Вы находитесь на странице: 1из 12

Researches and Applications in Mechanical Engineering (RAME) Volume 2 Issue 1, March 2013

Thermal Sinusoidal Vibration and Transient Response of Magnetostrictive Functionally Graded Material Plates without Shear Effects

C.C. Hong

Department of Mechanical Engineering, Hsiuping University of Science and Technology, Taichung, 412 Taiwan, ROC

Abstract

The study of laminated magnetostrictive functionally graded material (FGM) plate without shear deformation under thermal sinusoidal vibration and transient response is calculated by using the generalized differential quadrature (GDQ) method. In the thermoelastic stress-strain relations that containing a power-law function of a two-material FGM plate , the linear temperature rise and the magnetostrictive coupling terms with velocity feedback control. Four edges of rectangular laminate d Terfenol-D FGM plate with simply supported boundary conditions are considered. The suitable product value of coil constant and control gain can be used to reduce the amplitude of center displacement into a smaller value .

Keywords

Magnetostrictive; FGM; Shear Deformation; Thermal Vibration; GDQ; Velocity Feedback Control

Introduction

is

usually made of different phases constituent materials, for example, the ceramic and metal used in

withstand

the

Typical functionally

graded

material

to

(FGM)

engine

combustion

chamber

ultra-high-temperature

and

to

reduce

the

stress

singularities,

respectively.

Chi

and

Chung

(2006)

presented the mechanical analysis

subjected to transverse load. There were several

of FGM plates

studies

in

the

transverse

displacements

for

the

laminated

plate

including

the

shear

deformation

effect. Amabili and Farhadi (2009) made the research of the shear deformable versus classical theories for nonlinear vibrations of rectangular isotropic and laminated composite plates. The over-prediction of natural frequencies in the solution without shear deformation was found. Ray and Shivakumar (2009)

the

analyzed the effect

of

shear

deformation

on

piezoelectric fiber-reinforced composite plate by using the finite element method (FEM). Nguyen et al. (2008) obtained the static numerical results for the FGM plate with the effect of shear deformation.

Magnetostrictive material Terfenol-D has the magneto-electric coupling property under the action of magnetism and mechanism. Hong (2009) used the computational generalized differential quadrature (GDQ) method to study the transient responses of magnetostrictive plates under thermal vibration. Thermal stresses and center displacement with and without shear effect were calculated in the thin and thick plate, respectively. Ramirez et al. (2006) obtained the free vibration solution for magneto-electro-elastic laminates through the Ritz approach. Lee and Reddy (2005) analyzed the non-linear response of laminated plate of magnetostrictive material under thermo-mechanical loading by using the FEM. Lee et al. (2004) obtained the transient vibration values of displacement for the Terfenol-D plate including the effect of shear deformation by using the FEM. Hong (2007) used the GDQ method to make the thermal vibration study for the Terfenol-D magnetostrictive laminated plate with the first-order shear deformation. Hong (2012) used the GDQ method to make the Terfenol-D FGM plate analyses under rapid heating induced vibration with the shear deformation effect. It is interesting to study thermal vibration of the transverse center displacement and thermal stress in the Terfenol- D FGM plate without the shear deformation effect by using the GDQ method.

Formulation

 

FGM

Most

materials

of

FGM

can

be

used

in

the

environment of higher

temperature

and

can

be

Researches and Applications in Mechanical Engineering (RAME) Volume 2 Issue 1, March 2013

expressed in

series

form

as

follows

by

Chi and

Chung (2006).

κ fgm

(

=κ +κ

2

1

)/ 2

.

(3e)

where E is the Young’s modulus, ν is the

is the thermal

expansion coefficients, κ is the thermal conductivity,

the subscript fgm represents the FGM plate, the

m subscripts 1 and 2 represent the constituent material

1 and 2, respectively. The property terms

=

n

m

∑

P

fgm

PV

i

i

n

. Poisson’s ratio, ρ

is the density, α

i = 1

is the material properties of FGM,

(1)

where

is the number of materials mixed to form the FGM,

P

fgm

E

1

E

2

V i is the volume fractions, and

n

m

∑

V

i

= 1

for all

ν

1

ν

2

ρ

1

ρ

2

α

1

α

2

κ 1 are expressed

κ

2

i = 1

corresponding to term

P

i

in equation (2).

constituent materials,

P i is the individual

constituent material properties, usually with the form as follows.

GDQ Method

derivatives

(

x

i

,

y

j

)

of

The GDQ method approximates the derivative of

. function, and the first-order and the second-order

coordinates

(i, j) can be discretized in

as

P

i =

P

0

(

P

1

T

1

+ 1 +

P , P

0

1

, P , P

1

2

T

PT

1

and

+

P T

2

P

3

2

+

P T

3

are

)

(2)

the

3

function

f

*

(

x

,

y

)

at

in

temperature coefficients, environment.

which

of grid point

is the temperature of

series forms by Shu and Du (1997 ) and rewritten follows:

= 2)

We denote the parameters for a two-material (

FGM plate as follows. a and b is the length in the

x , y direction of the plate, h * is the total thickness

of magnetostrictive layer and FGM plate,

thickness of magnetostrictive layer,

are the thickness of FGM material 1 and FGM material 2, respectively, p 1 and p 2 are the in-plane distributed forces, q is the applied pressure load.

The sum of volume fractions is in the form :

n

m

h

3

and

is the

h

2

h

1

V

1

+ V = 1

2

, the variation form of

V

2

= (

z

V

+ h / 2

2

)

power-law function is

h

used in the

R

n

, where

z

is the thickness of

FGM plate,

material properties for equation (1) can be assumed for the simple calculation and expressed as follows by Hong (2012).

is the power-law index. And the

is the thickness coordinate,

h

R

n

E

fgm

=

(

E

2

E

1

)(

z

+

h / 2

h

)

R

n

+

E

1

,

(

ν =ν +ν

fgm

2

1

ρ

fgm

α

fgm

= (

ρ +ρ

2

1

(

=α +α

2

1

(3a)

,

(3b)

,

(3c)

,

) / 2

)/ 2

)/ 2

 

(3d)

shear

stresses

and

shear

strains

in

the

laminate

* ∂ f ∂ x
*
f
x

shear strains in the laminate * ∂ f ∂ x i , i , j j

i

,

i

,

j

j

N

A

l = 1

(1)

i , l

M

m = 1

B

(1)

j , m

f

f

*

l

, j

,

(4a)

*

i m

,

,

(4b)

 
2 * ∂ f 2 i , j ∂ x
2
*
f
2
i
,
j
x

N

A

(2)

f

*

i , l

 

l

, j

,

l = 1

 

(4c)

 
2 * ∂ f 2 i , j ∂ y
2
*
f
2
i
,
j
y

M

B

(2)

i

,

m

 

j , m

,

m = 1

 

(4d)

 
2 * ∂ f ∂∂ x y ( m and B i , j )
2
*
f
∂∂
x
y
( m
and
B i
, j )

i

,

j

N

M

∑ ∑

A

(1)

i l

,

B

(1)

j , m

l

,

m

.

 

l = 1

m = 1

 

(4e)

where

, j )

( m

A i

 

denote

the

weighting

coefficients

for

 

the

m th -order

derivative

of

the

function

f

*

(

x

,

y

)

with

 

respect

to

x

and

 

y

directions.

 

Thermo

Magnetostrictive Effect

We

magnetostrictive FGM plate of the length a, b in the x ,

Elastic

Stress-Strain

a

Relations

with

consider

rectangular

laminated

uniformly

distributed load and thermal effect. There are no

without shear effect assumption. The plane stresses

y

direction,

respectively,

under

Researches and Applications in Mechanical Engineering (RAME) Volume 2 Issue 1, March 2013

in a laminated FGM plate with magnetostrictive

effect

following

equations by Lee and Reddy (2005):

for

the

k th

layer

are

in

the

ε ,ε

x

y

and

ε

xy

are

Q

16

the

= Q = Q = 0

26

in-plane

.

(6d)

strains,

45

σ

σ

x

y

σ

xy

(

k

)

=

Q

Q

11

Q

12

16

Q

12

Q

22

Q

16

Q

26


ε

ε

T

x

y

− − ∆ T

α

x

y

α

T = T x y t +

(

,

)

z

h

*

T

1

(

x

,

y

,

t )

ij

is the temperature

0

,

difference between the FGM plate and curing area,

z is the coordinate in the thickness direction, h * is

the

magnetostrictive

magnetic field intensity, expressed in the following

(2004),

feedback

equation

e ~ is the transformed

is the

Q

26

Q

0

0

0

− 

66

(

k

0

0

0

)

ε

~

e

~

e

~

e

xy

31

32

36

− ∆ T

α

xy

(

k

)



0

0

~

H

z

(

(

k

k

)

)

z

plate

total

(

x

,

y t

,

) =

by

k

c

~

I

(

x

,

y t

,

~

I

(

thickness,

coupling

Lee

~

I

(

x

,

y t

,

)

) =

( )

c t

w

t

modulus,

et

with

~

H

z

(5)

α y are the coefficients of thermal

expansion, α xy is the coefficient of thermal shear,

where α x and

al.

velocity

~

H

control

, in which

k

is the coil

c

Q ij is the stiffness of magnetostrictive FGM plate,

as

the simpler

 

follows.

 

B

Where

f , f

1

2

11

A

A

11

16

+

B

16

=

, f

f

 

1

f

2

f

3

FGM are

Q

11

=

Q

22

Q

12

Q

=

16

A

=

B

A

16

55

2

12

+

A

+

B

12

Q

66

+

66

B

2

u

0

2

u

0

∂∂ x y

=

66

=

=

1

0

0

0

1

0

ν

+ν

fgm

)

,

(6c)

A

A

66

26

+

B

26

2

v

0

B

2

x

0

2

A

A

16

0

66

v = v

0

(

B

12

2

v

A

+

0

y

2

x

,

x

,

y , t

in

)

+

A

12

2

B

66

A

26

+

66

2

2

w

x

2

y t

,

the

B

26

2

B

w

∂∂ x y

+ H

0

0

1

A

A

26

22

26

+

B

22

2

w

(

y

0

0

1

p

1

,

2

t

0

0

0

2

0

)

0

0

0

forms

of

Q ij for

given

E fgm

constant,

control gain.

)

is the coil current, c(t) is the

Dynamic Equilibrium Differential Equations

, Without shear deformation effect, the time dependent of tangential displacement equations is

1 ν

2

fgm

(6a)

fgm

E

fgm

assumed

and the transverse displacement

equation of the middle-plane is assumed in the form:

w = w(x, y,t) in which t is time. The dynamic

equilibrium differential equations in terms of displacements including the magnetostrictive loads are expressed in the following matrix forms by Hong

(2009):

forms:

u = u

0

(

x

,

y , t

)

1

ν fgm

2

,

(6b)

E fgm

2(1

Q

44

0

0

0

16

66

+

2

B

v

∂∂ x y

2

B

2

u

66

0

y

2

0

0

1

t

2

x

2

 ∂ u

 ∂ v

t

t

t

2

2

0

2

 ∂ w

2

2

 ∂ u

v

 ∂ t

2

0

2

p

2

,

q

+

ρ

(7)

3

are the expressions of thermal loads

(N , M ) , mechanical loads

and

magnetostrictive loads

~

(N

,

~

M )

,

14

Researches and Applications in Mechanical Engineering (RAME) Volume 2 Issue 1, March 2013

(

N

x

,

M

x

)

(

N

y

(

N

xy

,

M

y

)

,

M

xy

f

1

f

2

=

=

=

h

*

2

h

*

2

h

*

=

2

h

*

2

(

(

)

=

h

*

2

h

*

2

N

x

x

+

N

xy

y

N

xy

x

+

N

y

y

+

+

p

p

1

2

+

+

~

N

x

+

~

N

xy

x

N ~

xy

x

+

y

N ~

y

y

,

,

T

= [

T

0

(

x

,

w

=

y

) +

w

z

(

h

*

x

,

y

) sin(ω

mn

t

),

(8c)

T ( x, y)]sin( t )

1

γ

.

(8d)

f

3

=

M

x

x

+ 2

+

~

M

x

x

Q

11

α +

x

Q

12

α +

y

M

xy

+

M

y

where the term ω mn is the natural frequency of

plate, γ is the frequency of applied heat flux.

y

+ 2

~

M

xy

y

y

+

~

M

y

y

Q

16

α

xy

)(

T

0

,

z

z

h

*

T

1

)

+ q

We apply the weighting coefficients of discretized

, equations (4) in the two-dimensional GDQ method to

discrete the differential equations (7). And we use the

following non-dimensional parameters under sinusoidal temperature

dz

(T 0 = 0,

T

1

= T

1

sin(

π x

/

a

) sin(

π y

/

b

)

).

Q

12 α

x

(

Q

16

α

x

+ Q

22

α y

+

Q

26

α

y

~

( N

x

~

( N

y

,

,

~

M

x

~

M

y

)

)

=

=

+ Q

26

α xy

)(

T

0

,

+

Q

66

α

xy

)(

h

*

*

2

h

h

2

*

*

2

h

~

e

31

~

e

32

H

z

H

z

T

0

(1,

(1,

,

z

z

z

z

h *

z

h

*

,

T 1 )

dz

,

T

1

,

)

dz

2

)

dz

,

z

2

)

dz

,

= x / a ,

(9a)

Y = y /b,

(9b)

,

(9c)

,

(9d)

.

(9e)

Thus, we can obtain the dynamic discretized matrix

equations as in the paper by Hong (2009).

X

U

=

u

0

/ a

/b

0

=

10

V v

=

*

h w

/(

α

x

T a

1

2

)

W

2

(

A

ij

,

B

ij

,

D

ij

)

=

~

( N

h

*

*

2

h

2

xy

,

~

M

xy

)

Q

ij

(1,

z

,

=

z

2

h

*

*

2

h

2

) dz

~

e

36

H

z

(1,

z

2

)

dz

,

,

(i, j = 1,2,6),

(

ρ

,

H

)

=

h

*

*

2

h

2

ρ

0

(1,

z ) dz

.

in which

the in-plane distributed forces,

pressure load.

ρ

0

is the density of ply,

q

p 1 and

is

p 2 are

the applied

Dynamic Discretized Equations

Without the shear deformation effect and under the following vibration of time sinusoidal displacement and temperature:

u

v

= [

=

[

u

v

0

0

(

(

x

x

,

,

y

y

)]sin(

t

ω

mn

,

(8a)

,

(8b)

)

)]sin(

ω

mn

t

)

Some Numerical Results and Discussions

We like to consider the FGM plate consisting of two

materials, of which the FGM material 1 is SUS304

(Stainless

(Silicon Nitride). The temperature-dependent

used to calculate

coefficients

Steel),

P , P

0

the

FGM

material

P

3

2

is

Si N

3

4

1

, P , P

1

2

and

material property terms, and

E

1

E

2

ν

1

ν ρ

2

1

ρ

2

α

1

α

2

and

κ 1 of these two typical

κ

2

constituent materials are listed in Table 1 by Shariyat

(2008). The upper surface magnetostrictive layer of

the three-layer (0 / 0 / 0 )

under four sides simply supported is considered, the

superscript of m denotes magnetostrictive layer.

The elastic modules, material conductivity and

specific heat of the Terfenol-D magnetostrictive material are used the same value as in the paper by Hong (2007). We use the grid points for the GDQ

computation as in the following coordinates:

laminates FGM plate

°

m

°

°

14

Researches and Applications in Mechanical Engineering (RAME) Volume 2 Issue 1, March 2013

x

i

= 0.5[1

cos(

i

N

1

1

π )]

,

a i

= 1,2,

, N

,

(10a)

y j

=

0.5[1

cos(

j 1

M 1

b

π )] ,

j

=

1,2,

,

M

.

(10b)

The dynamic convergence results are obtained for

center displacement amplitude w (a / 2,b / 2)

without shear effects in the thermal vibration of

sinusoidal temperature only (T 0 = 0,

time

shape

T

1

= 100

°

mode

K ,

p 1 =

m = n = 1 , with

p 2 = q = 0 )

at

t =

6s,

value

control

gain

k c(t) = 0 ,

c

aspect ratio

ratio

a / b =

0.5, 1 and 2, side-to-thickness

mm,

,

a / h

3

= 1

(unit

h

T

= 300

mm)

21× 21

*

=

100, 50, 20, 10

mm,

h

1

=

h

2

=

°

K

. Table

in

the

2

grid

shows

point

29 × 29 and

and 5,

0.1

the

h

mm,

*

= 1.2

R

n

= 1

w (a / 2,b / 2)

N × M = 17 ×17 33 × 33 of GDQ

25 × 25

method for the laminated Terfenol-D FGM plate at

time t = 6 s. We find the N × M = 33 × 33 grid

point have the good w (a / 2,b / 2) convergence

result and use further in the GDQ analyses of time

responses for displacement and stress for 20 and 100.

of applied heat

Terfenol-D

1s,

frequency

We obtain the lowest frequency γ

= 5,

a / h

*

flux

FGM

and

plate

vibration

ω

11

time

of

t =

( m = n = 1 ), at

0.001s,

2s,…and

9s,

N × M

=

33 × 33 ,

h * =

1.2

mm,

h

3

= 1

mm,

h

1

=

h 0.1 mm,

2

=

a / b = 1 ,

q = 0 ,

shown in

Table 3. For the magnetic coil constant

the w (a / 2,b / 2) (unit mm) of vibration under the

constant gain value c(t) = 1, time t = 6 s,

without shear, we obtain the sketch of w(a / 2,b / 2)

vs.

R

n

= 1 ,

T

= 653

°

K ,

T

1

= 100

°

K

,

as

k

c

effect on

a

/

h

*

=

5

k

c

as shown in Fig. 1. The suitable product

and c(t) can be used to reduce the

(a / 2,b / 2) into a smaller value

value of

k

c

amplitude

of

w

near 0.00, there is an amplitude peak value at k c(t)

= 7.16E08, we find k c(t) = 1.56E09 for

c

a

/

h

*

=

5

c

thick plate can be chosen as the best suitable product values.

Firstly, thermal sinusoidal vibration is investigated with time step equal to 0.1s, the suitable chosen

product k c(t)

of coil constant and controlled gain

c

values

versus

time

t

for

R

n

a

/

h

*

=

5

and thin plate

a

T

=

653

°

K

, as shown in Table 4.

= 1

/

*

h

, thick

=

plate

at

20, 100

Fig. 2 shows the w (a / 2,b / 2) (unit mm) versus

time t of GDQ method for the laminated Terfenol- D

FGM plate

without shear effects. At time t = 0.001s, there is a

great amplitude value of displacement with

uncontrolled value ( k c(t)

,

plate

w (a / 2, b / 2) = -0.22975 mm for thin

of

are

smaller than the amplitudes of w (a / 2,b / 2) with

-0.156429 mm for

, 20 and 100, respectively

a

/

h

*

=

5

= 0), w (a / 2,b / 2) =

plate

mm

for

a

/

h

*

thin

=

5

the

amplitudes

k c(t)

c

values

c

thick

w (a / 2,b / 2) = -4.31174

a h

/

*

plate

=

a

/

20

h

*

,

=

w (a / 2,b / 2)

100

.

with

We

find

controlled

uncontrolled value ( k c(t) = 0), generally by using

c

the GDQ method. We can use the suitable product

to reduce the amplitude

values of controlled k c(t)

c

of w (a / 2,b / 2) into a smaller value near 0.00.

Fig. 3 shows the time response of the dominated

dimensional stress

of upper surface

for the laminated Terfenol-D FGM plate

20 and 100, respectively without shear effects

. We find the maximum

response values of

and 20,

controlled k c(t)

case are almost equal to the

response values of

uncontrolled case ( k c(t) = 0), generally by using the

GDQ method.

Fig. 4 shows the compared non-dimensional

, 20 and 100

without / with shear effects, for Terfenol-D FGM plate, at time t = 6s, m = n = 1, N × M = 33 × 33

for the case of without shear and 17 ×17 for the

=

,

0. In the shear

effect case, we use the YNS first-order shear deformation theory for the time dependent of displacement field, the value for shear correction

and

σ

x

=

c

σ

x (unit GPa) at center position

Z = 0.5h

*

with respective to time

a

/

h

*

=

5

,

σ

xy

= 0

σ

x

= -9.13E-04GPa for

a

/

h

*

=

5

for

a

/

h

*

=

100

with

= -9.12E-04GPa with

-9.12E-04GPa

σ

x

c

a

/

h

*

=

5

α

xy

= 0

W(X ,b 2,6) versus X of

x c a / h * = 5 α xy = 0 W ( X ,

case of with shear, h * = 1.2 mm,

h

2

T

=

0.1

= 653

°

mm,

K

,

T

1

a / b = 1

= 100

°

K ,

,

h

3

= 1

mm,

R

n

h

1

= 1

q = 0

,

k c(t)

c

=

15

Researches and Applications in Mechanical Engineering (RAME) Volume 2 Issue 1, March 2013

used in the dynamic

equilibrium differential equations by Hong (2012).

For

of

without

shear case are smaller than that with shear case. The maximum value W(X ,b 2,6) = 0.0821655 occurs at

coefficients

k

α

=

k

β

plate

= 5 / 6

a

/

h

*

the

thick

=

5

the

X

values

, displacement W(X ,b 2,6) versus

, displacement W ( X , b 2 , 6 ) versus X = 0.5 of
, displacement W ( X , b 2 , 6 ) versus X = 0.5 of

X = 0.5 of the with shear case, and W(X ,b 2,6) =

the with shear case, and W ( X , b 2 , 6 ) = -0.0325304

-0.0325304 occurs at X = 0.691342 of the without

, the values

shear case. For the thin plate

of displacement