1

DYNAMICS OF ROTATING
MACHINES


Michael I Friswell, John E T Penny, Seamus D Garvey and
Arthur W Lees


SOLUTION MANUAL

Version 1. July 2011










The authors welcome any comments and corrections.

Supporting MATLAB scripts and functions have been written with emphasis on
clarity, not necessarily on efficiency or compactness.
2

Chapter 2

Problem 2.1
From Equation (2.24), the response is given by ( ) ( )
0 0
cos sin
n
t
d d
x t e a t b t

= + .
Thus ( ) ( )
0 0 0 0
cos sin sin cos
n n
t t
n d d d d d
x e a t b t e a t b t

= + + + .
When 0 t = , ( )
0 0
0 x x a = = and ( )
0 0 0
0
n d
x x a b = = + .
Hence ( )
0 0 0 n d
b x x = + .
Now 1kg m = , 9N m k = ,
3
0
10 m x

= ,
0
0 x = .
Thus 0 = and
0
0 b = , 9 3rad s = = = =
n d
k m . So ( ) ( )
3
10 cos3 m

= x t t .
If 1Ns m c = , 2 1
n
c m = = . Then ( ) ( ) 2 1 2 3 0.1666
n
c m = = = and
2 2
1 3 1 0.1666 2.958rad s = = =
d n
. ( )
0 0 0
0.1666 3 2.958 0.169 = = b x x .
Hence
( ) ( ) ( ) { }
3
10 cos 2.958 0.169sin 2.958 m x t t t

= +

Problem 2.2
For the shaft,
4 10 4
0.01m, 0.40m, 32 9.818 10 m d L J d

= = = = .
9 10
80 10 9.818 10 0.15 523.60Nm
sh
k GJ L

= = = .
For the disk,
4 2
32 0.1960kgm I hD = = . Thus
523.60 0.1960 51.68rad = = =
sh
k I s . ( ) ( ) 2 51.68 2 8.23Hz = = .

Problem 2.3
2
k m = . Thus ( )
2
1 2 700 27635N m 27.635kN m k = = = .

Problem 2.4
50 2 314.159rad s = = .
3
0
1mm 10 m x

= = .
3
0 0
314.159 10 0.3142ms v x

= = = ,
2 2 3 2
0 0
314.159 10 98.7m s a x

= = = .

Problem 2.5
Critical damping 1 = . Now 2
n
c m = so that
( ) 2 2 1 2 700 8796.5Ns m
c n
c m = = = .

Problem 2.6
( ) ( )
0
cos
n
t
d
x t e c t

= and ( )
( )
( ) ( )
0
cos
n
t T
d
x t T e c t T
+
+ = + . Now
( ) ( ) ( ) cos cos
d d
t T t + = so that
( )
( )
( )
n
n
n
t T
T
t
x t T
e
e
x t
e
+

+
= = .
Hence
( )
( )
n
T
x t
e
x t T

=
+
. Thus
( )
( )
2
2 2
log
1
e n n
d
x t
T
x t T
| |

= = = =
|
|
+
\ .
.
3

Problem 2.7
Let ( ) cos sin
s s s
x t A t B t = + . Substitute into equation (2.29) we have
( ) ( )
( ) ( )
2
2
0
cos sin 2 sin cos
cos sin cos
s s s s
n s s
A t B t A t B t
A t B t f m t
+ + +
+ + =

Collecting together terms in cos t and sin t and letting
0
1 f = gives
( ) { }
( )
( ) { }
2 2
2 2
2 cos 1 cos
2 sin 0
n s n s
n s n s
A B t m t
B A t
+ =
=

Dividing by
2
n
and noting that generally, cos 0 t and sin 0 t we have
( ) ( )
( )
2 2
2
1 2 1
1 2 0
s s n
s s
r A rB m
r B rA
+ =
=

Solving this pair of equations gives expressions for
s
A and
s
B identical to Equation
(2.33).

Problem 2.8
In Equation (2.35) let
2
r = .
Then,
( )
( ) ( )
( ) ( )
2
12
2 2
2
2 2
1
1
1 2
1 2
n
s
n
m
C
m

(
= = +

+
.
Differentiating this expression gives
( )
( ) ( ) ( )( ) ( )
3 2
2 2 2
2
1
1 2 2 1 1 2
2
s
n
d C
d
m

( (
= + +

.
Setting this expression to zero to determine the value of for the maximum response,
we have
( )( ) ( ) ( )
2
2
2 1 1 2 2 1 4 0
(
+ = + =

. Thus at maximum response
(resonance)
2
1 2 = and hence
2
1 2 r = . Thus the resonant frequency
r
is
given by
2
1 2 =
r n
. Note that
r
is not equal to
d
.

Problem 2.9
5Hz 10 s. 10 11 1.1
n
r = = = = .
(a) From Equation (2.34) it is clear that the response is proportional to the input
force. Doubling input force doubles response. No change in phase.
(b) If the phase of the input force changes, the phase of the response changes by
the same amount. Magnitude of response unchanged.
(c) If the forcing frequency is halved, r is halved. Using Equation (2.35) with
0 = ,
4

then
( )
( )
( )
( )
2 2
1.1
2
2
1 1
0.21
1 1.1
n n
s
r
m m
C
=

= =

,
( )
( )
( )
( )
2 2
0.55
2
2
1 1
0.6975
1 0.55
n n
s
r
m m
C
=

= =

.
Thus
( )
( )
0.55
1.1
0.21
0.3011
0.6975
s
r
s
r
C
C
=
=
= = - a reduction to 30% of the original response.
Because the forcing frequency is now below the natural frequency, the phase
changes by 180 to zero.
(d) Initially,
( )
( )
( )
( )
2 2
0
2
2
1 1
0.21
1 1.1
n n
s
m m
C
=

= =

.
With damping,
( )
( )
( )
( )
( )
2 2
0.1
2
2
2
1 1
0.3041
1 1.1 2 0.1 1.1
n n
s
m m
C
=

= =
+
,
Thus
( )
( )
0.1
0.0
0.21
0.6905
0.3041
s
s
C
C
=
=
= = - a reduction to 69% of the original response.
( )
1 1
0.1 2 2
2 2 0.1 1.1
tan tan 46
1 1 1.1
s
r
r

=
| | | |
= = =
| |
\ . \ .
. i.e. 46 180 134 + = .

Problem 2.10
In this case ( ) f t t T = . Thus from Equation (2.43)
0
0 b = and
( )
2
0
0 0 0
1 1 1 1
2 2
T
T T
t t
a f t dt dt
T T T T T
= = = =


From Equation (4.42) ( ) ( )
0
0 0
2 2 2
sin sin
T T
n
nt
b f t nt dt f t dt
T T T
| |
= =
|
\ .

.
Let 2 nt T x = so that ( ) 2 dt T n dx = . When 0, 0 t x = = and when , 2 t T x n = = .
Hence
( ) ( )
( )
2
2
0 2 2
0
2 2
sin sin cos 1
2 2
n
n
n
b x x dx x x x n
n n

= = =

.
Similarly,
( ) ( )
2
2
0 2 2
0
2 2
cos cos sin 0
2 2
n
n
n
a x x dx x x x
n n

= = + =

.
Thus the Fourier series for the function is
1
1 1 2
sin
2
n
t nt
T n T

=
| |
=
|

\ .

.
In this example, 2 T = . The MATLAB script, Problem_02_10.m, uses this series to
determine the Fourier approximation of the function using 10, 50 and 100 terms in the
series and provides the following graphical output.
5

-2 0 2 4
0
0.5
1
10 term Fourier series
Time (s)
F
o
r
c
e

(
N
)
-2 0 2 4
0
0.5
1
50 term Fourier series
Time (s)
F
o
r
c
e

(
N
)
-2 0 2 4
0
0.5
1
100 term Fourier series
Time (s)
F
o
r
c
e

(
N
)

Problem 2.11
For this system
1 0
kg
0 2
(
=
(

M and
5 4
N m
4 5
(
=
(

K . The equations of motion are
+ = Mq Kq 0 . Solving these equations for free vibrations leads to an eigenvalue
problem of the form = Ku Mu where
n
= . This equation can be solved by
finding roots of the quadratic
2
2 15 9 0 + = . Alternatively the eigenvalue problem
can be solved directly, see the MATLAB script Problem_02_11.m. Running this
script gives the following output:

1st mode u2/ u1 = 1. 0856
2nd mode u2/ u1 = - 0. 46058

1st nat ur al f r equency = 0. 81097 r ad/ s
2nd nat ur al f r equency = 2. 6158 r ad/ s

1st nat ur al f r equency = 0. 12907 Hz
2nd nat ur al f r equency = 0. 41632 Hz

Problem 2.12
The radial force due to each out of balance mass is
2
u u
f m r = , Thus when the
masses have rotated from the vertical by an angle , then in the vertical direction the
sum of the forces is
2 2 2
cos cos 2 cos
vert u u u
f m r m r m r t = + = , since
t = . (Refer to Figure 2.36). In the horizontal direction the sum of the forces is
2 2
sin sin 0
hz u u
f m r m r = = . Thus there is no horizontal force and, so, in the
x direction the equation of motion becomes
2
2 cos
u
mx kx m r t + = .
6

If a vibration absorber is added, and the displacement of this mass is
a
x then the
equation of motion for free vibrations are
0 0
0 0
a a a a a
a a
k k m x x
k k k m x x
( (
+ =
` ` `
( (
+
) ) )

.
Thus, assuming a harmonic solution, we have
2
2
0
0
a a a a
a a
k m k x
x
k k k m
(


= (
` `
( ) ) +

. Thus
( )( )
2 2 2
0
a a a a
k m k k m k + = .
Multiplying out, and dividing by
a
m m gives
4 2
0
a a a
a a
k k k k k
m m m m
| | +
+ + =
|
\ .
.
The forced vibration equations are
2
2
0
cos
a a a a
vert
a a
k m k x
t
f x
k k k m
(

= (
` `
( ) + )


Inverting the 2 2 coefficient matrix, we have
( )
2
2
2
0
1 1
cos cos

(
+
= = (
` ` `

( ) )


)
a vert
a a a
vert a a vert
a a a
k f
k k m k x
t t
f k m f x D D
k k m
.
where D is the determinant of coefficient matrix. The response x is zero if
2
a a
k m = .
The response of the absorber is ( ) cos
a vert
k D f t where
( )( )
2 2 2
a a a a
D k m k k m k = + . When
2
a a
k m = ,
2
a
D k = .
Thus ( )
2
2
1 cos cos

= =
u
a a vert
a
m r
x k f t t
k


Problem 2.13
The equation of motion are Mq +Kq = 0 and assuming a harmonic solution,
= Ku Mu where
n
= . The system mass and stiffness matrices are
10 0 0
0 10 0 kg
0 0 40
(
(
=
(
(

M ,
2 3 3 2
3 3 4 4
2 4 1 2 4
20 10 10
kN
10 20 10
m
10 10 40
k k k k
k k k k
k k k k k
+ ( (
( (
= + =
( (
( ( + +

K
2 3
1.1861 1
1.1861 , 1
1.0000 0


= =
` `

) )
u u . The products
T
2 3
u Mu and
T
2 3
u Ku are both zero.
Now, the triple product
T
3 3
20 = u Mu .
Thus { } { }
T T
3 3
20 1 1 0 4.47214= 0.22361 0.22361 0
N
= = u u .

7

Problem 2.14
(a) ( ) ( )
3
1
cos
Ni i i i
p
t c t
=
= +

q u and since the masses are initially at rest,

1 2 3
0 = = = . When 0 t = ,
3
0
1
Ni i N
p
c
=
= =

q u U C, where C is a column vector of

the coefficients
i
c . Thus
1
0 N

= C U q .
Now
0.17132 0.14369 0.22361
1
0.17132 0.14369 0.22361
kg
0.10161 0.12114 0
N
(
(
=
(
(

U
and
1
1.71324 1.71324 4.06432
1.43703 1.43703 4.84584 kg
2.23603 2.23603 0
N

(
(
=
(
(

U . Given
3
0
1
5 10 1 m
1

=
`

)
q , then
1
0
0.03745
0.00986 kgm
0
N

= =
`

)
C U q . Since ( ) ( )
3
1
cos
Ni i i
p
t c t
=
=

q u then
( ) ( ) ( )
1 2
0.17132 0.14369
0.17132 0.03745cos 0.14369 0.00986cos
0.101611 0.12114
t t t


= +
` `

) )
q . Hence

( ) ( ) ( )
1 2
0.006417 0.001417
0.006417 cos m 0.001417 cos m
0.003806 0.001194
t t t


= +
` `

) )
q ,
or
( ) ( ) ( )
1 2
6.42 1.42
6.42 cos mm 1.42 cos mm
3.81 1.19
t t t


= +
` `

) )
q .

(b) We have ( ) t + = Mq Kq Q . Letting
N
= q U p and pre-multiplying by
T
N
U we
have ( )
T T T
N N N N N
t + = U MU p U KU p U Q or ( ) t + = pp P .
This set of uncoupled equations can then be solved individually for
i
p .
( ) ( ) ( ) ( )
T
0.17132 0.17132 0.10161 0 10.16
0.14369 0.14369 0.12114 0 cos 12.11 cos
0.22361 0.22361 0 100 0
N
t t t t
(

(
= = =
` `
(

(
) )
P U Q
2 2 2 2
1
20.1725 406.9297rad s = = ,
2 2 2 2
2
42.9310 1843.0707rad s = = ,
2 2 2 2
3
54.7723 3000.0048rad s = = so that
8

2 2
406.9297 0 0
0 1843.0707 0 rad s
0 0 3000.0048
(
(
=
(
(

. Hence the uncoupled equations of
motion are ( )
1 1
406.9297 10.16cos p p t + = , ( )
2 2
1843.0707 12.11cos p p t + = and
3 3
3000.0048 0 p p + = . Note 6Hz 6 2 37.699rad s = = = . Solving these three
second order differential equations gives the following steady state solutions:
1
0.0100 p = ,
2
0.0287 p = and
3
0 p = . Thus
( ) ( ) ( )
( )
0.17132 0.14369 0.22361 0.010
0.17132 0.14369 0.22361 0.0287 cos m
0.10161 0.12114 0 0
0.00584
0.00584 cos m
0.00246
(

(
= =
`
(

(
)


=
`

)
N
t t t
t
q U p

1 1 3
30000 0.00246 73.8N
k
F k q = = =
( ) ( )
2 2 1 3
10000 0.00584 0.00246 83N
k
F k q q = = = .
( ) ( )
3 3 2 1
10000 0.00584 0.00584 0
k
F k q q = = + = .
( ) ( )
4 4 3 2
10000 0.00246 0.00584 83N
k
F k q q = = + = .

(c) Response at coordinate 1 due to a harmonic force of 100 N at coordinate 3 is
0.00584m . Therefore the response at coordinate 1 due to a harmonic force of 1 N is
5
5.84 10 m

. Thus the receptance

5
13
5.84 10 mN

= .
Mobility
( )
5 3
13 13
37.6991 5.84 10 2.2016 10 mNs Y j j j

= = = .
Inertance,
( )
2 2 5 2 2
13 13
37.6991 5.84 10 8.3039 10 m Ns

= = = A
The MATLAB script Program_02_14.m repeats these calculations and gives the
following output:

q_1( t ) = {6. 4167*cos( omega1*t ) }+ {- 1. 4167*cos( omega2*t ) } ( mm)
q_2( t ) = {6. 4167*cos( omega1*t ) }+ {- 1. 4167*cos( omega2*t ) } ( mm)
q_3( t ) = {3. 8056*cos( omega1*t ) }+ {1. 1944*cos( omega2*t ) } ( mm)

Response at q1 = - 5. 8428 mm
Response at q2 = - 5. 8428 mm
Response at q3 = 2. 4611 mm

For ce i n spr i ng 1 = 73. 8337 N
For ce i n spr i ng 2 = - 83. 0392 N
For ce i n spr i ng 3 = - 0. 0000 N
For ce i n spr i ng 4 = 83. 0392 N

Recept ance = - 5. 8428e- 005 m/ N
Mobi l i t y = - 2. 2027e- 003 j m/ Ns
I ner t ance = 8. 3039e- 002 m/ Ns^2

9

Problem 2.15
The equations of motion are Mq +Kq = 0 and assuming a harmonic solution,
= Ku Mu where
n
= . The system mass and stiffness matrices are
1
2
3
4
1 0 0 0
0 1 0 0
, kg
0 0 1 0
0 0 0 1
(
(

(
= =
`
(

(

)
q
q
q
q
q M ,
1 2 2
2 2 3 3 3
3 3 4 4
4 4
0 0 20 10 0 0
0 10 14 4 0
N
10
0 0 4 14 10 m
0 0 0 0 10 10
k k k
k k k k
k k k k
k k
+ ( (
( (
+
( (
= =
( ( +
( (
(

K .

(a) Remove
3
q and
4
q : Partition the matrices thus
1 0 0 0
0 1 0 0
kg
0 0 1 0
0 0 0 1
(
(
(
=
(
(

M ,
3
20 10 0 0
10 14 4 0
N
10
0 4 14 10 m
0 0 10 10
(
(

(
=
(
(

K .
With reference to Equation (2.114)

3 1 3 3
ss ss sm
14 10 0.25 0.25 0 4
10 , 10 , 10
10 10 0.25 0.35 0 0

( ( (
= = =
( ( (

K K K
Thus, using Equation (2.115),
1
ss sm
1 0
0 1
0 1
0 1
s

(
(
(
(
= =
(
(

(

(

I
T
K K
.
From Equation (2.113)
T T 3
1 0 20 10
N
kg, 10
0 3 10 10 m
( (
= = = =
( (

r r
M T MT K T KT .
The eigenvalue problem of this reduced system can be solved to give two natural
frequencies.

(b) Stage 1: The ratio of
ii
k to
ii
m is 20000, 14000, 14000 and 10000. The coordinate
with the highest ratio (i.e. coordinate 1) is the first coordinate to be eliminated from
the systems as follows. Reorder the rows and columns so that the new order is
| |
2 3 4 1 . Hence
2
3
3
0 0 0
4
1
1 0 0 0 14 4 0 10
0 1 0 0 4 14 10 0
N
kg 10
0 0 1 0 0 10 10 0
m
0 0 0 1 10 0 0 20
r r r
q
q
q
q
( (
( (


( (
= = =
`
( (

( (

( (
)
M K q .
Thus
| |
3
ss sm
20000, 10 0 0 10 = = K K . Hence,
| |
1
0.5 0 0
ss sm

= = P K K
and
10

1 0 0
0 1 0
0 0 1
0.5 0 0
(
(
(
(
= =
(
(

(
(

I
T
P
. Hence,
T
1
1.25 0 0
0 1 0 kg
0 0 1
(
(
= =
(
(

r
M T MT and
2
T 3
1 1 3
4
9 4 0
N
4 14 10 10 and
m
0 10 10
(

(
= = =
`
(

(
)
r r
q
q
q
K T KT q (from Equation (2.115) ).
Stage 2: The ratio of
ii
k to
ii
m is 7200, 14000 and 10000 for coordinates 2, 3 and 4.
Thus we remove coordinate with the highest ratio, coordinate 3. Reordering the rows
and columns into the sequence
| |
2 4 3 we have
2
3
1 1 1 4
3
1.25 0 0 9 0 4
N
0 1 0 kg 0 10 10 10
m
0 0 1 4 10 14
r r r
q
q
q
( (

( (
= = =
`
( (

( (

)
M K q .
Thus
| |
3
ss sm
14000, 4 10 10 = = K K , and so
| |
1
0.2857 0.7143
ss sm

= = P K K
and
1 0
0 1
0.2857 0.7143
(
( (
= =
( (

(

I
T
P
. Hence, using Equation (2.115) gives
2 3
2 2 2
4
1.3316 0.2041 7.8571 2.8571
N
kg 10
0.2041 1.5102 2.8571 2.8571 m
r r r
q
q
( (
= = =
`
( (

)
M K q .
The eigenvalue problem of this reduced system can be solved to give two natural
frequencies.

(c) Adding the masses at coordinates 1 and 2 together, and also at 3 and 4 together
gives the following mass and stiffness matrices.
2 0
kg
0 2
(
=
(

M ,
1 4 4 3
4 4
14 4
N
10
4 4 m
k k k
k k
+ ( (
= =
( (


K .
The eigenvalue problem of this reduced system can be solved to give two natural
frequencies.

(d) If
1 2
q q = and
3 4
q q = then
1
1 2
3 3
4
1 0
1 0
0 1
0 1
q
q q
q q
q
(
(


(
=
` `
(
)

(

)
.
Applying Equation (2.113) to the original mass and stiffness matrices gives
2 0
kg
0 2
(
=
(

M and
3
14 4
N
10
4 4 m
(
=
(

K . These matrices are identical to those of
part (c).
The MATLAB script Problem_02_15.m solves this problem and gives the following
output:

11

Ful l syst emmodel
1st nat ur al f r equency = 4. 8538 Hz
2nd nat ur al f r equency = 13. 2727 Hz
3r d nat ur al f r equency = 23. 5933 Hz
4t h nat ur al f r equency = 26. 6981 Hz

Ret ai ni ng q1 and q2
1st nat ur al f r equency = 6. 2228 Hz
2nd nat ur al f r equency = 23. 5014 Hz

Reduci ng syst emby el i mi nat i on coor d wi t h hi ghest k/ mt er m
t hus Ret ai ni ng q2 and q4
1st nat ur al f r equency = 4. 9189 Hz
2nd nat ur al f r equency = 13. 8693 Hz

Const r ai ni ng q1 = q2 and q3 = q4
1st nat ur al f r equency = 5. 735 Hz
2nd nat ur al f r equency = 13. 9672 Hz

Note that the reduction in stages, eliminating the coordinate with the highest
ii
k to
ii
m ratio gives the most accurate estimate for the two lowest natural frequencies.

Problem 2.16
If the amplitude of motion is
0
x then the energy dissipated over a quarter of a cycle
by the force
dry
f is
0 dry
f x . Hence over one complete cycle the energy dissipated is
0
4
dry
W f x = . The energy dissipated over one cycle by viscous friction is
2
2
0
t
t
W cx dt
=
=
=

. Let ( )
0
sin x x t = , ( )
0
cos x x t = and so ( )
2 2 2 2
0
cos x x t = .
Hence letting t = ,
2
2 2 2
0 0
0
cos W cx d c x

= =

. The equivalent viscous

damping can be determined by equation energies, thus
2
0 0
4
dry
W c x f x = = . Hence
the equivalent viscous damping is ( )
0
4
eq dry
c f x = . The equation of motion is
0
cos
eq
mx c x kx f t + + = . Thus ( ) ( )
0
2 4 2
eq n dry n
c m f x m = = . Hence
( )
0
2 4
dry
r f x k = where
n
r = . The amplitude of the response of a forced
system is
( )
( )
0
0
2
2
2
1 2
f k
x
r r
=
+
. Thus, substituting for for a system with a dry
friction damper, we have
( )
2 2
2 0
0
2
2
2
0
4
1
dry
f k
x
f
r
x k
=
| |
+
|

\ .
. Rearranging we have
( )
2
2
2
2 2 0
0
2
4
1
dry
f
f
x r
k
k
| |
+ =
|

\ .
. Further rearrangement gives
12

( )
2
2
2
2 2 0
0
2
0
4
1 1
dry
f
f
x r
f
k
(
| |
(
=
|
(
\ .

. Thus
( )
2
0
0
2
1
1
f
x
k r

where
0
4
dry
f
f
=

. To ensure
that
0
x is real,
2
1 0 > and hence
0
4
dry
f f > .
We can also determine the phase of the response because for a single degree of
freedom system with damping
2
2
tan
1
r
r

. Thus
( )
0
2
4
tan
1
dry
f x k
r

, but
( )
2
0
0
2
1
1
f
x k
r

and hence
2 2
0
4
tan
1 1
dry
f
f

= =

. Thus we see that the phase
angle is not a function of frequency for this system. For the above analysis to be valid,
1 < , and this implies that
0
4
dry
f f < . In the example, 1kg m = ,
3
100 10 N m k = ,
25Hz = , 10N
dry
f = and
0
50N f = . Thus 316.23rad/s = =
n
k m ,
25 2 157.08rad/s = = and hence 0.4967 = =
n
r .
0
4
4 10
0.2546
50
dry
f
f

= = =

.
Thus
( ) ( )
2 2
4 0
0
2 5 2
1 50 1 0.2546
6.42 10 m 0.642mm
1 10 1 0.4967


= = =

f
x
k r
.
1
2
180
tan 14.75
1

| |
| |
= = |
|
|

\ .

\ .
. The MATLAB script Problem_02_16.m repeats
these calculations and gives the following output and graphical output.

Nat ur al f r equency = 50. 3292 Hz
del t a = 0. 2546
Exci t at i on f r equency = 25
Response = 0. 6419 mmat 14. 7527 degr ee

Exci t at i on f r equency = 55
Response = 2. 4895 mmat - 14. 7527 degr ee

0 0.5 1 1.5 2
10
0
Frequency ratio r
A
m
p
l
i
t
u
d
e
Coulomb damping. F(dry) =10 N
0 0.5 1 1.5 2
0
50
100
150
Frequency ratio r
P
h
a
s
e

(
d
e
g
r
e
e
s
)

13

Problem 2.17
Equation of motion is
3
0
cos mx kx hx f t + + =
(a) If we let
1
cos x x t = ,
2
1
cos x x t = and
3 3 3
1
cos x x t = . Expanding this
expression in terms of multiple angles gives ( )
3 3
1
cos3 3cos 4 x x t t = + .
Considering only the cos t terms, we have
3 3
3
1
4
cos x x t = . Substituting in the
equation of motion and cancelling the cos t factor we obtain
2 3
3
1 1 1 0
4
mx kx hx f + + = . Thus,
( )
3 2
3
1 1 0
4
0 hx k m x f + = .
Given 50 2 314.1593/s = =
6
10 N m k = and
6 3
40 10 N m h = we can solve
this cubic equation for various values of
0
f .(see MATLAB function
Problem_02_17.m) to obtain
0 1
0 1
0 1
0 1
30N 2.27mm
60N 4.40mm
120N 8.02mm
240N 13.2mm
f x
f x
f x
f x
= =
= =
= =
= =

If 55Hz = ,
0
240N f = ,
1
81.1and 1.24mm x =
When
0
0 f = ,
( )
3 2
3
1 1
4
0 hx k m x + = . Thus
( )
2 2
3
1
4
0 hx k m + = and, dividing
through by m we have
( )
2 2 2
3
1 0
4
0 x + = where h m = and
2
0
k m = .
Therefore
2 2 2
3
0 1
4
x = + . Note that if 0 > ,
2
3
1
4
0 x > and hence
0
> .
When
6 6 2 2
40 10 10 4 10 / ms = = ,
2 6 5 2 2
0
10 10 10 rad /s = = 0.010m x = .
Thus
( )
2 5 6 2 2 2
3
4
10 4 10 0.01 100300rad /s = + = and so
( ) 100300 2 50.4046Hz = = .

(b) Letting
1 3
cos cos3 x x t x t = + ,
2 2
1 3
cos 9 cos3 x x t x t =
and
3 3 3 2 2 2 2 3 3
1 1 2 1 2 3
cos 3 cos cos3 3 cos cos 3 cos 3 x x t x x t t x x t t x t = + + + .
Expanding the powers of the trigonometric functions in terms of multiple angles, and
neglecting terms in cosp t , where 5, 7and9 p = we have
( ) ( )
3 3 2 2 3 2 3
1 1
1 1 3 1 3 1 1 3 3
4 4
3 3 6 cos 6 3 cos3 x x x x x x t x x x x t = + + + + +
Substituting in the equation of motion we have
( ) ( ) { }
( ) ( ) { }
2 3 2 2
1
1 1 1 3 1 3
4
2 3 2 3
1
3 1 1 3 3 0
4
3 3 6 cos
9 6 3 cos3 cos
k m x h x x x x x t
k m x h x x x x t f t
+ + + +
+ + + =

Thus, equating coefficients of cos t and cos3 t gives a pair of equations thus:
( ) ( ) { }
( ) ( ) { }
2 3 2 2
1
1 1 1 1 3 1 3 0
4
2 3 2 3
1
2 3 1 1 3 3
4
3 3 6 0
9 6 3 0
f k m x h x x x x x f
f k m x h x x x x
= + + + =
= + + + =

14

Let
2 2
1 2
f f = + . Clearly when the values of
1
x and
3
x are the roots,
1 2
0 f f = =
and 0 = . The MATLAB function f mi nsear ch. miterates from initial values to
minimise and hence solve the equations. This function is used in the MATLAB
function Problem_02_17.m. From this function we have the following output

Si ngl e t er msol ut i on. For ci ng f r equency = 50Hz
f or ce = 30 N, x1 = 2. 2736 mm
f or ce = 60 N, x1 = 4. 4048 mm
f or ce = 120 N, x1 = 8. 0172 mm
f or ce = 240 N, x1 = 13. 1609 mm

Si ngl e t er msol ut i on: For ci ng f r equency = 55Hz
f or ce = 240N. x1 = 81. 0725 mmand - 1. 2360 mm

Magni t ude of vi bn, 10 mm. Fr equency of unf or ced r esponse = 50. 4046 Hz

Two t er msol ut i on, f or ce = 240N
omega = 50 Hz x1 = 13. 1604 mm, x3 = 2. 8954e- 003 mm
omega = 55 Hz ( 1st sol ut i on) x1 = 80. 7894 mm, x3 = 0. 5636 mm
omega = 55 Hz ( 2nd sol ut i on) x1 = - 1. 2360 mm, x3 = - 1. 9390e- 006 mm

Note that to obtain the second solution, when 55Hz = , the initial values were zero.
15

Chapter 3

Equations of motion for a r igid body, pinned at one end.
The equations of motion for the system shown in Fig 3.26 can be developed by using
Newtons 2
nd
law. A free body diagram for the system is shown below.



Moments acting in the direction:
y d p
f L I I = +

where
d
I is the moment of
inertia of the rotor about the left end.
Moments acting in the direction:
x d p
f L I I =

. In general,
x
f is the
force to extend a spring or give velocity to a viscous damper. Similarly for
y
f . Now
v L = and u L = . Hence
( ) ( )
y y y y y
f k v c v k L c L = + = +

and
( ) ( )
x x x x x
f k u c u k L c L = + = + . Hence the equations of motion are
2 2
2 2
0
0
d p y y
d p x x
I I c L k L
I I c L k L
+ + + =
+ + =

Pr oblem 3.1
In this problem there is no damping in the bearing and the stiffness is the same in the
x and y directions. Hence
2
2
0
0
d p
d p
I I kL
I I kL
+ + =
+ =

or in matrix notation,
2
2
0
0 0 0
0 0 0
0
p
d
d p
I
I kL
I I
kL
( (
(
+ + = ( ( ` ` ` `
(

) ) ( ( ) )



.
Seeking solutions of the form
0
j t
e

= and
0
j t
e

= , results in the equations
2 2
0
2 2
0
0
0
(
+

(
=
` `
(
) ) +

d p
p d
I L k j I
j I I L k
. This matrix must be singular and hence its
determinant is zero. Thus is the solution of
( ) ( )
2
2
2 2
0 + + =
d p
I L k I or
16

( ) ( )
2
2
2 2
+ =
d p
I L k I . Hence
( ) ( )
2 2
+ =
d p
I L k I and so
2 2
0 =
d p
I I L k . For example, solving this pair of quadratics for
10,000rev min = we have
2 2 6
10000 2
10 0.6 0.5 10 0
60
| |
+ =
|
\ .
. And
2 2 6
10000 2
10 0.6 0.5 10 0
60
| |
=
|
\ .
. Solving these equations give
129.7888and 192.6206 = and 129.7888and192.6206 = . Thus, taking the
positive roots, 129.7888rad sand192.6206rad s = or, converting to Hz by
dividing by 2 gives 20.6565Hzand30.6565Hz = .
Alternatively, we can solve the eigenvalue problem as described by equations (3.48),
(3.50), (3.51) and (3.52). The MATLAB script Pr oblem_03_01.m gives the user the
choice of solving either the characteristic equation or the eigenvalue problem to
determine the system natural frequencies. Of course, both methods give the same
numeric values for the frequencies which are as follows

Sol ut i on of t he char act er i st i c equat i on
Rot or speed = 0 r ev/ mi n
Nat ur al f r equency = 25. 1646 Hz
Nat ur al f r equency = 25. 1646 Hz

Rot or speed = 3000 r ev/ mi n
Nat ur al f r equency = 23. 7093 Hz
Nat ur al f r equency = 26. 7093 Hz

Rot or speed = 10000 r ev/ mi n
Nat ur al f r equency = 20. 6565 Hz
Nat ur al f r equency = 30. 6565 Hz

Once is calculated the relative displacements of the rotor are determined from the
matrix equation above as
2 2
0
2 2
0


= =

p
d
p
d
j I
L k I
j I
L k I
. Note that for 0 = the modes are not unique.
At 3000rev min, 148.9699rad sand167.8195rad s = .
0 0
and = j j
respectively. Since v L = and u L = , then
0 0
and = v u j j .
For the first mode,
( )
( )
1 1 cos
2
sin
j j
u t
e e
v t j j


= + =
` ` ` `

) ) )
)
where t = .
For the second mode,
( )
( )
1 1 cos
2
sin
j j
u t
e e
v t j j


= + =
` ` ` `

) ) )
)
.
The above is identical to Equations (3.34) and (3.35) and the text with these equations
explains why the first mode (23.71 Hz) is a backward rotating mode and the second
mode (26.71 Hz) is a forward rotating mode. The orbits are circular.
17

Pr oblem 3.2.
In this problem there is no damping in the bearing. Hence
2
2
0
0
d p y
d p x
I I k L
I I k L
+ + =
+ =

or in matrix notation,
2
2
0 0
0 0
0 0 0
0
p y
d
d p
x
I k L
I
I I
k L
(
(
(
( + + =
( ` ` ` `
(
(
) ) ( ) )




.
Seeking solutions of the form
0
j t
e

= and
0
j t
e

= , results in the equations
2 2
0
2 2
0
0
0
(
+

(
=
` `
(
) ) +

d y p
p d x
I L k j I
j I I L k
. This matrix must be singular and hence
its determinant is zero. Thus is the solution of
( )( ) ( )
2
2 2 2 2
0 + + =
d y d x p
I L k I L k I . Letting
2
= this leads to a
quadratic in thus:
( ) ( )
2 2 2 2 2 2
0 + + + =
d p d x y x y
I I I L k k L k k which is easily
solved for given parameters. For example, when
3000rev min 3000 2 60 314.159rad s = = = we have
2 6 9
100 6.1448 10 81.25 10 0 + = . Solving this equation gives
23997and 33858 = and hence 138.77 and 205.40 s j j = = Taking the
positive roots we have
( ) ( ) 138.77 2 and205.40 2 24.6548Hzand29.2854Hz = =
Alternatively, we can solve the eigenvalue problem as described by equations (3.48),
(3.50), (3.51) and (3.52). The MATLAB script Pr oblem_03_02.m gives the user the
choice of solving either the characteristic equation or the eigenvalue problem to
determine the system natural frequencies. Of course, both methods give the same
numeric values for the frequencies which are as follows
Solution of the characteristic equation

Rot or speed = 0 r ev/ mi n
Nat ur al f r equency = 25. 1646 Hz
Nat ur al f r equency = 28. 6921 Hz

Rot or speed = 3000 r ev/ mi n
Nat ur al f r equency = 24. 6548 Hz
Nat ur al f r equency = 29. 2854 Hz

Rot or speed = 10000 r ev/ mi n
Nat ur al f r equency = 22. 0866 Hz
Nat ur al f r equency = 32. 6906 Hz

Pr oblem 3.3.
In this problem there is damping and the stiffness in the bearing and the properties are
the same in the x and y directions. Hence
2 2
2 2
0
0
d p
d p
I I cL kL
I I cL kL
+ + + =
+ + =

or in matrix notation,
18

2
2
2 2
0 0 0
0 0
0
p
d
d
p
cL I
I kL
I
I cL kL
(
(
(
(
+ + = (
` ` ` `
(
(
) ) ( ) )





Since damping is included we now look for solutions of the form
0
=
st
e and
0
=
st
e , where s is complex, results in the equations
2 2 2
0
2 2 2
0
0
0
(
+ +

(
=
` `
(
) ) + +

d p
p d
s I sL c L k s I
s I s I sL c L k

Setting the determinant of the above array to zero we have
( ) ( )
2
2
2 2 2
0 + + + =
d p
s I sL c L k s I . Hence
2 2 2
+ + =
d p
s I sL c L k js I . This gives the following quadratic equation (with one
complex coefficient)
( )
2 2 2
0 + + =
d p
I s L c j I s L k . For example, when
3000rev min 3000 2 60 314.159rad s = = = we have
( )
2 4
10 125.00 188.50 25 10 0 + + = s j s
Each equation has two roots so the two equations have together four roots (forming
complex conjugate pairs) thus 6.62 167.70 and 5.88 148.85 s j j = . From these
roots we can obtain the damped natural frequency by taking the imaginary part of s,
the natural frequency by taking the absolute values of s, and the damping, , by
changing the sign of the real value divided by the absolute value of s. For example,
when 6.62 167.70 s j = , ( ) 167.7 2 26.69Hz
d
= = , 6.62167.7 0.395 = =
and
( )
2 2
6.62 167.7 2 26.71Hz
n
= + = .
Alternatively, we can solve the eigenvalue problem as described by equations (3.48),
(3.50), (3.51) and (3.52). The MATLAB script Pr oblem_03_03.m gives the user the
choice of solving either the characteristic equation or the eigenvalue problem to
determine the system damped and natural frequencies. Of course, both methods give
the same numeric values for the frequencies which are as follows

Sol ut i on of t he char act er i st i c equat i on
Rot or speed = 0 r ev/ mi n
Damped nat ur al f r eq ( Hz) Nat ur al f r eq ( Hz) Zet a
25. 1449 25. 1646 0. 0395
25. 1449 25. 1646 0. 0395

Rot or speed = 3000 r ev/ mi n
Damped nat ur al f r eq ( Hz) Nat ur al f r eq ( Hz) Zet a
23. 6897 23. 7082 0. 0395
26. 6897 26. 7105 0. 0395

Rot or speed = 10000 r ev/ mi n
Damped nat ur al f r eq ( Hz) Nat ur al f r eq ( Hz) Zet a
20. 6380 20. 6535 0. 0388
30. 6380 30. 6610 0. 0388
19

Pr oblem 3.4.
In this case the force applied to the support in the x and y directions is
= + =
x c c
f ku k v kL k L and = + =
y c c
f k u kv k L kL . The moment acting on
the rotor in the direction is
2 2
=
y c
Lf k L kL . Similarly the moment acting on
the rotor in the direction is
2 2
= +
x c
Lf kL k L and the equations of motion
become
2 2
2 2
0
0
+ + =
+ =

d p c
d p c
I I L k L k
I I L k L k
or, in matrix notation,
2 2
2 2
0
0 0
0 0 0
(
(
(
+ + = (
( ` ` ` `
(

( ) ) ( ) )




p
d c
d p
c
I
I kL k L
I I
k L kL

Seeking solutions of the form
0
st
e = and
0
st
e = , gives the following equation
for s
2 2 2
2 2 2
det 0
(
+
(
=
(
+

d p c
p c d
s I L k s I L k
s I L k s I L k
or
( ) ( )( )
2
2 2 2 2
0 + =
d p c p c
s I L k s I L k s I L k and hence
( ) ( )
2 4 2 2 2 2 4 2 2
2 0 + + + =
d p d c
I s I I L k s L k k , a quadratic in
2
s .
For example, when 3,000rev min = and letting
2
s = , we have
2 6 10
100 5.0355 10 6 10 0 + + = . Solving this quadratic gives
4 4
1.935 10 and 3.100 10 = . Hence 139.12 and176.07 s j j = and so
( ) ( ) 139.12 2 22.14Hzand 176.07 2 28.02Hz
n
= = = .
Alternatively, we can solve the eigenvalue problem as described by equations (3.48),
(3.50), (3.51) and (3.52). The MATLAB script Pr oblem_03_04.m gives the user the
choice of solving either the characteristic equation or the eigenvalue problem to
determine the system natural frequencies. Of course, both methods give the same
numeric values for the frequencies which are as follows

Sol ut i on of t he char act er i st i c equat i on
Rot or speed = 0 r ev/ mi n
Nat ur al f r equency = 22. 5079 Hz
Nat ur al f r equency = 27. 5664 Hz

Rot or speed = 3000 r ev/ mi n
Nat ur al f r equency = 22. 1415 Hz
Nat ur al f r equency = 28. 0227 Hz

Pr oblem 3.5.
This machine has the equations of motion given by Equation (3.9), with 2
T
k k = ,
0
C
k = and
2
2 =
R
k a k , where L is the distance between the bearings (hence
2 a b L = = ). Thus the equations of motion are
20

2
2
2 0
2 0
2 0
2 0
+ =
+ =
+ + =
+ =

d p
d p
mu ku
mv kv
I I a k
I I a k
or in matrix notation,
2
2
2 0 0 0 0 0 0 0
0 0 0 0
0 2 0 0 0 0 0 0
0 0 0 0
0 0 0
0 0 0 0 0 2 0 0
0 0 0 0 0 0 0
0 0 0 2
( (
(
( (
(

( (
(
+ + =
` ` ` `
( (
(

( (
(


( ( ( ) ) ) )





p
d
d p
k
m u u u
k
m v v v
I
I a k
I I
a k
The translational and rotational equations of motion decouple. Thus, from
translational equations of motion, natural frequencies are 2k m (twice).
We now consider the 3
rd
and 4
th
equations and seek solutions of the form
0
st
e =
and
0
st
e = . Thus
2 2
2 2
2
det 0
2
d p
p d
I s a k I s
I s I s a k
(
+
(
=
(
+

gives the following equation for s

( ) ( )
2
2
2 2
2 0 + + =
d p
I s a k I s or
2 2
2 + =
d p
I s a k jI s giving the pair of
quadratics
2 2
2 0 + =
d p
I s jI s a k . At 3,000rev min = ,
2 6
50 9424.8 2 10 0 + = s j s and solving this equation gives
315.34 and 126.85 s j j = . Taking the positive roots,
( ) ( ) 315.34 2 50.19Hzand 126.85 2 20.19Hz = = = .
Alternatively, we can solve the eigenvalue problem as described by equations (3.48),
(3.50), (3.51) and (3.52). The MATLAB script Pr oblem_03_05.m gives the user the
choice of solving either the characteristic equation or the eigenvalue problem to
determine the system natural frequencies. Of course, both methods give the same
numeric values for the frequencies which are as follows

Sol ut i on of t he char act er i st i c equat i on,
Nat ur al f r equency = 14. 2353 Hz ( t wi ce)
Nat ur al f r equency = 20. 1882 Hz
Nat ur al f r equency = 50. 1882 Hz

Pr oblem 3.6.
This system is described by Equation (3.84). In matrix notation, we have
0 0 0 0
0 0 0 0 0 0
0 0 0 0
0 0 0 0 0 0
0 0 0
0 0 0 0 0 0
0 0 0 0 0 0 0 0 0
(
( (
(
( (

(
( (
+ + =
` ` ` `
(
( (

(
( (


( (
( ) ) ) )





T C
T C
p
d C R
d C R p
m k k u u u
m k k v v v
I
I k k
I k k I

Using the stiffness formulae gives in Appendix 2, Table A2.1, System 5, we have
21

( )
( )
3
12 3
4 3
T uu vv
EI a b
k k k
b a b
+
= = =
+
,
( )
( )
12
4 3
R
EI a b
k k k
b a b

+
= = =
+
and
( )
( )
2
6 2 3
4 3
C u v
EI a b
k k k
b a b

+
= = =
+
. Letting
0
=
st
u u e etc. we can solve these
equations either by forming the characteristic equation or by solving the eigenvalue
problem. The characteristic equation for this system is
4 3 2
0 A B C D E + + + + = where
2
s = and
2 2
d
A m I = ,
( )
2 2 2
2 2
p d R d T
B mI mI k I k m = + + ,
4 2 2 2
2
C T R T R C
E k k k k k k = +
2 2 2 2 2 2 2
4 2 2
R d T R p T d C d T
C m k mI k k mI k mI k I k = + + + and
2 2 2 2 2 2 2
2 2 2 2
T R C R p T d T C d T R
D mk k mk k I k I k k I k k = + + . Clearly, these expressions are
complicated and deriving them requires a fair amount of tedious algebra. Finally a
quartic equation must be solved and that requires a numerical procedure. In this
problem it is probably easier to solve the eigenvalue problem directly as described by
equations (3.48), (3.50), (3.51) and (3.52). To determine the values of
0 0
u we
can either (i) extract the information from the eigenvalues, or (ii) substitute the values
of s into one of the system equations. Thus, from the first system equation, since
2
0 0 0
0
T C
ms u k u k + + = , then
0
2
0
| |
=
|
|

+
\ .
C
T
u k
ms k

The MATLAB script Pr oblem_03_06.m gives the user the choice of solving either
the characteristic equation or the eigenvalue problem to determine the system natural
frequencies and
0 0
u . Of course, both methods give the same numeric values
which are as follows

Sol ut i on of t he EVP
Rot or speed = 0 r ev/ mi n
Nat ur al f r equency = 15. 239 Hz. u/ psi = 0. 30996
Nat ur al f r equency = 15. 239 Hz. u/ psi = 0. 30996
Nat ur al f r equency = 67. 1535 Hz. u/ psi = - 0. 21508
Nat ur al f r equency = 67. 1535 Hz. u/ psi = - 0. 21508

Rot or speed = 1000 r ev/ mi n
Nat ur al f r equency = 9. 9593 Hz. u/ psi = 0. 28808
Nat ur al f r equency = 22. 1203 Hz. u/ psi = 0. 36325
Nat ur al f r equency = 60. 0598 Hz. u/ psi = - 0. 3347
Nat ur al f r equency = 79. 1488 Hz. u/ psi = - 0. 12689

Using the values of
0 0
u we can make sketches of the mode shapes as shown
below. The left column diagrams are for the stationary rotor, the right column
diagrams are for the spinning rotor.
22

0 0.5 1
-0.5
0
0.5
0 0.5 1
-0.5
0
0.5
0 0.5 1
-0.5
0
0.5
0 0.5 1
-0.5
0
0.5


Pr oblem 3.7.
The equations of motion are the same as Equation (3.4), except that here a force
1 x
f
is required to enforce the constraint
1
0 u u a = = . The equations of motion are
then (noting that 2 a b L = = )
( )
( ) ( )
( ) ( )
( )
0 1
1 2
1 2
0 1
0
0
x
d p
d p x
mu k u a f
mv k v a k v a
I I ak v a ak v a
I I ak u a af
+ + =
+ + + =
+ + + =
+ + =

Adding a times the first equation to the last equation, gives
( ) ( )
( ) ( )
( )
1 2
1 2
0
0
0
2 0
d p
d p
mv k v a k v a
I I ak v a ak v a
mau I I ak u a
+ + + =
+ + + =
+ + + =

Now applying the constraint u a = to eliminate u gives
( ) ( )
( ) ( )
( )
1 2 1 2
2
1 2 1 2
2 2
0
0
0
4 0
d p
d p
mv k k v a k k
I I a k k v a k k
I ma I a k
+ + + =
+ + + + =
+ + =

There is an alternative way of deriving these equations. In the y-z plane the system
identical to the system shown in Fig 3.6 and described by Equation (3.6). In the x-z
pane the system is identical to the system described in Problem 3.1.
Thus in the y-z plane (from Equations (3.6) and noting that 2 a b L = = ) we have
( ) ( )
( ) ( )
1 2 1 2
2
1 2 1 2
0
0
d p
mv k k v a k k
I I a k k v a k k
+ + + =
+ + + + =

In the x-z plane (from Problem 3.1) we have
2
1 0
0
d p
I I k L + =

where
1 d
I is the moment of inertia bearing 1 and
2
1 d d
I I ma = + so that
23

( ) ( )
( ) ( )
( )
1 2 1 2
2
1 2 1 2
2 2
0
0
0
4 0
d p
d p
mv k k v a k k
I I a k k v a k k
I ma I a k
+ + + =
+ + + + =
+ + =

, as before.
Letting
0
=
st
e , etc, we obtain
( )
( ) ( )
( )
2
1 2 1 2
2 2
1 2 1 2
2 2 2
0
0
det 0
0 4
d p
p d
ms k k a k k
a k k I s a k k I s
I s I ma s a k
(
+ +
(
(
+ + =
(
(
+ +
(


From this determinant we can obtain the characteristic equation thus

Part 1 When 0 = . Here the 3
rd
equation is uncoupled from the other two and so we
obtain
( )
2 2 2
0
4 0
d
I ma s k a + + = and
( )
( )
4 2 2 2
1 2 1 2
4 0
d d
mI s I ma k k s a k k + + + + = .
The linear equation in
2
s and the quadratic equation in
2
s can easily be solved.

Part 2. When
1 2
k k k = = . Here the 1
st
equation uncouples from the other two and so
we obtain
( ) ( ) { }
2 4 2 2 2 2 2 2 4
0 0
2 4 8 0
d d d d p
I I ma s ka I ma k a I I s a k k + + + + + + =
and
2
2 0 ms k + = . The linear equation in
2
s and the quadratic equation in
2
s can
easily be solved.

Alternatively, in matrix notation we have
( )
( ) ( )
2
1 2 1 2
2
1 2 1 2
2
0
0 0 0 0 0
0 0 0 0
0 0 0 0
0
0
0 0
0
0 0 4
d p
p d
m v v
I I
I I ma
k k a k k
v
a k k a k k
a k
(
(

(
(

+
( ` ` (
( (

+ ) )
(

| ( +

|
(

+ + = |
( ` `
|
(
|
) )
(
.




These equations can be rearranged into the form = Ax + Bx 0 , and the eigenvalue
problem can be solved as described by equations (3.48), (3.50), (3.51) and (3.52). The
MATLAB script Pr oblem_03_07.m gives the user the choice of solving either the
characteristic equation or the eigenvalue problem to determine the system natural
frequencies. Of course, both methods give the same numeric values for the
frequencies which are as follows

Sol ut i on of t he char act er i st i c equat i on
Rot or speed = 0 r ev/ mi n
St i f f ness k1 = 1800000 N/ m
St i f f ness k2 = 2200000 N/ m
Nat ur al f r equency = 28. 8598 Hz
Nat ur al f r equency = 36. 5126 Hz
Nat ur al f r equency = 56. 3715 Hz

24

Rot or speed = 9550 r ev/ mi n
St i f f ness k1 = 2000000 N/ m
St i f f ness k2 = 2000000 N/ m
Nat ur al f r equency = 29. 0576 Hz
Nat ur al f r equency = 33. 8302 Hz
Nat ur al f r equency = 60. 7316 Hz

Pr oblem 3.8.
Using the stiffness formulae gives in Appendix 2, Table A2.4, System 6 with
2 a b L = = and
1 2
k k = , we have
( )
2
3
6
3 D EI a k = + , 0
C u v
k k k

= = = ,
( )
3 2
6
6
3
T uu vv
EI
k k k EIk a k
D
= = = + and
( )
2 5 2
6
6EI
k k EIa k a k
D

= = + . The
equations of motion for the system are
0
0
0
0
+ =
+ =
+ =
+ =

T
T
d R
d R
mu k u
mv k v
I k
I k

Note that because 0
C
k = and gyroscopic effects are ignored the four equations are
uncoupled from each other and can be solved independently of each other. Thus,
letting
0

=
j t
u u e etc. we have we have
n T
k m = (twice) and
n R d
k I = .
Simply supported beam with central disk: From Appendix 2, Table A2.1, System 1
with 2 a b L = = we have
3
6
ss
T
k EI a = , 6
ss
R
k EI a = and 0
ss
C
k = . Thus we can
compute the natural frequencies as above, i.e.
ss
n T
k m = (twice) and
ss
n R d
k I = .
Rigid rotor: The stiffness coefficients for this system are given (3.7) so that
2
rgd
T
k k = ,
2
2
rgd
R
k a k = and 0
rgd
C
k = . Again, the equations uncouple so that the
frequency equations are as above. Although these calculations are simple we have
written a MATLAB script to carry them out. Thus The MATLAB script
Pr oblem_03_08.m determines the system natural frequencies for the three models, as
follows.

Bear i ng st i f f ness = 50000N/ m
Ful l model ( Hz) Si mpl y suppor t ed beam( Hz) Ri gi d r ot or ( Hz)
1st Mode: 13. 2231 72. 1384 13. 4510
2nd mode: 34. 9851 190. 8604 35. 5881

Bear i ng st i f f ness = 1000000N/ m
Ful l model ( Hz) Si mpl y suppor t ed beam( Hz) Ri gi d r ot or ( Hz)
1st Mode: 46. 1998 72. 1384 60. 1549
2nd mode: 122. 2332 190. 8604 159. 1549

Bear i ng st i f f ness = 100000000N/ m
Ful l model ( Hz) Si mpl y suppor t ed beam( Hz) Ri gi d r ot or ( Hz)
1st Mode: 71. 6252 72. 1384 601. 5491
2nd mode: 189. 5026 190. 8604 1591. 5494

These results show that when the bearing stiffness is low the system behaves like a
25

rigid rotor on flexible supports. When the bearing stiffness is high, the system behaves
like a simply supported flexible beam. At the intermediate stiffness (1MN/m) only the
full model adequately describes the system.

Pr oblem 3.9.
The data of Table 3.4 gives the eigenvalues and eigenvectors for a rotor supported by
(a) isotropic bearings and (b) anisotropic bearings. The natural frequencies of the
system are determined from the eigenvalues, i.e. ( ) ( ) imag 2
n
s = . The method to
determine the shape of an orbit and direction of whirl is described in Section 3.6.1.
Equation (3.55) requires amplitude and phase of the displacements in the x and y
directions. From this the shape of the orbit can be determined (, given in Equation
(3.60)) and the direction of whirl given by the sign of .
The displacements in the x and y directions at the two ends of the rotor are given by
Equation (3.61). The MATLAB script Pr oblem_03_09.m analyses the isotropic and
the anisotropic bearing case. The script begins by regenerates the data of Table 3.4 and
then determines T (Equation (3.56)),
T
= H TT , and then the eigenvalues of H.
From these eigenvalues the script determines and the direction if rotation. The
MATLAB script Pr oblem_03_09.m determines the system natural frequencies and
values of , thereby giving the shape and direction of the orbit at various locations
along the rotor.

[ - ve kappa: backwar d whi r l , +ve kappa: f or war d whi r l ]

I sot r opi c bear i ngs
omega_n ( Hz) 47. 4044 47. 4233 88. 8962 154. 9324
x- y or bi t at r t r cent r e kappa: - 1. 0000 1. 0000 - 1. 0000 1. 0000
Rot at i on at r t r cent r e kappa: - 1. 0000 1. 0000 - 1. 0000 1. 0000
x- y or bi t at bear i ng 1 kappa: - 1. 0000 1. 0000 - 1. 0000 1. 0000
x- y or bi t at bear i ng 2 kappa: - 1. 0000 1. 0000 - 1. 0000 1. 0000

I sot r opi c bear i ngs
omega_n ( Hz) 47. 4140 50. 3203 91. 6220 157. 8805
x- y or bi t at r t r cent r e kappa: - 0. 0067 0. 0076 - 0. 4441 0. 3827
Rot at i on at r t r cent r e kappa: - 0. 2558 - 0. 2926 - 0. 8889 0. 9348
x- y or bi t at bear i ng 1 kappa: 0. 0012 0. 0297 - 0. 9120 0. 9286
x- y or bi t at bear i ng 2 kappa: - 0. 0212 - 0. 0286 - 0. 8770 0. 9380

As expected, for the isotropic bearings, the orbits are circular. For the anisotropic
bearings, the orbits are elliptic and some of the modes are forward, some are backward
and some are a mixture. For example the mode at frequency 47.414 Hz has a forward
whirl at bearing 1, but backward at the centre of the rotor and at bearing 2.

Pr oblem 3.10.
The equations of motion of this system are given by Equation (3.9) (or Equation
(3.26)) with 2 =
T
k k , 0
C
k = and
2
0.5
R
k L k = (since 2 = = a b L ). Thus
2
2
2 0
2 0
0.5 0
0.5 0
d p
d p
mu ku
mv kv
I I L k
I I L k
+ =
+ =
+ + =
+ =

.
We know that 30kg m = . The first and second equations are uncoupled from the
26

third and fourth equations and are independent of speed. Thus one natural frequency
is independent of speed and from the 1
st
or 2
nd
equation is given by
1
2k m = .
Hence ( )
2
2
1
0.5 0.5 30 2 7.1 29.8516kN m k m = = =
The other natural frequency, when 0 = , is obtained from the 3
rd
or 4
th
equation is
given by
2
2
0.5
d
L k I = Thus
( )
2 2 3
2
2 2
2
0.5 0.5 0.8 29.8516 10
0.53838kgm
2 21.2
d
L k
I

= = =

When the rotor spins the second pair of natural frequencies are given by the solutions
to equation (3.29), that is
( ) ( )
2
2
2
0
d R p
I k I + = . Thus
2
d R
p
I k
I
+
=

where the natural frequency is 2 46.7rad s = and the rotor spin speed is
2000 2 60 209.44rad s = = .
2 2
0.5 0.5 0.8 29851.6 9552.5Nm = = =
R
k L k
and hence
2
0.59883kgm
p
I = .

Pr oblem 3.11.
The development in Section 3.6.1 is reworked with the eigenvalue
i i
s j = and the
corresponding mode. Consider first the calculation of the lengths of the semi-major
and semi-minor axes - the mode considered here will be the complex conjugate of the
mode used in Section 3.6.1, and hence the matrix T will be post multiplied by
1 0
0 1
(
(

. Thus H will be as given in Section 3.6.1 and thus so will be the lengths of
the semi-major and semi-minor axes. The only question that remains is to determine
the direction of whirl. If
u
and
v
are the angle corresponding to the new mode
with eigenvalue
i i
s j = then the equivalent equations to Equation (3.59) is
( )
( )
( )
( )
( )
( ) ( )
cos cos
cos cos
u i u i
u v u i u v u i
r t u t r t
r t v t r t


= =
` ` `
+ +

) ) )

and hence the same conclusions can be draw about the direction of whirl by making
( )
v u
negative. Thus in this case ( ) 0
v u
< < implies forward whirl and
( ) 0
v u
< < implies backward whirl.

Pr oblem 3.12
Since the rotor is stationary and the bearings are also isotropic we only need to
consider motion in, say, the x direction. Using the stiffness formulae gives in
Appendix 2, Table A2.4, System 6, and letting
1
k gives
( )
3 3
6 1 2
3 D a k EI b k = + ,
( )
3 3
1 1 2
6
3
3
uu
EI
k EIk a b k k
D
(
= + +

,
( )
( )
2 2
1 1 2
6
3
3
u
EI
k EI ak ab a b k k
D

(
= +

,
27

( )
2 2 2
1 1 2
6
3
3
EI
k EIa k a b a b k k
D

(
= + +

.
Since
1
k is a factor in both the numerators and the denominator of the above
equations, these simplify to
( )
3 3
6 2
3 D a EI b k = + ,
( )
3 3
2
6
3
3
T uu
EI
k k EI a b k
D
(
= = + +


( )
2 2
2
6
3
3
C u
EI
k k EIa ab a b k
D

(
= = +

,
( )
2 2 2
2
6
3
3
R
EI
k k EIa a b a b k
D

(
= = + +


Thus the equations of motion are
0 0
0 0
T C
d C R
m k k u u
I k k
( (
+ =
` ` `
( (

) ) )

.
From this equation we can obtain + = Mq Kq 0 and hence the eigenvalue
2
0 0
s + = Mq Kq 0 . This can be solved to obtain the two system natural frequencies.
Using the reduction formulae of Appendix 2,
2
red T C R
k k k k = and
2 2
red D C R
m m I k k = + . Hence the frequency of the reduced model is
_ n red red red
k m = . The MATLAB script Pr oblem_03_12.m solves the eigenvalue
problem to determine the system natural frequencies and the natural frequency of the
reduced model, which are as follows

Di sk di amet er / t hi ckness 0. 650/ 0. 065 1. 200/ 0. 120
1st nat ur al f r equency ( Hz) 17. 4505 6. 4971
2nd nat ur al f r equency ( Hz) 68. 5833 15. 8084
Reduced model , nat f r eq ( Hz) 17. 4785 6. 6074

Pr oblem 3.13
(a) Using the formulae gives in Appendix 1:
For the cylinder,
2 2 2
8 100 0.4 8 2kgm
p
I MD = = = ,
2 2 2
2 12 2 2 100 0.6 12 4kgm
d p
I I Mh = + = + = .
For the disk,
2 2 2
8 100 1.4 8 24.5kgm
p
I MD = = =
2 2 2
2 12 24.5 2 100 0.1 12 12.3333kgm
d p
I I Mh = + = + =
The center of gravity of the system, relative to bearing number 1 is given by
Mz z M = and hence 100 0.8 100 1.6 200z + = . Thus 1.2m z = .
For the total system,
( ) ( )
2
2 24.5 26.5kgm
p p cyl p dsk
I I I = + = + =
( ) ( )
( ) ( )
2 2
2 2 2
4 100 0.4 12.3333 100 0.4 48.3333kgm
d dC dC
cyl dsk
I I Mh I Mh = + + +
= + + + =

(b) ( ) 48.3333 5 241.66Nm
x d
M I = = =
7000 2
26.5 2 38,851Nm
60
y p
M I
| |
= = =
|
\ .

(c) Acceleration of the centre of gravity is 0.4
c
a g z g = + . Thus
28

( )
2
0.4 9.81 1.2 5 9.81 19.7340m s
c
a = + + = . Denoting a vertical force by V,
200 19.7340 3946.8N
c c
V Ma = = = . The distance from bearing 1 to bearing 2 is
2
1.3m
b
z = .
2 2
1.3 1.2 0.1m
b
z z z = = = .
( ) ( )
1 2 2
241.66 0.1 3946.8 1.3 117,7N
b x c b
V M z V z = + = + =
2 1
3946.8 117.7 3829.1N
b c b
V V V = = =
Denoting a vertical force by H,
2
388511.3 29,885N
b y b
H M z = = = The
horizontal force acting n the bearings are equal and opposite.
The MATLAB script Pr oblem_03_13.m repeats these calculations thus:

Cof G. Di st ance f r ombr g 1 = 1. 2000 m
Combi ned syst em, I d = 48. 3333 kg m^2
Combi ned syst em, I p = 26. 5000 kg m^2
Moment _y = 38851. 0291 Nm
Moment _x = - 241. 6667 Nm
Ver t i cal l oad br g 1 = 117. 7026 N
Ver t i cal l oad br g 2 = 3829. 0974 N
Hor i zont al l oads = 29. 8854 kN and - 29. 8854 kN
29

Chapter 4

In this chapter almost all the solutions require either the element matrices for the axial
deflection of a bar or the lateral deflection of a beam. These are given in text and are
repeated here for convenience.
For a bar in axial vibration (see Equation (4.12) for details)
2 1 1 1
,
1 2 1 1 6
e e e e e
e e
e
A l E A
l
( (
= =
( (

M K .

For a beam in bending (see Equations (4.24) and (4.25) for details)
2 2 2 2
3
2 2 2 2
156 22 54 13 12 6 12 6
22 4 13 3 6 4 6 2
54 13 156 22 12 6 12 6 420
13 3 22 4 6 2 6 4
e e e e
e e e e e e e e
e e e e e
s s
e e e e
e
e e e e e e e e
l l l l
l l l l l l l l
A l E I
l l l l
l
l l l l l l l l
( (
( (

( (
= =
( (

( (
( (


M K .

Pr oblem 4.1
Assembling two axial deflection elements of equal length ( 2
e
l L = ) gives:
2 1 0 0 0 0 2 1 0
1 2 0 0 2 1 1 4 1
6 2 6 2 6 2
0 0 0 0 1 2 0 1 2
A L A L A L
( ( (
| | | | | |
( ( (
= + =
| | |
( ( (
\ . \ . \ .
( ( (

M .
1 1 0 0 0 0 1 1 0
2 2 2
1 1 0 0 1 1 1 2 1
0 0 0 0 1 1 0 1 1
EA EA EA
L L L
( ( (
| | | | | |
( ( (
= + =
| | |
( ( (
\ . \ . \ .
( ( (

K .
With free-free boundary conditions there are no constraints to be applied. Forming
and solving the eigenvalue problem = Ku Mu , where
2
= , gives three natural
frequencies. The lowest one is zero, because the unconstrained bar can move as a
rigid body.
With fixed-free boundary conditions we must eliminate the first row and column from
the system matrices. Thus
4 1 2 1
2
,
1 2 1 1 6 2
A L
EA
L
( ( | | | |
= =
| | ( (

\ . \ .

M K .
Forming and solving this eigenvalue problem gives two natural frequencies.
The MATLAB script Pr oblem_04_01.m solves the above eigenvalue problems to
give

I n model , r ho = A = E = L = 1
Fr ee- f r ee syst em- 1st el ast i c f r eq = 3. 4641 r ad/ s
Fi xed- f r ee syst em- 1st f r eq = 1. 6114 r ad/ s

Adding a force ( ) f t to the free end gives
( )
0
f t

+ =
`
)
Mq Kq .
30

Pr oblem 4.2
Assembling two axial deflection elements of equal length gives:
2 1 0 0 0 0 2 1 0
1 2 0 0 2 1 1 4 1
6 2 6 2 6 2
0 0 0 0 1 2 0 1 2
A L A L A L
( ( (
| | | | | |
( ( (
= + =
| | |
( ( (
\ . \ . \ .
( ( (

M .
1 1 0 0 0 0 1 1 0
2 2 2
1 1 0 0 1 1 1 2 1
0 0 0 0 1 1 0 1 1
EA EA EA
L L L
( ( (
| | | | | |
( ( (
= + =
| | |
( ( (
\ . \ . \ .
( ( (

K .
Note that 2
e
l L = .
With fixed-fixed ends we must eliminate the first and third rows and columns to give
| | | |
2
4 , 2
12
AL EA
L

= = M K . These are, of course, scalar quantities and so

2 2
12 3.4641
n
E L E L = =

Assembling three axial deflection elements of equal length gives
2 1 0 0 0 0 0 0 0 0 0 0
1 2 0 0 0 2 1 0 0 0 0 0
6 3 0 0 0 0 6 3 0 1 2 0 6 3 0 0 2 1
0 0 0 0 0 0 0 0 0 0 1 2
A L A L A L
( ( (
( ( (
| | | | | |
( ( (
= + +
| | |
( ( (
\ . \ . \ .
( ( (

M
1 1 0 0 0 0 0 0 0 0 0 0
1 1 0 0 0 1 1 0 0 0 0 0
3 3 3
0 0 0 0 0 1 1 0 0 0 1 1
0 0 0 0 0 0 0 0 0 0 1 1
EA EA EA
L L L
( ( (
( ( (

| | | | | |
( ( (
= + +
| | |
( ( (
\ . \ . \ .
( ( (

K .
Hence
2 1 0 0 1 1 0 0
1 4 1 0 1 2 1 0
3
,
0 1 4 1 0 1 2 1 6 3
0 0 1 2 0 0 1 1
( (
( (

| | | |
( (
= =
| |
( (
\ . \ .
( (

A L
EA
L
M K .
Eliminating the first and fourth rows and columns gives
4 1 2 1
3
,
1 4 1 2 18
AL EA
L
( (
= =
( (

M K . The eigenvalue problem can be formed and
solved to give two natural frequencies

Assembling four axial deflection elements of equal length gives
2 1 0 0 0 1 1 0 0 0
1 4 1 0 0 1 2 1 0 0
4
, 0 1 4 1 0 0 1 2 1 0
6 4
0 0 1 4 1 0 0 1 2 1
0 0 0 1 2 0 0 0 1 1
( (
( (

( (
| | | |
( ( = =
| |
\ . \ . ( (

( (
( (

A L
EA
L
M K
Eliminating the first and fifth rows and columns gives
31

4 1 0 2 1 0
4
1 4 1 , 1 2 1
24
0 1 4 0 1 2
s s
AL EA
L
( (

( (
= =
( (
( (

M K . The eigenvalue problem can be
formed and solved to give three natural frequencies.
The MATLAB script Pr oblem _04_02.m solves the above formulations to give

I n model , r ho = A = E = L = 1
cl amped- cl amped, 2 el ement - 1st nat f r eq = 3. 4641 r ad/ s
cl amped- cl amped, 3 el ement - 1st nat f r eq = 3. 2863 r ad/ s
cl amped- cl amped, 4 el ement - 1st nat f r eq = 3. 2228 r ad/ s

Pr oblem 4.3
Modelling the system with two elements
1 1 0 0 0 0 0 0 0
2 2
1 1 0 0 1 1 0 0 0
0 0 0 0 1 1 0 0
EA EA
L L
k
( ( (
| | | |
( ( (
= + +
| |
( ( (
\ . \ .
( ( (

K
2 1 0 0 0 0
1 2 0 0 2 1
6 2 6 2
0 0 0 0 1 2
A L A L
( (
| | | |
( (
= +
| |
( (
\ . \ .
( (

M . Thus
*
2 1 0 1 1 0
2
1 4 1 , 1 2 1
6 2
0 1 2
0 1 1
A L
EA
L
k
(
(
( | | | |
(
= =
| | (
(
\ . \ .
(
(
+

M K where
*
2 k kL EA = .
The bar is clamped at the left hand end, so we must eliminate the first row and
column to give
*
2 1
4 1
2
,
1 2 12
1 1
AL EA
L
k
(
(
= =
(
(
+


M K . Since 2 k EA L = ,
*
0.25 k = and hence
4 1 2 1
2
,
1 2 1 1.25 12
AL EA
L
( (
= =
( (

M K . Solving the eigenvalue problem (see
MATLAB script Pr oblem_04_03.m gives

I n model , r ho = A = E = L = 1, k0 = E*A/ ( 2*L)
Cl amped- f r ee syst emwi t h spr i ng - 1st f r eq = 1. 9027 r ad/ s

Pr oblem 4.4
Modelling the system with two elements
1 1 0 0 0 0
2 2
1 1 0 0 1 1
0 0 0 0 1 1
EA EA
L L
( (
| | | |
( (
= +
| |
( (
\ . \ .
( (

K
2 1 0 0 0 0 0 0 0
1 2 0 0 2 1 0 0 0
6 2 6 2
0 0 0 0 1 2 0 0
A L A L
m
( ( (
| | | |
( ( (
= + +
| |
( ( (
\ . \ .
( ( (

M . Thus
32

*
2 1 0 1 1 0
2
1 4 1 , 1 2 1
6 2
0 1 1
0 1 2
A L
EA
L
m
(
(
( | | | |
(
= =
| | (
(
\ . \ .
(
(
+

M K where
*
12 m m AL = .
The bar is clamped at the left hand end, so we must eliminate the first row and
column to give
*
4 1
2 1
2
,
1 1 12
1 2
AL EA
L
m
( (
= =
(
(

+


M K . Since 3 m AL = ,
*
36 m = and hence
4 1 2 1
2
,
1 38 1 1 12
AL EA
L
( (
= =
( (

M K . Solving the eigenvalue problem must be
solved, (see MATLAB script Pr oblem_04_04.m) gives

I n model , r ho = A = E = L = 1, m0 = 3*r ho*A*L
Cl amped- f r ee syst emwi t h mass - 1st f r eq = 0. 54733 r ad/ s

Pr oblem 4.5
Assembling an axial deflection element of length 3 4 L and an axial deflection
element of length 4 L gives
2 1 0 0 0 0 6 3 0
3
1 2 0 0 2 1 3 8 1
6 4 6 4 6 4
0 0 0 0 1 2 0 1 2
A L A L A L
( ( (
| | | | | |
( ( (
= + =
| | |
( ( (
\ . \ . \ .
( ( (

M
1 1 0 0 0 0 1 1 0
4 4 4
1 1 0 0 1 1 1 4 3
3 3
0 0 0 0 1 1 0 3 3
EA EA EA
L L L
( ( (
| | | | | |
( ( (
= + =
| | |
( ( (
\ . \ . \ .
( ( (

K
Since the left hand end is clamped, we must eliminate the first row and column, thus
8 1 4 3
4
,
1 2 3 3 24 3
AL EA
L
( (
= =
( (

M K . The eigenvalue problem can be formulated
and solved, see MATLAB script Pr oblem_04_05.m.
Assembling two axial deflection elements, each of length 2 L , and applying the
boundary conditions has been carried out in Problem 4.1, part (b) and the system
matrices are
4 1 2 1
2
,
1 2 1 1 12
AL EA
L
( (
= =
( (

M K . Again the eigenvalue problem can be
formulated and solved. Pr oblem_04_05.m gives

I n model , r ho = A = E = L = 1
Cl amped- f r ee wi t h equal l engt h el ement s - 1st f r eq = 1. 6114 r ad/ s
Cl amped- f r ee wi t h unequal l engt h el ement s - 1st f r eq = 1. 6157 r ad/ s

Pr oblem 4.6
The arrangement of rows and columns for a lateral displacement element is
| |
1 1 2 2
u u . Thus using a single element to model a pinned-pinned beam, we
must eliminate the first and third rows and columns. Hence
33

2 2 2 2
3
2 2 2 2
4 3 4 2
420
3 4 2 4
L L L L
AL EI
L
L L L L
( (

= = ( (
( (

M K
We can formulate and solve the eigenvalue problem to determine the system natural
frequencies, see MATLAB script Pr oblem_04_06.m. Running this scripts gives

I n model , r ho = A = E = I = L = 1
Pi nned- Pi nned syst emwi t h one el ement - 1st f r eq = 10. 9545 r ad/ s

Pr oblem 4.7
Assembling two lateral deflection elements of equal length gives:
2 2
2 2
2 2
2 2
0 0 0 0 0 0 156 22 54 13 0 0
0 0 0 0 0 0
22 4 13 3 0 0
0 0 156 22 54 13
54 13 156 22 0 0
0 0 22 4 13 3 420
13 3 22 4 0 0
0 0 54 13 156 22
0 0 0 0 0 0
0 0 13 3 22 4 0 0 0 0 0 0
e e
e e e e
e e
e e
e
e e e e
e e e e
e e
e e e e
l l
l l l l
l l
l l
Al
l l l l
l l l l
l l
l l l l
(
(

(
(

( = +

(

(

(
(


M
| | (
|
(
|
(
|
(
|
(
|
(
|
(
|
(
|
| (
\ .

2 2
2 2
3
2 2
2 2
0 0 0 0 0 0 12 6 12 6 0 0
0 0 0 0 0 0
6 4 6 2 0 0
0 0 12 6 12 6
12 6 12 6 0 0
0 0 6 4 6 2
6 2 6 4 0 0
0 0 12 6 12 6
0 0 0 0 0 0
0 0 6 2 6 4 0 0 0 0 0 0
e e
e e e e
e e
e e
e e e e
e
e e e e
e e
e e e e
l l
l l l l
l l
l l
EI
l l l l
l
l l l l
l l
l l l l
| ( (

( (

( (
( (

( ( = +

( (

( (

( (
( (

\
K
|
|
|
|
|
|
|
|
|
|
.


where 2
e
l L = . Note that pairs of rows and columns are overlapped because we are
making both the slope and deflection at the right hand end of the first element to be
equal to the corresponding slope and column of the left hand end of the second
element. For a clamped-clamped beam the first, second, fifth and sixth rows and
columns must be eliminated to give
2 2 3
312 0 24 0
,
420
0 8 0 8
e
e e e
Al EI
l l l
( (

= =
( (
( (

M K .
Thus
( )
11
1
4 4
11
24 420
22.7359
2 312
k EI EI
m
AL
A L

= = =

The MATLAB script Pr oblem_04_07.m) repeats this calculation thus:

I n model , r ho = A = E = I = L = 1
Cl amped- cl amped syst emwi t h t wo el ement - 1st f r eq = 22. 7359 r ad/ s

Pr oblem 4.8
The assembly of two lateral deflection elements of equal length is shown in Problem
4.7 (above) where 2
e
l L = . For a clamped-pinned beam the first, second and fifth
rows and columns must be eliminated to give
34

2 2
2 2
312 0 13
0 8 3
420
13 3 4
e
e
e e
e e e
l
Al
l l
l l l
(
(
=
(
(


M
2 2
3
2 2
24 0 6
0 8 2
6 2 4
e
e e
e
e e e
l
EI
l l
l
l l l
(
(
=
(
(

K . Solving the resultant
eigenvalue problem (see MATLAB script Pr oblem_04_08.m) gives

I n model , r ho = A = E = I = L = 1
Cl amped- pi nned syst emwi t h t wo el ement - 1st f r eq = 15. 5608 r ad/ s

Pr oblem 4.9
Using a single element,
e
l L = . Because the beam is clamped at the left hand end, the
first and second rows and columns of the element matrices must be eliminated.
Introducing an extra coordinate (
3
q in Figure 4.21) and adding the effect of a
concentrated mass m at
3
q , we have
2 2
*
156 22 0 156 22 0
0 0 0
22 4 0 0 0 0 22 4 0
420 420
0 0 0 0 0
0 0
L L
AL AL
L L L L
m
m
( (
(
( (

(
= + =
( (
(
( (
(


M where
*
420 m m AL = . Adding a spring of stiffness k between
1
q and
3
q , gives
*
1
2 2
2
3 3
* *
3
12 6 0 12 * 6
0
6 4 0 0 0 0 6 4 0
0 0 0 0
0
(
+ (
(
(
(

(
( = + = =
` (
(
(
(
(
)
(


L k L k
k k q
EI EI
L L L L q
L L
k k q
k k
K q
where
* 3
k kL EI = . Thus the equation of motion is + = Mq Kq 0 . Note that there is
an error in the stiffness matrix given in the solution for this problem in the text book.

Pr oblem 4.10
Consider Equation (4.37). Since the cross sectional area is now a function of we
must rewrite the integral in the first equation of (4.37) as
( ) ( ) ( )
( )
( )
( )
( )
T
1 1
1 1 2 2
0
2 2
e
l
e e
e e
e e
N N
A N A N d
N N


+
` `


) )

Hence
( ) ( )
( ) ( )
2
1 1 2 2 1 1 1 2 2 1 2
2 0
1 1 2 2 1 2 1 1 2 2 2
e
l
e e e e e e e
e e e e e e e
A N A N N A N A N N N
d
A N A N N N A N A N N
(
+ +
(
( + +

where ( )
1 1 e e
N N = , etc.
Now ( )
1
1
e
e
N
l
= and ( )
1
1
e
e
N
l
= .
Consider (for example) the element in row 1, column 2.
2 2
1 2 1 2
2
0
0
1 2
1 2
2
1 1 1
1
2 2
1 1
2 2 2 2
e
e
l
l
e e e e e e
e
e e
e
e
e
A A d A A
l l l l l l
l
l l A A
A l A
l
l
( | | | | | | | || |
+ = + ( | | | | |
| |
( \ . \ .\ . \ . \ .
| | | | | |
= + = +
| | |
\ . \ . \ .

35

Completing all the integrations gives
1 2
1 1
1 1 2
e
e
A A E
l
( + | |
=
| (

\ .

K
Rewriting the integral in the first equation of Equation (4.39) to account for the
variation of cross section area,
( ) ( ) ( )
( )
( )
( )
( )
T
1 1
1 1 2 2
0
2 2
e
l
e e
e e
e e
N N
A N A N d
N N


+
` `


) )

Hence
( ) ( )
( ) ( )
2
1 1 2 2 1 1 1 2 2 1 2
2 0
1 1 2 2 1 2 1 1 2 2 2
e
l
e e e e e e e
e e e e e e e
A N A N N A N A N N N
d
A N A N N N A N A N N
(
+ +
(
( + +

Consider (for example) the element in row 1, column 2.
( )
( ) ( ) ( )
2 2
2 2
2 2
1 1 2 2 1 2 1 2
0 0
1 2 1
e e
e e e
e e
l l
e e e e
l l l
l l
A N N A N N d A A d

+ = + +

.
Competing the integration process gives ( )
1 2
12
e
l A A + . Completing all the
integrations, we have
1 2 1 2
1 2 1 2
3
3 12
e
e
A A A A
l
A A A A
+ + (
=
(
+ +

M

(a) Using tapered elements
Let
1 2 3
, , A A A be the cross section area at the clamped end, at the mid point and at the
free end of the bar. Then the assembled matrices are
2 3 1 2
1 1 0 0 0 0
1 1 0 0 1 1
2 2
0 0 0 0 1 1
e e
A A A A E E
l l
( (
+ + | | | |
( (
= +
| |
( (
\ . \ .
( (

K and
1 2 1 2
1 2 1 2 2 3 2 3
2 3 2 3
3 0 0 0 0
3 0 0 3
12 12
0 0 0 0 3
e e
A A A A
l l
A A A A A A A A
A A A A
+ + ( (

( (
= + + + + +
( (
( ( + +

M where 2
e
l L = .

(a) Using uniform elements
Let
1 2 3
, , A A A be the cross section area at the clamped end, at the mid point and at the
free end of the bar. The mean cross sectional area of the first element is ( )
1 2
2 A A +
and for the second element is ( )
2 3
2 A A + . Then the assembled matrices are
2 3 1 2
1 1 0 0 0 0
1 1 0 0 1 1
2 2
0 0 0 0 1 1
e e
A A A A E E
l l
( (
+ + | | | |
( (
= +
| |
( (
\ . \ .
( (

K and
2 3 1 2
2 1 0 0 0 0
1 2 0 0 2 1
6 2 6 2
0 0 0 0 1 2
e e
l l A A A A
( (
+ + | | | |
( (
= +
| |
( (
\ . \ .
( (

M where 2
e
l L =
Note that in both models the assembled stiffness matrices are identical, In each case
we apply the boundary condition we must eliminate the first row and column of the
36

assembled matrices. The MATLAB script Pr oblem_04_10.m develop both models
for the linearly tapered bar and solves the resulting eigenvalues as follows:

I n model , r ho = E = I = L = 1. A var i es l i near l y f r om0. 2 t o 0. 1
Cl amped- f r ee syst em: t wo t aper ed el ement - 1st f r eq = 1. 8279 r ad/ s
Cl amped- f r ee syst em: t wo uni f or mel ement - 1st f r eq = 1. 7933 r ad/ s

Pr oblem 4.11
(a) Modelling the bar with three elements gives
2 1 0 0 0 0 0 0 0 0 0 0
1 2 0 0 0 2 1 0 0 0 0 0
0 0 0 0 0 1 2 0 0 0 2 1 6 3 6 3 6 3
0 0 0 0 0 0 0 0 0 0 1 2
A L A L A L
( ( (
( ( (
| | | | | |
( ( (
= + +
| | |
( ( (
\ . \ . \ .
( ( (

M
1 1 0 0 0 0 0 0 0 0 0 0
1 1 0 0 0 1 1 0 0 0 0 0
3 3 3
0 0 0 0 0 1 1 0 0 0 1 1
0 0 0 0 0 0 0 0 0 0 1 1
EA EA EA
L L L
( ( (
( ( (

| | | | | |
( ( (
= + +
| | |
( ( (
\ . \ . \ .
( ( (

K
Applying the boundary condition by eliminating the first row and column gives
2 1 0 4 1 0
3
1 2 1 and 1 4 1
6 3
0 1 1 0 1 2
A L
EA
L
( (
| | | |
( (
= =
| |
( (
\ . \ .
( (

K M
The eigenvalue problem can be formulated and solved, see the MATLAB script
Pr oblem_04_11.m.

(b) Using a single element
2 1 1 1
,
1 2 1 1 6
AL EA
L
( (
= =
( (

M K
Applying the boundary condition by eliminating the first row and column gives
2 6, m AL k EA L = = and thus
2
3 k m E L = =

(c) To reduce the model
2 1 0
3
1 2 1
0 1 1
EA
L
(

(
| |
=
| (
\ .
(

K . Thus
2 1
3
1 2
ss
EA
L
(
| |
=
| (

\ .

K and hence
1
2 1
1
1 2 3 3
ss
L
EA

(
| || |
=
| | (
\ .\ .

K .
0
3
1
sm
EA
L

| |
=
`
|

\ .
)
K and thus
1
2 1 0 13
1
1 2 1 2 3 3
ss sm

(
= =
` `
(

) )
K K . Thus we have
13
2 3
1


=
`

)
T
T
r
EA L = = K T KT and
T
3
r
AL = = M T MT .
Now
2
3 k m E L = = . This is identical to the frequency given by a single
degree of freedom system.
37

The MATLAB script Pr oblem_04_11.m solves .the three cases of this problems and
gives the following results

I n model , r ho = A = E = L = 1
Cl amped- f r ee: 3 el ement s - 1st f r eq = 1. 5888 r ad/ s
Cl amped- f r ee: 1 el ement - 1st f r eq = 1. 7321 r ad/ s
Cl amped- f r ee: Reduced model - 1st f r eq = 1. 7321 r ad/ s


Pr oblem 4.12
Assembling a large number of elements can only be done realistically using a
computer. The MATLAB script Pr oblem_04_12.m assembles a number of axial
deflection elements and applies the boundary conditions, The first three system
natural frequencies are computed using 3, 4, 6 and 8 elements, thus:

I n model , r ho = A = E = I = L = 1
No of el ement s 1st nat f r eq 2nd nat f r eq 3r d nat f r eq ( r ad/ s)
3 9. 8776 39. 9451 98. 5901
4 9. 8722 39. 6342 90. 4495
6 9. 8701 39. 5104 89. 1770
8 9. 8698 39. 4887 88. 9407
Exact 9. 8696 39. 4784 88. 8264

The also script computes the first three natural frequencies using up to 64 elements
and plots the percentage error against the number of elements,

0 10 20 30 40 50 60 70
10
-8
10
-6
10
-4
10
-2
10
0
10
2
Number of Elements
%

E
r
r
o
r

38

Chapter 5

Note. Solving Problems 5.1, 5.2, 5.3, 5.8, 5.9 and 5.11 requires a finite element
analysis that allows shafts to be modelled and include gyroscopic effects, etc. Here
we use the rotordynamics software developed to accompany this book, but other
appropriate software can be used.

Problem 5.1
Modelling this problem requires finite element analysis (FEA) and can only be solved
using appropriate FEA software. Here MATLAB script Problem_05_01.m makes
use the Rotordynamics Software package to model and analyse the system. The
diagrams below, generated by the script, shows the solid and hollow rotors modelled
with 16 elements.
Brg Type 1 Brg Type 1
N
o
d
e

1
N
o
d
e

2
N
o
d
e

3
N
o
d
e

4
N
o
d
e

5
N
o
d
e

6
N
o
d
e

7
N
o
d
e

8
N
o
d
e

9
N
o
d
e

1
0
N
o
d
e

1
1
N
o
d
e

1
2
N
o
d
e

1
3
N
o
d
e

1
4
N
o
d
e

1
5
N
o
d
e

1
6
N
o
d
e

1
7

Solid shaft modelled with 16 elements.
Brg Type 1 Brg Type 1
N
o
d
e

1
N
o
d
e

2
N
o
d
e

3
N
o
d
e

4
N
o
d
e

5
N
o
d
e

6
N
o
d
e

7
N
o
d
e

8
N
o
d
e

9
N
o
d
e

1
0
N
o
d
e

1
1
N
o
d
e

1
2
N
o
d
e

1
3
N
o
d
e

1
4
N
o
d
e

1
5
N
o
d
e

1
6
N
o
d
e

1
7

Hollow shaft modelled with 16 elements

The output of the script is as follows:
39

Sol i d shaf t , 16 Ti moshenko el ement s
Nat ur al Fr equency 1 = 16. 1328 Hz
Nat ur al Fr equency 2 = 17. 1523 Hz
Nat ur al Fr equency 3 = 64. 0008 Hz
Nat ur al Fr equency 4 = 68. 8696 Hz
Nat ur al Fr equency 5 = 144. 2202 Hz

Sol i d shaf t , 16 Eul er - Ber noul l i el ement s
Nat ur al Fr equency 1 = 16. 1527 Hz
Nat ur al Fr equency 2 = 17. 1776 Hz
Nat ur al Fr equency 3 = 64. 3551 Hz
Nat ur al Fr equency 4 = 69. 3417 Hz
Nat ur al Fr equency 5 = 144. 601 Hz

Hol l ow shaf t , 16 Ti moshenko el ement s
Nat ur al Fr equency 1 = 16. 6622 Hz
Nat ur al Fr equency 2 = 17. 7754 Hz
Nat ur al Fr equency 3 = 64. 7167 Hz
Nat ur al Fr equency 4 = 69. 5014 Hz
Nat ur al Fr equency 5 = 145. 6799 Hz

Hol l ow shaf t , 16 Eul er - Ber noul l i el ement s
Nat ur al Fr equency 1 = 16. 7702 Hz
Nat ur al Fr equency 2 = 17. 9158 Hz
Nat ur al Fr equency 3 = 66. 6536 Hz
Nat ur al Fr equency 4 = 72. 1014 Hz
Nat ur al Fr equency 5 = 147. 602 Hz

Sol i d shaf t : di f f er ence i n nat f r eq 1 = 0. 1234%
Sol i d shaf t : di f f er ence i n nat f r eq 2 = 0. 1479%
Sol i d shaf t : di f f er ence i n nat f r eq 3 = 0. 55356%
Sol i d shaf t : di f f er ence i n nat f r eq 4 = 0. 68546%
Sol i d shaf t : di f f er ence i n nat f r eq 5 = 0. 26401%

Hol l ow shaf t : di f f er ence i n nat f r eq 1 = 0. 64794%
Hol l ow shaf t : di f f er ence i n nat f r eq 2 = 0. 78969%
Hol l ow shaf t : di f f er ence i n nat f r eq 3 = 2. 993%
Hol l ow shaf t : di f f er ence i n nat f r eq 4 = 3. 7408%
Hol l ow shaf t : di f f er ence i n nat f r eq 5 = 1. 3194%

The inside and outside diameters of the hollow shaft are chosen to make the natural
frequencies of the hollow shaft system very close to those of the solid shaft. However,
the differences between the natural frequencies obtained using Euler-Bernoulli and
the Timoshenko elements are greater for the hollow shaft than for the solid shaft.
40

Problem 5.2
Modelling this system requires finite element analysis (FEA) and can only be solved
using appropriate FEA software. Here MATLAB script Problem_05_02.m makes
use the Rotordynamics Software package to model and analyse the system. The
diagram below, generated by the script, shows the rotor, modelled with 16 elements.
Brg Type 1 Brg Type 1
N
o
d
e

1
N
o
d
e

2
N
o
d
e

3
N
o
d
e

4
N
o
d
e

5
N
o
d
e

6
N
o
d
e

7
N
o
d
e

8
N
o
d
e

9
N
o
d
e

1
0
N
o
d
e

1
1
N
o
d
e

1
2
N
o
d
e

1
3
N
o
d
e

1
4
N
o
d
e

1
5
N
o
d
e

1
6
N
o
d
e

1
7

Rotor modelled with 16 elements

The output of the script is as follows:

Fi r st f i ve nat ur al Fr equenci es ( Hz)
4 el ement s 8 el ement s 16 el ement s
25. 0929 25. 0928 25. 0928
30. 1380 30. 1378 30. 1378
59. 4160 59. 4145 59. 4144
65. 9957 65. 9937 65. 9935
162. 5512 162. 5205 162. 5182

Note that as the number of elements is increased the change in a particular estimated
frequency decreases.


Problem 5.3
Modelling this problem involves finite element analysis (FEA) and can only be
solved using appropriate FEA software. Here MATLAB script Problem_05_03.m
makes use the Rotordynamics Software package to model and analyse the system.
The diagrams below, generated by the script, shows the rotor models. In model 1,
four tapered elements and three uniform elements are used, in model 2, seven uniform
elements of different diameters are used and in model 3, eleven uniform elements if
different diameters are used. In models 2 and 3 the diameters of the uniform elements
are chosen to be the average of the shaft diameters at each end of the element.

41

Brg Type 1 Brg Type 1
N
o
d
e

1
N
o
d
e

2
N
o
d
e

3
N
o
d
e

4
N
o
d
e

5
N
o
d
e

6
N
o
d
e

7
N
o
d
e

8

Model 1. Model with 4 tapered elements and 3 uniform elements.
Brg Type 1 Brg Type 1
N
o
d
e

1
N
o
d
e

2
N
o
d
e

3
N
o
d
e

4
N
o
d
e

5
N
o
d
e

6
N
o
d
e

7
N
o
d
e

8

Model 2. Model with 7 uniform elements of three different diameters.
Brg Type 1 Brg Type 1
N
o
d
e

1
N
o
d
e

2
N
o
d
e

3
N
o
d
e

4
N
o
d
e

5
N
o
d
e

6
N
o
d
e

7
N
o
d
e

8
N
o
d
e

9
N
o
d
e

1
0
N
o
d
e

1
1
N
o
d
e

1
2

Model 3. Model with 11 uniform elements of different diameters.
42


The output of script is as follows:

Model 1. Seven t aper ed and uni f or mel ement s
Nat ur al f r equency 1 = 27. 7103 Hz, Er r or = 0. 0183
Nat ur al f r equency 2 = 32. 5623 Hz, Er r or = 0. 0343
Nat ur al f r equency 3 = 40. 8942 Hz, Er r or = 0. 1192
Nat ur al f r equency 4 = 50. 9622 Hz, Er r or = 0. 2002
Nat ur al f r equency 5 = 93. 1906 Hz, Er r or = 2. 2136

Model 2. Seven uni f or mel ement s
Nat ur al f r equency 1 = 27. 4093 Hz, Er r or = 0. 2827
Nat ur al f r equency 2 = 32. 1892 Hz, Er r or = 0. 3388
Nat ur al f r equency 3 = 39. 6128 Hz, Er r or = 1. 1622
Nat ur al f r equency 4 = 48. 3880 Hz, Er r or = 2. 3740
Nat ur al f r equency 5 = 109. 0774 Hz, Er r or = 18. 1004

Model 3. El even uni f or mel ement s
Nat ur al f r equency 1 = 27. 6201 Hz, Er r or = 0. 0719
Nat ur al f r equency 2 = 32. 4496 Hz, Er r or = 0. 0784
Nat ur al f r equency 3 = 40. 5212 Hz, Er r or = 0. 2538
Nat ur al f r equency 4 = 50. 2079 Hz, Er r or = 0. 5541
Nat ur al f r equency 5 = 95. 7711 Hz, Er r or = 4. 7941

Problem 5.4
The equations of motion for this system are
2 2 2 2
2 2 2 2
0
0
d p uu uv uu uv
d p vu vv vu vv
I I c L c L k L k L
I I c L c L k L k L
+ + + + + =
+ + + + =





These equation are similar to those given in the solution of Problem 3.3, except that
the stiffness and damping properties are different to the x and y directions and there is
cross coupling between the stiffness and damping in the x and y directions. Thus we
have in matrix notation,
2 2
0
0 0
0 0 0
(
( ( (
+ + + =
( ` ` ` ` `
( ( (

) ) ( ) ) )



p
d uu uv uu uv
d vu vv vu vv p
I
I c c k k
L L
I c c k k I

The stiffness and damping coefficients are based on the hydrodynamic bearing
properties. To determine the stiffness and damping of the bearing at 3000 rev/min,
from Equation (5.84)
( )
( )
( )
3 3
2 2
3
0.1 3000 2 60 0.03 0.020
0.0393
8
8 1200 2 0.2 10
b
s
D L
S
fc

= = =

where.
From Equation (5.85), the Somerfeld number,
2
0.3125
s
b
S D
S
L
| |
| |
= =
|
|

\ .
\ .
.
To determine the eccentricity it is necessary to solve the quartic equation given by
Equation (5.83).
( ) ( ) ( )
4 3 2 2 2 2 2
4 6 0.0400 16 4 0.0400 1 0 + + + = where
2
= .
The equation must be solved numerically, and the smallest root is 0.6557 = . Thus
0.8098 = . Using Equation (5.61)
43

( )
3 2
2
2 2
521.47N
2 1
b
r
D L
f
c

= =

and
( )
3
3 2
2 2
296.76N
8 1
b
t
D L
f
c

= =

From Equation (5.86)
2
1
tan 0.5691
4

= =

. Hence
( )
1
tan 0.5726 0.5174rador 0.5201 180 29.7974

= = = =
To determine the stiffness and coefficients for the bearings we must evaluate
Equation (5.87). This is a tedious process and the details are not shown. The results
are 10348Ns m, 18184Ns m, 79640Ns m
uu uv vu vv
c c c c = = = =
5.508MN m, 2.188MN m, 16.323MN m, 28.684MN m.
uu uv vu vv
k k k k = = = =
Using the data provided and computed, the equations of motion in matrix form when
the rotor spins at 3000 rev/min are

=
`

)
q ,
2
10 0
Nms
0 10
(
=
(

M ,
0 188.496
Nms
188.496 0
(
=
(

G ,
4
0.2587 0.4546
10 Nms
0.4546 1.9910
(
=
(

C ,
6
1.3770 0.5470
10 Nm
4.0809 7.1711
(
=
(

K
where ( ) Mq + G+C q +Kq = 0 . This equation can be transformed into an eigenvalue
problem, see section 5.8. The MATLAB script Problem_05_04.m determines the
bearing stiffness and damping matrix elements and solves the resultant eigenvalue
problem. The output of the script is as follows:

Rot or speed = 3000 r ev/ mi n: - - -
Somer f el d number = 0. 3125
Eccent r i ci t y = 0. 80978
Radi al f or ce = 521. 4747N
Tangent i al f or ce = 296. 756N
Gamma = 29. 6429degr ees

Root s = ( - 1605. 9467) r ad/ s and ( - 563. 3804) r ad/ s
Root s = ( - 40. 18492+287. 8334i ) r ad/ s and ( - 40. 18492- 287. 8334i ) r ad/ s

Nat ur al f r eq ( Hz) Damped nat f r eq ( Hz) Zet a
46. 2544 45. 8101 0. 13827
46. 2544 45. 8101 0. 13827

Rot or speed = 6000 r ev/ mi n: - - -
Somer f el d number = 0. 625
Eccent r i ci t y = 0. 73606
Radi al f or ce = 486. 3917N
Tangent i al f or ce = 351. 3163N
Gamma = 35. 8402degr ees

Root s = ( - 504. 1876+573. 1688i ) r ad/ s and ( - 504. 1876- 573. 1688i ) r ad/ s
Root s = ( - 22. 53692+325. 6684i ) r ad/ s and ( - 22. 53692- 325. 6684i ) r ad/ s

Nat ur al f r eq ( Hz) Damped nat f r eq ( Hz) Zet a
121. 4935 91. 2226 0. 66048
121. 4935 91. 2226 0. 66048
51. 9557 51. 8317 0. 06904
51. 9557 51. 8317 0. 06904

44

Problem 5.5
Using Equation (5.97), the equations of motion for this system are
0
0
s
s
mu cu ku k v
mv cv k u kv
+ + + =
+ + =


. Let
0
st
u u e = , etc to give
2
2
0
s
s
ms cs k k
k ms cs k
+ +
=
+ +
.
Hence (taking square roots)
2
s
ms cs k jk + + = . Let s j = (i.e. assuming the real
part of the solution is zero) gives
2
0 + =
s
k m jc jk . Considering the real and
imaginary parts we have
2
0 k m = and 0
s
k c + = . Thus
n
k m = = and
s n crit
c k c = = . Since ( )
s m b
k D L = and P = then at the critical conditions,
( )
s m b crit n
k P D L c = = .
Thus ( )
( ) max crit n m b crit m b
P c D L c D L k m = =
Given that the power 30000W P = at a rotor speed of 9600rev min = , then
9600 2 60 1005.3rad s = = and 29.8416Nm P = = . For this system, 3 = ,
and the blades are of length 0.05m
b
L = with a mean diameter of 0.15m
m
D = .
Thus ( ) ( ) 3 29.8416 0.15 0.05 11937N m
s m b
k D L = = = .
The rotor is of length 0.3m L = (not 300m as stated in question!) with a diameter
0.015m d = carries a central mass (m) of 3kg. Thus, for the shaft,
4 4 9 4
64 0.015 64 2.4850 10 m I d

= = = .
Since
11
2 10 Pa E = , at the mid-span of the shaft,
3 11 9 3
48 48 2 10 2.4850 10 0.3 883570N m k EI L

= = = . Thus we have
883570 3 542.7rad s
n
k m = = = and hence 21.9948Ns m
crit s n
c k = =
These calculations are repeated in the MATLAB script Problem_05_05.m.
In order to determine the system properties for various values of damping, it is
necessary to solve an eigenvalue problem derived from the system equations. The
eigenvalue problem is solved in the MATLAB script Problem_05_05.m. From the
roots s, we have
( ) imag 2
d
s = , 2
n
s = and ( ) real s s = . To determine
the direction of the orbit we must compute This is done in Problem_05_05.m
using the MATLAB function whi r l . mwhich is provided in the Rotordynamics
software package. The output from the script is

k_sw = 11936. 6207 N/ m
Cr i t i cal dampi ng = 21. 9948 Ns/ m

Dampi ng ( Ns/ m) Damped f r eq ( Hz) Nat f r eq ( Hz) zet a kappa
0 86. 3755 86. 3775 0. 006754 - 1
0 86. 3755 86. 3775 - 0. 006754 1

20 86. 3739 86. 3739 - 0. 000613 1
20 86. 3739 86. 3811 0. 012896 - 1

40 86. 3690 86. 3703 0. 005529 1
40 86. 3690 86. 3846 0. 019037 - 1

45

Note that (i) the value of 1 = implying that the forward or backward whirl orbit is
circular, (2) that when 0 c = and 20Ns m c = one value of is negative, i.e. the
system is unstable, as we expect, since the critical damping is 21.995 Ns m.


Problem 5.6
We now introduce an error into Equation (5.97), the equations of motion for this
system by changing the sign of
s
k . Thus we have
0
0
+ + =
+ + + =


s
s
mu cu ku k v
mv cv k u kv
. Let
0
st
u u e = , etc to give
2
2
0
+ +
=
+ +
s
s
ms cs k k
k ms cs k
.
Hence (taking square roots)
2
s
ms cs k jk + + = . Let s j = (i.e. assuming the real
part of the solution is zero) gives
2
0 + =
s
k m jc jk . Considering the real and
imaginary parts we have
2
0 k m = and 0
s
k c + = . Note that these equations are
identical to those of Problem 5.5, so the frequencies computed will be identical to
those of Problem 5.5. However, where as
2
0
0 s
u ms cs k
v k
+ +
= in the correctly
formulated analysis,
2
0
0 s
u ms cs k
v k
+ +
=

in the incorrectly formulate analysis. Thus the

sign of
0 0
u v will be reversed in the incorrect formulation. So if
0 0
u v is positive
the incorrect formulation will be negative and vice-versa. The direction of whirl will
be reversed and in error.

Problem 5.7
From Equation (5.96)
0 0
0 0
2
1 1
0
4 2
2
1 1
0
2 4
0
0
0
0
d d d
d d d
d d d
d d d
m m c c m u u
m m m c c v v
k k m c
u
v
c k k m
+ + ( (
+
` `
( (
+ +
) )
(
+

(
+ =
` `
(
) ) +




Let
0 d
m m m = + ,
0
= +
d
k k k ,
0
= +
d
c c c and
0
st
u u e = , etc. Hence,
( )
( )
2 2
1 1
4 2
2 2
1 1
2 4
0
d d d
d d d
ms cs k m c m s
c m s ms cs k m
+ + +
=
+ + +

To determine the stability boundary, the real part of the solution is zero, and hence we
let s j = . Under these conditions,
c
and
c
(where subscript c implies
the critical value at the boundary of stability. Thus
( )
( )
2 2
1 1
4 2
2 2
1 1
2 4
0
+ + +
=
+ + +
c c d c d c d c
d c d c c c d c
m j c k m c j m
c j m m j c k m
. Thus
( ) ( )
2
2
2 2 2
1 1
4 2
0 + + + + =
c c d c d c d c
m j c k m c j m
46

( )
2 2
1 1
4 2
+ + = +
c c d c d c d c
m j c k m j c j m . Hence
( )
2 2
1
4
+ =
c d c c d c
m k m m and
1
2
=
c d c
c c
Taking the positive sign in the imaginary equation, the limit of stability is when
( ) 2 2 = =
c c d c
c c r where
d
r c c = . Substituting this value of
c
into the real
part of the equation (and taking the positive sign) gives
( ) ( )
2 2
2
1
4
2 2 2 + =
c d c d c
m r k m r rm and hence
2
2
2
=
+
c
d d
k
mr m m r
.
Note that when
0
0 c = , ( )
0
1
d d
r c c c = + = .
Thus 2 =
c c
and
0
2 2 = =

c
d
k k
m m m

Note also that as
0
c , 0 r and then 2 =
c
d
k
m
and 0 =
c
.
For the shaft,
4 7 4
64 3.068 10 m I d

= = . The shaft stiffness at the centre is (see
Appendix 2, Table A2.1, system 1) is
3 7
0
48 1.3635 10 N m k EI L = = .
7 7 7
1.3635 10 2 10 3.3635 10 N m k = + = .
0
600kg m = , 120kg
d
m = ,
200Ns m
d
c = and 20kN m
d
k = ,
For
0
0 c = ,
7
0
3.3635 10
2 2 473.53rad s 4521.9rev min
600

= = =
c
k
m
and
2 473.53 2 236.76rad s 37.68Hz = = =
c c
.
The critical speed and frequency can be easily calculated for other system damping.
However, MATLAB script Problem_05_07.m calculates these speeds and
frequencies (see output below) and also plots various parameters against rotor speed
for a system damping of
0
160Ns m c = .

Syst emdampi ng c0 = 0. Condi t i ons at l i mi t of st abi l i t y
Shaf t speed = 4521. 9312 r ev/ mi n
Whi r l f r equency = 37. 6828 Hz

Syst emdampi ng c0 = 80 Ns/ m. Condi t i ons at l i mi t of st abi l i t y
Shaf t speed = 6231. 7804 r ev/ mi n
Whi r l f r equency = 37. 0939 Hz

Syst emdampi ng c0 = 160 Ns/ m. Condi t i ons at l i mi t of st abi l i t y
Shaf t speed = 7663. 7602 r ev/ mi n
Whi r l f r equency = 35. 4804 Hz

Syst emdampi ng c0 t end t o i nf i ni t y. Condi t i ons at l i mi t of st abi l i t y
Shaf t speed = 10111. 3456 r ev/ mi n
Whi r l f r equency = 0 Hz
47

Problem 5.8
This problem involves finite element analysis (FEA) and can only be solved using
appropriate FEA software.
Case 1. When the bearings are rigid, the bearings act in their nominal position, i.e.
0.1m from end of the shaft. There is no axial tension in the shaft.
Case 2. When contact angle is 20 the point of contact of the bearing on the shaft is
shifted by
( ) 0.035tan 20 0.0127m = = so that the bearings are now 0.0873m from the ends of
the shaft.. A tensile force of 500N acts on the shaft.
Case 3. If 0.0127m = a tensile force on the shaft can be applied to the shaft which
will negate the effect of this shift in bearing position. To do this it is necessary to
calculate the first natural frequency for a range of tensile forces and interpolate to find
the force that gives a first natural frequency identical to that of Case 1 (above).

Here MATLAB script Problem_05_08.m makes use the Rotordynamics Software
package to model and analyse the system. The diagrams below, generated by the
script, shows the rotor modelled with 12 elements.

Brg Type 1 Brg Type 1
N
o
d
e

1
N
o
d
e

2
N
o
d
e

3
N
o
d
e

4
N
o
d
e

5
N
o
d
e

6
N
o
d
e

7
N
o
d
e

8
N
o
d
e

9
N
o
d
e

1
0
N
o
d
e

1
1
N
o
d
e

1
2
N
o
d
e

1
3

The output of the script is as follows:

Tensi l e f or ce appl i ed = 0 N
Di spl acement of t he bear i ngs out war ds = 0mm
Fr equency 1 = 23. 2513 Hz
Fr equency 2 = 24. 3349 Hz
Fr equency 3 = 79. 8648 Hz
Fr equency 4 = 127. 6658 Hz
Fr equency 5 = 168. 3636 Hz

Tensi l e f or ce appl i ed = 500 N
Di spl acement of t he bear i ngs out war ds = 12. 739mm
Fr equency 1 = 22. 3558 Hz
Fr equency 2 = 23. 3372 Hz
Fr equency 3 = 77. 8035 Hz
Fr equency 4 = 123. 1271 Hz
Fr equency 5 = 163. 4651 Hz

For ce t o negat e t he ef f ect of a cont act angl e of 20 deg: 14905. 5202 N

48

Problem 5.9
This problem involves finite element analysis (FEA) and can only be solved using
appropriate FEA software. In order to compute the bearing properties we require the
static load acting on each bearing. From the data of Table 5.11, we can compute the
volume and hence the mass of the shaft. Similarly, the question gives the details of
the disks and the volume and hence the mass can be determined. Thus we have
101.914kg
sft
m = , 34.184kg
dsk
m = and adding these together the total mass of the
rotor is 136.097kg. Thus total force on the bearings is 136.097 13.35kN F g = = .
This can then be divided in ratios 1:1:1, 1:3:1 and 2:1:2, i.e 445N: 445N: 445N,
267N:801N:267N and 534N:267N:534N.
The MATLAB script Problem_05_09.m makes use the Rotordynamics Software
package to model and analyse the system. The diagrams below, generated by the
script, shows the rotor modelled with 28 Timoshenko elements. Note that all the
hydrodynamic bearing configurations give a stable system.

Brg Type 3 Brg Type 3 Brg Type 3
N
o
d
e

1
N
o
d
e

2
N
o
d
e

3
N
o
d
e

4
N
o
d
e

5
N
o
d
e

6
N
o
d
e

7
N
o
d
e

8
N
o
d
e

9
N
o
d
e

1
0
N
o
d
e

1
1
N
o
d
e

1
2
N
o
d
e

1
3
N
o
d
e

1
4
N
o
d
e

1
5
N
o
d
e

1
6
N
o
d
e

1
7
N
o
d
e

1
8
N
o
d
e

1
9
N
o
d
e

2
0
N
o
d
e

2
1
N
o
d
e

2
2
N
o
d
e

2
3
N
o
d
e

2
4
N
o
d
e

2
5
N
o
d
e

2
6
N
o
d
e

2
7
N
o
d
e

2
8
N
o
d
e

2
9


Rot or spd = 0, I sot r opi c br gs
Nat ur al f r equency 1 = 19. 0763 Hz
Nat ur al f r equency 2 = 19. 0763 Hz
Nat ur al f r equency 3 = 38. 7344 Hz
Nat ur al f r equency 4 = 38. 7344 Hz
Nat ur al f r equency 5 = 67. 2388 Hz
Nat ur al f r equency 6 = 67. 2388 Hz

Rot or spd = 3000r ev/ mi n, I sot r opi c br gs
Nat ur al f r equency 1 = 18. 9058 Hz, kappa = - 1. 0000
Nat ur al f r equency 2 = 19. 2469 Hz, kappa = 1. 0000
Nat ur al f r equency 3 = 38. 4893 Hz, kappa = - 1. 0000
Nat ur al f r equency 4 = 38. 9753 Hz, kappa = 1. 0000
Nat ur al f r equency 5 = 65. 8526 Hz, kappa = - 1. 0000
Nat ur al f r equency 6 = 68. 5804 Hz, kappa = 1. 0000

Rot or spd = 0, Ani sot r opi c br gs
Nat ur al f r equency 1 = 18. 8252 Hz
Nat ur al f r equency 2 = 19. 0763 Hz
Nat ur al f r equency 3 = 36. 9948 Hz
Nat ur al f r equency 4 = 38. 7344 Hz
Nat ur al f r equency 5 = 63. 3190 Hz
Nat ur al f r equency 6 = 67. 2388 Hz

49

Rot or spd = 3000r ev/ mi n, Ani sot r opi c br gs
Nat ur al f r equency 1 = 18. 7426 Hz, kappa = - 0. 4948
Nat ur al f r equency 2 = 19. 1590 Hz, kappa = 0. 5010
Nat ur al f r equency 3 = 36. 9661 Hz, kappa = - 0. 1596
Nat ur al f r equency 4 = 38. 7591 Hz, kappa = 0. 0978
Nat ur al f r equency 5 = 62. 9348 Hz, kappa = - 0. 2357
Nat ur al f r equency 6 = 67. 5727 Hz, kappa = 0. 2983

Hydr odynami c br g, f or ce = 1: 1: 1
Damped f r equency 1 = 18. 8745 Hz, Zet a = 0. 0024, kappa = 0. 1463
Damped f r equency 2 = 19. 6652 Hz, Zet a = 0. 0045, kappa = - 0. 0047
Damped f r equency 3 = 34. 9988 Hz, Zet a = 0. 0628, kappa = 0. 2970
Damped f r equency 4 = 45. 7624 Hz, Zet a = 0. 0387, kappa = - 0. 1739
Damped f r equency 5 = 52. 7040 Hz, Zet a = 0. 3367, kappa = 0. 4842
Damped f r equency 6 = 53. 0997 Hz, Zet a = 0. 4313, kappa = 0. 4302

Hydr odynami c br g, f or ce = 1: 3: 1
Damped f r equency 1 = 18. 9386 Hz, Zet a = - 0. 0016, kappa = 0. 2121
Damped f r equency 2 = 19. 6256 Hz, Zet a = 0. 0072, kappa = - 0. 0632
Damped f r equency 3 = 35. 6859 Hz, Zet a = 0. 0747, kappa = 0. 4200
Damped f r equency 4 = 45. 9458 Hz, Zet a = 0. 0359, kappa = - 0. 1442
Damped f r equency 5 = 44. 2690 Hz, Zet a = 0. 3387, kappa = 0. 5882
Damped f r equency 6 = 49. 1626 Hz, Zet a = 0. 3617, kappa = 0. 1765

Hydr odynami c br g, f or ce = 2: 1: 2
Damped f r equency 1 = 18. 6005 Hz, Zet a = - 0. 0093, kappa = 0. 3250
Damped f r equency 2 = 19. 5302 Hz, Zet a = 0. 0148, kappa = - 0. 1593
Damped f r equency 3 = 32. 4314 Hz, Zet a = 0. 0560, kappa = 0. 2676
Damped f r equency 4 = 45. 5308 Hz, Zet a = 0. 0650, kappa = - 0. 2367
Damped f r equency 5 = 56. 5182 Hz, Zet a = 0. 2383, kappa = 0. 5638
Damped f r equency 6 = 55. 7581 Hz, Zet a = 0. 4457, kappa = 0. 4325

Problem 5.10
Part a: Solve problem 5.1(b).
For the shaft,
( )
4 4 7 4
0
6.1662 10 m
64
s i
I d d

= = .
5 2
1.2332 10 Nm
s
EI =
For the shaft,
( )
2 2
0
6.5611kg m
4
i
A d d

= =
For each disk,
2 2
0
( ) 75.2776kg
4
d d d
M d d t

= =
Assume the shape of the first mode of vibration can be approximated by
( ) ( )
1
sin u z z L = .
Thus ( ) ( ) ( )
1
cos u z L z L = and ( ) ( ) ( )
2
1
sin u z L z L = . Note that when
0 z = and z L = , both ( )
1
u z and ( )
1
u z are zero. Thus the boundary conditions are
satisfied.
( ) ( )
{ }
4
2
2
5
1
max
2
0
sin 7.3321 10
2 2
| || | | |
= = =
| | |
\ . \ . \ .

L
s
s
EI L
U EI L z L dz
L
and
50

( ) { } ( ) { } ( ) { }
( ) ( )
{ }
{ }
2 2 2
2 2 2
1 1 1
max 1 1 1 1 2
2 2 2
0
2 2 2
1
1
2
2
2 1
1
sin sin sin
sin 0.5 sin
2
1.6
6.5611 75.2776 0.6913 0.8536 60.7724
2 2
L
d D d d
d
T A z L dz M z L M z L
L
A M L L
= + +
( | |
= + +
| (
\ .
( | |
= + + =
| (
\ .

max max
T U = . Thus
2 5
1
60.7724 7.3321 10 = so that
5
1
7.3321 10 60.7724 109.64 s = = .
Thus the approximate natural frequency ( ) 109.64 2 17.4816Hz = = .

Part b: Solve Problem 5.2
For the shaft,
4 6 4
1.5532 10 m
64
s s
I d

= = .
5 2
3.1063 10 Nm
s
EI =
For the shaft,
2
34.4593kg m
4
s
A d

= =
For each disk,
2 2
s
( ) 75.6574kg
4
d d d
M d d t

= =
Assume that the shape of first mode of vibration is approximated by
( )
4 5
1
0.2148 0.8815 1.6667 u z z z z = + .
Thus ( )
( )
3 4
1
0.8815 6.6667 5 u z z z L = + and ( )
( )
2 3 2
1
20 20 u z z z L =
When 0.4 = z then 0.25 z = and when 1.6 z = then 1 z = . Substituting these values
into ( )
1
u z , we have ( ) ( ) = =
1 1
0.4 1.6 0 u u so that there is no deflection at the
bearings. When 0 z = then 0 z = and when 1.6 z = then 1 z = . Substituting these
values into ( )

1
u z , we have ( ) ( )
1 1
0 1.6 0 u u = = . Thus there is no bending moment at
the ends of the shaft.
{ }
2
2 2 4 3 5 5
1
max
2 3
0
400 1
20 1.4445 10
2 105
| || || |
= = =
| | |
\ .\ . \ .

L
s
s
EI
U EI z L z L dz
L
and
( ) { } ( ) { } ( ) { } ( ) { }
( )
{ }
2 2 2 2
2 2
1 1
max 1 1 1 1 1 1 2 1 3
2 2
0
2
2 2 2 2
1
1 1
2
34.4593 1.6 0.0158 75.6574 0.2148 0.1530 +0.1563 3.9896
(
= + + +
(

(
= + + =
(

L
d d d d
T A u z dz M u z u z u z
max max
T U = . Thus
2 5
1
3.9896 1.4445 10 = so that
5
1
1.4445 10 3.9896 190.28 s = = .
Thus the approximate natural frequency ( ) 190.28 2 30.2844Hz = = .
51

Problem 5.11
Part (a): For the 2 degree of freedom system the mass matrix is
2
312.9kg 0
0
0
0 12.71kgm
dsk
dsk
d
m
I
(
(
= =
(
(


M
From Appendix 2, Table A2.1, system 1,
( )
3 3 3 3 6
3 1.60 10 N m
T uu
k k EI a b a b = = + =
( )
2 2 2 2 5
3 2.75 10 N
C u
k k EI a b a b

= = =
( )
5
3 3.30 10 Nm
R
k k EI a b ab

= = + = and so
5
16.0N m 2.75N
10
2.75N 3.30Nm
(
=
(

K .
The eigenvalue problem can then be solved.

Part (b). For the 2 degree of freedom system mass coefficients are given in Appendix
2, Table A2.2, system 1. ( ) 17 35 4.76kg
uu
m A a b = + = ,
( )
2 2
13 35 0.168kgm
u
m A b a

= = ,
( )
3 3 2
2 105 0.052kgm m A a b

= + = .
Hence
2
4.76kg 0.168kgm
0.168kgm 0.052kgm
shft
(
=
(

M . Thus
shft dsk
M = + M M , giving
2
317.63kg 0.168kgm
0.168kgm 12.765kgm
(
=
(

M . Using K from (a), the eigenvalue can then be
solved.
Part (c) The MATLAB script Problem_05_11.m uses the rotordynamics package to
solve the 28 degree of freedom mode. The script also solves the two, 2 degree of
freedom models and also computes the errors in the 2 degree of freedom models,
compared with the 20 element model. The output of Problem_05_11.m is given
below:
Brg Type 1 Brg Type 1
N
o
d
e

1
N
o
d
e

2
N
o
d
e

3
N
o
d
e

4
N
o
d
e

5
N
o
d
e

6
N
o
d
e

7
N
o
d
e

8
N
o
d
e

9
N
o
d
e

1
0
N
o
d
e

1
1
N
o
d
e

1
2
N
o
d
e

1
3
N
o
d
e

1
4
N
o
d
e

1
5
N
o
d
e

1
6
N
o
d
e

1
7
N
o
d
e

1
8
N
o
d
e

1
9
N
o
d
e

2
0
N
o
d
e

2
1

20 El ement FE model
Di sk di am/ t hi ckness 1st Nat f r eq ( Hz) 2nd Nat f r eq ( Hz)
1 0. 3000/ 0. 0300 40. 3017 226. 6899
2 0. 4000/ 0. 0400 28. 0620 136. 5622
3 0. 5000/ 0. 0500 20. 5934 81. 8612
4 0. 6000/ 0. 0600 15. 8173 52. 6797
5 0. 7000/ 0. 0700 12. 5873 36. 1118
6 0. 8000/ 0. 0800 10. 2940 26. 0200
52


2dof model : Massl ess shaf t , no cent r al hol e i n di sk
Di sk 1st Nat f r eq ( Hz) 2nd Nat f r eq ( Hz) 1st %er r or 2nd %er r or
1 45. 7859 298. 3024 13. 6077 31. 5905
2 29. 6902 145. 5528 5. 8020 6. 5835
3 21. 1981 83. 5017 2. 9363 2. 0041
4 16. 0807 53. 0839 1. 6657 0. 7673
5 12. 7161 36. 2349 1. 0235 0. 3409
6 10. 3627 26. 0641 0. 6667 0. 1693

2dof model : I ncl udi ng shaf t mass, di sk wi t h a cent r al hol e
Di sk 1st Nat f r eq ( Hz) 2nd Nat f r eq ( Hz) 1st %er r or 2nd %er r or
1 40. 3289 242. 1845 0. 0676 6. 8351
2 28. 0665 137. 4105 0. 0158 0. 6212
3 20. 5944 81. 9180 0. 0046 0. 0694
4 15. 8175 52. 6853 0. 0016 0. 0106
5 12. 5874 36. 1125 0. 0006 0. 0021
6 10. 2941 26. 0201 0. 0003 0. 0005

2dof model : I ncl udi ng shaf t mass, no cent r al hol e i n di sk
Di sk 1st Nat f r eq ( Hz) 2nd Nat f r eq ( Hz) 1st %er r or 2nd %er r or
1 40. 0549 242. 0939 - 0. 6123 6. 7952
2 27. 9428 137. 3867 - 0. 4247 0. 6037
3 20. 5332 81. 9091 - 0. 2924 0. 0585
4 15. 7842 52. 6809 - 0. 2087 0. 0024
5 12. 5678 36. 1099 - 0. 1546 - 0. 0051
6 10. 2819 26. 0184 - 0. 1182 - 0. 0062

This output show that the results are substantially improved by accounting for the
mass of the shaft, particularly where the disk is small.

The question asks the reader to (a) model the system assuming a mass-less shaft and a
disk without a central hole, and (b) model the system accounting for the shaft mass
and modelling the disk with a central hole. By mistake the solutions in the text for
part (b) gives the results of modelling the system accounting for the shaft mass and
modelling the disk without a central hole. Clearly the mass at the centre of the disk is
accounted for twice. However, the above results show that this over-estimation of
mass makes very little difference to the results.


53

Chapter 6

Note. Solving Problems 6.10, 6.11 and 6.12 requires a finite element analysis that
allows shafts to be modeled and include gyroscopic effects, etc. Here we use the
rotordynamics software developed to accompany this book, but other appropriate
software can be used.

Problem 6.1
The equations of motion for free vibration of this rotor are developed in Problem 3.1
and are as follows:
2
2
0
0
d p
d p
I I L k
I I L k
+ + =
+ =

To obtain the critical speed we can solve these equations to determine the roots
(natural frequencies) and then equate the speed of rotation to the natural frequencies.
A slightly different but equivalent approach is to combine the pair of equations above
by letting j = . Then subtracting j the second equation from the first we
have
2
0
d p
I j I kL + = . We can now solve this equation by letting
0
j t
e

=
etc. to obtain
( )
2 2 2 2
0
0
d p
I j I kL + = and hence
( )
2
d p
L k I I = . This is
a forward whirling critical speed. If we let
0
j t
e

= in the previous equation we
obtain
( )
2
d p
L k I I = + . Given that 0.5m L = ,
2
0.6kgm
p
I = ,
2
10kgm
d
I = and
6
10 N m k = . Then
2 2 6
2 4 4
0.5 10
2.3585 10 and2.6596 10
10 0.6
d p
L k
I I

= = =

.
Hence the critical speeds are 153.57 and 163.08rad s or 1467rev min(backwards)
and 1557rev min (forwards).
To determine the response of the rotor to an unbalance we must add the out of balance
forcing term to the equations of motion. The out of balance force at the right end of
the rotor due to an out of balance mass
0
m at a radius is
2
0 ub
f m = . This
rotating force can be resolved into two components in the x and y directions, i.e
2
0
cos
x
f m t = and
2
0
sin
y
f m t = . (Here it is assumed that the out of
balance is in the x direction when 0 t = ). A force in the x direction causes a moment
2
0
cos
x
M Lf m L t

= = to act about the left end of the rotor in the direction . A

force in the y direction causes a moment
2
0
sin
y
M Lf m L t

= = to act about
the left end of the rotor in the direction . Thus the forced equations of motion are
2 2
0
2 2
0
sin
cos
d p
d p
I I L k m L t
I I L k m L t
+ + =
+ =

Again letting j = etc. we obtain
( )
2 2
0
cos sin
d p
I j I kL m L t j t + = + and hence
2 2
0
j t
d p
I j I kL m L e

+ =
54

Letting
0
j t
e

= we have
( )
2 2 2 2
0 0
j t j t
d p
I I kL e m L e

+ + = and hence
( ) { }
2
0
0
2 2
d p
m L
kL I I

=

. The orbit at the left end of the rotor
0
r is given by
0 0
r L = . Now = 1500 2 / 60rad/s and thus
( )
2
3
0
2 6 2
0.1 0.1 0.5
6.8295 10 rad
0.5 10 10 0.6


= =

and hence the radius of the orbit
at the right end is =
0
3.41mm L . The MATLAB script Problem_06_01.m repeats
these calculations and gives the following output.

Cr i t i cal speeds ar e 1466. 52 and 1557. 32 r ev/ mi n
Radi us of or bi t = 3. 4147 mm

Problem 6.2.
The equations of motion for free vibrations are
( )
( )
2 2
0 1 2
2 2
0 1 2
0
0
d p
d p
I I L k k k
I I L k k k
+ + + + =
+ + + =

To obtain the critical speed we can solve these equations to determine the roots
(natural frequencies) and then equate the speed of rotation to the natural frequencies.
A slightly different but equivalent approach is to combine this pair of equations by
letting j = . Then subtracting j the second equation from the first we have
( )
2 2
0 1 2
0
d p
I j I kL k k k + + + = .
We can now solve this equation by letting
0
j t
e

= etc. to obtain
( ) { }
2 2 2 2 2
0 1 2 0
0
d p
I j I k k k L + + + = and hence
( ) ( ) { }
2 2 2 2
2 1 0 0
0
d p
k L I I k L k L + + + = .
Thus
( ) ( ) { }
2 2 2 2
2 1 0
0
d p
k L I I k L k L + + + = . Here is a backward critical
speed. Solving this quadratic equation using the system data gives 179.48 = and
141.41 Similarly, if we let
0
j t
e

= , then we have
( ) ( ) { }
2 2 2 2
2 1 0
0
d p
k L I I k L k L + + = Here is a forward critical speed.
Solving this quadratic equation using the system data gives 193.06and 149.70 = .
Taking the positive roots we have 179.48and193.06rad s = or 1714 and 1844
rev min.
To determine the response we must add the out of balance moments to the system
thus:
( )
( )
2 2 2
0 1 2
2 2 2
0 1 2
sin
cos
d p
d p
I I L k k k M t
I I L k k k M t
+ + + + =
+ + + =

55

To determine the response, let j = . Then subtracting j the second equation
from the first we have
( )
( )
2 2 2
0 1 2
cos sin
d p
I j I k k k L M t j t + + + = +
and hence
( )
2 2 2
0 1 2
j t
d p
I j I k k k L M e

+ + + = .
Letting
0
j t
e

= we have
( ) ( ) { }
2 2 2 2
0 1 2 0 d p
I I k k k L M + + + = and
hence
( ) ( ) { }
2
0
2 2 2
0 1 2 d p
M
I I k k k L

=
+ + +
. The radius of the orbit at the
right end of the rotor is
0 0
r L = . At 1500rev min the bearing stiffness is
1.3096MN m. Thus
( )
( )
3 2 3 2
3
0
2 2 6 2 2 2
1
5 10 5 10
1.2922 10 rad
10 0.6 0.5 1.3096 10
d p
I L k I

= = =
+ + +
and hence the radius of the orbit at the right end is 0.646 mm.
The MATLAB script Problem_06_02.m repeats these calculations and gives the
following output. It also plots the Campbell diagram.

Backwar d cr i t i cal speed = 1713. 93 r ev/ mi n
For war d cr i t i cal speed = 1843. 57 r ev/ mi n
Radi us of or bi t = 0. 6461 mm

0 500 1000 1500 2000
0
5
10
15
20
25
30
Rotor speed rev/min
N
a
t
u
r
a
l

f
r
e
q
u
e
n
c
y

(
H
z
)


Problem 6.3.

From the solution to Problem 3.2 the natural frequencies are given by
( ) ( )
4 2 2 2 2 2 4
0
d p d x y x y
I I I L k k L k k + + + = . The critical speeds are obtained by
setting = so that
( ) ( )
4 2 2 2 2 4
0
d p d x y x y
I I I L k k L k k + + = which is a
56

quadratic in
2
. Hence the solutions for
2
are 24717.4 and 32990.4, thus the
critical speeds are 157.22 and 181.63rad s or 1501 and 1734rev min.
From Problem 3.2 we have the natural frequencies at 0, 3000 and 10,000 rev min.
This data provides 3 points for each frequency line in the Campbell diagram.
Furthermore, we have one extra point for each frequency line because we know that at
a critical speed = . The MATLAB script Problem_06_03.m computes the critical
speeds and plots the Campbell diagram, using 4 points to define each frequency line.
The points are joined by straight lines. The dotted line shows the exact frequency
lines. The difference between the exact and approximate lines is very small. The
output is as follows:

Cr i t i cal speeds ar e 1501. 3181 and 1734. 4623 r ev/ mi n


0 2000 4000 6000 8000 10000
22
24
26
28
30
32
34
N
a
t
u
r
a
l

f
r
e
q
u
e
n
c
i
e
s
,

H
z
Rotor speed rev/min


Problem 6.4.
The equations of motion for free vibration of this system are
0
0
0
0
uu u
vv v
d p v
d p u
mu k u k
mv k v k
I I k k v
I I k k u

+ + =
+ + =
+ + + =
+ + =

where the stiffnesses are obtained from Appendix 2, Table A2.1, System 4, and are
3 2
12 6 4
, ,
T uu vv C u u v v R
EI EI EI
k k k k k k k k k k k
a
a a

= = = = = = = = = = =
Writing the equations in matrix notation we have + + = Mq Gq Kq 0 where
( ) diag , , ,
d d
m m I I = M and

57

0 0 0 0
0 0 0 0
0 0 0
0 0 0
p
p
I
I
(
(
(
=
(
(

(

G ,
0 0
0 0
0 0
0 0
T C
T C
C R
C R
k k
k k
k k
k k
(
(

(
=
(
(
(

K . To solve the eigenvalue
problem we must let
0
j t
e

= q q and thus: ( )
2
0 0
j = M C q Kq . Solving this
eigenvalue problem with the problem data gives the following critical speed, The
resulting values of critical speed are 150.47, 175.32, 669.93 and 1002.1j rad/s - note
that the last critical speed is imaginary and hence there are only 3 critical speeds for
this 4 dof system. These are 1436.9, 1674.2, and 6397.4 rev/min. Since the supports
are isotropic only the critical speed at 1674.2 rev/min will be excited by unbalance.

+ + =
+ + =
+ + + =
+ + =

2
2
cos
sin
0
0
uu u
vv v
d p v
d p u
mu k u k m t
mv k v k m t
I I k k v
I I k k u

Alternatively to obtain the forward whirling critical speeds excited by unbalance, we
write the equations of motion in terms of complex co-ordinates r = u + jv and
= j , then

+ + =
+ + =

2
0
j t
uu u
d p u
mr k r k m e
I I j k k r

The critical speeds are then given by
( )

(
+
(
=
(
+

2
2
det 0
uu u
u d p
m k k
k I I k

which gives the quadratic in
2
thus
( ) ( ) ( )
+ + =
4 2
0
d p d p uu uu u
m I I I I k mk k k k with solutions

2
= 3.073810
4
or 1.004110
6
. Note the second solution is not real - that is
this critical speed cannot be attained. The response to the unbalance is given by
q = q
0
e
jt
where
( )

(
+


(
=
`
(
+
)

1
2
2
0
2
0
uu u
u d p
m k k
m
k I I k
q which gives

(
=

T 4 3
0
3.6058 10 , 1.2997 10 q , and hence the maximum orbit has radius
0.361mm.
From Figure 6.57, for r = 3/ 0.1= 30 we have

=
3
2.4 10 and hence
= =
2 2
73.77rad/s
n
. The torque required is =11.44Nm
p
I .
The steady state response for an angular misalignment of the disk is
( )
( )

(
+

(
=
`
(
+

)

1
2
0 2
2
0
uu u
d p
u d p
m k k
I I
k I I k
q
58

If

= 1
o
, then q
0
= 1.226410
3
,3.879810
3

, and hence the maximum in

degrees is 0.361/1.2264=0.285.
The MATLAB script Problem_06_04.m repeats these calculations and gives the
following output:

Cr i t i cal speeds = 1437, 1674 and 6397 r ev/ mi n
Radi us of or bi t due t o o/ b = 0. 3606 mm
Equi vql ent angul ar mi sal i gnment = 0. 2852 degr ee

r _bar = 30. 00
For t hi s val ue of r _bar obt ai n al pha_bar = 2. 4e- 3 f r omFi g 6. 57
Accel er at i on t hr o cr i t i cal speed = 73. 7715 r ad/ s^2
Tor que t o accel er at e t hr o cr i t i cal speed = 11. 4395 Nm


Problem 6.5.
Consider Equation (6.32). The stiffness coefficients for this system are given in
Appendix 2, Table A2.1, System 1. When the disk is at the mid-span, a b = and
hence 0
C
k = . Furthermore, there is no damping and thus 0
T C R
c c c = = = . Thus, for
pure angular misalignment of disk in complex coordinates
( )
2
0
T
j t
d p R d p
mr k r
I jI k I I e

+ =
+ =

The first equation gives no response. Letting
0
j t
e

= we obtain
( )
( )
2
0
2
d p
R d p
I I
k I I

=

. Now
12
126kNm/rad
R
EI
k k
L

= = = . Thus, when =

1 ,
3
0
3.3923 10

= . Thus the maximum value of

0
or
0
is
3
0 0
3.3923 10 180 0.1944

= = = .
The MATLAB script Problem_06_05.m repeats these calculations and gives the
following output

Response of a J ef f cot t r ot or at 6000 r ev/ mi n = 0. 1944 degr ees

Problem 6.6.
Let r = u + jv . Then we have the response, from Equation (6.47), for the bend and the
unbalance
( ) +

+ = +
0
2
0
j t
j t
b
mr kr kr e m a e where k =
48EI
L
3
= 381.70kN/m,
6.9360kg m = and
0
is the phase of the unbalance relative to the bend. With just the
bend the response at speed 1000rev min = , or 104.72rad s, is
r =
kr
b
m
2
+ k
= 0.6244mm
To balance the bend at speed , the right side (i.e. excitation) of the equation must be
zero, and hence we must have =

0
180 and m
0
=
kr
b

2
a
=
k 0.510
3

2
0.075
= 34.81g
At 1500 rev/min the response is then r =
kr
b
m
0
a
2
m
2
+ k
= 1.1330mm
59

The MATLAB script Problem_06_06.m repeats these calculations and gives the
following output:

Mass of shaf t = 5. 55 kg
Mass of di sk = 6. 94 kg
Response t o bent shaf t at 1000 r ev/ mi n = 0. 6244 mm
Requi r ed bal ance mass = 0. 2320 kg
Response at 1500 r ev/ mi n = 1. 1330 mm

Note that mass of the shaft is similar to the mass of the disk. Thus, neglecting the
mass of the shaft is likely to lead to large errors in computed natural frequency.

Note also that in the text book the solution contains an error, the balance mass is
incorrect.

Problem 6.7.
Using the stiffness formulae gives in Appendix 2, Table A2.4, System 6 with
2 a b L = = , gives
( )
3
48
24
b
uu
b
EIk
k
EI L k
=
+
. For this rotor,
8 4
4.3216 10 m I

= and
hence
3 2
8.6431 10 Nm EI = . Hence
5
3.9834 10 N m
uu
k = . If the mass of the shaft
is neglected then the first critical speed is
crit uu
k m = where m is the disk mass.
Evaluating, we have 240.41rad s
crit
= or 2296 rev/min. When an out of balance
acts,
2
0
cos
uu
mu k u m t + = Letting
0
cos u u t = ,
( )
2 2
0 0 uu
m k u m + = .
Thus the unbalance response is given by
( )
2
0
0
2
uu
m
u
m k

=
+
. At a rotor speed of
1500rev min or 157.08rad s,
0
1.62mm u = .
The force required to produce this deflection, f is given by
0 uu
k u . This force is
transmitted to the ground via the bearings, so the deflection at each bearing is 2 f .
Thus
( ) ( )
0
2 2
brg brg uu brg
u f k k u k = = .
Thus
5
6
3.9834 10
1.62mm 0.065mm
2 5 10
brg
u

= =


The stiffness of the system with short bearings,
uu
k , can also be derived by
recognising that the shaft stiffness
( )
uu
ss
k is effectively in series with the stiffnesses
at the bearings. Using Equation (2.9) we have
( ) 1 1 1 1
uu uu b b
ss
k k k k = + + . From
Appendix 2, Table A2.1, System 1, with 2 a b L = = , we have
( )
3
48
uu
ss
k EI L = .
Using this expression in Equation (2.9) gives the same expression for
uu
k as above.
The same line of reasoning can be applied to a shaft with long bearings. (i.e. clamped-
clamped supports). From Appendix 2, Table A2.1, System 2, with 2 a b L = = , we
have
( )
3
192
uu
cc
k EI L = . Substituting in Equation (2.9), ( )
( )
3
192
96
b
uu
cl
b
EIk
k
EI L k
=
+
.
60

Evaluating this system stiffness,
( )
6
1.423310 N m
uu
cl
k = . This has raised the
stiffness by a factor of 3.573 and hence the critical speed of 3.573 1.89 = . The new
critical speed is 4340 rev/min. The MATLAB script Problem_06_07.m gives the
following output:

Cr i t i cal speed = 2295. 7921 r ev/ mi n
Def l ect i on at t he di sk = 1. 6212 mm
Def l ect i on at t he bear i ng = 0. 064579 mm
St i f f ness i ncr eased by a f act or of 3. 573
Fi r st cr i t i cal r ai sed t o 4339. 6013 r ev/ mi n

Note that the solution to this problem given in the text book contains errors.

Problem 6.8.
We will begin by plotting a Campbell diagram. This system is described by Equation
(3.84). In matrix notation, the equations of motion for the system are
0 0 0 0
0 0 0 0 0 0
0 0 0 0
0 0 0 0 0 0
0 0 0
0 0 0 0 0 0
0 0 0 0 0 0 0 0 0
(
( (
(
( (

(
( (
+ + =
` ` ` `
(
( (

(
( (


( (
( ) ) ) )





T C
T C
p
d C R
d C R p
m k k u u u
m k k v v v
I
I k k
I k k I

or + + = Mq Gq Kq 0 . Using the stiffness formulae gives in Appendix 2, Table
A2.1, System 5, we have
( )
( )
3
12 3
4 3
T uu vv
EI a b
k k k
b a b
+
= = =
+
,
( )
( )
12
4 3
R
EI a b
k k k
b a b

+
= = =
+
and
( )
( )
2
6 2 3
4 3
C u v
EI a b
k k k
b a b

+
= = =
+
. Letting
0
=
st
e q q , then
2
0 0 0
s s + + = Mq Gq Kq 0 . This leads to an eigenvalue problem of the form
s
s s
( (
=
` `
( (

) )
G M q K 0 q
M 0 q 0 M q
. This eigenvalue problem can be solved for a
particular value of to give the eigenvalues s. From s we can obtain the system
natural frequencies and these frequencies can be plotted against rotor speed to create a
Campbell diagram, see below. Examining this diagram shows immediately why there
are two backward critical speeds but only one forward critical speed. The highest
natural frequency (a forward whirl) is increasing with rotor speed (due to the
gyroscopic effect) faster than the increase in the excitation frequency with rotor
speed, so that the two can never be equal, however high the system rotation speed.
We now compute the critical speeds directly. From + + = Mq Gq Kq 0 , we let
0
j t
e

= q q to give
2 2
0 0 0
j + + = Mq Gq Kq 0 . This leads to the eigenvalue
problem
( )
2
0 0
j = M G q Kq where is a critical speed. Solving this equation
numerically gives the 4 critical speeds, 3 real and one imaginary, see below.
Alternatively, we can go back to the equations of motion and let r u jv = + and
j = . Adding j the second equation to the first and subtracting j the third
61

equation from the fourth, we have
0
0
T C
d p C R
mr k r k
I j I k r k
+ + =
+ + =

.
Letting
0
j t
r r e

= etc gives
( )
( )
2
0 0
2 2
0 0
0
0
T C
R d p C
k m r k
k I I k r
+ =
+ + =
To solve this pair of
equations we must determine the roots of
( ) ( ) { } ( )
4 2 2
0
d p T d p R T R C
m I I k I I mk k k k + + =
The roots of this equation are the forward whirling critical speeds. To determine the
backward whirling critical speeds we let
0
j t
r r e

= and this leads to
( ) ( ) { } ( )
4 2 2
0
d p T d p R T R C
m I I k I I mk k k k + + + + =
Both of these equation are quadratics in
2
and if the value of
2
is positive, then
is a real value
The MATLAB script Problem_06_08.m computes the critical speeds and plots the
Campbell diagram. The output is shown below.

Sol vi ng ei genval ue pr obl em
Cr i t i cal speeds r ev/ mi n 683+ 0j
Cr i t i cal speeds r ev/ mi n 1537+ 0j
Cr i t i cal speeds r ev/ mi n 0+2562j
Cr i t i cal speeds r ev/ mi n 3183+ 0j

For war d cr i t i cal speeds ( r ev/ mi n)
1537

Backwar d cr i t i cal speeds ( r ev/ mi n)
3183
683

Note that one of the four critical speeds is imaginary so the solution exists
mathematically but not in reality. Note also that in the text book solutions the forward
critical speed is incorrect.
0 1000 2000 3000 4000 5000 6000
0
50
100
150
Rotor speed (rev/min)
N
a
t

f
r
e
q
u
e
n
c
y

(
H
z
)

62

Problem 6.9
On Figure 6.62 we must draw the 1x line, i.e. from the origin to the point (12000
rev/min, 200 Hz), a 3x line, i.e. from the origin to the point (12000 rev/min, 600 Hz)
and a line at a constant speed of 10000 rev/min. Where these lines cuts the natural
frequency lines we can obtain the following data (approximately).
(1) From the left hand diagram of Figure 6.62 (which we can identify as the rotor
on isotropic bearings because when the rotor is stationary all frequencies are
repeated because the frequency in the x and y directions is the same) we have
(a) 2,200 and 5,000 rev/min due to synchronous unbalance. (b) 720 and 1,200
rev/min due to the 3x force. (c) At 10,000 rev/min, 22, 38, 119, 244 and 573
Hz.
(2) From the right hand diagram of Figure 6.62 (which we can identify as the
rotor on anisotropic bearings because when the rotor is stationary the
frequencies are distinct because the frequencies in the x and y directions are
different due to the bearing stiffnesses) we have (a) 2,200, 2,600, 2,900 and
5,600 rev/min due to synchronous unbalance. (b) 740, 900, 1,000, 1,400,
5,800 rev/min due to the 3x force. (c) At 10,000 rev/min, 26, 38, 46, 124, 244,
and 573Hz.

Note that in the case of the rotor supported by isotropic bearings only the forward
critical speeds are excited. In the case of the rotor supported by anisotropic bearings
forward and backward critical speeds can be excited.

Problem 6.10.
Modeling this problem requires finite element analysis (FEA) and can only be solved
using appropriate FEA software. Here MATLAB script Problem_06_10.m makes use
the Rotordynamics Software package to model and analyze the system. The diagrams
below, generated by the script, shows mode shapes, the Campbell diagram and
response plots for the three sets of out of balance.

Fi r st f our cr i t i cal speeds 1577 1746 3531 4027 r ev/ mi n

Only the forward whirling critical speeds will be excited (i.e. 1,746 and 4,027
rev/min).
0 1000 2000 3000 4000 5000
0
10
20
30
40
50
60
70
80
90
Rotor spin speed (rev/min)
U
n
d
a
m
p
e
d

n
a
t
u
r
a
l

f
r
e
q
u
e
n
c
i
e
s

(
H
z
)


63

Nat Freq =26.7733Hz Nat Freq =61.5787Hz

Mode shapes
0 500 1000 1500 2000 2500 3000 3500 4000 4500
10
-5
Rotor spin speed (rev/min)
R
e
s
p
o
n
s
e

m
a
g
n
i
t
u
d
e

(
m
)


Node 1, x
Node 5, x
Node 7, x
0 500 1000 1500 2000 2500 3000 3500 4000 4500
-200
-100
0
100
200
Rotor spin speed (rev/min)
P
h
a
s
e

(
d
e
g
r
e
e
s
)
o/b case (a)


Node 1, x
Node 5, x
Node 7, x

Note this mass distribution only weakly excites the first critical speed (above).
0 500 1000 1500 2000 2500 3000 3500 4000 4500
10
-5
Rotor spin speed (rev/min)
R
e
s
p
o
n
s
e

m
a
g
n
i
t
u
d
e

(
m
)


Node 1, x
Node 5, x
Node 7, x
0 500 1000 1500 2000 2500 3000 3500 4000 4500
-200
-100
0
100
200
Rotor spin speed (rev/min)
P
h
a
s
e

(
d
e
g
r
e
e
s
)
o/b case (b)


Node 1, x
Node 5, x
Node 7, x

Note this mass distribution only weakly excites the second critical speed (above).

64

0 500 1000 1500 2000 2500 3000 3500 4000 4500
10
-5
Rotor spin speed (rev/min)
R
e
s
p
o
n
s
e

m
a
g
n
i
t
u
d
e

(
m
)


Node 1, x
Node 5, x
Node 7, x
0 500 1000 1500 2000 2500 3000 3500 4000 4500
-200
-100
0
100
200
Rotor spin speed (rev/min)
P
h
a
s
e

(
d
e
g
r
e
e
s
)
o/b case (c)


Node 1, x
Node 5, x
Node 7, x

Note this mass distribution excites both the first and second critical speed (above).


Problem 6.11
This problem is based on Problem 5.9 and since it requires a finite element analysis
(FEA) it can only be solved using appropriate FEA software. In order to compute the
bearing properties we require the static load acting on each bearing. From the data of
Table 5.11, we can compute the volume and hence the mass of the shaft. Similarly,
the question gives the details of the disks and the volume and hence the mass can be
determined. Thus we have 101.914kg
sft
m = , 34.184kg
dsk
m = and adding these
together the total mass of the rotor is 136.097kg. Thus total force on the bearings is
136.097 13.35kN F g = = . This can then be divided in ratios 1:1:1. The MATLAB
script Problem_06_11.m makes use the Rotordynamics Software package to model
the system with 28 Timoshenko elements and analyze the system.
The system is initially modeled assuming that the outer bearings are placed at the
extremities of the rotor. (This is the model used to generate Figures 6.63 and 6.64).
The MATLAB script gives the following output:

I sot r opi c br gs. Cr i t i cal speeds ( r ev/ mi n)
1015. 60 1021. 15 2037. 86 2053. 11

Ani sot r opi c br gs. Cr i t i cal speeds ( r ev/ mi n)
1006. 85 1019. 01 1966. 53 2046. 06

Hydr odynami c br g, f or ce = 1: 1: 1. Cr i t i cal speeds ( r ev/ mi n)
I ni t i al val ues
600. 0 700. 0 1000. 0 1050. 0 1800. 0 2210. 0 2350. 0 2720. 0
Fi nal val ues
600. 4 703. 0 986. 8 1055. 7 1825. 1 2201. 6 2362. 7 2716. 8

In the above computation, it is difficult to accurately locate the critical speeds at 2,201
and 2,716 rev/min, even when the initial values of critical speed used in the iteration
are close to these values. The MATLAB script Problem_06_11e.m models the same
65

system, except that the bearings are now moved to the same position as in Problem
5.9 and gives the following output:

I sot r opi c br gs. Cr i t i cal speeds ( r ev/ mi n)
1140. 69 1148. 49 2312. 75 2335. 34

Ani sot r opi c br gs. Cr i t i cal speeds ( r ev/ mi n)
1128. 63 1145. 48 2218. 74 2324. 96

Hydr odynami c br g, f or ce = 1: 1: 1. Cr i t i cal speeds ( r ev/ mi n)
I ni t i al val ues
820. 0 1020. 0 1105. 0 1195. 0 2025. 0 2755. 0
Fi nal val ues
819. 5 1018. 8 1107. 7 1196. 0 2023. 0 2754. 5

The relatively small change in the position of the outer bearings make some quite
significant change to the critical speeds, particularly in the case of the hydrodynamic
bearings.


Problem 6.12
Modeling this problem requires finite element analysis (FEA) and can only be solved
using appropriate FEA software. Here MATLAB script Problem_06_12.m makes use
the Rotordynamics Software package to model and analyze the system. The diagrams
below, generated by the script, show the Campbell diagram and response plots for the
of out of balance and the bent shaft.

Fi r st f our cr i t i cal speeds 989 1010 3795 4180 r ev/ mi n

The graphs below show the response due to out of balance and bend. In this problem
the responses at the first critical speed is similar.
0 500 1000 1500 2000 2500 3000 3500 4000 4500
10
-5
Rotor spin speed (rev/min)
R
e
s
p
o
n
s
e

m
a
g
n
i
t
u
d
e

(
m
)


Node 6, x
Node 11, x
0 500 1000 1500 2000 2500 3000 3500 4000 4500
-200
-100
0
100
200
Rotor spin speed (rev/min)
P
h
a
s
e

(
d
e
g
r
e
e
s
)
out of balance


Node 6, x
Node 11, x

66


0 500 1000 1500 2000 2500 3000 3500 4000 4500
10
-4
10
-2
Rotor spin speed (rev/min)
R
e
s
p
o
n
s
e

m
a
g
n
i
t
u
d
e

(
m
)


Node 6, x
Node 11, x
0 500 1000 1500 2000 2500 3000 3500 4000 4500
-200
-100
0
100
200
Rotor spin speed (rev/min)
P
h
a
s
e

(
d
e
g
r
e
e
s
)
bent shaft


Node 6, x
Node 11, x



Problem 6.13.
We begin with the proof of the expression for the solution of ( )
i
p t , starting with
Duhamels integral given in the question. Thus
( ) ( )
( ) ( )
0
0 0
i i
t t
s t s t
i i g i
p t f u e d f u e e d

= =

.
Thus ( )
( )
0 0
0 0
i i i
t t
s t s t s
i i i
p t f u e d f u e e d
+
= =

.
Hence
( )
( )
( )
( )
0
0
0
1
1
i
i i i
t
s t
s s t s t i
i i
i i
f u e
p t f u e e e
s s

(
(
= =
(
(

+


Finally ( )
( )
( )
0
i
s t t i
i
i
f u
p t e e
s

=
+

From the equations of motion developed for Problem 3.1 we have to add the forcing
term. The forced equations of motion are (with k
x
= k
y
= k )
( )
2
2
0
d p
d p g
I I c L k
I I c L k kLu t
+ + + =
+ + =

. where
2
c kL .
Or in matrix notation
( )
2
2
0 0 0
0
0
p
d
g d p
c I
I L k
kLu t I I c
L k
( (
(
+ + = ( ( ` ` ` `
(

) ( ( ) ) )



,
i.e.
( )
g
u t + + = Mq Gq Kq Q where
0
kL

=
`
)
Q .
67

Here
10 0
0 10
(
=
(

M ,
6
0.25 0
10
0 0.25
(
=
(

K ,
20 188.5
188.5 20
(
=
(

C ,
6
0
10
0.5

=
`
)
Q .
Writing these equations in state space form gives
( )
g
d
u t
dt
( (
+ =
` ` `
( (

) ) )
G M q K 0 q Q
M 0 q 0 M q 0
. Let
(
=
(

G M
A
M 0
,
(
=
(

K 0
B
0 M
and

=
`
)
q
u
q
. Setting the right side of the above equations to zero,
and letting
st
R
e = u u leads to the eigenvalue problem
R R
s = Au Bu . Solving this
eigenvalue problem (with 4 degrees of freedom) provides the eigenvalues or roots
i
s
and the eigenvectors
Ri
u , 1, ,4 i = and the subscript R denotes that this is the right
eigenvector. Alternatively, we have
L L
s = u A u B or
T T
L L
s = A u B u . Solving this
eigenvalue problem provides identical eigenvalues, but the left eigenvectors,
Li
u ,
1, ,4 i = . (See Section 5.8). Note that
T
= B B but
T
A A . Consider the equation
for the forced system,
( )
g
u t + = Au Bu g where

=
`
)
Q
g
0
. In this example
6
0 0.5 10 0 0
T
(
=

g .
Let
| |
1 2 3 4 R R R R R
= U u u u u ,
| |
1 2 3 4 L L L L L
= U u u u u and note that
R
U
and
L
U are square matrices. In this example
0.0002 0.0057 0.0002 0.0057 0.0062 0.0004 0.0062 0.0004
0.0057 0.0002 0.0057 0.0002 0.0004 0.0062 0.0004 0.0062
0.9534 0.0466 0.9534 0.0466 0.0728 0.9272
R
j j j j
j j j j
j j j
+ +
+ +
=
+
U
0.0728 0.9272
0.0466 0.9534 0.0466 0.9534 0.9272 0.0728 0.9272 0.0728
j
j j j j
(
(
(
( +
(
+ +



0.0001 0.0058 0.0001 0.0058 0.0061 0.0006 0.0061 0.0006
0.0058 0.0001 0.0058 0.0001 0.0006 0.0061 0.0006 0.0061
0.9773 0.0227 0.9773 0.0227 0.0913
L
j j j j
j j j j
j j
+ +
+ +
=
+
U
0.9087 0.0913 0.9087
0.0227 0.9773 0.0227 0.9773 0.9087 0.0913 0.9087 0.0913
j j
j j j j
(
(
(
( +
(
+ +

Note that the modes shapes are not unique, they can be scaled in amplitude and a
phase angle can be added to each element of the vector.
Let ( ) ( )
R
t t = u U p and pre-multiply the equation of motion by
T
L
U . Thus we have
( )
T T T
L R L R L g
u t + = U AU p U BU p U g . Letting
T *
L R
= U AU A and
T *
L R
= U BU B .
*
A
and
*
B are diagonal matrices. In this system,
( )
*
0.0126 0.2095
0.0126 0.2095
diag
0.0399 0.2392
0.0399 0.2392
j
j
j
j
(
(
+
(
=
(
(
+

A .
Furthermore
* *
= B A s (where s is a diagonal matrix of the roots) so that
( )
( ) ( )
1
* T
L g g
u t u t

= = p sp A U g f and hence
( )
1
* T
L

= f A U g . In this example,
68

3
2.9121 0.0493
2.9121 0.0493
10
0.2870 3.0520
0.2870 3.0520
T
L
j
j
j
j
+ (
(

(
=
(
(
+

U g and
4
0.0595 1.3862
0.0595 1.3862
10
1.2608 0.0905
1.2608 0.0905
j
j
j
j
+ (
(

(
=
(
(
+

f .
This set of differential equations in p comprise four uncoupled equations of the form
( )
i i i i g
p s p f u t = where 1, ,4 i = and
( )
{ }
T *
i L i
i
f A = U g . Each equation can be
solved using for the function
( )
g
u t using Duhamels integral to obtain ( )
i
p t . Given
( ) t p we can determine ( ) t u from ( ) ( )
R
t t = u U p . From ( ) t u we can derive ( ) t
and ( ) t . Hence we can compute the displacements at the flexible bearing since
( ) ( ) u t L t = and ( ) ( ) v t L t = . Clearly, to solve ( )
( )
( )
0
i
s t t i
i
i
f u
p t e e
s

=
+
to
obtain at large number of values in the time series requires a computer.
The MATLAB script Problem_06_13.m solves this problem and gives the following
numeric and graphical output:

Root = - 1. 0595+/ - j 167. 8162
Root = - 0. 9405+/ - j 148. 9666
0 1 2 3
-0.5
0
0.5
1
1.5
2
R
e
s
p
o
n
s
e

i
n

d
i
r
e
c
t
i
o
n

o
f

g
r
o
u
n
d

d
i
s
t
u
r
b
a
n
c
e

m
m
Time, s
0 1 2 3
-1
-0.5
0
0.5
1
R
e
s
p
o
n
s
e

o
r
t
h
o
g
o
n
a
l

t
o

g
r
o
u
n
d

d
i
s
t
u
r
b
a
n
c
e

m
m
Time, s


-0.2 -0.1 0 0.1 0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
d
i
s
p
l
a
c
e
m
e
n
t

m
m
displacement mm
Orbit during the time 2 to 3 sec
-1 -0.5 0 0.5 1
-0.5
0
0.5
1
1.5
2
d
i
s
p
l
a
c
e
m
e
n
t

m
m
displacement mm
Orbit during the time 0 to 1 sec


Removing the damping in the system obviously removes the decay in the time series.
69

Chapter 7

Note. Solving Problems 7.10 and 7.11 requires a finite element analysis that allows
shafts to be modeled and include gyroscopic effects, etc. Here we use the
rotordynamics software developed to accompany this book, but other appropriate
software can be used.

Problem 7.1
For a J effcott rotor, ignoring the mass of the shaft, the only mass is the disc and the
stiffness is the stiffness of the rotor at mid-span. Thus in thex and y-directions,
0
uu
mu k u + = and 0
vv
mv k v + = . In rotation about the x and y-directions,
0
d p
I I k

+ + =

and 0
d p
I I k

+ =

. These equations are based on

those given in Equation (3.83). Letting,
uu vv T
k k k = = and
R
k k k

= = . Thus in
coordinates fixed in space we have Mq + Gq +Kq = 0 where in this case
0 0 0
0 0 0
0 0 0
0 0 0
d
d
m
m
I
I
(
(
(
=
(
(
(

M ,
0 0 0 0
0 0 0 0
0 0 0
0 0 0
p
p
I
I
(
(
(
=
(
(

(

G ,
0 0 0
0 0 0
0 0 0
0 0 0
T
T
R
R
k
k
k
k
(
(
(
=
(
(
(

K ,
u
v



=
`

)
q .
Here,
3 6
48 1.1685 10 N m
T
k EI L = = and
5
12 1.4314 10 Nm
R
k EI L = =
Letting
0
st
e = q q , (and
0
st
s e = q q ,
2
0
st
s e = q q ) gives
( )
2
0
s s = M + G +K q 0 . We
can either solve the resulting eigenvalue problem with 4 degrees of freedom or, obtain
( ) ( )
2 2
0 0
0, 0 + = + =
st st
T i T i
ms k u e ms k v e ,
2
0
2
0
0
0
(
+

(
=
` `
(
) ) +

d R p
st
p d R
I s k I s
e
I s I s k
.
From the first pair of equations,
T
s k m = (twice) and from the second set of
equations
2
2
0
d R p
p d R
I s k I s
I s I s k
+
=
+
. Thus
( ) ( )
2
2
2
0
d R p
I s k I s + + = . Rearranging this
equation gives
2
d R p
I s k j I s + = or
2
0
d p R
I s j I s k + = . We can solve this
quadratic equation to find the roots. The MATLAB script Problem_07_01.m carries
out these calculations. It also solves the eigenvalue for the complete system.

Converting the equations of motion from the fixed coordinates to coordinates rotating
with the shaft using the transformation given in Equations (7.5) (7.10) we have
( ) ( ) { }
2
1 2 1
+ + + Mq + M G q + M G K q = 0

where = M M

, = G G

, = K K

,
2
= M M

,
1
0 2 0 0
2 0 0 0
0 0 0 2
0 0 2 0
d
d
m
m
I
I
(
(
(
=
(
(
(

M

and
1
0 0 0 0
0 0 0 0
0 0 0
0 0 0
p
p
I
I
(
(
(
=
(
(
(

G

70

Letting
0
st
e = q q , we have
0
st
s e = q q

and
2
0
st
s e = q q

we can either solve the

resulting eigenvalue problem with 4 degrees of freedom or, obtain a pair of 2 degree
eigenvalue problems thus:
2 2
2 2
2
0
2
T
T
ms m k m s
m s ms m k
+
=
+
and

( ) ( )
( ) ( )
2 2
2 2
2
0
2
d p d R d p
d p d p d R
I s I I k I I s
I I s I s I I k
+ +
=
+ +

Thus we have
2 2
2 0 + =
T
ms j m s k m and
( ) ( )
2 2
2 0 + + =
d d p R p d
I s j I I s k I I . We
can solve these quadratic equations to find the roots. The MATLAB script
Problem_07_01.m gives the user the choice of solving either the characteristic
equations or the eigenvalue problem to determine the system natural frequencies. Of
course, both methods give the same numeric values for the frequencies which are as
follows

Sol ut i on of char act er i st i c equat i ons
I n f i xed coor di nat es at 3000 r ev/ mi n
Nat ur al f r equency 1 = 16. 4507 Hz
Nat ur al f r equency 2 = 31. 4102 Hz
Nat ur al f r equency 3 = 31. 4102 Hz
Nat ur al f r equency 4 = 110. 2007 Hz
I n r ot at i ng coor di nat es at 3000 r ev/ mi n
Nat ur al f r equency 1 = 18. 5898 Hz
Nat ur al f r equency 2 = 60. 2007 Hz
Nat ur al f r equency 3 = 66. 4507 Hz
Nat ur al f r equency 4 = 81. 4102 Hz

Problem 7.2
In rotating coordinates we have
( ) ( ) { }
2
1 2 1
+ + + Mq + M G q + M G K q = 0


where from Equations (7.80), (7.83) and (7.84)
0 0 0
0 0 0
0 0 0
0 0 0
(
(
(
=
(
(
(

dx
dy
m
m
I
I
M ,
( )
( )
1
0 2 0 0
2 0 0 0
0 0 0
0 0 0
(
(
(
=
(
+
(
(
+
(

dx dy
dx dy
m
m
I I
I I
M and
2
0 0 0
0 0 0
0 0 0
0 0 0
(
(

(
=
(
(

dy
dx
m
m
I
I
M , From (7.85)
1
0 0 0 0
0 0 0 0
0 0 0
0 0 0
p
p
I
I
(
(
(
=
(
(
(

G

and
0 0 0 0
0 0 0 0
0 0 0
0 0 0
p
p
I
I
(
(
(
=
(
(

(

G

.
T
0 0
0 0
0 0
0 0
T C
T C
C R
C R
k k
k k
k k
k k
(
(

(
= = =
(
(
(

K T KT K

71

For a simply supported shaft, from Appendix 2, Table A2.1, system 1 we have
( )
3 3
3 3
3
T uu vv
EI a b
k k k
a b
+
= = = ,
( )
2 2
2 2
3
C u v
EI a b
k k k
a b


= = = and
( ) 3
R
EI a b
k k k
ab

+
= = = , where 0.2m a = and 0.5m b = . To solve the equations
of motion we must rearrange them as a set of state space equation thus
( ) ( )
2
1 2 1
d
dt
( (
+ + +

= ( (
` `
( ( ) )

M G M M G K 0 q q
q q
M 0 0 M





Letting
0
st
e = q q , (and
0
st
s e = q q

etc.) leads to an 8 8 eigenvalue problem in s. The

MATLAB script Problem_07_02.m formulates and solves this eigenvalue problem
to determine the eigenvalues. The script also checks that the real part of the
eigenvalues are positive implying an unstable system. The output of the script is as
follows. Note that the system is unstable under certain conditions.

Root s at a r ot or speed of 2400 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
- 0. 0000 14. 4102 - 0. 0000 - 14. 4102 0
- 0. 0000 365. 3374 - 0. 0000 - 365. 3374 0
0. 0000 484. 6082 0. 0000 - 484. 6082 0
0. 0000 662. 7558 0. 0000 - 662. 7558 0
Root s at a r ot or speed of 2600 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
- 7. 5741 0. 0000 7. 5741 0. 0000 1
- 0. 0000 380. 7460 - 0. 0000 - 380. 7460 0
0. 0000 492. 6906 0. 0000 - 492. 6906 0
- 0. 0000 678. 9032 - 0. 0000 - 678. 9032 0
Root s at a r ot or speed of 2800 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
0. 0000 17. 7441 0. 0000 - 17. 7441 0
- 0. 0000 396. 4589 - 0. 0000 - 396. 4589 0
0. 0000 501. 3517 0. 0000 - 501. 3517 0
0. 0000 695. 5383 0. 0000 - 695. 5383 0

Problem 7.3
In rotating coordinates,
( ) ( ) { }
2
1 2 1
+ + + Mq + M G q + M G K q = 0

where from
Equations (7.80), (7.83) and (7.84)
0 0 0
0 0 0
0 0 0
0 0 0
d
d
m
m
I
I
(
(
(
=
(
(
(

M

,
1
0 2 0 0
2 0 0 0
0 0 0 2
0 0 2 0
d
d
m
m
I
I
(
(
(
=
(
(
(

M

and
2
0 0 0
0 0 0
0 0 0
0 0 0
(
(

(
=
(
(

d
d
m
m
I
I
M . From (7.85)
1
0 0 0 0
0 0 0 0
0 0 0
0 0 0
p
p
I
I
(
(
(
=
(
(
(

G

and
72

0 0 0 0
0 0 0 0
0 0 0
0 0 0
p
p
I
I
(
(
(
=
(
(

(

G and
T
0 0 0
0 0 0
0 0 0
0 0 0

(
(
(
= = =
(
(
(

uu
vv
k
k
k
k
K T KT K
For a simply supported shaft, from Appendix 2, Table A2.1, system 1 we have
3
48
xx
uu
EI
k
L
= ,
3
48
yy
vv
EI
k
L
= ,
12
xx
EI
k
L

= and
12
yy
EI
k
L

= where, for an
elliptical cross section
3
64
xx y x
I d d = etc. Since the first pair of equations is
uncoupled from the second pair we can solve them separately. Thus, letting
0
st
u u e =
etc., we have
2 2
2 2
2 0
0
2
uu st
vv
ms k m m s u
e
v
m s ms k m
(
+

= (
` `
( ) ) +

.
Setting the determinant of the coefficients to zero gives
( )
{ } ( )( )
2 4 2 2 2 2 2
2 0
uu vv uu vv
m s m k k m s k m k m + + + + = . This is a quadratic
in
2
s and can readily be solved. From the second pair of equations and letting
0
st
e = etc, we have
( ) ( )
( ) ( )
2 2
2 2
2
0
0
2
d d p d p
st
d p d d p
I s k I I I I s
u
e
v
I I s I s k I I

(
+

(
=
` `
(
) ) +

.
Setting the determinant of the coefficients to zero gives
( ) ( ) { }
( ) ( ) ( ) ( )
2 4 2 2 2 2
2 2
2 2
0
d d d d p p
d p d p
I s I k k I I I I s
k I I k I I


+ + + + +
=

This is a quadratic in
2
s and can readily be solved. Alternatively the 4 equations of
motion can be solved by forming the state equations and solving the 8 8 eigenvalue
problem in s (see Problem 7.2). The MATLAB script Problem_07_03.m gives the
user the choice of solving either the characteristic equations or the eigenvalue
problem to determine the system natural frequencies. The script also checks that the
real part of the roots or eigenvalues are positive implying an unstable system. The
output of the script is as follows. Note that the system is unstable under certain
conditions.

Sol ut i on of t he char act er i st i c equat i ons
Root s at a r ot or speed of 1900 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
0. 0000 9. 4646 0. 0000 - 9. 4646 0
0. 0000 326. 4443 0. 0000 - 326. 4443 0
0. 0000 355. 7124 0. 0000 - 355. 7124 0
0. 0000 409. 7592 0. 0000 - 409. 7592 0
Root s at a r ot or speed of 2000 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
6. 6870 0. 0000 - 6. 6870 0. 0000 1
0. 0000 331. 4147 0. 0000 - 331. 4147 0
0. 0000 361. 6804 0. 0000 - 361. 6804 0
0. 0000 420. 2268 0. 0000 - 420. 2268 0
73

Root s at a r ot or speed of 2100 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
0. 0000 6. 3311 0. 0000 - 6. 3311 0
0. 0000 336. 5616 0. 0000 - 336. 5616 0
0. 0000 367. 8495 0. 0000 - 367. 8495 0
0. 0000 430. 6949 0. 0000 - 430. 6949 0

Problem 7.4
In rotating coordinates, the internal shaft viscous damping is given by
0 0 0
0 0 0
0 0 0
0 0 0
iT
iT
d
iR
iR
c
c
c
c
(
(
(
= =
(
(
(

f Cq

. Now in rotating coordinates,
( )
T T
d d i i
= = = + f Tf TC q TC T q T q

. Thus
T
i i i
= = C TC T C

and
T
1
0 0 0
0 0 0
0 0 0
0 0 0
iT
iT
i i
iR
iR
c
c
c
c
(
(

(
= =
(
(

(

C TC T

. Thus, in fixed coordinates, the equations of
motion are ( ) ( )
1 e i i
+ + + + + = Mq C C G q C K q 0 where
0 0 0
0 0 0
0 0 0
0 0 0
d
d
m
m
I
I
(
(
(
=
(
(
(

M ,
0 0 0 0
0 0 0 0
0 0 0
0 0 0
p
p
I
I
(
(
(
=
(
(

(

G ,
0 0 0
0 0 0
0 0 0
0 0 0
T
T
R
R
k
k
k
k
(
(
(
=
(
(
(

K and
0 0 0
0 0 0
0 0 0 0
0 0 0 0
e
e
e
c
c
(
(
(
=
(
(

C
When the system is at rest, ( ) + + + =
e i
Mq C C q Kq 0. Considering the first pair of
equations we have
( )
( )
0
0
+ + + =
+ + + =


e iT T
e iT T
mu c c u k u
mv c c v k v
. Hence, for these separate single degree of freedom
equations, we have
2
e iT
T
c c
mk
+
= (twice). Considering the second pair of equations,
0
0
+ + =
+ + =


d iR R
d iR R
I c k
I c k
. Hence, , for these separate single degree of freedom equations,
we have
2
iR
d R
c
I k
= (twice). Consider now the set of system equation. Letting
0
=
st
e q q , we have ( ) ( )
2
1 0
(
+ + + + + =

st
e i i
s s e M C C G C K q 0.Because the
first pair and the second pair of equations are uncouple from each other we can
proceed as follows:
74

( )
( )
2
2
0
e iT T iT
iT e iT T
ms c c s k c
c ms c c s k
+ + +
=
+ + +
and thus
( )
{ }
2
2 2 2
0 + + + + =
e iT T iT
ms c c s k c . Hence ( )
2
0 + + + =
e iT T iT
ms c c s k jc .
Similarly
( )
( )
2
2
0
d iR R iR p
iR p d iR R
I s c s k c I
c I I s c s k
+ + +
=
+ + +
and thus
( ) ( )
2
2
2 2
0 + + + + =
d iR R iR p
I s c s k c I . Hence
( )
2
0 + + + =
d iR R iR p
I s c s k j c I .
We can solve these two quadratic equations to obtain the 4 system roots.
Alternatively we can the 4 equations together, (see Problem 7.2). The MATLAB
script Problem_07_04.m gives the user the choice of solving either the characteristic
equations or the eigenvalue problem to determine the system natural frequencies. The
script also checks that the real part of the roots or eigenvalues are positive implying
an unstable system. The output of the script is as follows. Note that the system is
unstable under certain conditions.

zet a = 0. 005067 ( t wi ce) and 0. 0074759 ( t wi ce)
Sol ut i on of t he char act er i st i c equat i ons
Root s at a r ot or speed of 2200 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
- 2. 0838 127. 8392 - 2. 0838 - 127. 8392 0
- 0. 0272 197. 3557 - 0. 0272 - 197. 3557 0
- 1. 9728 197. 3557 - 1. 9728 - 197. 3557 0
- 1. 9162 559. 8082 - 1. 9162 - 559. 8082 0
Root s at a r ot or speed of 2300 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
- 2. 0860 124. 2722 - 2. 0860 - 124. 2722 0
0. 0170 197. 3560 0. 0170 - 197. 3560 1
- 2. 0170 197. 3560 - 2. 0170 - 197. 3560 0
- 1. 9140 575. 8762 - 1. 9140 - 575. 8762 0

Problem 7.5
In rotating coordinates ( ) ( )
{ }
2
1 2 1
+ + + = Mq + M G q + M G K q Q

. An out of
balance force rotates with the shaft so that in rotating coordinates the out of balance
force is in a fixed direction. Thus, since Q

is constant, there will be no velocities or

accelerations in the rotating coordinate system, so that ( )
{ }
2
2 1
+ + = M G K q Q

.
Separating the translation and rotation coordinate gives
2
0 0 0 0
0 0 0
0
T T
T
p d R R
| | ( ( (

+ + =
|
` ` ` ( ( (
|
) ) \ . )
m k q
Q
I I k q

Since there is no excitation or coupling to the rotation coordinates, we have
2
2
0
0
Tx x
y
Ty
k m Q u
Q v
k m
(


( =
` `
( )
)

75

If the out of balance force is in the x-direction,
2
2
0
2
0
1
0
0
(


( =
` `
( ) )

Tx
Ty
k m
u
m
v
k m
and hence
2
0
2

Tx
m
u
k m
. Similarly
when the out of balance force is in the y direction
2
0
2

Ty
m
v
k m
. Finally, when the
out of balance force is at 45 to the x direction is
2
2
0
2
0
1 2
0 1 2
(


( =
` `
( )
)

Tx
Ty
k m
u
m
v
k m
, Hence
( )
2
0
2
2

=

Tx
m
u
k m
and
( )
2
0
2
2

=

Ty
m
v
k m
. Thus the response is
2
2 2 0
2 2
1 1
2
Tx Ty
m
r u v
k m k m

= + = +

The MATLAB script Problem_07_05.m gives the following output:

Rot or speed = 1900 r ev/ mi n
Response due t o o/ b i n di r ect i on of maj or axi s = 85. 8894 mu_m
Response due t o o/ b i n di r ect i on of mi nor axi s = 336. 9978 mu_m
Response due t o o/ b at 45 degr ee t o maj or axi s = 245. 9111 mu_m

Problem 7.6
3
48
xx
uu
EI
k
L
= ,
3
48
yy
vv
EI
k
L
= . From Equation (7.42),
( )
2 2
2
2 2
2

=
+


x y
x y
where is
the gravity critical speed.
From Equation (7.22),
( ) ( )( )
4 2 2 2 2 2 2 2 2
2 0
x y x y
s s + + + + =


Let s j = , then
2 2
s = and
4 4
s = . Substituting in Equation (7.22) gives
( ) ( )( )
4 2 2 2 2 2 2 2 2
2 0
x y x y
+ + + =


( ) ( )
2 2 2 2 2
2 0
x y x y
+ + =

and thus
( )
2 2
2
2 2
2

=
+


x y
x y
. Now, for this rotor,
2 4
4.7270 10
x Tx
k m = =

and
2 4
4.1546 10
y Ty
k m = =

. Thus
( )
( )
2 2
8
2 4
4
2 2
4.7270 4.1546 10
1.1056 10
2 4.7270 4.1546 10
2


= = =
+
+


x y
x y
. Hence
( )
4
1.1056 10 60 2 1004.1rev min = =
The MATLAB script Problem_07_06.m calculates the gravity critical thus:

Gr avi t y cr i t i cal = 1004. 0798 r ev/ mi n
76

Problem 7.7
For a J effcott rotor, ignoring the mass of the shaft, the only mass is the disc and the
system stiffness is the stiffness of the shaft at mid-span. Let
uu vv T
k k k = = and
R
k k k

= = .

Fixed Coordinates: In coordinates fixed in space we have Mq + Gq +Kq = 0
where, in this case,
0 0 0
0 0 0
0 0 0
0 0 0
d
d
m
m
I
I
(
(
(
=
(
(
(

M ,
0 0 0 0
0 0 0 0
0 0 0
0 0 0
p
p
I
I
(
(
(
=
(
(

(

G ,
0 0 0
0 0 0
0 0 0
0 0 0
T
T
R
R
k
k
k
k
(
(
(
=
(
(
(

K ,
u
v



=
`

)
q .
Letting
0
st
e = q q , (and
0
st
s e = q q ,
2
0
st
s e = q q ) gives
( )
2
0
s s = M + G +K q 0 .
Noting that the first two equations are uncoupled from each other and from the third
and fourth equations, the first two equations give
( ) ( )
2 2
0 0
0, 0 + = + =
st st
T T
ms k u e ms k v e and hence
1 2 T
s s j k m = = ,
5 6 T
s s j k m = = . Since
n
s j = then
1 2 n n T
k m = = . The normal
convention is to order the natural frequencies using the amplitude; here we have
assumed that the lower natural frequencies arise from the translational modes, which
is often the case for realistic systems but is not guaranteed.
From the third and fourth equations,
2
0
2
0
0
0
(
+

(
=
` `
(
) ) +

d R p
st
p d R
I s k I s
e
I s I s k
. Hence
2
2
0
d R p
p d R
I s k I s
I s I s k
+
=
+
. Thus
( ) ( )
2
2
2
0
d R p
I s k I s + + = . Rearranging,
2
0
d p R
I s j I s k + = . Dividing by
d
I we
have
2 2
0
0 s j s + = where
p d
I I = and
2
0 R d
k I = . Solving the quadratic
equations gives
( ) ( )
2 2 2 2 2 2
0 0
4 2 and 4 2 s j s j = =
or
( )
( )
2
2 2 2 2
1 1
0 0
2 2
4 2 s j j
| |
= = +
|
\ .
. Since the natural
frequency is defined to be positive we have,
( )
2
2
1 1
3 0
2 2
n
| |
= + +
|
\ .
and
( )
2
2
1 1
4 0
2 2
n
| |
= + +
|
\ .
, where
3 4 n n
. The eigenvalues may be written
as
3 3 n
s j = ,
4 4 n
s j = ,
7 3 n
s j = ,
8 4 n
s j = .

Rotating Coordinates: Converting the equations of motion from the fixed
coordinates to coordinates rotating with the shaft using the transformation given in
Equations (7.5) (7.10) we have
( ) ( ) { }
2
1 2 1
+ + + Mq + M G q + M G K q = 0

where = M M

, = G G

, = K K

,
77

2
= M M

,
1
0 2 0 0
2 0 0 0
0 0 0 2
0 0 2 0
d
d
m
m
I
I
(
(
(
=
(
(
(

M

and
1
0 0 0 0
0 0 0 0
0 0 0
0 0 0
p
p
I
I
(
(
(
=
(
(
(

G

.
Letting
0
=


st
e q q etc. we obtain four equations. (Note we are writing the root as s to
distinguish it from the roots in the fixed coordinates, s). The first pair of equations are
uncoupled from the second pair of equations and each pair can be solve separately
from the other pair. Thus, from the first pair of equations,
2 2
2 2
2
0
2
+
=
+


T
T
ms m k m s
m s ms m k
, and hence
2 2
2 0 + =
T
ms j m s k m .
Dividing by m gives
( )
2 2 2
1
2 0 + =
n
s j s since
2
1
=
n T
k m. Solving these
two quadratic equations, and simplifying we have ( )
1 n
s j s j = = . Now
n
s j = and so
1 1 n n
= and
2 1 2 n n n
= + = + . Note that taking the
absolute value ensures that the natural frequencies are positive. Furthermore the
ordering is arbitrary as the relative amplitudes of the natural frequencies will vary
with rotor spin speed. From the second pair of equations, we have
( )
( )
2 2 2
2 2 2
2
0
2
+ +
=
+ +


d d p R p d
p d d d p R
I s I I k I I s
I I s I s I I k
. Hence,
( ) ( )
2 2
2 0 + =
d p d R d p
I s j I I s k I I . Dividing by
d
I and noting that
p d
I I = and
2
0 R d
k I = we have ( ) ( )
2 2 2
0
2 1 0 + + = s j s . Solving
these two quadratic equations gives
( ) ( )
2
2
1 1
0
2 2
1 s j s j
| |
= + =
|
\ .

from the definition of s above. Also, from the definition of
3 n
and
4 n
above we
have
( ) ( )
2
2
1 1
3 0 3
2 2
1
n n
= + + = + and
( ) ( )
2
2
1 1
4 0 4
2 2
1
n n
= + + = . Since
3 4 n n
we know that
3 n

corresponds to a backward whirling mode (see Section 3.6.1), and we have to add
to obtain the natural frequency in the rotating frame. Conversely
4 n
is a forward
whirling mode and we have to subtract to obtain the natural frequency in the
rotating frame. This is consistent with the discussion at the end of Section 7.3.1,
although note that this discussion considers the transformation from the rotating to the
stationary coordinates, whereas here we consider the transformation from the
stationary to rotating coordinates. In Problem 7.1 the natural frequencies equal to
16.4507, 31.4102 (twice) and 110.2007 Hz. Applying these results to these natural
frequencies in fixed coordinate computed to convert to rotating coordinates gives
31.4102 50 81.4102Hz + = , 31.4102 50 18.5898Hz = ,
110.2007 50 60.2007Hz = and 16.4507 50 66.4507Hz + = (as in Problem 7.1).

78

Problem 7.8
In rotating coordinate the equations of motion are of the form
( ) ( ) { }
2
1 2 1
+ + + Mq + M G q + M G K q = 0

where for this system
2
2
0
0
kL
kL
(
= (
(

K

0
0
dx
dy
I
I
(
=
(

M

( )
1
0
0
dx dy
dx dy
I I
I I
(
+
( =
( +

M

2
0
0
dy
dx
I
I
(
=
(

,
0
0
p
p
I
I
(
=
(

(

G

1
0
0
p
p
I
I
(
=
(
(

G

and

=
`

)
q
For this system, 0.5m L = ,
6
1 10 N m k = ,
2
10.6kgm
dx
I = ,
2
10.2kgm
dy
I = and
2
0.8kgm
p
I = and hence
( )
( )
1
0
0 20
20 0
0
p dx dy
dx dy p
I I I
I I I
(
+
(
(
+ = =
(
(
+

M G

,
2 1
0
9.4 0
0 0 9.8
p dy
p dx
I I
I I
(
(
+ = =
(
(

(

M G

,
2 6
2 6
0 0.25 10 0
0 0 0.25 10
kL
kL
( (

= = ( (
( (

K

. Letting
0
st
e = q q we have
( ) ( ) { }
2 2
1 2 1 0
s s
(
+ + + =

M + M G + M G K q 0

. The system roots can be
determined from the determinant of the coefficient matrix thus
( ) ( ) { }
2 2
1 2 1
0 s s + + + = M + M G + M G K


This simplifies to
2 2 6
2 2 6
10.6 9.4 0.25 10 20
0
20 10.2 9.8 0.25 10
s s
s s
+
=
+

This leads to a quadratic in
2
s . Alternatively we can solve an eigenvalue problem.

The gravity critical speed can be derived following the approach of Equations (7.37)
and (7.38). Thus we have
( ) ( ) { }
2
1 2 1
sin
cos

+ + +
`

)


t
mg
t
Mq + M G q + M G K q = . Thus
( )
( )
( )
( )
2 2
2 2
0
0
0
0
0
sin
cos
0
(
+
(
(
+ +
(
(
+


(
+

(

`
(
) +

p dx dy
dx
dy
dx dy p
p dy
p dx
I I I
I
I
I I I
kL I I
t
mg
t
kL I I
q q
q =

Let
1 0
sin q t = =

and
2 0
cos q t = = . Then
79

( ) ( )
( ) ( )
2 2 2
0
2 2 2
0
1
1
(
+

(

` `
(
) ) +

p dy dx p dx dy
p dx dy p dx dy
kL I I I I I I
mg
I I I kL I I I
=
The maximum response occurs when the determinant of the coefficient matrix is zero.
Now
( ) { } ( )
2
2
2 2 4
= +
p dy dx p dx dy
D kL I I I I I I and hence, if 0 D = ,
( ) ( )
2 2 2
+ =
p dy dx p dx dy
kL I I I I I I . This leads to gravity critical speed is
given by
( )
2
2
gr
dx dy p
kL
I I I
=
+
.
Out of balance
( ) ( ) { }
2 2
1 2 1 0
1 2
2
1 2


+ + +
`

)


L
m r Mq + M G q + M G K q = . Thus
( )
( )
2 2
2
0
2 2
0
1 2
2
1 2 0
(
+


(

`
(
+
)


p dy
p dx
kL I I
L
m r
kL I I
q = . Hence we can
determine the response q.
The MATLAB script Problem_07_08.m carries out these calculations and gives the
following output

Sol ut i on of t he char act er i st i c equat i ons
Root s at a r ot or speed of 1500 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
0. 0000 3. 8210 0. 0000 - 3. 8210 0
0. 0000 306. 2299 0. 0000 - 306. 2299 0
Root s at a r ot or speed of 1540 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
- 1. 6091 0. 0000 1. 6091 0. 0000 1
0. 0000 310. 2638 0. 0000 - 310. 2638 0

Gr avi t y cr i t i cal = 754. 9382 r ev/ mi n
At 1500 r ev/ mi n, r esponses ar e 0. 96584 mmand 2. 1291 mm
Or bi t r adi us = 2. 3379 mm


Problem 7.9
In this case the force applied to the support in the x and y directions is
= + =
x c c
f ku k v kL k L and = + =
y c c
f k u kv k L kL . The moment acting
on the rotor in the direction is
2 2
c
k L kL . Similarly the moment acting on the
rotor in the direction is
2 2
+
c
kL k L and, in fixed coordinates the equations of
motion become
2 2
2 2
0
0
+ + + =
+ =

d p c
d p c
I I L k L k
I I L k L k
or, in matrix notation,
2 2
2 2
0
0 0
0 0 0
(
(
(
+ + = (
( ` ` ` `
(

( ) ) ( ) )




p
d c
d p
c
I
I kL k L
I I
k L kL

80

Seeking solutions of the form
0
st
e = and
0
st
e = , gives the following equation
for s
2 2 2
2 2 2
det 0
(
+ +
(
=
(
+

d p c
p c d
s I L k s I L k
s I L k s I L k
or
( ) ( )
2 2
2 2 2
0 + + + =
d p c
s I L k s I L k and hence
( )
2 2
0 + =
d p c
I s jI s L k jk ..
Alternatively, we can solve the eigenvalue problem as described by equations (3.48),
(3.50), (3.51) and (3.52). The MATLAB script Problem_07_09.m gives the user the
choice of solving either the characteristic equation or the eigenvalue problem to
determine the system natural frequencies. Of course, both methods give the same
numeric values for the frequencies which are as follows

Sol ut i on of t he char act er i st i c equat i ons
Root s at a r ot or speed of 0 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
- 15. 7337 158. 8948 - 15. 7337 - 158. 8948 0
15. 7337 158. 8948 15. 7337 - 158. 8948 1
Nat ur al f r equenci es at a r ot or speed of 0 r ev/ mi n
25. 2889 Hz 25. 2889 Hz
Root s at a r ot or speed of 3000 r ev/ mi n
Real par t I mag par t Real par t I mag par t Unst abl e = 1
- 15. 7063 149. 7466 - 15. 7063 - 149. 7466 0
15. 7063 168. 5961 15. 7063 - 168. 5961 1
Nat ur al f r equenci es at a r ot or speed of 3000 r ev/ mi n
23. 8329 Hz 26. 8329 Hz

Problem 7.10
Modelling this system requires finite element analysis (FEA) and can only be solved
using appropriate FEA software. The rotor is the same as that of Problem 5.2 except
that now the rotor has internal damping. This can lead to instability at high speeds.
The MATLAB script Problem_07_10.m makes use of the Rotordynamics Software
package to model and analyse the system. The model has 8 elements and the stability
is determining by checking that all the eigenvalues have negative real parts. This is
done every 0.5 rev/min, until any one eigenvalue has a positive real part.. The output
of the script is as follows:

Rot or i s unst abl e above 4521. 5 r ev/ mi n
Brg Type 3 Brg Type 3
N
o
d
e

1
N
o
d
e

2
N
o
d
e

3
N
o
d
e

4
N
o
d
e

5
N
o
d
e

6
N
o
d
e

7
N
o
d
e

8
N
o
d
e

9

81

0 1000 2000 3000 4000 5000
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
Rotor Speed (rev/min)
R
e
a
l
(
e
i
g
e
n
v
a
l
u
e
s
)


Problem 7.11
This is the same system as in Problem 5.2 and 7.10 except that now the rotor is
asymmetric. There shaft stiffness in one direction is 10% higher than in the other
direction. Modelling this system requires finite element analysis (FEA) and must be
solved using appropriate FEA software. Here the MATLAB script Problem_07_11.m
uses the Rotordynamics Software package to model and analyse the system in
rotating coordinates and plots the real part of the eigenvalues. When the real part of
the eigenvalues are positive the system is unstable. It can be seen there are three
speed ranges where that is the case. The MATLAB script gives the following output:

Unst abl e above 1746. 5 r ev/ mi n
Unst abl e bel ow 1831. 5 r ev/ mi n
Unst abl e above 2944. 5 r ev/ mi n
Unst abl e bel ow 2949. 5 r ev/ mi n
Unst abl e above 4030 r ev/ mi n
Unst abl e bel ow 4226. 5 r ev/ mi n

At 3000 r ev/ mi n
Nat ur al f r eq 1 = 17. 4679 Hz
Nat ur al f r eq 2 = 19. 1392 Hz
Nat ur al f r eq 3 = 75. 8073 Hz
Pesudo- nat f r eq f r omnat f r eq 1 = 32. 5321 Hz
Pesudo- nat f r eq f r omnat f r eq 1 = 67. 4679 Hz
Pesudo- nat f r eq f r omnat f r eq 2 = 30. 8608 Hz
Pesudo- nat f r eq f r omnat f r eq 2 = 69. 1392 Hz
Pesudo- nat f r eq f r omnat f r eq 3 = 25. 8073 Hz
Pesudo- nat f r eq f r omnat f r eq 3 = 125. 8073 Hz

82



0 1000 2000 3000 4000 5000
-10
-5
0
5
10
Rotor Speed (rev/min)
R
e
a
l
(
e
i
g
e
n
v
a
l
u
e
s
)
83

Chapter 8

Problem 8.1
From Equation (8.1) we have
( )
( ) ( ) ( ) ( ) { }
{ }
4
2
1
2 0 76 180 132 180 212 180
2
2
0.4 0.2 0.2 0.2 0.7 0.1 0.2 0.3
0.080 0.0097 0.0388 0.0468 0.0520 0.0509 0.0318
0.0080 0.0590 kgm
ub i i
i
j j j j
f m
e e e e
j j j
j
=

=
= + + +
= + + + + + +
= +

( ) ( ) {
( ) ( ) }
{ }
4
2
1
2
2 2
0.080 0 0.0097 0.0388 0.5
0.0468 0.0520 1.0 0.0509 0.0318 1.5
0.1183 0.0237 kgm
ub i i i
i
M m z
j
j j
j
=
=
= + + +
+ + +
= +

Since for balance ( )
2
0
B D ub
b b f + + = and ( )
2
0
B B D D ub
b z b z M + + = .
Cancelling the
2
terms from the above equations, then from Equation (8.4) we have
= Ab v where
1 1 1 1
0.5 1.5
B D
z z
( (
= =
( (

A and
0.0080 0.0590
0.1183 0.0237
ub
ub
f j
M j

= =
` `

) )
v .
Thus
1
0.1063 0.0648
kgm
0.1143 0.0058
j
j

= =
`
+
)
b A v . Hence
0.1245
kgm
0.1144

=
`
)
b and
146.6149
180 degree
2.8974

= =
`
)
b . Thus the unbalance correction masses are
0.1245 1.2447
1
kg
0.1144 1.1445 0.1

= = =
` `
) )
m b
The MATLAB script Problem_08_01.m repeats these calculations to give.

Requi r ed bal ance mass at B = 1. 2447 kg at - 148. 6149 degr ee
Requi r ed bal ance mass at D = 1. 1445 kg at 2. 8974 degr ee

Problem 8.2
From Equation (8.1) we have
( )
( ) ( ) ( ) { }
{ } { }
3
2
1
2 120 180 15 180 45 180
2
2 2 3
10 0.15 50 0.10 20 0.20
0.7500 1.2990 4.8296 1.2941 2.6264 2.6264
6.9081 0.2353 kgmm= 10 6.9081 0.2353 kgm
ub i i
i
j j j
f m
e e e
j j j
j j
=

=
= + +
= + + + +
= + +

84

( ) {
( ) ( ) }
{ }
3
2
1
2 3
2 3 2
10 0.7500 2.2990 0.4
4.8296 1.2941 0.8 2.6264 2.6264 1.6
10 8.0892 2.9706 kgm
ub i i i
i
M m z
j
j j
j
=

=
= + +
+ +
=

Part (a) Since for balance ( )
2
1 2
0
b b ub
b b f + + = and
( )
2
1 1 2 2
0
b b b b ub
b z b z M + + = . From Equation (8.4) we have = Ab v where
1 2
1 1 1 1
0 1.2
b b
z z
( (
= =
( (

A and
2 3
6.9081 0.2353
10
8.0892 2.9706
ub
ub
f j
M j

= =
` `
+
) )
v .
Thus
2 1 2 3
0.1671 2.2402
10 kgm
6.7140 2.4755
brg
j
j


= = =
`
+
)
f b A v .
At 800rev min, the forces at the bearings are
( )
2
3
0.1671 2.2402 1.1725 15.7226
800 2 60 10 N
6.7140 2.4755 47.3109 17.3740
brg
j j
j j

= =
` `
+ +
) )
f .
Thus
15.7662
N
50.4001
brg

=
`
)
f and
94.2651
180 degree
159.8353

= =
`
)
b
Part (b). Since for balance ( )
2
1 3
0
D D ub
b b f + + = and
( )
2
1 1 3 3
0
D D D D ub
b z b z M + + = . Cancelling the
2
terms from the above equations,
then from Equation (8.4) we have = Ab v where.
1 3
1 1 1 1
0.4 1.6
D D
z z
( (
= =
( (

A and
3
6.9081 0.2353
10
8.0892 2.9706
ub
ub
f j
M j

= =
` `
+
) )
v .
Thus
1 3
2.4698 2.1618 2.4698 2.1618
10 kgm= gm
4.4383 2.3971 4.4383 2.3971
j j
j j


= =
` `
+ +
) )
b A v .
Hence the unbalance corrections are
3.2822
gm
5.0442

=
`
)
b and
138.8044
180 degree
151.6272

= =
`
)
b
The MATLAB script Problem_08_02.m repeats these calculations to give

For ce at br g 1 = 15. 7662 N at - 94. 2651 degr ee
For ce at br g 2 = 50. 4001 N at 159. 8353 degr ee

Requi r ed bal ance at dsk 1 = 3. 2822 g mat - 138. 8044 degr ee
Requi r ed bal ance at dsk 3 = 5. 0442 g mat 151. 6272 degr ee

Problem 8.3
At 3000 rev/min, the initial rotor response is:
( )
30 180
0
0.02 0.0173 +0.0100 mm
j
r e j

= =
Adding the trial balance mass
180 180
1
10 10gm
j
b e

= = gives the response
( )
60 180
1
0.03 0.0150 0.0260 mm
j
r e j

= =
Thus
85

( ) ( )
1 0
0.0173 0.0100 0.0150 0.0260 0.0023 0.0360 mm
d
r r r j j j = = + =
From Equation (8.14)
( ) ( ) ( )
( )
0 1
0.0173 +0.0100 10 0.0023 0.0360
3.0769 4.6154 gm
c d
b r b r j j
j
= =
= +

Thus the product of the balance mass times the radius at which it acts is
5.5470gm
c
b = and 180 123.69
b c
b = = . The MATLAB script
Problem_08_03.m repeats these calculations and give the following output

Requi r ed bal ance = 5. 547 g mat 123. 6901 degr ee

Problem 8.4
The initial rotor response at 1500 rev/min in planes 1 and 2 is
64 180
0
89 180
0.35 0.1534 0.3146
mm
0.0063 0.3599
0.36
j
j
e j
j
e




= =
` `

)

)
r The first trial balance mass added in
plane 1 is
0
3 3
1
0.5
0.5
10 10 kgm
0
0
j
e



= =
` `
)
)
b . The rotor response is then
94 180
1
122 180
0.35 0.0305 0.3487
mm
0.2385 0.3816
0.45
j
j
e j
j
e




= =
` `

)

)
r
Then
1 1 0
0.1839 0.0341
mm
0.2447 0.0217
d
j
j

= =
`

)
r r r
The second trial mass added in plane 2 is
3 3
2
90 180
0
0
10 10 kgm
0.5
0.5
j
j
e

= =
` `
)
)
b
The rotor response is then
73 180
2
86 180
0.59 0.1725 0.5642
m
0.0384 0.5487
0.55




= =
` `

)

)
j
j
e j
j
e
r
Then
2 2 0
0.0191 0.2496
mm
0.0321 0.1887
d
j
j

= =
`

)
r r r
Defining
| |
1 2
0.1839 0.0341 0.0191 0.2496
mm
0.2447 0.0217 0.0321 0.1887
d d d
j j
j j
(
= =
(


r r r then
1
6.8256 1.7542 9.1008 1.4630
1
1.5437 8.9132 1.8132 6.6441 mm
d
j j
j j

+ (
=
(
+

r
Defining,
| |
3
1 2
0.5 0
10 kgm
0 0.5j

(
= =
(

b b b , then from Equation (8.26)
1 3
0
0.5076 0.6943
10 kgm
0.5794 0.0935
c d
j
j

(
= =
(

b br r
Thus
0.8601
gm
0.5869
c
(
=
(

b and
126.1743
180 degree
9.1636
c c
(
= =
(

b
The MATLAB script Problem_08_04.m repeats these calculations and gives the
following results

86

Requi r ed bal ance, pl ane 1 = 0. 86007 g mat - 126. 1743 degr ee
Requi r ed bal ance, pl ane 2 = 0. 58688 g mat - 9. 1636 degr ee

Problem 8.5
The initial rotor response at 3000 rev/min at bearings 1 and 2 is
49 180
0
52 180
0.18 0.1181 0.1358
mm
0.3817 0.4886
0.62
j
j
e j
j
e

+
= =
` `
+
)

)
r . The first trial balance mass added at
disk 1 is
0
3 3
1
0.5
0.5
10 10 kgm
0
0
j
e



= =
` `
)
)
b . The rotor response is then
54 180
1
40 180
0.17 0.0999 0.1375
mm
0.5822 0.4885
0.76
j
j
e j
j
e

+
= =
` `
+
)

)
r .
Then
1 1 0
0.0182 0.0017
mm
0.2005
d
j +
= =
`
)
r r r . The second trial mass is added at disk
2 (and the first trial mass at disk 1 is left in place). Thus
0
3 3
2
90 180
0.5 0.5
10 10 kgm
0.5
0.5
j
j
e
j
e

= =
` `
)

)
b
The rotor response is then
22 180
2
24 180
0.10 0.0927 0.0375
mm
0.5755 0.2562
0.63
j
j
e j
j
e

= =
` `
+
)

)
r
Thus
2 2 0
0.0254 0.1733
mm
0.1938 0.2323
d
j
j

= =
`

)
r r r
Defining
| |
1 2
0.0182 0.0017 0.0254 0.1733
mm
0.2005 0.1938 0.2323
d d d
j j
j
+ (
= =
(

r r r then
1 1
5.6519 5.2157 4.4320 0.4241
mm
0.2534 5.0898 0.0197 0.4634
d
j j
j j

(
=
(
+ +

r
Defining
| |
3
1 2
0.5 0.5
10 kgm
0 0.5j

(
= =
(

b b b , then from Example (8.26)
1 3
0
0.4999 0.6864
10 kgm
0.3766 0.4701
c d
j
j

(
= =
(
+

b br r
Thus
0.8492
gm
0.6023
c
(
=
(

b and
126.0642
180 degree
51.3043
c c
(
= =
(

b .
The MATLAB script Problem_08_05.m repeats these calculations and gives the
following output:

Requi r ed bal ance, pl ane 1 = 0. 84919 g mat - 126. 0642 degr ee
Requi r ed bal ance, pl ane 2 = 0. 60233 g mat 51. 3043 degr ee

This machine is flexible and we have performed a 2 plane balance. Therefore above
3000rev/min the residual unbalance will excite the higher flexible modes and hence
the machine will not be balanced at higher speeds. Indeed the machine will not be
exactly fully balanced at any speeds, and is only balanced at the sensor location for
3000rev/min.
87


Problem 8.6
From Equation (6.55), the response of a complex rotor to an out of balance is
( )
1
2 2
0 0 u
j

(
= + +

q K M G C b where
2
0
b is a vector of unknown
unbalance forces. From Equation (6.60), the response of a bent rotor is
( )
1
2
0 0 b b
j

(
= + +

q K M G C Kq where
0 b
q is a vector defining the shape of
the bent rotor. Let us assume that a bent rotor spins at
0
, but fact that the rotor is
bent is unknown and the response is assumed to be due to unbalance. Then this
unbalance can be corrected, and the response reduced to zero, by introducing
appropriate balance masses thus:
( )
( )
1
2 2
0 0 0 0 0 0 b
j

(
= + +

0 K M G C Kq b or
( )
2
0 0 0 b
= Kq b 0 . Thus the
rotor is successfully balanced at speed
0
. When the speed changes to
1
, for
example, then
( )
2
0 1 0 b
Kq b and clearly this is not equal to a vector of zeros. Thus
when the rotor speed changes the response is no longer zero and the rotor is not
balanced.

Problem 8.7
At 1500 rev/min the initial response at bearings 1 and 2 is
150 180
0
120 180
0.20 0.1732 0.1000
mm
0.1500 0.2598
0.30
j
j
e j
j
e

+
= =
` `
+
)

)
r
The first trial mass added to disk 1 is
0
3 3
1
15
0.1 150
10 10 kgm
0
0
j
e



= =
` `
)
)
b
The rotor response at bearings 1 and 2 is then
140 180
1
150 180
0.10 0.0766 0.0643
mm
0.1732 0.1.000
0.20
j
j
e j
j
e

+
= =
` `
+
)

)
r
Then
1 1 0
0.0966 0.0357
mm
0.0232 0.1598
d
j
j

= =
`

)
r r r
The second trial mass added to disk 2 is
0
3 3
2
90 180
0.1 150 15
10 10 kgm
15
0.1 150
j
j
e
j
e

= =
` `
)

)
b
The rotor response at bearings 1 and 2 is then
90 180
2
300 180
0.45 0.4500
mm
0.2000 0.3464
0.40
j
j
e j
j
e

= =
` `

)

)
r
Then
2 2 0
0.1732 0.3500
mm
0.3500 0.6062
d
j
j
+
= =
`

)
r r r
Defining
| |
1 2
0.0966 0.0357 0.1732 0.3500
mm
0.0232 0.1598 0.3500 0.6062
d d d
j j
j j
+ (
= =
(


r r r then
88

1 1
2.6420 12.9045 6.8085 2.7647
mm
2.3220 1.9600 0.9139 1.7090
d
j j
j j

+ + (
=
(
+

r
Defining
| |
3
1 2
15 15
10 kgm
0 15j

(
= =
(

b b b , then from Equation (8.26)
1 1
0
0.4795 0.1505
10 kgm
0.0580 0.0437
c d
j
j

+ (
= =
(


b br r
Thus
0.3350
kg
0.0484
c c
(
= =
(

m b and
17.4255
180 degree
143.0003
c c
(
= =
(

b .
At 3000 rev/min the initial response at bearings 1 and 2 is
70 180
0
110 180
0.10 0.0342 0.0940
mm
0.0684 0.1879
0.20
j
j
e j
j
e

+
= =
` `
+
)

)
r
Adding the first trial mass described by
1
b (above), the rotor response at bearings 1
and 2 is then
90 180
1
150 180
0.25 0.2500
mm
0.2598 0.1500
0.30
j
j
e j
j
e

= =
` `
+
)

)
r
Then
1 1 0
0.0342 0.1560
mm
0.1914 0.0379
d
j
j
+
= =
`

)
r r r
Adding the second trial mass described by
2
b (above), the rotor response at bearings
1 and 2 is then
130 180
2
300 180
0.30 0.1928 0.2298
mm
0.2000 0.3464
0.40
j
j
e j
j
e

+
= =
` `

)

)
r
Then
2 2 0
0.2270 0.1358
mm
0.2684 0.5343
d
j
j
+
= =
`

)
r r r
Defining
| |
1 2
0.0342 0.1560 0.2270 0.1358
mm
0.1914 0.0379 0.2684 0.5343
d d d
j j
j j
+ + (
= =
(


r r r then
1 1
5.1845 5.1720 0.7086 3.1617
mm
1.1757 2.0804 1.6834 0.9965
d
j j
j j

(
=
(
+

r
Thus from Equation (8.26),
1 2
0
1.3270 0.4402
10 kgm
0.4313 0.1001
c d
j
j

+ (
= =
(
+

b br r
Thus
0.0932
kg
0.0295
c c
(
= =
(

m b and
161.6485
180 degree
13.0647
c c
(
= =
(

b .
The MATLAB script Problem_08_07.m repeats these calculations and gives the
following results

Rot or speed 1500 r ev/ mi n
Requi r ed bal ance, pl ane 1 = 0. 33501 kg at 17. 4255 degr ee
Requi r ed bal ance, pl ane 2 = 0. 048397 kg at - 143. 0003 degr ee

Rot or speed 3000 r ev/ mi n
Requi r ed bal ance, pl ane 1 = 0. 093205 kg at 161. 6485 degr ee
Requi r ed bal ance, pl ane 2 = 0. 029517 kg at 13. 0647 degr ee

89

Problem 8.8
The initial response at bearings 1 and 2 in the x and y directions is
0
10 180
0.0420
0.0420
mm
0.0443 0.0078
0.045
x
j
j
e



= =
` `
+
)
)
r
40 180
0
30 180
0.065 0.0498 0.0418
mm
0.0580 0.0335
0.067
j
y
j
e j
j
e




= =
` `

)

)
r
The first trial mass added to disk 1 is
0
3 3
1
7.5
0.03 250
10 10 kgm
0
0
j
e



= =
` `
)
)
b
The rotor response at bearings 1 and 2 in the x and y directions is then
20 180
1
10 180
0.076 0.0714 0.0260
mm
0.0788 0.0139
0.080
j
x
j
e j
j
e




= =
` `

)

)
r
60 180
1
60 180
0.120 0.0600 0.1039
mm
0.0600 0.1039
0.120
j
y
j
e j
j
e




= =
` `

)

)
r
Hence
1 1 0
0.0294 0.0260
mm
0.0345 0.0217
d x x x
j
j

= =
`

)
r r r
1 1 0
0.0102 0.0621
mm
0.0020 0.0704
d y y y
j
j

= =
`

)
r r r
The second trial mass added to disk 2 is
3 3
2
0
0
0
10 10 kgm
7.5
0.03 250
j
e



= =
` `

)
)
b
The rotor response at bearings 1 and 2 in the x and y directions is then
10 180
2
0.0473 0.0083
0.048
mm
0.0500
0.0500
j
x
j
e


= =
` `
)
)
r
40 180
2
40 180
0.074 0.0567 0.0476
mm
0.0575 0.0482
0.075
j
y
j
e j
j
e




= =
` `

)

)
r
Hence
2 2 0
0.0053 0.0083
mm
0.0057 0.0078
d x x x
j
j

= =
`

)
r r r
2 2 0
0.0069 0.0058
mm
0.0006 0.0147
d y y y
j
j

= =
`

)
r r r
Thus
1 2
1 2
0.0294 0.0260 0.0053 0.0083
0.0345 0.0217 0.0057 0.0078
mm
0.0102 0.0621 0.0069 0.0058
0.0020 0.0704 0.0006 0.0147
d x d x
d
d y d y
j j
j j
j j
j j
(
(
(
(
= =
(
(

(


r r
r
r r

The 4 2 matrix
d
r cannot be inverted, but we can obtain the pseudoinverse (see
Equation (8.37) and the following text) to give

2.930 5.503 4.359 6.403 9.258 16.119 2.701 1.890

100
1.897 42.16 7.414 45.28 54.35 61.71 13.53 21.86
d
j j j j
j j j j
+ + (
=
(
+ + +

r
90

Defining
| |
3
1 2
7.5 0
10 kgm
0 7.5

(
= =
(

b b b then we have

0
0.0060 0.0055
kgm
0.0052 0.0007
c d
j
j
(
= =
(

b br r
Thus
0.0324
kg
0.0208
c c
(
= =
(

m b and
137.6634
180 degree
8.0893
c c
(
= =
(

b .
Using x direction data only
| |
1 2
0.0294 0.0260 0.0053 0.0083
mm
0.0345 0.0217 0.0057 0.0078
d d x d x
j j
j j
(
= =
(


r r r
1 1
2.0602 0.7769 2.0467 0.9277
100 mm
9.2800 0.1796 8.8545 1.2691
d
j j
j j

+ (
=
(
+ +

r
Thus, using Equation (8.26),
1
0
0.0086 0.0056
kgm
0.0094 0.0041
c d
j
j

(
= =
(
+

b br r
Thus
0.0410
kg
0.0410
c c
(
= =
(

m b and
146.6913
180 degree
23.3087
c c
(
= =
(

b .
The MATLAB script Problem_08_08.m repeats these calculations and gives the
following results and also gives a plot of the rotor orbit at the bearings, shown below.

Usi ng x di r ect i on dat a onl y
Requi r ed bal ance, pl ane 1 = 0. 041016 kg at - 146. 6913 degr ee
Requi r ed bal ance, pl ane 2 = 0. 041016 kg at 23. 3087 degr ee

Usi ng x and y di r ect i on dat a
Requi r ed bal ance, pl ane 1 = 0. 032418 kg at - 137. 6634 degr ee
Requi r ed bal ance, pl ane 2 = 0. 020814 kg at - 8. 0893 degr ee
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
-0.1
-0.05
0
0.05
0.1
x direction (mm)
y

d
i
r
e
c
t
i
o
n

(
m
m
)

Initial orbit
1st trial orbit
2nd trial orbit


91

Problem 8.9
The initial rotor response at bearings 1 and 2 in the x and y directions is
0
9 180
0.0420
0.0420
mm
0.0444 0.0070
0.045
x
j
j
e



= =
` `
+
)
)
r
39 180
0
34 180
0.065 0.0505 0.0409
mm
0.0555 0.0375
0.067
j
y
j
e j
j
e




= =
` `

)

)
r
The first trial mass added to disk 1 is
0
3 3
1
7.5
0.03 250
10 10 kgm
0
0
j
e



= =
` `
)
)
b
The rotor response at bearings 1 and 2 in the x and y directions is then
22 180
1
13 180
0.076 0.0705 0.0285
mm
0.0779 0.0180
0.080
j
x
j
e j
j
e




= =
` `

)

)
r
61 180
1
56 180
0.120 0.0582 0.1050
mm
0.0671 0.0995
0.120
j
y
j
e j
j
e




= =
` `

)

)
r
Hence
1 1 0
0.0285 0.0285
mm
0.0335 0.0250
d x x x
j
j

= =
`

)
r r r
1 1 0
0.0077 0.0640
mm
0.0116 0.0620
d y y y
j
j

= =
`

)
r r r
The second trial mass added to disk 2 is
3 3
2
0
0
0
10 10 kgm
7.5
0.03 250
j
e



= =
` `

)
)
b
The rotor response at bearings 1 and 2 in the x and y directions is then
7 180
2
3 180
0.048 0.0476 0.0058
mm
0.0549 0.0026
0.0500
j
x
j
e j
j
e

= =
` `
+
)

)
r
45 180
2
40 180
0.074 0.0523 0.0523
mm
0.0575 0.0482
0.075
j
y
j
e j
j
e




= =
` `

)

)
r
Hence
2 2 0
0.0056 0.0058
mm
0.0055 0.0044
d x x x
j
j

= =
`

)
r r r
2 2 0
0.0018 0.0114
mm
0.0019 0.0107
d y y y
j
j

= =
`

)
r r r
Thus
1 2
1 2
0.0285 0.0285 0.0056 0.0058
0.0335 0.0250 0.0055 0.0044
mm
0.0077 0.0640 0.0018 0.0114
0.0116 0.0620 0.0019 0.0107
d x d x
d
d y d y
j j
j j
j j
j j
(
(
(
(
= =
(
(

(


r r
r
r r

The 4 2 matrix
d
r cannot be inverted, but we can obtain the pseudoinverse (see
Equation (8.37) and the following text) to give

0.632 0.922 0.645 0.060 0.490 0.0687 0.236 0.447

100
3.691 5.290 3.452 0.204 2.778 0.0812 1.272 2.198
d
j j j j
j j j j
+ + + (
=
(
+

r
92

Defining
| |
3
1 2
7.5 0
10 kgm
0 7.5

(
= =
(

b b b then we have

0
0.0073 0.0060
kgm
0.0097 0.0025
c d
j
j
(
= =
(
+

b br r
Thus
0.0376
kg
0.0401
c c
(
= =
(

m b and
140.5128
180 degree
14.5691
c c
(
= =
(

b .
Using x direction data only,
| |
1 2
0.0285 0.0285 0.0056 0.0058
mm
0.0335 0.0250 0.0055 0.0044
d d x d x
j j
j j
(
= =
(


r r r
1 1
0.9753 0.7759 1.2276 0.7479
100 mm
5.6155 4.8156 6.0137 3.8134
d
j j
j j

+ (
=
(
+

r
Thus, using Equation (8.26),
1
0
0.0063 0.0070
kgm
0.0034 0.0072
c d
j
j

(
= =
(
+

b br r
Thus
0.0375
kg
0.0318
c c
(
= =
(

m b and
131.8990
180 degree
64.3837
c c
(
= =
(

b .
The MATLAB script Problem_08_09.m repeats these calculations and gives the
following results and also gives a plot of the rotor orbit at the bearings, shown below.

Usi ng x di r ect i on dat a onl y
Requi r ed bal ance, pl ane 1 = 0. 037452 kg at - 131. 899 degr ee
Requi r ed bal ance, pl ane 2 = 0. 031837 kg at 64. 3837 degr ee

Usi ng x and y di r ect i on dat a
Requi r ed bal ance, pl ane 1 = 0. 03761 kg at - 140. 5128 degr ee
Requi r ed bal ance, pl ane 2 = 0. 04009 kg at 14. 5691 degr ee

-0.15 -0.1 -0.05 0 0.05 0.1 0.15
-0.1
-0.05
0
0.05
0.1
x direction (mm)
y

d
i
r
e
c
t
i
o
n

(
m
m
)

Initial orbit
1st trial orbit
2nd trial orbit


93

Problem 8.10
The initial rotor response at bearings 1 and 2 in the x and y directions is
0
20 180
0.0310
0.0310
mm
0.0479 0.0174
0.051
x
j
j
e



= =
` `
+
)
)
r
30 180
0
30 180
0.050 0.0433 0.0250
mm
0.0624 0.0360
0.072
j
y
j
e j
j
e




= =
` `

)

)
r
The first trial mass added to disk 1 is
0
3 3
1
7.5
0.03 250
10 10 kgm
0
0
j
e



= =
` `
)
)
b
The rotor response at bearings 1 and 2 in the x and y directions is then
1
20 180
0.034
0.0340
mm
0.0498 0.0181
0.053
x
j
j
e



= =
` `
+
)
)
r
40 180
1
30 180
0.055 0.0421 0.0354
mm
0.0658 0.0380
0.076
j
y
j
e j
j
e




= =
` `

)

)
r
Hence
2
1 1 0
0.3000
10 mm
0.1879 0.0684
d x x x
j

= =
`
+
)
r r r
2
1 1 0
0.1169 1.0353
10 mm
0.3464 0.2000
d y y y
j
j

= =
`

)
r r r
The second trial mass added to disk 2 is
3 3
2
0
0
0
10 10 kgm
7.5
0.03 250
j
e



= =
` `

)
)
b
The rotor response at bearings 1 and 2 in the x and y directions is then
2
20 180
0.032
0.0320
mm
0.0479 0.0174
0.0510
x
j
j
e



= =
` `
+
)
)
r
30 180
2
30 180
0.052 0.0450 0.0260
mm
0.0624 0.0360
0.072
j
y
j
e j
j
e




= =
` `

)

)
r
Hence
2
2 2 0
0.1000
10 mm
0
d x x x

= =
`
)
r r r
2
2 2 0
0.1732 0.1000
10 mm
0
d y y y
j

= =
`
)
r r r
Thus
1 2
1 2
0.3000 0.1000
0.1879 0.0684 0
mm
0.1169 1.0353 0.1732 0.1000
0.3464 0.2000 0
d x d x
d
d y d y
j
j j
j
(
(
( +
(
= =
(
(

(

r r
r
r r

The 4 2 matrix
d
r cannot be inverted, but we can obtain the pseudoinverse (see
Equation (8.37) and the following text) to give

0.189 0.982 0.483 0.176 0.327 0.378 2.701 1.890

100
5.323 2.945 0.423 2.243 2.762 0.106 3.980 2.236
d
j j j j
j j j j
+ + (
=
(
+ + +

r
94

Defining
| |
3
1 2
7.5 0
10 kgm
0 7.5

(
= =
(

b b b then we have

0
0.0760 0.0044
kgm
0.0412 0.3004
c d
j
j
+ (
= =
(


b br r
Thus
0.3046
kg
1.2128
c c
(
= =
(

m b and
176.6739
180 degree
97.8016
c c
(
= =
(

b .
Using x direction data only
| |
1 2
0.3000 0.1000
mm
0.1879 0.0684 0
d d x d x
j
(
= =
(
+

r r r
1 3 1
0 0.4698 0.1710
10 mm
1.0000 1.4095 0.5130
d
j
j

(
=
(
+

r
Using Equation (8.26),
1
0
0.1913
kgm
0.3413
c d

(
= =
(

b br r
Thus
0.7650
kg
1.3650
c c
(
= =
(

m b and
180.0000
180 degree
0.0000
c c
(
= =
(

b
The MATLAB script Problem_08_10.m repeats these calculations and gives the
following results and also gives a plot of the rotor orbit at the bearings, shown below.

Usi ng x di r ect i on dat a onl y
Requi r ed bal ance, pl ane 1 = 0. 765 kg at 180 degr ee
Requi r ed bal ance, pl ane 2 = 1. 365 kg at - 7. 4562e- 014 degr ee

Usi ng x and y di r ect i on dat a
Requi r ed bal ance, pl ane 1 = 0. 3046 kg at 176. 7 degr ee
Requi r ed bal ance, pl ane 2 = 1. 213 kg at - 97. 8 degr ee

-0.06 -0.04 -0.02 0 0.02 0.04 0.06
-0.05
0
0.05
x direction (mm)
y

d
i
r
e
c
t
i
o
n

(
m
m
)

Initial orbit
1st trial orbit
2nd trial orbit


95

Problem 8.11
The rotor system is described in EXAMPLE 8.5.5, except that the diameter of the left
hand disk is increase to 0.35 m. The MATLAB script Problem_08_11.m models and
simulates the rotor, including the inherent unbalance, using the Rotordynamics
Software Package. From this script Problem_08_11.m we obtain the mode shapes
and the response due to the inherent out of balance. For th first mode of the stationary
rotor, the ratio of displacements between the right hand disk to the left hand disk (
1
a )
is 2.5241 . For the second mode, this ratio (
2
a ) is 2.1588. Thus
1
1
1
a

=
`
)
u and
2
2
1
a

=
`
)
u . Since we wish to balance the effect of the first two modes, we must to
define the vectors
1
e and
2
e according to Equation (8.43).
Fir the first mode we must define
1
e so that
T
1 1
1 = u e . If
2
1
1
a
=
`
)
e then
T
1 1 1 2
a a = u e . Thus if we make
2
1
1 2
1
1
a
a a

=
`

)
e then
T
1 1
1 = u e . There is in fact a
wide range of possible forms of
1
e .
For the second mode we define
2
e so that
T
2 2
1 = u e and
T
1 2
0 = u e . If
1
2
1
a
=
`
)
e ,
then
T
1 2
0 = u e and
T
2 2 2 1
a a = u e . Thus, if we make
1
2
2 1
1
1
a
a a

=
`

)
e then
T
2 2
1 = u e . (Note that the trial forces
1
b and
2
b are determined by
1 1 1
= b e and
2 2 2
= b e . The parameters
1
and
2
are chosen to make the trail forces a suitable
size, i.e. not so small that it is difficult to measure the response accurately, but not so
large as to risk damaging the rotor due to an excessive level of vibration. It does not
matter if
1
e is chosen such that
T
1 1
u e is not equal to unity, because the scale factor in
1
e is multiplied by
1
. The parameter
1
is still chosen to make
1
b have a suitable
magnitude. The same argument applies to
2
e ,)
Initial Response. From the rotor system model, we find that the response at 1110
rev/min at nodes 5 and 13 due to the inherent unbalance is
01
33.9991 8.5685
mm
85.0928 21.4472
j
j

=
`
+
)
r
Balance the first mode.
2
1
1 2
2.1588 0.4610
1 1
1 1 0.2135 4.6829
a
a a

= = =
` ` `

) ) )
e .
Let
1
0.001 = . Then
1 1 1
= b e .
Thus
1
0.4610
gm
0.2135

=
`
)
b and
1
0
180 degrees
180

=
`
)
b
With the trial balance
1
b added to the rotor, the simulation gives the following
response at 1110 re/min at nodes 5 and 13:
96

1
1.8516 8.5685
mm
4.6511 21.4472
j
j

=
`
+
)
r and hence
1 1 01
32.1475
mm
80.4417
d

= =
`
)
r r r
( )
H
3 1 01
1 1
H
1 1
10 1.0578 0.2666
d
d d
j

= =
r r
r r
and
1 1 1
0.4876 0.1229
gm
0.2259 0.0569
c
j
j

= =
`
+
)
b e .
Hence,
1
0.5029
gm
0.2329
c

=
`
)
b and
1
345.85
180
165.85
c

=
`
)
b
With the trial balance forces
1
b removed and replaced by the balance forces
1 c
b , the
first mode is balanced and the simulation gives the following response at 2,315
rev/min at nodes 5 an 13:
02
0.4532 0.1620
100 mm
1.0134 0.3623
j
j

=
`

)
r
Balance the second mode.
1
2
2 1
2.5241 0.5390
1 1
1 1 0.2135 4.6829
a
a a

= = =
` ` `

) ) )
e . Let
2
0.001 = . Then
2 2 2
= b e .
Thus
2
0.5390
gm
0.2135

=
`
)
b and
1
0
180 degrees
0

=
`
)
b
With the trial balance forces
2
b added to the rotor (in addition to the balance forces
1 c
b ), the simulation gives the following response at xxx re/min at nodes 5 and 13:
2
0.6799 0.1620
100 mm
1.5205 0.3623
j
j

=
`

)
r and thus
2 2 02
113.3145
mm
253.3900
d

= =
`

)
r r r
( )
H
3 2 02
2 2
H
2 2
10 0.3999 0.1430
d
d d
j

= =
r r
r r
.
2 2 2
0.2156 0.0771
gm
0.0854 0.0305
c
j
j

= =
`

)
b e .
Hence,
2
0.2289
gm
0.0907
c

=
`
)
b and
2
340.33
180
340.33
c

=
`
)
b
Combining
1 c
b and
2 c
b gives
1 2
0.7032 0.2000
gm
0.1405 0.0264
c c c
j
j

= + =
`
+
)
b b b .
Hence,
0.7311
gm
0.1429
c

=
`
)
b and
344.13
180 degrees
169.36
c

=
`
)
b
The MATLAB script Problem_08_11.m simulates the rotor system and repeats these
calculations and gives the following results:

Requi r ed bal ance, mode 1 at di sk 1: 0. 50288 g mat - 14. 1463 degr ees
Requi r ed bal ance, mode 1 at di sk 2: 0. 23295 g mat 165. 8537 degr ees

Requi r ed bal ance, mode 2 at di sk 1: 0. 22893 g mat - 19. 6714 degr ees
Requi r ed bal ance, mode 2 at di sk 2: 0. 090699 g mat - 19. 6714 degr ees

Tot al bal ance, modes 1 & 2 at di sk 1: 0. 73109 g mat - 15. 874 degr ees
Tot al bal ance, modes 1 & 2 at di sk 2: 0. 14294 g mat 169. 3564 degr ees

97

Problem 8.12
The rotor system is described in EXAMPLE 8.5.4. The MATLAB script
Problem_08_12.m models and simulates the system, including the inherent
unbalance, using the Rotordynamics Software Package. From this script we obtain the
response due to the inherent out of balance. However, the mode shapes are assumed to
be have been measured inaccurately and are
1
1
0.9

=
`
)
u and
2
0.95
1

=
`

)
u . If we make
1
1
0.95
t

=
`
)
e then
T
1 2
0
t
= e u . Letting
( )
T
1 1 1 1
0.5391
0.5121
t t

= =
`
)
e e e u , then
T
1 1
1 = e u . Of
course,
T
1 2
0 = e u .
(Note that the trial forces
1
b and
2
b are determined by
1 1 1
= b e and
2 2 2
= b e . The
parameters
1
and
2
are chosen to make the trail forces a suitable size, i.e. not so
small that it is difficult to measure the response accurately, but not so large as to risk
damaging the rotor due to an excessive level of vibration. It does not matter if
1
e is
chosen such that
T
1 1
u e is not equal to unity, because the scale factor in
1
e is
multiplied by
1
. The parameter
1
is still chosen to make
1
b have a suitable
magnitude. The same argument applies to
2
e .)
Initial Response. From the rotor system model, we find that the response at 820
rev/min at nodes 5 due to the inherent unbalance is
( )
01
1.0100 1.8433 mm = r j
Balance the first mode.
1
0.5291
0.5121

=
`
)
e . Let
3
1
0.4 10

= . Then
1 1 1
0.2156
gm
0.2049

= =
`
)
b e . Thus
1
0.2156
gm
0.2049

=
`
)
b and
1
0
180 degrees
0

=
`
)
b
With the trial balance forces
1
b added to the rotor, the simulation gives the following
response at 820 rev/min at nodes 5:
( )
1
0.0844 2.1567 mm = r j and hence ( )
1 1 01
0.9255 0.3133 mm = =
d
r r r j
Using the data from node 5 only, ( )
4 1 01
1 1
1 1
10 1.4965 8.4731

= =
d
d d
r r
j
r r
.
1 1 1
0.0807 0.4568
gm
0.0766 0.4339

= =
`

)
c
j
j
b e . Hence,
1
0.4638
gm
0.4406

=
`
)
c
b and
1
79.9841
180
79.9841

=
`

)
c
b degrees. With the trial balance forces
1
b removed and
replaced by the balance forces
1 c
b , the first mode is balanced and the simulation
gives the following response at 3075 rev/min at nodes 5:
( )
02
0.6355 0.1399 mm = + r j
98

Balance the second mode.
1
1
0.9

=
`
)
u . If we make
2
0.9
1
t

=
`
)
e then
T
2 1
0
t
= e u . Letting
( )
T
2 2 2 2
0.4852
0.5391
t t

= =
`

)
e e e u , then
T
2 2
1 = e u . Of course,
T
2 1
0 = e u .
Let
3
2
0.4 10

= . Then
2 2 2
0.1941
gm
0.2156

= =
`

)
b e . Thus
2
0.1941
gm
0.2156

=
`
)
b and
2
0
180 degrees
180

=
`
)
b
With the trial balance forces
2
b added to the rotor (in addition to the balance forces
1 c
b ), the simulation gives the following response at 3075 rev/min at nodes 5:
( )
2
0.6425 0.2428 mm = r j and thus ( )
2 2 02
0.0070 0.3827 mm = =
d
r r r j
Using that data from node 5 only, ( )
4 2 02
2 2
2 2
10 1.3405 6.6663

= =
d
d d
r r
j
r r
.
2 2 2
0.0650 0.3234
gm
0.0723 0.3594

= =
`
+
)
c
j
j
b e . Hence,
2
0.3299
gm
0.3666

=
`
)
c
b and
2
78.6300
180 degrees
101.3700

=
`
)
c
b
Combining
1 c
b and
2 c
b gives
1 2
0.1457 0.7802
gm
0.0044 0.0746

= + =
`

)
c c c
j
j
b b b .
Hence
0.7937
gm
0.0747

=
`
)
c
b and
79.4213
180 degrees
86.6443

=
`

)
c
b .
The MATLAB script Problem_08_12.m simulates the rotor system and repeats these
calculations and gives the following results:

Usi ng appr oxi mat e mode shapes
e1 = [ 0. 5391 0. 5121 ]
Requi r ed bal ance, mode 1 at di sk 1: 0. 4638 g mat 280. 0159 degr ees
Requi r ed bal ance, mode 1 at di sk 2: 0. 4406 g mat 280. 0159 degr ees

e2 = [ 0. 4852 - 0. 5391 ]
Requi r ed bal ance, mode 2 at di sk 1: 0. 3299 g mat 281. 3700 degr ees
Requi r ed bal ance, mode 2 at di sk 2: 0. 3666 g mat 101. 3700 degr ees

Tot al r equi r ed bal ance at di sk 1: 0. 7937 g mat 280. 5787 degr ees
Tot al r equi r ed bal ance at di sk 2: 0. 0747 g mat 273. 3557 degr ees

Response at di sk 1 at 820 r ev/ mi n = 87. 6393 mu_m
Response at di sk 1 at 3075 r ev/ mi n = 0. 0000 mu_m

99

Usi ng t he exact mode shapes
e1 = [ 0. 5000 0. 5000 ]
Requi r ed bal ance, mode 1 at di sk 1: 0. 4522 g mat 280. 0167 degr ees
Requi r ed bal ance, mode 1 at di sk 2: 0. 4522 g mat 280. 0167 degr ees

e2 = [ 0. 5000 - 0. 5000 ]
Requi r ed bal ance, mode 2 at di sk 1: 0. 3598 g mat 281. 4885 degr ees
Requi r ed bal ance, mode 2 at di sk 2: 0. 3598 g mat 101. 4885 degr ees

Tot al r equi r ed bal ance at di sk 1: 0. 8120 g mat 280. 6689 degr ees
Tot al r equi r ed bal ance at di sk 2: 0. 0930 g mat 274. 3107 degr ees

Response at di sk 1 at 820 r ev/ mi n = 2. 6806 mu_m
Response at di sk 1 at 3075 r ev/ mi n = 0. 0000 mu_m

Problem 8.13
Using four measurements. (This is a restatement of the solution of EXAMPLE 8.6.1).
Initial estimate for the required balance. The largest difference between the original
response magnitude and the modified response occurs when the trial balance is fitted
at 240. i.e
3 0
2.6759 2.0 0.6759mms r r = = . Therefore, to balance the as found
response of 2.0mms, the initial estimate of the corrective balance is
0
20 2.0 0.6759 59.18gm
c
b = = . Because the trial unbalance at 240 increases the
unbalance response, to reduce the unbalance response the initial estimate for the
position of the corrective balance must be 240 180 60 = .
The initial estimate for ( )
1 1
0 0 0
2.0 59.18 0.0338kg s
c
R r b

= = = at 60 .
We must now determine the values of the three unknown parameters, i.e. ( ) R and
the real and imaginary parts of
c
b to achieve balance. Letting
( )
1
R = ,
( )
2
real
c
b = and ( )
3
imag
c
b = , we must adjust these parameters to make
( )
1 2 3
0, where 1, , 4
i i
r b j i + = = .
To estimate the unknown parameters, we minimise the sum of the squares of the
errors, i.e. ( ) ( )
( )
3
2
1 2 3 1 2 3
0
minimize , ,
i i
i
J r b j
=
= +

To minimize this function requires an iterative scheme. Thi sis done in MATLAB
script Problem_08_13.m, together with the MATLAB user defined function bnpr . m.
Using 4 measurements with an initial balance of 59.18 g m at 60 degrees gives a final
balance of 57.139g m at 42.4897 degrees.
Using three measurements. Case 1. Starting with the same initial balance as in
Example 8.6.1 (above), i.e. 59.18gm at 60 and minimising the function
( )
1 2 3
, , J gives a balance of 57.1353g m at 42.491 degrees.
Using three measurements. Case 2. Starting with an initial balance of 25gmat 40
and minimising the function ( )
1 2 3
, , J gives a balance of 29.654g m at 51.0166
degrees.
Thus to obtain a unique solution this non-linear problem of three unknowns requires
four equations, based on the as found conditions and the response to three trial
balance masses. Three equations, (as found and two trial balance masses) do not give
a unique (correct) solution. The apparent solution depends on the initial conditioned.
Using script Problem_08_13.m gives
100

Usi ng 3 t r i al masses. I ni t i al bal ance = 59. 18 g mat 60 degr ees
Requi r ed bal ance = 57. 139g mat 42. 4897 degr ees
Fi nal val ue of | R( Omega) | = 0. 035002 1/ ( kg s)

Usi ng onl y 2 t r i al masses. I ni t i al bal ance = 59. 18 g mat 60 degr ees
Requi r ed bal ance = 57. 1353g mat 42. 491 degr ees
Fi nal val ue of | R( Omega) | = 0. 035005 1/ ( kg s)

Usi ng onl y 2 t r i al masses. I ni t i al bal ance = 25 g mat 40 degr ees
Requi r ed bal ance = 29. 654g mat 51. 0166 degr ees
Fi nal val ue of | R( Omega) | = 0. 067444 1/ ( kg s)

Problem 8.14
The rotor system is described in EXAMPLE 8.5.4. The MATLAB script
Problem_08_14.m models and simulates the rotor, including the inherent unbalance,
using the Rotordynamics Software Package.
Initial Response
From the rotor system model, we find that the response at 1000 rev/min at nodes 5
and 13 due to the inherent unbalance is
01
1.2173 2.0176
mm
1.2184 2.0227
j
j

=
`

)
r

Balance the first mode.
Since we wish to balance the effect of the first mode by adding a mass to the LH disk
only, we must to define the vector
1t
e as
1
1
0
t

=
`
)
e . The first mode shape is
1
1
1

=
`
)
u .
Hence
( )
T
1 1 1 1
1
0
t t

= =
`
)
e e e u and then
T
1 1
1 = e u . The parameters
1
is chosen to make
the trial forces a suitable size, i.e. not so small that it is difficult to measure the
response accurately, but not so large as to risk damaging the rotor due to an excessive
level of vibration. It does not matter if
1
e is chosen such that
T
1 1
u e is not equal to
unity, because the scale factor in
1
e is multiplied by
1
. The parameter
1
is still
chosen to make
1
b have a suitable magnitude. (The same argument applies to
2
e ,)
Let
3
1
0.4 10

= . Then
3
1 1 1
0.4
10
0

= =
`
)
b e and hence
1
0.4
gm
0

=
`
)
b and
1
0
180 degrees
0

=
`
)
b . With the trial balance forces
1
b added to the rotor, the
simulation gives the following response at 1000 re/min at nodes 5 and 9:
1
0.2466 2.3925
mm
0.2447 2.3976
j
j

=
`

)
r and hence
1 1 01
0.9707 0.3749
mm
0.9737 0.3749
d
j
j

= =
`

)
r r r
( )
H
4 1 01
1 1
H
1 1
10 1.5717 8.9177
d
d d
j

= =
r r
r r
.
1 1 1
0.1572 0.8918
gm
0
c
j
= =
`
)
b e .
Hence,
1
0.9055
gm
0
c

=
`
)
b and
1
80.0045
180 degrees
0
c

=
`
)
b
101

With the trial balance
1
b removed and replaced by the balance
1 c
b , the first mode is
balanced and the simulation gives the following response at 3075 rev/min at nodes 5
an 9:
02
0.1738 0.0177
mm
0.1731 0.0172
j
j

=
`
+
)
r
Balance the second mode.
1
1
1

=
`
)
u and
2
1
1

=
`

)
u If we make
2
1
1
t

=
`
)
e then
T
2 1
0
t
= e u . Letting
( )
T
2 2 2 2
0.5000
0.5000
t t

= =
`

)
e e e u , then
T
2 2
1 = e u . Of course,
T
2 1
0 = e u .
Let
3
1
0.4 10

= . Then
2 2 2
0.2
gm
0.2

= =
`

)
b e and hence
2
0.2
gm
0.2

=
`
)
b and
2
0
180 degrees
180

=
`
)
b . With the trial balance forces
2
b added to the rotor (in
addition to the balance forces
1 c
b ), the simulation gives the following response at
3000 rev/min at nodes 5 and 9:
2
0.1633 0.3911
mm
0.1626 0.3906
j
j

=
`
+
)
r and thus
2 2 02
0.0104 0.3734
mm
0.0104 0.3734
d
j
j

= =
`
+
)
r r r
( )
H
4 2 02
2 2
H
2 2
10 0.1353 1.8614
d
d d
j

= = +
r r
r r
.
2 2 2
0.0068 0.0931
gm
0.0068 0.0931
c
j
j
+
= =
`

)
b e .
Hence,
2
0.9331
gm
0.9331
c

=
`
)
b and
2
94.1568
180 degrees
85.8432
c

=
`

)
b
Combining
1 c
b and
2 c
b gives
1 2
0.1504 0.7987
gm
0.0068 0.0931
c c c
j
j

= + =
`

)
b b b . Hence,
0.8127
gm
0.0933
c

=
`
)
b and
280.6647
180 degrees
274.1568
c

=
`
)
b
The MATLAB script Problem_08_14.m simulates the rotor system and repeats these
calculations and gives the following results (very similar to EXAMPLE 8.5.4):

Requi r ed bal ance, mode 1 at di sk 1: 0. 90552 g mat - 80. 0045 degr ees
Requi r ed bal ance, mode 1 at di sk 2: 0 g mat 0 degr ees

Requi r ed bal ance, mode 2 at di sk 1: 0. 093314 g mat 94. 1568 degr ees
Requi r ed bal ance, mode 2 at di sk 2: 0. 093314 g mat - 85. 8432 degr ees

Tot al bal ance, modes 1 & 2 at di sk 1: 0. 81274 g mat 280. 6647 degr ees
Tot al bal ance, modes 1 & 2 at di sk 2: 0. 093314 g mat 274. 157 degr ees

102

Problem 8.15
Balance mode 1 at 830 rev/min. Converting the as found response at 830 rev/min and
response to the trial balance mass to complex numbers gives
81 180
85 180
01
98 180
1.616
0.2528 1.5961
0.633 0.0552 0.6306 mm
0.5861 4.1700
4.211
j
j
j
e
j
e j
j
e

+


= =
` `

+
)

)
r
72 180
94 180
1
87 180
1.726
0.5334 1.6415
0.651 0.0454 0.6494 mm
0.2246 4.2861
4.292
j
j
j
e
j
e j
j
e

+


= =
` `

+
)

)
r
Thus
1 1 01
0.2806 0.0454
0.1006 0.0188 mm
0.8107 0.1161
d
j
j
j
+

= =
`

+
)
r r r
Now
1
b has been chosen to be proportional to
1
e . Since
1
0
5 gm
0


=
`

)
b then
1
e must
be of the form
0
0
p


`

)
. The element p must be chosen to make
T
1 1
1 = u e .
Since
1
0.3780
0.1526
1


=
`

)
u ,
1
0 0
1 0.1526 6.5531
0 0


= =
` `

) )
e
( ) ( )
1 1 1 1 1
2 2
5 6.5531 0.7630 = = = = b e b e .
( )
H H
1 1 1 01 1 1
0.1590 + 3.9549
d d d
j = = r r r r .
1 1 1
0
1.0419 25.9168 gm
0
c
j


= =
`

)
b e
and hence
1
0
25.94 gm
0
c


=
`

)
b and
1
0
92 degrees
0
c


=
`

)
b
Balance mode 2 at 1080 rev/min. Converting the as found response at 1080 rev/min
and response to the trial balance mass to complex numbers gives
156 180
15 180
02
25 180
9.142
8.3516 3.7184
0.447 0.4318 0.1157 mm
2.0836 0.9716
2.299
j
j
j
e
j
e j
j
e

= = +
` `

+
)

)
r
103

160 180
15 180
2
21 180
13.664
12.8400 4.6734
0.682 0.6588 0.1765 mm
3.2068 1.2310
3.435
j
j
j
e
j
e j
j
e

= = +
` `

+
)

)
r . Thus
2 2 02
4.4883 0.9550
0.2270 0.0608 mm
1.1232 0.2594
d
j
j
j


= = +
`

+
)
r r r . Now
2
b has been chosen to be
proportional to
2
e . Since
2
5
0 gm
1.89


=
`

)
b ,
2
e must be of the form
1
3
0
p
p


`

)
. Also, we
must make
1 2
0 =
T
u e and
2 2
1 =
T
u e where
2
1
0.0514
0.2513


=
`

)
u . Thus
1 T
1
T
2
3
0
0
1
p
p

(

= (
` `
( )


)
u
u
or
( ) ( )
( ) ( )
1 1
1 1 3
3 2 2
1 3
0
1
p
p
(

= (
` `
) ( )

u u
u u

( ) ( )
( ) ( )
1 1
1 3
2 2
1 3
0.3780 1
1 0.2513
(
(
= = (
(

(


u u
U
u u
. Hence
1 1
3
0 0.9132
1 0.3452
p
p

= =
` ` `

) ) )
U
Thus
2
0.9132
0
0.3452


=
`

)
e , Check on orthogonality: | |
T
1
1 2
T
2
1 0
0.3368 1

(
=
`
(


)
u
e e
u

( ) ( )
2 2 2 2 2
1 1
5 0.9132 5.4750 = = = = b e b e .
Thus
( )
H H
2 2 2 02 2 2
10.6695 2.2603
d d d
j = = r r r r .
2 2 2
9.7440 2.0642
0 gm
3.6832 0.7803
c
j
j


= =
`

+
)
b e
or
2
9.9602
0 gm
3.7650
c


=
`

)
b and
2
168
0 degrees
12
c


=
`

)
b
Balance mode 3 at 2900 rev/min. Converting the as found response at 2900 rev/min
and response to the trial balance mass to complex numbers gives
156 180
152 180
03
152 180
0.791
0.7226 0.3217
10.290 9.0855 4.8309 mm
0.7867 0.4183
0.891
j
j
j
e
j
e j
j
e

+


= = +
` `

+
)

)
r
104

169 180
164 180
3
165 180
0.823
0.8079 0.1570
10.731 10.3153 2.9579 mm
0.8973 0.2404
0.929
j
j
j
e
j
e j
j
e

+


= = +
` `

+
)

)
r
Thus
3 3 03
0.0853 0.1647
1.2298 1.8730 mm
0.1106 0.1779
d
j
j
j


= =
`


)
r r r
We must make
1 3
0 =
T
u e ,
2 3
0 =
T
u e and
3 3
1 =
T
u e . Thus,
T
1
T
2 3
T
3
0
0
1




=
` `

)

)
u
u e
u
. Letting
T
1
T
2
T
3
0.3780 0.1526 1
1 0.0514 0.2513
0.0763 1 0.0866

(

(

=
`
(

(


)
u
U u
u
. then
1
3
0 0.0806
0 0.9835
1 0.1196

= =
` `

) )
e U .
Check on orthogonality:
| |
T
1
T
2
T
3
1 0 0
0.3368 1 0
6.5531 0.0398 1

(


(
=
`
(

(


)
1 2 3
u
u e e e
u

Now
3
b should have been chosen to be proportional to
3
e .
3
0.410
5 gm
0.608


=
`

)
b and this
implies that
3
0.4100 0.0806or 5 0.9835or 0.608 0.1196 = . Thus
3
5.0862or 5.0839or 5.0832 = . This means that
3
b proportional to
3
e to an
accuracy of 3 significant figures. The mean value is for
3
is 5.0844.
( )
H H
3 3 3 03 3 3
2.1402 23.2255
d d d
j = = + r r r r .
3 3 3
0.1725 1.8722
2.1048 22.8421 gm
0.2560 2.7780
c
j
j
j
+

= = +
`

+
)
b e
or
3
1.881
22.939 gm
2.790
c


=
`

)
b and
3
95
95 degrees
95
c


=
`

)
b

The MATLAB script Problem_08_15.m repeats these calculations and gives the
following results:

105

Mode 1 at 830 r ev/ mi n
Vect or e1 = [ 0. 0000 - 6. 5531 0. 0000]
del t a = - 0. 763
al pha = 3. 9581 at 87. 6979 degr ee
Ampl i t ude of bc1 = 0. 0000 25. 9378 0. 0000
Phase ( deg) of bc1 = 0. 0000 - 92. 3021 0. 0000

Mode 2 at 1080 r ev/ mi n
Vect or e2 = [ 0. 9132 0. 0000 - 0. 3452]
del t a = 5. 475
al pha = 10. 9063 at - 168. 0391 degr ee
Ampl i t ude of bc2 = 9. 9602 0. 0000 3. 7650
Phase ( deg) of bc2 = - 168. 0391 0. 0000 11. 9609
Or t hogonal i t y t est
[ 1. 0000 0. 0000 ]
[ 0. 3368 1. 0000 ]

Mode 3 at 2900 r ev/ mi n
Vect or e3 = [ 0. 0806 0. 9835 0. 1196]
del t a = 5. 0844
al pha = 23. 3239 at 95. 2648 degr ee
Ampl i t ude of bc3 = 1. 8801 22. 9388 2. 7898
Phase ( deg) of bc3 = 95. 2648 95. 2648 95. 2648
Or t hogonal i t y t est
[ 1. 0000 0. 0000 - 0. 0000 ]
[ 0. 3368 1. 0000 - 0. 0000 ]
[ - 6. 5531 0. 0398 1. 0000 ]
106

Chapter 9

Problem 9.1
(a) Axial vibration. In this case,
2 2
4 0.0020m A d = = ,
8
1 1
9.8175 10 N m = = k EA L ,
8
2 2
6.545 10 N m = = k EA L ,
3
50 10 N m =
c
k
and
3
30 10 N m =
b
k .
2
4 22.054kg
D
M hD = = . The equations of motion (with
4 degrees of freedom) for both axial vibrations are Mq +Kq = 0 where
1 1
1 1
2 2
2 2
0 0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0 0
b D
c c D
c c D
D
k k k M
k k k k M
k k k k M
k k M
+ ( (
( (
+
( (
= =
( ( +
( (

( (

K M
This leads to the eigenvalue problem
0 0
= Mq Kq where
2
n
= . This is solved in
the MATLAB script Problem_09_01.m and the natural frequencies are shown below.
If we divide the system into to parts then the system matrices are shown below.
1 1
1 1
1 1
0
0
D
r r
D
k k M
k k M
( (
= =
( (

K M ,
2 2
2 2
2 2
0
0
D
r r
D
k k M
k k M
( (
= =
( (

K M
Thus we have two systems with 2 degrees of freedom and each resultant eigenvalue
problem can be solved separately. See the output of the MATLAB script below.
If we join the central inertias together then
2 0
0 2
b c c D
rb rb
c c D
k k k M
k k M
+ ( (
= =
( (

K M . This is a 2 degree of freedom system
and the resultant eigenvalue can be solved, see the output of the MATLAB script
Problem_09_01.m as follows:

Axi al vi br at i ons
Nat ur al f r equency 1 = 2. 711 Hz
Nat ur al f r equency 2 = 8. 204 Hz
Nat ur al f r equency 3 = 1226. 1684 Hz
Nat ur al f r equency 4 = 1501. 7444 Hz

Tr eat i ng each r ot or separ at el y
Syst emA, nat ur al f r equency 1 = 0 Hz
Syst emA, nat ur al f r equency 2 = 1501. 7291 Hz
Syst emB, nat ur al f r equency 1 = 0 Hz
Syst emB, nat ur al f r equency 2 = 1226. 1566 Hz

Tr eat i ng r ot or as t wo r i gi d bodi es coupl ed t oget her
Syst emA & B j oi ned, nat ur al f r equency 1 = 2. 7111 Hz
Syst emA & B j oi ned, nat ur al f r equency 2 = 8. 204 Hz

The coupling stiffness is very low compared to the shaft axial stiffnesses. For the two
highest frequencies, the coupling has little effect so treating the two rotors separately
gives accurate results. Conversely, joining the two rotors together predict the two
lowest frequencies very accurately.

107

(b) Torsional vibration. In this case
4 7 4
32 6.136 10 m

= = J d ,
5
1 1
1.227 10 Nm = = k GJ L ,
4
2 2
8.181 10 Nm = = k GJ L ,
5
6 10 Nm =
c
k and
2 2
8 0.248kgm = =
D D
I M D . The equations of motion (with 4 degrees of freedom)
for both axial vibrations are Mq +Kq = 0 where
1 1
1 1
2 2
2 2
0 0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0 0
D
c c D
c c D
D
k k I
k k k k I
k k k k I
k k I
( (
( (
+
( (
= =
( ( +
( (

( (

K M
This leads to the eigenvalue problem
0 0
= Mq Kq where
2
n
= . This is solved in
the MATLAB script Problem_09_01.m and the natural frequencies are shown below.
If we divide the system into to parts then the system matrices are shown below.
1 1
1 1
1 1
0
0
D
r r
D
k k I
k k I
( (
= =
( (

K M ,
2 2
2 2
2 2
0
0
D
r r
D
k k I
k k I
( (
= =
( (

K M
Thus we have two systems with 2 degrees of freedom and each resultant eigenvalue
problem can be solved separately. See the output of the MATLAB script below.
If we join the central inertias together then
2 0
0 2
c c D
rb rb
c c D
k k I
k k I
( (
= =
( (

K M . This is a 2 degree of freedom system and
the resultant eigenvalue can be solved, see the output of the MATLAB script
Problem_09_01.m as follows:

Tor si onal vi br at i ons
Nat ur al f r equency 1 = 0 Hz
Nat ur al f r equency 2 = 94. 5794 Hz
Nat ur al f r equency 3 = 146. 2986 Hz
Nat ur al f r equency 4 = 365. 9606 Hz

Tr eat i ng each r ot or separ at el y
Syst emA, nat ur al f r equency 1 = 0 Hz
Syst emA, nat ur al f r equency 2 = 158. 2961 Hz
Syst emB, nat ur al f r equency 1 = 0 Hz
Syst emB, nat ur al f r equency 2 = 129. 2483 Hz

Tr eat i ng r ot or as t wo r i gi d bodi es coupl ed t oget her
Syst emA & B j oi ned, nat ur al f r equency 1 = 0 Hz
Syst emA & B j oi ned, nat ur al f r equency 2 = 247. 5005 Hz

Note that because the torsional stiffness of the coupling is similar to the shaft
stiffnesses, analysing the rotors separately, or joining then with a rigid connection
give poor prediction of the natural frequencies of the 4 degree of freedom system.

108

Problem 9.2
For this relatively simple system the equations of motion are Mq +Kq = 0 where the
stiffness and inertia matrices are
1 1
1 1 1
1 1 2 2
2 2 2
2 2
0 0 0 0
0 0 0
0 0 0 0
2 0 0
0 0 0 0
0 0
0 0 0 0
0 0 2
0 0 0
0 0 0 0
cr
cg
te
cg
gr
I
k k
I
k k k
I
k k k k
I
k k k
k k
I
(
(
(
(

(
(
(
( = = +
(
(

(
(
(
(

(

K M
Substituting numerical values give
6
1 1 0 0 0 100 0 0 0 0
1 2 1 0 0 0 10 0 0 0
25 10 0 1 2 1 0 0 0 50 0 0
0 0 1 2 1 0 0 0 10 0
0 0 0 1 1 0 0 0 0 80
( (
( (

( (
( ( = =
( (

( (
( (

K M
Solving these equations to determine the natural frequencies leads to an eigenvalue
problem. This eigenvalue problem is solved in the MATLAB script
Problem_09_02.m. The output is as follows:

Nat ur al f r equency 1 = 0 Hz
Nat ur al f r equency 2 = 58. 7123 Hz
Nat ur al f r equency 3 = 124. 8542 Hz
Nat ur al f r equency 4 = 360. 9772 Hz
Nat ur al f r equency 5 = 378. 8684 Hz

Not that the first natural frequency is zero, because the system is unconstrained.

Problem 9.3

Suppose the rotation of the small gear is . The kinetic energy of gear is
2
1
2
2
D
T I = .
But
2
g = and hence the kinetic energy of the gear is
2 2
1
2 2
2
D
T I g =

. Thus, referred
to
2
, the moment of inertia of the small gear is
2
2 D
g I . Similarly, considering the
strain energy stored in the right side shaft,
( ) ( )
2 2
1 1
2 3 2 2 3
2 2
V k k g = = .

1

2

g
2

3

k
1

k
2

I
D2

I
D1

I
D1

I
D1

109

Hence
( )
2 2 2
1
2 2 2 3 3
2
2 V k g g = + . This implies that there are terms in
2
2
g k and
2
gk appear in the equation of motion see the stiffness matrix below. Thus, for this
relatively simple system the equations of motion are Mq +Kq = 0 where the stiffness
and inertia matrices are
1 1 1
2 2
1 1 2 2 1 2
2 2 1
0 0 0
0 0
0 0 0
D
D D
D
k k I
k k g k gk I g I
gk k I
( (
( (
= + = +
( (
( (

K M .
Using the data given, the second moment of area for the shaft is
7 4
6.136 10 m

6
and hence
5
1
1.227 10 Nm k = and
4
2
8.181 10 Nm k = ,
2
1
0.248kgm
D
I = ,
2
2
0.0155kgm
D
I = and 2 g = . The MATLAB script Problem_09_03.m solves
the eigenvalue problem for this problem and gives the following output:

Nat ur al f r equency 1 = 0 Hz
Nat ur al f r equency 2 = 106. 2585 Hz
Nat ur al f r equency 3 = 215. 2714 Hz

Note that since this torsional system is unconstrained, the first natural frequency is
zero.

Problem 9.4
Consider the system with all the constraints removed, i.e. there are no gear
interactions. Thus (see Equation (9.10)),
U U U U U C
+ = + M q K q Q Q where
C
Q are
the internal forces that enforce the constraints between the degrees of freedom. The
unconstrained mass and stiffness matrices are

1
2
3
4
1
2
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
g
g
g
U
g
m
e
m
I
I
I
I
I
I
I
(
(
(
(
(
( =
(
(
(
(
(

M
1 1
2 2
1 1
2 2
0 0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
m m
e e
m m U
m m
e e
m m
k k
k k
k k
k k
k k
k k
(
(
(
(
(
=
(
(

(
(

K

110

Let us refer the unconstrained coordinates
1
,
2
and
4
to
3
. From Figure 9.16 we
can see that
2 23 3
= where
23 3 2
r r = and
4 43 3
= where
43 3 4
r r = .
1 12 2
= where
12 2 1
r r = . In matrix notation
23 1
12
43 7
0 1 0 0 0 0 0
1 0 0 0 0 0 0
0 0 1 0 0 0 0
(

(
=
` `
(

(
) )
or
T
U
= E q 0 . (Equation (9.12). Thus
23
T
12
43
0 1 0 0 0 0
1 0 0 0 0 0
0 0 1 0 0 0
(
(
=
(
(

E
Using
1 12 2
= and
2 23 3
= we have ( )
1 12 23 3
= . Let
U R
= q Tq where
1 12 23
2 23
3
3
5
4 43
6
5
7
6
7
0 0 0
0 0 0
1 0 0 0
0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
(
(

(

(

(
=
` `
(


(

(
)

(

(

)
.
Thus when we apply constraints the system has 4 degrees of freedom. Note that
T
E T = 0 . Note also that in this system
12 43
2.8571 = = and
23
1 = . Since
U U U U U C
+ = + M q K q Q Q , (Equation (9.12)) and
U R
= q Tq ,
T
=
U R
T Q Q ,
T
=
C
T Q 0 and
T
=
R U
M T M T etc, we have
R R R R R
+ = M q K q Q This gives
5
5.469 0.257 4.000 0.257 0.063 0 0 0
0.257 0.090 0 0 0
10
4.000 0 4.000 0
0.257 0 0 0.090
(
(

(
= =
(
(

R R
K M
0.180 0 0
0 0 20 0
0 0 0 0.180
(
(
(
(
(


We can determine the system natural frequencies by solving the eigenvalue problem
2
0 0
=
R R R R
M q K q . (For the system natural frequencies, see the output from
MATLAB script Problem_09_04.m, below).

Gear error: The gear error is
( )
5
2 3
1 10 sin 3000t

+ = . This can be combined

with the existing constraints by ( )
T
U
t = E q e (Equation (9.30)) where in this case
( )
5
1 10
0 sin
0

=
`

)
t t e . (Recall that
23
1 = ). Let
U R ref
= + q Tq q (Equation
(9.31)). Here
ref
q is chosen so that
( )
T
ref
t = E q e . Thus
( )
T
ref
t = EE q Ee and
hence
( )
( )
1
T

=
ref
t q E E E e (Equation (9.32)).
111

In this problem
{ }
T 4
10 0.143 0.050 0.050 0.143 0 0 0 sin
ref
t = q .
Substitute
U R ref
= + q Tq q into
U U U U U C
+ = + M q K q Q Q (Equation (9.12)) and
premultiplying by
T
T , noting that and
T
=
C
T Q 0 and
T
=
R U
M T M T etc. gives
( )
T
+ = +
R R R R R U ref U ref
M q K q Q T M q K q , Equation (9.33). This can be written
as + = +
R R R R R Rref
M q K q Q Q where
( )
T
= +
Rref U ref U ref
Q T M q K q . In this
problem
R
= Q 0 . Now since
ref
q is a harmonic,
2
ref ref
= q q and hence
( )
T 2
Rref U U ref
= Q T K M q . In this problem
2.000
0.129
sin
2.000
0.129
Rref
t

=
`

)
Q where 3000rad s =
Solving Equation (9.33), we have
( )
1
2
R R R Rref

= q K M Q . In this problem
3
0.1316
0.0020
10 sin t
0.0003
0.0022
R

=
`

)
q
Now
U R ref
= + q Tq q .and in this problem
{ }
T 3
10 0.3618 0.1266 0.1366 0.3904 0.0020 0.0003 0.0022 sin
U
t

= q
We can determine ( )
6 3 ke e
k = where
6
etc. is a component of
U
q . From
Equation (9.12), assuming a harmonic solution and
U
= Q 0 , then the torques due to
the constraints and gear error are
( )
2
C U U U
= Q K M q . This gives, for this
problem
{ }
T
3.238 17.096 36.331 3.494 0 0 0 sin
C
t = Q
Hence
12 1 1 C
f Q r = ,
34 4 4 C
f Q r = ,
23 2 2 12 C
f Q r f = or
23 3 3 34 C
f Q r f = .
The MATLAB script Problem_09_04.m solves this problem and gives the following
output.

Nat ur al f r equency 1 = 0. 0000 Hz
Nat ur al f r equency 2 = 32. 5860 Hz
Nat ur al f r equency 3 = 35. 5881 Hz
Nat ur al f r equency 4 = 470. 9944 Hz

Max t or que on shaf t ke = - 54. 7769 Nm
Max t angent i al f or ce = 263. 4817 N

Problem 9.5
From Equation (9.31)
U R ref
= + q Tq q , we know that
T
=
U
E q e (Equation (9.30))
and hence
T T T T
U R ref ref
= + = E q E Tq E q E q (since
T
= E T 0 ).
Thus
T
ref
= E q e . Equation (SM.1)
112

Suppose we have two choices for
ref
q , denoted by
1 ref
q and
0 ref
q . Thus from
Equation (SM.1)
( )
T
1 0 ref ref
= = E q q e e 0 . Equation (SM.2)
Since
T
= E T 0 and
| |
E T is full rank, we can write
1 0 ref ref
q q in terms of the
columns of
| |
E T . Thus
1 0 ref ref
= + q q Th Eg Equation (SM.3)
for some vectors h and g . From Equation (SM.2)
( )
T T T
1 0 ref ref
= + = E q q E Th E Eg 0 Equation (SM.4)
Since
T
= E T 0 , this implies that
T
= E Eg 0 Equation (SM.5)
Since E is full rank we must have = g 0 and hence
1 0 ref ref
= q q Th Equn (SM.6).
The two possible
ref
q vectors will lead to different reduced states
R
q , denoted by
0 R
q and
1 R
q . From Equation (9.33) these are given by
( )
( )
T
0 0 0 0
T
1 1 1 1
+ = +
+ = +


R R R R R U ref U ref
R R R R R U ref U ref
M q K q Q T M q K q
M q K q Q T M q K q

Subtracting, and using Equation (SM.6), we have
( ) ( ) ( )
T
1 0 1 0
+ = +
=

R R R R R R U U
R R
M q q K q q T M Th K Th
M h K h

and hence
( ) ( )
1 0 1 0 R R R R R R
+ + + = M q q h K q q h 0

Equation (SM.7)
Initial conditions: From Equations (9.31) and (SM.6),
( )
1 0 1 0 u u R R
= + = q q T q q h 0 and thus
1 0 R R
+ = q q h 0 at 0 t = . (Similarly for
the initial velocities). Then from Equation (9.31)
0 0 0 1 0
1 1 0 0 1
u R ref R ref
R ref ref ref u
= + = + +
= + + =
q Tq q Tq Th q
Tq q q q q


Problem 9.6
The equations of motion can be derived from energy principles or by applying
Newtons 2
nd
Law.

For example, considering mass
3
m in the free body diagram shown above, then
3 3 3 4
m q f f = where
3
f and
4
f are spring forces and are given by
( )
3 3 3 2
f k q q = and ( )
4 4 3 4
f k q q = . Form these three relationships, and
rearranging we obtain ( )
3 3 3 2 3 4 3 4 4
0 m q k q k k q k q + + = . Applying Newtons 2
nd

Law to each mass gives the following equation, expressed here in matrix notation
thus + = Mq Kq 0 where
m
3
m
2

q
2
q
3

m
4

q
4

f
3
f
4

113

| |
1 2 3 4 5 6 7 8 9
diag m m m m m m m m m = M and
1 2 2
2 2 3 3
3 3 4 4
4 4 5 5
5 5 6 6
6 6 7 10 7
7 7 8 8
8 8 9 9
9 9
0 0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0 0
k k k
k k k k
k k k k
k k k k
k k k k
k k k k k
k k k k
k k k k
k k
=
+ (
(
+
(
( +
(
+
(
(
+
(
+ +
(
(
+
(
+
(
(

(

K

To obtain the first three axial; natural frequencies we must solve the eigenvalue
problem
0 0
= Kq Mq where
2
n
= . The MATLAB script Problem_09_06.m
formulates the mass and stiffness matrices from the given data, solves the eigenvalue
problem and gives the following output:

Nat ur al f r equency 1 = 1. 5479 Hz
Nat ur al f r equency 2 = 11. 9264 Hz
Nat ur al f r equency 3 = 17. 9239 Hz

1 2 3 4 5 6 7 8 9
0
0.005
0.01
Mode 1
Axial position
1 2 3 4 5 6 7 8 9
0
0.05
Mode 2
Axial position
1 2 3 4 5 6 7 8 9
-0.05
0
0.05
Mode 3
Axial position

Problem 9.7
Although the constraint is posed in the question in terms of angular velocity the same
constrain must exist in terms of angular position. Since the torques are specified in
terms of shafts 1 and 2, then it is sensible to write the position of shaft 3 in terms of
shafts 1 and 2. Let
1 1
q = ,
2 2
q = . The constraint is
1 2 3
5 0 + = . Thus
114

3 1 2
5 q q = + . In matrix for this becomes = Tq where
1
5
1 0
0 1
1
(
(
=
(
(

T . Now
T
R
= M T MT. Here
| |
1 2 3
diag J J J = M . Expressing the torques applied to shafts
1 and 2 as a vector we have
1
2

=
`

)
and hence
R
= M q . Thus
1
R

= q M . This
gives the accelerations of shafts 1 and 2 only. Hence to obtain the accelerations of all
three shafts we have = Tq

. These relatively simple calculations can be done by

hand but the MATLAB script Problem_09_07.m implements these calculations and
gives the following output:

Angul ar accel er at i on of shaf t 1 = 64. 5361 r ad/ s^2
Angul ar accel er at i on of shaf t 2 = 371. 134 r ad/ s^2
Angul ar accel er at i on of shaf t 3 = 9. 6907 r ad/ s^2

Problem 9.8
To refer a inertia from one shat speed to another we consider the kinetic energy of the
inertia. Let
m
I be the polar moment of inertia of the motor,
g
I be the polar moment
of inertia of the small gear and
c
I be the polar moment of inertia of the large gear and
cone combined. Let
s
be the angular velocity of the shaft and
c
be the angular
velocity of the cone. Thus the kinetic energy of the cone and large gear, is
2
1
2
c c
T I = .
Now,
c s
g = where g is the gear ratio and in this case is equal to 1 6. Thus the
kinetic energy of the cone and large gear can be expressed as
2 2
1
2
c s
T I g = . Hence
the polar moment of inertia of the cone and large gear is
2
c
I g when it is referred to
the shaft, i.e. it is assumed to be rotating at the same speed as the shaft.
If the torque acting on the crusher cone is
c
, then the work done by this torque is
c c
where
c
is the angular displacement of the cone. Since
c s
g = the work
done by the torque is
c s
g when related to the shaft. The stiffness of the shaft can be
derived from
1 2
1 1/ 1 k k k = + . The polar moment of inertia of the cone and large
gear combined is the sum of the polar moments of inertia of the individual
components. Thus, relating the cone inertia and torque to the shaft, and applying
Newtons 2
nd
Law it is easy to show that + = M K

where
2
0
,
0
(
=
(
+
(

m
m c
I
I g I
M
(
=
(

k k
k k
K and
0
c
g

=
`

)
.

The MATLAB script Problem_09_08.m implements these calculations and gives the
following output:

Ref er r ed i ner t i a of cone and l ar ge gear = 50kg m^2
equi val ent shaf t st i f f ness = 1. 2MN/ r ad
Nat ur al f r equency = 28. 4705 Hz
Mode shape = [ - 0. 040825 0. 12247]

115

Peak shaf t t or que at 10 Hz = 427. 7746 Nm
Peak shaf t t or que at 20 Hz = 740. 3462 Nm
Peak shaf t t or que at 30 Hz = 3398. 879 Nm
Peak shaf t t or que at 40 Hz = 385. 0415 Nm
Peak shaf t t or que at 50 Hz = 179. 9207 Nm

Problem 9.9
Let
e
n is the number of elements and the coordinates are numbered 1 to 1
e
n + . The
driving machine is located at coordinate 1 and the drill bit is at coordinate 1
e
n + . We
can model the drill pipe using the stiffness and inertia elements given in Equations
(4.45) and (4.48) respectively. Using these matrices we can assemble the inertia
matrix (M) and the stiffness matrix (K) for the drill string. The inertia of the driving
machine must be added to the inertia matrix at coordinate 1 and the inertia of the drill
bit is added to the inertia matrix at coordinate 1
e
n + . This analysis has been
implemented in the MATLAB script Problem_09_09.m and gives the following
output:

Fi r st f i ve nat ur al Fr equenci es ( Hz)
5 el ement s 8 el ement s 50 el ement s
0. 0000 0. 0000 0. 0000
1. 9417 1. 9225 1. 9106
4. 0719 3. 9187 3. 8232
6. 5437 6. 0625 5. 7394
9. 1547 8. 4185 7. 6613

Problem 9.10
The equations of motion for this system are ( ) t + = Mq Kq Q where M and K are
given in Problem 9.6. We now need to determine the vector of forces acting on the
system. The axial force due to the cylinders firing once per cycle is
( ) ( ) ( ) ( ) 6,000cos 10 1,500cos 30 500sin 30 f t t t t = +
Only force terms at a frequency of 15Hz (i.e. 30) will excite the system at 15 Hz,
i.e. at this frequency the force is effectively ( ) ( ) ( ) cos 30 sin 30 f t a t b t = + where
1,500 a = and 500 b = . Due to the firing timing of each cylinder the axial force
separating the masses either side of the cylinder are
( ) ( ) ( ) cos 30 sin 30 f t a t n b t n = + , where 2 5 = (i.e. 72). and n
depends on the cylinder number, see below. Expanding this force function gives
( ) ( ) ( ) { } ( ) ( ) ( ) { } ( ) cos sin cos 30 sin cos sin 30 f t a n b n t a n b n t = + + .
When the system runs at 5Hz (i.e. 10 ) then only the terms at this frequency excite
the system, i.e. the force is effectively ( ) ( ) cos 10 f t a t = where 6,000 a = . Due to
the firing timing of each cylinder the force will be ( ) ( ) cos 10 f t a t n = where n
depends on the cylinder number. Thus
( ) ( ) ( ) { } ( ) ( ) ( ) { } ( ) cos sin cos 10 sin cos sin 10 f t a n b n t a n b n t = + + where
in this case, 0 b = .
Since the firing order is 1-3-5-2-4 then 0,1, ,4 n = . Thus, for cylinders 1, 2, ,5 the
values of n are 0, 3, 1, 4 and 2 in that order.
Each coordinate (corresponding to a mass) has a force acting on each side.

116


For example, the axial force acting at coordinate 3 is ( ) ( ) ( )
3 3 4
Q t f t f t = and hence
( ) ( ) ( ) { } ( ) ( ) ( ) { } ( )
( ) ( ) { } ( ) ( ) ( ) { } ( )
3
cos 3 sin 3 cos 10 sin 3 cos 3 sin 10
cos sin cos 10 sin cos sin 10
Q t a b t a b t
a b t a b t
= + +
+
Si
mplifying this expression
( ) ( ) ( ) ( ) ( ) ( ) ( ) { } ( )
( ) ( ) ( ) ( ) ( ) ( ) { } ( )
3
cos 3 cos sin 3 sin cos
sin 3 sin cos 3 cos sin
Q t a b t
a b t
= +
+

or ( ) ( ) ( )
3 3 3
cos sin
C S
Q t Q t Q t = + .
No forces act at coordinates 7, 8 and 9. The response is determined from
( ) ( )
1 1
2 2
and
C C C C

= = q K M Q q K M Q
The response at the rim (coordinate 9) is
2 2
9 9 9 C S
q q q = + . The axial force between
the rim and hub (coordinate 8) is
2 2
89 89 89 C S
F F F = + where ( )
89 9 9 8 C C C
F k q q =
and ( )
89 9 9 8 S S S
F k q q = .
The MATLAB script Problem_09_10.m determines the system response at the rim
and the axial force between the hub and rim. The output is as follows:

At 5 Hz, Axi al r eponse at gen r i m= 5. 9878 mu_m
Axi al f or ce bet ween gen hub and r i m= 130. 0146 N

At 15 Hz, Axi al r eponse at gen r i m= 5. 0966 mu_m
Axi al f or ce bet ween gen hub and r i m= 995. 9699 N
m
3
m
2

q
2
q
3

m
4

q
4

f
2
f
3