Vlasov Computation of Prismatic SHells
Vlasov Computation of Prismatic SHells
Vlasov Computation of Prismatic SHells
R=20050243020 2018-01-26T10:15:07+00:00Z
NATIONALADVISORY COMMITTEE
FOR AERONAUTICS
Washington
\ June 1949
~llllllllllll
31176014415781
1. Fundamental Assumptims
where the functians Ui(z) and Vk(z) depending cmly on z are the
required functions and the functicns ~(s) and ~k(s) me subJect to
a preklmimry choice. The magnitudes m and n will be explalned
,,
below.
n.~—c (1.3)
By choosing in so.?ne
manner n independent magnitudes Vk(z) for
the displacements of en elementary rod system .inthe plene of the
section z = constant and giving in succession to each of tlnesemagnitudes
unit values snd assdn~ the others equal to zero we csn, by considering
the elementary displacements thus obtained of this system, detemnine all
the required functions v~(s ). Each of these functions will’’rbpres~rita
contour displacement of a point s in correspondence with the elementary
state Vk* =1 !311dVh*= o for h+k (fig: 3). The function $k(s)
within the limits of each straight se@nent of the contour of the shell
maintains a constant value (does not depend on s) end represents an
axial displacement of the corresponding hinged Tod model. We shell ,
4 NACA TM 1234
After choosing the functions Qi(s) and ~k(s) in the sums (1.1)
:>uI?I
(1.2) the problem reduces to the determination of the functions Uf(z)
ml Vk(Z) (i = 1,2,3, .... m, k = 1,2,3, .... n).
.=++,
‘=4$+9 (2.1)
.
H,~nf~e
making Imseof equations (1.1) and (1.2) we have
(2.3)
where E and G are the moduli of elasticity of the material of the shell
in tension.and shear.
(2.4,2.5)
where the thickness of the shell 5 = 5(s) is assumed a given (in the
aeneral case not continuous) function of a sinale coordinate s and the
&gnitudes p = p(z,s), and” q = q(z,s) repre~ent the given external
surface forces.
(J = 1,2,3,..., m) (2.6)
(h = 1,2,3,..., n) (2.7)
The integrals above and those given in the following are definite and
are taken over the entire contour of the section z = constant.
The middle term e~resses the work of the internal sheartig forces.
For an element ds this work is determined as the product (with reversed
sign) of “theshearing force T5 de by the shear deformation which in the
case considered of the variation of the defomned state is equal to the
derivative Pj%) of the function pi(s).
Iq(s) d=
by the angle of rotation — J of two adjacent sections bounding this
element. The letters Mk=&s) and ~=~l(s) denote the bending
HA(2ATM 1234 7
53
J =— (2.10)
12
(
i,il= 1,2,..., m
h,k= 1,2,..., n )
xi
Chiuit +
xk r~Vk” - y~s~vk+~qh’o
k
1
(2.11)
E
Y =- (2.12)
G.
—
8 NACA TM X234
where the integral-sare taken over all the elements of the cross+ection
of the shell. These coefficients possess the properties of s?ymmetry
NACA TM 1234 9
(2.15)
~h’&& the centour integrals are taken over the parts of the contour for
wljichthe expressions under the integrsl sigu are different from zero.
...
The magnitudes pj(z ) and q+) corresponding to their physical
n&ning srieing from the method of their definition cen be termed the
unit (referring to unit length of the shell) generalized external forces.
The longitudinal one, p~(z), is computed as the work of the external
longitudinal surface forces, p(z,s)j over the longitudinal displacements,
9$s), of the elementary (unit) state of defomnation of the lamina
for U3* . 1, while the transverse one, qh(z), iS computed as the work of
the etiernal surface contour forces q(z,s) obtained for the contour
displacements of the points of the elementary lsmina ~h(s), determined
bY the wit displacement Uh* = 1.
10 lUU2ATM 1234
(2.17)
For these conditions eqyatims (2.11) assume the fomn
(2.18)
—— . .—:
.
12 NACA TM 1234
independent kinematic conditions which csn be given for the end sections
of the she~ z = O, z = 2 (where 2 is the length of the shell in the
direction of the generator). The positia of alllthe points of any of
these sections in space for a shell treated in the section z = constant
as a discr6te system is detemined by m + n independent magnitudes of
which m magnitudes U,(z) detemn5ne the position of all potits of
juncture of the section’on a prismatic surface (along the generator of
the shell) and n magnitudes Vk(z) detemine the position of these
points in the plsne of the cross-ection. For a single end section of
the shell m + n arbitrsry magnitudes may thus be assigned. For the
two sections z = O and z = 2 which bound the given shell over its
length the number of independent ccmditions is equal to 2(m + n) which
corresponds to the number of arbitrary constants in the integrated
equations (2.11).
Let us now consider the problem of the equilibrium of the shell for
which the boundary conditions in the secticms Z =0 ~d Z. I me
given ti terms of stresses or, in the case of the mixed boundsry problem,
partly in terms of stresses snd partly in terms of displacements. By
definition of the functions Ui(Z) and Vk(z) from eqUa.tiOn5(2.~)
the stresses a = Cl(z,a) ~d T = T(Z,5) at any point of the cross—
sections z = constant will be found according to equations (2.2) and
(2.3) also with sn accuracy up to 2(m + n) arbitrery constants. The
stresses u and T at the section z . constant for chosen polygonal
functions pi(s) ~d ~k(s) csn be expressed in terms of m + n
independent generalized static magnitudes. These magnitudes we shall
introduce by generalizing the fundamental concepts of the elementary
theorem of the bending of be- as was done in our work on the general
theory of the thin-led elastic rods based on the law of sectorial
areas (reference 3).
Starting from the idea of the virtural work of the normal and
shearing forces 05 and T5 of the cross-ection z = constsnt on
each of the m + n possible displacements of the points of this section
in space we introduce the following magnitudes
P3(Z) =
JUcpjdF
.
(j = 1,2,..., m)
where the contour integrals are taken over the entire crom+sectional
element of the shell. The _tudes Pj, ~ represent generalized
longitudinal and transverse forces of the section z . constant.
Considering these magnitudes as the internal forces of the shell we
express’them in terms of the fundamental functions Ui ~d ~k. @
the
,, basis of equations (2.2), (2.3), and (3.1) we have
,,
Pj = E xajiui’
i
i,il= 1,2,..., m (3.2)
( h,k=l,2,..., n)
.~’G~chiui+x~vk’
i k
or in expended form
P (z) P2(Z) pm(Z)
U(z,s) = *M(S) + ~T2(s)+ ... + ~ !?m(s) (3.5)
,
N~A TM l,jh
,
and the straight lines joining the ends of this ~c with the center of
flexure (fig. 4) we obtain for the generalized forces Pi,..., P4, the
first four terms of the series (3.4), the followtag values (reference 3,
p. 48).
Of these conditions the first three coincide with those lmown from
the theory of the bending of beams and which detezmine the principal
central axes of the cross-section of the shell. The second group of
equations (3.8) refers to the sectorial geometric characteristics of
the cross-section of the Bhell and determines the sectorial origin ~
(the starting point for computing the sectorial area) and the center of
flemre of the shell. For the conditions (3.8) the bimoment B = B(z)
represents a generalization statically equivalent to the vanishing of
the longitudhal force.
(3.9)
The equatim (3.5) for the thti+willed open shell, keep only the
first four terms, assumes cm the basis of equatims (3.7) ~dY 3.9) the
form
(3.10)
cs~
B “ -vB+&=O (3.11)
JLrJ(Jj JhJu
M ~ Pm
N+~x+% ~y+ ‘5
aM~+Eq5+... +=q)m (3.12)
‘=F. ‘Y x
where the first three terms refer to the stresses distributed over the
section accordtig to the law of plane deformation and the correspond~ng
elementary theory of the bending of beams. The remaining generalized
longitudinal.forces pk(z), p5(z), .... Pm(z) for the contour of the shell
possessing in the longitudhal directim a number of degrees of freedom
greater than three will represent the internal “longitudinal” forces
havtng essentiauy the sane character as the forces N, ~, and My of a
beam, the only difference being that these forces, for a choice of an
orthogonal system of functions qi(s), each represent over the cross-
section a balsmced system of longitudinal forces and arise as a result of’
the deplanation of the section. Tn contrast to the longitudinal force N
16 NACA TM 1234
(3.13)
NACA TM X234 . 17
The above equations together tith equations (3.2) lead, for the
section z = Zo, to the equations
%.= 6’
2d12 IJ1 + ‘1$
(4.1)
,,
where dl end ~ are the widths of %he vetiiced.and horizontal plates,
Jl and J2, the vertical and horizontal mommlm of inertia of the
eleimntary frame.
(4.3)
M12 ds
Su.
J —.
J
(4.4)
513 t523
J1=— (4.5)
12 ‘ J2=75--
-,
20 NACA TM 1234
(4.6)
where
S11=B4=0 (4.8)
antieqyation (4.6) for the zero moment shell goes over into the
simpler equaticm
in which now
bllru – c~2
A2 . (4.10)
4a~r~
I
\‘ — —
—
22 NACA TM X234
(4.11)
The first and third conditions are kinematic and express the fact that
in the plene of action of the force P the deplsmation of the section
is equal to zero (section z = O by symmetry remains plane) ~d at
infinity the defamation of the contour becomes zero. The second and
fourth conditions are static snd refer to the generalized (in the sense
of virtual work) transverse force Q (the transverse bimoment) which
in the sectim z = O must be in equilibrium with the external transverse
load and at hfinity becomes zero.
Jn the first row of table 3 sre given the values of the relative
coordinate C = ~ (d is the width of a plate), in the second row are
given the.magnitudes proportional to the longitudinal displacements U1
of the right jo3nt of the “frame” of the cross-section; the third row
refers to the transverse displacements VI (deflection of the horizontal
plate), de~ez?niningthe deformation of the contour of the secticm of the
shell. from these displacements are determined the transverse bending
moments of the shell arising from the stiffness of the joint of the
plates (horizontal snd vertical). The maximum bending moments, according
to the graph showp in figure 9, arise at the joints of the transverse frame
and ace determined by the formula
in which the plus sign refers to the left aad the minus si~ to the
rl@t joint. LU the fo~fi.hZVY;Of the table me given the magnitudes
proportional to the longitudinal normal stresses IS= a({) referring
to the points of juncture of the cross+ection. The graph of these
stresses over the section z = constant agrees with the graph of the
depknation of the section, shown in figure 6. The last row refers to
the tangential stresses T = T(c) forming in the section z = Constmt
the flow of shearing forces s = Tb.
NACA TM 1234 23
where
where x(s), y(s) are the coordinates of an arbitrary point of the contour
determined by the distance s along the contour, h(s) is the length of
the perpendicular dropped from the center of the rectangle on tineline of
the contour; the primes denote derivatives with respect to the parameter s.
magnitudes U2(Z) and U3(Z) are the angles of rotation of the section
z = constant relative to the axes Oy, Ox, respectively, and the
magnitude U4 is the generalized d.eplsnationof the section. The
magnitudes Vk(Z) (k = 1,2,3,4) represent: VI(Z) – the angle of
rotation of the section relative to the axis Oz (angle of torsion),
VP(Z), V3(Z) – the deflections la the direction of the axes Ox
and Oy, and V4(Z) - the generalized deformation of the contour of the
section.
M= 12
(5.3)
dl @J1) + d2 (FJJ2)
NACA TM 1234
E
M= (.5.4)
I
dll513 + d2 823
96
v4+q4.o
dll(EJl) +~1(~2)
i
t
%= @z), the generalized external forces referring to unit length of
/i
).
! the shell and computed by the formulas
ill
~ P4=JP(Z+)’W(S)-
dq(zs~+)dsQ+=fq(zs)*@d
(5.6)
1
‘,
1’ ‘
:, ,
;, in which p(z,s) ,and q(z,s) are the loads referred to unit mea of the
II
/1. middle surface of the shell end acting in the direction of the generator
and contour line respectively.
/1
/1
Ii
II
26 NACA TM 1234
(5.8)
FZ + F2 ~,,-— ~ 1
—q =0
6 al%? + 2d~
(
1
4 --—
Gd12F2 ~2F1
)M+-q
2Gl?1~
A2 . 24 1+ 1
(5.12)
al (EJI) + d2 (W2) (q2F1 Gd12F2
I I )
B4 = 3U 1
d12~2(Fl +F2) dl/J1 + %IJ2
(5.14)
e= x ~
n
sin (al- l)fiz
n=l,2. .. 2
(5.15)
2q
= O (n= 1,2,3,.. .)
cx@pc(2n -1)
—.-—.... . . ,. . . .
..———.
NAJ2ATM1234 29
From these e uations and series (5.14) the required functicns u(z),
M(z), aud 9Tz) may be determined with any initially given degree of
accuracy in the case where the shell is acted upon by a vertical
uniformly distributed load q in the plane of any vertical plate
(fig.‘12). We may remark”that”the series (5.24) for equations (5.15)
possess very good convergence and for practical purposes may be
restricted to the two or three first terms of the expansion, equation (5.14).
II -
30 NACATM 1234
w Ult+EFd
38 ~l?__ 1096
p FU2+:
z 62
I/5
FdU2s+z 33 FdV2
2 11=– -—
3&. ,fim +J&
IT 81 T 2~3i
( )
NACA TM 1234 31
=0 (6.2)
U1=Alcos ~,..., 4
pl=;Plsh;
V3 = B3 Sti ~, (6.3)
we shalJ have
~ au
a= —= E(Ul*~ +U2tg2) (6.5)
az
—
32 NACA TM X234
NACA TM 1234 33
Let the shelJ with the cross-ection shown in figure 11 have the
dimensions
63
Jl = J2=Z (7.2)
Il.. — —
34 NACA TM 1234
shell with rigid contour, the additional stresses connected with the
deviation from the hypothesis of plane sections for longitudinal b~~anced
load will have a local character. This remark refers to shells of
closed profile. ti the case of shells end thin-walled rods of open
profile the stresses due to the bimoment, even in the absence of
deformation of tiiecmltour of the section, extend over a considerable
pert of the length of the shell and do not have the cheracter of “local”
stresses.
Multiply<onnected Shells
‘1%.e
general differential equations (2.11) for the coefficients of
these equations (2.13) snd free terms (2.15) refer to the problem of the
equilibrium of thin-walled spatial multiply-connected prismavic shell.
These equations are generalized also to the theory of the vibraticm of
such shells. For this it is necessary in equations (2.11) to understand
by the forces pi and ~ tnertia forces arising from the vibrations
of the shell for given forms, with an accuracy up to a certain number of .
parameters, of the longitudinal and transverse displacements.
78 al 75 a2v
q=-—— (8.1)
F=–——
E!’#’ g #
From the above formlas are dete-ed the surface inertia forces, that
is the forces referred to unit surface area. These forces act as
follows: ~ in the direction of the generator, and T in the direction
of the tangent to the contour line of the tangent to the contour line
of the cross-section. Setting, accordingto equations (1.1) and (1.2)
(8.2)
Substituting these values in (8.1) and computing the free terms of the
equaticms (2.1.1)by formulas (2.15) we obtain
y ~- a2
-& Jf(z,t)
pj(Z,t) = ‘;Lat P+) @ m
J
(8.3)
32
qh(z,t) = ‘;~ +@jt)
at
J $k(s)ij-(s) ~
32
Pj(z,t) = ‘Z ~“a~i -&i(z,t) (i = 1,2,..., m)
g at-
(8.4)
,
a2
qh(z,t) = ‘~~rhk _&j+,t) (k= 1,2,..., n)
at
Equations (2.11) in the case of the dynamic problem will.have the form
NACA TM 1234 37
38 NACA TM 1234
9. General Remarks
The method here developed can be used to obtain the solution of the
plane problem of the theory of elasticity for a rectangular regime Thus,
for ex&ple, for the case of the plane persXlel state of the rectangular
plate the b~hsmmnic problem leads to a system of symmetrically constructed
ordinary differential equations
referrtig to a narrow rectangular plate which over its entire lower edge
is rigidly fixed agdnst vertical and horizontal displacements, etc.
(9.2)
E%2(p
+ G@’ + G*C = O
ri?m
82 d2 Id d2 d
G@+m?$+ CR— -— —R 3_z_=o (9.4)
12(1-+) ( ‘s ds2 ‘Rds )[ ds’2 ~+ 1
.
40 NACA TM 1234
a(m) + b(Tb) _ ~ 1 b
——-
la’r
.— + ~“xf+ ~“yt + e?’~t = o
(9.5)
az as
>
E & G az
P.
(9.6)
Ja20
dT
udX+m+ —u =0
2 [1
dz
Translation by S. Reiss
Nationsl Advisory Committee
for Aeronautics
NACA TM 1234
I-
U1 U* ... % ‘2 ““”
7~)D2 - b21 Ya22D2 - b22 .0. 7~D2 -b% -C21D +22D ... -cmD
C21D C22D ... caD r21D2 - y,21 r22$ -7022 >.0 rmD2 - 7s~
. . . . . . .*. ..* ● ✎☛ ✎ ✎ ✎ ● ** ● ** ● 9*
TABLE 2
U1 .*. Um VI ..*
‘2 — I —
ya11D2 -bll %2
9.. .**
.0. . . . . . . ● ☛☛
● ** ● o* . . . ✎☛☛
. . . ..9
-b
— —
C21D C22D ■ ,. ● .*
. . . . . . . . . ● **
● . . ● . . ● *. .,.
(2.19)
I
.,
44 NACA TM X234
1 2 3 4 8 12 10 02
*==’=’’+==’====?
i.
I
Figure 1.
Figure 2.
J?igur,e
3.
IL —– —
NACA TM X234
$-- 9-
m 7
\
Figure 4.
1 l-r-q
? .—
l-- ‘—=
J
I
/
I
1’ /1
// ///1
Figure 6. Figure 7.
1111111111111
-T--
Y, (s)
.
Ili
NACATM 1234 47
P
.
T
1<
I
?/{ /////////////////// //
.—-d+
Figure 10.
L
z x
tY
.
..—
—.I
I
h)
Figure 12.
I
48 , I’iJWATM X234
d,a“
-7
4 dz
z
V,=l
n
I---
If -)
I
1’
[L--
I
I
i
/’
/ I
Figure 14.
dz
-z
Figure 15.
NACATM 1234
M
., -— .
Figure 16.
.—
Figure 17.
q
I
i- U’-or.-+-+
Figure 18.
I
I
NACA TM 1234
(d
Figure 19.
\ \
\
\ I
\\
\ . I
\\ J \
\
I
\
I
+ 8
Figure 20.
—.
NACATM 1234 51
-, .-. , .
---- —-
Figure21.
Figure22.
V2
p
/2 /4
R /0
—
Figure23.
.,- ,,
-.. :;’&:m
Illlllllllw’immmillllllllllllllll
31176014415781
i
:
.. . .
,,
,.
,,
~.
.’,,
,“
..
:, -
..
./’”