You are on page 1of 45

TRANSVERSE VIBRATION OF A BEAM VIA THE FINITE ELEMENT METHOD

Revision E
By Tom Irvine
Email: tomirvine@aol.com
November 18, 2008
_______________________________________________________________________
Introduction
Many structures are too complex for analysis via classical method. Closed-form solutions
are thus unavailable for these structures.
For example, a structure may be composed of several different materials. Some of the
materials may be anisotropic. Furthermore, the structure may be an assembly of plates,
beams, and other components.
Consider three examples:
1. A circuit board has numerous chips, crystal oscillators, diodes, connectors,
capacitors, jump wires, and other piece parts.
2. A large aircraft consisting of a fuselage, wing sections, tail section, engines, etc.
3. A building has plies, foundation, beams, floor sections, and load-bearing walls.
The finite element method is a numerical method that can be used to analyze complex
structures, such as the three examples.
The purpose of this tutorial is to derive for a method for analyzing beam vibration using
the finite element method. The method is based on Reference 1.
Theory
Consider a beam, such as the cantilever beam in Figure 1.
EI,

Figure 1.
where
E is the modulus of elasticity.
I is the area moment of inertia.

L is the length.
is mass per length.
The product EI is the bending stiffness.
The vibration modes of the cantilever beam can be found by classical methods.
Specifically, the fundamental frequency is
.
187510
2 EI
1 =

(1)

This problem presents a good opportunity to compare the accuracy of the finite element
method to the classical solution.
Let y(x,t) represent the displacement of the beam as a function of space and time.
The free, transverse vibration of the beam is governed by the equation:

2
d2
2 y( x , t )
(
)
(
)
EI
x
y
(
x
,
t
)
=

x 2
x 2
t 2

(2)

Equation (2) neglects rotary inertia and shear deformation. Note that it is also independent
of the boundary conditions, which are applied as constraint equations.
Assume that the solution of equation (1) is separable in time and space.

y( x , t ) = Y ( x )f ( t )

(3)

2
2
2 Y ( x )f ( t )
EI(x ) 2 Y( x )f ( t ) = (x )

x 2
x
t 2

(4a)

2
2
2 f (t)
f (t)
EI(x ) 2 Y( x ) = Y( x ) (x )

x 2
x
t 2

(4b)

The partial derivatives change to ordinary derivatives.

d 2 d 2
d 2 f (t)
(
)
f (t)
Y( x ) = Y( x ) x
EI

dx 2 dx 2
d t2

(5)


1
d 2 d 2
1 d 2 f (t )
Y( x ) =
EI
Y( x ) (x ) dx 2 dx 2
f (t) d t 2

(6)

The left-hand side of equation (6) depends on x only. The right hand side depends on t
only. Both x and t are independent variables. Thus equation (6) only has a solution if both
sides are constant. Let 2 be the constant.

1
d 2 d 2
1 d2 f (t)
EI
Y
(
x
)
=

= 2

2
Y( x ) (x ) dx 2 dx 2
f
(
t
)

dt

(7)

Equation (7) yields two independent equations.

d 2
d2
EI( x )
Y( x ) ( x ) 2 Y( x ) = 0

dx 2
dx 2

d2
d t2

f ( t ) + 2 f ( t ) = 0

(8)

(9)

Equation (8) is a homogeneous, forth order, ordinary differential equation.


The weighted residual method is applied to equation (8). This method is suitable for
boundary value problems. An alternative method would be the energy method. The
energy method is introduced in Appendix A.
There are numerous techniques for applying the weighted residual method. Specifically,
the Galerkin approach is used in this tutorial.
The differential equation (8) is multiplied by a test function ( x ) . Note that the test
function ( x ) must satisfy the homogeneous essential boundary conditions. The essential
boundary conditions are the prescribed values of Y and its first derivative.
The test function is not required to satisfy the differential equation, however.
The product of the test function and the differential equation is integrated over the domain.
The integral is set equation to zero.

d 2

d2
Y( x ) ( x ) 2 Y( x ) dx = 0
( x )
EI( x )
2

dx 2
dx

(10)

The test function ( x ) can be regarded as a virtual displacement. The differential equation
in the brackets represents an internal force. This term is also regarded as the residual.
Thus, the integral represents virtual work, which should vanish at the equilibrium
condition.
Define the domain over the limits from a to b. These limits represent the boundary points
of the entire beam.
d 2

d2
( x )
Y( x ) ( x ) 2 Y( x ) dx = 0
EI( x )
2
a

dx 2
dx

d 2

d2
( x )
Y( x )
EI( x )
2
a

dx 2
dx

(11)

b
2
dx ( x ) ( x ) Y( x ) dx = 0
a

(12)
Integrate the first integral by parts.
b d

d
d2
(
x
)
EI
(
x
)
Y
(
x
)

2
dx
a dx

dx

b d
d2
d
dx
(
x
)
EI
(
x
)

Y
(
x
)

a dx
dx

dx 2

a (x) {(x)
b

dx

2 Y ( x ) dx = 0

(13)

d
d2

(
x
)
EI
(
x
)
Y
(
x
)

2
dx

dx

b d
d2
d

(
x
)
EI
(
x
)
Y
(
x
)

a dx
dx

dx 2

a (x) {(x)
b

dx

2 Y( x ) dx = 0

(14)

d
d2

(
x
)
EI
(
x
)
Y( x )

dx

dx 2

b d
d2
d

(
x
)
EI
(
x
)
Y( x )

a dx
dx

dx 2

dx

a (x) {(x)
b

2 Y ( x ) dx = 0

(15)
Integrate by parts again.

d
d2
Y( x )
EI( x )
( x )
dx

dx 2

b d
d

( x )

a dx dx

d2
Y( x )
EI( x )

dx 2

dx

2


b
b
d2
d
2
2 ( x ) EI( x ) 2 Y( x ) dx ( x ) ( x ) Y( x ) dx = 0
a
a dx


dx

(16)

d
d2

(
x
)
EI
(
x
)
Y( x )

dx

dx 2

( x )

dx
a

d2
Y( x )
EI( x )

dx 2

2


b
b
d2
d

(
x
)
EI
(
x
)
Y
(
x
)

dx ( x ) ( x ) 2 Y( x ) dx = 0

a
a dx 2


dx 2

(17)
The essential boundary conditions for a cantilever beam are
Y (a ) = 0

(18)

dY
=0
dx x =a

(19)

Thus, the test functions must satisfy


(a ) = 0

(20)

d
=0
dx x =a

(21)

The natural boundary conditions are

d2
d
Y( x )
EI( x )
dx

dx 2

=0

(22)

x =b

d2
Y( x )
=0
EI( x )

dx 2
x =b

(23)

Equation (23) requires

d2
EI( x )
=0
( x )
2

dx
x =b

(24)

Apply equations (20), (21), and (24) to equation (17). The result is

2


b
b
d2
d
2
2 ( x ) EI( x ) 2 Y( x ) dx ( x ) ( x ) Y( x ) dx = 0
a
a dx


dx

(25)

Note that equation (25) would also be obtained for other simple boundary condition cases.
Now consider that the beam consists of number of segments, or elements. The elements
are arranged geometrically in series form.
Furthermore, the endpoints of each element are called nodes.

The following equation must be satisfied for each element.

d2
( x )

dx 2

d2
Y( x )
EI( x )

dx 2

dx

( x ) ( x ) 2 Y( x ) dx = 0
(26)

Furthermore, consider that the stiffness and mass properties are constant for a given
element.

EI

d2
( x )

dx 2

d2
Y( x )

dx 2

2
dx

( x ) Y( x ) dx = 0
(27)

Now express the displacement function Y(x) in terms of nodal displacements y j1 and y j
as well as the rotations j1 and j .
Y( x ) = L1 y j1 + L 2 y j + L3 h j1 + L 4 h j1 ,

( j 1) h < x < j h

(28)

Note that h is the element length. In addition, each L coefficients is a function of x.


Now introduce a nondimensional natural coordinate .
= j x/h

(29)

Note that h is the segment length.


The displacement function becomes.
Y() = L1 y j1 + L 2 h j1 + L3 y j + L 4 h j1 ,

0 < <1

(30)

The slope equation is


Y ' () = L1' y j1 + L 2 ' h j1 + L3 ' y j + L 4 ' h j1 ,

0 < <1

The displacement function is represented terms of natural coordinates in Figure 2.

(31)

j1

Y(x)

y j1

yj

(j-1) h

jh

Figure 2.

Represent each L coefficient in terms of a cubic polynomial.


Li = ci1 + ci 2 + ci 3 2 + ci 4 3 ,

i =1, 2, 3, 4
(32)

{
}
+ { c 21 + c 2 2 + c 23 2 + c 2 4 3 }h j1
+ { c31 + c3 2 + c3 3 2 + c3 4 3 }y j
+ { c 41 + c 4 2 + c 4 3 2 + c 4 4 3 } h j ,

Y() = c11 + c1 2 + c13 2 + c1 4 3 y j1

0 < <1
(33)

{
}
+ { c 2 2 + 2 c 23 + 3 c 2 4 2 } h j1
+ { c3 2 + 2 c3 3 + 3 c3 4 2 }y j
+ { c4 2 + 2 c4 3 + 3 c4 4 2 }h j ,

Y ' () = c1 2 + 2 c13 + 3 c1 4 2 y j1

0 < <1
(34)

Solve for the coefficients c i j . The constraint equations are


Y ( 0) = y j

(35)

Y(1) = y j1

(36)

Y ' ( 0) = h j

(37)

Y ' (1) = h j1

(38)

Evaluate the displacement at = 0.

{ }

{ }

{ }

{ }

Y(0) = c11 y j1 + c 21 h j1 + c31 y j + c 41 h j

(39)

Boundary condition (35) requires

{ c11} y j1 + { c21}h j1 + { c31}y j + { c41} h j

= yj

(39)

c11 = 0

(40)

c 21 = 0

(41)

c31 = 1

(42)

c 41 = 0

(43)

The displacement equations becomes

}
+ { c 2 2 + c 23 2 + c 2 4 3 }h j1
+ {1 + c3 2 + c3 3 2 + c3 4 3 }y j
+ { c 4 2 + c 4 3 2 + c 4 4 3 } h j ,

Y() = + c1 2 + c13 2 + c1 4 3 y j1

0 < <1
(44)

The slope equations becomes

}
+ { c 2 2 + 2c 23 + 3c 2 4 2 }h j1
+ { c3 2 + 2c3 3 + 3c3 4 2 } y j
+ { c 4 2 + 2c 4 3 + 3c 4 4 2 } h j ,

Y ' () = c1 2 + 2c13 + 3c1 4 2 y j1

0 < <1
(45)

Evaluate the slope at = 0.


Y ' ( 0) =

{ c12 } y j1 + { c2 2 }h j1 + { c3 2 }y j + { c4 2 } h j

(46)

Boundary condition (37) requires.

{ c12 } y j1 + { c2 2 }h j1 + { c3 2 }y j + { c4 2 } h j = h j

(47)

c12 = 0

(48)

c2 2 = 0

(49)

c32 = 0

(50)

c 4 2 = 1

(51)

10

The displacement equations becomes

}
+ { c 23 2 + c 2 4 3 }h j1
+ {1 + c3 3 2 + c3 4 3 }y j
+ { + c 4 3 2 + c 4 4 3 } h j ,

Y() = + c13 2 + c1 4 3 y j1

0 < <1

(52)
The slope equations becomes

} y j1 + { 2 c2 3 + 3c2 4 2 }h j1
+ { 2 c3 3 + 3 c3 4 2 }y j + { 1 + 2 c 4 3 + 3 c 4 4 2 } h j ,

Y ' () = 2 c13 + 3 c14 2

0 < <1

(53)

Y(1) = c13 + c1 4 y j1 + c 2 3 + c 2 4 h j1
+ 1 + c 3 3 + c3 4 y j + 1 + c 4 3 + c 4 4 h j

(54)
Boundary condition (36) requires

{ c13 + c14 }y j1 + { c2 3 + c2 4 }h j1
+ { 1 + c3 3 + c 3 4 } y j + { 1 + c 4 3 + c 4 4 } h j

= y j1
(55)

c13 + c1 4 = 1

(56)

c13 = c1 4 + 1

(57)

c2 3 + c2 4 = 0

(58)

c2 3 = c2 4

(59)

11

1 + c33 + c3 4 = 0

(60)

c33 = 1 c3 4

(61)

1 + c4 3 + c4 4 = 0

(62)

c4 3 = 1 c4 4

(63)

The displacement equation becomes

{ [1 c14 ] 2 + c14 3} y j1
+ { [ c 2 4 ] 2 + c 2 4 3 } h j1
+ {1 + [ 1 c3 4 ] 2 + c3 4 3 } y j
+ { + [1 c 4 4 ] 2 + c 4 4 3 } h j ,

Y ( ) = +

0 < <1

(64)
The slope equation becomes

{[ ]
}
+ { 2 c 2 4 + 3 c 2 4 2 } h j1
+ { 2 [ 1 c 3 4 ] + 3 c3 4 2 } y j
+ { 1 + 2 [1 c 4 4 ] + 3 c 4 4 2 } h j ,

Y ' () = + 2 1 c1 4 + 3 c1 4 2 y j1

0 < <1
(65)

12

The slope equation becomes

{ [
]
}
{
}
+ { 2 [ 1 c 3 4 ] + 3 c3 4 } y j
+ { 1 + 2 [1 c 4 4 ] + 3 c 4 4 } h j

Y ' (1) = + 2 1 c1 4 + 3 c1 4 y j1
+ 2 c 2 4 + 3 c 2 4 h j1

(66)
Boundary condition (38) requires

{ [
]
}
{
}
+ { 2 [ 1 c3 4 ] + 3 c 3 4 } y j
+ { 1 + 2 [1 c 4 4 ] + 3 c 4 4 } h j = h j1

+ 2 1 c1 4 + 3 c1 4 y j1
+ 2 c 2 4 + 3 c 2 4 h j1

(67)

{ [ 2 2c14 ] + 3 c14 } y j1
+ { 2 c 2 4 + 3 c 2 4 } h j1
+ { [ 2 2c3 4 ] + 3 c3 4 } y j
+ { 1 + [ 2 2 c 4 4 ] + 3 c 4 4 } h j = h j1

(68)

2 + c1 4 = 0

(69)

c1 4 = 2

(70)

c 2 4 = 1

(71)

2 + c3 4 = 0

(72)

c3 4 = 2

(73)

13

1 + c 44 = 0

(74)

c 44 = 1

(75)

The displacement equation becomes

{[1 ( 2)] 2 2 3} y j1
+ { [ ( 1)] 2 13 } h j1
+ {1 + [ 1 2] 2 + 2 3 } y j
+ { + [1 ( 1) ]2 + ( 1)3 } h j ,

Y ( ) = +

0 < <1
(76)

{
} { 2 3}h j1
+ {1 3 2 + 2 3 } y j + { + 2 2 3 } h j ,

Y() = + 3 2 2 3 y j1 +

0 < <1

(77)
Recall

= j x/h

(78)

Thus

d = dx / h

(79a)

h d = dx

(79b)

d
= 1 / h
dx

(80)

d
d d
=
dx dx d

(81)

Note

14

{
} { 2 3}h j1
+ {1 3 2 + 2 3 } y j + { + 2 2 3 } h j ,

Y( x ) = + 3 2 2 3 y j1 +

( j 1)h x jh , = j x / h , 0 < < 1

(82)

{ [ 6 6 2 ] y j1 + [2 32 ] h j1
+ [ 1 6 + 6 2 ] y j + [ 1 + 4 3 2 ] h j },

d
Y( x ) = { 1 / h}
dx

( j 1)h x jh , = j x / h , 0 < < 1

(83)

d2
dx 2

{ }{ [ 6 12 ] y j1 + [2 6 ] h j1

Y( x ) = 1 / h 2

+ [ 6 + 12 ] y j + [ 4 6 ] h j ,
( j 1)h x jh , = j x / h, 0 < < 1
(84)

Now Let
Y( x ) = L T

a,

( j 1)h x jh , = j x / h

(85)

where
L1 = 3 2 2 3

(86)

L 2 = 2 3

(87)

L 3 = 1 3 2 + 2 3

(88)

L 4 = + 2 2 3

(89)

15

a=

y j1 h j1

yj hj

]T

(90)

The derivative terms are


1
d
Y( x ) = L' T a ,
dx
h

d2
dx 2

1
Y( x ) = L" T
h2

( j 1)h x jh , = j x / h

a,

( j 1)h x jh , = j x / h

(91)

(92)

Note that primes indicate derivatives with respect to .


In summary.
3 2 2 3

2 3

L=
2
3
1 3 + 2
+ 2 2 3
(90)
6 6 2

2 3 2

L' =
2
6 + 6
1 + 4 3 2
(91)
6 12
2 6

L" =
6 + 12

4 6
(92)

16

Recall
EI

d2

( x )

dx 2

d2

Y( x )

dx 2

2
dx

( x ) Y( x ) dx = 0
(93)

The essence of the Galerkin method is that the test function is chosen as
( x ) = Y( x )

(94)

Thus
EI

d2

Y( x )

dx 2

d2

Y( x )

dx 2

2
dx

[Y( x )] 2

dx = 0
(95)

Change the integration variable using equation (79b). Also, apply the integration limits.

h EI

2
d2

1
1
d
Y
(
x
)
Y
(
x
)

d h 2 [Y( x )] 2 d = 0

0 dx 2
0
dx 2

(96)

h EI

1
1
T
2 L"
0 h

h 2

0 [L
1

1
T
2 L"
h

][

a LT

d = 0
(97)

17

T
{ [ L"

1
1
EI
0
h3

h 2

] [ L" T a ] }d

0 [L
1

][

a LT

d = 0

(98)
T
T
{ [ a L" ] [ L"

1
1
EI
0
h3

h 2

0 [ a
1

T L

] } d

][ L T a ]

d = 0

(99)

1
1
EI
0
h3

a T L" L" T

}d

h 2

0 {
1

a T L LT

} d = 0
(100)

T }d h 2 1 { L LT } da = 0
{
L
"
L
"

EI 1
a T
3
h 0

T }d h 2 1 { L LT } d = 0
{
L
"
L
"

EI 1

h3 0

(101)

(102)

For a system of n elements,


K j 2 M j = 0 ,

j = 1, 2, ..., n

(103)

where

{ L"

EI 1
K j =
h3 0

L" T d

18

(104)

M j = h

T
0 { L L } d
1

6 12 [ 6 12
2 6
T

L" L" =
6 + 12

4 6

2 6

(105)

6 + 12

4 6]

(106)

L " L "T =
(6 12 )(6 12 )

(6 12 )(2 6) (6 12 )( 6 + 12 ) (6 12 )(4 6)
(2 6)(2 6) (2 6 )( 6 + 12 )
(2 6 )(4 6)
( 6 + 12 )( 6 + 12 ) ( 6 + 12 )(4 6)
(4 6)(4 6)
(107)

Note that only the upper triangular components are shown due to symmetry.
L " L "T =
(6 12 )(6 12 )

(6 12 )(2 6) (6 12 )( 6 + 12 ) (6 12 )(4 6)
(2 6)(2 6) (2 6 )( 6 + 12 )
(2 6 )(4 6)
( 6 + 12 )( 6 + 12 ) ( 6 + 12 )(4 6)
(4 6)(4 6)
(108)

19

L " L "T =
(6 12 )(6 12 )

(6 12 )(2 6) (6 12 )( 6 + 12 ) (6 12 )(4 6)
(2 6)(2 6) (2 6 )( 6 + 12 )
(2 6 )(4 6)
( 6 + 12 )( 6 + 12 ) ( 6 + 12 )(4 6)
(4 6)(4 6)
(109)

L " L "T =
36 144 + 1442 12 60 + 722

4 24 + 36 2

36 + 144 144 2
12 + 60 722
36 144 + 1442

24 84 + 722

8 36 + 362

24 + 84 722
16 48 + 362

(110)
Kj =

EI

h3

36 144 + 144 2 12 60 + 72 2

1
4 24 + 362
0

36 + 144 144 2
12 + 60 722
36 144 + 1442

24 84 + 722

8 36 + 36 2
d
24 + 84 722
16 48 + 36 2

(111)

20

Kj =

EI

h3

36 72 2 + 483 12 30 2 + 243

4 12 2 + 123

36 + 72 2 48 3
12 + 30 2 24 3
36 72 2 + 48 3

24 42 2 + 243

8 18 2 + 123

24 + 42 2 243
16 24 2 + 123

(112)

Kj =

EI

h3

36 72 + 48 12 30 + 24 36 + 72 48

4 12 + 12 12 + 30 24

36 72 + 48

24 42 + 24
8 18 + 12
24 + 42 24

16 24 + 12

(113)

EI
K j =
h3

12 6 12

4 6

12

6
2
6

(114)

21

3 2 2 3 3 2 2 3

2 3

T
LL =
2
3
1 3 + 2
+ 2 2 3

2 3

1 3 2 + 2 3

+ 2 2 3

(115)

L LT =

) (3 2 2 3 )(2 3 ) (3 2 2 3 )(1 3 2 + 2 3 ) (3 2 2 3 )( + 2 2 3 )
(2 3 )2
(2 3 )(1 3 2 + 2 3 ) (2 3 )( + 2 2 3 )
(1 3 2 + 2 3 )2 (1 3 2 + 2 3 )( + 2 2 3 )
( + 2 2 3 )2

2 2 3 2
3

(116)

a11 a12 a13

a 22 a 23
T
LL =

a 33

a14
a 24
a 34

a 44
(117)

a11 = 9 4 125 + 46

(118)

a12 = 3 4 55 + 26

(119)

a13 = 3 2 23 9 4 + 125 46

(120)

22

a14 = 33 + 8 4 75 + 26

(121)

a 22 = 4 25 + 6

(122)

a 23 = 2 3 3 4 + 55 26

(123)

a 24 = 3 + 34 35 + 6

(124)

a 33 = 1 6 2 + 43 + 9 4 125 + 46

(125)

a 34 = + 2 2 + 23 8 4 + 75 26

(126)

a 44 = 2 43 + 64 45 + 6

(127)

Recall
M j = h

T
0 { L L } d
1

M j , 11 = h

(128)

4
5
6
0 { 9 12 + 4 }d
1

4
9
M j , 11 = h 5 26 + 7
7 0
5

(129)

(130)

4
9
M j , 11 = h 2 +
7
5

(131)

13
M j, 11 = h
35

(132)

156
M j , 11 = h

420

(133)

23

M j , 12 = h

4
5
6
0 { 3 5 + 2 }d
1

(134)

5
2
3
M j , 12 = h 5 6 + 7
6
7 0
5

(135)

3 5 3
M j, 12 = h +
5 6 7

(136)

11
M j , 12 = h

210

(137)

22
M j , 12 = h

420

(138)

M j , 13 = h

2
3
4
5
6
0 { 3 2 9 + 12 4 }d
1

(139)

9
4
1
M j, 13 = h 3 4 5 + 26 7
7 0
2
5

(140)

1 9
4
M j , 13 = h 1 + 2
7
2 5

(141)

9
M j , 13 = h
70

(142)

54
M j , 13 = h

420

(143)

24

M j , 14 = h

3
4
5
6
0 { 3 + 8 7 + 2 }d
1

(139)

3
8
7
2
M j, 14 = h 4 + 5 6 + 7
5
6
7 0
4

3 8 7 2
M j , 14 = h + +
4 5 6 7

(141)

13
M j , 14 = h

420

M j , 22 = h

(140)

(142)

4
5
6
0 { 2 + }d
1

(143)

1
1
1
M j , 22 = h 5 6 + 7
3
7 0
5

1 1 1
M j , 22 = h +
5 3 7

(144)

(145)

1
M j , 22 = h

105

(146)

4
M j , 22 = h

420

(147)

M j, 23 = h

2
3
4
5
6
0 { 3 + 5 2 }d
1

25

(148)

1
1
3
5
2
M j, 23 = h 3 4 5 + 6 7
4
5
6
7 0
3

(149)
1 1 3 5 2
M j , 23 = h +
3 4 5 6 7

(150)

13
M j , 23 = h

420

(151)

M j , 24 = h

3
4
5
6
0 { + 3 3 + }d
1

(152)

1
3
1
1
M j, 24 = h 4 + 5 6 + 7
5
2
7 0
4

(153)

1 3 1 1
M j , 24 = h + +
4 5 2 7

(154)

3
M j , 24 = h

420

(155)

M j, 33 = h

2
3
4
5
6
0 {1 6 + 4 + 9 12 + 4 }d
1

(156)

9
4
M j, 33 = h 2 3 + 4 + 5 26 + 7
5
7 0

9
4
M j, 33 = h 1 2 +1 + 2 +
5
7

26

(157)

(158)

156
M j , 33 = h

420

(159)

2
3
4
5
6
0 { + 2 + 2 8 + 7 2 }d
1

M j , 34 = h

(160)

1
2
1
8
7
2
M j , 34 = h 2 + 3 + 4 5 + 6 7
5
6
7 0
3
2
2

(161)
1 2 1 8 7 2
M j , 34 = h + + +
2 3 2 5 6 7
(162)
22
M j , 34 = h

420

M j , 44 = h

(163)

2
3
4
5
6
0 { 4 + 6 4 + }d
1

(164)

1
6
2
1
M j , 44 = h 3 4 + 5 6 + 7
5
3
7 0
3

1
6 2 1
M j , 44 = h 1 + +
5 3 7
3
4
M j , 44 = h

420

(165)

(166)

(167)

27

Recall

{ L"

EI 1
K j =
h3 0

M j = h

L" T d

(168)

T
0 { L L } d

(169)

12 6 12 6

4 6
2
EI
K j =
12 6
h3

(170)
156 22 54 13

4 13 3
h
Mj=

156 22
420

(171)
Example 1
Model the cantilever beam in Figure 1 as a single element using the mass and stiffness
matrices in equations (170) and (171). The model consists of one element and two nodes
as shown in Figure 3.
E1

N2

N1

Figure 3.

Note that h=L.

28

The eigen problem is


6 12 6 y1
12
6
4
2 h 1 L 2
6
EI

=

L3 12 6 12 6 y 2 420

2
4 h 2
6
6

156 22 54 13 y1
22
4
13
3 h 1

54 13 156 22 y 2

4 h 2
13 3 22
(172)

6 12 6 y1
12
6
6
4
2 h 1

=
12 6 12 6 y 2

2
4 h 2
6
6

156 22 54 13 y1
22
4
13
3 h 1

54 13 156 22 y 2

4 h 2
13 3 22
(173)

where
L4 2

=
420 EI

(174a)

420 EI
=

L4

(174b)

The boundary conditions at node 1 are


y1 = 0

(175)

1 = 0

(176)

The first two columns and the first two rows of each matrix in equation (173) can thus be
struck out.

29

The resulting eigen equation is thus


12 6 y 2
156 22 y 2
6 4 h = 22
4 h 2

(177)
The eigenvalues are found using the method in Reference 2.
1 0.029715
= 2.8846

(178a)

1 0.1724

2 1.6984

(178b)

The finite element results for the natural frequencies are thus
1
=
2

420 EI 0.1724
1.6984
L4

(179)

1
=
2

3.5331

L4 34.807

(180)

EI

The finite element results are compared to the classical results in Table 1.
Table 1. Natural Frequency Comparison, 1 Element
Finite Element
Model

Classical
Solution

Index

L4
EI

L4
EI

3.5331

3.5160

34.807

22.034

30

The classical results are taken from Reference 3. The finite element results thus overpredicted the natural frequencies. Nevertheless, good agreement is obtained for the first
frequency.

Example 2
Model the cantilever beam in Figure 1 with two elements using the mass and stiffness
matrices in equations (170) and (171). Let each element have equal length.
The model consists of two elements and three nodes as shown in Figure 4.

E1

N1

E2

N2

N3

Figure 4.

There are several keys to this problem. One is that h=L/2. The other is that node N2
receives mass and stiffness contributions from both elements E1 and E2. Thus, the
resulting global matrices have dimension 6 x 6 prior to the application of the boundary
conditions.
The local stiffness matrix for element 1 is
6 12 6 y1
12

EI 6
6
4
2 h 1

(L / 2)3 12 6 12 6 y 2

6
2
4 h 2
6
(181)
The displacement vector is also shown in equation (181) for reference.

31

The local stiffness matrix for element 2 is


6 12 6 y 2
12

EI 6
6
4
2 h 2

(L / 2 )3 12 6 12 6 y 3

6
2
4 h 3
6

(182)

The local mass matrix for element 1 is


156 22 54 13 y1

3 h 1
4
13
(L / 2) 22

420 54 13 156 22 y 2

4 h 2
13 3 22

(183)

The local mass matrix for element 2 is


156 22 54 13 y 2

3 h 2
4
13
(L / 2) 22

420 54 13 156 22 y 3

4 h 3
13 3 22

(184)

The global eigen problem assembled from the local matrices is


6 12 6
0
0 y1
12
6
6
4
2
0
0 h 1

EI 12 6 24
0 12 6 y 2

(L / 2 )3 6
6
2
0
8
2 h 2

0
0 12 6 12 6 y 3

6
0
6
2
4 h 3
0
0
0 y1
156 22 54 13
22
4
13 3
0
0 h 1

0
54 13 y 2
(L / 2) 2 54 13 312
=

3 h 2
8
13
420 13 3 0
0
0
54
13 156 22 y 3

0 13 3 22
4 h 3
0
(185)

32

Again, the boundary conditions at node 1 are


y1 = 0

(186)

1 = 0

(187)

The first two columns and the first two rows of each matrix in equation (185) can thus be
struck out.
54 13 y 2
0 12 6 y 2
312 0
24

EI 0
3 h 2
8
13
6
8
2 h 2 (L / 2) 2 0

(L / 2 )3 12 6 12 6 y 3 420 54 13 156 22 y 3

4 h 3
6
2
4 h 3
13 3 22
6
(188)
Let
(L / 2) 2

420

=
EI

(L / 2)3

(189)

(L / 2)(L / 2 )3 2

420
EI

(190)

L4 2

=
6720 EI

(191a)

6720 EI
=

L4

(191b)

33

The eigen problem in equation (188) becomes


54 13 y 2
0 12 6 y 2
312 0
24

0
0
8
13
8
2 h 2
3 h 2
6

=
54 13 156 22 y 3
12 6 12 6 y 3

4 h 3
2
4 h 3
6
13 3 22
6
(192)
The eigenvalues are found using the method in Reference 2. Equation (192) yields four
eigenvalues.
1 0.0018414
0.073481
2 =

3 0.84056

4 7.0810

(193a)

1 0.042912

2 0.27107
=
3 0.91682

4 2.6610

(193b)

The finite element results for the first two natural frequencies are thus
1
=
2

6720 EI 0.042912
0.27107
L4

1
EI 3.5177
=

L4 22.221
2

(194)

(195)

34

The finite element results are compared to the classical results in Table 2.
Table 2. Natural Frequency Comparison, 2 Elements
Classical
Finite Element
Solution
Model
Index

L4
EI

L4
EI

3.5177

3.5160

22.221

22.034

Excellent agreement is obtained for the first two roots.


The next step would be to solve for the eigenvectors, which represent the mode shapes. A
greater number of elements would be required to obtain accurate mode shapes, however.
Example 3
Repeat example 1 with 16 elements. Let each element have equal length. The global
stiffness and mass matrix are omitted for brevity.
The eigenvalue scale factor is
420 (16 4 ) EI

=
4

(196)

The model yields 32 eigenvalues. The first four are


1 4.4913e - 007
1.7639e - 005
2 =

3 0.0001383

4 0.00053121

(197)

35

1

2 =
3

4

1 0.00067017

2 0.0041999
=
3 0.01176

4 0.023048

(198)

0.00067017

420 (16 4 )EI 0.0041999


0.01176
L4

0.023048

1

2 =
3

4

(199)

3.5160

EI 22.035
L4 61.698

120.92

(200)

The finite element results are compared to the classical results in Table 3.
Table 3. Natural Frequency Comparison, 16 Elements
Classical
Finite Element
Solution
Model
Index

L4

EI
3.5160

L4

EI
3.5160

22.035

22.034

61.698

61.697

120.92

120.90

Excellent agreement is obtained for the first four roots.


The corresponding mode shapes are shown in Figures 5 through 8. Note that each mode
shape is multiplied by an arbitrary amplitude scale factor. The absolute amplitude scale is
thus omitted from the plots.

36

Displacement

CANTILEVER BEAM, 16 ELEMENT MODEL, MODE 1

L
x

Figure 5.

Displacement

CANTILEVER BEAM, 16 ELEMENT MODEL, MODE 2

L
x

Figure 6.

37

Displacement

CANTILEVER BEAM, 16 ELEMENT MODEL, MODE 3

L
x

Figure 7.

Displacement

CANTILEVER BEAM, 16 ELEMENT MODEL, MODE 4

L
x

Figure 8.

38

Additional examples are given in Appendix C.

Reference
1. L. Meirovitch, Computational Methods in Structural Dynamics, Sijthoff & Noordhoff,
The Netherlands, 1980.
2. T. Irvine, The Generalized Eigenvalue Problem, 1999.
3. W. Thomson, Theory of Vibration with Applications, Second Edition, Prentice-Hall,
New Jersey, 1981.
4. K. Bathe, Finite Element Procedures in Engineering Analysis, Prentice-Hall, New
Jersey, 1982.

39

APPENDIX A

Energy Method
The total strain energy P of a beam is
2
d2y
1 L

P=
EI
dx
2
2 0
dx

(A-1)

The total kinetric energy T of a beam is

T=

1 2 L
n [ y ]2 dx
0
2

(A-2)

Again let
Y( x ) = L T

a,

( j 1)h x jh , = j x / h

1
d
Y( x ) = L' T a ,
dx
h

d2

1
Y( x ) = L" T
dx 2
h2

( j 1)h x jh , = j x / h

a,

( j 1)h x jh , = j x / h

d = dx / h

(A-3)

(A-4)

(A-5)

(A-6)

Assume constant mass density and stiffness.


The strain energy is converted to a localized stiffness matrix as

{ L"

EI 1
K j =
h3 0

L" T d

40

(A-7)

The kinetic energy is converted to a localized mass matrix as


1
0

M j = h

{ L L T } d

(A-8)

The total strain energy is set equal to the total kinetic energy per the Rayleigh method. The
result is a generalized eigenvalue problem.
For a system of n elements,
K j 2 M j = 0 ,

j = 1, 2, ..., n

(A-9)

where
12 6 12 6

4 6
2
EI
K j =
12 6
h3

(A-10)
156 22 54 13

4 13 3
h
Mj=

156 22
420

(A-11)

41

APPENDIX B

Beam Bending - Alternate Matrix Format


The displacement matrix for beam bending is
y1

1
y2

(B-1)

The stiffness matrix for beam bending is

12 6h

4h 2
EI

Kj =
h3

12

6h
6h 2h 2
12 6h

4h 2

(B-2)

The mass matrix for beam bending is


156 22h 54 13h

4h 2 13 3h 2
h
Mj=

156 22h
420

4h 2

42

(B-3)

APPENDIX C

Free-Free Beam
Repeat example 1 from the main text with a single element but with free-free boundary
conditions.

The eigen problem is


6 12 6 y1
12
6
6
4
2 h 1 L 2
EI

=

L3 12 6 12 6 y 2 420

6
2
4 h 2
6

156 22 54 13 y1
22
3 h 1
4
13

54 13 156 22 y 2

4 h 2
13 3 22
(C-1)

6 12 6 y1
12
6
6
4
2 h 1

=
12 6 12 6 y 2

6
2
4 h 2
6

156 22 54 13 y1
22
3 h 1
4
13

54 13 156 22 y 2

4 h 2
13 3 22
(C-2)

where
L4 2

=
420 EI

(C-3)

420 EI
=

4
L

(C-4)

43

The eigenvalues are found using the method in Reference 2. Equation (C-1) yields four
eigenvalues.

1 0
0
2 =

3 1.7143

4 20

(C-5)

1 0

2 0
=
3 1.3093

4 4.4721

(C-6)

The finite element results for the first four natural frequencies are thus
1

2 =
3

4
1

2 =
3

4

420 EI 0
L4 1.3093

4.4721

EI
L4

0
0

26.833

91.652

44

(C-7)

(C-8)

The finite element results are compared to the classical results in Table C-1.

Table C-1. Natural Frequency Comparison, 2 Elements


Classical
Finite Element
Solution
Model
Index

L4
EI

L4
EI

26.833

22.373

91.652

61.673

45

You might also like