Composite Structures: A.M. Zenkour, M. Sobhy
Composite Structures: A.M. Zenkour, M. Sobhy
Composite Structures: A.M. Zenkour, M. Sobhy
Composite Structures
journal homepage: www.elsevier.com/locate/compstruct
a r t i c l e i n f o a b s t r a c t
Article history: The sinusoidal shear deformation plate theory is used to study the thermal buckling of functionally
Available online 30 June 2010 graded material (FGM) sandwich plates. This theory includes the shear deformation and contains the
higher- and first-order shear deformation theories and classical plate theory as special cases. Material
Keywords: properties and thermal expansion coefficient of the sandwich plate faces are assumed to be graded in
Sinusoidal plate theory the thickness direction according to a simple power-law distribution in terms of the volume fractions
Thermal buckling of the constituents. The core layer is still homogeneous and made of an isotropic material. Several kinds
FGM sandwich plate
of symmetric sandwich plates are presented. Stability equations of FGM sandwich plates include the
thermal effects. The thermal loads are assumed to be uniform, linear and non-linear distribution
through-the-thickness. Numerical examples cover the effects of the gradient index, plate aspect ratio,
side-to-thickness ratio, loading type and sandwich plate type on the critical buckling for sandwich plates.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction ega [8] studied the behavior of uniformly heated plates for the
situation where a patch is adhered over a region of the structure.
Thermal buckling of composite structures has long been of vivid Sandwich plates have been extensively used in construction of
interest to researchers because it is an undesired phenomenon and many components in automotive, aerospace, aeronautic, ship,
some cases of it cannot be avoided under special conditions. Ther- underwater, and building structures. These plates are generally
mal buckling of plates, beams and shells was numerously studied fabricated from three layers. The two face sheets adhesively bonded
in several reports based on the various plate theories, depending to the core. They are utilized for many purposes, among them:
upon the through-thickness displacement pattern considered. It avoiding noise or vibrations, thermal insulation, construction of
is well known that the classical plate theory, which is based on light-weight structures with higher strength to weight ratios, and
Kirchhoff’s hypothesis, overestimates the buckling load when ease of manufacturing. Some reports have been performed to ana-
applied to even moderately thick plates. This is particularly true lyze the thermal buckling of sandwich plates. Shiau and Kuo [9]
for composite plates in which transverse shear moduli are small developed a higher-order triangular plate element for the thermal
in comparison to the in-plane Young’s moduli. In such cases, it be- buckling analysis of a rectangular composite sandwich plate.
comes necessary to take into account shear deformation effects. Dynamic buckling of imperfect sandwich plates subjected to
Thus, various improved plate theories such as first- and higher- thermo-mechanical loads was investigated by Shariyat [10]. Based
order shear deformation plate theories and sinusoidal shear defor- on finite element method, thermal buckling and vibration of sand-
mation plate theory have been developed to predict the behavior of wich plates and beams with composite facings and viscoelastic core
plates with thickness shear deformation (for more details see, Noor were investigated by Ganesan and Pradeep [11,12], Vangipuram
and Burton [1], Thornton [2] and Zenkour [3,4]). Shariyat [5] ana- and Ganesan [13] and Pradeep et al. [14]. However, Matsunaga
lyzed thermal buckling of rectangular composite multilayered [15–17] employed the global higher-order deformation theory to
plates under uniform temperature rise using a layerwise plate the- analyze the thermal buckling of composite sandwich plates and
ory. A thermal postbuckling analysis was presented by Shen [6] shells.
and Shen and Williams [7] for a simply supported plates using The concept of FGM was proposed in 1984 by a group of mate-
first-order shear deformation plate theory and classical one. The rials scientists, in Sendai, Japan, for thermal barriers or heat shield-
plates are subjected to uniform or non-uniform parabolic temper- ing properties [18]. Inhomogeneous materials such as functionally
ature loading and resting on elastic foundations. Karlsson and Bott- graded materials (FGMs) have been developed and have attracted
technical interest as new functional and intelligent materials for
a super-high-temperature environment. Such inhomogeneous
* Corresponding author at: Department of Mathematics, Faculty of Science, Kafr
El-Sheik University, Kafr El-Sheikh 33516, Egypt. materials are developed to be utilized as thin structural elements
E-mail addresses: [email protected], [email protected] (A.M. Zenkour). such as a plate or shell. When the structural elements become thin
0263-8223/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compstruct.2010.06.012
94 A.M. Zenkour, M. Sobhy / Composite Structures 93 (2010) 93–102
(a) (b)
(c) (d)
(e)
Fig. 2. Through-thickness distributions of volume fraction function for various values of the power-law index k and various types of FGM sandwich plates: (a) the (1–0–1)
FGM sandwich plate, (b) the (2–1–2) FGM sandwich plate, (c) the (1–1–1) FGM sandwich plate, (d) the (1–2–1) FGM sandwich plate, and (e) the (2–2–1) FGM sandwich plate.
2.1.3. The (1–1–1) FGM sandwich plate 2.1.4. The (1–2–1) FGM sandwich plate
As shown in Fig. 2c the plate is made of three equal-thickness Here the core thickness is twice the face thickness (see Fig. 2d).
layers. So, one gets Then
h h h h
h1 ¼ ; h2 ¼ : ð5Þ h1 ¼ ; h2 ¼ : ð6Þ
6 6 4 4
96 A.M. Zenkour, M. Sobhy / Composite Structures 93 (2010) 93–102
Moreover, the displacement of the classical plate theory (CPT) The principle of virtual work for the present problem may be
can easily be obtained by setting W(z) = 0. All of the generalized expressed as follows:
displacements u0, v0, w0, u1, v1 are independent of z. Z h
The non-linear von Karman strain–displacement equations are Nxx deð0Þ ð0Þ ð0Þ ð1Þ ð1Þ ð1Þ
xx þ N yy deyy þ N xy dcxy þ M xx dexx þ M yy deyy þ M xy dcxy
X
as follows: i
þ Pxx deð2Þ ð2Þ ð2Þ ð0Þ ð0Þ
xx þ P yy deyy þ P xy dcxy þ Q yz dcyz þ Q xz dcxz dX ¼ 0;
2
@u 1 @w ð18Þ
exx ¼ þ ;
@x 2 @x
2 where (Nxx, Nyy, Nxy) denote the total in-plane force resultants,
@ v 1 @w (Mxx, Myy, Mxy) denote the total moment resultants, (Pxx, Pyy, Pxy) are
eyy ¼ þ ;
@y 2 @y the additional stress couples associated with the transverse shear
@w effects and (Qyz, Qxz) are transverse shear stress resultants and they
ezz ¼ ; are defined as
@z ð11Þ
8 9 8
@ v @u @w @w > N
< xx > 3 Z hn >
r 9
< xx >
ðnÞ
cxy ¼ þ þ ; = X =
@x @y @x @y Nyy ¼ ryy dz;
>
: >
; n¼1 hn1 > : >
;
@w @ v Nxy rxy
cyz ¼ þ ;
@y @z 8 9 8
> M
< xx > = X 3 Z hn >
r 9
< xx > =
ðnÞ
@w @u
cxz ¼ þ : Myy ¼ ryy zdz;
@x @z > >
: ; n¼1 hn1 > : >
;
M xy rxy ð19Þ
On the basis of the displacement field given in Eq. (8), Eq. (11)
8 9 8
becomes > P
< xx > = X 3 Z hn >
r 9
< xx > =
ðnÞ
8 9 8 ð0Þ 9
8 ð1Þ 9 8 ð2Þ 9 Pyy ¼ ryy Wdz;
> exx > > > e >
exx > > e > >
: ; n¼1 hn1 >
> : >
;
>
< >
= >
< >
< xx >
>
== >
< xx >= Pxy rxy
eyy ¼ eyy þ z eyy þ WðzÞ eð2Þ
ð0Þ ð1Þ
yy ; ( ) ( )ðnÞ
>
> > > > > > > > 3 Z hn
:c > ; >
: ð0Þ >
; >
: ð1Þ >; >
: ð2Þ >
; Q yz X ryz dWðzÞ
xy cxy cxy cxy ð12Þ b
¼K dz;
( ) ( ) Q xz n¼1 h n1 r xz
dz
ð0Þ
cyz dWðzÞ cyz
ezz ¼ 0; ¼ ; b is the shear correction factor of FPT; hn and hn1 are the top
where K
cxz dz ð0Þ
cxz and bottom z-coordinates of the nth layers.
A.M. Zenkour, M. Sobhy / Composite Structures 93 (2010) 93–102 97
The stress resultants are related to the strains by the relations where
8 9 2 38 9 8 T 9 3 Z
X hn
aðnÞ ðzÞEðnÞ ðzÞT
< fNg = ½A ½B ½C < feð0Þ g = >
< fN g >
= N0xx ¼ N0yy ¼ dz: ð27Þ
fMg ¼ 4 ½B ½D ½F 5 feð1Þ g fM T g n¼1 hn1 1m
: ; : ð2Þ ; > : T > ;
fPg ½C ½F ½H fe g fP g ð20Þ For the CPT, the stability equations are reduced to the first three
( ð0Þ )
Q yz J 0 cyz equations of Eq. (26).
¼ 44 ;
Q xz 0 J 55 cð0Þ
xz
5. Trigonometric solution for the thermal buckling
where
where k = mp/a, l = np/b and U 1mn ; X 1mn ; V 1mn ; Y 1mn and W 1mn are arbi-
4. Stability equations trary parameters. Substituting Eq. (30) into Eq. (26), one obtains
½SfKg ¼ 0; ð31Þ
The stability equations of the plate may be derived by the adja-
cent equilibrium criterion. Assume that the equilibrium state of the where {K} denotes the columns
FGM plate under thermal loads is defined in terms of the displace-
fKg ¼ fU 1 ; V 1 ; W 1 ; X 1 ; Y 1 gT ; ð32Þ
ment components (u00 ; v 00 ; w00 ; u01 ; v 01 ). The displacement compo-
nents of a neighboring stable state differ by (u10 ; v 10 ; w10 ; u11 ; v 11 ) for the shear deformation theories, and
with respect to the equilibrium position. Thus, the total displace-
fKg ¼ fU 1 ; V 1 ; W 1 gT ; ð33Þ
ments of a neighboring state are
for the CPT.
u0 ¼ u00 þ u10 ; v 0 ¼ v 00 þ v 10 ; w0 ¼ w00 þ w10 ;
ð25Þ The elements Sij of the symmetric matrix [S], for the shear defor-
u1 ¼ u01 þ u11 ; v 1 ¼ v 01 þ v 11 ; mation plate theories, are given by:
where the superscript 1 refers to the state of stability and the super- S11 ¼ ½2k2 þ ð1 mÞl2 A;
script 0 refers to the state of equilibrium conditions.
S12 ¼ ð1 þ mÞklA;
Substituting Eqs. (13), (14) and (25) into Eq. (18) and integrat-
ing by parts and then equating the coefficients of du10 ; dv 10 ; dw10 ; S13 ¼ 2kðk2 þ l2 ÞB;
du11 and dv 11 to zero, separately, the governing stability equations S14 ¼ ½2k2 þ ð1 mÞl2 C;
are obtained for the shear deformation plate theories as S15 ¼ ð1 þ mÞklC;
@N1xx @N 1xy S22 ¼ ½ð1 mÞk2 þ 2l2 A
þ ¼ 0;
@x @y S23 ¼ 2lðk2 þ l2 ÞB; S24 ¼ S15 ;
@N1xy @N 1yy ð34Þ
þ ¼ 0; S25 ¼ ½ð1 mÞk2 þ 2l2 C
@x @y S33 ¼ 2ðk2 þ l2 Þ2 D 2k2 N0xx 2l2 N0yy ;
@ 2 M1xx @ 2 M 1xy @ 2 M1yy @ 2 w10 @ 2 w10
þ2 þ þ N0xx þ N0yy ¼ 0; ð26Þ S34 ¼ 2kðk2 þ l2 ÞF;
@x 2 @x@y @y 2 @x 2 @y2
1 1 S35 ¼ 2lðk2 þ l2 ÞF;
@Sxx @Sxy
þ Q 1xz ¼ 0; S44 ¼ ½2k2 þ ð1 mÞl2 H 2J;
@x @y
@S1xy @S1yy S45 ¼ ð1 þ mÞklH;
þ Q 1yz ¼ 0: S55 ¼ ½ð1 mÞk2 þ 2l2 H 2J;
@x @y
98 A.M. Zenkour, M. Sobhy / Composite Structures 93 (2010) 93–102
where where
1m 3 Z
X
R12 ¼ mR; 2 1 ¼
hn
aðnÞ ðzÞEðnÞ ðzÞ
R11 ¼ R22 ¼ R; R66 ¼ R; P ¼ a2 þ b p2 ; b dz: ð38Þ
2 ð35Þ hn1 1m
n¼1
J 44 ¼ J 55 ¼ J; ðR ¼ A; B; C; D; F; HÞ: Also, for the CPT, the critical buckling DTcr can be obtained by solv-
For the CPT, the components of [S] are reduced to be 3 3 matrix ing the following equation:
with elements Sij (i, j = 1, 2, 3) and they are mentioned above in Eq.
S11 S12 S13
(34). For non-trivial solution of Eq. (31), the determinant jSj should
S12 S22 S23 ¼ 0; ð39Þ
be zero. S
13 S23 S33
Table 2
Critical buckling temperature (103a0DTcr) of a homogeneous isotropic plate under uniform temperature rise.
Table 4 Table 5
Critical buckling temperature Tcr of FGM sandwich square plates under linear Critical buckling temperature Tcr of FGM sandwich square plates under non-linear
temperature rise. temperature rise (c = 5).
Scheme k Theory a/h = 5 a/h = 10 a/h = 15 a/h = 25 a/h = 50 Scheme k Theory a/h = 5 a/h = 10 a/h = 15 a/h = 25 a/h = 50
1–0–1 1/2 SPT 5.69553 1.55657 0.68008 0.21589 0.01680 1–0–1 1/2 SPT 21.62877 5.91108 2.58262 0.81985 0.06380
HPT 5.69147 1.55627 0.68002 0.21588 0.01680 HPT 21.61337 5.90995 2.58239 0.81982 0.06380
FPT 5.62013 1.55073 0.67888 0.21573 0.01679 FPT 21.34245 5.88890 2.57804 0.81924 0.06376
CPT 6.64118 1.62279 0.69346 0.21764 0.01691 CPT 25.21986 6.16255 2.63342 0.82651 0.06422
2 SPT 5.21919 1.38631 0.59924 0.18578 0.00916 2 SPT 23.06830 6.12734 2.64858 0.82115 0.04051
HPT 5.21036 1.38566 0.59911 0.18576 0.00917 HPT 23.02926 6.12449 2.64800 0.82107 0.04052
FPT 5.09710 1.37714 0.59736 0.18553 0.00915 FPT 22.52869 6.08684 2.64029 0.82005 0.04044
CPT 5.87400 1.43100 0.60822 0.18696 0.00924 CPT 25.96247 6.32487 2.68827 0.82634 0.04083
2–1–2 1/2 SPT 5.61388 1.53464 0.67020 0.21231 0.01590 2–1–2 1/2 SPT 21.35073 5.83656 2.54893 0.80746 0.06048
HPT 5.61059 1.53440 0.67015 0.21231 0.01590 HPT 21.33821 5.83566 2.54875 0.80744 0.06048
FPT 5.54350 1.52919 0.66908 0.21216 0.01589 FPT 21.08306 5.81584 2.54466 0.80689 0.06044
CPT 6.55131 1.60032 0.68347 0.21405 0.01601 CPT 24.91597 6.08637 2.59941 0.81408 0.06090
2 SPT 4.74906 1.25196 0.53793 0.16341 0.00354 2 SPT 22.38252 5.90053 2.53532 0.77017 0.01668
HPT 4.74274 1.25150 0.53784 0.16340 0.00354 HPT 22.35275 5.89838 2.53488 0.77011 0.01668
FPT 4.64467 1.24420 0.53635 0.16320 0.00353 FPT 21.89054 5.86398 2.52785 0.76918 0.01662
CPT 5.31032 1.29008 0.54559 0.16441 0.00360 CPT 25.02775 6.08019 2.57139 0.77488 0.01698
1–1–1 1/2 SPT 5.61662 1.53926 0.67268 0.21329 0.01615 1–1–1 1/2 SPT 21.13243 5.79146 2.53095 0.80248 0.06078
HPT 5.61448 1.53912 0.67265 0.21328 0.01615 HPT 21.12437 5.79091 2.53084 0.80247 0.06078
FPT 5.55460 1.53446 0.67169 0.21315 0.01614 FPT 20.89907 5.77339 2.52722 0.80199 0.06075
CPT 6.57686 1.60671 0.68631 0.21507 0.01626 CPT 24.74530 6.04523 2.58225 0.80921 0.06121
2 SPT 4.67391 1.23506 0.53063 0.16082 0.00289 2 SPT 22.00152 5.81379 2.49783 0.75703 0.01363
HPT 4.66999 1.23477 0.53057 0.16081 0.00289 HPT 21.98303 5.81247 2.49756 0.75699 0.01363
FPT 4.58474 1.22842 0.52927 0.16064 0.00288 FPT 21.58175 5.78254 2.49144 0.75619 0.01358
CPT 5.24611 1.27402 0.53845 0.16184 0.00296 CPT 24.69501 5.99723 2.53467 0.76185 0.01393
1–2–1 1/2 SPT 5.68984 1.56851 0.68682 0.21860 0.01751 1–2–1 1/2 SPT 20.80527 5.73535 2.51143 0.79933 0.06402
HPT 5.68943 1.56850 0.68682 0.21860 0.01751 HPT 20.80375 5.73532 2.51144 0.79933 0.06402
FPT 5.64318 1.56490 0.68608 0.21850 0.01750 FPT 20.63465 5.72216 2.50871 0.79897 0.06400
CPT 6.71309 1.64077 0.70145 0.22052 0.01763 CPT 24.54685 5.99959 2.56491 0.80635 0.06446
2 SPT 4.80799 1.28377 0.55379 0.16944 0.00508 2 SPT 21.54917 5.75380 2.48205 0.75946 0.02279
HPT 4.80746 1.28375 0.55378 0.16944 0.00508 HPT 21.54679 5.75368 2.48202 0.75946 0.02279
FPT 4.74083 1.27872 0.55275 0.16931 0.00507 FPT 21.24818 5.73116 2.47740 0.75885 0.02275
CPT 5.46601 1.32900 0.56289 0.17064 0.00516 CPT 24.49836 5.95651 2.52284 0.76480 0.02312
100 A.M. Zenkour, M. Sobhy / Composite Structures 93 (2010) 93–102
(a) (b)
(c) (d)
Fig. 3. Critical buckling temperature difference Tcr versus the power-law index k for various types of FGM sandwich square plates (a/h = 10): (a) the (1–0–1) FGM sandwich
plate, (b) the (2–1–2) FGM sandwich plate, (c) the (1–1–1) FGM sandwich plate, and (d) the (1–2–1) FGM sandwich plate.
isotropic plate k = 0, E(z) = E0, a(z) = a0, m = 0.3. Critical buckling the ceramic constituents in the plate. In general, the results of
temperature change (103a0DTcr) for different values of the side- the SPT is very agreement with those of the HPT. Moreover, similar
to-thickness ratio a/h and aspect ratio b/a of a homogeneous plate to the case of the homogeneous plate, the differences between the
is illustrated in Table 2. With the increase of the side-to-thickness higher-order shear deformation theories (SPT and HPT) and FPT are
ratio a/h, severe decrement for critical buckling temperature can be lower than those between any of them and CPT.
clearly observed. Also, it can be observed that the critical buckling Fig. 3 shows the effect of the inhomogeneity parameter k on the
temperature for the homogeneous plate decreases gradually as the critical buckling temperature Tcr for different types of sandwich
plate aspect ratio b/a increases. The difference between the shear plates under uniform, linear and non-linear temperature change
deformation plate theories and the CPT decreases as the ratios through-the-thickness using the HPT. It is clear that the critical
a/h or b/a increase because the plate becomes thin or long. buckling temperature Tcr for the plates under a non-linear temper-
The critical buckling temperature difference (Tcr = 103 DTcr) for ature change is higher than that for the plates under uniform tem-
various types of FGM sandwich plates versus the side-to-thickness perature change. While Tcr for the plates under linear temperature
ratio a/h under uniform, linear and non-linear temperature rise change is intermediate to the two previous thermal loading cases.
through-the-thickness are illustrated in Tables 3–5, respectively. It is further observed that, for the plate without core or the core of
In these tables, the power-law index is set as k = 1/2,2. It is seen the plate is half the face thickness, the critical buckling Tcr de-
that, regardless of the sandwich plate types and the thermal creases rabidly to reach minimum values and then increases grad-
loading cases, the critical temperature difference decreases mono- ually as the inhomogeneity parameter k increases as shown in
tonically as the side-to-thickness ratio a/h increases. The critical Fig. 3a and b. However, for the (1–1–1) and (1–2–1) FGM sandwich
buckling predicted for k = 1/2 are upper bounds of that predicted plates, Tcr decreases smoothly as k increases (see Fig. 3c and d).
for k = 2 except the results of the non-linear thermal loading case The critical buckling temperature change Tcr versus the aspect
c = 5. It is important to observe that the disagreements among ratio b/a of FGM sandwich plates under various thermal loading
the present theories decrease slightly with the increase of the types is exhibited in Fig. 4. It can be observed that with the
thickness of the core or, by another mean, with the increase of increase of the plate aspect ratio b/a, the critical buckling Tcr
A.M. Zenkour, M. Sobhy / Composite Structures 93 (2010) 93–102 101
(a) (b)
Fig. 4. Critical buckling temperature difference Tcr versus the plate aspect ratio b/a for two types of FGM sandwich plates (k = 1, a/h = 10): (a) the (1–0–1) FGM sandwich plate
and (b) the (1–2–1) FGM sandwich plate.
(a) (b)
Fig. 5. Critical buckling temperature difference Tcr versus the side-to-thickness ratio a/h and the aspect ratio a/b of the nonsymmetric (2–2–1) sandwich plate.
decreases gradually wherever the loading type is. It is also noticed lower one are made of FGM (Ti 6Al 4V+ZrO2) whereas the core is
from Fig. 4 that the Tcr increases with the increase of the non-lin- still ZrO2. To clarify the effect of shear deformation on the critical
earity parameter c. temperatures, the results obtained by SPT as well as HPT and FPT
Moreover, the results for the nonsymmetric FGM sandwich are compared with those obtained by CPT. It is shown through
plate is displayed in Fig. 5. It can be seen from this figure that, the numerical examples that the results of the shear deformation
regardless of the loading type and the power-law index k, the crit- plate theories are lower than those of the CPT, indicating the shear
ical buckling temperature difference Tcr decreases as the side-to- deformation effect. The following conclusions may be drawn from
thickness ratio a/h and the aspect ratio b/a increase. The critical the present analysis:
buckling temperature for the ceramic plate is higher than that
for the FGM sandwich plate. This is because the ceramic plate is The critical buckling temperature difference decreases as the
stronger than the other. The differences between the loading types side-to-thickness ratio and the plate aspect ratio increases.
decrease with the increase of a/h because the plate becomes thin. The critical buckling temperature differences of FGM sandwich
plates are generally lower than the corresponding values for
8. Conclusions homogeneous ceramic plates.
The critical buckling temperature of FGM sandwich plate under
Critical temperatures of simply supported homogeneous and non-linear temperature rise across the thickness increases as
inhomogeneous composite plates have been analyzed by using the temperature exponent c increases.
the sinusoidal shear deformation plate theory. The inhomogeneous With the increase of the core thickness of the sandwich plates,
plates are considered as FGM sandwich plates. The upper layer and the critical buckling temperature decreases.
102 A.M. Zenkour, M. Sobhy / Composite Structures 93 (2010) 93–102