Free vibration characteristics of a functionally graded beam by nite

element method
Amal E. Alshorbagy, M.A. Eltaher
*
, F.F. Mahmoud
Faculty of Engineering, Zagazig University, Zagazig, Egypt
a r t i c l e i n f o
Article history:
Received 20 March 2010
Received in revised form 10 June 2010
Accepted 5 July 2010
Available online 27 July 2010
Keywords:
FG beam
EulerBernoulli beam
Free vibration
Finite element
a b s t r a c t
This paper presents the dynamic characteristics of functionally graded beam with material
graduation in axially or transversally through the thickness based on the power law. The
present model is more effective for replacing the non-uniform geometrical beam with axi-
ally or transversally uniform geometrical graded beam. The system of equations of motion
is derived by using the principle of virtual work under the assumptions of the EulerBer-
noulli beam theory. The nite element method is employed to discretize the model and
obtain a numerical approximation of the motion equation. The model has been veried
with the previously published works and found a good agreement with them. Numerical
results are presented in both tabular and graphical forms to gure out the effects of differ-
ent material distribution, slenderness ratios, and boundary conditions on the dynamic
characteristics of the beam. The above mention effects play very important role on the
dynamic behavior of the beam.
2010 Elsevier Inc. All rights reserved.
1. Introduction
Recently, a new class of composite materials known as functionally graded materials (FGMs) has drawn considerable
attention. This material is a type of material whose composition designed to change continuously within the solid. The gra-
dient compositional variation of the constituents from one surface to the other provides an elegant solution to the problem
of high transverse shear stresses that are induced when two dissimilar materials with large difference in material properties
are bonded. The FGM concept originated in Japan in 1984 during the space-plane project, in the form of a proposed thermal
barrier material capable of withstanding a surface temperature of 2000 K and a temperature gradient of 1000 K across a
cross-section <10 mm [1]. Since FG beams are used in aerospace, automotive industries and machine elements, understand-
ing their dynamic behavior is important. Compared with FG plates and shells, studies for FG beams are relatively less [2].
Cheng and Batra [3] exploited Reddys third-order plate theory to study buckling and steady state vibrations of a simply
supported functionally gradient isotropic polygonal plate resting on a Winkler Pasternak elastic foundation and subjected to
uniform in-plane hydrostatic loads. Sankar [4] presented an elasticity solution for simply supported FG beams subjected to
sinusoidal transverse loading. The Youngs modulus is assumed to vary in an exponential fashion through the thickness. Sna-
kar found when the softer side is loaded; the stress concentrations are less than that in a homogeneous beam, and the re-
verse is true when the harder side is loaded. Chakraborty et al. [5] developed a new beam element to study the thermoelastic
behavior of functionally graded beam structures. The element is based on the rst-order shear deformation theory and it
accounts for varying elastic and thermal properties along its thickness. Both exponential and power-law variations of mate-
rial property distribution are used to examine different stress variations. Static, free vibration and wave propagation
0307-904X/$ - see front matter 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.apm.2010.07.006
* Corresponding author. Tel.: +20 101089561.
E-mail address: [email protected] (M.A. Eltaher).
Applied Mathematical Modelling 35 (2011) 412425
Contents lists available at ScienceDirect
Applied Mathematical Modelling
j our nal homepage: www. el sevi er . com/ l ocat e/ apm
problems are considered. Goupee and Senthil [6] proposed a genetic algorithm methodology to optimize the natural fre-
quencies of functionally graded structures by tailoring their material distribution for three model problems. In the rst prob-
lem, the material distributions that maximize each of the rst three natural frequencies of a functionally graded beam were
required. The goal of the second model problem was to minimize the mass of a functionally graded beam while constraining
its natural frequencies to lie outside certain prescribed frequency bands. The last problem aimed to minimize the mass of a
functionally graded beam by simultaneously optimizing its thickness and material distribution such that the fundamental
frequency is greater than a prescribed value. Aydogdu and Taskin [2] investigated analytically free vibration of simply sup-
ported FG beam. Youngs modulus of beam varies in the thickness direction according to power law and exponential law. It is
found that classical beam theory (CBT) gives higher results and difference between CBT and higher order theories is increas-
ing with increasing mode number.
Semi-analytical elasticity solutions for static bending and thermal deformation of bi-directional FG beams are presented
using a hybrid state space-based differential quadrature method by Lu et al. [7]. The results illustrated the ability of the bi-
directional functionally graded properties along the axial direction to reduce thermal stresses instead of the conventional
unidirectional functionally graded materials. Li [8] presented a new unied approach for analyzing the static and dynamic
behaviors of functionally graded beams (FGB) with the rotary inertia and shear deformation. The Timoshenko beam theory is
extended to treat FGB as well as layered beams. Different from previous approaches, a single fourth-order partial differential
equation has been derived. Xiang and Yang [9] investigated free and forced vibration of a thermally prestressed, laminated
functionally graded beamof variable thickness using the Timoshenko beamtheory and the differential quadrature method. It
is shown that the use of thicker FGM layers with a smaller volume fraction index in the laminated beam structure effectively
increases natural frequencies and lowers the vibration amplitude. Ying et al. [10] presented two-dimensional elasticity solu-
tions for bending and free vibration of functionally graded beams resting on WinklerPasternak elastic foundations. Trigo-
nometric series are adopted for the fully simply-supported beams to transform the partial differential state equation into an
ordinary one, thus making exact solutions possible. A third-order zigzag theory based model for layered functionally graded
beams in conjunction with the modied rule of mixtures (MROM) for effective modulus of elasticity is validated through
experiments for static and free vibration response by Kapuria et al. [11]. This study demonstrated the capability of the zigzag
theory in accurately modelling the mechanics of layered beams with the ceramic content.
Piovan and Sampoia [12] studied the dynamic behavior of rotating beams made of functionally graded materials. The
model has been deduced employing a formulation accounting for shear-deformability and nonlinear straindisplacements
relationships. The inuence of the graded properties in the damping effects and geometric stiffening of the rotating beam
by using the nite element method are gured out. Sina et al. [13] developed a new beam theory used to analyze free vibra-
tion of functionally graded beams. The beam properties are assumed to be varied through the thickness following a simple
power law distribution. The results showed that the new theory is a little different in natural frequency from the traditional
rst-order shear deformation beam theory and the mode shapes of the two methods are coincidental. Oyekoya et al. [14]
exploited Mindlin-type element and Reissner-type element for the modelling of FG composite plate subjected to buckling
and free vibration. Vibration and buckling analysis were then undertaken for different ber distribution cases and the effects
of ber distribution were studied. Simsek and Kocaturk [15] analyzed free vibration characteristics and the dynamic behav-
ior of a FG simply-supported beam under a concentrated moving harmonic load. The system of equations of motion is de-
rived by using Lagranges equations under the assumptions of the EulerBernoulli beam theory. It is observed that, the
effects of the different material distribution, velocity of the moving harmonic load, the excitation frequency on the dynamic
responses of the FG beam play very important role on the dynamic behavior of the FG beam.
Simsek [16] investigated dynamic behavior of a functionally graded beam under a moving mass is within the framework
of EulerBernoulli, Timoshenko and the third-order shear deformation beam theories. Malekzadeh et al. [17] developed a
formulation for the out-of-plane free vibration analysis of functionally graded (FG) circular curved beams in thermal envi-
ronment. The formulation is based on the rst-order shear deformation theory (FSDT). The material properties are assumed
to be temperature dependent and graded in the direction normal to the plane of the beam curvature. It was shown that the
temperature dependence of the material properties has signicant effects on the natural frequency parameters. Ke et al. [18]
investigates the nonlinear free vibration of functionally graded nanocomposite beams reinforced by single-walled carbon
nanotubes (SWCNTs) based on Timoshenko beam theory and von Karman geometric nonlinearity. The material properties
of functionally graded carbon nanotube-reinforced composites (FG-CNTRCs) are assumed to be graded in the thickness direc-
tion and estimated though the rule of mixture. Results showed that; an increase in CNT volume fraction leads to higher linear
and nonlinear frequencies for both uniform distribution and FG-CNTRC beams. Both linear and nonlinear frequencies of FG-
CNTRC beam with symmetrical distribution of CNTs are higher than those of beams with uniform or unsymmetrical distri-
bution of CNTs. Huang and Li [19] studied free vibration of axially functionally graded beams with non-uniform cross-sec-
tion. They transformed the governing equation with varying coefcients to Fredholm integral equations to nd the natural
frequencies of beams with variable exural rigidity and mass density.
In the present study, the free vibration analysis of FG beams is investigated using numerical nite element method. The
equations of motion of FG beams are derived using EulerBernoulli beam theory and virtual work principle. The material
constituents of beams assumed to be varying through the thickness or longitudinal directions according to a simple power
law. The present model is effective for comparing the tapered beam of linearly variable width or depth and graded beams of
special polynomial non-homogeneity. Finally, the effects of various boundary conditions (BCs), power-exponent index and
beams slenderness ratio are investigated.
A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425 413
2. The mathematical model formulation
A functionally graded simply-supported beam of length L, width b, thickness h, with co-ordinate system (Oxyz) having the
origin O is shown In Fig. 1.
2.1. Material properties
In this study, it is assumed that the material properties of the beam such as, Youngs modulus E and mass density q vary
continuously according to power-law form.
(a) If the material properties varied through the beam thickness as shown in Fig. 2, they can be described by a simple
power law distribution [15] as follow:
Pz P
U
P
L

z
h

1
2
_ _
k
P
L
; 1
O
z
x
I
I
L
Fig. 1. A functionally graded simply-supported beam.
2 2.25 2.5 2.75 3 3.25 3.5 3.75 4
E(GPa)
-0.5
-0.25
0
0.25
0.5
z
/
h
k=0.1
k=0.2
k=0.5
k=1
k=2
k=5
k=10
4000 4500 5000 5500 6000 6500 7000 7500 8000
(kg/m
3
)
-0.5
-0.25
0
0.25
0.5
z
/
h
k=0.1
k=0.2
k=0.5
k=1
k=2
k=5
k=10
Fig. 2. The variation of Youngs modulus and the mass density through the thickness direction of the beam.
0 0.2 0.4 0.6 0.8 1
Normilazed beam length
2
2.25
2.5
2.75
3
3.25
3.5
3.75
4
E
(
G
P
a
)
k=0.1
k=0.2
k=0.5
k=1
k=2
k=5
k=10
0 0.2 0.4 0.6 0.8 1
Normilazed beam length
4000
4500
5000
5500
6000
6500
7000
7500
8000
(
k
g
/
m
3
)
k=0.1
k=0.2
k=0.5
k=1
k=2
k=5
k=10
Fig. 3. The variation of Youngs modulus and the mass density through the axial direction of the beam.
414 A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425
where P
U
an P
L
are the corresponding material properties of the upper and the lower surfaces of the beam, and k is the
non-negative power-law exponent which dictates the material variation prole through the thickness of the beam.
(b) If the material properties varied through the beam axis as presented in Fig. 3, the material characteristics can be
assumed to be varied also a power law distribution as:
Pz P
L
P
R
1
x
L
_ _
k
P
R
; 2
where P
R
and P
L
are the corresponding material properties of the right and the left side of the beam, and k is the non-
negative power-law exponent which dictates the material variation prole through the axis of the beam.
2.2. The conditions of geometric t
Based on the EulerBernoulli beam theory, the axial displacement u and the transverse displacement of any point of the
beam, w, are given by [15]
ux; z; t u
o
x; t z
@w
o
x; t
@x
; 3
Wx; z; t w
o
x; t; 4
where u
o
and w
o
are the axial and the transverse displacement of any point on the mid-plane, and t denotes time. Eqs. (3) and
(4) can be rewritten as
fd
s
g
u
w
_ _

1 0 z
0 1 0
_ _
u
0
w
0
@w
0
@x
_

_
_

_
; 5
where d
s
is the displacement vector. By assuming the small deformations, the displacementstrain relation can be repre-
sented by
e
xx

@u
@x

@u
0
x; t
@x
z
@
2
w
0
x; t
@x
2
6a
or
e
xx
1 z
@u
0
@x
@
2
w
0
@x
2
_ _
; 6b
where e
xx
is the normal strain in the x-direction.
2.3. The material law
Considering the material of FGM beam obeys Hookes law, the strainstress constitutive equation can be written as
following:
r
xx
Eze
xx
Ez 1 z
@u
0
@x
@
2
w
0
@x
2
_ _
; 7
when the variation of the Youngs modulus through the thickness direction
r
xx
Exe
xx
Ex 1 z
@u
0
@x
@
2
w
0
@x
2
_ _
; 8
when the variation of the Youngs modulus through the axial direction.
2.4. The condition of equilibrium
The condition of equilibrium of dynamical structural for free vibration based on the principle of virtual work stated that
[20],
dW
s
dW
I
0; 9
where dW
S
is the virtual work done by a stress eld r on a virtual stain eld e is equal to
dW
s

___
v
r
T
dedxdydz: 10
A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425 415
By substituting Eqs. (6), and (7) or (8) into Eq. (10), yields
dW
s
b
_
L
0
@uo
@x
@
2
wo
@x
2
_ _
_ h
2

h
2
E zE
zE z
2
E
_ _
dz
_ _
d
@u
0
@x
@
2
w
0
@x
2
_ _
dx: 11
The virtual work done the inertia forces through the virtual displacement eld can be presented as
dW
I

___
v
q

d
s
dd
s
dxdydz: 12
By substituting Eq. (5), into Eq. (9) yields
dW
I
b
_
L
0
_
u
0
_
w
0
@ _ wo
@x
_ _
_ h
2

h
2
q 0 qz
0 q 0
qz 0 qz
2
_

_
_

_dz
_

_
_

_d
_
u
0
_
w
0
@ _ w
0
@x
_

_
_

_
dx: 13
3. Finite element formulation
The displacement components at the mid-plane of a beam element shown in Fig. 4 can be categorized into the following:
In-plane components
U
e
0
x; t N
3i2
U
i
t N
3j2
U
j
t: 14
Transverse components
w
e
0
x; t N
3i1
W
i
t N
3i
h
i
t N
3j1
W
j
t N
3j
h
j
t; 15
where U
i
, W
i
and h
i
are in plane displacement, transverse displacement and slope at the nodal, respectively. Ns are the shape
functions, which are presented in Appendix. By substituting Eqs. (14) and (15) into Eqs. (11) and (13), then substituting into
Eq. (9) and integrate over domain we nd the following equation of motion:
M

d
s
kd
s
0; 16
where M is the global mass-matrix of beam and K is the global stiffness matrix of beam, which are dened in Appendix.
4. Numerical results
This section divided mainly to two subsections; one for that the material properties of the beam vary continuously in the
thickness direction and the other is devoted to material properties is changed continuously in axial direction of the beam.
Functionally graded material (FGM) of the beam is composed of steel and alumina (Al
2
O
3
) and its properties vary according
to power-law. Therefore, the bottomsurface of the beam is pure steel, whereas the top surface of the beam is pure alumina in
case of variation through thickness. If the variation through the axial direction, the right side is pure steel and left side is pure
alumina. The material properties of steel an alumina are given in Table 1.
The parameters of the beam are: b (width) = 0.4 m and L (length) = 20 m. The non-dimensional quantities used here are

x
x
L
; E
ratio

E
u
E
l
; q
ratio

q
u
q
l
and k
2
xL
2

q
l
E
l
A
I

;
where I = bh
3
/12 is the moment of inertia of the cross-section of the beam. In the free vibration analysis, the dimensionless
frequencies of the beam are calculated and given in tabular form for different material distribution, the ratio of the Youngs
Fig. 4. Beam element.
Table 1
Material properties of FGM constituents.
Properties Unit Steel Alumina (Al
2
O
3
)
E GPa 210 390
q Kg/m
3
7800 3960
416 A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425
modulus, and the slenderness ratio. The mode shapes are given in graphical form for each case. The beam is simply sup-
ported through analysis.
4.1. FGM through thickness direction
This section is devoted mainly to the FG beam with continuously variation through the thickness direction to verify the
present proposed model. The effects of variation of Youngs modulus, material distributions, non-negative power-law expo-
nent and slenderness ratio on the fundamental frequencies are illustrated.
4.1.1. The effect of Youngs modulus variation
The effects of Young modulus ratios, slenderness ratios, and non-negative power-law exponent on the rst three dimen-
sionless frequencies are presented in Tables 24. It can be observed that, the natural frequencies increase with an increase in
power exponent (when E
ratio
< 1), and decrease with an increase in power exponent (when E
ratio
> 1). For a constant power
exponent, an increase in E
ratio
causes the increase in fundamental frequencies. No signicant variation on the frequencies
with respect to variation in slenderness ratio. Comparison between proposed model and published work is presented also
in Table 2. It can be concluded that; the present results agree reasonably well with the published work.
The variation of the rst frequency with the variation of E
ratio
and power exponent (k) is shown in Fig. 5, which illustrates
that the variation of the E
ratio
is more effective on the frequency than the variation in power exponent (k).
The rst three dimensionless mode shapes are illustrated in Figs. 68. As expected no effect on dimensionless mode
shapes with respect to variation in E
ratio
, power exponent, or slenderness ratio because no change in material properties
through the axial direction. The slenderness has not effect on the frequencies due to the proposed model (Euler beam) which
assumed no shear effect that play a vital role in the variation of frequencies for thick beam (small slenderness ratio). To point
out the effect of slenderness ratio, the higher-order shear deformation or Timoshenko beam theory should be developed.
Table 2
The rst dimensionless frequency parameters k
1
for different material distribution (E
ratio
= E
u
/E
l
, q
ratio
= q
u
/q
l
= 1).
L/h E
ratio
K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 Present 0.25 2.2203 2.3746 2.4614 2.5979 2.7041 2.8057 2.9302 3.0085
[15] 2.2203 2.3739 2.4606 2.7035 2.8053 3.0084
Present 0.5 2.6404 2.7107 2.7576 2.8363 2.8946 2.9461 3.011 3.0563
[15] 2.6403 2.7104 2.7573 2.8944 2.9459 3.0562
Present 1.0 3.14 3.14 3.14 3.14 3.14 3.14 3.14 3.14
[15] 3.1399 3.1399 3.1399 3.1399 3.1399 3.1399
Present 2.0 3.7341 3.6773 3.63 3.5296 3.4423 3.3768 3.3196 3.2726
[15] 3.734 3.6775 3.6301 3.4421 3.3765 3.2725
Present 4.0 4.4406 4.3366 4.2455 4.0346 3.8241 3.6496 3.5326 3.4549
[15] 4.4406 4.3370 4.2459 3.8234 3.6485 3.4543
100 Present 0.25 2.2214 2.3798 2.4683 2.6074 2.7159 2.8071 2.9317 3.01
[15] 2.2213 2.3752 2.4621 2.7053 2.8071 3.01
Present 0.5 2.6417 2.7121 2.759 2.8377 2.8961 2.9476 3.0125 3.0578
[15] 2.6416 2.7117 2.7587 2.896 2.9475 3.0578
Present 1.0 3.1415 3.1415 3.1415 3.1415 3.1415 3.1415 3.1415 3.1415
[15] 3.1415 3.1415 3.1415 3.1415 3.1415 3.1415
Present 2.0 3.7359 3.6791 3.6317 3.5313 3.444 3.3784 3.3213 3.2743
[15] 3.7359 3.6793 3.632 3.444 3.3784 3.2742
Present 4.0 4.4428 4.3388 4.2476 4.0366 3.826 3.6514 3.5343 3.4566
[15] 4.4427 4.3392 4.2481 3.8259 3.6513 3.4565
Table 3
The second dimensionless frequency parameters k
2
for different material distribution (E
ratio
= E
u
/E
l
, q
ratio
= q
u
/q
l
= 1).
L/h E
ratio
K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 0.25 4.4338 4.7419 4.9152 5.1877 5.3997 5.6028 5.8514 6.0078
0.5 5.2727 5.4132 5.5068 5.6638 5.7804 5.8832 6.0128 6.1032
1.0 6.2703 6.2703 6.2703 6.2703 6.2703 6.2703 6.2703 6.2703
2.0 7.4567 7.3432 7.2488 7.0483 6.874 6.7431 6.6291 6.5352
4.0 8.8676 8.6599 8.478 8.0567 7.6363 7.2877 7.0541 6.8991
100 0.25 4.4425 4.7592 4.9362 5.2145 5.4314 5.6139 5.8629 6.0196
0.5 5.2831 5.4238 5.5176 5.675 5.7918 5.8948 6.0246 6.1152
1.0 6.2827 6.2827 6.2827 6.2827 6.2827 6.2827 6.2827 6.2827
2.0 7.4714 7.3577 7.263 7.0622 6.8876 6.7564 6.6421 6.5481
4.0 8.885 8.677 8.4947 8.0726 7.6515 7.3023 7.0683 6.9128
A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425 417
4.1.2. The effect of material distribution
Results are obtained for an FG beam constructed from steel and alumina according to power-law. The rst ve non-
dimensional frequencies for different slenderness ratios and power exponent are given in Table 5. Fig. 9 shows the effect
of the variation of power-law exponent on the fundamental frequency of the beam. It is seen that, when the power exponent
Fig. 5. Variation of the rst dimensionless frequency with respect to E
ratio
and k.
Table 4
The third dimensionless frequency parameters k
3
for different material distribution (E
ratio
= E
u
/E
l
, q
ratio
= q
u
/q
l
= 1).
L/h E
ratio
K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 0.25 6.6338 7.0947 7.3537 7.7611 8.0783 8.3823 8.7546 8.9888
0.5 7.889 8.0991 8.2392 8.474 8.6483 8.8022 8.9962 9.1316
1.0 9.3817 9.3817 9.3817 9.3817 9.3817 9.3817 9.3817 9.3817
2.0 11.157 10.987 10.846 10.546 10.285 10.089 9.9182 9.7779
4.0 13.268 12.957 12.685 12.054 11.424 10.902 10.553 10.322
100 0.25 6.6631 7.1381 7.4036 7.8209 8.1462 8.4199 8.7935 9.0285
0.5 7.9238 8.1349 8.2756 8.5116 8.6868 8.8413 9.036 9.1719
1.0 9.423 9.423 9.423 9.423 9.423 9.423 9.423 9.423
2.0 11.206 11.035 10.893 10.592 10.33 10.134 9.9622 9.8212
4.0 13.326 13.014 12.741 12.108 11.476 10.952 10.601 10.368
0 0.2 0.4 0.6 0.8 1
Normalized beam length
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
N
o
r
m
a
l
i
z
e
d

d
e
f
l
e
c
t
i
o
n

a
n
d

s
l
o
p
e
1
st
Mode
Deflection
Slope
Fig. 6. The normalized rst mode shape for material variation through thickness.
418 A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425
k, starts to deviate from zero and it grows up, the beam deviates from pure alumina to steel. This deviation causes reduction
in the non-dimensional frequency. The nonlinear diminishing occurs in frequency when k increased from 0 to 2 and the lin-
ear reduction occurs when k is increased from 2 to 10 as shown in Fig. 9. The present result is conformable with the previous
example where E
ratio
> 1. It is notably that, the change in dimensionless frequency with respect to slenderness ratio is
0 0.2 0.4 0.6 0.8 1
Normalized beam length
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
N
o
r
m
a
l
i
z
e
d

d
e
f
l
e
c
t
i
o
n

a
n
d

s
l
o
p
e
2
nd
Mode
Deflection
Slope
Fig. 7. The normalized second mode shape for material variation through thickness.
0 0.2 0.4 0.6 0.8 1
Normalized beam length
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
N
o
r
m
a
l
i
z
e
d

d
e
f
l
e
c
t
i
o
n

a
n
d

s
l
o
p
e
3
nd
Mode
Deflection
Slope
Fig. 8. The normalized third mode shape for material variation through thickness.
Table 5
The variation of non-dimensional frequency for different material distributions and slenderness ratios.
L/h k
i
K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 i = 1 4.3425 4.2817 4.2315 4.1262 4.0359 3.9684 3.9075 3.8572
i = 2 8.6716 8.5503 8.45 8.2397 8.0595 7.9245 7.803 7.7026
i = 3 12.975 12.793 12.643 12.328 12.058 11.856 11.675 11.525
i = 4 17.239 16.997 16.798 16.379 16.02 15.752 15.511 15.312
i = 5 21.453 21.153 20.904 20.382 19.934 19.599 19.301 19.054
50 i = 1 4.3444 4.2836 4.2333 4.1279 4.0377 3.9701 3.9092 3.8589
i = 2 8.6866 8.565 8.4646 8.2539 8.0734 7.9382 7.8165 7.7159
i = 3 13.025 12.842 12.692 12.376 12.105 11.902 11.72 11.569
i = 4 17.356 17.113 16.912 16.491 16.131 15.861 15.617 15.417
i = 5 21.679 21.376 21.125 20.599 20.148 19.811 19.507 19.256
100 i = 1 4.3446 4.2838 4.2336 4.1282 4.0379 3.9703 3.9094 3.8591
i = 2 8.6887 8.5671 8.4666 8.2559 8.0754 7.9402 7.8184 7.7178
i = 3 13.032 12.849 12.699 12.383 12.112 11.909 11.726 11.575
i = 4 17.373 17.13 16.929 16.508 16.147 15.876 15.633 15.432
i = 5 21.712 21.409 21.157 20.631 20.18 19.842 19.537 19.286
A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425 419
0 2 4 6 8 10
Power-exponent (k)
3.8
3.9
4
4.1
4.2
4.3
4.4
N
o
n
d
i
m
e
n
s
i
o
n
a
l

F
r
e
q
u
e
n
c
y

(

1
)
Fig. 9. Variation of the rst nondimensisnal frequency with the power-law exponent (k).
Table 6
The rst dimensionless frequency parameters k
1
for different material distribution (E
ratio
= E
u
/E
l
, q
ratio
= q
u
/q
l
= 1).
L/h E
ratio
K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 0.25 2.2203 2.3285 2.4106 2.5821 2.7533 2.9278 3.0834 3.1265
0.5 2.6404 2.6868 2.7258 2.8148 2.9104 3.0122 3.1052 3.1316
1.0 3.14 3.14 3.14 3.14 3.14 3.14 3.14 3.14
2.0 3.7341 3.6988 3.6653 3.5758 3.4611 3.3244 3.1923 3.1531
4.0 4.4406 4.3768 4.3144 4.1387 3.8937 3.5795 3.2668 3.1726
100 0.25 2.2214 2.3297 2.4118 2.5834 2.7546 2.9293 3.0849 3.1281
0.5 2.6417 2.6881 2.7271 2.8162 2.9119 3.0137 3.1067 3.1332
1.0 3.1415 3.1415 3.1415 3.1415 3.1415 3.1415 3.1415 3.1415
2.0 3.7359 3.7006 3.6671 3.5775 3.4628 3.326 3.1939 3.1547
4.0 4.4428 4.3789 4.3166 4.1408 3.8957 3.5812 3.2684 3.1742
Table 7
The second dimensionless frequency parameters k
2
for different material distribution (E
ratio
= E
u
/E
l
, q
ratio
= q
u
/q
l
= 1).
L/h E
ratio
K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 0.25 4.4338 4.6693 4.8374 5.1675 5.473 5.7674 6.0636 6.1987
0.5 5.2727 5.3752 5.4573 5.6327 5.8048 5.9739 6.1459 6.2263
1.0 6.2703 6.2703 6.2703 6.2703 6.2703 6.2703 6.2703 6.2703
2.0 7.4567 7.3774 7.3039 7.1176 6.9031 6.6783 6.4483 6.3365
4.0 8.8676 8.7236 8.5853 8.2113 7.7399 7.2209 6.6902 6.4302
100 0.25 4.4425 4.6785 4.847 5.1778 5.484 5.7789 6.0756 6.2109
0.5 5.2831 5.3858 5.4681 5.6438 5.8162 5.9858 6.158 6.2386
1.0 6.2827 6.2827 6.2827 6.2827 6.2827 6.2827 6.2827 6.2827
2.0 7.4714 7.3919 7.3183 7.1316 6.9167 6.6915 6.4611 6.349
4.0 8.885 8.7407 8.6022 8.2276 7.7555 7.2356 6.7036 6.443
Table 8
The third dimensionless frequency parameters k
3
for different material distribution (E
ratio
= E
u
/E
l
, q
ratio
= q
u
/q
l
= 1).
L/h E
ratio
K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 0.25 6.6338 6.9949 7.247 7.733 8.1753 8.5989 9.0262 9.2305
0.5 7.889 8.0472 8.1715 8.4317 8.6814 8.9246 9.1708 9.2901
1.0 9.3817 9.3817 9.3817 9.3817 9.3817 9.3817 9.3817 9.3817
2.0 11.157 11.033 10.92 10.638 10.324 10.001 9.6735 9.5139
4.0 13.268 13.043 12.829 12.258 11.562 10.815 10.059 9.6942
100 0.25 6.6631 7.0259 7.2792 7.7678 8.2123 8.6377 9.0664 9.2713
0.5 7.9238 8.0827 8.2076 8.4691 8.7199 8.9642 9.2114 9.3311
1.0 9.423 9.423 9.423 9.423 9.423 9.423 9.423 9.423
2.0 11.206 11.082 10.968 10.685 10.37 10.045 9.7163 9.5559
4.0 13.326 13.101 12.886 12.313 11.614 10.864 10.105 9.7373
420 A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425
relatively small and approximately neglected for higher slenderness ratio. This results agree very well with analytical solu-
tion [13]. Seeing as the material properties are constant through the beam axis, no change occurs in mode shapes and their
proles as presented in Figs. 68.
0.5 1 1.5 2 2.5 3 3.5 4
E
ratio
2
2.5
3
3.5
4
4.5
N
o
n
-
d
i
m
e
n
s
i
o
n
a
l

f
r
e
q
u
e
n
c
y

(

1
)
k=0.0
k=0.1
k=0.2
k=0.5
k=1
k=2
k=5
k=10
Fig. 10. Variation of the non-dimensional frequency (k
1
) with Youngs modulus ratio.
N
o
n
-
d
i
m
e
n
s
i
o
n
a
l

f
r
e
q
u
e
n
c
y

(

1
)
0 2 4 6 8 10
Power exponent (k)
2
2.5
3
3.5
4
4.5
E
ratio
=4
E
ratio
=2
E
ratio
=1
E
ratio
=0.5
E
ratio
=0.25
Fig. 11. Variation of the non-dimensional frequency (k
1
) with power exponent (k).
Table 9
The variation of non-dimensional frequency for different material distributions and slenderness ratios.
L/h k
i
K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 i = 1 4.3425 4.3047 4.2689 4.1741 4.0537 3.9113 3.7741 3.7334
i = 2 8.6716 8.5867 8.5084 8.3115 8.087 7.8528 7.6132 7.4969
i = 3 12.975 12.842 12.722 12.424 12.096 11.759 11.418 11.252
i = 4 17.239 17.06 16.898 16.501 16.069 15.627 15.183 14.971
i = 5 21.453 21.228 21.024 20.53 19.995 19.45 18.901 18.644
50 i = 1 4.3444 4.3065 4.2708 4.1759 4.0555 3.913 3.7757 3.735
i = 2 8.6866 8.6015 8.5231 8.3259 8.101 7.8664 7.6264 7.5098
i = 3 13.025 12.892 12.771 12.472 12.143 11.804 11.462 11.295
i = 4 17.356 17.176 17.013 16.614 16.179 15.734 15.286 15.073
i = 5 21.679 21.452 21.246 20.747 20.207 19.656 19.101 18.841
100 i = 1 4.3446 4.3068 4.271 4.1761 4.0558 3.9132 3.7759 3.7353
i = 2 8.6887 8.6036 8.5252 8.3279 8.103 7.8684 7.6282 7.5116
i = 3 13.032 12.899 12.778 12.479 12.149 11.811 11.468 11.301
i = 4 17.373 17.193 17.029 16.63 16.195 15.75 15.302 15.088
i = 5 21.712 21.485 21.279 20.779 20.238 19.686 19.13 18.87
A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425 421
4.2. FGM through axial direction
It is noted that most of the previous analyses are related to FGMs with material properties varying in the thickness direc-
tion. However, there are practical occasions which require tailored grading of macroscopic properties in two or three direc-
tions [7]. So, this section is concerned with the variation of FG beam through the axial direction.
4.2.1. The effect of Youngs modulus variation
The rst three non-dimensional natural frequencies for the FG beam for modulus ratios, power exponent, and slenderness
ratios are tabulated in Tables 68. The effect of the modulus ratios on the frequency is presented in Fig 10. It is observed that,
the frequency increased signicantly with increasing of modulus ratio especially for small power exponent. On the other
hand, no signicantly changes occur in fundamental frequency for different modulus ratios for higher value of power expo-
Table 10
The rst non-dimensional frequency parameters (k
1
) for different boundary conditions.
L/h BC K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 SS 4.3425 4.3047 4.2689 4.1741 4.0537 3.9113 3.7741 3.7334
CF 2.5926 2.5292 2.482 2.3824 2.3007 2.247 2.2236 2.2211
FC 2.5926 2.5855 2.5786 2.5586 2.5282 2.4783 2.3895 2.3263
CC 6.5373 6.4364 6.3673 6.2208 6.087 5.9591 5.8315 5.7704
50 SS 4.3444 4.3065 4.2708 4.1759 4.0555 3.913 3.7757 3.735
CF 2.5931 2.5297 2.4825 2.3828 2.3012 2.2474 2.2241 2.2216
FC 2.5931 2.586 2.5791 2.5591 2.5287 2.4789 2.39 2.3268
CC 6.5408 6.4399 6.3707 6.2241 6.0902 5.9624 5.8348 5.7736
100 SS 4.3446 4.3068 4.271 4.1761 4.0558 3.9132 3.7759 3.7353
CF 2.5932 2.5298 2.4826 2.3829 2.3012 2.2475 2.2241 2.2216
FC 2.5932 2.5861 2.5792 2.5592 2.5288 2.4789 2.3901 2.3269
CC 6.5413 6.4404 6.3712 6.2246 6.0907 5.9628 5.8352 5.774
Table 11
The second non-dimensional frequency parameters (k
2
) for different boundary conditions.
L/h BC K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 SS 8.6716 8.5867 8.5084 8.3115 8.087 7.8528 7.6132 7.4969
CF 6.4808 6.372 6.2948 6.1246 5.9567 5.7825 5.6106 5.5623
FC 6.4808 6.436 6.3934 6.2792 6.1312 5.9529 5.7831 5.7181
CC 10.835 10.68 10.572 10.333 10.093 9.8504 9.606 9.4992
50 SS 8.6866 8.6015 8.5231 8.3259 8.101 7.8664 7.6264 7.5098
CF 6.49 6.3809 6.3034 6.1328 5.9646 5.7902 5.6183 5.5701
FC 6.49 6.4452 6.4026 6.2884 6.1404 5.962 5.7917 5.7266
CC 10.857 10.701 10.593 10.353 10.114 9.8705 9.6257 9.5187
100 SS 8.6887 8.6036 8.5252 8.3279 8.103 7.8684 7.6282 7.5116
CF 6.4913 6.3822 6.3047 6.134 5.9658 5.7913 5.6194 5.5712
FC 6.4913 6.4465 6.404 6.2898 6.1417 5.9633 5.793 5.7278
CC 10.86 10.704 10.596 10.356 10.116 9.8734 9.6285 9.5215
Table 12
The third non-dimensional frequency parameters(k
3
) for different boundary conditions.
L/h BC K = 0.0 K = 0.1 K = 0.2 K = 0.5 K = 1 K = 2 K = 5 K = 10
20 SS 12.975 12.842 12.722 12.424 12.096 11.759 11.418 11.252
CF 10.82 10.661 10.549 10.297 10.038 9.7619 9.457 9.323
FC 10.82 10.724 10.635 10.407 10.136 9.8499 9.5908 9.4852
CC 15.129 14.922 14.776 14.442 14.096 13.739 13.38 13.226
50 SS 13.025 12.892 12.771 12.472 12.143 11.804 11.462 11.295
CF 10.856 10.696 10.584 10.33 10.07 9.793 9.4875 9.3536
FC 10.856 10.761 10.672 10.443 10.172 9.8848 9.6242 9.518
CC 15.194 14.985 14.839 14.504 14.157 13.798 13.439 13.284
100 SS 13.032 12.899 12.778 12.479 12.149 11.811 11.468 11.301
CF 10.861 10.701 10.589 10.335 14.135 9.7975 9.4919 9.358
FC 10.861 10.766 10.677 10.448 10.177 9.8898 9.6291 9.5227
CC 15.203 14.995 14.848 14.513 14.166 13.807 13.447 13.292
422 A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425
nent. Fig. 11 illustrates the variation in non-dimensional frequency with respect to power exponent. As seen in Fig. 11, the
increase in power exponent causes the increase in frequency for E
ratio
< 1, the decrease in frequency for E
ratio
> 1 and no
changes occur for E
ratio
= 1.
4.2.2. The effect of material distribution
Results are obtained for an FG beam constructed from steel and alumina according to power-law through the axial direc-
tion of beam. The rst ve non-dimensional frequencies for different slenderness ratios and power exponent are given in
0 2 4 6 8 10
Power-exponent (k)
0
1
2
3
4
5
6
7
N
o
n
-
d
i
m
e
n
s
i
o
n
a
l

f
r
e
q
u
e
n
c
y

(

1
)
CC
SS
FC
CF
Fig. 12. Variation of the non-dimensional frequency (k
1
) with power exponent (k) for different BCs.
0 0.2 0.4 0.6 0.8 1
Normalized beam Length
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
N
o
r
m
a
l
i
z
e
d

d
e
f
l
e
c
t
i
o
n

a
n
d

s
l
o
p
e
1
st
Mode (CC)
k=0
k=2
Deflection
Slope
0 0.2 0.4 0.6 0.8 1
Normalized beam Length
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
N
o
r
m
a
l
i
z
e
d

d
e
f
l
e
c
t
i
o
n

a
n
d

s
l
o
p
e
1
st
Mode (CF)
k=0
k=2
Deflection
Slope
0 0.2 0.4 0.6 0.8 1
Normalized beam Length
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
N
o
r
m
a
l
i
z
e
d

d
e
f
l
e
c
t
i
o
n

a
n
d

s
l
o
p
e
1
st
Mode (FC)
k=0
k=2
Slope
Deflection
0 0.2 0.4 0.6 0.8 1
Normalized beam Length
-1.2
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
N
o
r
m
a
l
i
z
e
d

d
e
f
l
e
c
t
i
o
n

a
n
d

s
l
o
p
e
1
st
Mode (SS)
k=0
k=2
Deflection
Slope
Fig. 13. Variation of the mode shape with power exponent (k) for different BCs (L/h = 100).
A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425 423
Table 9. It is noted that; the increasing in power exponent cause the decrease in the fundamental frequency. This decreasing
due to transition of the beam from alumina which has higher elastic modulus and lower density to steel which has the lower
elastic modulus and higher density. The change of slenderness ratio from 20 to 50 causes the increase of frequency with 4%.
On the other hand, approximately no variation occurs in frequency when the slenderness ratio increased from 50 to 100.
4.2.3. The effect of boundary conditions
The geometry constraints or boundary conditions (BCs) have signicant effects on the dynamic characteristics of the
structures. So, the effects of four different beam constraints on the fundamental frequencies of the FG beam are presented
in Tables 1012. The boundary conditions are clamped-clamped (CC), clamped-free (CF), free-clamped (FC), and simply-sim-
ply (SS). Fig. 12 presents decrease in frequencies as k increased for the different (BCs). This reduction in frequencies due to
the increase of steel relative to alumina in the beam constitution. It is also noted, CF and FC have the same frequency only for
k = 0, but have different frequencies for other values of k. Approximately, no signicant effect of the slenderness ratio on the
fundamental frequencies especially at the higher values of L/h.
Non-dimensional mode shapes for the different BCs are presented in Fig. 13 at L/h = 100. It is noted that, the variation of
material distribution (variation of k) along the beam length has signication effects on the mode shapes.
5. Conclusions
Free vibration characteristics and dynamic behavior of a functionally graded beam for different material distributions are
analyzed numerically by nite element method. The equation of motion for free vibration is derived by principle of virtual
work under the assumptions of the EulerBernoulli beam theory. It is assumed that the material properties of the beam vary
continuously through the thickness or axial direction according to power-law form. Numerical comparisons are presented to
validate the convergence and accuracy of the presented model. Numerical results show that:
(1) The variation of material distribution along the axial direction means the variation of the beam stiffness along spatial,
which effects on the frequencies and mode shape. On the other hand, the beam with stiffness along spatial not chan-
ged with the variation of material along the thickness. So, the mode shape not varied with in this case.
(2) Slenderness ratio has not effects on the fundamental frequencies or mode shapes. This due to the deciency in Euler-
beam theory for consideration the shear effect, which affects signicant on the frequencies especially for the short
beam. To gure out the effect of slenderness ratio, the other advanced theories such as Timoshenko or Reddy theories
should be considered.
(3) The natural frequencies increase with an increase in power exponent (when E
ratio
< 1), and decrease with an increase in
power exponent (when E
ratio
> 1).
Appendix
The shape functions are
N
3i2
1
x
l
_ _
; A1
N
3i1

1
l3
_ _
l
3
3lx
2
2x
3
; A2
N
3i

1
l2
_ _
l
2
x 2lx
2
2x
3
; A3
N
3j2

x
l
_ _
; A4
N
3j1

1
l3
_ _
3lx
2
2x
3
; A5
N
3j

1
l2
_ _
lx
2
x
3
; A6
where l is the beam element and x is the local co-ordinate of the beam element.
The global stiffness matrix can be depicted by
K

n
0
K
e
; A7
where n is the total number of discretized elements. The element stiffness matrix is
K
e
b
_
l
0
B
T
D
E
Bdx; A8
424 A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425
where
B
dN
3i2
dx
0 0
dN
3j2
dx
0 0
0
d2N
3i1
dx
2
d2N
3i
dx
2
0
d2N
3j1
dx
2
d2N
3j
dx
2
_ _
A9
and
D
E

_ h
2

h
2
E zE
zE z
2
E
_ _
dz: A10
The global mass-matrix is
M

n
0
M
e
A11
and the element stiffness matrix is
M
e
b
_
l
o
N
T
D
R
Ndx; A12
where
N
N
3i2
0 0 N
3j2
0 0
0 N
3i1
N
3i
0
d2N
3j1
dx
2
d2N
3j
dx
2
0
dN
3i1
dx
dN
3i
dx
0
d2N
3j1
dx
d2N
3j
dx
_

_
_

_
A13
and
D
R

_ h
2

h
2
q 0 qz
0 q 0
qz 0 qz
2
_

_
_

_dz: A14
References
[1] X.L. Chen, K.M. Liew, Buckling of rectangular functionally graded material plates subjected to nonlinearly distributed in-plane edge loads, Smart Mater.
Struct. 13 (2004) 14301437.
[2] M. Aydogdu, V. Taskin, Free vibration analysis of functionally graded beams with simply supported edges, Mater. Des. 28 (2007) 16511656.
[3] Z.Q. Cheng, R.C. Batra, Exact correspondence between eigenvalues of membranes and functionally graded simply supported polygonal plates, J. Sound
Vib. 229 (4) (2000) 879895.
[4] B.V. Sankar, An elasticity solution for functionally graded beams, Compos. Sci. Technol. 61 (2001) 689696.
[5] A. Chakraborty, S. Gopalakrishnan, J.N. Reddy, A new beam nite element for the analysis of functionally graded materials, Int. J. Mech. Sci. 45 (2003)
519539.
[6] A.J. Goupee, S.V. Senthil, Optimization of natural frequencies of bidirectional functionally graded beams, Struct. Multidiscip. Optim. 32 (2006) 473484.
[7] C.F. Lu, W.Q. Chen, R.Q. Xu, C.W. Lim, Semi-analytical elasticity solutions for bi-directional functionally graded beams, Int. J. Solids Struct. 45 (2008)
258275.
[8] X.-F. Li, A unied approach for analyzing static and dynamic behaviors of functionally graded Timoshenko and EulerBernoulli beams, J. Sound Vib. 318
(2008) 12101229.
[9] H.J. Xiang, J. Yang, Free and forced vibration of a laminated FGM Timoshenko beam of variable thickness under heat conduction, Composites: Part B 39
(2008) 292303.
[10] J. Ying, C.F. Lu, W.Q. Chen, Two-dimensional elasticity solutions for functionally graded beams resting on elastic foundations, Compos. Struct. 84 (2008)
209219.
[11] S. Kapuria, M. Bhattacharyya, A.N. Kumar, Bending and free vibration response of layered functionally graded beams: a theoretical model and its
experimental validation, Compos. Struct. 82 (2008) 340390.
[12] M.T. Piovan, R. Sampoia, A study on the dynamics of rotating beams with functionally graded properties, J. Sound Vib. 327 (2009) 134143.
[13] S.A. Sina, H.M. Navazi, H. Haddadpour, An analytical method for free vibration analysis of functionally graded beams, Mater. Des. 30 (2009) 741747.
[14] O.O. Oyekoya, D.U. Mba, A.M. El-Zafrany, Buckling and vibration analysis of functionally graded composite structures using the nite element method,
Compos. Struct. 89 (2009) 134142.
[15] M. Simsek, T. Kocaturk, Free and forced vibration of a functionally graded beam subjected to a concentrated moving harmonic load, Compos. Struct. 90
(2009) 465473.
[16] M. Simsek, Vibration analysis of a functionally graded beam under a moving mass by using different beam theories, Compos. Struct. 92 (2010) 904
917.
[17] P. Malekzadeh, M.R. Golbahar, M.M. Atashi, Out-of-plane free vibration of functionally graded circular curved beams in thermal environment, Compos.
Struct. 92 (2010) 541552.
[18] L.L. Ke, J. Yang, S. Kitipornchai, Nonlinear free vibration of functionally graded carbon nanotube-reinforced composite beams, Compos. Struct. 92
(2010) 676683.
[19] Y. Huang, X.F. Li, A new approach for free vibration of axially functionally graded beams with non-uniform cross-section, J. Sound Vib. 329 (2010)
22912303.
[20] L. Meirovitch, Methods of Analytical Dynamics, McGraw-Hill, Inc., 1970.
A.E. Alshorbagy et al. / Applied Mathematical Modelling 35 (2011) 412425 425