gi ven in Section 3.1 i t is suggest ed t hat equat i on (13) be used to cal cul at e
t he wave drag provi ded t hat t he pressure di st ri but i on is not required. Ot herwi se it woul d
pr obabl y be bet t er to det ermi ne t he pressure di st ri but i on first and t hen t he wave drag. If t he
pressure di st ri but i on is requi red it can be found usi ng t he met hod of Thomson 28.
Nevert hel ess, it is appar ent t ha t t he area rul e (equations (2) and (3)) offers. an al t er nat i ve
met hod for cal cul at i ng t he wave drag of an i sol at ed wing. However, a difficulty arises if any
st r ai ght lines, along whi ch t he wi ng pl anar source di st ri but i on is di scont i nuous, are i ncl i ned so t ha t
t he component of free-stream vel oci t y nor mal to t he line is supersonic (for example, along sharp
edges t he source di st ri but i on is discontinuous). But , since al most all t he wing el ement al area
di st ri but i ons have source represent at i ons cont i nuous everywhere, t he wave drag ma y be cal cul at ed
using t hese el ement al area di st ri but i ons onl y and it ma y not be necessary to use t he results of
Section 3.2 for bodies wi t h di scont i nuous source represent at i ons. Lord ~9 has poi nt ed out t ha t
a numeri cal appl i cat i on of t hi s procedure to wings wi t h supersonic edges is not sat i sfact ory.
The area-rule met hod is par t i cul ar l y useful for wings wi t h curved pl an-forms a nd has been
'used by Lord and Bennet t . Det ai l s may be found in Refs. 30 and 31.
4.2. Isolated Bodi e s . - - - Re c e nt l y a number of aut hors have i nvest i gat ed numeri cal met hods for
eval uat i ng t he double i nt egral of equat i on (1), i. e. , met hods have been developed for cal cul at i ng
t he wave drag of slender bodies wi t hout di scont i nui t i es in surface slope. The more general case
when t he source represent at i on is not cont i nuous will be exami ned l at er in t he present section.
There are two mai n met hods for eval uat i ng D{S}, equat i on (1), namel y, t he Fourier-series
met hod and t he opt i mum met hod of Emi nt on 31. Considering a body of uni t l engt h, t he
Fourier-series met hod relies upon expressing S ' ( x ) as a sine series in t he new vari abl e
= cos -1 (1 -- 2x). The di sadvant age of t hi s met hod lies in t he det er mi nat i on of S ' ( x ) , whi ch
requires a numeric~il di fferent i at i on of t he cross-sectional area di st ri but i on S(x). Fur t her mor e,
it is not clear how ma ny t erms in t he Fouri er series are requi red to give reasonabl e accuracy in
t he answer for t he wave drag.
The difficulties of t he Fourier-series met hod can be overcome most sat i sfact ori l y by using t he
met hod of Emi nt on 4, 3~ for eval uat i ng t he double i nt egral of equat i on (1). Thi s met hod depends
upon some resul t s in t he t heor y of opt i mum area di st ri but i ons and details can be found in t he
papers cited 4,31. The gi ven area di st ri but i on is repl aced by t he opt i mum area di st ri but i on wi t h
mi ni mum wave drag for a large number of fixed areas at equal l y spaced poi nt s al ong t he body.
Thus t he eval uat i on of t he wa ve drag, gi ven by t he double i nt egr al of equat i on (1), requires
onl y a knowl edge of t he areas at a large number of points, usual l y ni net een, al ong t he body.
The deri vat i ve S' ( X ) is not requi red for t hi s met hod.
Slender bodies wi t h di scont i nuous source represent at i ons cannot be t r eat ed in general, but
can be t r eat ed provi ded t hat t he cross-section is circular at and wi t hi n t he nei ghbourhood of a
di scont i nui t y in surface slope, i . e. , t he wave drag is gi ven by equat i on (15) and depends upon t he
cross-sectional area di st ri but i on only.
A numeri cal procedure was developed first by Di ckson and Jones a2, who ut i l i sed a t ransforma-
t i on of equat i on (15) wher eby Fouri er series could be used di rect l y to eval uat e t he i nt egral s
i nvol ved. Thi s approach would be expect ed to have a count er par t in t he t heor y of opt i mum
area di st ri but i ons if such opt i ma could be .found for t hi s more general case. Emi nt on is
i nvest i gat i ng an ext ensi on of t he opt i mum met hod to t he present problem, t hus det ermi ni ng t he
opt i mum area di st ri but i on wi t h mi ni mum wave drag (equat i on (15)) for prescri bed values of
17
AS,' at the points x -. x~where the cross-sectional area is S(x~) (i -- 1, 2, . . . n). It appears
t hat a simple and illuminating extension may be possible and t hat the opt i mum met hod will
be ext ended to the evaluation of equation (15). However, since the met hod of Dickson and
Jones 3~ seems to be the only published met hod and since their report 3~ may not be generally
available, their met hod will be summarised here.
Consider a body of unit length and introduce the function P(x) defined by
/ . . !
P ' ( , ) = S 6 ) - ~ ~ S , + x ~ &' , x; ~< x ~ . %1 .
/ =I / =1
The function P(x) is chosen by Dickson and Jones 3~ because P'(x) is continuous everywhere.
Now introduce the Fourier-series variable $ = cos -~ (1 -- 2x) and put
co
P'(x) = ~ A,, sin m~,
so t hat equation (15) becomes
D q 2 _ ~ . \ ~ 1 [ " z l & , ) '_ "
1 " 2
BR,
-1- ~ mA,. 2 -/ AS, ~ [ _ ( 2 m) 2 1 ]
i = l ~ , t = l
-1- A S~' A,. cos m~
i = 1 11 1
+ - 1 - l o g ( 1 - + . , x ; }
Yg , = 1 j = l
+ I x , - ' " " " "
where x~ = 1(1 -- cos %) and =R, 2 = S(x,). This result is in a somewhat different form from
t hat given by Dickson and Jones a'.
4.3. Wing-Body Combinations.--In Sections 2 and 3 it has been shown t hat t he calculation of
the wave drag of a wing-body combination can be reduced either to .a wing problem together
with two ' body' problems by using the transfer rule or to a number of ' body' problems by
using the area rule. Hence the numerical met hods for wings and bodies given in Sections 4.1
and 4.2 enable the wave drag of a wing-body combination to be found. Procedures for calculating
t he transferred area distribution of the wing are described in the Appendix.
5. Some Miscellaneous Remarks.--5.1. Configurations with Round-Nose Wing Sections.--Ward 10
has pointed out t hat the results of linearised t heory are incorrect if the Wing of the Wing-body
combination has rounded edges where T'(R) is infinite. Nevertheless, the linearised-tlheory
result for wings with rounded subsonic leading edges can be obtained by allowing for the effect
of the leading-edge singularity in the way suggested by Jones3L The result obtained is t hat to
t he existing result (equation (6), for example), there must be added a leading-edge drag F per
unit length given by
F ~r cosaA
~- - [1 - ( M cos AT] 1 " '
where q is t he kinetic pressure, A is the local sweepback angle of the leading edge, and r is the
local nose radi us. .
18
I t should be not ed t hat this result is rest ri ct ed to wings for whi ch t here are no abr upt changes
in t he curvat ure of t he leading edge. Moreover, t he result shows t hat t he l i neari sed-t heory wave
drag of a wing wi t h a rounded leading edge is infinite if t he leading edge is sonic (i.e., M cos A -- 1).
Thi s infinite result for a wing wi t h a sonic leading edge is obvi ousl y i ncorrect and emphasises t he
fact t hat t he linearised t heor y is not sat i sfact ory for wings wi t h rounded edges. Thus, i n t he
absence of rion-linear t heory, experi ment al results are requi red for det ermi ni ng t he wave drag of
such wings. However, since t he t ransfer rule separates t he wing wave drag from t he t ot al wave
drag of a wi ng-body combi nat i on, it can be used t o est i mat e t he i nt erference Wave drag.
5.2. Systems of Wings and Bodies.--The discussion s o far has been rest ri ct ed to wi ng-body
combi nat i ons, al t hough it could also have been given for wi ng- syst em- body- syst em combi nat i ons.
The wing system includes wing-like elements of t he configuration as well as any subsi di ary bodies
si t uat ed on these elements while t he body system is essent i al l y axi al and may include fins and
tailplanes. Since t he area rule, moment of area rule and t ransfer rule are all equal l y appl i cabl e
t o systems of wings and bodies as t o wi ng-body combi nat i ons, no essentially new probl em is
i nt roduced.
5.2.1. Arrangements Of bodies.--Arrangements of bodies have been exami ned by Fr i edman
and Cohen 3~ and Rennemann 35. I n bot h cases t he anal ysi s does not use any o f t he met hods
given in .Section 2 but such an analysis could be carried out using either t he area rule or t he
t ransfer rule. Rennemann 35 has exami ned t he shape of a body of revol ut i on in t he pressure fi el d
of a larger body and has shown t ha t t he opt i mum shape for mi ni mum wave drag can be cal cul at ed
sufficiently ~ exact l y by ignoring t he larger body, i.e., t he Sears-Haack body based upon t he gi ven
l engt h a n d vol ume is t he best body shape to use for t he smaller body.
Fr i edman and Cohen 3~ have used t hi s result to exami ne t he posi t i oni ng of auxi l i ary bodies
alongside a mai n body. For t wo-body or t hr ee- body syst ems t he most favourabl e l ocat i on is
where t he maxi mum cross-section of t he auxi l i ary body is j ust upst r eam of t he Mach cone from
t he t ai l of t he mai n body while tile least favourabl e l ocat i on is wi t hi n t he Mach di amond enclosing
t he mai n body. These conclusions are not absol ut el y definite and so a f ur t her anal ysi s of t hi s
probl em ma y. be needed.
A suggested appl i cat i on of an analysis of t he above t ype is as follows : Fi rst l y, consi der t he
mai n and auxi l i ary bodies of t he gi ven syst em and posi t i on these for mi ni mum wave drag at t he
Mach number under consideration. Secondly, bs) assuming t hat t he wave drag of these bodies is
consi derabl y great er t han t he wave drag of a wing system, i t follows t ha t t he next step is to
i nt roduce a wing syst em cont ai ni ng these bodies and shape all t he bodies according t o tile nat ur e
of tile wing system. This l at t er procedure can be carried out using t he met hods discussed i n
Section 2. Details are given in Section 5.2.2 below.
5.2.2. Combinatio~cs with low wave drag.---Consider, for example, t he probl em of shapi ng and
posi t i oni ng a number of subsi di ary bodies such as ext ernal stores on t he wing of a confi gurat i on
so t ha t t he overall wave drag is a mi ni mum. To a first appr oxi mat i on t he most i mpor t ant effect
of a change in t he wing syst em wi l l be in t he wave drag of t he auxi l i ary bodies alone and t he
i nt erference wave drag between t hem and t he wing (this result assumes t ha t t he su.bsidiary-
bodies are not close to t he body and t o each ot her ) . By considering separat el y each subsi di ary
body and t hewi ng, t he i nt erference probl em can be t r eat ed by using equat i on (10) and so, for a
given wing, it follows t ha t t he vari abl e wave drag t er m will be. D{A1 + 5 v} where Al(x) is t he
t ransferred area di st ri but i on of t he wing system, less t he appropri at e subsi di ary body, on to t he
auxi l i ar y body axis a n d N(x) is the cross-sectional area di st ri but i on of t he subsidiary, body.
Thus t i l e opt i mi sat i on procedure for each subsi di ary body has been reduced t o mml ml sl ng
D{A1-]-kv} subj ect t o certairi prescribed conditions. Fi nal l y t he opt i mi sat i on procedure is
compl et ed by exami ni ng t he mai n body, whi ch is shaped using t he t ransfer-rul e procedure given
in Section 2.4.2.
19
In many appl i cat i ons it may be necessary to repeat t hi s opt i mi sat i on procedure because any
var i at i on in Al(x), due to subsequent modifications of ot her subsi di ary bodies, is not considered
when mi ni mi si ng D{AI q- ~Y}. Thus repet i t i on may be essential i n t hose cases where t he applica-
t i on of a s ymmet r y condi t i on implies t hat t here will be correspondi ng changes in Al(x) for any
al t erat i on to ~(x). However, no repet i t i on at all will be necessary if each subsi di ary body does
not influence, or is not influenced by, any ot her subsi di ary body.
I t will be not ed t hat t hi s opt i mi sat i on procedure exami nes fi rst l y t he subsi di ary bodies and
t hen t he mai n body. But , since t he subsi di ary bodies may requi re a repet i t i on of t he procedure
i t is recommended t hat t he whole wing syst em be considered before any shapi ng of t he mai n
body is i nvest i gat ed.
The most i mpor t ant exampl e of wi ng- syst em- body- i nt er f er ence occurs when t here are two
subsi di ary bodies only. In t hi s case t he assumpt i on t hat t he bodies are not ' close ' to each ot her
is unnecessary since t he wave drag can be found by br eaki ng clown t he wi ng syst em into a number
of simple wi ng- body problems. I t can be shown t hat t he i nt erference wave drag for t he i sol at ed
wi ng syst em consists of t hree t er ms: t he i nt er f er ence wave drag bet ween each subsi di ary body
and t he wi ng (two terms) t oget her wi t h t hat bet ween t he two subsi di ary bodies. The l ast t er m
is eval uat ed by selecting one subsi di ary body as a ' mai n ' body and t hen appl yi ng t he t ransfer
rule, equat i on (10). I t is t hi s body- body i nt erference t er m whi ch would be neglected, somet i mes
wi t hout j ust i fi cat i on, b y t he simplified opt i mi sat i on procedure present ed for a general wing
syst em. However, al t hough t he inclusion of t hi s t er m does make any opt i mi sat i on procedure more
complex, i t does enable, also, est i mat es of t he wave drag of t he complete configuration to be made.
5.3. Problems Associated with I~#akes and Exhaus t s . - - The presence of engine i nt akes and
exhaust s wi t hi n an ai rcraft has not been considered in t he previ ous discussions whi ch are applic-
able onl y to ideal confi gurat i ons wi t h no i nt akes and exhaust s. I t will be shown t hat t he use of
two si mpl i fyi ng assumpt i ons enables t he t ot al ext er nal .wave drag to be est i mat ed. As well as
t he wave drag arising from t he pressure forces act i ng upon t he ext er nal surface of t he ai rcraft
i t is necessary now to consider t he ext er nal wave drag due to t he pressure forces act i ng upon
t he pr e- ent r y and post -exi t st ream tubes. The pr e- ent r y st ream t ubes ext end from i nf i ni t y
upst r eam to each engine i nt ake and separat e t he air flowing i nt o t he i nt akes from t hat flowing
ar ound t hem while t he post -exi t st r eam t ubes ext end from each engine exhaust to i nf i ni t y down-
s t r e a m and separat e t he hi gh-speed gases flowing out of t he exhaust s from t he air whi ch passes
over t he ext er nal surfaces of t he configuration, i.e., t he pr e- ent r y and post - exi t st ream t ubes
separat e t he ext ernal flow from t he i nt er nal engine flow:
Thus, in principle, an effective ' sol i d' combi nat i on, i nchl di ng all t he pr e- ent r y and post -exi t
st r eam t ubes can be found. The met hods of wave drag est i mat i on present ed in Sections 2 and 3
are not necessari l y st r i ct l y appl i cabl e to t hi s ' s o l i d ' confi gurat i on because large per t ur bat i on
velocities woul d be associ at ed wKh any regions of large ' sur f ace' slopes on ei t her t he i nt ake
cowls or t he pr e- ent r y and post -exi t st r eam robes. Nevertheless, par t of t he following discussion
is based upon t he linearised t heor y and an appr oxi mat e met hod for t r eat i ng engine i nt akes and
exhaust s is described.
The usual met hod of cal cul at i ng t he t ot al ext ernal wave drag requi res i t s division i nt o t hree
component s: t he ext ernal wave drag proper, whi ch is associated wi t h t he act ual ext ernal surfaces
of t he ai r cr af t ; t he pr e- ent r y wave drag, whi ch is associated wi t h t he pr e- ent r y st r eam t ubes;
t he post -exi t wave drag, whi ch is associated wi t h t he post -exi t st r eam tubes. The ext er nal wave
drag proper cannot be eval uat ed di rect l y and t herefore i t is conveni ent to use t he l i neari sed
t heor y met hods of Sections 2 and 3 in or der to cal cul at e i t for an i nt er nal flow in whi ch all t he
pr e- ent r y and post -exi t st r eam t ubes are of const ant cross-sectional area. But , since t he pr e- ent r y
and post -exi t Stream t ubes are not of const ant cross-sectional area, i n general, t here are a number
of t erms to be added before t he act ual ext ernal wave drag proper is found. The l argest of t hese
addi t i onal t erms is l i kel y t o be t hat associated wi t h t he flow over t he i nt ake cowl and can be
est i mat ed using exi st i ng t echni ques for i sol at ed i nt akes. The r emai ni ng addi t i onal t erms are
negl ect ed since t hey are assumed to be small.
20
I t is suggest ed t ha t cal cul at i ons of ext er nal wave drag be carri ed out usi ng t hi s assumpt i on
t oget her wi t h t he addi t i onal assumpt i on t ha t t he we - e nt r y and post -exi t wave drags can be
est i mat ed by exami ni ng each i nt ake- engi ne- exhaust syst em i n isolation. There is, however, one
case for whi ch t hese assumpt i ons ma y not be required. For slender bodies, t he l i neari sed t heor y
t echni ques of Fr aenkel and Por t noy "4 and War d ~1 woul d enable t he i nt ake probl em to be solved
exact l y, provi ded t ha t t he i nt ake has shar p edges and t here is no spillage ar ound t hese edges.
The t r eat ment of an exhaust upst r eam of t he t ai l is difficult because of t he hi gh-speed exhaust
gases. For ar bi t r ar y confi gurat i ons t he assumpt i ons are necessary and, since t hey may give
rise to incorrect wave-drag est i mat es, an i nvest i gat i on of t he problems i nvol ved in t r eat i ng
engine i nt akes and exhaust s ma y be desirable.
5.4. Non- l i mar Effects.--Since t he present report deals wi t h t he linearised t heory, non-l i near
effects are not exami ned. Confining at t ent i on to ~ t he t ransfer rule, i t is seen t hat onl y t he i nt er-
ference wave drag need be considered, ~ince t he l i neari sed-t heory resul t s for t he i sol at ed wi ng
and body wave dr ags can be modified by usi ng correlations bet ween exper i ment and linearised
t heor y as a guide to t he magni t ude of any non-l i near effects. Fur t her mor e, h' om t he form of
equat i on (10) i t appears t hat sat i sf act or y modifications may be able to be carried out by
modi f yi ng onl y t he t ransferred area di st ri but i on A (x).
However, at t he present time, t he est i mat i on of non- l i near effects must be based upon known
effects on i sol at ed wings and i sol at ed bodies. The above suggestion is pur el y t ent at i ve and is
i nt ended to emphasi se t he need for some satisfactorsr met hod of est i mat i ng non-l i near effects on
t he wave drag of wi ng-body combi nat i ons. -
I n t he t ransoni c range, non-l i near effects on wave dr ag have been exami ned by Owat i t sch and
Keune 3~ for l ow-aspect -rat i o wings. I n t he supersonic speed range t he effects of st rong shock
and expansi on waves have not been considered except in cert ai n par t i cul ar cases such as two-
di mensi onal wings.
5.5. Boundary-Layer Correc~ions.--An i mpor t ant aspect of wave-drag cal cul at i ons for a specific
confi gurat i on is t he det er mi nat i on of corrections t hat allow for t he boundar y- l ayer di spl acement
t hi ckness. Consider a slender body whose wave drag is gi ven by t he double i nt egr al of equat i on
(1). I n equat i on (1), S"(xl) ma y be regarded as a surface slope t er m and S"(x2) as a local pressure
term. War r en has poi nt ed out t hat t he nat ur e of t he boundar y- l ayer correction to t he wave
drag D{S} requi res careful consideration because onl y t he local pressure t er m is altered. Let
~( x) be t he i ncr ement in S(x) due to t he boundar y- l ayer di spl acement t hi ckness. Then t he wave
drag is gi ven by equat i on (1) wi t h S"(x d unal t er ed and S"(x2) repl aced by S"(x~) + ~"(x2). But ,
since t he boundar y l ayer correction is small, i t can be shown t hat t he wave drag is gi ven approxi -
mat el y by D{S + ~}. Si mi l ar consi derat i ons appl y to equat i on (13) whi ch gives t he wave drag
of a wing. Thus, in order to obt ai n an effective solid body, hal f t he boundar y- l ayer di spl acement
t hi ckness is added everywhere to t he profile of a wi ng-body combi nat i on.
6. Coml usi on. - - The wave drag of a combi nat i on of a t hi n wing and a smoot h sl ender body
can be est i mat ed subj ect to t he usual l i mi t at i ons of l i neari sed t heor y and, i t shoul d be noted,
subj ect to t he val i di t y of t he assumpt i on t ha t t he i nt erference vel oci t y pot ent i al is negligible.
For smoot h combi nat i ons, whi ch are aer odynami cal l y slender, t he wave drag is gi ven by t he
sonic area rule, whi ch st at es t ha t t he wave drag depends onl y on t he axi al di st ri but i on of cross-
sect i onal area. For smoot h combi nat i ons whi ch are ' not-so-slender ', i t is necessary to use t he
moment of area rule, whi ch st at es t hat t he wave drag depends upon t he axi al di st r i but i on of tile
second moment of area as well as upon t he cross-sectional area di st ri but i on. For ar bi t r ar y
combi nat i ons t he wave drag is gi ven by. t he supersonic area rule, whi ch expresses t he wave drag
as t he mean of t he wave drags associ at ed wi t h t he so-called el ement al area di st ri but i ons. If t he
wave drag of tile i sol at ed wi ng of an ar bi t r ar y confi gurat i on is known, i t is conveni ent to use
t he t r ansf er rule, whi ch expresses t he sum of t he body wave drag and t he i nt erference wave
drag in a simple form.
21
The met hods of wave-drag est i mat i on can be i nt er pr et ed t o yield pract i cal met hods of designing
for low wave drag under prescribed conditions. For a given wing, t he general probl em of designing
a body so. t hat t he combi nat i on has low wave drag t hr oughout a range of Mach number can be
solved. The met hod of designing t he body is simple at all Mach numbers, but t he possible
benefits of an opt i mum design decrease qui t e r api dl y wi t h increase of Mach number. The achieve-
ment of low wave drag is faci l i t at ed by allowing t he wing to be altered, an example of whi ch is
t he design for low wave drag according to t he moment of area rule.
Previ ous discussions of t he met hods of est i mat i on of t he wave drag of wi ng-body combi nat i ons
have been confined to combi nat i ons wi t h smoot h slender bodies. Here an ext ensi on of these
met hods is gi ven for t he cases when t he combi nat i ons i ncorporat e bodies wi t h di scont i nui t i es in
surface slope.
All t he t heorei i cal met hods pr esent ed her e require sui t abl e numeri cal t echni ques for t hei r
pract i cal appl i cat i on and i t is found t ha t t h e existing t echni ques are ver y sat i sfact ory.
No numeri cal results or part i cul ar appl i cat i ons are described but since most of t he met hods
discussed in t hi s report are somewhat idealised some miscellaneous remarks are made to help in
t hei r appl i cat i on t o pract i cal aircraft. I n part i cul ar, it is poi nt ed out t hat t he present met hod
of allowing for t he air whi ch passes t hr ough t he engines of an ai rcraft is far from sat i sfact ory
and a more adequat e procedure is essential.
aO, (g2
A
AMi
Adx)
A(x)
A
J~
A,,( o, M)
B
C
C~
CD
D
NOTATI ON
Coefficients of B , B 2 in series expansi on of D
U/2 mul t i pl i ed by di scont i nui t y in T' al ong t he line x = c;,(y)
Aspect rat i o
Transferred area di st ri but i on appropri at e t o t he Mach number M~
Transferred area di st ri but i on (equation (7))
Transferred area di st ri but i on of wing syst em on to axis of subsi di ary body
Cont ri but i on of streamwise st ri p of wing to A (x)
(Transferred) area di st ri but i on of port i on of gross wing syst em blan' keted by
t he body
Fouri er coefficient in expansi on of S' (x) using t he vari abl e
Fouri er coefficient in expansi on of S' (x, O, M) using t he vari abl e ~
{M ~ -- 1}~/=
Fouri er coefficient in expansi on of M, , ' (x) using t he vari abl e ~0
Wi ng-root c h o r d
Defines line along whi ch discontinuities in T' occur
Generalised moment of area (see Section 2.3.1)
Cont our l yi ng in t he pl ane x = x, and al ong whi ch t he body surface slope
~7is discontinuous
Wave-drag coefficient D/qS, S bei ng an appropri at e area
Wave drag
22
D{S}
D(O, M)
DWB
DB--AA
g_
/ ( M)
F
[,+
k
K
l
z,
M
m,~(x)
N
p(x)
q
Q
' QAA
Q(a)
f
r , 0
R
R+, Rj
R t
R
R(~,o)
s,'(x)
s(x, o, M)
S ( x )
S
s(x; o, M)
.<,~(x)
N O T A T I O N - - c o n t i n u e d
Wave drag of ' b o d y ' with the cross-sectional area distribution S ( x ) and
the prescribed shape
Wave drag associated with an elemental area distribution
interference wave drag
Wave drag of body less portion of gross wing blanketed by the body
Eccentricity of elliptic cross-section
Mach-number function used in Section 2.4.2
Leading-edge drag force per unit length due to curvature of wing section
Int egral defined in Appendix
Slenderness paramet er
-(2N + 1)
Length of wing-body combination
Length of pt h moment of area distribution
Mach number
nt h moment of area distribution
Number of discontinuities ill S ' ( x ) (also asubscri pt )
Number of elements into which streamwise strip of wing is divided
Modified form of S ( x ) (see Section 4.2)
Kinetic pressure .vp0U ~
Source distribution
Source distribution representing A A (x)
1/po multiplied by the source density at the point R
Radial co-ordinate (Fig. 1) or leading-edge nose radius
Vector (y, z)
Unit vector (sin O, cos 0)
Distance of body from axis
Radius of body of revolution at the points of discontinuity x ---- xi , x~
Typical body radius
Position vector (x, y, z) or (x, r)
Position vector (x + B r0 r, r)
Defines cross-section of configuration at the station x or/ axi s
Increment in S ( x ) due to boundary-l ayer displacement thickness
cosec l+ S ( x , O, M) +
Cross-sectional area distribution of body (and also of wing-body com-bination)
Const ant value of S,,
Constant element of 7'(x) in the range X,, ~< x ~< X~c_,,
Elemental area distribution
Generalised moment of area (see Section 2.3.1)
23
d S
AS i '
t
T( x , y), T(R)
U
V
d V
W
Wo
A W -
( x , y , z)
X, i , X j
X ~
0 =
A
#
p
Po
d~
"c o
d ,
9 =
d ~
' i , j , n , m, p, r ,
A, B, W, WO, WN
1 , 2
dashes
bar
N O T A T I O N - - c o n t i n u e d
El ement of area used in sense t hat dS dx is an el ement of vol ume
Increase in S ' ( x ) at a poi nt of di scont i nui t y x = x~
Body t hi ckness rat i o
Wi ng t hi ckness at t he poi nt (x, y)~ R
Free-st ream vel oci t y
Wi ng vol ume
pt h moment vol ume of t he configuration (see Section 2.3.2)
El ement of vol ume
Wi ng
Sui t abl y chosen gross wing
Wo -- W
Rect angul ar Cart esi an co-ordinates (see Fig. 1)
Poi nt s at whi ch body surface slope is di scont i nuous
One of 2N division poi nt s for t he area di st ri but i on 7~(x), X1 <~ x K X~N
Body surface slope rel at i ve to free-stream direction
Increase in ;~ at t he poi nt of di scont i nui t y x = x~
t an -1 (z/y) (see Fig. 1)
Angle of leading-edge sweepback
Mach angle (----- sin -1 (l / M))
d '
t an -1 (B cos 0)
Free-st ream densi t y
El ement of area of t he wing surface
Wi ng-root t hi ckness rat i o
El ement of arc l engt h on t he cont our C~
c o s - 1 ( 1 - 2 x )
cos -1 (1 -- 2x~)
That port i on of t he sl ender-body per t ur bat i on vel oci t y pot ent i al whi ch can
be det er mi ned from t he t wo-di mensi onal i ncompr essi bl e cross-flow
Increase i n 6~ at t he poi nt of di scont i nui t y x = x~
Cross-sectional area di st ri but i on of a body, or a st reamwi se wing strip, at
a di st ance R from axis
El ement used in t he sense t hat d ~ dx is an el ement of wing vol ume
Subscri pt s denot i ng different member s of a family, e.g., M~(x) is t he i t h
moment of area
Subscri pt s de not i ng t ransferred area di st ri but i on, body, wing,, gross wing,
net wi ng respect i vel y
Subscri pt s usual l y appl i ed to vari abl es of i nt egr at i on
Denot e deri vat i ves wi t h respect to x, e.g., S ' ( x ) = ~S/ ~x
Denot es a ' me a n ' value, e.g., _/i(x)
24
No . Aut hor
1 R. T . Whi t comb . . . . . .
2 W. T. Lor d . . . . . .
3 W. T. Lor d . . . . . .
4 E. Emi nt on and W. T. Lor d
5 R. T. Jones . . . . . .
6 W. T . Lor d . . . . . .
7 C. H. E. - Wa r r e n . . . . . .
8 G. N. War d . . . . . .
9 G. N. War d . . . . . .
10 G. N. War d . . . . . .
11 B. S . Bal dwi n and R. R. Di ckey
12 W. D. Hayes . . . . . .
13 H. Lomax and M. A. Heasl et . .
14 L. M. Sheppar d . . . . . .
15 R. Legendre . . . . . .
16 M. A. Heasl et , H. Lomax and
J. R. Sprei t er
17 E . W. Gr aham, 13. J. Beane and
R. M. Li cher
18 R. C. Lock . . . . . .
19 R. T . Jones . . . . . .
20 M. C. Adams and W. R. Sears . .
21 G. N. War d . . . . . .
22 M. J . Li ght hi l l . . . . . .
23 W. T. Lord, A. J. Ross and
E. Emi nt on
RE F E RE NCE S '
Title, etc.
A s t udy of t he zero-lift drag-rise charact eri st i cs of wi ng- body combi nat i ons
near t he speed of sound. N.A.C.A. Research Memo. L52H08 ( TI B 3279).
Sept ember , 1952.
Not es on t he dr ag of wi ng- body combi nat i ons at t ransoni c and supersonic
speeds. R. A. E. Tech. Not e Aero. 2229. A.R.C. 16,784. J anuar y, 1953.
Exampl es of wi ng- body combi nat i ons of ver y smal l t heoret i cal dr ag rise
at t ransoni c speeds. R. A. E. Tech. Not e Aero. 2266. A.R.C. 16,695.
December , 1953.
Not e on t he numeri cal eval uat i on of t he wave dr ag of smoot h slender bodi es
using opt i mum area di st ri but i ons for mi ni mum wave drag. J . R. Ae. Soc.
Techni cal Not e. Vol. 60. No. 1. J anuar y, 1956.
Theor y of wi ng- body dr ag at supersonic speeds. N.A.C.A. Research Memo. "
A53H18a ( TI B 3890). Sept ember , 1953.
Unpubl i shed R. A. E. Not e. 1954.
Unpubl i shed R. A. E. Tech. Not e. 1954.
On t he cal cul at i on of dr ag at supersonic speeds. I - - Th e dr ag of source
and vor t ex- doubl et di st ri but i ons. (Unpublished.) 1954.
On t he cal cul at i on of dr ag at supersonic speeds. I I - - - Fusel ages of mi ni mum
d r a g . (Unpublished.) 1954.
The dr ag of source di st ri but i ons in linearised supersonic flow. College of
Aeronaut i cs Repor t 88, Febr uar y, 1955. A. R. C. 17,478. Febr uar y, 1955.
(Also A.R.C. 17,836 (Unpublished.).)
Appl i cat i on of wi ng- body t heor y to dr ag reduct i on at low supersonic speeds.
N.A.C.A. Research Memo. A54J19 ( TI B 4534). J anuar y, 1955.
Li neari sed supersonic flow. Nor t h Amer i can Avi at i on Inc. Repor t AL 222.
June, 1947.
Recent devel opment s in t he t heor y of wi ng-body wave ctrag. J . Ae. Sci.,
Vol. 23, No. 12. December, 1956.
The wave dr ag of non-lifting combi nat i ons of t hi n wings and ' non- s l ender '
bodies. R. & M. 3076. March, 1957.
Lhni t e sonique de la resi st ance d' ondes d' un aeronef. Comptes Rendus,
Tome 236, No. 26, pp. 2479 t o 2480. June, 1953.
Li neari sed compressi bl e, flow t heor y for sonic flight speeds. N.A.C.A.
Repor t 956. ( For mer l y N.A.C.A. Tech. Not e 1824.) ' 1950.
The dr ag of non- pl anar t hi ckness di st ri but i ons in supersonic flow. Aero.
Quart., Vol. VI. , Par t 2. May, 1955.
A not e on t he appl i cat i on of t he supersonic ar ea rule t o t he det er mi nat i on
of t he wave dr ag of r ect angul ar wings. J . Fl ui d Mech. , Vol. I I , Par t 6.
August , 1957.
Theoret i cal det er mi nat i on of t he mi ni mmn dr ag of airfoils at supersonic
speeds. J . Ae. Sci., Vol. 19, No. 12. December, 1952.
Sl ender-body t he or y- - Re vi e w and extension. J . Ae. Sci., Vol. 20, No. 2.
Febr uar y, 1953.
Li neari sed t heor y of s t eady hi gh-speed flow. Cambri dge, London, 1955.
Methods for predi ct i ng phenomena in t he hi gh-speed flow of gases. J . Ae.
Sci., Vol. 16, No. 2. Febr uar y, 1949.
Theoret i cal dr ag rise of wings at t ransoni c speeds. R. A. E. Tech. Not e
Aero. 2267. A.R.C. 16,785. April, 1954.
25
No. Author
24 L. E. Fraenkel and H. Por t noy. .
25. L. E. Fraenkel . . . . . .
26 M . J . Lighthill . . . . . .
27 R. A. Bishop and E. G. Cane . .
28 K. D. Thomson . . . . . . .
29
3O
31
. 32
33
34
35
36
37
W. T. Lord . . . . . .
W. T. Lord and A. A. Bennett ..
E. Eminton . . . .
R. Dickson and L. B. Jones
R. T. Jones . . . . .
M. D. Friedman and D. Cohen
C. Rennemann . . . .
K. 0swatitsch and F. Keune
i
M. A. Faget . . . . . .
R E F E R E N C ES- - cont i ~cued
Title, etc.
Supersonic flow past slender bodies with discontinuous profile slope. Aero.
Quart., Vol. VI, Part 2. May, 1955.
Supersonic flow past slender bodies of elliptic cross-section. R. & M. 2954.
May, 1952.
Supersonic flow past slender bodies of revolution the slope of whose meridian
section is discontinuous. Quart. J. Mech. App. Math., Vol. 1, Part 1.
March, 1948.
Charts of t he theoretical wave drag of wings at zero lift. C.P. 313. June,
1956.
A numerical method for determining the linearised flow over wings at
supersonic speeds. Aero. Res. Lab. (Australia)Tech. Note HSA TN20.
January, 1955.
Wave drag of ' smoot h' configurations. R.A.E. Tech. Note Aero. 2417.
November, 1955.
Wave drag of unswept wings. R.A.E. Unpublished.
On the minimisation and numerical evaluation of wave drag. R.A.E.
Report Aero. 2564. A.R.C. 19,2i2. December, 1955.
Computation of the wave drag of bodies of revolution. Englisll Electric
Co. Ltd. Aero. Tech. Memo. Am 15. 1955.
Leading-edge singularities in thin-aerofoil theory. J. A e. Sci., Vol . 17,
No. 5. May, 1950.
Arrangement of fusiform bodies to reduce tile wave drag at supersonic
speeds. N.A.C.A. Tech. Note 3345. November, 1954.
Minimum drag bodies of revolution in a non-uniform supersonic flow field.
N.A.C.A. Tech. Note 3369. February, 1955.
An equivalence theorem for low-aspect-ratio wings in transonic flow.
R.A.E. Library Translation 545. August, 1955. (Translated by K. W.
Mangler from Z. Flugwiss, Vol. 3, No. 2, 1955.)
Unpublished N.A.C.A. Report. 1955.
A P P E N D I X
Nu me r i c a l Pr ocedur es f o r Cal cul at i ng the Me a n or Tr ans f e r r e d
Ar e a Di s t r i but i on o f a Wi n g
Si nc e t h e t r a n s f e r r ul e h a s b e e n r e c o mme n d e d f or wa v e - d r a g c a l c u l a t i o n s i t is a d v i s a b l e t o
d e v e l o p a s i mp l e n u me r i c a l p r o c e d u r e f o r c a l c u l a t i n g t h e me a n or t r a n s f e r r e d a r e a d i s t r i b u t i o n
A( x ) (see e q u a t i o n (7)). Th e e l e me n t a l a r e a d i s t r i b u t i o n s ( Se c t i on 2. 2. 1) c o u l d b e c a l c u l a t e d a n d
t h e n t h e me a n f o u n d . Thi s me t h o d ha s t h e a d v a n t a g e t h a t al l t i l e e l e me n t a l a r e a d i s t r i b u t i o n s
c a l c u l a t e d f or a g i v e n Ma c h n u mb e r c a n b e i n c o r p o r a t e d i n c a l c u l a t i o n s f or hi ghe r Ma c h n u mb e r s ;
h o we v e r , i t i s l i ke l y t o b e u s e f u l o n l y wh e n a s ma l l n u mb e r of e l e me n t a l a r e a d i s t r i b u t i o n s ar e
r e q u i r e d t o r e p r e s e n t t h e wi ng. A n u me r i c a l p r o c e d u r e de a l i ng wi t h o n l y t h e me a n a r e a di s t r i -
b u t i o n ha s b e e n g i v e n b y Faget 37; a simple d e r i v a t i o n of t hi s me t h o d i s p r e s e n t e d be l ow.
Suf f i ci ent de t a i l s a r e g i v e n f or t h e me t h o d t o b e u s e d wi t h o u t r e f e r e n c e t o F a g e t ' s p a p e r '~7, wh i c h
ma y n o t b e r e a d i l y a va i l a bl e .
Fi r s t l y , t h e wi n g i s d i v i d e d i nt o a n u mb e r of s t r e a mwi s e s t r i ps so t h a t t i l e c r o s s - s e c t i o n a l a r e a
d i s t r i b u t i o n of e a c h s t r i p def i nes a b o d y wh i c h i s t o b e p l a c e d a l ong t h e c e nt r e - l i ne of t h e s t r i p.
Ta k e one of t h e s e b o d i e s a n d l et i t s d i s t a n c e f r o m t h e axi s b e R a n d i t s c r os s - s e c t i ona l a r e a
d i s t r i b u t i o n b e ~/~(x). T h e n T( x ) i s r e p l a c e d b y a n u mb e r ; N, of f i ni t e o v e r l a p p i n g e l e me n t s ,
26
each of constant cross-sectional area. Let the nt h element, having the cross-sectional area S,,,
ext end from x = X, to x = XI~_,, (n = 1, 2 , . . . , N ; K = 2N + 1). Thus T( x ) vanishes outside
the range X1 ~< x ~< X2N and
N
, Z ~ ( x ) - ~ , S , , ; s , , . = S f or X;,. ~< x ~< X,~ . . . . ( K = 2N + ] ) .
Faget 37 chooses the areas S,~ so t hat t hey are equal to each other. Now, equation (7) can be
written as
A ( x ) = 1 - t
Jxi V{ B2R - (x -
where ~A (x) is the contribution of a streamwise strip to the transferred area distribution A (x).
Therefore
S N
( } A( ) = ~ - , ~ 1 _ l ' , ' f
where, for X~_,, ,-- B R <~ X~ - / B R
q- si n -1 ~ , X , - - B R < x <~ X I , _ , , - - B R
si n-~ ( X K c ~ R X) + s i n - i ( x B ~ " ) , X~(_n - - B R <~ x <~ X,~ q- B R I , , =
si n-' X ~ X + 2 ' X , + B R <~ x <~ XK_ n @ B R
and, for X.v_,, - - B R >~ X~ + B R
2- + sin -~ " , X , , - B R <~ x <~ X, , + B R
I , , = ~ , X, , + B R <~ x <~ X x _ , , - - B R
sin_ , X~: x Jr- ~ , X K _ , , - B R <~ x <~ XI(_,, + B R
(Note t hat I, , is defined for X, , - - B R <~ x <~ X, : _ , + B R and t hat I,, = 0 when x = X,, -- B R ,
X K _ n -Jr- B R ) .
The function I, , / ~ given here can be expressed in terms of inverse cosines ; t hen it is identical
with one of Faget' s functionsaL Furt her details of the numerical procedure for evaluating A (x)
may be found in Ref. 37. The fact t hat I,, is symmetrical about x ---- I(X, + X~_,,) may enable
t he calculations to be simplified.
In any application of Faget' s procedure ~7 it is not recommended t hat N be large or t hat a
large number of streamwise strips be taken. For example, the wi dt h of the streamwise strips
may conveniently be chosen to be of the same order as the body diameter while the value of N
used will depend upon t he wing section shape. Furthermore, the approximation to A (x) obtained
by this procedure can always be improved since the end points of the transferred area distribution
are defined exactly by the Mach diamond enclosing the wing system. The volume condition
is also useful.
At low supersonic speeds the procedure can be simplified and, from Section 2.4.1, it follows t hat
This result is valid when B R is ' not-so-small '. When B R is small it becomes $A (x) ~ T( x ) ,
which is the slender-body result for ~.A (x).
27
X
J
M> B
FIG. 1.
Z
r
> Y
PLANE x = CONSTANT
System of co-ordinates for wing-body combination.
28
(71666) Wt . 52]8943 K.7 4158 Hw. PRINTED ]N GRnAT BRITAIN
R. & M. No . 3 0
Publications of the
Aeronautical Research Council
A N N U A L T E C HN I C A L REP ORTS OF THE A E R ON A U T I C A L
R E S E A R C H COUNCI L ( B OUND VOL UME S )
I 939 Vol. I. Aerodynamics General , Performance, Airscrews, Engines. 5os. (52s.).
Vol. I I . Stability and Control, Fl ut t er and Vibration, Instruments, Structures, Seaplanes, etc.
63s. (65s.)
I 94o Aero and Hydrodynami cs, Aerofoils, Airscrews, Engines, Hut t er , Icing, Stability and Control,
Structures, and a miscellaneous section. 5os. (52s.)
I 94I Aero and Hydrodynami cs, Aerofoils, Airscrews, Engines, Fl ut t er , Stability and Control,
Structures. 63s. (65s.)
1949- Vol. I. Aero and Hydrodynami cs, Aerofoils, Airscrews, Engines. 75s. (77s.)
Vol. I I . Noise, Parachutes, Stability and Control, Structures, Vibration, Wi nd Tunnel s.
47 s. 6d. (49 s. 6d.)
1943 Vol. I. Aerodynamics, Aerofoils, Airscrews. 8os. ( 8 2 s . )
Vol . I I . Engines, Flutter, Materials, Parachutes, Performance, Stability and Control, Structures.
9os. (92s. 9d.)
1944 Vol. I. Aero and Hydrodynami cs, Aerofoils, Aircraft, Airscrews, Controls. 84s. (86s. 6d.)
Vol. II. Fl ut t er and Vibration, Materials, Miscellaneous, Navigation, Parachutes, Performance,
Plates and Panels, Stability, Structures, Te s t Equi pment , Wi nd Tunnels.
84~. (86s. 6d.)
1945 Vol. I. Aero and Hydrodynami cs, Aerofoils. 13os. (132s. 9d.)
Vol. I I . Aircraft, Airscrews, Controls. 13os. (132s. 9d.)
Vol. I I I . Fl ut t er and Vibration, Instruments, Miscellaneous, Parachutes, Plates and Panels,
Propulsion. I3os. (132s. 6 d . )
Vol. I V. Stability, Structures, Wi nd Tunnel s, Wi nd Tunne l Techni que. 13os. ( I 32s. 6d.)
Annual Reports o f the Aeronauti cal Research Counc i l - -
I 9 3 7 2s. (2s. 2d.) 1938 rs. 6d. (i s. 8d.) 1939-48 3 s. (3 s. 5d.)
I ndex to all Reports and Memoranda published in the Annual
Techni cal Reports, and s eparat el y- -
A p r i l , 195o - R. & M. 2600 2s. 6d. ( 2n l od. )
Author I ndex t o all Report s and Memoranda o f the Aeronauti cal
Res earch C o t m c i l - -
I 9O9- - Januar y, I 954
I ndexes to the Te c h n i c a l Reports
Counc i l - -
D e c e m b e r I, I 9 3 6 - - J u n e 3o, 1939
July I , I 9 3 9 - - J u n e 3 o, 1945
July I, I 9 4 5 - - J u n e 3 o, 1946
July I, I 9 4 6 - - De c e mb e r 3 I , 1946
January I , I 9 4 7 - - J u n e 3% 1947
Publ i shed Report s and Memoranda
Co unc i l - -
B e t w e e n Nos. 225 I - 2 3 4 9
Between Nos. 2351- 2449
Between Nos. 2451- 2549
Between Nos. 2551- 2649
Between Nos. 2651- 2749
R. & M. No. 2570 I SS. (ISS. 8d.)
o f the Aeronauti cal Research
R. & M. No. 1850 IS. 3 d. (rs. 5d.)
R. & M. No. 195o is. (i s. 2d.)
R. & M. No. 2050 is. ( i s. 2d.)
R. & M. No. 215o IS. 3 d. (IS. 5d.)
R . & M . No. 2250 IS. 3 d . ( l s . S d . )
o f the Aeronaut i cal Research
R. & M. No. 2350 IS. 9 d. (i s. I l d . )
R. & M. No. 2450 2s. (2s. 2d.)
R. & M. No. 2550 2s. 6d. (2s. l od. )
R. & M. No. 2650 2s. 6d. (zs. l od. )
R. & M. No. 2750 2s. 6d. (2s. l od. )
Prices in hrackets include postage
HER MAJESTY'S STATIONERY OFFICE
York House, Kingsway, London W.C.2 ; 423 Oxford Street, "London W.i ; Ha Castle Street, Edinburgh 2;
39 King Street, Manchester z ; z Edmund Street, Birmingham 3 ; I9 St. Mary Street, Cardiff; Tower Lane, Bristot ~ ;
8o Chiehester Street, Belfast, or through any bookseller.
S.O. Code No. z3-3o77
R. & M. No . 3 0

You might also like