R. ' & M. No.

2817
(11,148)
A.R.C. Teeh~_~ical Report
MI NI STRY OF SUPPLY
AERONAUTICAL RESEARCH COUNCIL
REPORTS AND MEMORANDA
3 0 3!}~ 1 9 5 4
Aileron Reversal and Wing Divergence
of Swept Wings
By
E. G. BROADBENT, B.A.,
and
OLA MANSFIELD, B. Sc.
Crown Copyright Reserved
LONDON" HER MAJESTY'S STATIONERY OFFICE
1954
PRICE 7s 6d NET
Aileron Reversal
of
and
Swept
By
Wing
Wings
E. G. BROADBENT, B.A.,
and ,,
OLA MANSFIELD, B. Sc .
Divergence
COMMUNICATED BY THE PRI NCI PAL DIRECTOR OF SCIENTIFIC RESEARCH (AIR),
MINISTRY OF SUPPLY
Reports Memoranda No. 2817
Septemher, 194 7
3 0 ,]UN 1954
L ] B R A k Y
Summary.--A met hod of solution for t he aileron reversal speed of a swept wing (with emphasis on sweepback) is
developed on the lines of strip and semi-rigid theories.
The influence of the following par amet er s is investigated : - -
(a) The degree of sweep.
(b) Wing torsional and flexural stiffness.
(c) Wing plan-form.
(d) Aileron plan form.
Families of curves are given for extended variation of these parameters which may be used for the direct estimation
of the reversal speed of a given wing by interpolation.
A solution is given for the wing divergence speed of a swept wing.
The general results have been obtained using simple modes of wing deformation but equations are quoted for any
given modes of deformat i on and the adopted modes are compared wi t h the act ual deformations produced by the aero-
dynami c loading for an extreme case. A suggestion is put forward for improving the accuracy of the semi-rigid approach
by an iterative met hod of solution and the flexural mode of distortion is investigated for a particular case.
1. I nt roduct i on. - - Swept wings can produce serious st ruct ural problems owing to elastic
effects t hat are uni mport ant in an unswept wing 1. Among these problems is the design of a
wing for high aileron (or elevon) reversal speed and for adequate rolling power at operational
speeds, since loss of aileron power is caused not only by twisting t but also by bending t of the
wings. In swept-back wings of conventional layout both the bending and twisting deformations
produce changes of aerodynamic incidence t hat are unfavourable from the point of view of
aileron reversal. Some relief from these effects is probabl y obtained from the distortion of the
wing section, the upward bending of the wing producing camber t hat is concave upwards and
therefore tends to counteract the reduction of the incidence.
The present report covers the theoretical consideration of aileron reversal and wing divergence
for swept wings with particular emphasis on aileron reversal of swept-back wings and examines
the effects of the following paramet ers : - - t he degree of sweep, wing stiffness, wing and aileron
plan f o r m. The results are presented as families of curves which may be used for the direct
estimation of the reversal speed for any given wing.
* R. A. E. Report Structures 9, received 14th Januar y, 1948.
t In the present report t he terms bending and flexure refer to bending along the flexural axis, and twisting and
torsion refer to rot at i on about this axis.
1
A
I t will be clear t hat quant i t at i ve accuracy of t he results is l i mi t ed by t he errors of t he aero-
dynami c and elastic derivatives. The inaccuracies of t he l at t er, using semi-rigid t heor y and
simple modes of deformat i on are well un.derstood ~,~,'~,~ and lead to an est i mat e of t he aileron
reversal speed of about 85 to 95 per cent of t he speed appropri at e to t he correct elastic deforma-
tion. The errors in reversal speed i nt roduced by t he aerodynami c assumptions, using st ri p
t heor y and simple sweep and compressi bi l i t y corrections, are more uncer t ai n but would n o t b e
expect ed to exceed 2 5 per cent. If t he aerodynami c deri vat i ves and elastic deformat i ons
for any specific wing were avai l abl e from exper i ment al or t heoret i cal sources t hen great er
accuracy t han t he above woul d be obt ai nabl e.
The results gi ven in t he families of curves of t he var i at i on of t he paramet ers enabl e a com-
pariso n to be made of t he aileron reversal speeds of a swept and unswept wing. As a rough
appr oxi mat i on for wings of t he same pl an form and of average stiffness rat i o t he wing t orsi onal
stiffness would have to be increased by 40 per cent t o mai nt ai n t he same reversal speed if t he
wing were swept back t hr ough 45 deg.
The following factors are favourabl e for increasing t he reversal speed of swept -back wings : -
hi gh wing stiffnesses in torsion and flexure, sweepback of t he flexural axis, hi gh wing t aper in
pl an form, large aileron chord and t he use of bal ance-t abs. The effect of change of wing camber
due t o wing bendi ng in t he case of an average wing is to increase t he reversal speed by not more
t han 5 per cent.
Wi ng divergence is not liable to be serious for swept -back wings, as t he wing bendi ng provides
a stabilising effect, but wilI he serious for a swept-forward wing. The following factors are
favourabl e for increasing t he divergence speed of swept-forward wings : - - hi gh wing stiffness,
especially in flexure, forward posi t i on of flexural axis and hi gh wi ng-t aper in pl an form.
2. Aerodynamic Assumpt i ons. - - A di agram of t he st rai ght t apered wing considered is shown
in Fig. 1.
The aer odynami c assumpt i ons appl y t o bot h t he semi-rigid t r eat ment (section 3) and t he
i t er at i ve process (Appendix I) and are as follows : - -
(a) The aer odynami c axis lies along t he quart er-chord.
(b) The rat i o of t he aileron chord t o t he wing chord is const ant over t he ai l eron span. (The
i t er at i ve process (Appendix I) could t ake account of var i at i ons of aileron chord.)
(c) The aileron is t orsi onal l y rigid.
(d) Aerodynami c forces on strips parallel to t he ai rcraft centre-line are expressible in t erms
of t he local chord and incidence change. The l at t er consists of t he component s due t o
wing t orsi on and t he change of incidence due to wing bendi ng.
(e) Change of wing camber due to wing bendi ng has in t hi s work been considered separat el y
(section 4) because of t he uncer t ai nt y rel at i ng t o t hi s effect. I t is assumed t hat dis-
t ort i ons of t he aerodynami c section ot her t han those whi ch can be r epr esent ed as a simple
change of camber do not occur.
(f) The t heoret i cal incompressible st ri p-t heory coefficients al, a~ and m have been used
and assumed const ant al ong t he span.
(g) Allowance for compressi bi l i t y can be made by appl yi ng t he Gl auert correct i on t o t he
reversal speed as gi ven by Figs. 2 to 12. In a specific cal cul at i on corrections shoul d be
appl i ed t o t he i ndi vi dual coefficients.
(h) Allowance for sweepback is made by reducing t he aerodynami c coefficients by @cos/3.
This l aw of a/cos/~ is not ri gi dl y justifiable on ei t her a t heoret i cal or experi ment al basis.
I t was first suggested, so far as t he aut hors are aware, by McKi nnon Wood 7 and may be
perhaps as much as 20 per cent out.
2
I t is hoped t hat a more accurat e aer odynami c t r eat ment will be developed on t he lines of
Fal kner ' s approach (R. & M.19106). No det ai l ed t hought has yet been gi ven to t he appl i cat i on
of Fal kner ' s met hod but i t seems t hat its use would i mmensel y ext end t he comput at i onal labour,
par t i cul ar l y of an i t er at i ve met hod. Some refi nement could, of course, be obt ai ned wi t hout
compl i cat i on by appl yi ng simple factors to t he st r i p- t heor y coefficients, t hese factors bei ng
deri ved by compari son of st ri p t heor y and Fal kner ' s met hod appl i ed to cert ai n t ypi cal wi ng
configurations.
3. The Semi - ri gi d Treat ment of the Pr obl e m. - - The wi ng shown in Fig. 1 is consi dered to be
bui l t -i n at t he root ands i nce t he present i nvest i gat i on is l i mi t ed to t he eval uat i on of t he reversal
speed when t he rolling moment and rolling vel oci t y under appl i ed aileron are bot h zero t hi s
assumpt i on is justified. The ot her elastic assumpt i ons are as follows : - -
(a) The fl exural axis is st rai ght . For gener al i t y t he flexural axis is assumed to be at a
di st ance eoeo behi nd t he quart er-chord at t he wing root and to be at an angle of e to t he
aer odynami c sweepback, e bei ng smal l compared wi t h ft.
( b) The fl exural mode of di st ort i on is t aken to be parabolic*.
(c) The t orsi onal mode of di st ort i on is t aken to be linear-~.
(d) The reference section is at t he mi d-span of t he aileron.
I n Appendi x I I t he equat i ons are gi ven for general modes of di st ort i on, and a t ri al of t hose
adopt ed is made for a par t i cul ar const ruct i on.
3.1. Deri vat i on of the Eq u a t i o n . - - I t is assumed t hat t he aer odynami c forces on a wi ng st ri p
paral l el to t he ai rcraft centre-line may be expressed i n t erms of t he local incidence and chord.
The forces on such a st ri p are gi ven in t he form of a lift and a pi t chi ng moment (both act i ng in
t he vert i cal pl ane t hr ough t he strip) and t he l at t er is resolved i nt o two moment s about and
al ong t he axis of bendi ng.
v
O~, o[
The y ' co-ordinate is t aken to be al ong t he axis of bendi ng and s' is t he val ue of y ' at t he tip.
We t hen have on t he appl i cat i on of aileron, for a st ri p dy ;
dL = qc dy' (a~o: q- as~) cos/3 . . . . . . . . . . . . . . (1)
where L represent s t he lift force, q t he dynami c pressure, c t he chord, ~ t he local i ncr ement of
incidence and ~ t he local aileron angle (measured paral l el to t he centre-line). The coefficients
al and as have t hei r usual significance (al = ~CL/ ~ and as = ~ CJ ~ ) and are assumed const ant
over t he span.
* Some justification of this assumption is given in Appendi x I.
t This assumpt i on has been shown t o be sat i sfact ory in t he case of t he unswept wing (R. & M. 21864).
3
I n a s i mi l ar ma n n e r
dM = - - qc ~dy' {m~ - e(a2S + a l ~ ) } c o s ~ . . . . . . . . . . ( 2 ) i
whe r e M denot es t he mo me n t a bout an axi s t hr ough t he f l exur al cent r e nor ma l t o t he ai r cr af t
cent r e- l i ne ; nose- up mo me n t s are pos i t i ve. The di s t ance of t he f l exur al axis aft of t he quar t et !
chor d is de not e d by ec, and m = - - ( a C , , , / a S ) C L cons t ant .
The mo me n t M is now r esol ved i nt o t wo c o mp o n e n t s - - a mo me n t M1 about t he f l exur al axis,,
and a mo me n t M~ about a l i ne nor ma l t o t hi s axis. I n a ddi t i on t he equat i ons are wr i t t e n in.
non- di me ns i ona l form, and be c ome
d L = qCoS' (1 - - ~ ) ( a l c + a=S) cos/~ d~7 . . . . . . . . . . (3)
d M 1 = - - qCo2S'(1 - - ,~)2{mS - - e ( a 2 S + a~c)} cos2fl &] "~
d M a = - - qCo~S'(1 - - z~)~{mS - - e( a=S + alc0} cos/~ si n fl d,~ -_, . . . . (4)
% ,
He r e ~ = y / , and t he t aper ~ is def i ned by
c = Co( 1 - ~ )
wher e co is t he r oot chor d.
I n or der t o r el at e t he a e r odyna mi c l oads gi ven by equat i ons (3) a nd (4) t o t he s t r uct ur al !:
di st or t i ons, t wo semi - r i gi d mode s are now chosen. For Si mpl i ci t y t he be ndi ng mode is t a ke n
t o be par abol i c, and t he t or s i onal mode t o be l i near ; t he equat i ons for gener al modes are gi ven ii
i n Appe ndi x I I .
Thus if 0 denot es t he angl e of t wi s t ( about t he f l exur al axis)
o = O o 0 / , 7 o ) . . . . . . . . . . . . . . . . . . ( s )
and if z is t he ver t i cal di s pl a c e me nt of a poi nt on t he f l exur al axi s
z = ~ 0 ( ~ / ~ 0 ) ~ . . . . . . . . . . . . . . . . . . ( 6 )
wher e t he suffix o refers t o a r ef er ence sect i on. By di f f er ent i at i on of e qua t i on (6) t he mode of
be ndi ng is obt a i ne d i n t er ms of t he sl ope W
l d z l 2 z o ( ~ )
' P - - s ' d ~ - - S ' ~ o ~v = ~0 . . . . . . . . . . . . . . (7)
The l ocal i nci dence is gi ven by
= 0 cos/3 + ~0 sin/~ . . . . . . . . . . . . . . . (8)
a nd t he l ocal ai l er on angl e by
= S ~ - 0 c o s / ~ . . . . . . . . . . . . . . . . . ( 9 )
wher e Sl is a c ons t a nt equal t o So + 0o cos ~, and So is t he val ue of t he ai l er on angl e at t he
r ef er ence sect i on.
The a e r odyna mi c l oads gi ven by equat i ons (3) a nd (4) can now be t r ans f er r ed by t he pr i nci pl e
of wor k t o e qui va l e nt l oads at t he r ef er ence sect i on. The r el at i ons are
d L ' = (rj/rjo) ~ d L )'- (10)
a nd d M ' = (rl / ~o) d M * f . . . . . . . . . .
Af t er car r yi ng out t he s ubs t i t ut i ons (3), (4), (5), (7), (8) and (9) i n t he equat i ons (10), a nd
ma ki ng t he f ur t her s ubs t i t ut i on
CoCo + s ' v e sec 3
e = co(1 - - Tn) . . . . . . . . . . . . . . . . (11)
* Thi s e q u a t i o n hol ds f or b o t h c ompone nt s of M si nce b o t h mode s (in t e r ms of t he l ocal angl e) ar e l i near i n y.
4
he i nt egr at i ons can be effect ed to gi ve L' , MI ' a nd M( .
~'ritten
/ a , l A l ( ~
L ' = q c o s ' cos/3 ~ ~ ~
0 c s / 3 [ ! - ' ~ ~ o 3 4
5
r ( 1 - 4 ) 1
The r esul t s of t hi s process ma y be
{ I " ' ] J 1 ' = qCo 2 s ' c o s 2 f l - - m ~ 1 - - 7 2 2 , ( 1 - - y3) + (1 y 4)
2 3
q mOo~o ~ c s ~ 1 3r ~ 2 - c 4 ( i _ r ~) + g ( 1 - - r ~) + ~o L 2
q_ 1 a ~ t s ' s sec/3 + e_o ( a i A ~ - - a # o cos/3)
Co ~o 3 4
+ (a~A~ - - a 2 0 o cos/3) t y~
~o2Co 4 5
nd M( = M~' t an/ 3 ;
; her e A~ = 0o cos/3 + ~o si n ft.
- < ) l
5
. . ( 1 2 )
The expr essi ons for L' , MI ' and M, ' ma y now be r el at ed to t he el ast i c st i ffness b y t he pr i nci pl e
f work. The equat i ons are
moOo = M I ' ; (13)
. . , , . . . .
I~*?o = 4 M ( - - 2 L ' r l o S '
J
I t r emai ns to equat e t he t ot al r ol l i ng mo me n t to zero, for t he r ever s al . condi t i on
- 1
[ ~ dL = 0 . . . . . . . . . . . . . . . . . . . ( t 4 )
d O
The onl y unknowns occur r i ng i n equat i ons (13) a n d (14) are q a nd t he r at i os ~1/00, ~Oo/00.
"he equat i ons can t her ef or e be sol ved for q to gi ve t he r ever s al speed.
3.2. N u m e r i c a l S o l u t i o n . - - S o l u t i o n s ha ve been obt a i ne d for a series of va r i a t i ons of t he
~eometric par amet er s , as shown i n Tabl e 1, whi ch also gi ves t he cor r espondi ng Fi gur e numbe r .
TABLE 1
Geomet ry Correspondi ng Fi gure
Taper rat i o r = 0. 75
Aileron span (1 - - y) y = 0-6
Aileron chord E = 0. 25
Aspect rat i o A = 6
Fl exur al axis f ~ = 0
\ eo = 0
Vari at i on of r
---- 0. 75, 0. 5, 0. 25 2, 4, 3
Vari at i on of y
y ---- 0. 5, 0. 6, 0. 7 7, 2, 8
$
TABLE 1 - - c o , # i m b e d
Ge ome t r y Cor r es pondi ng Fi gur e
Var i at i on of E
E = 0. 2, 0. 25, 0 . 3 9, 2, 10
Var i at i on of e
s = 0. 0, 0. 02, 0. 04 2, 11, 12
Var i at i on of A
A = 4, 6 , 8 5 , 2 , 6
The following values have been assumed for t he st andar d wing : -
Aspect rat i o A = 6
i
Wi ng t aper ~ = 0" 75 ( i . e . , rip ~hora
rooo~chor~ -- 0"25)
Aileron span ~ = 0- G ( i . e . , ~o,.on sp~ _ O. 4)
W i n g s p ~ n - -
Aileron chord E = 0. 25 (i.e., Ailoro,~oho,l
Wi.g ,l,or~ -- 0"25)
Fl exur al axis ~( ~o= 0 ( i . e . flexural axis at a-chord)*
= 0 ' il
Reference section ~0 = 0-8 ( i . e . , Reference section at mi d-ai l eron position)
Usi ng t he above shown st andar d val ues t he equat i ons for t he condi t i ons at reversal for 40 deg
sweep reduce to
M o = 0. 127 + 0.107/5 ~
L, = 0. 498 + 0 . 5 9 3 / 1 5 f . . . . . . . . . . . . . . ( 1 5 )
where M o = q c , 2 s , L, -- p : and c , , is t he mean chord.
~CmS 2
The equat i ons (15) cl earl y represent a r ect angul ar hyper bol a wi t h asympt ot es M o = 0.12']:
and L, = 0. 498. The stiffnesses are here expressed in non-di mensi onal form dependi ng onl
t he reversal speed.
The hyper bol a appropri at e to equat i on (15) is pl ot t ed in Fig. 2 and i ncl uded i n t he same
gr aph are curves represei l t i ng angles of sweep of 0, 30 deg, 35 deg and 45 deg. A series of
si mi l ar curves covering t he var i at i on i n t he par amet er s of Tabl e 1 are gi ven in Figs. 3 to 12.
I t will be not i ced t hat all t he curves possess t he same f or m- - t ha t of r ect angul ar hyperbol ae
except where/3 = 0, when, of course, t he reversal condi t i on is i ndependent of flexural stiffness.
The i nt er pr et at i on of t he curves is si mpl y t hat any condi t i on above and to t he ri ght of t he
curves corresponds to posi t i ve control. Thus for posi t i ve cont rol to be avai l abl e at all fl yi ng
speeds i t is necessary t ha t t he poi nt ( M o , L , ) appropri at e to t he design di vi ng speed shoul d lie,
above and to t he ri ght of t he curve referri ng to t he geomet r y of t he par t i cul ar ai rcraft concerned.i
The pract i cal region for t he var i at i on of tile non- di mensi onal par amet er s is shown in Fig. 2.1
Thi s region has been obt ai ned st at i st i cal l y, al t hough of necessi t y t he st at i st i cs appl y chi efl y
* Thi s a s s umpt i on was ma de t o si mpl i f y t he cal cul at i ons and is s hown l at er (sect i on 3.3.3) t o be j ust i f i ed b y the=
i ns ens i t i vi t y of r ever sal t o e o.
6
to unswept ai rcraft . The lower l i mi t s shown for Mo and L, are appr opr i at e to t he design di vi ng
speeds (not reversal speeds) of t he appr opr i at e ai rcraft and represent t he lowest values gi ven
by t he survey. There does seem, however, to be a t endency for swept ai rcraft in t he desi gn
st age at present to have lower values of L,. Thi s applies par t i cul ar l y to hi gh speed long range
ai rcraft .
I t is not easy to draw det ai l ed inferences from t he families of curves about t he influence of
sweepback, as t here is no common assumpt i on whi ch can easi l y be made for t he geomet r y of
wings wi t h different amount s of sweep. The vari ous curves allow direct conl pari son for wings
of t he same area and same aspect rat i o (except for t he case in whi ch aspect rat i o is varied, when
direct compari son is impossible).
3.3. Effect of Parameters on Aileron Reversal of Swept-bac~ Wi ngs. --In Fig. 13 are pl ot t ed
a series of graphs of t he t orsi onal stiffness par amet er (Mo) agai nst vari ous geomet ri c par amet er s
for a selection of t hree stiffness ratios. The stiffness rat i o r is defined by
L~ l~ c (16)
r ~ - - - . . . . . . . . . . . . . . .
M o - - m o S
and in pract i ce usual l y lies i n t he region of 1. 5 to 3. The val ues of r for whi ch t he graphs are
pl ot t ed, are 1.25, 2. 5 and 5, and t he resul t s are gi ven for 45 deg sweep and compared wi t h zero
sweep. The effects of t he vari ous par amet er s can be exami ned in t ur n ; in t hi s case t he com-
parisons are made in t erms of M0, whi ch means t hat if t he stiffness rat i o is fixed by ot her con-
si derat i ons t he var i at i on of t he requi red t orsi onal stiffness will be as shown. If bot h stiffnesses
are t r eat ed as variables, i t does not, of course, follow t hat hi gh val ues of r are par t i cul ar l y
desirable, for, al t hough t hey lead to lower values of t orsi onal stiffness, t he fl exural stiffness
itself is higher. I n practice, common sense suggests t hat if t he probl em of loss of cont rol is to
det ermi ne bot h stiffnesses t he opt i mum region (from, say, a wei ght poi nt of view) in whi ch to
work, is t hat ne a r t he axis of t he hyperbol a. I t has been suggest ed by Lee s t hat qui t e a shar p
wei ght mi ni mum occurs in t hi s region.
3.3.1. The effect of swe@bac~.--Fig. 13t~ shows how Mo depends on 5 for different val ues of
r, but t he effect on design is not so easy to predi ct , and needs careful i nt er pr et at i on. I ndeed
i t is doubt ful whet her any definite conclusions can be drawn wi t hout act ual l y t r yi ng to desi gn
to some gi ven specification bot h wi t h and wi t hout sweepback, and compar i ng t he results. If
t he flexural stiffness is high, for example, sweepback appears to have a beneficial effect on
reversal due to t he fact t hat t he aer odynami c coefficients are all reduced. But , in t he first
place t he wi ng area will pr obabl y have to be i ncreased due to t he fact t hat CL ..... is less for t he
swept -back wing, and secondly, if t he aspect rat i o is kept const ant , t he t orsi onal stiffness will
have to be obt ai ned over a great er rat i o of l engt h to br eadt h. On t he ot her hand, t he aspect
rat i o will pr obabl y not be kept const ant as t he wi ng is swept back. For more nor mal val ues
of t he stiffness rat i o t he requi red t orsi onal stiffness increases wi t h sweepbac k , because of t he
i mpor t ance of t he bendi ng effect.
3.3.2. The effect of wing taper.--Fig. 13A shows t hat increase of t aper is beneficial, t hough
appar ent l y to a somewhat smaller ext ent for t he swept wi ng t han for t he unswept wing.
3.3.3. The effect of sweepback of the flexural axi s. --It has been shown (R. & M. 21864) t ha t
for an unswept wi ng t he posi t i on of t he fl exural axis is not ver y i mpor t ant , and i t mi ght be
expect ed t hat t hi s would also be t rue for a swept -back wing, as from t he poi nt of vi ew of t or si on
t he two are essent i al l y similar. Thi s has been checked by cal cul at i on and found to b e correct
as regards bodily movement of t he fl exnral axis. A numeri cal exampl e may be gi ven for t he
st andar d case for zero and 35 deg sweep, and for t he flexural axis on t he quar t er - chor d and
10 per cent chord aft of this'. The reversal equat i ons are gi ven in Tabl e 2.
TABLE 2
e o = 0 e o = 0- 1
[3 = 0 Mo = 0' 247 Mo = 0" 278
Mo = 0. 150 + 0. 105p M0 = 0. 169 + 0. 118p
/3 = :350
L~ = 0. 425 + 0" 607/ p L~ : 0"437 + 0. 624/ p
When i t is remembered t ha t t he reversal speed depends on t he square root of t he val ues for
Mo and L, i t becomes evi dent t h a t t he par amet er eo is compar at i vel y uni mpor t a nt - - a t l east
for nor mal posi t i ons of t he flexural axis.
On t he ot her hand, rel at i ve di spl acement o~ t he flexural axis along t he wi ng span can have
a ver y pronounced ef f ect - - much more so for t he swept wing t han t he unswept wing. Thi s is
shown in t he gr aph (Fig. 13B) of Mo and s, where t he flexural axis t hough assumed st r ai ght is
swept back an angle s rel at i ve to t he quart er-chord. For i nst ance, i n t he case r = 5, if e changes
from 0 to 0. 04 radi ans, t he val ue of Mo is hal ved, or for a gi ven t orsi onal stiffness (too) t he
reversal speed woul d be i ncr eased from, say, 350 m. p. h, to 500 m.p.h. Of course, t hi s has
been worked out for a hi ghl y t apered wing so t hat s = 0. 04 is equi val ent to a large di spl acement
of t he fl exural axis t owards t he wing t i p ; at t he t i p itself t he axis is displaced t hr ough 60 per
cent of t he chord. I t is pr obabl y too much to hope for such a large benefit for a pract i cabl e
wi ng const ruct i on, but t ypes of const ruct i on whi ch t end in t he reverse di rect i on (i.e., t he flexural
axis forward at t he t i p and aft at t he root) shoul d be avoi ded if possible.
3.3.4. The effect of aspect r at i o. - - Fi g. 13A shows t hat increase of aspect rat i o for gi ven stiffness
is sl i ght l y beneficial, but t he probl em of pr ovi di ng t he stiffness pr obabl y more t han count er-
bal ances t he gain.
3.3.5. The effect of aileron geomet r y. - - The effect of span changes consi derabl y wi t h sweepback
and stiffness ratio, t he change bei ng due to t he rel at i ve change in i mpor t ance of t he bendi ng
deformat i ons. For no sweepback, or for ver y hi gh flexural st i f f ness, short ailerons appear
to be f avom' abl e- - t hough agai n i t must be remembered t hat t he stiffness has to be provi ded
over a great er l engt h of wing ; but for hi ghl y swept wi ngs of lower stiffness ratio, t he l ong-span
aileron is beneficial. The effect of aileron chord is smal l in all cases, and is of t he same sign as
for t he unswept wing, namel y t ha t increase of chord is sl i ght l y beneficial.
4. Change of Camber on Wi n g Be n d i n g . - - I n t he precedi ng anal ysi s i t has been t aci t l y assumed
t hat t he aer odynami c section of t he wi ng r emai ned unchanged by t he di st ort i on of t he wi ng as
a whole. To what ext ent t hi s appr oxi mat i on is t rue i n pract i ce will depend upon t he t ype of
const ruct i on adopt ed, but if t he wing is desi gned wi t h a st andar d t orsi on-box and single spar
t hen t he wing section in t he line of flight will di st ort as t he wi ng deforms.
Thi s di st ort i on mani fest s itself as a change of camber of t he section, and t he magni t ude of
t he change may be est i mat ed from t he assumed nlodes of deformat i on. As t he mode of t wi st
is assumed to be l i near t hi s mode will not, of itself, produce any camber change ; on t he ot her
hand t he parobolic mode of bendi ng will give rise to a const ant change of camber over t he wing
span. Moreover t he effect of t hi s camber change is beneficial. For, if an aileron is appl i ed
downwards, t he wi ng bends up and acquires camber i n t he concave sense seen from above whi ch
t ends to offset t he nose-down pi t chi ng- moment from t he aileron.
The magni t ude of t he effect has been est i mat ed for a wing swept back at 45 deg, and of const ant
chord, for t hree val ues of t he stiffness ratio. The change in t he aer odynami c coefficients is
based on Gl auert ' s st andar d t wo-di mensi onal t heor y , and t he magni t ude of t he effect in t erms
of reversal speed was in all cases bet ween 3 and 4 per cent.
8
5. W i n g D i v e r g e n c e . - - T h e wi ng considered is shown in Fig. 1 and t he same aer odynami c
and elastic assumpt i ons are made as for t he semi-rigid t r eat ment of aileron reversal in section 2
wi t h t he except i on t hat t he fl exural axis is at a const ant fract i on of t he chord (e) aft of t he
quart er-chord, for si mpl i ci t y. When t he wi ng acquires a posi t i ve i ncr ement of incidence,
t he lift forces are i ncreased and t he wi ng bends upwards, and t wi st s nose up. I n t he absence
of bendi ng t he t wi st of t he wi ng woul d agai n increase t he lift forces and a st abl e condi t i on or
di vergence woul d result, dependi ng on t he di st ur bi ng ai r forces and t he elastic rest ori ng forces.
The upwar d bendi ng reduces t he effective i nci dence and so for a swept -back wi ng t he bendi ng
provi des a st abi l i si ng effect.
The equat i ons for l i ft and moment are : - -
d L = qCoS'(1 - - ~) a l ~c o s ~ d~ "~
J
d M = eqco2s'(1 - - ,~)2a~c~cos? d~ .
By assumi ng tile modes gi ven by equat i ons (5) and (6) t hese ma y be wr i t t en
-%
d L = qCoS'(1 - - ~ )
()
~ 7
d M = eqco2S'(1 - - "c~']) ~ ~ a i A ~ c o s f i d ~ J
(17)
( 1 8 )
The equi val ent
equat i ons (12) are t herefore
loads at t he reference section correspondi ng to t hose for reversal gi ven by t he
L ' = qcos ' al A~ cos/~(-} -- ~/5)/~o 3
eqc?s'alA1 cos ( 1 - + 2
M / t a n fl
equi l i bri um are, as before,
4 M~ ' - - 2 L ' v o S '
M1 .
M
2 I
and t he equat i ons of
l~o o =
moO o =
J
. . . . . . ( 1 9 )
For a t ypi cal numeri cal solution, we agai n resort to t he di mensi ons of Fig. 1 ; choosing e = 0. 2
we t hen have t he equat i ons, for/~ = 45 deg.
Mo = 0.212(1 + p) "~
%
L,~ = -- 2.24(1 -1- 1/25 ) f
(20)
Agai n t he gr aph of Mo and L~ is a r ect angul ar hyperbol a, but in t hi s case ont y a smal l par t of
t he posi t i ve quadr ant is enclosed by it. A compari son wi t h t he reversal curve is gi ven in Fig. 14
and t he compar at i ve uni mpor t ance of di ver gence is qui t e obvious. The graphs are shaded on
t he danger side.
The divergence cal cul at i ons are pr obabl y ver y pessi mi st i c as t he assumpt i on of a bui l t - i n
wi ng is not justified. If an assumpt i on of const ant lift were made, for exampl e, t he cri t i cal
speeds would be much higher. But it would not be ver y sat i sf act or y for an ai rcraft to fly at
speeds in excess of i t s ' cl assi cal ' divergence speed ( i . e. , assumi ng bui l t - i n wings) as i t woul d
be ver y suscept i bl e to gusts.
9
6. The Effects of Sweep Forward.--The met hod of sol ut i on of sections 3 and 6 has been
appl i ed to a wing similar to t hat in Fig. 1 but havi ng forward sweep and t he results pl ot t ed i n
Fig. 15. The most not i ceabl e feat ure is t hat t he divergence and aileron reversal curves now
a ppe a r in t he opposite quadr ant s from t he swept -back wing, so t hat divergence becomes more
critical t han aileron reversal for t he swept-forward wing. In part i cul ar t he flexural stiffness
requi red t o avoi d divergence is ver y high, bei ng double t hat whi ch would be adequat e to pr event
aileron reversal on a swept -back wing. Aileron reversal for a swept-forward wing is not ver y
sensi t i ve to flexural stiffness and is avoi ded ent i rel y wi t h a t orsi onal stiffness appreci abl y less
t han woul d be necessary for an unswept or swept -back wing.
As always for divergence problems i t is beneficial to keep t he flexural axis as well forward as
possible, but i t shoul d be not ed t hat whereas an unswept wing wi t h flexural axis at t he quart er-
chord is stable at all speeds, t hi s is not t rue for t he swept -forward wing, and t he flexural stiffness
requi red to pr event divergence can still be ver y hi gh for appreci abl e angles of sweep. Thi s
stiffness is mar ked in Fig. 15 (for 45 deg sweep).
7. Co~tdusions.--The conclusions are l i mi t ed quant i t at i vel y by t he limits of t he aer odynami c
t r eat ment but t he following qual i t at i ve conclusions can be drawn from t he results obt ai ned.
Loss of rolling power t hr ough st r uct ur al flexibility is, in general, more serious for sweep-
back t han for sweep-forward or for no sweep. A possible except i on is t he case of ver y hi gh
stiffness ratio. The following factors are beneficial and give hi gh aileron reversal speed.
(a) Hi gh wing-torsional-stiffness.
(b) Hi gh wing-flexural-stiffness.
(c) Sweepback of flexural axis behi nd aerodynami c axi s.
(d) Hi gh wing t aper in pl an form.
(e) Large aileron chord.
(f) The use of bal ance tabs.
Increase of aileron span is beneficial in reduci ng t he loss of incidence due to bendi ng but
det r i ment al wi t h reference to direct t orsi on effects and t hus t he opt i mum val ue would depend
upon t he stiffness rat i o and degree of sweep of t he par t i cul ar wing under consi derat i on.
The effect of change of wing camber due t o t he bendi ng of t he wing is small and sl i ght l y
increases t he reversal speed.
Wi ng divergence is not liable to be serious wi t h swept -back wings but is for swept-forward.
The following factors are found to be beneficial and give rel at i vel y hi gh divergence speeds for
swept -forward wings.
(i) Hi gh st i ffnesses--especi al l y flexurai.
(ii) Wel l -forward flexural axis.
(iii) Hi gh wing t aper in pl an form.
8. Further Devel@ments.--The met hod used in t he present report is t o be ext ended to
i nvest i gat e t he effects of wing deformat i on on t he avai l abl e roiling power at speeds below t he
aileron reversal speed. For considerations of rolling power, i t is i ncorrect to assume t he wing
hel d fixed at t he root since t he dampi ng t erms are i mpor t ant , but t he overall effects due to
sweep shoul d be similar to t hose described in t he present paper, as t he dampi ng. can har dl y
produce any abr upt changes. For t he unswept wing i t has been shown (R. & M. 2186 *) t ha t
t he mode of t wi st suffers ver y l i t t l e change over t he speed range up to reversal.
The precise requi rement s for t he cont rol power of fut ure high-speed civil and mi l i t ar y ai rcraft
and guided weapons are not clear but since positive cont rol will always be necessary at all
flying speeds i t seems Certain t hat it will be necessary t o predi ct cont rol power and reversal
10
speeds accurately for swept wings as in the case of unswept wings. The principle uncertainty
(as for all aero-elastic phenomena) is t hat of the aerodynamic data. This means t hat much
more information is required on the derivatives a,, a~, m and the location of the aerodynamic
centre with special reference to their dependence on sweep and Mach number. In particular,
the reversal speed is very sensitive to the value of m. Data~0,1~,~2 are becoming available but
more are required.
A
A1
E
L , L '
L~
M
M1, MI '
M~, M~'
Mo
P
C, Co, C t
e, e 0
f ( r i )
~ " t 4
/ i ~ o
P
q
t "
S , S '
y, y'
Z, Z o
O~
/3
O, Oo
r] , rio
~, ~o
T
LIST OF PRINCIPAL SYMBOLS
Aspect ratio
Defined by equation (12)
Ratio aileron-chord/wing-chord
Flexural mode
Lift force and equivalent at reference section
Non-dimensional stiffness (Z~/qc,,~s ~)
Pitching moment
Twisting component of M and equivalent at reference section
Bending component of M and equivalent at reference section
Non-dimensional stiffness mo/qc,,~s
The flexural rigidity (EI)
C L l a
Chord, root value and tip value
Position of flexural axis aft of quarter-chord, and value at wing root
Torsional mode
Flexural mode in slope due to distortion in.flexure
Flexural stiffness
--(~CJO~)CL
Torsional stiffness
Amplitude ratio ( = %/00)
Dynamic pressure (p V 2)
Stiffness ratio ( = @,/smo)
Semi-span, and length of wing from root to tip (= s sec ~)
Spanwise co-ordinates perpendicular to centre-line and parallel
flexural axis respectively
Vertical deflection, and value at reference section
Incidence
Angle of sweep
Inboard end of aileron is ?s from centre-line
Angle between flexural axis and quarter-chord line
Angle of twist
Non-dimensional spanwise co-ordinate ( = y/ s)
Aileron angle ; ~1 relative to fixed axes and ~ relative to wing
Angle of bending
Taper (= 1 -- c,/co)
11
to
No. Author
1 A. R. Collar . . . . . . . .
2 D. M. Hirst . . . . . . . .
3 Mary Victory . . . . . . . .
4 A. R. Collar and E. G. Broadbent . .
5 E. G. Broadbent and A. R. Collar ..
s
6 V. M. Falkner . . . . . . . .
7 R. McKinnon Wood . . . . . .
8 G. H. Lee . . . . . . . .
9 H. Glauert . . . . . . . .
10 R. Dickson . . . . . . . .
11 H. M. Lyon and M. Gdaliahu ..
12 C. F. Bethwaite . . . . . .
13 A. R. Collar, E. G. Broadbent and
Elizabeth Put t i ck
RE F E RE NCE S
Title, etc.
S.ome Problems in the Design of Swept-back Wings. A.R.C. 9264.
January, 1946. (Unpublished.)
Calculation of the Critical Reversal Speed of Wings. R. & M. 1568.
September, 1933.
The Calculation of Aileron Reversal Speed. R. & M. 2059. January,
1944.
The Rolling Power of an Elastic Wing, Part I Compress'ibility Effects
Absent. R. & M. 2186. October, 1945.
The Rolling Power of an Elastic Wing, Part I I : Compressibility
Effects and Results for Supersonic Speeds. R. & M. 2186. October,
1945.
The Calculation of Aerodynamic Loading on Surfaces of any Shape.
R. & M. 1910. August, 1943.
Notes on Swept-back Wings for High Speeds. A.R.C. 8806.
September, 1945. (Unpublished.)
Tailless Aircraft Design Problems. J.R.Ae.S. February; 1947.
Aerofoil a~2d A#screw Theory. Cambridge University Press. 1926.
The Relationship between the Compressible Flow Round a Swept-back
Aerofoil and the Incompressible Flow Round Equivalent Aerofoils.
A.R.C. 9986. August, 1946. (Unpublished.)
The Estimation of Aerodynamic Loads on Swept-back Wings. A.R.C.
9951. June, 1946. (Unpublished.)
Preliminary Note on High Speed Tunnel Tests on a Swept-back Wing.
A.R.C. 9751. November, 1946. (Unpublished.)
An Elaboration of the Criterion for Wing Torsional Stiffness. R. & M~
2154. January, 1946.
AP P E NDI X I
The Fl e x ur al Modes of Di st ort i on and It erat i ve Tr eat ment
A s o l u t i o n of t h e i n t e g r a l e q u a t i o n f or t h e r ol l i ng p o we r of a n u n s we p t e l a s t i c wi n g a i l e r on
c o mb i n a t i o n ha s b e e n r e p o r t e d ( R. & M. 21864,5) i n wh i c h t h e s o l u t i o n wa s a c h i e v e d b y me a n s
of a n i t e r a t i v e p r o c e s s a p p l i e d t o a s er i es of c h o r d wi s e s t r i ps a l ong t h e wi n g s pa n. Thi s . pr oces s
c a n b e e x t e n d e d f ul l y t o c o v e r t h e cas e of a s we p t - b a c k wi n g s u b j e c t e d t o b o t h f l e xur a l a n d
t o r s i o n a l d e f o r ma t i o n . Th e n u me r i c a l s o l u t i o n of t hi s mo r e c o mp l e x p r o b l e n l wo u l d , h o we v e r ,
b e s o me wh a t t e di ous , a n d t h e p r e s e n t i n v e s t i g a t i o n i s r e s t r i c t e d t o a n ar t i f i ci al c a s e f or wh i c h
t h e t o r s i o n a l s t i f f nes s i s a s s u me d i nf i ni t e. Th e p l a n f o r m of t h e wi n g c o n s i d e r e d is s h o wn i n
Fi g. 1.
A.1.
l oc a l f l e xur a l r i g i d i t y ( EI ) a t a s e c t i o n i s p r o p o r t i o n a l t o t h e c u b e of t h e c h o r d ;
t h u s E I = P = Po(1 - - ,r~)~ . . . . . . . . . . .
Al s o
= ~1 = c o n s t a n t
a n d ~ = dz/ dy'
Th e e x p r e s s i o n s f or d M a n d dL n o w b e c o me
d M = - - qc ~ dy rn$l
The I nt egral Eq u a t i o ~ . - - I t is a s s u me d f or t h e p u r p o s e of t hi s i l l u s t r a t i o n t h a t t h e
. . . . ( a l )
dL = qc dy ( a~ sin/3 + a~1) t
. . . . ( A 2 )
12
and furt her t he rolling moment due to t he st ri p is
d R = qcy dy (al v sin/3 + a2~). . ( A a )
The t ot al rolling moment is zero, so from (18) we may write
j~ (1 -- ~)~(aW sin/3 + a~l)d,~ = 0
0
whi ch gives a rel at i on for ~,.
( 1 4 )
The bendi ng moment (tip up) at t he section y is obt ai ned from equat i on (A2) as
J c=,n sin/3~ dy , + c( y, ' - - y ' ) ( a , v sin 5 + a ~, ) d y / q
yl y!
dv
and t hi s is eciual t o - - P dy'
The i nt egral equat i on for v is now obt ai ned by i nt egr at i on of t hi s expression.
of solution, we replace v by a funct i on h(~) such t hat
= v / v ,
where V, is t he val ue of v at t he reference section.
di mensi onal form as
- P0cos /3 ( 1 -
S 1 ~ 1
q- co
*?
where, from equat i on (A4),
For conveni ence
The equat i on may t hen be wr i t t en i n non-
f
~ (1 - - r~,) 2 A~m sin/3, d.v,
(~ -- V)(1 -- , ~) ( a~A2 - - azh sin/3) sec/3, d~ . . . (A5)
~1 f l ai~(1 - - r~)hsi nfl d~
A ~ = - - - - 1 . . . . . . . . . . ( A 6 )
The equat i ons as wr i t t en are di rect l y appl i cabl e onl y to t he st r ai ght t apered wing wi t h a
stiffness di st ri but i on similar to t ha t assumed. They could, of course, be obt ai ned qui t e general l y
in t erms of t he local values of t he chord and t he stiffness wi t hout affecting t he i t er at i ve process
of solution.
A.2. Sol ut i on of the E q u a t i o n s . - - T h e i t er at i ve met hod of solving equat i ons of t he t ype of
(AS) and (A6) has been described in previous report s 4,~, but in t he present case t he i t erat i ons
are a l i t t l e longer because of t he t er m (~1 -- ~) in equat i on (A5). The mai n steps in one i t er at i on
are gi ven i n Tabl e 3, and bri efl y expl ai ned below.
The first par t of t he i nt egr at i on from ~ to 1 (i.e., t he par t dependi ng on m) is carried out in
t he usual manner for an assumed mode h(~) i n col umn (4) and t he result gi ven i n col umn (6).
Column (7) gives t he pr oduct of t he t erms aft er t he fact or (~2 -- ~) i n equat i on (A5). Col umn
(8) r epr esent s t he bendi ng moment at st ri p 5 due to t he aer odynami c lift on st ri p 6. I t is
accordi ngl y 0. 0530 (from col umn 7) mul t i pl i ed by ( ~ o - ~5). I n t he same way col umn 9
represent s t he bendi ng moment at st ri p 4 due t o t he lift out boar d of t hi s st ri p ; i t is obt ai ned
by gumming t he bendi ng moment s due to each out boar d strip, whi ch sum is gi ven at t he bot t om
of tile column. Columns (10), (11) and (12) repeat columns (8) and (9) for t he bendi ng moment s
at strips 3, 2 and 1 respectively. The moment s on t he vari ous strips are ar r anged vert i cal l y
in col umn (13) and aft er mul t i pl yi ng by S/Co are added to those al ready found from t he pi t chi ng
13
moment deri vat i ve in col umn (6) to give col umn (14). Column (15) represent s t h e final i nt e-
grat i on, and col umn (16) t he new appr oxi mat i on to t he mode. Tabl e 3 gives, in fact, t he fi nal
i t er at i on, and i t can be seen t hat col umn (16) repeat s col umn (4) to a sufficient degree of accuracy.
The critical val ue of q is gi ven by
O. 4578qCo"S ~ = Po c o s " / 3 . . . . . . . . . . . . . . . . (A7)
I t will be not ed t hat , by hypot hesi s, t he bendi ng moment at t he ai rcraft centre-line (due t o t he
lift forces) must be zero. Thi s may be seen to be t rue by ext r apol at i ng t he number s in col umn
(13) to t he val ue ~ = 0 whi ch is done in Fig. 16. The ver y fact t hat t he r esul t ant rolling moment
is zero may per haps be t aken as an i ndi cat i on t hat t he lift forces alone can have l i t t l e influence
on t he reversal speed. That t hi s is not necessari l y so is shown by t he present exampl e where
( par t l y due to t he hi gh aspect rat i o and hi gh taper) it will be not ed t hat t he fact ored number s
of col umn (13) are appr eci abl y great er t han the correspondi ng number s of col umn (6). The
t wo are compared in Fig. 16".
As a more direct compari son of t he magni t ude of t he effect of t he lift forces on t he reversal
speed t he i t er at i ons have been repeat ed wi t h t he t er m in t he square bracket s of equat i on (AS)
made zero. The coefficient correspondi ng to 0. 4578 i n equat i on (A7) was in t hi s case 0" 1154.
A.3. The A p p r o x i ma t e Fl e x u r a l Mo d e . - - T h e mode h(~) of Tabl e 1 is not t he t rue mode of
t he wi ng di st ort i on in flexure, but is a mode of t he angul ar di st ort i on i n flexure. The act ual
mode of vert i cal deflection is obt ai ned by i nt egr at i on in Tabl e 4.
TABLE 3
General d a t a a 1 = 5 . 5 , a 2 = 3 . 3 7 , m = 0 . 5 7 , fi = 4 0 , s / c o = 2 . 2 8 6 , y = 0 . 6 , z = 0 . 7 6 4
1 2 3 4 5 6 7 8
Summa t i on f 1 (1 - z'~]) f l
.d~ . d~71 St r i p d~ ~ h(r]) f or A~ , . . . . ~
" " " &7 1 " " "
0 . 2
0 . 2
0 . 2
0. 133
0. 133
0- 133
0. 1
0 . 3
0- 5
0. 666
0 . 8
0. 933
0. 0491
0. 2265
0 . 6 t 2 3
0. 8401
1. 000
1. 072
0"0079
0"0912
0"3294
0- 3182
0"3602
0"3326
Az = 1 " 4 3 9 5
0. 03332
0- 03332
0. 03332
0. 03332
0. 01639
' 0 . 0 0 5 7 8 5
- - 0 " 0 4 1 8 4
- - 0" 1612
- - 0 " 3 4 9 4
0. 1604
0. 08882
0. 05296 0. 007043
~ = 0 . 0 0 7 0 4 3
9 10 11 12 13 14 15 16
..@1 . . . @1 . . . @1 . . @~ ...~@1 (6)+s/c0(13) . . . @ h(~)
'~1 0
0- 01190
0. 01414
Z = 0 . 0 2 6 0 2
0. 02663
0. 02665
0. 02293
Z = 0 . 0 7 6 2 1
- - 0 . 0 6 9 8 8
0. 05872
0. 04441
0. 03352
Z ' = 0 . 0 6 6 7 8
- - 0 . 0 3 2 2 4
- - 0. 13975
0-09081
0. 06218
0. 04411
Z = 0 - 0 2 5 1 1
0-02511
O' 06678
O. 07621
O- 02602
O. 007043
0 ' 0
0. 09072
0. 1860
0. 2075
0. 09279
0. 03249
0. 005785
0. 02304
0. 1043
0. 2801
0. 3842
0- 4578
0. 4902
0. 0503
0- 2277
0. 6119
0. 8394
1. 000
1. 071
* Il l t he i t er at i ons t he pi t chi ng=moment l oads are c onc e nt r a t e d at t he cent r es of t he st ri ps, b u t i n pr act i ce, of course,
t h e y onl y become fully oper at i ve at t he i nboar d ends of t he st ri ps. Thi s modi f i cat i on has been ma de i n dr awi ng t he
second cur ve of Fi g. 16.
14
TABLE 4
1 2 3 4 5
t
St ri p h(~) fl (h~7)d'7 F(~) 1"6472"3
0"0503
0. 2277
0"6119
0-8394
1-000
1-071
0. 0024
0"0280
0"1t 20
0. 2340
0"3570
0-4930
0. 0067
0. 0784
0. 3137
0. 6555
1. 000
1. 380
0. 0082
0. 0877
0-3330
0. 6439
0. 9816
1. 398
The first four col umns of Tabl e 4 are sel f-expl anat ory, and in col umn (4) t he t rue ver t i cal
mode is given. The i nt egr at i on of col umn (3) was performed graphi cal l y. An appr oxi mat e
mode of t he form
has been deri ved by t he met hod of l east squares (see for exampl e Ref. 13) and is gi ven i n col umn
5. I t ma y be compared wi t h t he t rue mode of col umn 4 ; t he two modes are also compared i n
Fig. 17.
I t must be emphasi sed t ha t t he work i n t hi s section refers onl y to a wi ng whi ch is t orsi onal l y
rigid, and t he precise effect of t he i nt r oduct i on of t he t orsi onal freedom on t he fl exural mode is
not known. However, t he wri t ers feel t hat t he resul t does at least i ndi cat e t ha t t he assumpt i on
of a parabol i c mode in flexure is reasonabl e if no bet t er i nf or mat i on is available. The equat i ons
for t he sol ut i on of t he reversal probl em by semi-rigid t heor y wi t h general modes are gi ven i n
Appendi x II.
A P P E N D I X I I
Semi-rigid Solution f or General .Modes
For si mpl i ci t y we assume t he fl exural axis to coincide wi t h t he quart er-chord.
reversal speed is gi ven by sol ut i on of t he equat i on*
mo I1 cos = fl
qCo2S
Ia - - co I~ sin fl cos
s
I s COS fl
0 , G cos fl
l~ ~- I5 t an fi I6 sec fl + Co I7 sin fl
' S
q C o S ~ l o 2
I 9 s i n fl , 1,0
w h e r e t h e t o r s i o n m o d e i s g i v e n b y
0 = Oof(~)
a n d t h e m o d e o f f l e x u r e b y
z = zoF (, )
Then t he
= 0
* The equat i on is devel oped f r om t he met hod out l i ned in t he mai n t ext .
15
where 00 and z0 are the values of the twist and vertical displacement respectively at the reference
section. The integrals I1 . . . . . . I10 are given by
I1 = m ~ dv I6 =
Y
I2 = m f dr] I 7 =
y
I a = ( a t - - a=) c f F dr] I s=
o
I ~ = m ' dr] I 9 =
; i
Z5 = al ~ F F ' dr] I l o =
7
Here F' represents d F / d r ] .
a2 F d~
(cy
m ~o F ' d ~ l
r ] f d~
gl r ] F ' dg]
( o )
a~ ~ dr]
In the expressions which include the (al -- a2) term it must be remembered t hat a2 is zero
except over t he aileron ; ~, denotes the inboard end of the aileron which is assumed to ext end
to the wing tip.
As a test of the modes assumed in the text, the calculated aerodynamic loading for those
modes has been applied to a wing designed with a complete torsion box and approxi mat el y
constant bendi ng stress under normal lift distribution. Thus for this idealised wing, all the
bending strength is effective in torsion ; the skin thickness has been varied linearly from root
to tip in such a manner as to satisfy the constant stress requirement. The deformation of this
wing under these loads is t hen shown in Fig. 18 and compared with the assumed algebraic modes.
The agreement achieved for the flexural mode is satisfactory, but t hat for the torsional mode
is not so good. This fact is exaggerated in t hat the idealised wing provides an extreme case
with high taper, and an unbroken torsion box over the very stiff root sections. In most practical
cases the mode should be much more linear, as experience has shown it to be for unswept wings 4.
Moreover, since the aerodynamic effect of quite large changes in torsional mode (linear to cubic)
have been shown to be small (see Ref. 2) tile overall error of the semi-rigid hypothesis should
be no greater for a swept wing t han for a similar unswept wing.
16
Co
FL EXURAL A X I ~
QUA RT g R CHORD
FK;. 1. Diagram of wing considered.
I
25
~,0
i
~ t.5
1.0
0-5
STANDARD C A S E
~ C = o. 75
~- = 0 " 0
A = G, O
~o
30
/ / 3 = 55
j / 3 : 4 AN.
A \ I N P~ACnc~ " x, \ \ %.
/ - - e:
k,
O 0-2 0. 4 O,G O.B I,O
# = A N G L E O F B W E E P B A C K
'T" = T A P E m R A T I O ( I - TIP C H O R D / R O O T C H O R D )
~_. = A N G L E ; B F _ T W E E N q g A R T E R C H O R D ~, F L E U R A L A ~ I 5
A = ASPECT RATIO ( TWO WING, S)
~' = POSITION OF I NBOARD END OF AILERON
E ~ AILER, ON CHORD/ WING CHORD
FIG. 2, Variation of aileron reversal speed with wing stiffness.
St andard wing.
3.o
2-0
i
+ , - 5
1.0
0.5
2"5
FIG. 3
" C = 0 ' 2 5 .
_( 3 = 0 ~ = 0"O
_/ 3 = 3n o A ," G.O
/ - / 3 : 3 5 ~ = O ' ~
~ - ) 3 4 0 " F __. o o ~ 5 [
/ /3 4 5
5 1 ~ N D A I E ~
2..-0
i+ 1 .5
}, 0
0"5
0 0 " 2 O. 4 O' G 0 , 8
- - ~ H o
3 = 0
! 1 . _ ~ / 3 = 3 0
-~ 4 0
4 - ~
I ' 0
I
9. 7= 0 . 5
~ = O0
A = G. O
~ = O. G
E. = 0 " ~ 5
1,2
' I ',
, 5 T A N D,~,R D
0
0
FIGS. 3 a n d 4.
FIG. 4
I
0,~ 0.4 0 " 6 o. s t-o
: - M0
Vari at i on of aileron reversal speed wi t h wing stiffness.
1,2
Effect of wi ng t aper.
18
3-0
~'0
I'0
0,C
0
FI G. S
I
I I i
ASPECT~'ATfO 4
~------# : o ~C : o , 7 ~
~
-- ~o ~ : o.o
-- 55--
4 0 ~ = 0 " ~
/ f l : 45 E : 0 " 2 5
I r
I STA
O. ~ 0"4 0 ~ 0 " 8 IO 1"2
_ _ ~ . M g
NDA~D
2.0
FIGS. 5 a n d 6.
t ' 0
0 , 0
0
I
/ # . q : o o ' 7 ~ 1
I ~0 o
~ i o I - - - - ? . I = ! .
FIG 6
0.a 0, 4 0'G 0"8 1"0 I,~
,.-M g
Va r i a t i o n of ai l er on r e ve r s a l s pe e d wi t h wi ng st i f f ness. Ef f ect of wi ng a s pe c t r a t i o.
19
L
t_~
= 3 0
/ ~ = 3 5 ~
~ 0 = 45
1
~_ = 0"5
= 0 ' 7 5 ]
~5. = O , O
A ,= G ' O 5 T A N
E = 0 " 2 , ~
2.0
I-o
FIG. 7
o I
0, O 0 " 2
2 , 0
1"0
O
O
0" 4
= 0 ~
~
f 3 = 3 5
/3 ~ 4 0
/ / 3 = , 5
0,G ~'0 1-2 O ' 8
~' = 0 " 7 .
"C o , 7 5 I
; = o . o J
A ~- o
FI G. 8
I
O ' 2
S T A N D A R D
0' 4 O*G. D' B 1-o I"2
~- M 0
Variation of aileron reversal speed with wing stiffness. Effect o5 aileron span. Fi c~s. 7 a n d S.
20
5"0
L ~
2.0
1.0
i I E : O . 2
'12 = 0 " 7 5
/ / 3 -- O
/ / ~ 30 g : o . o
. ~ / ~ : , , o ) : , . o
- ~ / 4o" : o . $
I
F I G . 9
I
O' 2
' 5TANDAI:'D
0
o 0.4 o.G ,Me 0.8 11.o
E = 0"3
0
~ : ' I ~ : 0"75
3O =
: 3 5 J_. = 0 - 0
~
: 4 0 A = ~ . o
#
1-o ' k ~ ~
o-o F I G DO
0 0.2 0"4 0"~, 0"8 1"0
~-M o
t.2
t 5TANr
I-2
FIGS. 9 and 10. Vari at i on of aileron reversal speed wi t h wi ng stiffness. Effect of aileron chord.
21
3 . 0
2 . 0
T
1.o
O ' 0 .
O
2. 0
1.o
0 " 0
0
FI Gs. 11 a n d 12.
= 0
..- 3 0
-- 3 5 "
/ = 4 0
; - ~ / 3 = 4 5 "
/
/
/ 1"
1"
f / /
I
& o o ~
" T = 0 . 7 5
A ~ G'O
~ = O'G
E = . O- 2 5
J
!
F I G . ml
0. ~ 0, 4 O' G 0"8
= O
,, - / / 3 4 5
f
S T A NDA RD
'I tO f . 2
1
F~ . o . o 4
' u = 0 . 7 5
A = 6 . 0
)~ = O, G
F. = 0 - 2 5
I
5 T A NDA RD
F O G . nz
0" 2 0 " 4 0 ' G 0 ' 8 1, 0
Va r i a t i on of ai l er on r e ve r s a l s pe e d wi t h wi ng st i f f ness.
1'2
Ef f ect of wi ng f l exur al axi s.
2 2
i.o
M
e
0.5
M O m e
. ~ = 45
~ r : 2. #
-/3=O
o o. 5 t . o
A . VARI ATI ON WITH T A P E R
(~'C = '1 - TI P CHORD/ ROOT CHORD)
M 0
B. VARIATION
1,0:
~ : 4"-~ o
f3 = 45" r : ~ ' 5
po45 r : 5
- 0 , 0 ~ 0'{3 0, 0~ 0 . 0 4 0. 0~
W I T H S T R U C T U R A L ~ WE E P
(~., : ANGLE BETWEEN AERODYNAMIC AND
r:LEX UP,#,L A~l~5)
'1,0
M o
0"5
C.
O
2
r - - - - - - - - . J . _ _ _ _ _ . _ . " - - " - - - - - - " - - - - - - - - .
4"
'3= 45 r : t.a5
~ : 4 5 r : z' 5
G= o
G : 45~ r= 5
VARI ATI ON
8 tO
- - ~ " A
WITH ASPECT RATIO
M G
I
I
/
0
O4 0"5 0-G
J3 : 4 5 r : I?--~
'VARIATION WI TH
AILERON
45 r : 2"5
- ~ ~ =45 r : 5
~3-- 0
i
(>7 008
AILERON SP/~N
..~PA N )
SPAN
I-0
M o
0.5
o
0"15 O.Z
3= 45 ~ r-- 1-25
~= 45 r = 2. 5
r 3 = 4 a r : 5
O'Z5 0"3 0-35
E. V A R I A T I O N W I T H AILERON C H O R D
(E= : AI L ERON E l 5 0 R D )
WING, CHOF~D
FIG. 13.
M I'C
/
o l o 2O
E VARI ATI ON
i . . . ~ ,. " ~ "
P= 5"0
3 0 4 0 5 0
W I T H .~WEEP B A C K
Var i at i on of wi ng t orsi onal stiffness par amet er wi t h wing and aileron geomet ri cal par amet er s.
9,3
l~m. 14
2-B
L~
- - 2 - 0
!G, ENCE / ~
#
0.5I
- O g - O d Y O O. 1
I
\
CONTROL
POSITIVE
CONTROL
REVERSED
I
DIVERGENCE FOR
UNSWEPT W~N~
0"2 0 '2~ 0"4
- - ~ M g
Reversal and divergence diagram for flexural axis at 45 per cent chord with 45 deg sweepback.
(The curves are shaded on the danger side.)
l O
D I V E R G E N C E . LI NE FOR S W E F ~
FORWARD WING- F L E X U R A L
AXIS AT 45 % CHORD
D I V E R G E N C E . . LIN~ FOP~
S W g P T F O R W A R D W li',J(~
FLgXURAI~ A~,CIS AT
C~UARTER C H O R b
r
IRF--VERSAL CURVE FOR STANDARD
WING 5WE, PT FORWARD TO 45
FzG. 15.
J DIVERGENCF., FOR
J UNSWEPT WING
L
-oo2 Ol ~ . 0. ~ 0"4
0.6 0"8
Stiffness diagram for a wing swept forward at 45 deg.
(The curves are shaded on the danger side.)
2 4
bO
O ]
o'18
o,16
0 o.14
0II
0.1l
0. 0
0 " 0 2
o o . 2 0 - 4 O - G
LIF,T COMPONENT
/
0 - 8 , 1 - 0
Fro. 16. Rel at i ve magni t udes of lifting and pi t chi ng
component s on t he bendi ng moment at a section ~1.
(Torsional stiffness infinite.)
1. 4
1"2
1,0
O.~
O-E
0-4
0. 2
O
FIG. 17.
TRUE OI 5~LACEMENT OF'THE
MIO-POIKIT,S OF" THESECTI ON5
S H O W N X
APPROXI MATE MOOF-.
rc~). I.o4 ~ a . s .
/
/
J
0,'~ 0,4- 0-.0 0 ' 8 I,O
Compari son of flexural mode wi t h appr oxi mat e
anal yt i cal mode.
(Torsional stiffness infinite.)
FI G. 18
~.o
1.5
bO
O.S
TORSION
ASSUMED LINEAR
/ /
J
J
0 0"~ 0"4
S
~ C w A L C U L A T m M O D E ~O R ,~ A L m E O
,N ~ O F A ~ E N D ,X a TAP~ "C- o . 7 s
!
0"~ 0"8 1, 0
,?..O
1.5
F"
90
0"5
0
FLEXUI ~E
/
/
A~3UHED P A R A B O L I C /
. p -
0- 2 0- 4 0"6 0"~
~-r2
CALCULAT{ MODE ~'OR IDEALISED
Wl~Ei {:)F APPENDIX T;[ TAPER eE= 0"75
I-(3
A c h e c k o n t h e s h a p e of t h e a s s u me d mo d e s f or a p a r t i c u l a r e x t r e me cas e.
Ap p e n d i x I I .
T h e i d e a l i s e d wi n g of
J4296 Wt.17/680 K9 4/54 D&Co. 34/263
26
P R I N T I S D I N G R E A T B R I T A I ' N
_~
No & Mo Noo 28t7
Publications of the
Aeronautical Research Council
ANNUAL TECHN]ICAL REPORTS OF THE AERONAUTI CAL RESEARCH COUNCI L
(BOUND VOLUMES)
1936 Vol. I. Aerodynamics General, Performance, Airscrews, Fl ut t er and Spinning. 4os. (4os. 9d.)
Vol. n . St abi l i t y and Control, Structures, Seaplanes, Engines, etc. 5os. (5os. ied.)
1937 Vol. I. Aerodynamics General, Performance, Airscrews, Fl ut t er and Spinning. 4os. (4os. led.)
Vol. I I . St abi l i t y and Control, Structures, Seaplanes, Engines, etc. 6os. (6is.)
1938 Vol. I. Aerodynamics General, Performance, Airscrews. 5os. (5IS.)
Vol. II. St abi l i t y and Control, Fl ut t er, Structures, Seaplanes, Wi nd Tunnels, Materials. 3os. (3os. 9d.)
1939 Vol. I. Aerodynamics General, Performance, Airscrews, Engines. 5os. (5os. i i d. )
Vol. 1I. St abi l i t y and Control, Fl ut t er and Vibration, Inst rument s, Structures, Seaplanes, etc.
63 s. (64s. 2d.)
194o Aero and Hydrodynami cs, Aerofoits, Airscrews, Engines, Fl ut t er, Icing, St abi l i t y and Control,
Structures, and a miscellaneous section. 5os. (5IS.)
I941 Aero and Hydrodynami cs, Aerofoils, Airscrews, Engines, Fl ut t er, St abi l i t y and Control, Structures.
63 s. (64 s. 2d.)
I942 Vol. I. Aero and Hydrodynami cs, Aerofoils, Airscrews, Engines. 75 s. (76s. 3d.)
Vol. I1. Noise, Parachutes, St abi l i t y and Control, Structures, Vibration, Wi nd Tunnels. 47 s. 6d.
(48s. 5 d . )
I943 Vol. I. ( I n thepress. )
Vol. II. ( I g the press.)
ANNUAL PdlgFORTS OF 77-1E AERONAD-TECAL RESEARCH COUNC~L- -
I933-34 is. 6d. (IS. 8d.) I937 2s. (2s. 2d.)
1934-35 xs. 6d. (IS. 8d.) 1938 IS. 6d. (IS. 8d.)
April I, 1935 to Dec. 3 I, 1936. 4 s. (4 s. 4 d.) 1939-48 3 s. (3 s. 2d.)
]INDEX TO ALL REPORTS AND MEMORANDA ]PUIi~LIISHND tIN THE ANNUAL TECHNI CAL
REPORTS~ AND SEPARATELY- -
April, I95o . . . . R. & M. No. 16oo. 2s. 6d. (~s. 7d.)
AUTHOR ]INDEX TO ALL REPORTS AND MEMORANDA OF THE AERONAUT]ICAL RESEARCH
COUNCNL--
19o9-1949 . . . . . R. & M. No. 2570. I5 s. (ISS. 3d.)
]INDEXES TO THE TECHNI CAL REPORTS OF THE AERONAUTECAL RESEARCH COUNCI L- -
December I, 1936 - - June 30, 1939.
J ul y i , 1939 - - June 30, 1945.
J ul y I, 1945 - - June 30, 1946.
J ul y I, 1 9 4 6 - December 31, 1946.
J anuar y I, 1947 - - June 30, 1947.
Iul y, 1 9 5 1 . . . .
R. & M. No. 185o.
R. & M. No. 195o.
R. & M. No. 2050.
R. & M. No. 215o.
R. & M. No. 2250.
R. & M. No. 235 O.
is. 3 d. (IS. 4d.)
i s . ( i s . i ~ . )
i s . ( i s . i { d . )
IS. 3 d. (IS. 4d.)
IS. 3 d. (IS. 4d.)
IS. 9 d. (IS. Iod.)
Prices i,~ brackets include postage.
Obtainable from
HER MAJ ES TY' S S TATI ONERY OF F I CE
Yo r k Hous e , Ki n g s wa y , Lo n d o n W. C. 2 ; 423 Ox f o r d St r eet , Lo n d o n W. I ( Pos t Or de r s -"P. O. Box
No . 569, Lo n d o n S. E. 1) ; 13h Cas t l e St r eet , Ed i n b u r g h 2 ; 39 ICi ng St r eet , Ma nc he s t e r 2 ; 2 E d mu n d
St r eet , Bi r mi n g h a m 3 ; 1 St. An d r e w' s Cr es cent , Ca r d i f f ; To we r La ne , Br i s t ol I ; 80 Cki a he s t e r
St reet , Bel f as t O R T H R O U G H A N Y B O O K S E L L E R
S.O. Code No. 23-2817
Ro & Mo Neo 28~17