NOISE & VIBRATION

MULTI DEGREE OF FREEDOM SYSTEMS

DR. MUHD. HAFEEZ BIN ZAINULABIDIN

FACULTY OF MECHANICAL & MANUFACTURING
ENGINEERING
UTHM
 CHAPTER 2
Multi Degree of Freedom Systems

Figure source: http://cite.iiit.ac.in/vlab/Expr6.htm

Multi Degree of Freedom Systems
1. Mass – Spring System
2. Generalized Coordinates and Generalized
Forces
3. Lagrange’s Equations
4. General Equations in Matrix Form
5. Eigenvalues and Eigenvectors
Mass – Spring System
a) Newton’s 2nd Law Approach

Figure 2(a) and 2(b)

 The application of Newton’s 2nd law of
motion to mass mi gives
mi xi  ci x i 1   ci  ci 1  x i  ci 1 x i 1  ki xi 1   ki  ki 1  xi
 ki 1 xi 1  Fi ; i  2, 3, ..., n  1

Equations of motion of the masses m1 and mn

where i = 1, xo = 0 and i = n, xn+1 = 0.
m1 x1   c1  c2  x 1  c2 x 2   k1  k 2  x1  k 2 x2  F1
mn xn  cn x n 1   cn  cn 1  x n  k n xn 1   k n  k n 1  xn  Fn
Equations of motion can be expressed in
matrix form as 
 m x   c  x   k  x  F

  

where [m], [c] and [k] are the mass, damping

and stiffness matrices, respectively.
m1 0 0 0
0 m2 0 0 


 m   0 0 m3  0 
 
    
 0 0 0  mn 
  c1  c2   c2 0  0 0 
 c c  c3   c3  0 0 
 2 2 
 0  c3 c  c4   0 0 
 c   3

 0 0  c4    
       
 
 0 0 0   cn  cn  cn 1  

 k1  k 2   k2 0  0 0 
 k k  k3   k3  0 0 
 2 2 
 0  k3 k  k4   0 0 
k   3

 0 0  k4    
       
 
 0 0 0   kn  kn  kn 1  
    
… x , x , x and F are the displacement,
velocity, acceleration and force vectors.

 x1  t    x 1  t    x1  t    F1  t  
x  t    x  t    x  t   F  t  
   2  
  
2  
  
2    2 
x    , x    , x    , F    .
           
       
 xn  t    x n  t   xn  t    Fn  t  
 In general form, the mass, damping and
stiffness matrices are
 m11 m12 m13  m1n 
m m22 m23  m2 n 
 12
 m  m13 m32 m33  m3 n 
 
     
m1n m2 n m3 n  mnn 

 c11 c12 c13  c1n   k11 k12 k13  k1n 

c c22 c23  c2 n  k k 22 k 23  k 2 n 
 12  12
 c  c13 c32 c33  c3 n   k   k13 k32 k33  k3 n 
   
         
c1n c2 n c3n  cnn  k1n k2n k3 n  k nn 
b) Influence Coefficients Approach
The equations of motion of a multi degree
of freedom system can also be written in
terms of influence coefficients.

There are 3 main influence coefficients:

i. Stiffness Influence Coefficients
ii. Flexibility Influence Coefficients
iii. Inertia Influence Coefficients
 Stiffness Influence Coefficients
 Defined as the force at point i due to a unit
displacement at point j when all the points
other than the point j are fixed.
 Total force at point i, Fi can be found by
summing up the forces due to all
displacements xnj (j = 1, 2, …, n).
Fi   kij x j , i  1, 2, ..., n
j 1

or

F  kx

Steps in determining the stiffness influence
coefficients.
1. Assume x1 = 1 and x2, x3, …, xn = 0. The set of
forces will maintain the system in the assumed
configuration.
2. Write static equilibrium equations for each
mass.
3. Solve the equations to find the influence
coefficients.
4. Repeat Step 1 to 3 by assuming x2 = 1 and so
on.
 Example:
Consider multi degree of freedom mass –
spring system shown in Figure 2(c).

Figure 2(c)
x1=1

Figure 2(d)

m1 : k1  k 2  k11
m2 : k 21   k 2
m3 : k31  0
x2=1

Figure 2(e)

m1 : k12  k 2  0
m2 : k 22  k3  k 2
m3 : k32  k3
x3=1

Figure 2(f)

m1 : k13  0
m2 : k 23  k3  0
m3 : k33  k3
Solution for x1=1
k11  k1  k 2 , k 21  k 2 , k31  0

Solution for x2=1

k12  k 2 , k 22  k 2  k3 , k32  k3

Solution for x3=1

k13  0, k 23   k3 , k33  k3

Stiffness matrix
 k1  k 2   k2 0 
 k     k2  k2  k3   k3 
 0  k3 k3 
Flexibility Influence Coefficients
 Defined as the deflection at point i due to a
unit load at point j when all the loads at
other points other than the point j are zero.
 Total deflection at point i, xi can be found by
summing up the contributions of all forces
Fj. n
xi   aij F j , i  1, 2, ..., n
j 1

 or 
x   a F
Steps in determining the flexibility influence
coefficients.
1. Assume F1 = 1 and F2, F3, …, Fn = 0.
2. Write static equilibrium equations for each mass.
3. Solve the equations to find the flexibility
coefficients.
4. Repeat Step 1 to 3 by assuming F2 = 1 and so on.
OR
5. Find inverse [k] if the stiffness matrix is available.
F1=1

Figure 2(g)
m1 : k1a11  k 2  a21  a11   1
m2 : k 2  a21  a11   k3  a31  a21 
m3 : k3  a31  a21   0
F2=1

Figure 2(h)
m1 : k1a12  k 2  a22  a12 
m2 : k 2  a22  a12   k3  a32  a22   1
m3 : k3  a32  a22   0
F3=1

Figure 2(i)

m1 : k1a13  k 2  a23  a13 

m2 : k 2  a23  a13   k3  a33  a23 
m3 : k3  a33  a23   1
Solution for F1=1
1 1 1
a11  , a21  , a31 
k1 k1 k1
Solution for F2=1
1 1 1 1 1
a12  , a22   , a32  
k1 k1 k 2 k1 k 2
Solution for F3=1
1 1 1 1 1 1
a13  , a23   , a33   
k1 k1 k 2 k1 k 2 k3

Flexibility matrix 1 1 1 
 
 k1 k1 k1 
1 1 1 1 1 
 a         
 k1  k1 k 2   k1 k 2  
1 1 1  1 1 1 
       
 k1  k1 k 2   k1 k 2 k3 
Inertia Influence Coefficients
 Defined as the set of impulses applied at point i
to produce a unit velocity at point j and zero at
every other points.
 Total impulse at point iF
,~ i can be found by
summing up the impulses causing the velocities x j
n
F   mij x j , i  1, 2, ..., n
~ i
j 1

 or 
F   m  x
~
Steps in determining the inertia influence
coefficients.
1. Assumex 1  1 andx 2 , x 3 , ..., x n  0 .
2. Write equilibrium equations for each mass.
3. Solve the equations to find the inertia
coefficients.
4. Repeat Step 1 to 3 by assuming x 2  1 and
so on.
 Generalised Coordinates &
Generalised Forces
 Equations of motion of a vibrating system can
be formulated in a number of different
coordinate systems.
 n independent coordinates are necessary to
describe the motion of a system having n d.o.f.
 Any set of n independent coordinates is called
generalised coordinates, designated by
q1, q2, …, qn.
 Generalised coordinates may be lengths, angles,
or any other set of numbers that define the
configuration of the system at any time uniquely.
 Generalised coordinates also independent of the
conditions of constraint.
 When external forces act on the system, the
configuration of the system changes.
 New configuration of the system can be
obtained by changing the generalised
coordinates qj by ∂qj, j = 1, 2, …, n.
 If Uj denotes the work done in changing the
generalised coordinate qj by the amount ∂qj,
the corresponding generalised force Qj can
be defined as
Uj
Qj  , j  1, 2, ..., n
q j

where Qj will be a force (moment) when qj

is a linear (angular) displacement.
Example: Triple Pendulum
 To illustrate the concept of generalized coordinates, consider
the triple pendulum system shown.
 The system can be specified by the six coordinates (xj, yj),
j = 1, 2, 3.
 However the coordinates are not independent but are
2 2 2
constrained by the  y1  l1
x1 relations
 x2  x1  2   y2  y1  2  l2 2
 x3  x2  2   y3  y2  2  l32
 Since the coordinates (xj, yj) are not independent, they cannot
be called generalized coordinates.

 The angular displacement j are used to specify the locations
of the masses without constraint. qj  j
 Thus, they form a set of generalized coordinates .
 Lagrange’s Equation
The equations of motion of a vibrating system can
often be derived in a simple manner in terms of
generalized coordinates by the use of Lagrange’s
equations.
Lagrange’s equations for n d.o.f system can be stated
d  T  T V
as:     Q jn  , j  1, 2, ..., n
dt  q j  q j q j

  x  y  z 
Q n
j    F xk
  q
k
 F yk
 q
k
 F zk
 q
k 

k  j j j 
Example:

The arrangement of the compressor, turbine and generator in

a thermal power plant is shown in figure above. This
arrangement can be considered as a torsional system where Ji
denote the mass moments of inertia of the three components,
Mti indicate the external moments acting on the components
and kti represent the torsional spring constants of the shaft
between the components. Derive the equations of motion of
the system using Lagrange’s equations by treating the angular
displacements of the components as generalized coordinates.
Kinetic energy of the system
1 2 1  2 1  2
T J11  J 2 2  J 3 3
2 2 2
Potential energy of the system
1 1 1
kt11  kt 2  2  1   kt 3  3   2 
2 2 2
V
2 2 2
External moments applied to the components
1  
Q1 n   M t1  M t 2 2  M t 3 3  M t1
1 1 1
 n 1  2 3
Q2  M t1  Mt2  M t3  Mt2
 2  2  2
 n 1  2  3
Q3  M t1  Mt2  M t3  M t3
 3  3  3
Substituting [1], [2] and [3] into Lagrange’s equations to
obtain 
J11   kt1  kt 2 1  kt 2 2  M t1
J 22   kt 2  kt 3  2  kt 21  kt 3 3  M t 2
J   k   k   M
3 3 t3 3 t3 2 t3

Equations ofmotion expressed in matrix form

 J1 0  
0  1  kt1  kt 2   kt 2 0  1   M t1 
0         
 J2 0   2     kt 2  kt 2  kt 3   kt 3   2   M t 2 
 0 0 J 3  3   0  kt 3 kt 3   3   M t 3 
General Equations in Matrix
Form
Equations of motion of a multi d.o.f system can be
derived in matrix form from Lagrange’s equations
d  T  T R V
      Fi , i  1, 2, ..., n
dt  x i  xi x i xi

The kinetic and potential energy in matrix form

1 T 
T  x  m x
2
1
V  xT k  x

2
Rayleigh dissipation function
1 T 
R  x  c  x
2
From the theory of matrices
T 1 T  1 T  T 
   m x  x  m    m  x
x i 2 2
T
 mi x , i  1, 2, ..., n
d  T   T  T
   mi x and  0, i  1, 2, ..., n
dt  x i  xi

V 1 T  1 T  T 
   k  x  x  k     k  x
xi 2 2
T 
 ki x , i  1, 2, ..., n
R 1 T  1 T  T 
   c  x  x  c     c  x
x i 2 2
T 
 ci x , i  1, 2, ..., n

  
Substituting
 m x   c xinto
   k  xLagrange’s

F equation to obtain
 Eigenvalues and Eigenvectors
Solution of Eigenvalue Problem
Consider equation of motion in the form of
 k     m X  0
  2

The equation can also be expressed as

  1
   k    m  X  0 where  
2
Premultiplying equation by [k]-1 to obtain
 
   I    D X  0
 D   k  1 m
where [I] is the identity matrix and

For nontrivial solution of
X
    I    D  0
Example:

Find the natural frequencies and mode shapes of the

system shown in figure above for k1 = k2 = k3 = k and m1
= m2 = m3 = m.
Solution:
The dynamical matrix is
 D   k  1 m   a m
Where the stiffness, flexibility and mass matrix can be
obtained using influence coefficient approach or
Lagrange’s1 equation
1 1 1 0 0
1
 a  1 2 2  m  m0 1 0
k
1 2 3 and 0 0 1

Thus 1 1 1
 D  m 0 2 2
k
1 2 3
Setting the characteristics determinant equal to zero to
obtain the frequency equation
 0 0  1 1 1
     I    D   0  0   m 1 2 2  0
k
 0 0   1 2 3
1
where  
 2

Dividing throughout by

1   
  1  2  2   3  5 2  6  1  0
  2 1  3

where m m 2
 
k k
The roots of cubic equation are
2
m1 k
1   0.19806, 1  0.44504
k m
m2
2
k
2   1.5553, 2  1.2471
k m
m3
2
k
3   3.2490, 3  1.8025
k m
Once the natural frequencies are known, the mode
shapes or eigenvectors can be calculated using
equation  i  I    D  X  i   0, i  1, 2, 3

where  X 1 i  
i  i 
X  X 2 
X i 
 3 
First Mode: Substituting the valueof
1  5.0489 k
m

 1
 1 0 0  1 1 1   X 1  0

5. 0489
m
0 1 0   m 1 2 2   X  1   0
 k   k   2   
 0 0 1 1 2 3   X 3 1  0

4.0489 1  1   X 1 1  0
 1 3 . 0489  2   X  1   0
  2   
  1 2 2.0489  X 3 1  0
From the first two rows of equation
X 2 1  X 3 1  4.0489 X 1 1

Solutions 3.0489 X 2 1  2 X 3 1  X 1 1

X 2 1  1.8019 X 1 1  1  1
andX 3  2.2470X 1
The mode shape
 X 1 1   1 
    
X  1   X 2 1   X 1 1 1.8019 
 X  1  2.2470
 3   
Second Mode: Substituting the value of
2  0.6430 k
m

 2
 1 0 0  1 1 1    X 1  0

0. 6430
m
0 1 0   m 1 2 2   X  2    0
 k   k   2   
 0 0 1 1 2 3   X 3 2   0

 0.3570 1  1   X 1 2   0
 1  1. 3570  2   X  2    0
  2   
  1 2  2.3570  X 3 2   0

From the first two rows of equation

 X 2 2   X 3 2   0.3570 X 1 2 

Solutions  1.3570 X 2 2   2 X 3 2   X 1 2 

X 2 2   0.4450 X 1 2  X 3 2   0.8020 X 1 2 
and
The mode shape
 X 1 2    1 
    
X  2   X 2 2    X 1 2   0.4450 
 X  2   0.8020
 3   
Third Mode: Substituting the value of
3  0.3078 k
m

 3
 1 0 0 1 1 1   X 1  0

0.3078
m
0 1 0   m 1 2 2   X  3   0
 k   k   2   
 0 0 1 1 2 3   X 3 3  0

 0.6922 1 1  X1 3  0

 1  1.6922  2  X  3   0
  2   
 1 2  2.6922 X 3 3  0

From the first two rows of equation

 X 2 3  X 3 3  0.6922 X 1 3 

Solutions  1.6922 X 2 3   2 X 3 3   X 1 3 

X 2 3   1.2468 X 1 3   3  3
andX 3  0.5544X 1
The mode shape
 X 1 3   1 
    
X  3   X 2 3   X 1 3   1.2468
 X  3   0.5544 
 3   
Mode Shapes
ALL THE BEST!