NASA Technical Paper 3568

Buckling Behavior of Long Anisotropic Plates

Subjected to Combined Loads
Michael P. Nemeth
Langley Research Center Hampton, Virginia

National Aeronautics and Space Administration

Langley Research Center Hampton, Virginia 23681-0001

November 1995
Available electronically at the following URL address: http://techreports.larc.nasa.gov/ltrs/itrs.html

Printed copies available from the following:

NASA Center for AeroSpace Information National Technical Information Service (NTIS)
800 Elkridge Landing Road 5285 Port Royal Road
Linthicum Heights, MD 21090-2934 Springfield, VA 22161-2171
(301) 621-0390 (703) 487-4650
 Abstract

A parametric study is presented of the buckling behavior of infinitely long sym-

metrically laminated anisotropic plates subjected to combined loads. The study
focuses on the interaction of a subcritical (stable) secondary loading state of constant
magnitude and a primary destabilizing load that is increased in magnitude until buck-
ling occurs. The loads considered in this report are uniform axial compression, pure
in-plane bending, transverse tension and compression, and shear. Results are pre-
sented that were obtained by using a special purpose nondimensional analysis that is
well suited for parametric studies of clamped and simply supported plates. In particu-
lar, results are presented for a [+45] s graphite-epoxy laminate that is highly anisotro-
pic and representative of a laminate used for spacecraft applications. In addition,
generic buckling-design charts are presented for a wide range of nondimensional
parameters that are applicable to a broad class of laminate constructions. These
results show the general behavioral trends of specially orthotropic plates and the
effects of flexural anisotropy on plates subjected to various combined loading condi-
tions. An important finding of the present study is that the effects of flexural aniso-
tropy on the buckling resistance of a plate can be significantly more important for
plates subjected to combined loads than for plates subjected to single-component
loads.

Introduction with a significant potential for reducing structural weight

of aircraft and launch vehicles. Thus, understanding the
Buckling behavior of laminated plates is a topic of buckling behavior of symmetrically laminated plates is
fundamental importance in the design of aerospace vehi- an important part of the search for ways to exploit plate
cle structures. Often the sizing of many subcomponents orthotropy and anisotropy to reduce structural weight.
of these vehicles is determined by stability constraints, in
addition to strength and stiffness constraints. One sub- In many practical cases, symmetrically laminated
component that is of practical importance in structural plates exhibit specially orthotropic behavior. However,
design is the long rectangular plate. These plates com- in some cases, such as [+45] s laminates, these plates
monly appear as subcomponents of stiffened panels that exhibit anisotropy in the form of material-induced cou-
are used for wing structures. In addition, long plates pling between pure bending and twisting deformations.
appear as subcomponents of semimonocoque shells that This coupling is referred to herein as flexural anisotropy,
are used for fuselage and launch vehicle structures. and it generally yields buckling modes that are skewed in
Buckling results for infinitely long plates are important appearance, as depicted in figure 1. The effects of flex-
because they often provide a useful conservative estimate ural orthotropy and flexural anisotropy on the buckling
of the behavior of finite-length rectangular plates, and behavior of long rectangular plates subjected to single
they provide information that is useful in explaining the and combined loading conditions are becoming better
behavior of these finite-length plates. Moreover, knowl- understood. For example, recent in-depth parametric
studies that show the effects of anisotropy on the
edge of the behavior of infinitely long plates can provide
insight into the buckling behavior of more complex buckling behavior of long plates subjected to compres-
structures such as stiffened panels. sion, shear, in-plane bending, and various combinations
of these loads have been presented in references 1
An important type of long plate that appears as a through 4. The results presented in these references indi-
subcomponent of advanced composite structures is the cate that the importance of flexural anisotropy on the
symmetrically laminated plate. In the present paper, the buckling resistance of long plates varies with the magni-
term, "symmetrically laminated plate," refers to plates in tude and type of the combined loading condition. How-
which every lamina above the plate midplane has a corre- ever, the extent of the influence of the combined loading
sponding lamina located at the same distance below the condition on the importance of neglecting flexural
plate midplane, with the same thickness, material proper- anisotropy in a buckling analysis is not well understood.
ties, and fiber orientation. Symmetrically laminated
plates remain flat during the manufacturing process and The objectives of the present paper are to present
exhibit flat prebuckling deformation states. These char- buckling results for specially orthotropic plates subjected
to combined loads in terms of a useful nondimensional
acteristics and the amenability of these plates to struc-
tural tailoring provide symmetrically laminated plates design parameter and to identify the effects of flexural
 anisotropy
on thebucklingbehaviorof long symmetri- compression, shear, and ir_-plane
Kxly = 6 = 0,
cally laminated plates subjected to combined loads in a
bending buckling coefficients
more explicit manner than has been given in references 1
Ks h' = 6 = 0, defined by equations (18), (20),
and 2. This second objective is accomplished by model- and (21), respectively, in which
ing various combined loads as a primary system of desta-
Kbly = 8 = o anisotropy is neglected in the
bilizing loads with a secondary system of subcritical analysis
loads instead of using the traditional approach in which
each component of a combined loading is treated as a
destabilizing load with a fixed relative magnitude. This L1, L2, L 3, L4 nondimensional load factors
approach permits the effects of flexural orthotropy, flex- defined by equations (14)
ural anisotropy, and combined loading characteristics on through (17), respectively
C
plate-buckling behavior to be obtained and presented in a
nxl, nyl, nxyl, nbl nondimensional membrane
direct manner. The primary destabilizing loads that are
stress resultants of system of
considered consist of uniform axial compression, shear,
destabilizing loads defined by
and pure in-plane bending loads; the secondary subcriti-
equations (10) through (13),
cal loads that are considered consist of transverse tension
respectively
or compression and shear loads. Results are presented for C

plates with the two opposite long edges clamped or sim- nx2, ny2, nxy2, nb2 nondimensional membrane
ply supported. A number of generic buckling curves that stress resultants of system .3f sub-
are applicable to a wide range of laminate constructions critical loads defined by eq_Jations
are also presented using the nondimensional parameters (10) through (13), respecti,,ely
described in references l, 2, and 5. N number of terms in series
representation of out-of-phme
Symbols
displacement field at buckling
Am, Bm displacement amplitudes (see (see eq. (22))
eq. (22)), in.
intensity of eccentric in-pie-he
b bending load distribution defined
plate width, (see fig. 1), in.
by equation (5), Ib/in.
Dtb DI2, D22, D66 orthotropic plate-bending Nxc intensity of constant-valuexl ten-
stiffnesses, in-lb
sion or compression load distribu-
D16, D26 anisotropic plate-bending tion defined by equation (5), Ib/in.
stiffnesses, in-lb N ,G,N. longitudinal, transverse, and
shear membrane stress resultants,
El, E 2, GI2 lamina moduli, psi
respectively (see eqs. (5), ('7),
K b =- (nbl)c r nondimensional buckling coeffi- and (8)), lb/in.
cient associated with critical value N c
xi ' Nyl, Nxyl, Nbl membrane stress resultants of
of an eccentric in-plane bending system of destabilizing loacs
load (see eq. (21) and fig. l(a)) (see eqs. (6) through (9)), lb/in.

Ks = (nxyl)cr nondimensional buckling coeffi- N C

x2 "Ny2, Nxy2, Nb2 membrane stress resultants of
cient associated with critical system of subcritical loads
value of a uniform shear load (see eqs. (6) through (9)), lb/in.
(see eq. (20) and fig. 1(a))
P' Pcr nondimensional loading parame-
K x =-(nCxl)cr nondimensional buckling coeffi- ter (see eqs. (14) through (17))
cient associated with critical value and corresponding value at buck-
of a uniform axial compression ling (see eqs. (18) through (:21 )),
load (see eq. (18) and fig. 1(a)) respectively

Ky =--(nyl)cr nondimensional buckling coeffi- out-of-plane displacement

cient associated with critical field at buckling defined by
value of a uniform transverse equation (22), in.
compression load (see eq. (19) x,y plate coordinate system
and fig. l(a)) (see fig. 1), in.
 nondimensional parameters
D16 (3)
defined by equations (1), (2),
,y- (D311D22)I/4
(3), and (4), respectively

E0,E1 symbols that define distribution of

in-plane bending load (see fig. 1 - D26 (4)
3 1/4
and eq. (5))
(D11D22)
nondimensional plate coordinates
where b is the plate width and _. is the half-wavelength of
half-wavelength of buckling the buckle pattern of an infinitely long plate (fig. 1). The
mode (see fig. 1), in. subscripted D-terms are the bending stiffnesses of classi-
cal laminated plate theory. The parameters ot.o and [_
k/b buckle aspect ratio (see fig. 1)
characterize the flexural orthotropy, and the parameters _,
V12 lamina major Poisson's ratio and _ characterize the flexural anisotropy.

_m(rl) basis functions used to represent The loading combinations included in the analysis
buckling mode (see eq. (22)) are uniform biaxial tension and compression, uniform
shear, and eccentric in-plane bending, as depicted in fig-
ure 1. The longitudinal stress resultant N x is partitioned
Analysis Description
in the analysis into a uniform tension or compression part
Often in preparing generic design charts for buckling and a linearly varying part that correspond to eccentric
of a single flat plate, a special purpose analysis is pre- in-plane bending loads. This partitioning is given by
ferred over a general purpose analysis code, such as a
finite element code, because of the cost and effort usually Nx = Nxc - Nb[e 0 + (E 1 - 0)'q] (5)
involved in generating a large number of results with a
general purpose code. The results presented herein were where Nxc denotes the intensity of the constant-valued
obtained using such a special purpose analysis. The anal- tension or compression part of the load, and the term
ysis details are lengthy, and only a brief description of containing N b defines the intensity of the eccentric in-
the analysis is presented. plane bending load distribution. The symbols E0 and e l
define the distribution of the in-plane bending load, and
Symmetrically laminated plates can have many dif- the symbol rl is the nondimensional coordinate given by
ferent constructions because of the variety of material r I =fib(fig. 1).
systems, fiber orientations, and stacking sequences that
can be used to construct a laminate. A way of coping The analysis is based on a general formulation that
includes combined destabilizing loads that are propor-
with the vast diversity of laminate constructions is to use
tional to a positive-valued loading parameter _ that is
convenient nondimensional parameters. The buckling
increased monotonically until buckling occurs and inde-
analysis that is used in the present paper is based on clas-
pendent subcritical combined loads that remain fixed at a
sical plate theory and the classical Rayleigh-Ritz method
specified load level below the value of the buckling load.
and is derived explicitly in terms of the nondimensional
Herein, use of the term, "subcritical load," includes
parameters defined in references 1, 2, and 5. This
loads, such as a transverse tension load, that do not cause
approach is an effective method of conducting generic
buckling to occur. In practice the subcritical loads are
in-depth parametric studies of buckling behavior in terms
of the minimum number of independent parameters applied to a plate first with an intensity that is below the
needed to characterize fully the buckling behavior and to intensity that will cause the plate to buckle. With the sub-
critical loads fixed, the destabilizing loads are applied by
obtain results that indicate overall trends and the sensitiv-
increasing the magnitude of the loading parameter until
ity of the results to changes in the parameters. The non-
buckling occurs. This approach permits combined-load
dimensional parameters used in the present paper are
interaction to be investigated in a direct and convenient
given by
manner.

b(Dll) 1/4 The distinction between the destabilizing and sub-

(1)
critical loading systems is implemented in the buckling
analysis by partitioning the prebuckling stress resultants
as follows:
D12 + 2D66 (2)
= 1/2 c C

Nxc = - Nxl + Nx2 (6)

(DllD22)
 Ny = -Ny 1 + Ny 2 (7) Nondimensional buckling coefficients are given by
the values of the dimensionless stress resultants of the
Nxy = Nxyi + Nxy 2 (8) system of destabilizing loads at the onset of buckling;
that is,
Nb = Nbl + Nb2 (9)
c .N c . .2
where the stress resultants with the subscript 1 constitute Kx=_(nxl)cr= ( xl)cr O (18)
2(DIID22)I/2 = LlPcr
the destabilizing loads, and those with the subscript 2
constitute the subcritical loads. The sign convention that
is used for positive values of these stress resultants is (Nyi)cr b2
shown in figure 1. The normal stress resultants of the K v = (nyi)cr - _ L2_cr (19)
system of destabilizing loads, N xl
c and N., " _2D22
yl are defined
to be positive valued for compression loads. This con-
vention results in positive eigenvalues being used to indi-
(Nxy I )cr b2
cate instability caused by compression loads.
Ks=(nxyl)cr = 3 1/4 = L3Pcr (20)
The buckling analysis includes several nondimen- x2(DI1D22)
sional stress resultants that are associated with equa-
tions (6) through (9). These dimensionless stress result- (Nb 1)cr b2
ants are given by (21)
K b - (nbl)c r n2(DIID22 )1/2 = LaPcr
c N cxj b. 2
(10) where Pcr is the magnitude of the loading parameter at
nxj = n2(D11D22)l/2 buckling. Positive values of the coefficients K x and Ky
correspond to uniform compression loads, and the coeffi-
cient Ks corresponds to uniform positive shear. The
(11) direction of a positive shear-stress resultant actir_g on a
plate is shown in figure 1. The coefficient K b corre-
sponds to the specific in-plane bending-load distribution
that is defined by the selected values of the parameters
Nx yjb 2 and E1.
nxyj = 2 3 1/4 (12)
n (DI1D22) The mathematical expression that is used in _:e vari-
ational analysis to represent the general off-centered and
skewed buckle pattern is given by
Nbjb 2
(13) N
nbj = 1t2(DIID22)l/2
WN(_' rl) = Z (Amsin x_ +BmcoS rc_)_m(rl) (22)
where the subscript j takes on the values of 1 and 2. In m=l

addition, the destabilizing loads are expressed in terms of where _ = xfL and 1] = ylb are nondimensional coordi-
the loading parameter _ in the analysis by
nates, w N is the out-of-plane displacement field, and A m
C and B m are the unknown displacement amplituc.:es. In
nxl = LlP (14) accordance with the Rayleigh-Ritz method, the basis
functions _m(rl) are required to satisfy the kinematic
ny 1 = L2/_ (15) boundary conditions on the plate edges at r I = 0 and
rl = 1. For the simply supported plates the basis functions
nxy I = L3P (16) used in the analysis are given by

nbl = L4P (17) _m(_q) = sinImxrl] (23)

where L 1 through L 4 are load factors that determine the for values ofm = 1, 2, 3 ..... N. Similarly, for the clamped
specific form (relative magnitude of the load compo- plates, the basis functions are given by
nents) of a given system of destabilizing loads. Typi-
_m('q) = cosI(m- 1)nrl]-cos[(m+ l)nrl] (24)
cally, the dominant load factor is assigned a value of l
and all other factors are given as positive or negative Algebraic equations that govern the buckling behav-
fractions.
ior of infinitely long plates are obtained by substituting
theseriesexpansion for thebucklingmodethatis given of parameters that is applicable to a broad class of lami-
nate constructions. The range of each nondimensional
by equation (22)into thesecondvariationof thetotal
parameter used herein is given by 0.1 <[t<3.0,
potentialenergyandthenby computing the integrals
0 < _, < 0.6, and 0 < 8 < 0.6. The results presented in ref-
appearing in the secondvariationin closedform.The
erences 1 and 2 indicate that a value of 0.6 for "y and
resultingequations constitutea generalized eigenvalue
problemthatdepends on theaspectratioof thebuckle corresponds to a highly anisotropic plate. For isotropic
patternX/b(fig. 1) andthenondimensional parameters plates, [3 = 1 and T = _5= 0. Moreover, for plates without
andnondimensional stressresultants thatare defined flexural anisotropy, T = _ = 0. Values of these parameters
herein.Thesmallesteigenvalue of theproblemcorre- that correspond to several practical laminates are given in
references 1 and 2.
spondsto bucklingandis foundby specifying a value
ofX/bandsolvingthecorresponding generalized eigen- To simplify the presentation of the fundamental
valueproblem foritssmallest eigenvalue.Thisprocess is generic behavioral trends, results are presented only for
repeatedfor successive valuesof L/b until the overall plates in which Y and 8 have equal values (e.g., [+45] s
smallest eigenvalue is found. laminates). However, these behavioral trends are applica-
Results that are obtained by using the analysis ble to laminates that have nearly equal values of T and 8
described herein have been compared with other results such as a [+35/-15] s laminate made of the typical
for isotropic, orthotropic, and anisotropic plates that have graphite-epoxy material described herein. For this
been obtained by other analysis methods. These compari- laminate _=1.95, T=0.52, and _5=0.51. Further-
sons are discussed in references 1 and 2; in every case the more, results showing the effects of _x_, or equivalently
analysis described herein was found to be in good agree- (D 11/D22) TM, on the buckling coefficients are not pre-
ment with the results obtained from other analyses. sented since it has been shown in references 1 and 2 that
variations in this parameter only affect the critical value
Results and Discussion of the buckle aspect ratio k/b and not the buckling coeffi-
cient. (The buckling coefficient remains constant.) A
Results are presented for clamped and simply sup- value of (D I 1[D22) TM = 1 was used in all the calculations
ported plates that are loaded by various combinations of presented herein. For clarity the compression, shear, and
axial compression, transverse tension or compression, in-plane-bending buckling coefficients, defined by equa-
pure in-plane bending, and shear. For loading cases that tions (18), (20), and (21), respectively, are expressed
involve shear, a distinction is made between positive and
as Kx[,t=8= 0, Ks].=_=0, Kb[_,=8=0 when
negative shear loads whenever flexural anisotropy is generic results are desbribed for which flexural aniso-
present. A positive shear load corresponds to the shear tropy is neglected in the analysis.
loads shown in figure 1. No distinction between positive
and negative pure in-plane bending loads is necessary for Importance of Anisotropy on Behavior of [+45] s
flexurally anisotropic plates. Plates
Results are presented first for [+45] s flexurally Results are presented in figure 2 for a clamped
anisotropic plates. The importance of flexural anisotropy [+45]s plate subjected to a destabilizing uniform uniaxial
on the buckling behavior of a [+45] s plate subjected to compression load. In addition, results are presented for a
uniform uniaxial compression, pure shear, or pure in- combined-loading condition that consists of uniform
plane bending loads is well-known and is documented in axial compression and either a uniform subcritical trans-
references 1 and 2. Additional results for this laminate verse tension load or a compression load. Similar results
are presented herein that show that the importance of are presented in figure 3 for a corresponding plate in
flexural anisotropy on the buckling behavior depends on which the destabilizing load is a uniform shear load.
the type of combined loading. This thin laminate is repre- In these figures the minimum value of the loading
sentative of spacecraft structural components and is parameter _ found by solving the generalized eigen-
made of a typical graphite-epoxy material with a longitu- value problem for a given value of L/b is shown for val-
dinal modulus E l = 127.8 GPa (18.5 106 psi), a trans- ues of 0 < L/b < 2. Moreover, the magnitude of the
verse modulus E 2 = 11.0 GPa (1.6 x 106 psi), an in-plane subcritical transverse load is indicated in the figures by
shear modulus G12 = 5.7 GPa (0.832 x 106 psi), a major the value of the nondimensional stress resultant ny 2 that
Poisson's ratio v12 = 0.35, and a nominal ply thickness of is defined by equation (11). A limiting value of the sub-
0.127mm (0.005 in.). The fiber orientation angle of a
critical load given by ny 2 =---4 corresponds to a wide-
lamina is measured from the plate x-axis to the lamina column buckling mode. The dashed lines shown in the
major principal axis denoted by the subscript 1. figures correspond to the solutions of the generalized
Generic results are presented next, in terms of the eigenvalue problem that are obtained when flexural
nondimensional parameters described herein, for a range anisotropy is included in the analysis. In contrast, the

5
 solidlinescorrespond
to thesolutions that are obtained
Results that are similar to the results presented in
when flexural anisotropy is neglected in the analysis (i.e., figures 2 and 3 are presented in figures 4 and 5 for a
DI6 = D26 = 0). The overall minimum value of the load-
clamped [+45]s plate that is loaded by a subcritical shear
ing parameter for each curve corresponding to ny2 = 0 is load and either a destabilizing uniform axial comaression
indicated by an unfilled circle, and the minimum for each
load or a pure in-plane bending load, respectively. Two
curve that corresponds to nonzero values of ny 2 is indi- groups of curves are shown in the figures that correspond
cated by a filled circle. These minimum values of the
to plates without a subcritical load and with eithe- a posi-
loading parameter correspond to the value of the buck-
tive or negative subcritical shear load with a m_gnitude
ling coefficient for each curve. The corresponding values equal to 75 percent of the corresponding shear buckling
of L/b are the critical values of the buckle aspect ratio. coefficient K s (eq. (20)). For the cases in which the

The results presented in figure 2 indicate that the anisotropy is included in the analysis, the buckling coef-
buckling coefficient (minimum loading parameter) of a ficients for positive and negative shear loads are _:iven by
K s = 6.12 and K s = -17.16, respectively. For the cases in
plate that is loaded by pure compression only (ny 2 = 0) is
overestimated by approximately 30 percent of the aniso- which the anisotropy is neglected, the buckling coeffi-
tropic buckling load and that the critical value of the cients for positive and negative shear loads are given by
buckle aspect ratio _b is slightly overestimated if the Ks= 12.03 and Ks=-12.03 , respectively. The results
shown in figures 4 and 5 therefore include the effects of
flexural anisotropy is neglected in the buckling analysis.
neglecting the anisotropy in the calculation of tne sub-
For a subcritical transverse tension load given by
ny 2 = 3.5 (87.5 percent of the magnitude for a wide- critical load nxy 2 and the actual calculation of the buck-
column buckling mode), the results show that the effect ling coefficient.

of the flexural anisotropy becomes slightly less impor- The results shown in figures 4 and 5 indicate that
tant; that is, the buckling coefficient is overestimated by neglecting the flexural anisotropy in a buckling _alysis
22 percent for this case. Neglecting the anisotropy for of the plate that is subjected only to uniform axial com-
this case also yields results that slightly overestimate the
pression or pure in-plane bending (nxy 2 = 0) yields a
critical value of the buckle aspect ratio. However, for
solution that overestimates the buckling coeffic,_ent by
a subcritical transverse compression load given by
approximately 30 percent of the anisotropic buckling
ny 2 =-3.5, the buckling coefficient is overestimated by load, and the critical value of the buckle aspect ratio is
76 percent when the flexural anisotropy is neglected in slightly overestimated. For the subcritical positiv. _ shear
the analysis, and the critical value of the buckle aspect load with nxy 2 = 0.75Ks, the results predict that the buck-
ratio is slightly underestimated.
ling coefficient is overestimated by approximately
The results shown in figure 3 for a shear-loaded plate 84 percent and 66 percent of the anisotropic buckling
with a subcritical transverse tension or compression load coefficient for the uniform axial compression loads and
and the results shown in figure 2 indicate that the effect the pure in-plane bending loads, respectively. Moreover,
for both loading conditions, the critical value of the
of neglecting the flexural anisotropy in the buckling anal-
ysis is much more pronounced for the shear-loaded plate buckle aspect ratio is also slightly overestimated when
than for the corresponding compression-loaded plate. In the anisotropy is neglected in the analysis. In contr._st, for
particular, the results presented in figure 3 indicate that the subcritical negative shear load with nxy 2 = 0.75Ks,
results that are obtained by neglecting the fiexural aniso- the results predict that the buckling coefficient is under-
tropy in the buckling analysis of a plate that is loaded by estimated by approximately 26 percent and 16 per,;ent of
the anisotropic buckling coefficient for the uniform axial
shear only (ny 2 = 0) overestimate the buckling coefficient
by approximately 97 percent of the anisotropic buckling compression and pure in-plane bending loading condi-
tions, respectively. In addition, the critical value of the
load compared with 30 percent for the corresponding
compression-loaded plate. For a subcritical transverse buckle aspect ratio is slightly underestimated for tl_e uni-

tension load given by ny 2 = 3.5, the results for the shear- form axial compression load and slightly overestimated
loaded plate predict that the buckling coefficient is over- for the pure in-plane bending load when the flexural
estimated by approximately 75 percent if flexural aniso- anisotropy is neglected in the analysis.
tropy is neglected. More significantly, for a subcritical
Generic Results for Specially Orthotropic Plates
transverse compression load given by ny2 =-3.5, the
results predict that the shear buckling coefficient is over-
A broad range of buckling results for specially ortho-
estimated by approximately 209 percent when the flex- tropic plates that are subjected to combined loacs and
ural anisotropy is neglected in the analysis. For each
that have either clamped or simply supported edges are
subcritical loading case, the critical values of the buckle presented in this section. These results are included to
aspect ratio are overestimated when the anisotropy is demonstrate the effects of flexural orthotropy on plate-
neglected in the buckling analysis. buckling behavior in a general manner and to plovide
designcurvesthatcanbeusedin conjunction withthe presented for plates that are subjected to a subcritical
transverse tension or compression load and destabilizing
resultspresented in thenextsectionto obtainbuckling
coefficients
of flexurally anisotropic plates. uniform axial compression, pure in-plane bending, and
shear loads, respectively. Then, results are presented for
Results are presented in figures 6 through 10 that plates subjected to a subcritical shear load and to destabi-
show the genetic effects of plate flexural orthotropy on lizing uniform uniaxial compression and pure in-plane
the buckling coefficients of clamped and simply sup- bending loads, respectively.
ported plates with T = _--0. The results in figures 6
through 8 are for plates that are subjected to a subcritical Plates subjected to subcritical transverse tension or
transverse tension or compression load and a destabiliz- compression loads. Results are presented in figures 11
ing uniform axial compression load, pure in-plane bend- and 12, 13 and 14, and 15 and 16, respectively, for plates
ing load, and shear load, respectively. Similarly, the subjected to a subcritical transverse tension or compres-
generic effects of flexural orthotropy on plates that are sion load and destabilizing uniform axial compression,
subjected to a subcritical shear load and a uniform axial pure in-plane bending, and shear loads. In figures 11
compression load or to a pure in-plane bending load are through 16, the ratio of the anisotropic buckling coeffi-
shown in figures 9 and 10, respectively. The results in cient to the corresponding orthotropic buckling coeffi-
figures 9 and 10 are also applicable to plates that are cient that is computed with T = 5 = 0 (figs. 6 through 10)
loaded in negative shear since the shear buckling coeffi- is given as a function of the orthotropic parameter 13 for
cients for specially orthotropic plates that are loaded in equal values of the anisotropic parameters (T = _5) rang-
positive or negative shear have the same magnitude. Two ing from 0.1 to 0.6. For each value ofT= 5 that is given
sets of curves are shown in figures 6 through 10 for val-
in figures 11 through 16, three curves are presented. The
ues of the orthotropic parameter 13= 0.5, 1, 1.5, 2, 2.5, solid lines correspond to a value of the nondimensional
and 3. The solid curves correspond to results for clamped
transverse load ny 2 = 0 that is used in the calculations (no
plates and the dashed curves are for simply supported subcritical load). Similarly, the finely dashed and
plates. These curves show the buckling coefficient as a coarsely dashed lines correspond to values of ny 2 = -0.5
function of the nondimensional transverse load ny 2 in fig- (compression) and 0.5 (tension), respectively, for the
ures 6 through 8 and as a function of the nondimensional simply supported plates, and to values of ny 2 = -2 (com-
shear load nxy 2 in figures 9 and 10. pression) and 2 (tension), respectively, for the clamped
plates. The magnitudes of these values correspond to
The results presented in figures 6 through 10 indicate
that the buckling coefficients of specially orthotropic 50 percent of Ky, the buckling coefficient that corre-
sponds to a wide-column buckling mode (Ky = 1 and 4
plates (T = _ = 0) increase substantially as the orthotropic
for simply supported and clamped plates, respectively).
parameter 13 increases. Furthermore, the results in fig-
ures 6 through 8 indicate that as the subcritical transverse
The results that are presented in figures 11 and 12,
load increases through positive values (increasing ten-
respectively, for simply supported and clamped plates
sion), the buckling coefficients increase substantially.
subjected to a subcritical transverse tension or compres-
This trend is shown to be more pronounced for the shear-
sion load and to destabilizing uniform uniaxial compres-
loaded plates than for the compression-loaded plates or
sion show that the anisotropic buckling coefficient is
for the plates that are loaded by pure in-plane bending.
always less than the corresponding orthotropic buckling
Moreover, the increase in buckling coefficient with
coefficient for the full range of parameters considered. In
increasing subcritical tension load is predicted to be
addition, these results predict that the effects of neglect-
slightly more pronounced for the simply supported plates
ing anisotropy are more pronounced for the plates with
than for the clamped plates. The results presented in fig-
the transverse compression loads than for the plates with-
ures 9 and 10 for the plates with subcritical shear loads
out a subcritical load (ny 2 = 0) or a transverse tension
indicate that the buckling coefficient is substantially
load. Moreover, these results predict that this trend is
reduced as the magnitude of the nondimensional shear
slightly more pronounced for the clamped plates than for
load increases. In addition, the results also predict this
the simply supported plates, as indicated by the increase
trend to be slightly more pronounced for the simply sup-
in the relative spacing of the three curves that correspond
ported plates than for the clamped plates.
to each fixed value of T and 5. The results also predict
that the reduction in buckling coefficient caused by
Interaction of Flexural Anisotropy and Loading
anisotropy is generally larger for the clamped plates with
Results are presented in this section that show the ny 2 =-2 (transverse compression at a magnitude of
50 percent of the wide-column buckling coefficient) than
generic effects of plate flexural anisotropy and combined
loading condition on the buckling coefficients of for the corresponding simply supported plates. This trend
is indicated by the observation that the curve for the
clamped and simply supported plates. First, results are
 clampedplatesgenerallyliesbelowthecorresponding bilizing pure in-plane bending loads. Moreover, these
curveforthesimplysupported platesforeachfixedvalue results predict that this trend is also slightly more pro-
of T and_i.Basedona similarobservation, theresults nounced for clamped plates than for simply sttpported
alsopredict,however,thatthe reductionin buckling plates. Furthermore, like the plates with the uniaxial
coefficient
caused byanisotropy is generally smallerfor compression loads, the results predict that the rcxluction
theclamped plateswith ny 2 = 0 or 2 (transverse tension in buckling coefficient caused by anisotropy is generally
at a magnitude of 50 percent of the wide-column buck- larger for clamped plates loaded in positive shear and
ling coefficient) than for the corresponding simply sup- with ny 2 = -2 (transverse compression) than for the cor-
ported plates.
responding simply supported plates. Similarly, the results
also predict that the reduction in buckling coefficient
The results presented in figures 13 and 14 for simply
supported and clamped plates, respectively, that are sub- caused by anisotropy is generally smaller for the damped
jected to a subcritical transverse tension or compression plates loaded in positive shear and with ny 2 = 0 or 2
load and a destabilizing pure in-plane bending load, also (transverse tension) than for the corresponding simply
show that the anisotropic buckling coefficient is always supported plates. For the plates loaded by negative shear,
less than the corresponding orthotropic buckling coeffi- these trends are reversed. Overall, the results presented in
cient for the full range of parameters being considered. figures 11 through 16 indicate that plates loaded by posi-
These results also predict that the effects of neglecting tive shear typically exhibit substantially larger reductions
anisotropy are more pronounced for plates with a trans- in the buckling coefficient when flexural anisotropy is
included in the analysis than do the corresponding plates
verse compression load than for plates with ny 2 = 0 or the
transverse tension load, but not to the extent that is loaded by uniform axial compression or pure in-plane
exhibited by the corresponding plates with the uniaxial bending. In contrast, the results for plates that are loaded
compression load. (Compare the difference in the rela- by negative shear predict sizable increases in the buck-

tive spacing of the three curves that correspond to each ling coefficients when the flexural anisotropy is included
in the calculations.
fixed value of 2( and 8 that is shown in figures 11
through 14.) Moreover, these results predict that this
trend is slightly more pronounced for clamped plates Plates subjected to subcritical shear load. Results
than for simply supported plates, but also not to the are presented in figures 17 through 20 and in figures 21
extent that is predicted for the corresponding plates with through 24 for plates that are subjected to a subcritical
a uniaxial compression load. Furthermore, like the plates shear load and to destabilizing uniform axial compres-
with a uniaxial compression load, the results also predict sion and pure in-plane bending, respectively. Moreover,
that the reduction in buckling coefficient caused by the results in figures 17, 18, 21, and 22 are for a positive
anisotropy is generally larger for clamped plates with subcritical shear load, and the results in figures 19, 20,
ny 2 =-2 (transverse compression) than for the corre- 23, and 24 are for a negative subcritical shear load. In
sponding simply supported plates. Similarly, the results figures 17 through 24, the ratio of the anisotropic buck-
also predict that the reduction in buckling coefficient ling coefficient to the corresponding orthotropic buckling
caused by anisotropy is generally smaller for clamped coefficient computed with 2(= 8 = 0 (figs. 6 through 10)
plates with ny 2 = 0 or 2 (transverse tension) than it is for is given as a function of the orthotropic parameter 13for
the corresponding simply supported plates. equal values of the anisotropic parameters (2(= 81_rang-
ing from O. 1 to 0.6. For each value of 2(= 8 that i., given
The results for simply supported plates subjected to a
in figures 17 through 24, four curves are presented. The
suberitical transverse tension or compression load and a
solid lines correspond to the value of the nondimensional
positive destabilizing shear load presented in figure 15
shear load nxy 2 = 0 (no subcritical load). In addition, the
and corresponding results for clamped plates presented in
finely dashed, moderately dashed, and coarsely dashed
figure 16 also show that the anisotropic buckling coeffi-
lines correspond to values of nxy2=O.25Ks, 0.5Ks,
cient is always less than the corresponding orthotropic
and 0.8Ks, respectively, where K s is the shear buckling
buckling coefficient. However, these results also show
coefficient. More specifically, Ks is the positive shear
that this trend is reversed for negative shear loads. These
buckling coefficient in figures 17, 18, 21, and 22, and the
results also predict that the effects of neglecting aniso-
negative shear buckling coefficient in figures 19, 20, 23,
tropy are generally more pronounced for plates with
and 24. For each case, the shear buckling coefficient is a
transverse compression loads than they are for plates function of 13,2(, and 8.
with ny 2 = 0 or with transverse tension loads. This trend
for the shear-loaded plates is slightly less pronounced Points on the curves shown in figures 17 through 24
than it is for the corresponding plates with the desta- are obtained by calculating the shear buckling coefT]cient
bilizing uniaxial compression loads and slightly more
first, assigning the appropriate subcritical value of nxy2,
pronounced for the corresponding plates with the desta- and then computing the buckling coefficient of the
 Moreover, these results show that the reductions in buck-
systemofdestabilizing loads.Sincethecalculation ofthe
shearbucklingcoefficient thatis usedin specifying the ling coefficients are generally more pronounced for the
subcriticalloadingmagnitude alsodepends on y and8, simply supported plates than for the clamped plates. This
theresultsshownin figures17through24represent the trend is also indicated by the observation that each curve
overalleffectof neglecting flexuralanisotropy in both for the clamped plates generally lies above the corre-
thecalculationof thesubcriticalshear
loadingmagnitude sponding curve for the simply supported plates for a
andthebucklingcoefficient of thesystemof destabiliz- fixed value of _' and 8. More importantly, these results
indicate that the reduction in buckling coefficient that is
ing loads.This computational procedure is described
associated with including anisotropy in the calculations
subsequently in detailby usingtheresultsshownin fig-
increases substantially as the magnitude of the positive
ures25and26forthecaseof nxy 2 = 0.8K s and for plates
subcritical shear load increases.
with [_ = 3 that are subjected to a subcritical shear load
and to uniform uniaxial compression and pure in-plane The results shown in figures 19 and 20 for simply
bending, respectively. supported and clamped plates, respectively, which are
Buckling interaction curves for plates that are sub- loaded by destabilizing uniform axial compression and a
negative subcritical shear load, predict monotonic reduc-
jected to a subcritical shear load and uniform uniaxial
tions in buckling coefficient with increasing values of the
compression and to a subcritical shear load and pure in-
anisotropic parameters for the values of nxy2=0
plane bending are shown in figures 25 and 26, respec-
and 0.25K s. However, for the larger values of nxy 2 shown
tively. The solid curve shown in each figure corresponds
to results for specially orthotropic plates with y = 8 = 0, in the figures, the results predict increases in buckling
and the dashed lines correspond to results for y = 8 = 0.6. coefficient with increasing values of the anisotropic
Furthermore, the unfilled and filled circles on the curves parameters, as anticipated from the previous results that
have been presented. This seemingly unusual trend for
correspond to results for nxy 2 = 0.8K s and for positive
the negative shear load is a manifestation of the horizon-
and negative shear loading, respectively. The ordi-
nates of the filled and unfilled circles on the curves tal shift in the compression-shear buckling interaction
curves caused by flexural anisotropy that has been
for T= 8 = 0 correspond to the buckling coefficients
reported in reference 1 and illustrated in figure 25. In
gxl - o and Kbly- 8- 0 that are indicated in fig-
addition, these results show this trend of increasing buck-
ure's 1-7 t_ough 20 afidqn figures 21 through 24, respec-
tively. Similarly, the ordinates of the filled and unfilled ling coefficient with increasing values of the anisotropic
circles on the curves for y = 8 = 0.6 correspond to the parameters to be, overall, slightly more pronounced for
buckling coefficients K x and K b that are also indicated in clamped plates than for simply supported plates. The
trend of a decreasing buckling coefficient with decreas-
figures 17 through 20 and in fig-ures 21 through 24,
respectively. The buckling coefficient ratios shown in ing values of the anisotropic parameters is predicted,
overall, to be slightly more pronounced for simply sup-
figures 17 through 24 for the case nxy2=O.8Ks are
obtained by dividing the value of the ordinate on the ported plates than for clamped plates. Furthermore, these
curve for y = 8 = 0.6 (in figs. 25 and 26) by the corre- results indicate that the importance of anisotropy on the
buckling resistance changes dramatically with the magni-
sponding value of the ordinate for the curve for y = 8 = 0
for either the filled or unfilled circles. For positive sub- tude of a negative subcritical shear load.
critical shear loading, the results in figures 25 and 26 Results similar to those presented in figures 17
indicate that the buckling coefficient ratios for plates and 18 are presented in figures 21 and 22 for simply sup-
subjected to uniform uniaxial compression and pure in- ported and clamped plates, respectively, that are loaded
plane bending have values less than 1. In contrast, for by pure in-plane bending and a positive subcritical shear
negative shear loading, the results in figures 25 and 26 load. These results also predict monotonic reductions in
indicate that the buckling coefficient ratios have values buckling coefficient with increasing values of the aniso-
greater than 1. This trend reversal is caused by the hori- tropic parameters and increasing magnitudes of a posi-
zontal shift in the buckling interaction curves that occurs tive subcritical shear load. However, these results show
when anisotropy is present. that the reductions in buckling coefficient are practically
The results presented in figures 17 and 18 for simply the same for the simply supported and clamped plates.
The results also indicate that the reduction in buckling
supported and clamped plates, respectively, loaded by
uniform axial compression and a positive subcritical coefficient associated with having anisotropy included
in the calculations increases substantially as the magni-
shear load, predict monotonic reductions in buckling
coefficient with increasing values of the anisotropic tude of the positive subcritical shear load increases.
parameters. This same trend is also predicted for increas- Comparing the results in figures 17, 18, 21, and 22 sug-
ing magnitudes of the positive shear load as indicated by gests that the reductions in buckling coefficient that are
the four different line types shown in figures 17 and 18. caused by including flexural anisotropy in the analysis

9
 arenearly the same for the plates loaded by uniform axial
the plates loaded by pure in-plane bending. Moreover,
compression and pure in-plane bending with nxy 2 = 0. the results indicate that the curves for plates with nonzero
However, the comparison also suggests that the reduc- subcritical shear loads, which show decreases in buck-
tions in buckling coefficient that are caused by including ling coefficient with increasing values of the anisotropic
flexural anisotropy in the analysis are more pronounced
parameters, also show that the decreases in buckling
for plates that are loaded by uniform axial compression coefficient are, overall, much more pronounced for the
than for those that are loaded by pure in-plane bending plates loaded by pure in-plane bending than for the
when a nonzero subcritical shear loading is present, espe- compression-loaded plates.
cially for the larger magnitudes of a subcritical shear
load. This trend is also indicated by the observation that
Concluding Remarks
each curve for the compression-loaded plates generally
lies below the corresponding curve for the plates that are A parametric study has been presented of the buck-
subjected to pure in-plane bending, for a fixed value of T ling behavior of infinitely long symmetrically lanainated
and 5.
anisotropic plates subjected to combined loads. A special
purpose nondimensional analysis that is well suited for
Results similar to those presented in figures 19 parametric studies of clamped and simply supported
and 20 are shown in figures 23 and 24 for simply sup- plates has been described, and its key features have been
ported and clamped plates, respectively, that are loaded discussed. The results presented herein have focused on
by pure in-plane bending and a negative subcritical shear the interaction of a subcritical (stable) secondary loading
load. These results also predict monotonic reductions state and a primary destabilizing loading state. The inter-
in buckling coefficient with increasing values of the action of uniform axial compression, pure in-plane bend-
anisotropic parameters for the values of nxy 2 = 0 and ing, transverse tension and compression, and shear loads
0.25K s. Unlike the results for compression-loaded plates, with plate flexural anisotropy and orthotropy have been
the results in figures 23 and 24 predict monotonic reduc- examined. In particular, results have been presented
tions in buckling coefficient with increasing values for [+45]s thin graphite-epoxy laminates that are repre-
of the anisotropic parameters for nxy2=O.5K s. For sentative of spacecraft structural components. In addi-
nxy2 = 0.8K s, the results predict increases in buckling tion, a number of generic buckling results have been
coefficient with increasing values of the anisotropic presented that are applicable to a broad class of laminate
parameters. This unusual trend for a negative shear load constructions and show explicitly the effects of tlexural
is also a manifestation of the horizontal shift in the pure orthotropy and flexural anisotropy on plate-buckling
in-plane bending-shear buckling interaction curves behavior under combined loads.
caused by flexural anisotropy that has been reported in
reference 2 and illustrated in figure 26. Moreover, the The most significant finding of the present study is
difference between the trends for plates that are loaded that the importance of flexural anisotropy on the buck-
ling behavior of a long plate is strongly dependent on the
by negative subcritical shear and pure in-plane bending
and by negative subcritical shear and uniform axial com- type and magnitude of the combined load that is applied.
pression is manifested by the pronounced difference in Specifically, the results presented show that significant
errors (on the order of 100 percent) can be made when
the shape of the corresponding buckling interaction
curves shown in figures 25 and 26. (See also refs. 1 calculating buckling coefficients of plates subjected to
and 2.) For the curves in figures 23 and 24 that show combined loads if flexural anisotropy is neglected. Over-
increases in buckling coefficient with increasing values all, the results presented herein show that the buckling
of the anisotropic parameters, the results predict that the coefficients increase significantly as the orthotropic
increases in buckling coefficient are generally more pro- parameter _i - D12 + 2D66 increases, and that they
nounced for the clamped plates than for the simply sup-
(DI1D22) 1/2
ported plates. In contrast, for the curves that show
decreases in buckling coefficient with increasing values decrease significantly as the anisotropic parameters
of the anisotropic parameters, the results predict that the D
16 D26
decreases in buckling coefficient are slightly more pro- ]t- 3 1/4 and 6- 1/4 with equal
nounced for the simply supported plates than for the (DIlD22) (DIID_2)
clamped plates. Comparing the results in figures 19, 20, values (T = 8) increase for all load combinations that are
23, and 24 indicates that the curves in the figures, which considered, except those involving negative shear loads.
show increases in buckling coefficient with increasing For the negative shear loading condition, the trend is
values of the anisotropic parameters, also show that the
generally reversed. The results that are presented herein
increases in buckling coefficient are, overall, much more
also show, in many cases, that the effects of plate ,miso-
pronounced for the compression-loaded plates than for tropy are more pronounced for clamped plates th;m for
10
 3. Zeggane, Madjid; and Sridharan, Srinivasan: Buckling Under
simply supported plates that are subjected to combined
Combined Loading of Shear Deformable Laminated Anisolxo-
loads.
pic Plates Using "Infinite" Strips. Int. J. Num. Meth. Eng.,

NASA Langley Research Center vol. 31, May 1991, pp. 1319-1331.

Hampton, VA 23681-0001
September 22, 1995 4. Zeggane, M.; and Sridharan, S.: Stability Analysis of Long
Laminate Composite Plates Using Reissner-Mindlin Infinite
References Strips. Comp. & Struc., vol. 40, no. 4, 1991, pp. 1033-1042.

1. Nemeth, Michael P.: Buckling Behavior of Long Symmetrically

Laminated Plates Subjected to Combined Loadings. NASA 5. Nemeth, Michael P.: Importance of Anisotropy on Buckling of
TP-3195, May 1992. Compression-Loaded Symmetric Composite Plates. AIAA J.
vol. 24, Nov. 1986, pp. 1831-1835.
2. Nemeth, M. P.: Buckling of Symmetrically Laminated Plates
With Compression, Shear, and In-Plane Bending. AIAA J.,
vol. 30, no. 12, Dec. 1992, pp. 2959-2965.

11
 Y

E]Nb]

EONbl

lllllllltllll
(a) Primary (destabilizing) loading system.

Ny2

elNb2

_-_yvb2

(b) Secondary (subcritical) loading system.

Figure 1. Sign convention for positive-valued stress resultants (e o > E ! > 0 shown above).

12
 @
e_
o 0
-ei

e_
E
@

13
m

_J

+&

13
 =
< <

] I
I
l
e

It,---

)4
 "i::

o
_J

o3

+_

"C

:::s

o.o
m , m m m I m m ' m

:::s

II

15
 0
-ei

I
I
_
I iI
-- ,-2
I

--a

+_

E
2

I I I , I I !

8 0 0 0
t'N
N

e-, ,.-,
_ _ _

o,._

II

16
 e_

r_

II

II

_._

o,,=

_'_

_'_

17
 J

I I
I I I jJ
I I
I I I
I I I
I I I
"0
I I I
'I I I L)
I I

I
I I
I I I
II
I I I
oO
I I I
I I I II

| | I
I I I I, j_ "-;
I _=_
I
I
I I I I
I I I I 0
I I I I
I 0
I I I
I I I
0
I I I
I I I
I I I
I I I
I I I

t 0

=.
_O
?
II

= =
o.o

oo

r-]

C'4

J o_

l-r-I u

r.2 "_

18
0

!
!
!
II
!
I
I
I
I

19
 @

E
@

\
II
\
@
\
\
\
II

II

I
I %
I
I
I
I 0
I
I 0

II

=
@

c_
,g

tit1 @

0
u_

o_

2O
 \
\

.I.-.L

/ il

il

I
I
I
I
I
I
I
0
I
\ 0

/ o
0
Ii
r_
_o

Y
0
ca
I ...... _ d '0 z
0
II I R
I I _ ,

\ I I I

II, I
I
-! t
_

I I I
I I I
I I
Y I I I
I I I
0
I I I
I h II II , I i

o_

2]
 0
o

{J

"13
_J
+
II
I-L

I
.I

=
\ tJp

11 II
It _J

Q
o

= 0

N._.

_E

I:::u ,..

_ _._
_'_,

_'_

_g

L,

2.2
 .

E
0

ii

-I-
ll II II

II

0
%,
In E
0
. ,Iiii
II II

IIIiil I"

II

IIiiI
0
II
II
II

IIIIiiIIi

I . I .

"_-, ;_

_8
0"!_

o_

23
 O
i

II II
H

i
|

II II
oO II

II

q i

I I I . I
oo

24
 + i

II II II
i <M

II e_
I : E
I '
I

1
I i
I i - _I-_

gg ej

I:m
II e-"
II
i

II

E
II 0
00

"0
_J

e_
0

O
N.-.
--

e_ e_
_._

ae
13

_N
I . I . I
0

@
_N

4_

2S
 o

II

_ _ q q
q
li II
o

II

II II II

I :
II ll'
I I :f._

i i II

I i
2

I I I . I I . I . I . I . I I I I
o

I
I
_.._
o_

P2 _

_Z

26
 =
c_

II

c_

Q _ _ Q
c_
II II

+ I
il ii ii

I
I i
I II
o

r-

0.-_

i I

0 0
._'_
I i I i I li I . I , I . I . I . I . I 0

0 _

'_._

o,.._

27
 m
(,_

I::

(",I

+__ @
_.)

II

.,)

,.=

=,.

.._ ...-

_,m ,.c:

I I ,

I::

_,.-.

e-,

__._

28
0

"0

e_

r,

_',_

x_

_e
N'_

_Q

29
 0

0
i

_D

ii

II I::
II

+I_
8
r,J
II

,.o

0
(,0
f..i
,_
r_
C)

-- N

_._ _-.
0 (-I

g=

I I , I , I I , I , I , I I I _ :._
0

"r, _.
_'_
llll 0 "_

0 _
N'--

_,._

3O
 @

E
@
_J

{J

li II II II

E
=I

: , I
|

, , I
| , _j
: , I
|
: , I
@

t 1 m_

a
t _ a

V_

, I , I , I i I I i I , I I ,
e_
_N

llll- .r,

3)
 rr_

I:::
:f @

II
cO

li II il II

i . I
|
I . '-I'-

i . I 0

| |
: . I
II

\
@

II

II @

@ N
.o. _ E
,-.-

0 >,
.%

_'_
I I I | !
e'_ "::
0 0 ._ -_;
_:_ C-

o:e

_7
;

t'N _

32
 ..=

o
"0

"0

E
II II II II

' . I
' I +

, . I 0
I li
' . I
II
\

"0

o_

I I I . I . I

33
 m

It 1::
@
c_

o _ c d
I1 II II II

: , I
i ' I
| , e_
|
0
! . I
: . I
\ II

c-
O

e_
e_

I *

__

I I I . I , I I . I . I , I
e_._
0

.._ --_ 'k

34
 "0

-4-
0
0
q c._

II

c.}

..0

0
_0
.-_
.o
0

c._

4'

I . I . I .

e_',_
0
o'_.
I== ..0

._

35
 II
c,O
II

II

II

il

ii
t,O
\
II
"E:

l_.s.

II % @

II

I
il
qi |li ilii I_illl lii ill iI ii li li li li 11 li i_
@

I . I n . I , I ,

>

r,i

,,i @

e,_ ;

36
 m_
II

_
-r"

_g
i:m

4if if

1 I 5_
I

! i I
e_

_1
e_

1 I, ,i. I..., I .i..l ....I ..'. I' ' ''

1
0
-,_

%
II

_d

37
 REPORT DOCUMENTATION PAGE FormApproved
OMB No. 0704-0188

Public r_ng burden for this .,.u;_._i_n of information is e=ii_rT#ted tOaverage 1 hour per response, Inch_ling the time for reviswmg instructions, searching existing data SOurces
gathering and matntatntng the data needed, and completing and reviewing the collection of Inkxmation. Send comments regaling this burden estimate or any otl_er aspect of thil
collection of Info(metfon, Inclu_ng suggestions for reducing this burden, to Was_llngton Hea_luarters Servmes, Directorate for Information Operations and Reports, 1215 Jeffe_cor
Davis Highway, Suite 1204, Arlington, VA 22202-4302, ancl to the Office of Management and Budget, Paperwork Reduction Project (0704--0188), Weet_ington, DC 20503.
1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

November 1995 Technical Paper

4. TITLE AND SUcJ/i/-E
5. FUNDING NUMBERS
Buckling Behavior of Long Anisotropic Plates Subjected to Combined Loads
WU 505-63-50-08

6. AUTHOR(S)
Michael P. Nemeth

7. PERFORMR;G
ORGANIZATION
NAME(S)ANDADDRF-_S(ES) 8. PERFORMING ORGANIZATION
REPORT NUMBER

NASA Langley Research Center

Hampton, VA 23681-0001 L-17504

g. SPONSGR;NCdMONITORING
AGENCYNAME(S)ANDADDRESS(ES) 10. SPONSORING/MONITORING
AGENCY REPORT NUMBER

National Aeronautics and Space Administration

NASA TP-3568
Washington, DC 20546-0001

11. SUPPLEM.":.'FARY NOiES

Presented at the 36th AIAA/ASME/ASCEIASC/AHS Structures, Structural Dynamics, and Materials Conference,
April 10-12, 1995, New Orleans, LA.

12a. DISTRIBUTION/AVAILABILITY STATEMENT

12b. DISTRIBUTION CODE

Unclassified-Unlimited
Subject Category 39
Availability: NASA CASI (301) 621-0390

13. ABSTRACT (Maximum 200 words)

A parametric study is presented of the buckling behavior of infinitely long symmetrically laminated anisotropic
plates subjected to combined loads. The study focuses on the interaction of a subcritical (stable) secondary loadir_g
state of constant magnitude and a primary destabilizing load that is increased in magnitude until buckling occurs.
The loads considered in this report are uniform axial compression, pure in-plane bending, transverse tension arid
compression, and shear. Results are presented that were obtained by using a special purpose nondimensional anal,r-
sis that is well suited for parametric studies of clamped and simply supported plates. In particular, results are pr];-
sented for a [+45] s graphite-epoxy laminate that is highly anisotropic and representative of a laminate used fi)r
spacecraft applications. In addition, generic buckling-design charts are presented for a wide range of nondimen-
sional parameters that are applicable to a broad class of laminate constructions. These results show the general
behavioral trends of specially orthotropic plates and the effects of flexural anisotropy on plates subjected to various
combined loading conditions. An important finding of the present study is that the effects of flexural anisotropy cn
the buckling resistance of a plate can be significantly more important for plates subjected to combined loads than
for plates subjected to single-component loads.

14. SUBJECTTERMS
15. NUMBER OF PAGES
Buckling; Combined loads; Anisotropy; Flat plates; Composite plates 38
16. PRICE CODE

A03
17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION
OF REPORT OF THIS PAGE OF ABSTRACT OF ABSTRACT
Unclassified Unclassified Unclassified
NSN 7540o01-280-5500
Standard Form 298 (Rev. 2-89)
Prescribed by ANSI Std. Z39-18
298-102