2 Rectangular plate with two opposite edges simply Supported, the third edge free and the fourth edge built-in or simply supported.. A18.14 Cambined Bending and Tension or Compression
Trang 1
TỶ ==:
The expressions for bending and twisting
moments are not so quickly convergent To
improve the solution another series solution
can be developed as follows:
b) Levy alternate Single series solution:
The method will be developed for uniform
load q, = const Levy suggested a solution in
the form:
20
W= 2 Ym (y) sin ⁄" - (16)
m=)
where Ym is a function of y only Each term
of the series satisfies the boundary conditions
a*w
W320, 35a 3 0 at xa It remains to
determine Yq so as to satisfy the remaining ^
ow =0aty=b, two boundary.conditions w = 0, =— ay*
A further simplification can bs made if
we take the solution in the form
where
(x* = 2ax* + a3x)
* 24D
is the deflection of a very long strip with
the long side in the x-direction loaded by 4
uniform load q, supported at the short sides
x = 0, X =a, and free at the two long sides,
Since (17b) gatisfies the differential
equation and the boundary conditions at x = 0,
xX =a, the problem is solved if we find the
Solution of:
with w, in the form of (16) and satisfying
together with w, of eq (17b) the boundary
conditions w =o, ahs Oatyst 2 (see
where for symmetry m=1,3,5
This equation can be satisfied for all values
sinh =" sin =
where m2 1,3,5 Substituting this
expression into the boundary conditions:
where dy = mnb/2a
From these equations we find:
Trang 2n°D mel,3,s m° 2 cos hay
The summation of the first series of terms
corresponds to the solution of the middle of
This series converges very rapidly Taking a
square plate, a/b = 1, we find from (21a):
We observe that only the first term of the
series in (2le)} need to be taken into con-
can represent w in double Pourler Series:
Vị -42/ Vodxay = ab
9
Let us mow examine the deflection of the plate
of Fig 6 with a concentrated load P at the
point with co-ordinates x =€ , y=n The
increment of strain energy due to the incre-~
ment of the deflection by:
OV, can Gert pa)” 6 Cun ~ - - ~ (240)
The increment of the work of the load P is:
OW = P6Can, sin SỐ sin SL
From 6V, - OW = 0 we obtain:
Trang 3ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES
which is 3-1/2% less than the correct value
(2) Rectangular plate with two opposite
edges simply Supported, the third edge free
and the fourth edge built-in or simply
supported
y Fig 8 8
Assume that the edges x = 0 and x = 4 are
simply supported, the edge y = b free and the
edge y = 0 built-in (Fig 8) In such a case
the boundary conditions are:
we write the deflection in the form:
WW + Wa
where w, is the deflection of a simply supported strip of length, a, which for the system of axes of Fig 8 can be written (see Levy’s method in previous section):
oo
Wa > a Ym (vy) sin TT” -~ ~ - (260)
m=1,3,5 where w, being a solution of
ow, = aw, 3*wW axe * ays * ® axtay? 7 o
is found as in the previous section:
Yn (Am cos has + By == sink +
Cy sinn +p, BY cosh MY) - ~(264)
It is obvious that the two first boundary
conditions are identically satisfied by w =
W, + W, The coefficients Am, Bg, Cm, Dy must be determined sc as to satisfy the last
four boundary conditions Using the conditions
occurs at the middle of the free adgs
A18.14 Cambined Bending and Tension or Compression of Thin Plates,
In developing the differential equations
of equilibrium in previous pages, it was assumed that the plate is bent by transverse
loads normal to the plate and the deflections
were so small that the stretching of the middle Plane can be neglected If we consider now the
case where only edge loads are active coplanar
Trang 4
A18, 1ậ THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS
with the middle surface (Fig 9} we nave a
plane stress problem If we assume that the
Stresses are uniformly distributed over the
thickness and denote by Ny, Nyy Nyy, Nyy the |
resultant force of these streSses ber Unit
length of linear element in the x and
y-directions (Fig 9) it 1s obvious that:
Ny = hoy , Ny * hay
Nxy = Nyx = hơxy = hoy,
Fig 9
The equations of equilibrium tn the absence
of body forces can be written now in terms of
these generalized stresses Ny, Ny, Nyy by
substituting from eq (27) to the equations
of equilibriun
aNx Nyy BNy êNxy
s+
ax * ay S9; ay ax
On the other hand tf the plate is loaded by
transverse loads the stresses give rise to
pure bending and twisting moments only The
equations of equilibrium for the latter have
been given in before (see eqs 6, 7, 8) If
both transverse loads and coplanar edge loads
are acting simultaneously, then for small
vertical deflections the state of stress is
the superposition of the stresses due to
Ny, Ny, N and My, My, Myy For large
were đễrlactien oF the plate, however,
there is interaction of the coplanar stresses
and the deflections These Stresses give rise
to additional bending moments due to the non-
zero lever arm of the edge loads from the
deflected middle surface, as in the case of
beams When the edge loads are compressive
this additional moments might cause instability
and failure of the plate due to excessive
vertical deflections
In this chapter the problem of instability
of plates will be examined
When the edge loads are compressive and give rise to additional bending moments sq (8)
Fig 10a
Consider an element of the middle surface
dx, dy (Fig, 104) The conditions of forse equilibrium in the x, y-directions are given
by eq (28) Consider now the projection of
the stresses Ny, Ny, Nyy in the z-direction:
(a) Projection of Ny: From Fig loa it follows that the resultant projection is:
hy ay
and neglecting terms of higher order:
aw aNy aw,
Oe BF + ax ex) TTT TTT (4)
(b) Projection of Ny: By similar argument
we find that this projection 1s equal to:
aw SS aNy aw ee ee ee Le
Fig 10b {c) Projection of Nxy and Nyy: From Fig lob
we find:
ew ON ow 37x
~ Nxy Ft (xy + TS ax) Sy Sxay 4x)
and neglecting terms of higher order:
ay Gxay * “gx: Gp) duty - - - fe)
Simtlarly we find the projection of
Trang 5ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES
ô3w ,„ôNxy ôw,
(xy Sxoy cây ox? axdy
Thus in eq (6a) the terms given by (a), (b},
(¢) and (d) should be added (divided of course
by dxdy):
+ ye ge Ny Se py SH oy
ay 9 * Nx xt * By ays XY axay
aN, Ay, BW, Ny, Way aH gL
Se Sy! ae Sy * Sea TO te)
But due to the equation of equilibrium (28) the
two terms inSide the parentheses in (e) are
zero Thus:
Se = _
Ny or Ta + #Ñxy quay Say?
Eq (29) replaces eq (6a) when edge loads are
present fqs (6b) and (6c) are, however, still
valid since they exnress moment equilibrium
of the element dxay in wnicn tne contriouvion
of Ny, Ny, Nyy is zero Thus eliminating Q,,
Qy between (6b), (6c) and (29) we find:
Viw = ae * 3y? S5 ax ay
T (q+ Ny Set ly Sự + 2Nxy Ea - - - -(30)
Eq (30) replaces eq (128) when edge loads
are present and the deflections are large so
that instability might occur
The distribution of the coplanar stresses
Nx; Ny, Nyy can be found from eqs (28) by
solving the plane stress problem In the
following the above theory will be applied to
rectangular plates
A18.15 Strain Energy of Plates Due to Edge
Compression and Bending
The energy expression for pure bending,
eq (11), must be complemented to include the
contribution of the edge coplanar loads
Assume that first the edge loads are applied
Obviously the strains due to the stresses
Nx, Ny, Nyy are:
ix -YNy), ey > Te (Ny -UNy),
Á18 19
Vin sass (Nxex + Nyey +Nxy/Ƒ xy) axdy =
shelf (Md + Ny ~ 2nE NY ROY 8UNNy +2(1/)0Nvy`) TOIT '¬ = (2)
During bending due to transverse loads or/and due to buckling we assume that the edge loads and consequently Ny, Ny, Nyy, remain constant Its variation 1s thus zero and we
do not consider it in the following Let us
apply now the transverse load that produces
bending (We can also consider bending due to other transverse disturbance, which is the case of buckling) If u, v, are the displace- ments of the middle surface due to the coplanar loads (which are assumed constant across the thickness) and w the bending deflection of the plate it can be shown that the strains are:
surface the energy is:
Jf (Nxex + Nyey + Nxy Syy) dxdy = - - - (c)
Introducing (b) into (c) and adding the strain
energy due to bending, eq (11), we find the
total change of strain energy due to bending which ts:
2(1-v) [= a ee} dxdy - - (4) Here u, v are the additional coplanar displace-
ments after bending has started It can be
shown by integrating by parts that the first
integral is the work done during bending by the edge loads For instance taking a rectangular
plate this integral becomes:
Trang 6
A18 20
Obviously the first two integrals represent
the work done by the edge loads while the
second integral is zero due to the equilibrium
equations (28) Thus the work of the edge
plate is negligible (This is the so-called
inextensional theory of plates) In this case
by zeroing the strains in eq (b) and substi-
tuting in (f) we find:
aw aw
Nụ ==È // [Xx Bat ery Bt + amy & Blanay
In the strain energy expression, eq (a) the first two terms cancel each other and the
strain energy 1s due only to bending:
„1 320, o2Wy a
Vị =2D0// lt-šÐ -
2a-) (oe oe cả} đxếy - - (32h) xẻy
In the absence of transverse loads the work
of external forces is simply due to the edge
leads:
Wom Wy
Expressions (32b) and (c) will be used in
Solving the buckling problem by means of the
principle of virtual work
Al8.16 Buckling of Rectangular Plates with Various Edge
Loads and Support Conditions
General discussion
In calculating critical values of edge
loads for which the flat form of equilibrium
becomes unstable and the plate begins to
buckle, the same methods and corresponding
Treasonings as for compressed bars will be
employed
The critical values can be obtained by assuming that the plate has a slight initial
curvature or a small transverse load These
values of the edge loads for which the lateral
deflection w becomes infinite are the critical
values (see Part 1 for similar treatment in
columns)
Another way of investigating such
instability is to assume that the plate buckles
due.to a certain external disturbance and then
to calculate the edge loads for which such a
THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS
buckled configuration (deflections different from zero) is possible It was found in the
case of column that this latter solution
(Euler’s solution) approaches asymptotically
the first at the limit where the deflections
become extremely large but for even small
deflections the edge load acquires a value
very near the Zuler’s critical value The
latter technique is mathematically more
convenient and it gives for plates also a very
good estimate of their compressive strength
In the following we shall use this latter
approach by assuming a plate with edge loads
and no transverse load zq (30) becomes in
this case:
atw atw 1w lạy SỔN aẦw aw y
ax 2 Otay? ay? — 0x y1! Ny ay? * Bhy Bay?
By solving eq (33) we will find that the
assumed buckling made is possible (w # 0) for
certain definite values of the edge loads, the
smallest of which determines the critical icad
The energy method can also be used in investi-~
gating buckling problems In this method we assume that the plate is initially under the plane stress conditions due to the edge loads
and the stress distribution is assumed as known
We then consider the buckled state as a possible
configuration of equilibrium The change of the work i8 given by eq (22a) We interpret here was @ Virtual displacement though we do not use the variation symbol 6 Thus the increment of work 6W is given by (32a) and the increment of strain energy 5Vi is given by eq (32b) If Gw<5V; for every possible shape of buckling the flat equilibrium is stable If 6W=éVy for a certain shape of buckling then the flat con- figuration is unstable and the plate will buckle under any load above the critical load If
OW = 6V;, the equilibrium is neutral and fron
this equation we find the critical load The
critical load therefore 1s found from the equation:
Here w ts a certain assumed deflection which
satisfies the boundary conditions (virtual
deflection)
A18,17 Buckling of Simply Supported Rectangular Plates
Uniformly Compressed in One Direction
Let a plate of sides a and > (Fig 11)
simply supported around its periphery be com- pressed by load Ny uniformly distributed along the sides x = 0 and x =a From the
Trang 7
1 \ F— x This fraction has some intermediate value
= between the maximum and the minimum of the
fractions (a) It follows that if we wish to
make the fraction (b), which is similar to the
Fig 11 fraction of eq 35d, 4 minimum, we must take
ty 8 only one term in the numerator and the corres- obvious solution of the corresponding plane
stress problem we find that the state of
stress 1s everywhere a simple compression
equal to Ny (the load at the periphery)
The deflection surface of a simply
supported plate when bending takes place have
been found previously (see eq 14a) Its
general expression can be written in a double
series form:
co of
Wa Š 3 mg sin TC sin TC - - - (88a)
me} n=l
The increment of strain enargy found by
substituting (35a) in the right-hand side of
The increment of work done by the external
forces is found by substituting (35a) into the
left-hand side of eq (34) and Ny = const,
Here Cmn 18 arbitrary We are interested,
however, to find that values of Cyy which make
Ny minimum To that effect we use the follow-
ing mathematical reaconing:
Imagine a series of fractions:
If we add the numerators and the denominators
we obtain the fraction:
ponding term in the denominator Thus to make
the fraction of eq (35d) minimm, we must put
all the parameters C,,, Cis, Cara except one, zero This is equivalent to assuming that the buckling configuration is of simple
sinusoidal form in both Gatecttons, 1.9
Won = Can sin, sin SY ,
expression for Cyn obtained by dropping all the terms except Cy becomes:
The minimum
71%4°D m2 | n#
Ny = 42e ta)?
It is obvious that the smallest value of Nx is
obtained by taking n=1 This means that the plate buckles always in such a way that there
can be several half-waves in the direction of compression but only are half-wave in the
perpendicular direction Thus for n= 1, eq
(35e) becomes:
Wy) op 5 = m+ = op)
The value of m (in other words the number of
half-waves) which makes this critical value
the smallest possible depends on the ratio
a/b and can be found as follows:
Let us express (36a) in the form:
(N_) 2p ed
x/er b2 TTT “TT (46p)
where k 1S a numerical factor depending on bi
From (36a) and (36b) we nave:
a
«225 (a+: oe pF" ae eee eee (36c) a2
If we plot k against 2 for various values of the integer m=1, 2, 3, we obtain the curves
of Fig (12) From these curves the critical
load factor k and the corresponding number of
half-waves can readily be determined It is
only necessary to take the corresponding point
es as the axis of abcissas and to choose that curve which gives the smallest k In Fig 12 the portion of-the various curves which give the critical values of k are shown by full
lines The transition trom m to (m + 1) halr- waves occurs at the intersection of the two
(36) we find:
corresponding lines From eq
Trang 8Thus the transition from one to two hal?-waves a 2 ar ¬
22/1 = Ve
from two to three for
¬ teens and so on
Tne number of half-waves increases with the
ratio a/b and for very long plates m 1s very
large
zon
This means that 4 very long plate buckles in
half-waves the lengths of which approach the
width of tha plate The buckled plate is
subdivided into squares
The critical value of the compression stress is:
= (xJor „ K"*E t®
yy = ker = 15
h 12(1-y”) b#
{t = thickness)
Á18.18 Buckling of Simply Supported Rectangular Plate
Compressed in Two Perpendicular Directions
Lat (Fig 15) Ny, Ny the uniformly distri- buted edge compressions Using the same as
before expression for the deflections (eq 35a)
and applying the energy equation (33) with
Nx, Ny = constants (which ts the solution of
the corresponding plane stress problem) we find:
Taking any integer m and n the corresponding
deflection surface of the buckled dlate is given
Dy:
stn SRY, TƯỚC
a » Sin 5
and the corresponding ox, Oy are given dy (37c)
wnich is a straight line in the diagram By, Oy
(Fig 14) By plotting such lines for various
pairs of mand n we find the region of stability
and the critical combination of ay, Oy which ts
on the periphery of the polygon rormed by the
full lines of Pte 14,
Fig 14 A18.19 Buckling of Simply Supported Rectangular Plate Under Combined Bending and Compression
Let us consider a simply supported
rectangular plate (Fig 15) Along the sides
xX = 0, X =a there are linearly distributed
edge loads given by the equation:
Trang 9ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES which is 4 combination of pure bending and
pure compression.- Let us take the deflection
azain in the form:
Substituting in the right-hand side of
eq (34) we find the variation of strain
energy:
abn+
while the increment of work is:
The coefficients now Cyn are so adjusted that
(Noler becomes minimum By taking the
derivative of expression (38e) with respect
to each coefficient Cyn and equating these to
We examine for each value of m the solutions of
the system (38f) Starting from m > 1 and
is of infinite number of equations (n= 1, 2, 3 ) A sufficient approximation is
obtained by taking a large but finite number
of terms and find the solution of the finite
determinant (using for example digital computers) Thus a curve of Ogr versus a/b 1s
obtained for m= 1 like that of Fig 12
Repeating the same calculation form =2,3 , atc., we find similar curves of two, three, etc
half-wave lengths The regions of the curves with minimum ordinates define the region of
stability as in Fig (12)
A18.20 Inelastic Buckling of Thin Sheets The problem of the inelastic buckling of
thin sheets has been extensively studied by
various authors The main difficulty in such
studies is in reference to the stress-strain relations of plasticity under complex states
ef stress Many controversial discusstons have appeared in literature without resolving
the theoretical difficulties For this reason
we Will not develop the theory of inelastic buckling in this chapter Some of the better references on this subject are listed below
Chapter C4 presents the plasticity
correction factors to use in calculating the inelastic buckling strength of thin sheets
A18,21 References, (1) Bleich, F: Buckling Strength of Metal Structures Book by McGraw-Hill
(2) Stowell, E.Z., A Unified Theory of Plastic
Buckling of Columns and Plates NACA Report 998, 1948
(3) Gerard and Becker: Handbook of Structurel Stability NACA T.N 3781, 1957
Trang 10
A18 24
a
Courtesy of The Boeing Company, Seattle, Washington
This multiple exposure photograph of a Boeing supersonic transport mode! shows the variable-sweep wing
in three configurations: forward for takeoff and landing, swept part way back for transonic flight, and swept completely back as an arrow wing for 1800-mile-an-hour supersonic cruise
SPECIFICATIONS (Basic Design) Gross Weight
Payload Range Takeotf Distance (Max gross weight) Landing Distance
Cruising Speed
Takeoff Speed
Approach Speed Wing Span (forward position)
Trang 11
CHAPTER Al9 INTRODUCTION TO WING STRESS ANALYSIS
A19.1 Typical Wing Structural Arrangement
For aerodynamic reasons, the wing cross-
section must nave a streamlined shape commonly
referred to as an airfoll section The aero-
dynamic forces in flight change in magnitude,
direction and location Likewise in the various
landing operations the loads change in magni-
tude, direction and location, thus the required
structure must be one that can efficiently
resist loads causing combined tension, com-
pression, bending and torsion To provide
torsional resistance, a portion of the airfoil
surface can be covered with a metal skin and
then adding one or more internal metal webs to
produce a single closed cell or a multiple cell
wing cross-section The external skin surface
Which is relatively thin for subsonic aircraft
is efficient for resisting torsional shear
stresses and tenston, but quite inefficient in
resisting compressive stresses due to bending
of wing To provide strength afflciency, scan-
wise scifgening units commonly referred to as
flange stringers are attached to the inside of
the surface skin To nold the skin surface to
airfoil shape and to provide a nedium for
trans?erring surface air pressures to the
cellular deam structure, chordwise formers and
ribs are added To transfer large concentrated
loads into the cellular beam structure, heavy
ribs, commonly referred to as bulkheads, are
used
Figs AlS.1 and Als.2 illustrate typical
structural arrangements of wing cross-sections
for subsonic aircraft The surface skin is
In general the wing structural
arrangement can be classifled into two (1) the concentrated Zlange type where material is connected directly to in- ternal weds and (2) the distributed f @ cype
where str rs are attached to sxin bet
internal webs
Fig AS.3 shows several structural 2 xa~
ments for wing cross-sections for superscnie aireraft Supersonic airfoil shapes are relatively thin compared to subsonic aircraft
Fig Al9.2 Common Types of W
Arranger Beam Flange
Trang 12
Al8.2
To withstand the high surface pressu:
obtain sufficient strength much
skins are usually necessary Modern milling
machines permit tapering of skin thicknesses,
To obtain more flange material integral flange
units are machineẻ on the thick sxin as tllus-
trated in Fig k
s and to Ker wing
In a cantilever wing, the wing bending
moments decrease rapidly spanwise frem the
maximum values at the fuselage support points
Thus thick skin construction must be rapidly
tapered to thin skin for welgnt efficiency, but
thinner skin decreases allowable compressive
stresses To promote better efficiency sand-
wich construction can be used in outer portion
of wing (Fig 1) A light wetght sandwich core
is glued to thin skin and thus the thin skin ts
capable of resisting high compressive stresses
since the core prevents sheet from buckling
A19.2 Some Factors Which Influence Wing Structural
Arrangements
(1) Light Weight: +
The structural designer always strives for
the minimum weight which 1s practical from.a
production and cost standpoint The hisher the
ultimate allowable stresses, the lighter the
structures The concentrated flange type of
wing structures as illustrated Fig (a, > and ¢}
of Fig Al9.1 permits high allowable compressive
flange stresses since the flange members are
Stabilized by both web and covering = , thus
eliminating column action, which permits design
stresses approaching the crippling stress of the
flange members Since the flange members are
few in number, the size or thickness required is
relatively large, thus giving a high crippling
stress On the other nand, this type of design
does not develop the effectiveness of the metal
covering on the compressive side, which must be
balanced against the saving in the weight of the
flange members
In the distributed type of 7 lange arrange-
ANALYSIS OF WING STRUCTURES
flange allcwatle compressive str
is not practical to space wi
12 to 18 times the Tlange stringer 7
17 there were no othsr controllin:
could easily make calculations to
which of the above would orovs general, if the torsional °0rees are small, thus requiring only 4 the concentrated
should prove tha
In general, placed to give the m the Z direction, » in general that
fla material should ? be placed ˆetuean the
and 50 per cent o ing chord from the lead
edge
be inert ta in
th
1
Tue secondary or distriouting str
f the structural box bea should be mad
light as possible and thus in general =:
forward
lighter
the rear closing web of the box
the wing structure as a whole
In the layout of the main spanwise ?lange
members tends or changes in direction should be
avoided as added weight is required in splicing
or in transverse stiffeners which are necessary
cO change the direction of the load in the flan:
members If flange members must de spliced, c2
should be taxen not to splice them in the region
of a maximum cross-section, Furthermore, in
general, the smaller the number of fittings,
nter the structure
igh wing type, the entire
continue in the way of the airplane pedy How-
ever, in the mid-wing type or semi-iow wing type, limitations may prevent extending the entire wing through the fuselage, and some of
the shear webs as well as the wing cover
must be terminated at the side of she fuselage
If a distributed flange type of cell structure were used, the axial load in the clange string-
ers would nave to oe transferred to the members
extending through the fuselage + provide for
this transfer of 2a: loads requires structural weight and thus a concentrated flange type of
box structure might prove the best type of
structure
the low wing or the
wing str acture can