ARTICLE IN PRESS

JOURNAL OF
SOUND AND
VIBRATION
Journal of Sound and Vibration 320 (2009) 322338
www.elsevier.com/locate/jsvi

Vibration and acoustic response of a composite plate with

inherent material damping in a thermal environment
P. Jeyaraj, N. Ganesan, Chandramouli Padmanabhan
Machine Design Section, Department of Mechanical Engineering, Indian Institute of Technology Madras, Chennai 600 036, India
Received 29 May 2008; received in revised form 25 July 2008; accepted 2 August 2008
Handling Editor: L.G. Tham
Available online 20 September 2008

Abstract
This paper presents numerical studies on the vibration and acoustic response characteristics of a ber-reinforced
composite plate in a thermal environment by considering the inherent material damping property of the composite
material. Initially the critical buckling temperature is obtained, followed by free and forced vibration analyses considering
the pre-stress due to the imposed thermal environment. The vibration response predicted is then used to compute the sound
radiation. The critical buckling temperature and vibration response are obtained using the nite element method based
on the Classical Laminate Plate Theory (CLPT) while sound radiation characteristics are obtained using a coupled
FEM/BEM technique. It is found that the vibration response of the structure reduces with an increase in uniform
temperature rise for both Glass Epoxy and PEEK/IM7 materials, but the overall sound radiation of the plate reduces only
marginally due to interaction between reduced stiffness and enhanced damping.
r 2008 Elsevier Ltd. All rights reserved.

1. Introduction
Damping is an important factor in the dynamic design of any engineering component as it inuences the
vibration and noise levels signicantly. It also controls the fatigue life and impact resistance of the structures.
Fiber-reinforced composite structures are extensively used in aerospace, automobile, ship-building industries
and other engineering applications due to their high ratio of strength and stiffness to weight. Fiber-reinforced
composite materials have inherent material damping because of bermatrix interaction which is usually more
than that of the conventional isotropic materials.
Structures can be exposed to moisture and heat during their service life. Thus, knowledge of the dynamic
behavior of a structure over a range of temperatures is essential for the design of structures in a thermal
environment. Thermal stresses due to aerodynamic heating may induce buckling and dynamic instability in
structures.The pre-stress effect due to thermal load will affect the dynamic behavior of the structure due to the
change in the stiffness of the structure.

Corresponding author. Tel.: +91 44 2257 4690.

E-mail address: [email protected] (C. Padmanabhan).

0022-460X/$ - see front matter r 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jsv.2008.08.013

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

323

Initial investigations on the damping analysis of ber-reinforced composite materials have been reviewed
extensively by Chandra et al. [1]. Rikards [2] used a modal strain energy-based nite element method to obtain
the loss factors of a laminated composite plate. Melo and Radford [3] analyzed the time and temperature
dependence of the viscoelastic properties of Carbon Fiber Reinforced Plastic (CFRP) material experimentally.
They found that there is a signicant change in the loss factor compared to the elastic modulus due to a change
in temperature.
The buckling behavior of ber-reinforced composite plates subjected to a thermal load has been analyzed
extensively with the help of both analytical [4] and nite element methods [5,6]. The pre-stress effect due to
thermal load will affect the free vibration behavior of the structure and the natural frequencies of the structure
decreases with an increase in thermal load. The modal loss factor increases with the temperature; when the
uniform temperature rise approaches the critical buckling temperature there is a signicant increase in the
modal loss factor [7].
Temperature and moisture variations lead to a thermo-elastic-dynamic response in a structure as well as
noise radiation from such structures [8]. Park et al. [9] investigated the effects of support properties on the
sound radiated from the plate and found that both the velocity response and the sound radiation are strongly
inuenced by dissipation of vibration energy at the edges. Qiao and Huang [10] analyzed six different
boundary conditions to investigate the inuence of boundary conditions on the sound radiation of a plate
under a harmonically excited point force. They found that boundary conditions have a large effect on the
sound radiated from rectangular plates. The aim of the present work is to analyze the vibration and acoustic
response characteristics of a composite plate under a thermal environment subjected to mechanical timevarying harmonic excitations by considering the inherent material damping associated with the composite
materials.
2. Formulation
The Finite Element Method (FEM) is used to predict the critical buckling temperature, natural frequencies,
loss factors and vibration response of the plate. A two-dimensional four-noded quadrilateral thin-plate
element having ve degrees of freedom per node formulated based on the Classical Laminate Plate Theory
(CLPT) is used in the present work, which leads to the following displacement eld; more details of the
formulation can be found in [11].
9
8
qw >
8 9 >
>
>
u

z

>
>
>
qx >
= >
<u>
=
<
qw
v
(1)
v z >
>
; >
: >
>
>
> 
qy >
w
>
>
;
:
w
where u; v and w are the components of the displacement along x, y and z directions with x and y on the plane
of the plate and z along the thickness direction. The subscript  refers to the reference plane. The strain
displacement relations are given by
9
9 8 2
8
q w >
qu > >
>
>
>
>
> z 2 >
>
>
>
>
> >
> qx >
>
qx
9 >
8
>
>
>
>
>
>
>
>
>
> 2
> qv

>
>
=
= >
< q w >
<
= >
< x >

z
y
2
(2)


qy
qy >
>
>
>
>
>
>
>
; >
:g >
>
>
>
>
2
>
xy
>
> >
> q w >
> qu qv >
>
>
>
>
>
>
> >
> 2zqxqy >
> qy
qx >
>
;
>
:
;
:
9 8 
8


>
= >
< x >
< x
y
y

>
>
>
: g ; : g
xy

9
>
=

9
8
k
>
=
< x>
z ky ;
>
>
;
;
:k >
xy
xy

fg f g z fkg

(3)

ARTICLE IN PRESS
324

P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

where  and k are the strain and the curvature of the reference plane. The constitutive relations for a
laminated plate accounting for thermal effects can be written as [11]
0

sx

Q11

Q12

B s C BQ
@ y A @ 12
sxy
Q16

Q22
Q26

Q16

10

x  ax DT

B
C
Q26 C
A@ y  ay DT A
xy  axy DT
Q66

(4)

where Qij is the transformed reduced stiffness, ax ; ay and axy are the coefcients of thermal expansion along
x, y and shear directions and DT is the temperature difference (assumed to be uniform). The stress resultants
and stress couples are given by
9
8
8
s
N
>
= Z h=2 >
< x>
< x
Ny
sy

>
h=2 >
;
:N >
:s
xy

xy

9
>
=
>
;

dz;

9
8
8
M
s
>
= Z h=2 >
< x>
< x
My
sy

>
h=2 >
;
:M >
:s
xy

xy

9
>
=
z dz
>
;

(5)

The total potential energy P of the plates is P U b V , where U b is the strain energy of bending and V is the
potential energy due to external loads. A four-noded quadrilateral plate element with ve degrees-of-freedom
(dof) is used for the nite element solution of the laminates. The shape function of the displacement eld is
dened as
u v w yx yy T Nfdg

(6)

where N and fdg are the shape function matrix and the nodal variable vector, respectively. The structure
stiffness and mass matrices of the plate are obtained by using the minimum potential energy principle. The
structure stiffness matrix K can be written as
K 

"

Z Z

B D B dx dy;

Dn 

Aij
Bij

Bij
Dij

#
(7)

where Aij , Bij and Dij are the coefcients of extensional, bendingextensional coupling and bending stiffness
matrices. These are obtained as
Z

h=2

Qij 1; z; z2 dz;

Aij ; Bij ; Dij

i; j 1; 2; 6

(8)

h=2

Since all the elastic moduli associated with the composite materials considered in the present analysis are
complex, the structural stiffness matrix K is complex and can be decomposed as
K K R  jK I 

(9)

where K R  and K I  are the real and imaginary parts of the structure stiffness matrix, respectively. The mass
matrix of the plate can be written as
M

h=2
h=2

h=2

NT P N dx dy

(10)

h=2

where P is the mass property matrix. The geometric stiffness matrix K s  is obtained from the work done by
the membrane forces developed due to thermal load. The membrane forces S x , Sy and Sxy are specied in
terms of membrane stresses sx , sy and shear stress txy developed due to thermal load and are calculated from a
static analysis. The nonlinear membrane strains associated with lateral deection o of the plate can be

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

written as

  9
1 qo 2 >
>
>
>
>
9
8
>
2 qx
>

x
>
>
>
 2 >
=
=
<
1 qo
y

>
>
2 qy
>
; >
:g >
>
>
   >
>
>
xy
>
>
>
qo qo >
>
>
>
>
>
>
: qx
qy ;
8
>
>
>
>
>
>
>
>
>
<

325

(11)

The change in membrane strain energy associated with constant membrane forces and the nonlinear
membrane strains can be written as
 2
 2
  #
Z "
1
qo
1
qo
qo qo
Um
Sx
Sy
S xy
dA
(12)
2
qx
2
qy
qx
qy

Um

9
8
qo >T "
>
>
Z Z >
<
qx = S x
1

qo > Sxy
>
2
>
>
;
:
qy

9
8
qo >
#>
>
>
<
S xy
qx =
qo >
Sy >
>
>
;
:
qy

dx dy

(13)

Now the displacement vector is given by

fqg

4
X
N i fqi ge

(14)

i1

The derivatives are:

9
8
qo >
>
>
=
< qx >
qo >
>
>
>
;
:
qy

(15)

Bg fqg

This, when substituted into Eq. (13), leads to

Um

1
fqgT K s fqg;
2

K s 

"

Z Z
Bg 

Sx

Sxy

S xy

Sy

#
Bg  dx dy

(16)

3. Analysis approach
FEM is used to nd the critical buckling temperature, the effects of thermal load on the natural frequencies
and the vibration response of the plate. The thermal load is assumed to be created in the plate due to a
uniform temperature distribution across the surface of the plate. The thermal load applied on the structure will
induce membrane forces, which in turn inuence the lateral deections associated with the plate. The
resistance to bending deformation is reduced when membrane forces are compressive. These pre-loads on the
plate due to the thermal environment are calculated using a static analysis. The pre-stressed modal and
harmonic analyses are carried out with critical buckling temperature as a parameter to analyze the effect of
thermal load on the natural frequencies and vibration response. The entire analysis approach is summarized in
Fig. 1. When the temperature of the plate is increased from the ambient by DT, thermal stresses develop in the
plate (for any boundary condition with at least one edge restrained). This stress state (static) is used to
calculate the geometric stiffness matrix K s . Following this a buckling analysis is carried out using the
structural and geometric stiffness matrices K and K s 
K li K s fci g 0

(17)

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

Thermal
environment
(uniform
temperature)

Pre-stress
(Static
analysis)

Critical Buckling
Temperature (Tcr)
(Buckling analysis)

Natural Frequencies and Modal

loss factors (Pre-stressed modal
analysis)

Vibration response (Pre-stressed

harmonic response analysis)

Evaluation of RMS velocity from

pre-stressed harmonic response
analysis results

T as a parameter

326

Sound radiation calculation using

coupled FEM/BEM method with
the help of average RMS velocity

Fig. 1. A owchart of the analysis approach.

where li is the ith eigen value and ci is the corresponding eigen vector. The product of the lowest eigenvalue l1
and the temperature rise DT yields the critical buckling temperature, T cr , that is T cr l1 DT. Physically, T cr
denes the temperature at which the plate buckles due to thermal stresses. Since the structure is pre-loaded due
to the thermal eld, the natural frequencies of the structure are modied as these loads produce stresses which
change the structural stiffness. Pre-stressed modal analysis is carried out to nd the natural frequency of the
pre-loaded structure. The natural frequency at any given temperature can be calculated by evaluating the
geometric stiffness matrix at that temperature and by solving the eigenvalue problem as given below:
K R  K s   o2k Mffk g 0

(18)

where [M] is the structural mass matrix, while ok is the circular natural frequency of the pre-stressed structure
and fk the corresponding mode shape. Similarly, the kth modal loss factor Zk at any given temperature can
be obtained from the following equation:
Zk

ffk gT K I ffk g
ffk gT K R K s ffk g

(19)

After the computation of the natural frequencies, loss factors and mode shapes a pre-stressed harmonic
analysis is carried out to determine the vibration response of the pre-loaded structure. The general equation of
motion for a pre-stressed structure is
CfUg
_ K K s fUg fF tg
(20)
MfUg
U_ and U are the acceleration,
where F t is the applied load vector (assumed to be time-harmonic) and U,
velocity and displacement vector of the plate. Using a set of modal coordinates yk , k 1; 2; . . . ; n, where n is
the total number of degrees of freedom. The transformation from physical coordinates to modal coordinates
can be written as
fUg

n
X
ffk gyk ffyg

(21)

k1

where f is the modal matrix. Eq. (20) can then be written as

Cffyg
_ K R  K s ffyg fF tg
Mffyg

(22)

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

327

Pre-multiplying the above equation by fT and using the orthogonality relations after applying the
orthogonal and normal conditions Eq. (22) becomes
yk 2ok xk y_k o2i yk F k ;

k 1; 2; . . . ; m

(23)

where F k f fF tg. Since xk Zk =2 the above equation becomes

yk ok Zk y_k o2i yk F k

(24)

The vibration response is obtained by solving the above uncoupled equation.

In the Boundary Element Method, the boundary is divided into N elements and the Helmholtz equation is
discretized for a given node i so that the discrete form of the integral equation can be written as
N Z
N Z
X
X
qg
qp
cpi 
(25)
p dS 
g dS
qb
n
qb
n
j1 S j
j1 S j
where c is a coefcient which is either 1, 0 or 12 depending on whether the eld point is in the volume, outside
the volume or on the smooth boundary, p is the acoustic pressure, Sj is the surface of the element j on the
boundary, g is the free space Greens function in the frequency domain and nb is the surface unit normal vector.
As the normal velocity on the surface v i=ro oqp=qb
n Eq. (25) becomes
Z
N
N Z
X
X
cpi 
pg dS iro o
gv dS
(26)
j1

Sj

j1

Sj

where g qg=qb
n. For a node on the constant element ( p and v are constant over each element) Eq. 26 can be
written as
N Z
N Z
X
X
1
p 
g dSpj iro o
g dSvj
(27)
2 i j1 Sj
j1 S j
R
R
If one denes the inuence coefcients H ij Sj g dS; G ij Sj g dS, which are integrals relating node i, where
the fundamental solution is acting, to any other node j. For a particular node i Eq. (27) takes the form
N
N
X
X
1
pi 
H ij pj iro o
G ij vj
2
j1
j1

(28)

by introducing H ij 12 dij  H ij where dij is Knoneckers delta. Eq. (28) can now be written as
N
X

H ij pj iro o

j1

N
X

G ij vj

(29)

j1

If Eq. (29) is repeated for every node point i, a system of equations obtained can be expressed in matrix
form as
Hfpg iro oGfvg

(30)

where H and G are the boundary integral inuence matrices, fpg and fvg are the acoustic pressure and the
normal velocity of the boundary element nodes of the uids, respectively, at the uidstructure interface and
ro is the density of the uid. The reader is referred to Ref. [12] for detailed information on the formulation of
direct BEM/FEM.
4. Validation studies
4.1. Validation for critical buckling temperature evaluation
A single-layer thin (side/thickness 40) Kevlar/Epoxy square plate with ber orientation angle 45  and
having all edges clamped analyzed by Huang and Tauchert [4] is considered for the validation of the critical
buckling temperature. Huang and Tauchert [4] used an analytical method to predict the critical buckling

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

328

temperature while the present method uses a four-noded quadrilateral element formulated based on the
Classical Laminate Plate Theory (CLPT). The material properties of Kevlar/Epoxy analyzed are as follows [4]:
E 1 76 GPa; E 2 5:5 GPa; G12 G 13 2:3 GPa; G 23 1:5 GPa; n12 0:34; a1 4  106 =  C; and a2
79  106 =  C where E 1 and E 2 are Youngs modulus of the lamina along major (ber direction) and minor
axes, respectively, G 12 , G 13 and G 23 are shear moduli of the lamina, while n12 is the major Poissons ratio. The
coefcients of thermal expansion along major and minor axes are a1 and a2 , respectively. The critical buckling
temperature evaluated using the present approach is 131  C which matches very well with the value reported
by Huang and Tauchert [4] which is 132  C.
4.2. Validation for natural frequencies and loss factors
A simply supported, four-ply laminated carbon ber reinforced plastic (CFRP) square plate with bers
oriented at an angle 45  analyzed by Alam and Asnani [13] is considered for the validation of frequency and loss
factor. The ratio of the length to the thickness of the plate considered is 150. The material properties of the CFRP
considered are E 1 211 GPa; E 2 5:31 0:5j GPa; G12 G 13 2:61 0:5j GPa; G23 1:31 0:5j GPa;
n12 0:25; and density r 1524 kg=m3 . Alam and Asnani [13] used a variational principle to obtain the resonant
frequency parameter and loss factors while FEM is used here. Table 1 shows the comparison of the resonant
frequency parameter lr for the exural mode I and the extensional modes II and III and the associated system loss
factor Z. The expression for resonant frequency parameter is lr rt2 o2 =E 1 where r is the mass density of the
material considered, t is the thickness of the plate and o is the resonant frequency in radians per second. From
Table 1 it is clear that both the natural frequencies and the modal loss factors obtained in the present work agree
very well with the results reported by Alam and Asnani [13].
4.3. Validation of sound radiation calculation
A simply supported isotropic rectangular plate with dimensions 0:455 m  0:379 m  0:003 m analyzed by Li
and Li [14] is considered for the validation of sound power-level calculation using SYSNOISE. The plate is
made of steel with Youngs modulus E 210 GPa, density r 7850 kg=m3 and Poissons ratio n 0:3. The
plate is assumed to be vibrating in air whose density is ra 1:21 kg=m3 with a speed of sound c 343 m=s.
A harmonic excitation of 1 N is applied at the center of the plate and a structural damping ratio of 0.01 for
all modes is assumed for harmonic response analysis. Li and Li [14] used four-noded isoparametric elements
for their structural nite element model and calculated sound power level using the Rayleigh integral.
In the present work, the plate is modeled using 16  16 plate elements formulated based on the Classical
Laminate Plate Theory. The computed natural frequencies are compared with those reported by Li and Li [14]
in Table 2. The average rms velocity is calculated for each frequency from the pre-stressed harmonic analysis
results. The sound power level of the plate is calculated using SYSNOISE by assuming that the entire plate is
vibrating with an average rms velocity at each frequency. Fig. 2 show the comparison of sound power level
obtained using average rms velocity (present) with the results reported by Li and Li [14]. The modes
corresponding to modal indices (2,1), (2,2) and (1,2) will not be excited as the harmonic excitation force is at
the center of the plate. Both the natural frequencies and the sound power levels match very well with those of
Li and Li [14].
Table 1
Comparison of resonant frequency parameter and loss factor
Mode

I
II
III

Loss factor Z

Resonant frequency parameter lr

Alam and Asnani [13]

Present work

Alam and Asnani [13]

Present work

0:204  108
0:688  106
0:284  104

0:203  108
0:682  106
0:281  104

0.0135
0.499
0.0124

0.0132
0.496
0.0121

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

329

Table 2
Comparison of natural frequencies (Hz) with Li and Li [14]
Modal indices

Li and Li [14]

Present work

(1,
(2,
(1,
(2,
(3,

87
193
241
345
374

87
193
241
346
372

1)
1)
2)
2)
1)

100
Present
Li and Li [14 ]

Sound Power Level (dB)

80

60

40

20

0
0

100

200

300
400
500
Frequency (Hz)

600

700

800

Fig. 2. Comparison of sound power level with Li and Li [14].

5. Results and discussion

A rectangular composite plate clamped at all its edges with dimensions 0:5 m  0:4 m  0:01 m is now
considered for the detailed investigation. The plate is analyzed for two different ber-reinforced composite
materials namely Glass Epoxy and PEEK/IM7. The elastic material properties of the Glass/Epoxy are
assumed to be temperature independent while the PEEK/IM7 properties are temperature dependent. The plate
is assumed to have only one layer of ber. The plate is assumed to be vibrating in air whose density is
ra 1:21 kg=m3 with a speed of sound c 343 m=s. The material properties of the Glass Epoxy composite are
E 1 37:781 0:0014j GPa; E 2 10:91 0:0008j GPa; G 12 G13 4:911 0:011j GPa; n12 0:3;
r 1870 kg=m3 ; a1 7  106 =  C; and a2 2:3  105 =  C and then temperature dependence of the
viscoelastic properties of PEEK/IM7 composite materials is given in Table 3. The values of a1 and a2 of the
PEEK/IM7 material are 0:15  106 =  C and 26:57  106 =  C, respectively. A convergence study has been
carried out for the critical buckling temperature and natural frequencies without considering any thermal load
for zero ber orientation as given in Table 4. Based on the convergence study the plate is modeled using a
16  16 mesh.
5.1. Evaluation of critical buckling temperature T cr
In the present work, the vibration and acoustic response of a ber-reinforced composite plate for different
ber orientations has been analyzed by assuming that the structure is subjected to a uniform temperature rise
above ambient temperature. The uniform temperature rise applied on the plate is varied from 0 to 100  C for
each ber orientation and the corresponding variation in natural frequencies, loss factors, vibration and

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

330

Table 3
Temperature dependence of the viscoelastic properties of PEEK/IM7 [3]
Temperature  C

E 1 GPa

E 2 GPa

G12 GPa

n12

n23 103

Z1 ; 103

Z2 ; 103

Z6 ; 103

20
0
20
40
60
80
100
120

155.4
155.1
154.9
154.7
154.5
154.3
154.0
153.8

10.2
10.1
10.0
10.0
9.9
9.8
9.7
9.6

7.4
7.3
7.2
7.1
7.1
7.0
7.0
6.9

0.34
0.34
0.34
0.34
0.35
0.35
0.35
0.35

0.48
0.49
0.49
0.49
0.49
0.5
0.5
0.5

4.7
4.5
5.1
5.2
5.7
6.8
7.2
7.0

7.8
7.5
7.2
7.7
8.5
9.3
10.1
12.1

8.8
10.5
10.6
11.5
13.5
14.6
14.8
14.9

Table 4
Convergence study for critical buckling temperature T cr and natural frequencies
Critical buckling temperature
T cr  C

Natural frequency (Hz)

(1, 1)
(1, 2)
(2, 1)
(2, 2)
(1, 3)

Mesh size

10  10

12  12

14  14

16  16

18  18

132

131

131

131

131

266
510
568
770
912

264
509
567
768
908

264
508
567
767
906

264
508
566
766
905

264
508
566
766
905

Table 5
Critical buckling temperature for different ber orientations for glass epoxy
Fiber orientation

Critical buckling temperature  C

0
15
30
45
60
75
90

133
129
123
127
144
174
186

acoustic response has been analyzed. The uniform temperature rise has been limited to 100  C in order to keep
the temperature below the glass transition temperature of the polymer material which is around 140  C. To
start with, the critical buckling temperature is obtained for different ber orientations for Glass/Epoxy
composite material and this is given in Table 5. In Table 5, it must be noted that the critical buckling
temperature exceeds the glass transition temperature of the polymer matrix for the ber orientations 60 , 75
and 90 . Thus, the plate is analyzed for different ber orientations ranging from 0 to 45 in steps of 15 .
Similarly, the critical buckling temperature of a PEEK/IM7 plate with a 0 ber orientation is 307  C for the
dimensions 0:5 m  0:4 m  0:01 m, which also exceeds the glass transition temperature. Hence a plate with
dimensions 0:5 m  0:4 m  0:005 m is considered to analyze the PEEK/IM7 composite material for which the
critical buckling temperature obtained is 74  C. The pre-stressed modal analysis is carried out for different
values of ber orientation by varying the uniform temperature rise applied on the plate from 0  C to 100  C, in
order to determine the inuence of the thermal environment on the natural frequencies and corresponding

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

331

Table 6
Variation of natural frequencies and loss factors with temperature
Uniform
temperature ( C)

First mode

Second mode

Third mode

Fourth mode

Fifth mode

0
30
60
90
100

264(0.0049)
234(0.0062)
197(0.0088)
151(0.0153)
132(0.0262)

510(0.0073)
471(0.0085)
426(0.0102)
377(0.0134)
359(0.0163)

573(0.0028)
543(0.0031)
510(0.0035)
475(0.0041)
462(0.0047)

770(0.0053)
732(0.0059)
691(0.0066)
648(0.0075)
633(0.0081)

912(0.0078)
871(0.0086)
825(0.0096)
777(0.0110)
760(0.0124)

15

0
30
60
90
100

265(0.0051)
233(0.0064)
195(0.0093)
146(0.0168)
126(0.0536)

514(0.0065)
496(0.0075)
432(0.0098)
383(0.0126)
365(0.0137)

568(0.0036)
515(0.0039)
498(0.0043)
459(0.0051)
445(0.0060)

768(0.0055)
730(0.0061)
688(0.0069)
643(0.0079)
628(0.0093)

938(0.0074)
897(0.0081)
851(0.0091)
803(0.0101)
786(0.0114)

30

0
30
60
90
100

268(0.0051)
235(0.0066)
194(0.0098)
141(0.0190)
117(0.0215)

504(0.0051)
467(0.0062)
425(0.0076)
379(0.0099)
361(0.0105)

583(0.0051)
544(0.0057)
501(0.0067)
453(0.0080)
436(0.0093)

765(0.0058)
727(0.0065)
684(0.0073)
638(0.0084)
622(0.0094)

947(0.0036)
908(0.0040)
866(0.0045)
822(0.0051)
807(0.0061)

45

0
30
60
90
100

279(0.0047)
245(0.0061)
204(0.0089)
152(0.0163)
130(0.0193)

478(0.0053)
440(0.0063)
397(0.0078)
347(0.0102)
329(0.0128)

643(0.0045)
604(0.0051)
561(0.0059)
514(0.0071)
497(0.0088)

753(0.0060)
713(0.0067)
669(0.0076)
623(0.0085)
606(0.0096)

889(0.0043)
854(0.0048)
804(0.0053)
757(0.0059)
740(0.0078)

0
10
30
50
70

241(0.0060)
233(0.0062)
214(0.0085)
190(0.0132)
129(0.0713)

343(0.0072)
324(0.0078)
278(0.0125)
215(0.0232)
157(0.0269)

542(0.0077)
519(0.0081)
453(0.0122)
379(0.0191)
293(0.0324)

614(0.0055)
608(0.0059)
601(0.0064)
590(0.0078)
573(0.0153)

683(0.0063)
673(0.0065)
647(0.0077)
624(0.0100)
585(0.0132)

Fiber orientation
rise ( C)
Glass/Epoxy
0

PEEK=IM-7

modal loss factors and mode shapes. The results obtained from the pre-stressed modal analysis are given in
Table 6. In Table 6, the values given in bracket represent modal loss factors associated with the corresponding
modes. As the temperature rise is limited to 100  C in the present study (i.e., the plate is not analyzed close to
the critical buckling temperature due to glass transition temperature issues) the natural frequencies of the
fundamental mode for different ber orientations are not near zero. However, from the pre-stressed modal
analysis carried out near the critical buckling temperature it is found that the fundamental mode frequency
does go to zero and the loss factor increases to a very high value. This is due to a decrease in stiffness due to an
increase in temperature. The mode shapes, however, are not signicantly affected by thermal environment as
seen in Fig. 3. For the PEEK/IM7 material, loss factors obtained by considering temperature-dependent and
temperature-independent elastic properties are compared and it is found that there is no signicant variation
in loss factor.
5.2. Vibration and acoustic response studies
In order to compare the vibration response and sound radiation characteristics, a frequency range of
021500 Hz is chosen for different ber orientations. Before carrying out the harmonic response analysis, an
appropriate point of location of excitation of the harmonic force is chosen using the mode shapes of the plate
for different ber orientations. The location of excitation is chosen in such a way that it should not lie on the
nodal lines of modes in the frequency range of 021500 Hz; this is done at room temperature but since
the mode shapes are independent of temperature the excitations would still not coincide with any nodal
lines. The pre-stressed harmonic analysis is carried out by applying a 1 N load at the location chosen for each

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

332

T = 0C

T = 60C

T = 100C

MODE 1

MODE 2

MODE 3

MODE 4
Fig. 3. Mode shapes at different values of uniform temperature rise.

Fig. 4. Displacement at the point of excitation for a 45 ber orientation of a Glass/Epoxy composite plate.

ber orientation. The displacement and velocity response at the point of excitation are obtained for each ber
orientation. The average rms velocity of the plate has been calculated for each frequency and for each ber
orientation to give as an input for the sound radiation calculation. The direct boundary element method with
the bafe option is used in SYSNOISE to obtain the acoustic characteristics. The sound power level and
radiation efciency for each ber orientation is obtained.
Figs. 46 and 8 show the displacement response,velocity response, average rms velocity and sound
power level, respectively, as the excitation frequency is varied for a 45  ber orientation. Two trends can be
seen: (i) the natural frequencies reduce with increasing temperature and (ii) the resonant amplitudes decrease

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

333

Fig. 5. Velocity at the point of excitation for a 45 ber orientation of a Glass/Epoxy composite plate.

Fig. 6. Average rms velocity for a 45; ber orientation of a Glass/Epoxy composite plate.

with an increase in temperature. The trend of reduced resonant amplitudes cannot be clearly seen in higher
modes of displacement. This, however, can be seen clearly in the velocity, average rms velocity and sound
power-level responses. Generally, the pre-stress due to thermal load will reduce the stiffness of the structure
which in turn increases the displacement amplitude of vibration. This is not seen in the response at resonant
frequencies. This implies that the modal damping has a signicant counterinuence. Since the modal damping
increases signicantly with the temperature it reduces the vibration amplitude at the resonant frequencies.
Similar responses are observed for other ber orientations.
Fig. 7 shows the radiation efciency for a 45  ber orientation as a function of frequency. Sound power is
directly proportional to the product of the square of the averaged surface normal velocity and radiation
efciency. Orthotropic plates have two coincidence frequencies: one corresponding to bending waves in the
major direction (ber direction) f cx and another one corresponding to bending waves in the minor direction
f cy . The values of f cx and f cy were obtained from the analytical expression given by Ohlrich and Hugin [15].

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

334

1.6
1.4

Radiation efficiency

1.2
1.0
0.8
0.6
T = 0C

0.4

T = 60C
T = 100C

0.2
0.0
0

600

1200

1800 2400 3000 3600 4200

Harmonic Frequency (Hz)

4800

5400

6000

Fig. 7. Radiation efciency for a 45 ber orientation of a Glass/Epoxy composite plate.

Fig. 8. Sound power level for a 45 ber orientation of a Glass/Epoxy composite plate.

For the composite plate analyzed in the present work these frequencies are 1598 and 2564 Hz, respectively.
Fig. 7 shows that radiation efciency crosses one at the lower coincidence frequency and decreases towards
one only after the second coincidence frequency.
From Fig. 7 it can be seen that the radiation efciency of the plate generally decreases (no signicant
variation in the low-frequency range) with an increase in temperature. The sound power level shown in Fig. 8
reects the average rms velocity response as sound radiation is directly related to the normal velocity of the
structure at a particular frequency.For further investigation average of mean square velocity is calculated for
the 45  ber orientation in the entire frequency band as shown in Fig. 9 and for constant bandwidth frequency
bands (250 Hz) as shown in Fig. 10. From Figs. 9 and 10 one can see that the velocity generally decreases with
temperature. When the uniform temperature rise reaches 100  C, which is closer to the critical buckling

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

335

Fig. 9. Overall average rms velocity for a 45 ber orientation of a Glass/Epoxy composite plate.

Fig. 10. Average rms velocity for a 45 ber orientation of a Glass/Epoxy composite in constant frequency bands.

temperature for the different ber orientations analyzed, there is a marginal reduction in the overall average
rms velocity.
From the bandwise representation also it is clear that the rms velocity is higher when there is no rise in
uniform temperature (except at the lower band) and is lower when the uniform temperature rise is 100  C. This
is due to a signicant increase in the modal loss factor when the uniform temperature rise is closer to the
critical buckling temperature of the structure. In the lower frequency band, the vibration response is
inuenced by the stiffness of the structure; it can be clearly seen in the displacement, velocity and average rms
velocity responses.
To obtain a clearer picture, the overall sound power level for the entire frequency band is computed for
different ber orientations and the results are shown in Fig. 11. For all ber orientations, it is clear that even
though the overall sound power level is not signicantly affected by the thermal environment, there is a

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

336

88
Overall Sound Power Level (dB)

Fiber orientation

0 deg
15 deg
30 deg
45 deg

87.5
87
86.5
86
85.5
85
0

20

40

60

80

100

T (C)
Fig. 11. Overall sound power level for different ber orientations of a Glass/Epoxy composite plate.

95

Sound Power Level (dB)

90

T = 0C
T = 60C
T = 100C

T = 30C
T = 90C

85
80
75
70
65
60
0-250

250-500

500-750
750-1000
Frequency Band (Hz)

1000-1250

1250-1500

Fig. 12. Output power level for a 45 ber orientation of a Glass/Epoxy composite in constant frequency bands.

distinct downward trend when the uniform temperature rise approaches the critical buckling temperature.
Fig. 12 shows the sound power level represented in constant bandwidth frequency bands for the 45  ber
orientation. The shift in natural frequencies towards the lower bands due to an increase in uniform
temperature rise will in turn reduce the normal velocity which is directly related to the sound power radiated.
This can be seen in Fig. 12, where the sound power level decreases with an increase in uniform temperature rise
beyond the initial increase from DT 0 to 30  C.
The resonant amplitude of vibration response and acoustic response increases with temperature rise for an
isotropic plate while the resonant amplitudes decrease with an increase in temperature rise for the composite
plate with inherent material damping. Even though the structural stiffness reduces with an increase in
temperature, the modal loss factor reduces the resonant amplitude as it increases signicantly with
temperature rise for the composite plate. Due to this reason, there is no signicant change in the overall sound
power level of the composite plate also. This would suggest that one could use a composite material having
inherent material damping for structures subjected to a thermal eld. Figs. 13 and 14 show the sound power

ARTICLE IN PRESS
P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

337

Fig. 13. Output power level for a PEEK/IM7 composite plate.

Fig. 14. Overall sound power level for a PEEK/IM7 composite plate.

response and overall sound power level of the PEEK/IM7 composite plate. From Fig. 13 one can see similar
changes and acoustic responses can be seen for the PEEK/IM7 composite plate also. From Figs. 13 and 14 it is
clear that there is no signicant inuence of temperature-dependent material properties on the vibration and
acoustic response characteristics. This is again due to the competing effects of increased damping and reduced
stiffness.

ARTICLE IN PRESS
338

P. Jeyaraj et al. / Journal of Sound and Vibration 320 (2009) 322338

6. Conclusion
The effect of a thermal environment on the vibration response and consequent sound radiation from a
composite plate for different values of ber orientation is investigated. The calculations are performed for
different ber orientations and the clampedclamped boundary conditions are assumed. Displacement and
velocity at the point of excitation, average rms velocity, radiation efciency and output sound power level are
computed for different boundary conditions to show the inuence of thermal loading on vibration response
and sound radiation. It is found that the amplitudes of vibration at the resonant frequencies decrease with the
increase in temperature. The resonant amplitude is less when the uniform temperature rise approaches the
critical buckling temperature of the structure. From the studies carried out on PEEK/IM7, it is found that the
inuence of temperature-dependent material properties is not signicant on the vibration and acoustic
response of structures in a thermal environment. There is no signicant change in the overall sound power
level of a composite plate with material damping compared to an isotropic plate because of complementing
effects of reduced stiffness and increased damping. The vibration and sound radiation of a composite plate is
signicantly inuenced by inherent material damping when compared with an isotropic plate.
References
[1] R. Chandra, S.P. Singh, K. Gupta, Damping studies in ber-reinforced composites a review, Composite Structures 46 (1999) 4151.
[2] R. Rikards, Finite element analysis of vibration and damping of laminated composites, Composite Structures 24 (1993) 193204.
[3] J.D.D. Melo, D.W. Radford, Time and temperature dependence of the viscoelastic properties of CFRP by dynamic mechanical
analysis, Composite Structures 70 (2005) 240253.
[4] N.N. Huang, T.R. Tauchert, Thermal buckling of clamped symmetric laminated plates, Impact and Buckling of Structures, ASME
Aerospace Division 20 (1990) 5359.
[5] H.R.H. Kabir, H. Askar, R.A. Chaudhuri, Thermal buckling response of shear exible laminated anisotropic plates using a threenode isoparametric element, Composite Structures 59 (2003) 173187.
[6] K.R. Thangaratnam, R. Palaninathan, J. Ramachandran, Thermal buckling of composite laminated plates, Composite Structures 32
(1989) 11171124.
[7] N. Ganesan, V. Pradeep, Buckling and vibration of rectangular composite viscoelastic sandwich plates under thermal loads,
Composite Structures 77 (2007) 419429.
[8] C.S. Lyrintzis, Hygrothermal effects on structure borne noise transmission of stiffened laminated composite plates, Journal of Aircraft
77 (1990) 722730.
[9] J. Park, L. Mongeau, T. Siegmund, Inuence of support properties on sound radiated from the vibrations of rectangular plates,
Journal of Sound and Vibration 264 (2003) 775794.
[10] Y. Qiao, Q. Huang, The effect of boundary conditions on sound loudness radiated from rectangular plates, Archieve of Applied
Mechanics 264 (2007) 2134.
[11] O.C. Zienkiewicz, The Finite Element Method in Engineering Science, McGraw-Hill Publishing Company Limited, London, 1971.
[12] Z. Tong, Y. Zhang, Z. Zhang, H. Hua, Dynamic behavior and sound transmission analysis of a uidstructure coupled system using
the direct-BEM/FEM, Journal of Sound and Vibration 299 (2007) 645655.
[13] N. Alam, N.T. Asnanai, Vibration and damping analysis of ber-reinforced composite materialplates, Journal of Composite
Materials 20 (1986) 218.
[14] .S Li, X. Li, The effects of distributed masses on acoustic radiation behavior of plates, Applied Acoustics 69 (2008) 272279.
[15] M. Ohlrich, C.T. Hugin, On the inuence of boundary constraints and angle bafe arrangements on sound radiation from
rectangular plates, Journal of Sound and Vibration 277 (2004) 405418.