ht. J. Solids Srrucfures, 1970. Vol. 6, pp. 1463 to 1481. Pergamon Press.

Printed in Great Britain

BENDING, VIBRATION AND BUCKLING OF SIMPLY

SUPPORTED THICK ORTHOTROPIC RECTANGULAR
PLATES AND LAMINATES
S. SRINIVAS~
and A. K. RAO$
Department of Aeronautical Engineering, Indian Institute of Science, Bangalore 12, India

Abstract-This paper presents a unified exact analysis for the statics and dynamics of a class of thick laminates.
A three-dimensional, linear, small deformation theory of elasticity solution is developed for the bending, vibration
and buckling of simply supported thick orthotropic rectangular plates and laminates. All the nine elastic constants
of orthotropy are taken into account. The solution is formally exact and leads to simple infinite series for stresses
anddisplacemeutsinflexure,forcedvibrationand “beam-column” ty~problemsand toclosedformcharacteristic
equations for free vibration and buckling problems. For free vibration of plates, the present analysis yields a
triply infinite spectrum of frequencies instead of only one doubly infinite spectrum by thin plate theory or three
doubly infiuite spectra by Reissner-Mindlin type analyses. Some numerical results are presented for plates and
laminates. Comparison of results from thin plate, Reissner and Mindlin analyses with these yield some important
conclusions regarding the validity and effects of the assumptions made in the approximate theories.

NOTATION

a, b, h length, width and total thickness

E an elastic modulus used in defining R
12 b-aP
buckling stress parameter = ;;i I; F
0 x
L differential operator $
M,N mxhla, nlrhlb
Px7 py normal stresses on the edges x = 0, a and y = 0, b
P&rm critical buckling stress when plate is loaded on edges x = 0, a only
x, K z xi% y/b, r/h
z eigenvalue = Jm in vibration problems
= ,/(PxM2 + P,,N’)/E in problems with normal edge loads
mass density
; angular frequencies-forced or free
thin plate theory value
% error -1 xl00
i value by present exact analysis I
For laminated plates
Subscript j (j = 1,2,. . , p) denotes corresponding ply

\iG-l I I
number of plies
modular ratio between top and middle plies of a three-ply laminate (with identical top and bottom
plies and also identical relative moduli within each ply)

Figure 1 indicates the coordinate system and the dimensions

t Senior Research Fellow.
$ Professor.

1463
1464 S. SRINIVAS and A. K. RAO

Orthotropic stress-stram relations [I 11:

THIN plate theory is based on several assumptions, the most important of which are the
neglect of transverse shear deformations and rotatory inertia. The errors in such a theory
naturally increase as the plate thickness increases. Further, due to the neglect of transverse
shear deformations and c,, one cannot take into account all the 9 elastic flexibility con-

FIG. 1. Coordinate system and dimensions.

 Bending, vibration and buckling of simply supported thick orthotropic rectangular plates and laminates 146.5

stants of orthotropy. Consequently, the errors increase as the magnitudes of the transverse
elastic constants (F,, F,, , F,,, l/G,,, l/G,,) increase relative to the in-plane constants
(F,, I$, FXy, l/G,,,). So, it is natural to seek some improved formulations which account, at
least approximately, for transverse shear deformations and rotatory inertia. Reissner [II
and Mindlin [2] initiated such analyses by their modified plate theories for flexure and
vibration of isotropic thick plates. They start with the standard thin plate assumptions for
variation of stresses (or displacements) across thickness, but include the resulting transverse
shear deformations and rotatory inertia. Both these analyses permit satisfaction of three
boundary conditions on each edge, but they do not satisfy the governing differential
equations of three-dimensional elasticity exactly. Medwadowski [3] extended Reissner’s
theory to orthotropic plates. Yang et aE. [4] have developed a Mindlin-type analysis for
heterogeneous plates or laminates with general anisotropy and elastic moduli varying
continuously across each ply. They have also presented numerical results for plane waves.
Whitney [5] has provided another analysis for anisotropic laminates taking into account
transverse shear deformations; he does not presume linear variation of displacements u
and v across thickness. Several other papers relevant to anisotropic laminates are published
in Ref. [6].
The next obvious step is to seek an exact analysis by three-dimensional theory of
elasticity. This was recently achieved for certain plates and laminates of isotropic materials
[7-lo]. The current interest in thick laminates made up of orthotropic layers, indicates
that a three-dimensional analysis for the statics and dynamics of thick laminates made up
of general orthotropic plies would be useful. In this paper such an analysis is developed
for simply supported rectangular plates and laminates. Flexure, forced and free vibrations,
and a class of buckling and beam-column type of problems, are all treated together.
In the analysis herein, the boundary conditions for a simple support on a straight edge,
x = cons& are specified as,
w = 0, e, = 0 and ci, = 0 for all z (1)
The combination of edge conditions in equation (1) amounts to providing an edge support
ill~nitely rigid in its own plane (w = 0, z’= 0 on x = const.), but completely flexible to
out of plane stresses (Go = 0, for x = const.). In case of plates under normal edge loads,
equation (1) refers to the quantities arising out of deviation from the original state of uniform
strain.
For a homogeneous plate, the solution is set up in the form of a double trigonometric
series in Cartesian coordinates, [aide equation (4) to follow], for each of the displacements
u, v, w such that the governing differential equations and all the edge conditions are identi-
cally satisfied. The terms corresponding to each harmonic in the series contain six arbitrary
constants and these are explicitly obtained by solving six simultaneous equations, which
result from satisfying the appropriate lateral surface conditions.
The analysis for laminates is a direct extension of that for a homogeneous plate; each
ply is treated as an individual homogeneous plate and, at the interface between any pair
of plies, a set of six homogeneous conditions representing equilibrium andcontinuity must
be satisfied. In a p-plied laminate, there are 6p arbitrary constants corresponding to each
harmonic, i.e. each combination of m and n, and these are explicitly derived by satisfying
all the relevant lateral surface conditions.
k special feature of this analysis is that all homogeneous surface conditions (which
include the interface conditions in laminates) are satisfied identically.
1466 S. SRINIVAS and A. K. RAO

In the eigenvalue problems of free vibration or buckling, a closed form characteristic

equation (in determinant form, of order 6~) is obtained for each combination of WIand ti
Each such equation yields an infinite number of eigenvalues, each one of them representing
a different “thickness mode”. The first of these is generally referred to as a “flexural mode”
and it is only this mode which thin plate theory can identify and yield approximate resuits
for. In many buckling problems one ofthe flexural modes is the primary mode of instability.
In thin plate theory, only w(x, y) is given freedom. In Mindlin’s, 8t@z(x. _r)and &/?z(.x. ~1
and in Reissner’s, M&X, y) and M&x, y) are also given freedom. In the present exact analysis
u, a, w at every point are given freedom. CorrespondingIy for free vibration. thin plate
theory, Reissner-Mindlin analyses and the exact solution respectively yield one, three and
an infinite number doubly infinite sets of eigenvalues.
As the number of variables in orthotropic plates or laminates is large, detailed study ot
the effects of individual variables on the physical aspects of the problems, or on the errors
due to thin plate assumptions, are not undertaken ; but some broad conclusions are drawn
from the numerical results for an example of orthotropy. The properties of this material
are listed in Table 1. Some numerical results are also presented for three-ply laminates
with identical top and bottom plies; the relative values of the moduli are the same in all
the plies, i.e. (E, : Ey :E, : E,, : E,, : Ey, : G,, : G,, : G,z) are identical.

TABLE 1. ORTHOTROPIC PROPERTIESASSUMED IN EXAMPLES

E,/E, = 0.543103 EJE, = O-5301 72

E,,/E, = O-23319 E,,/E, = 0.010776
E,,/E, = OdI98276 G,,IE, = 0.262931
G,,/E, = 0.1599 14 G,,/E, = 0.26681

These properties correspond to Aragonite

crystals [ 1 I].

GOVERNING EQUATIONS OF THREE-DIMENSIONAL ELASTICITY

The equations of equilibrium in terms of displacements for a homogeneous orthotropic
plate can be written in Cartesian co-ordinates as

(2)
 Bending, vibration and buckling of simply supported thick orthotropic rectangular plates and laminates 1467

where
(0) is a (3 x 1) null matrix, and (f) is a (3 x 1) matrix.
Further,
I;; [jj = {O) with no body forces;,
= - p(ii B G)) for motion wtth no body forces other than inertia forces,
(-p{ ii ti 6) = pR*{u v w}, for simple harmonic oscillations with frequency Q;
(c) ( f ) = F(u v wf, for general applied stresses on edges ; F is a function of local
stresses in the initial state and second order deferential operators.
[F = -(I?, d2/8x2 + P, a”,@y”)for an initial state of uniform strains E,, ayand g, = 0.
In homogeneous plates this corresponds to uniform applied normal edge stresses
PXonx=O,uandP,ony=O,b].
Conditions (a), (b) and (c) can be apphed together.

SIMPLY SUPPORTED R~CTAN~U~R PLATES

The edge boundary conditions for a simply supported rectangular plate may be specified
as :
Onx = Oanda; cr, = 0, w= 0 and v = 0,
(3)
Ony=Oandb; o, =O, w=O and u=O.
This set ofedge conditions is identically satisfied by choosing u, v, win the foliowing form ;

[~~ = h$ljl ~~~~~~~~~~~~ (4)

where cfi,11/and x are pure fun~ions of Z. Considering simple harmonic osculations and
uniform applied normal edge stresses, substitution of equation (4) in equation (2) yields
for each combination of m and n,
d,-t-d# d, c&L d,
d3 d5 + d6L2 d,L J/ = (01 ~~~
[ -d4L -d7L d,fd,L2 liI x
where,
L denotes the operator 5,

d, = E~2-E~M~-~~~~2, dz = G,,, d3 = ;- @,,+ G~~)MN,

d, = (Em + G,JM, d5 = EA2-E,N= - G,,M2, de = Gym
d, = V&f G,AN, ds = ER*- G,M2 -G,N2, dg = E,;
A2 = 0, if the plate is static and also free of body forces @a)
= 5_2’ph*/E, for simple harmonic oscillations (6b)
= (P&W2+ ~~N2)/E, for u~ifo~ applied normal edge stresses, P, on x = 0, ct.and
P,ony=O,b (6c)
For non-trivial solution of the homogeneous equation (5), the determinant iaf rhc (.i .; ,?!
matrix on its left hand side must be zero. This yields
$1 &j d,(L2)3 -k(l& d: + 61,Ii; -l-d, db d, -t d, d, d, -t c/z “1, d,)( L” I2
~(dgd~-2d~dQd,$dld~+d1~16dH+dZd5ifK-~-dltf5~21)-.~ffifo~f~~
4”d,(d, d, - d$ = 0 ;_
!Ib
for each combination of m and n. The six roots of this equation define the six values of 1..
and correspondingly there are six arbitrary constants A(‘j(i = 1,2, . .6). Each arbitrary
constant A(” is associated with an eigenvector ~~(~)~~~)~(~)~.
The eigenvectors depend on the
multiplicity of the roots and a procedure for finding them is discussed in [13]. The eigeu-
vector corresponding to a non-repeating root I. = (’ is
dqL.(f&-t d,c2) - (’tl, d7 _
d&d, t-d,c2)-i*dj d, A&

i_-(d, +dd,c’)(d, +d,r,‘)+d:

1
where A is an arbitrary constant. The sum of the six eigenvectors corresponding to the six
roots gives the total expression for 4, u’/and x :

Using the stress-strain relationships (given in the notation) the foliowing expression for
stresses are obtained.
5x -Earn - ~~~~~ + E&x

OY -E,,Mc/I--E~N$+E,,Lx sinmrcX sin nzY

f7: 1 =-f 5 j - E,,M4 - E,,N$ + &Lx _

j
(3
‘tXY m=t n=~ G,,(NQ,-bM~)cosmnX cosnnY I
zx G,,(M~ c L#) cos mzX sin rr7tY

1 _j
zY’ 1G,,(Nx + L$) sin mnX cos rutY
The set of six arbitrary constants for each (m, n) in $, + and x is used to satisfy the six lateral
surface conditions.
As mentioned earlier, extension of this anaiysis from homogeneous plates to laminates
is straightforward. Equations (4H9) are valid for individual plies, the appropriate quantities
for the jth ply being indicated by subscript j (j = I, 2, . . , p).

LATERAL SURFACE CCWDITIONS

If the top and bottom surfaces are subjected to normal and shear stresses q&, ~1).
qxz,(x, Y), qYzl(x, Y) and q&, Y), a&, Y), qY&, Y) respectively, (they can be either static
loads or amplitudes of forcing functions with frequency a), then the lateral surface conditions
 Bending, vibration and buckling of simply supported thick orthotropic rectangular plates and laminates 1469

to be satisfied are :
for homogeneous plates;

and ~out~nuity conditions

z Wj-“t;+t zzc 0
uj-"j+l -- Vj-Vj+l (11)
The loadings are best expressed in double Fourier series,

Q,,,,, sin m7cXsin nit Y.

Q,+, cos mnX sin nn Y
QYz_ sin mzX cos nn Y
(12)
Q,,,,, sin mxX sin nz Y
QXzb,,,”cos mnX sin nx Y
Q,,lb,mn
sin mxX cos nrcY

Then, satisfaction of the lateral surface conditions, equations (10) or (1 l), leads to a set of Sp
simultaneous equations for each combination of m and n. These can be put in the foIlowing
form.
(a) For homogeneous plates ;

(13)
1470 S. SRIWWAS and A. K Iho

(b) For three-ply laminates ;

(c) For multi-ply laminates;

Equation (14) is extended in a straightforward manner and symbolically,

iCl{Bi= {Ql i14a)

where [C] is a square matrix made up of [R], [S] and null matrices, (B) is a column matrix
consisting of {Aj;‘s, and (Q) is the load matrix.
In the foregoing equations,
fAjj = (A”‘A(~ ,.., /P’jj,

[R(Z)lj =

and [0] is a (3 x 6) null matrix. In the above equations, superscripts denote the correspond-
ing eigenvectors and subscripts the corresponding plies. It is simple to modify equation (13)
or (14) to include any displacement conditions on the exterior lateral surfaces.
In non-homogeneous problems, i.e. in those problems for which the right hand side of
equation (13) or (14) is not zero, results can be evaluated by summing infinite series whose
individual terms are known explicitly by solution of equation (13) or (14), to the desired
accuracy by retaining su&cient number of terms. In case of homogeneous problems the
determinant (order 6~) of the square matrix on the left hand side of equation (13) or (14)
 Bending, vibration and buckling of simply supported thick orthotropic rectangular plates and laminates 1471

must be zero and this is the characteristic equation for the corresponding (m, n) combina-
tion.

PLATE UNDER UNIFORM NORMAL STATIC LOAD ON TOP SURFACE

Consider a plate loaded only on the top surface by an uniformly distributed normal
static load q,, so that,
16qom and n odd
Q zt,mn = ___
mn7c2’
= 0,m or n even (15)
= Qwt,mn
Q-=t,mn = Qz,,,,
= Qxz+n= Qm,m = 0
The scheme for computation is :
(i) Determine the roots of equation (7) for each ply, for each (m, n).
(ii) Find the eigenvectors, taking into account any multiplicity of roots.
(iii) Solve equations (13) or (14) for each (m, n) to determine the arbitrary constants.
(iv) Sum the series (4) and (9) for the displacements and stresses, retaining sufficient
terms to achieve the desired degree of accuracy.
Numerical results are presented in Tables 2-4 for a homogeneous plate and a
three-ply laminate whose material properties are indicated in Table 1. In Table 2, exact,
Reissner and thin plate [12] values of (T.+_, CJ,,,~~~, z._,,~~ and w,, are given for homo-
geneous plates with a/b = 0.5, 1 and 2, h/a = 0.05, 0.1 and 0.14. In Table 3, the ox, cY,
7,. and w distributions across the thickness for a 14 per cent thick homogeneous square
plate are given. In Table 4, effects of modular ratio between plies (j?) are presented for three-
ply square laminates with identical top and bottom plies.
The following observations can be made from the data in Tables 2-4.
(i) Thin plate theory underestimates maximum deflection. The errors in both maxi-
mum stresses and maximum deflections predicted by thin plate theory, increase as plate
thickness increases. Deflections are more inaccurate than stresses.
(ii) Even for quite thick plates Reissner’s theory predicts the deflections almost exactly,
but for stresses it does not appear to be distinctly superior to thin plate theory.
(iii) The thin plate and Reissner’s direct stress distributions across thickness (Table 3) are
only slightly different from the true distributions.
(iv) Modular ratio between plies has significant effect on the errors in thin plate theory
for laminates. Mostly errors increase with increasing moduli of outer plies.
It was also observed that the number of terms needed to maintain a given level of
accuracy increases as plate thickness increases. To maintain 0.1 per cent accuracy, homo-
geneous square plates with h/a = 0.05, 0.1 and 0.14 require retention of terms up to
m( = n) = 25,87 and 167 respectively. That is, the significance of the higher harmonics of
loading increases with increasing thickness. Finally, from Table 3 one confirms that the
mid-surface is not really a neutral surface.

FREE VIBRATION
In this case the exterior lateral surfaces are stress free and therefore,
zzQxz,,,n
QZt,"ln = Qyzt,,,
= Qz,,,,
= Qxz+,n
= Qyzt,,,,
=0 (16)
1472

Present Thtn
h/a exact Reissner‘s plate
analysis theory theory

0.05 -21.542 _. 21,542 ZI.ZOI -- l-58

0.10 - 1408.5 ~~~140x-4 -1325! --5.92
O-14 - 387.23 - 387.21 -- 344.93 - IO.92

0.05 10,443 - 10,442 10,246 - 1.x9

0.10 - 688.57 -688.37 --&IO.39 -- 7w
0.14 - 191-07 - 191-02 - 166.70 - 17.75

0.05 - 2048.7 - 2047-9 - 1988. I -- 2.96

0.10 - 139.08 - 138.93 - 124.26 -I’ IO.66
0.14 - 39.790 - 39-753 --- 32.345 _ IX.71
-_

y0 : normal stress on top surface {z = 0); w: deflection of central point (X = I‘ = % = 0.j): 0, ,tnd ea mmai
stresses at centre of top surface (X = Y = 0.5, Z = 0); I,~ : shear stress at centre of an edge (X L- 0. Y = % = 0 51;
elastic moduli are as per Table i.

Substitution of equation (16) in equation f 13) [or {14)j makes the latter homogeneous and
for non-trivial solution of the problem, the determinant of the square matrix (6~ x 6~) on
the left hand side must be zero. For each (WI,n), simultaneous solution of this characteristic
equation along with equation (7) yield an infinite number of frequencies, each correspond-
ing to a different thickness mode.
Typical numerical results are presented in Tables 5W. In Table 5 the first 8 exact eigen-
values along with the 3 by ~indlin’s theory (K’ = 5/6) and one by thin plate analysis are
given for n&/u = 0.1, 0.2, . . , , O-5 and r&/h = 0.1, 0.2,. . ,0-S. In Table 6, the stress and
displacement distributions across thickness for a homogeneous plate are given fol
mh/a = nh/b = 0.3. Table 7 gives the relative magnitudes of maximum stresses and dis-
placements for antisymmetric thickness modes. In Table 8 the flexural mode eigenvalues
for various p’s are tabulated for three-ply laminates with identical top and bottom plies.
From the data obtained the following observations are made.
(i) It is confirmed that thin plate theory frequencies are higher than true values.
The errors increase with increasing mh/a or nhlb.
(ii) For homogeneous plates, all the 3 eigenvalues given by Mindlin’s theory are close
to the corresponding exact values.
(iii) Thin plate and Mindlin’s stress and dispfacement distributions across thickness
are significantly different from the true distributions. For the flexural modes the deviations
increase with mh/u or nh/b.
(iv) Complexity of the stress and displacement distributions across thickness increases
for higher order thickness modes.
 Bending, vibration and buckling of simply supported thick orthotropic rectangular plates and laminates 1473

ORTHOTROPIC PLATESUNDER UNIFORMSURFACE LOAD

‘&J40 L.f%

“/:,error “i, err*r

Present Thin in thin Present Thin in thin
exact Reissner’s plate plate exact Reissner’s plate plate
analysis theory theory values analysis theory theory values
-.
0.5
79.545 79.337 79,121 -0.53 - 14.048 14.114 -- 14.154 0.75
20.204 2oQo1 19.780 -2.10 - 6.9266 -70611 -7.1328 2.98
10.515 10.312 10.092 -4.02 -4.8782 - 50445 - 5.1050 4.65

1.0
87.080 86.921 86.487 - 0.68 - 10.873 - 10,864 - 10.972 0.9 1
22.210 22J.M 21.622 - 2.65 -5.3411 - 54267 - 5.5642 4.18
It.615 i I-453 il.031 - 5.03 -3.7313 - 3.8741 - 3.9860 6.83

2.0
54.279 54.134 53.838 -0.81 - 6.2434 -6.2191 -6.4175 2.79
13.888 13.743 13.460 - 3.08 - 2.9573 - 3.0524 - 3.2402 9.57
7.2794 7.1358 6%71 - 5.66 - 1.9987 -2.1428 -2.3115 15.95

(v) As the order of the thickness modes increases, the in uences of transverse stresses
increase.
(4 odu~ar ratio between plies has signi~cant effect on the error in ~exural frequency
due to n plate assumptions and this error increases with increasing moduIi of outer
plies. density ratios have only slight effect for small m and t-z.

TABLE 3. VARIATIONS ACROSS THICKNESS IN HOMOGENEOUS ORTHOTROPIC PLATES UNDER UNIFORM NORMAL LOAL)
ON TOP SURFACE

Thin plate Thin plate Thin

z True Thin True and True and True plate and
distribution plate distribution Reissner distribution Reissner distribution Reissner
theory theories theories theories

0 1.00~ 1.0 3.0 1-O 1-O t-0 0 0

o,t I4iIO.5 I.0 0,784P 0.8 0.7914 0.8 0.4992 0.36
0.2 1mo7 1.0 0~5801 0.6 0.5888 06 0.7325 0.64
0.3 1~0007 1.0 a.3832 0.4 0.3906 0.4 0.8906 0.84
0.4 1J)oo4 1.0 0.1914 0~2 0.19S4 0.2 0.9789 0.96
0.5 1.0 I.0 - 0.0023 0 -0.0016 0 I.0 I.0
0.6 0.9994 1.0 -0.1869 - 0.2 -0.1921 -0.2 0.9540 0.96
0.7 0.9986 1.0 - 0.3786 -0.4 -0.3874 -0-4 0.8388 0.84
0.8 0.9976 1.0 -0.5755 -0,6 -0.5856 -0.6 0.~92 O&I
0.9 0.9964 t.0 - 0.7803 -0-8 -0-7882 -0.8 0.3758 0.36
1.0 0.9949 1.0 - 0.9954 - I.0 - 0.9968 -1.0 0 0

Variations of w, o, and oy along X = Y = 0.5 and t,, along X = 0, Y = 0.5. Elastic moduli are as per Table 1.
o,!h= 1, h/a = 0.14.
1474 S. SRINIVAS and A. K K,w

7 AHLF 4. DEFLECTIONS ANU 5 I KI:SSI:~w I‘~IKH :a,

XXI_. _~I- .___ -. _

Present exact analysts

1 5
-._ -~_-~__._ ._
- 688.58 - 258.97

Top ply at top surface 36.02 I 60.353 65.332 66.78 7

Top ply at interface 28.538 46.623 48.857 48.29Y
Mid ply at upper interface 28.538 9.3402 4.9030 3.237’)
Mid ply at lower interface - 28,454 - 92845 - 4.8600 ~~3.2009
Bottom ply at interface - 28.454 - 46426 - 48.609 -48.028
Bottom ply at bottom surface - 35.937 -60.155 .- 65.083 - 66.5 13

Top ply at top surface 22.210 38.49 1 43.566 46.424

Top ply at interface 17.669 30.097 33.413 34.955
Mid ply at upper interface 17.669 6.1607 3.4995 2.494 I
Mid ply at lower interface - 17.631 - 6.0574 3.3669 -- 2.3476
Bottom ply at interface - 17.631 - 30.322 33.756 - 75.353
Bottom ply at bottom surface -22.172 ~ 3x.715 -- 43.908 -~ 46.82 1

*,zi 40
At upper interface - 2.402’) - 3.7194 - 3.9285 - 3.Y55Y
At mid surface -5.3411 - 4.364 1 ~~4.0959 -. 3.9638
At lower interface - 1.9826 -- 3.2675 .~ 3.5154 _ 3.5768

q,): normal stress on top surface (z = 0); x at ceutre X = Y = L = 0.5, ox and (r). on A :: Y :T 0.5 and
rsz on X = 0, Y = 0.5 ; elastic moduli of all plies as per Table 1. Top and bottom plies are identlcai. /I, h -= 0. I,
h,.!h= O.&a/h = 1, hj'a= 0.1.

BUCKLING
The problem of buckling of simply supported’homogeneous plates is formally identical
IO that of free vibrations, except that the eigenvalue 1. should be interpreted according to
equation (6~) instead of equation (6b), as for example in Table 5. As already stated, only the
flexural mode eigenvalues are of significance in many buckling problems. The present
analysis can be directly extended for laminates which are initially in a state of uniform
strains ax, E, and crz = 0 throughout. Otherwise, a three-dimensional analysis of the initial
state will be involved and also the governing differential equation will be complicated by the
presence of variable terms (local stresses). The required initial state of uniform direct strains
can be realised in experiments by compressing by pairs of smooth rigid edge blocks moving
towards each other without rotation. This initial state is equivalent to one of direct stresses
Px,, P,,j uniform within and on the edge of each ply, such that the quantities (“,,PX, +-
Fxy,Pyj),(F,,,P,, + FyjPyj)are independent of j. In thin or moderately thick laminates, small
deviations from this condition at the edges, can be ignored by invoking Saint-Venant’s
principle and the present analysis can still be applied to obtain satisfactory results.
Numerical results are presented in Fig. 2 and Tables 9 and 10 for plates and laminates
loaded by stresses Pxi on sides x = 0 and a only. As such, (F,,P,,) and (F,,,P,,) are the same
for all plies, In other words, {~,~F,,,) are the same for all plies and (P,, : Px,: .) =
(l/F,, : l/F_ : .). The least buckhng stress P,,,,,, for any given plate dimensions and m is
 Bending, vibration and buckling of simply supported thick orthotropic rectangular plates and laminates 1475

ORTHOTROPIC LAMINATES UNDER UNIFORM SURFACE LOAD

Thin plate theory y< error in thin plate theory

5 10 15 1 5 10 15

-216.94 - 118.77 -81.768 - 7.00 - 16.23 - 25.48 - 32.82

61.141 66.947 69.135 0.21 1.31 2.47 3.52

48,913 53-557 55308 1.19 4.9 1 9.62 14.51
9.7826 5.3557 3.6872 I.19 4-74 9.23 13-88
- 9.7826 - 5.3557 - 3.6872 1.49 5.36 10.20 15.19
-48.913 - 53.557 - 55308 1.49 5.36 10.18 15.16
-61.141 - 66.947 - 69.135 0.45 I.64 2.86 3.94

36622 40.099 41.410 - 2.65 -4.86 - 7.96 - 10.80

29.297 32.079 33.128 -2.11 - 266 - 3-99 - 5.23
5.8595 3.2079 2.2085 -2.1t -489 - 8.33 - 1l-45
- 5.8595 - 3.2079 - 2.2085 - 189 - 3-27 -4.72 - 5.93
- 29.297 - 32.079 -- 33.128 - 1.89 - 3.38 -4.97 - 6.29
- 36.622 - 40.099 -41.410 - 2.48 -5.41 - 8.67 - Il.56

- 3.3860 - 3.7075 - 3.8287 - 1664 - 8.96 - 563 - 3.22

-4.5899 - 4.3666 - 4.2825 4.18 5.17 6.61 8.04
- 3.3860 - 3-7075 - 3,8287 I.03 3.63 5.46 7.04

real&d with only n = 1. For a given h/b, the buckling stress parameter k, achieves a
constant value for all integral values of a/b, and the corresponding m = a/b. That is, when
ujb is integral the plate buckles into square panels. This value of k, is also its asymptotic
value.
In Fig. 2 the buckling stress parameter k, is plotted against u/b for various h/b’s. In
Table 9 the asymptoticyalues of the buckling stress parameter k, are given for homogeneous
orthotropic plates with h/b = 0.05,O.l and 0.2. In Table 10, the asymptotic value of k, is
given for three-ply laminates. From these data the following observations on buckling of
orthotropic plates and laminates may be made.
(i) Thin plate theory gives optimistic values for the buckling stress, the errors increas-
ing rapidly as the thickness increases.
(ii) Mindlin theory can yield accurate buckling stresses even for thick homogeneous
plates.
(iii) Modular ratio between plies has significant effect on the errors in buckling stress
due to thin plate type assumptions. The errors increase as moduli of outer plies are in-
creased.
The stress and displacement distributions are similar to those of the corresponding
flexural mode in vibration.

CONCLUDING REMARKS
An exact three-dimensional unified analysis has been established for the bending,
vibration and buckling of simply supported, thick, orthotropic, rectangular plates and
 Exact analysis
----.. .__

I-A i-b”

..-.. __.. _
0 1 0.04742 O-21695 0.3Y405
0.2 0.10329 0.34501 0.56242
0-? 0.18881 O-49530 0.76004
04 0.29690 065 190 0.9690 I
O-S 0.42 124 0~81071 1‘1825
0. I 0~118x0 0.35 1jo 0.6727X
0.2 0.14942 0.43382 O-78796
0.3 O-24753 0.55201 W4433
0.4 0.34755 0.68957 1~12.71
0.5 0.46428 0.83699 I.3142
0. I 0~21804 O-5029 i O.Yi278
0.2 O-26244 o-560?7 1WI73
0.3 0.33200 O-65043 I.1814
0.4 042242 0.764 I 5 l-3321
0.5 0.52956 0.8936Q I .4994
0.1 0.33189 0.65908 1.2795
0-z 0.37066 0.70277 I .I3453
0.3 0~43225 0.77334 1.4463
0.4 0.5 1342 0.5666? I .5776
0.5 06 1092 0.977h’) L71YI
0 I 045265 @);i720 1~Sii90
0.2 048680 0.85223 I JA2.i
0.3 O-54 160 0.90962 l-7266
0.4 061465 U.98732 , .s, x;“- ’
0.5 0.70338 1.0823 I .x.559”‘ 1

* Pure thick- twist modes.

Elastic moduli as per_?_&!e I.
Eigenvalue i = fix/phZ/E,for free vibrations
,-.--- .-
C ‘P ,‘,,,, M?‘E, for buckling under uniform normal edge stresses I’, on .\ ; 0 and ~4.
\I
A and S denote modes which are antisymmetrtc and symmetric about mid-plane respectively

laminates. The solution is in series form and each term in it is explicitly defined by a set of
sinluItaneous equations. Thus it is easy to sum the series to any desired degree of accuracy
in non-homogeneous problems like plates under static or dynamic loads on lateral surfaces,
while in homogeneous problems like free vibration or buckling from a state of uniform
strain closed form characteristic equations are obtained to define natural frequencies o!
buckling stresses. This type of sotution was made possible by the choice of 2:=T-0, instead of
T,, = 0, as one of the “simple” support conditions along the straight edge x -I::const.
In fact equation (4) is only one of the twelve sets of series providing the general solution [ 141
of the governing differential equation. If the condition z,, = 0 is to be satisfied the series in
equation (41, should be augmented by additional series. Then the terms are no longer
explicitly determinable nor is the characteristic equation in closed form. There is consequent
loss of elegance and simplicity and enormous increase in computation. An example of an
isotropic square plate has been worked out f14] to examine whether the two modes oi
defining a simple support wouId lead to significantly different results. The L‘= 0 condition
 Bending, vibration and buckling of simply supported thick orthotropic rectangular plates and laminates 1477

AND BUCKLING OF IiOMOGENEOUS PLATES

Mindlin’s analysis
Thin plate y0 error in
theory thin plate
IV-P V-S I-A II-A* III-.4 I-A values
-.~
2.5479 3.2636 0.04740 1.3159 1.6646 0.04967 4.74
2,5604 3.3043 O-10317 1.3411 1.7305 0.11200 8.44
2.5824 3.3678 0.18839 I.3841 1.8306 0.21537 14-07
26145 3.4497 0.29588 I.4445 1.9560 0.35993 21.23
2,6565 3.5458 0.41927 1.5204 2.1001 0.54574 29.56
2.6264 3.2760 0.11873 1.4285 1.6921 0.13538 13.95
2.6334 3.3179 O-16918 1.4393 l-7655 0.19866 17.26
26481 3.3828 0.24692 1.467 1 1,8715 0.30289 22-37
2.6127 3.466t 0.34628 1.5142 2xlOO2 044802 28.91
2.708 1 3.5633 0.46201 1.5792 2.1456 0.63418 36.59
2.7383 3.2967 0.21776 1.5857 1.7450 0.27789 27.45
2.7392 3.3409 0.26192 l-5727 1.8341 0.34176 30.22
2.7457 3.4085 0.33101 1.5813 1.9480 044699 3464
2.7614 3.4942 0.42066 i.6127 2.0796 0.59313 40.4 I
2.7882 3.5933 0.52671 1.6648 2.2250 0.78011 47.31
2.8756 3.3261 0.33110 1.7266 1.8657 0.47734 43.82
2.8695 3.3743 0.36955 l-7022 1.9587 0.54152 46.09
2.8668 34463 0.43056 1.7006 2.0716 064750 49.80
2.8730 3.5356 0.51084 1.7218 2-2003 0.79466 54.78
2.8905 3.6373 0.60716 1.7638 2.3419 0.98269 60.85
3.0334 3.3648 0.45087 1.8150 2.0770 0.73374 62.10
3.0174 3.4204 O-48463 l-8067 2.3475 0.79809 63.95
3@040 3.4991 O-53877 1.8095 2.2459 ow459 67.02
3xxIOG 3.5931 0.61083 1.8292 2.3639 1.0526 71.25
3.0082 3.6979 0.69828 1.8665 2.4965 1.2416 76.52

is found to predict smaller central deflections, lower by 7.4 per cent, and 9.4 per cent for
10 per cent and 14 per cent thick plates.
The effect of thin plate assumptions is to increase the stiffness of the structure and there-
fore yield lower deflections, higher flexural frequencies and optimistic buckling stresses.
The errors increase as the thickness increases, and in vibrations, also with increasing order of
the mode as m or n increases. In laminates the modular ratios between pfies have very signifi-
cant effects on the errors due to thin plate type of assumptions and these errors increase
with increasing moduli of outer plies. In vibrations thin plate theory yields only one doubly
infinite spectrum of frequencies.
In orthotropic plates, the errors due to thin plate assumptions depend strongly also on
the extent of orthotropy, i.e. the relative values of various elastic moduli. The errors in thin
plate theory increase as the relative magnitudes of transverse elastic flexibility constants
(F,, F,,, F,,,, l/G,,, l/G,) increase relative to inplane constants (F,, F,, Fsy, l/G,,). In
isotropic plates, it was possible to define a “thin plate”. For example tolerating 1 per cent
inaccuracy, a plate could be called “thin” if it is under 5 per cent thick when subjected to
uniform loading or when {(~&/a)~+ (n!~/b)~}I l/250 for free vibrations [7,8]. Because of
the large number of parameters involved, it is rather difficult to define the term “thin” for
orthotropic plates.
1478 S. SRINIVAS ad A K. KAO

TABLE 6. VARIATWNS OF STRESSESAND DISPLACEMENTSACROSS THICKNESS FOR A ~1,,<i’b ‘. 1-c

It , i ‘j<I ;I(.; J: I.;,c
PLATE IN FREE VIBRATION
- . . I. ^“_-^-_ ..__.... -

c, 0, c: i,> T,; it:

z
4) u,,(O) bj(0.4) r,,(O) sJO.5) q,(O.Si u1;i: ,.(tj) \.J,!,
__-
(a) Thin plate theory and the three thickness-modes of Mindl!n’\ [heor)
0 1.0 I.0 . 1.0 0 0 ( ,!

0.1 04 0.8 0.x 0.36 0.36 i! x

0.2 0.6 0.6 0.6 0.64 0.64 (I.6
0.3 0.4 0.4 0.4 0.84 0.84 0.4
0.4 0.2 0.2 (1.2 0.96 0.96 !).-’
0.5 0 0 0 I.0 I0 II
Ib) Exact theory
First antisymmetric thickness mode (I-A)
0 I .o I.0 0 I .o 0 !I I-O I -li !4
0.1 0.7 134 0.7539 1.5685 0.7360 0.3943 0-3X35 0.6963 0.1644 I al54
0.2 0.4871 0.5396 2.0482 0.5 163 0.6738 ! 1.6632 0.4650 0.5532 I ul8?
0.3 0.3028 0.3479 1.7657 0.3278 0.8595 0 /YP34
J 0.2838 0.3595 I-00696
0.4 0.1450 0.1704 1.0 o-1591 0.9655 !I 0638 0.1343 0.1770 I~(9102
0.5 0 0 0 0 I .o i0 0 0 l.OlO3
Second antisymmetric thickness mode (II- 4 1
0 I.0 1.0 0 I.0 I4 I Xl i 0
0.1 o&J511 0.9509 1.3346 O.Y512 04090 0.951 I 0.95 IO 16.658
o-3 0~8090 0.8088 I .8798 0.8092 ii.5878 0.8090 0.8089 30.755
0.3 0.5878 0.5875 I .7099 O.5880 0.8090 0.5878 ii.5877 41.92-
0.4 0.3090 0.3089 1.0 0.3091 !).Y51 1 0.3090 0.3090 4Y.OY5
0.5 0 0 0 0 i 0 0 :,1 Cl $6
Third antisymmetric thickness mode (111-A)
0 I.0 1.0 0 I4 i! 0 Ia 141 : 81
0.1 0.9837 0.9516 I.3098 0.9789 O-2942 0.3021 0.9922 0.9721 O.Y(l_!
0.2 0.8573 0.8111 I.8614 O.X5Ul 0.5719 0.5804 08698 0.8399 (,“Y”‘,’
0.3 0.6333 0.5905 1.7037 0.6263 0.7992 0.8045 0.6450 0.6 167 0.6.t iy
0.4 0.3361 0.3109 I.0 0.3320 0.9482 O.Y497 0.343 I 0 3262 $55 T-7
0.s 0 0 0 II I.0 I .o 0 1) Ii.5 ?(I!>

z,-
First symmetric thickness mode (I-S) zJO.2)
0 I.0 I4 0 I-O 0 0 1.0 I -0 1.0
0.1 O.Y790 I.0130 0.3868 I .04 I 5 0.760’) 0.7614 0.9890 I .0074 0.81 1:
0.2 O-9636 I.0228 0.6777 I .07x 1.0 I.0 0.98 I 1 I 4 I 3i 94 141;
0.3 0.9533 I .0297 O-8804 I .0925 0.8661 0.8656 0.9757 I .0170 0.41’9
0.4 0.9472 1X)337 I .o I 1060 0.4919 0.49 15 0.9727 l-0194 $2073
0.5 0.9453 1.0350 I.0395 I.1 101 0 0 0.9717 I .02!)2 il
Second symmetric thickness mode (II-S)
0 14 I.0 0 I .o u 0 I.0 j iI I.li
0.1 I .0077 1.0062 0.3602 I.009 1 0.7241 0.7336 1.0073 ! 01’0 tl~83bl
0.2 1.0139 I.0114 0.6526 I.0165 I.0 1.0 1.0131 I.0221 11.6489
0.3 1.0184 I.0152 0.8680 I a?0 0.8968 0~8Y87 1.0173 I .0296 0443 1
o-4 1.0211 1.0176 1.o I .0254 Cl.5197 0.5122 1.019’) 1a342 O.‘24X
O-5 I.0221 1.0184 la444 1.0265 0 0 1.0208 i .035x o-

Elastic moduh as per Table I : mh/tr= n/~/h= 0.3.

The improved approximate theories like Reissner’s and Mindlin’s are good even for
quite thick plates for specific purposes. For example, Reissner’s theory predicts deflections
accurately, while Mindlin’s theory yields accuracte frequencies for the first 3 antisymmetric
thickness modes of free vibrations.
 Bending, vibration and buckling of simply supported thick orthotropic rectangular plates and laminates 1479

TABLE 7. RELATIVEMAGNITUDESOFMAXIMUMSTRESSESANDDISPLACEMENTSFORANTISYMMETRICTHICKNESS MODES

OFHOMOGENEOUSORTHOTROPICPLATESINFREEVIBRATION

Thin plate theory (I-A) 0.6142 0 0.4272 0.2932 0.2023 0.4712 0.4712
Mindlin’s theory
Thickness mode I-A 0.7693 0 0.4937 0.3306 0.2644 0.2290 0.3550
Thickness mode II-A 0.0174 0 0.1563 0.4716 0.3906 40.27 18.93
Thickness mode III-A 0.4804 0 0.3698 1.032‘ 0.5694 6.908 4.160
Exact theory
Thickness mode I-A 0.7278 0.0301 0.4759 0.2827 0.2290 0.2677 0.3121
Thickness mode II-A 0.0565 00002 0.1396 0.6128 0.5355 41.62 22.26
Thickness mode III-A 0.8624 0.1657 0.5336 04407 1.310 1.898 3.680

mh/a = nh/b = 0.3.

ORTHOTROPICLAMINATES

Density y0 error in
;zz /I = E,,/E,, Exact value Thin plate thin plate
1 value value

1 1 0.047419 0.049666 4.14

1 2 0.057041 0.060584 6.21
1 5 0.077148 0.085333 IO.61
1 10 0.098 104 0.115328 17.56
1 15 0.112034 0.138994 24.06
3 15 0.094548 0.117471 24.25

Elastic moduli of all plies as per Table 1. Top and bottom plies are
identical. h,/h = 0.1, h,/h = 0.8, mh/a = nhlb = 0.1. Eigenvalue =

8, I I I
i1
3.039, ALSO THIN PLATE
-I
ANALYSIS NEGLECTING
h/b EFFECTS
2.966 ,

4
m*

_-- ... -__-_-----.-_______

t

” 1 2 3 4
a/b

FIG. 2. Variationof buckling stressparameter k, with h/b and a/b for plateloaded on x = 0, a
kx = (P,,dE)(12/n2)(blh)‘.
1480 S. SRINIVA~ and A. K. KW

TARLEO. ASYMPTOTIC.~.FOR HUC'KL.INGDI~

HOMOC;ENEOIJSOR~~HOTKI~~~~~
PLATES UNDER UNIFORM NORMAL STRESS P, ON EDGES .\ = 0 h'v!>

I,
()CTlOi L,!
h;h Exaci Mindlin’s Thin plate thin pl,ttc
theory theory theory \alWl

0.05 2.966 2.965 3.039 246

0. I 2.770 2.768 3.039 9.7 1
0.2 3.210 2.204 3.039 37-i

12 P+‘~ 2
Elastic moduli as per Table 1. k, =
7rL E, Iih

TABLE 10.ASYMPTOTK~, FOR BIJCKLINGCS rHREE+I Y

ORI‘HOTROPIC LAMINATES FROM AN INITIAL.SIAI-EW
,,N,FORM STRAlh

‘lo error in
B Exact value rhin plate thin plate
value value

I 2.770 3-039 9.7

2 3.330 3.768 13.2
5 4.046 4.984 23.2
10 4.200 i&52 39.3
15 4.037 6.263 55. I

Elastic moduli of all plies are as per Table 1. Top

and bottom plies are identical.

/i = E,r,;E.,l = E,,.‘E,>

I’ = P,.‘,,,,$ EX, (I = I,.?. 3)

k, = 12/l(hjh)2k~

h,.lh = WI. h,;h = OX h, h = 0.1.

Thus, a three-dimensional elasticity solution is necessary when both stresses and

displacements are required in non-homogeneous problems or for establishing the full
spectrum of modes in free vibrations. The analysis presented here is for general orthotropic
laminates with arbitrary properties for each ply, except in buckling when the relative elastic
moduli must be identical for all the plies. Analysis of sandwich plates with core having only
shear rigidity is a special case of the above analysis. Tetragonal, cubic and isotropic materials
are special cases of orthotropic materials and the present analysis is applicable.

Acknowledgemenf-This research was sponsored by Directorate of Research Labs., Ministry of Defence. Govt. 01
India as a Grant-in-Aid Scheme and was carried out with Prof. C. V. Joga Rao as the Chief investigator.

REFERENCES

[IIE. REISSNER,On the effect of transverse shear deformation on the bending of elastic plates. /. appl. MPCII.12.
A69 (1945).
PI R. D. MINDLIN, Influence of rotatory inertia and shear on flexural motions of isotropic elastic plates. J. uppl.
Mech. 18, 31 (1951).
 Bending, vibration and buckling of simply supported thick orthotropic rectangular plates and laminates 1481

c31 S. J.MEDWADOWSKI, A refined theory of elastic, orthotropic plates. 1. appl. Mech. 25,437 (1958).
[41 P. C. YANG, C. H. NORRIS and Y. STAVSKY,EI~S~~Cwave propagation in heterogeneousplates. Int.J. Solids
Struct. 2, 665 (1966).
is1 J. W. WHISKY, The effect of transverseshear deformation on the bending of laminatedplates.J. cump.
Mater. 3, 534 (1969).
[61 1. camp. Mater. 3 (1969).
Fl S. SRINIVAS, A. K. RAOand C. V. JOCA RAO, Flexure of simply supported thick homogeneous and laminated
rectangular plates. 2. angew. Math. Mech. 49, 449 (1969).
[81 S. SRINIVAS, C. V. JOGARAO and A. K. RAO, An exact analysis for vibrations of simply supported homo-
geneous and laminated thick rectangular plates, Jnl Suund & Vibm. to be published.
[91 S. SRINIVAS, C. V. JOGA RAO and A. K. RAO, Some results from an exact analysis of thick laminates in vibra-
tion and buckling.J. appl. Meeh. to be publish~.
DOI S. SRINIVASand A. K. RAO, Buckling of thick rectangular plates. AIAA Jnl7, 1645 (1969).
[111 R. L. BISPLINGHOFF, J. W. MAR and T. H. H. PIAN,Statics of deformable solids, Chapter 7. Addison-Wesley
(1965).
t121 S. TIMOSHENKO and S. WOINOWSKY-KREIGER, Theory ofplates and shells, 2nd edition, Chapter 1I. McGraw-
Hill (1959).
1131 W. E. BOYCE and R. C. DIPRIMA,Elementary differential equations and Boundary value problems, Chapter 7.
John Wiley (1965).
[14] S. SRINIVAS,Thr~-dimensional analysis of some plates and laminates and a study of thickness effects.
Ph.D. thesis, Dept. of Aero. Engng., Indian Institute of Science, Bangalore.

(Received 22 July 1969; revised 18 December 1969)

AficTpaKT-&iWX ~~AUHeHH~~ TOWIbIiiaHaJIii3CTBTNKK II n~HaM~K~ HeKOTOpOfO lCIE%Xa TOlIClbtX

C~O~CTbiX~a~~KOB.~bI5O~~TC~ T~XMepHaK,~~He~Ha~ TeOpH5IMaJIbIXW#IOpMaLQ3I%, BpaMKaX TeOpHH
YIIpyrOCTH AJIRH3rIi6a,Kone6atM W nOTepIi YCTO#'IIIBOCTH,CBO6OAHO OIIepTbIX TOJlCTbIX Opl’OTpOEIHbIX
IlpxMOyrOJIbHbIX IIJIaCTIlH H CJIOHCTMX MSLCTHICOB. hIleHIte IlBJIReTCR ~OpMWlbHO TOYHbIM W IIpKB0AH-r
K HeCJlOW(HbIM6eCKOHe'IHbIM pXnaM AJIa HarIpaxeHHfi W Ilept?MeUeHMti IIjW H3rA6e, BblH)‘~J@HHbIM
KOAe6aHHRM H 3aAWElM TllIE3 “6aJXKa-KOnOHHa”. &IX CBO6OAHbIX me KOJIe6aHHk II3aAau )'CTOtiYABOCTH
lWIIeHIieIIpRIWLViTK XapaKTepIiCTWeCKKM ypaBHeHHIM B 3aMKHYTOM BUAe. B CJiyWe CBO6OJ@ibIX
Kone6aHu~ nnacTiiHoK, n~~araeMb1~ aHam AaeT -rpo&n& ~KoHeqHbi~ cneIcrpAn51 YacToT, BMecTe
TOJIbKo ABOBHOrO 6ecKOHe%HOrO CIIeKTpa,B PaMKaX T’ZOPUU TOHKHX IIJlaCfBHOK HJII~ TWX jQ3ORHMX.
6ecKoHeYHbIx cnexrpos npir atiam3e wna Pei%ccHepaH MwHnntrHa. ,&tsi nnacTIfH w cnoHc-rbIxMaCTHKOB
JWOTCx HeKOTOpbIe~~CA~HHbIep~3Y~bTaTbI.~paBHeHAepe3ynbTaTOBTeOpllHTOHK~XnnaCTClHOKHTeOpAek
PefiCCHepaIIMminmiHac pe3YJIbTaTaMH IIl3eA.QO)KeHHblMHB pa6oTe &X!T H~KOTObleBa~Hjlblt2 3aKnIo9eHIiR
OTHOCWTeJlbHO BaXCHOCTW H3'$+2KTOB IIWAIIO3IOXCeHIik,CAeAaHHbIX B npH6nH~eHHbIX TeOpYIRX.