EFFECTIVE MODAL MASS & MODAL PARTICIPATION FACTORS Revision E

By Tom Irvine
Email: tomirvine@aol.com

October 6, 2010
_______________________________________________________________________

Introduction

The effective modal mass provides a method for judging the “significance” of a vibration
mode.

Modes with relatively high effective masses can be readily excited by base excitation.
On the other hand, modes with low effective masses cannot be readily excited in this
manner.

Consider a modal transient or frequency response function analysis via the finite element
method. Also consider that the system is a multi-degree-of-freedom system. For brevity,
only a limited number of modes should be included in the analysis.

How many modes should be included in the analysis? Perhaps the number should be
enough so that the total effective modal mass of the model is at least 90% of the actual
mass.

Definitions
The equation definitions in this section are taken from Reference 1.
Consider a discrete dynamic system governed by the following equation

M x  K x  F (1)

where

M is the mass matrix

K is the stiffness matrix
x is the acceleration vector
x is the displacement vector
F is the forcing function or base excitation function

1
A solution to the homogeneous form of equation (1) can be found in terms of eigenvalues
and eigenvectors. The eigenvectors represent vibration modes.
Let  be the eigenvector matrix.

The system’s generalized mass matrix m̂ is given by

m̂  T M  (2)

Let r be the influence vector which represents the displacements of the masses resulting
from static application of a unit ground displacement.

Define a coefficient vector L as

L  T M r (3)

The modal participation factor matrix  i for mode i is

Li
i (4)
m̂ i i

The effective modal mass m eff , i for mode i is

Li 2
m eff , i  (5)
m̂ i i

Example
Consider the two-degree-of-freedom system shown in Figure 1, with the parameters
shown in Table 1.

2
 x2
m2

k3

x1
k2 m1

y
k1

Figure 1.

Table 1. Parameters
Variable Value
m1 2.0 kg
m2 1.0 kg
k1 1000 N/m
k2 2000 N/m
k3 3000 N/m

The homogeneous equation of motion is

m1 0   x1  k1  k3  k3   x1  0

 0 m  x     k 
k 2  k3  x 2  0
(6)
 2  2  3

The mass matrix is

 2 0
M   kg (7)
0 1 

3
The stiffness matrix is

 4000  3000
K   N/m (8)
 3000 5000 

The eigenvalues and eigenvectors can be found using the method in Reference 2.

The eigenvalues are the roots of the following equation.

det K  2M  0 (9)

 

The eigenvalues are

12  901.9 rad / sec2 (10)

1  30.03 rad / sec (11)

f 1  4.78 Hz (12)

22  6098 rad / sec2 (13)

2  78.09 rad / sec (14)

f 2  12.4 Hz (15)

The eigenvector matrix is

0.6280  0.3251
  (16)
0.4597 0.8881 

The eigenvectors were previously normalized so that the generalized mass is the identity
matrix.

m̂  T M  (17)

4
  0.6280 0.4597 2 0 0.6280  0.3251
m̂      (18)
 0.3251 0.8881 0 1 0.4597 0.8881 

 0.6280 0.4597 1.2560  0.6502

m̂     (19)
 0.3251 0.8881 0.4597 0.8881 

1 0
m̂    (20)
0 1

Again, r is the influence vector which represents the displacements of the masses
resulting from static application of a unit ground displacement. For this example, each
mass simply has the same static displacement as the ground displacement.

1
r (21)
1

The coefficient vector L is

L  T M r (22)

 0.6280 0.4597 2 0 1

L    (23)
 0.3251 0.8881 0 1 1

 0.6280 0.4597  2
L  1 (24)
 0.3251 0.8881  

 1.7157 
L  kg (25)
 0.2379

5
The modal participation factor  i for mode i is

Li
i (26)
m̂ i i

The modal participation vector is thus

 1.7157 
  (27)
 0.2379

The coefficient vector L and the modal participation vector  are identical in this
example because the generalized mass matrix is the identity matrix.

The effective modal mass m eff , i for mode i is

Li 2
m eff , i  (28)
m̂ i i

For mode 1,

m eff , 1 
1.7157 kg2 (29)
1 kg

m eff , 1  2.944 kg (30)

For mode 2,

m eff , 2 
 0.2379 kg2 (31)
1 kg

m eff , 2  0.056 kg (32)

Note that

m eff , 1  m eff , 2  2.944 kg  0.056 kg (33)

6
 m eff , 1  m eff , 2  3 kg (34)

Thus, the sum of the effective masses equals the total system mass.

Also, note that the first mode has a much higher effective mass than the second mode.

Thus, the first mode can be readily excited by base excitation. On the other hand, the
second mode is negligible in this sense.

From another viewpoint, the center of gravity of the first mode experiences a significant
translation when the first mode is excited.

On the other hand, the center of gravity of the second mode remains nearly stationary
when the second mode is excited.

Each degree-of-freedom in the previous example was a translation in the X-axis. This
characteristic simplified the effective modal mass calculation.

In general, a system will have at least one translation degree-of-freedom in each of three
orthogonal axes. Likewise, it will have at least one rotational degree-of-freedom about
each of three orthogonal axes. The effective modal mass calculation for a general system
is shown by the example in Appendix A. The example is from a real-world problem.

Aside

An alternate definition of the participation factor is given in Appendix B.

References

1. M. Papadrakakis, N. Lagaros, V. Plevris; Optimum Design of Structures under

Seismic Loading, European Congress on Computational Methods in Applied
Sciences and Engineering, Barcelona, 2000.
2. T. Irvine, The Generalized Coordinate Method For Discrete Systems,
Vibrationdata, 2000.
3. W. Thomson, Theory of Vibration with Applications 2nd Edition, Prentice Hall,
New Jersey, 1981.
4. T. Irvine, Bending Frequencies of Beams, Rods, and Pipes, Rev M, Vibrationdata,
2010.

5. T. Irvine, Rod Response to Longitudinal Base Excitation, Steady-State and

Transient, Rev B, Vibrationdata, 2009.

7
 APPENDIX A
Equation of Motion

m, J
kz1

kz2

kx1 0

kx2
ky2
ky1

kz3

kz4

kx3 ky3 kx4 ky4

Figure A-1. Isolated Avionics Component Model

The mass and inertia are represented at a point with the circle symbol. Each isolator is
modeled by three orthogonal DOF springs. The springs are mounted at each corner. The
springs are shown with an offset from the corners for clarity. The triangles indicate fixed
constraints. “0” indicates the origin.

8
 y

a1 a2

C. G.

b 0

c1

c2

Figure A-2. Isolated Avionics Component Model with Dimensions

All dimensions are positive as long as the C.G. is “inside the box.” At least one
dimension will be negative otherwise.

9
The mass and stiffness matrices are shown in upper triangular form due to symmetry.

m 0 0 0 0 0
 m 0 0 0 0 

 m 0 0 0
M 
 Jx 0 0
 Jy 0
 
 J z 

(A-1)

 4k x 0 0 0 2k x  c1  c2  4k x b 
 4k y 0 2k y  c1  c2  0 2k y  a1  a 2  
 
 4k z  4k z b 2k z  a1  a 2  0 
 
K = 4k z b 2  2k y  c12  c22  2k z  a1  a 2 b k y  a1  a 2 c1  c2  
   
 4k x c2 2  2k z  a12  a 22  2k x  c1  c2 b 
   
 4k x b  2k y  a1  a 2  
2  2 2 
   

(A-2)

10
The equation of motion is

 x   x  0 
 y   y  0 
     
 z   z  0 
M  K   
      0 
      0 
      
      0
(A-3)

The variables α, β, and θ represent rotations about the X, Y, and Z axes, respectively.

Example

A mass is mounted to a surface with four isolators. The system has the following properties.

M = 4.28 lbm
Jx = 44.9 lbm in^2
Jy = 39.9 lbm in^2
Jz = 18.8 lbm in^2
kx = 80 lbf/in
ky = 80 lbf/in
kz = 80 lbf/in
a1 = 6.18 in
a2 = -2.68 in
b = 3.85 in
c1 = 3. in
c2 = 3. in

11
Let r be the influence matrix which represents the displacements of the masses resulting from
static application of unit ground displacements and rotations. The influence matrix for this
example is the identity matrix provided that the C.G is the reference point.

1 0 0 0 0 0
0 1 0 0 0 0

0 0 1 0 0 0
r 
0 0 0 1 0 0
0 0 0 0 1 0
 
0 0 0 0 0 1
(A-4)

The coefficient matrix L is

L  T M r (A-5)

The modal participation factor matrix  i for mode i at dof j is

Li j
ij  (A-6)
m̂ i i

Each m̂ i i coefficient is 1 if the eigenvectors have been normalized with respect to the mass
matrix.

The effective modal mass m eff , i vector for mode i and dof j is

m eff , i j 
 L i j 2
(A-7)
m̂ i i

The natural frequency results for the sample problem are calculated using the program:
six_dof_iso.m.

The results are given in the next pages.

12
six_dof_iso.m ver 1.2 March 31, 2005

by Tom Irvine Email: tomirvine@aol.com

This program finds the eigenvalues and eigenvectors for a

six-degree-of-freedom system.

Refer to six_dof_isolated.pdf for a diagram.

The equation of motion is: M (d^2x/dt^2) + K x = 0

Enter m (lbm)
4.28
Enter Jx (lbm in^2)
44.9
Enter Jy (lbm in^2)
39.9
Enter Jz (lbm in^2)
18.8

Note that the stiffness values are for individual springs

Enter kx (lbf/in)
80

Enter ky (lbf/in)
80

Enter kz (lbf/in)
80

Enter a1 (in)
6.18

Enter a2 (in)
-2.68

Enter b (in)
3.85

Enter c1 (in)
3

13
Enter c2 (in)
3

The mass matrix is

m =

0.0111 0 0 0 0 0
0 0.0111 0 0 0 0
0 0 0.0111 0 0 0
0 0 0 0.1163 0 0
0 0 0 0 0.1034 0
0 0 0 0 0 0.0487

The stiffness matrix is

k =

1.0e+004 *

0.0320 0 0 0 0 0.1232
0 0.0320 0 0 0 -0.1418
0 0 0.0320 -0.1232 0.1418 0
0 0 -0.1232 0.7623 -0.5458 0
0 0 0.1418 -0.5458 1.0140 0
0.1232 -0.1418 0 0 0 1.2003

Eigenvalues

lambda =

1.0e+005 *

0.0213 0.0570 0.2886 0.2980 1.5699 2.7318

14
 Natural Frequencies =
1. 7.338 Hz
2. 12.02 Hz
3. 27.04 Hz
4. 27.47 Hz
5. 63.06 Hz
6. 83.19 Hz

Modes Shapes (rows represent modes)

x y z alpha beta theta

1. 5.91 -6.81 0 0 0 -1.42
2. 0 0 8.69 0.954 -0.744 0
3. 7.17 6.23 0 0 0 0
4. 0 0 1.04 -2.26 -1.95 0
5. 0 0 -3.69 1.61 -2.3 0
6. 1.96 -2.25 0 0 0 4.3

Participation Factors (rows represent modes)

x y z alpha beta theta

1. 0.0656 -0.0755 0 0 0 -0.0693
2. 0 0 0.0963 0.111 -0.0769 0
3. 0.0795 0.0691 0 0 0 0
4. 0 0 0.0115 -0.263 -0.202 0
5. 0 0 -0.0409 0.187 -0.238 0
6. 0.0217 -0.025 0 0 0 0.21

Effective Modal Mass (rows represent modes)

x y z alpha beta theta

1. 0.0043 0.00569 0 0 0 0.0048
2. 0 0 0.00928 0.0123 0.00592 0
3. 0.00632 0.00477 0 0 0 0
4. 0 0 0.000133 0.069 0.0408 0
5. 0 0 0.00168 0.035 0.0566 0
6. 0.000471 0.000623 0 0 0 0.0439

Total Modal Mass

0.0111 0.0111 0.0111 0.116 0.103 0.0487

15
 APPENDIX B

The following definition is taken from Reference 3. Note that the mode shape functions are
unscaled. Hence, the participation factor is unscaled.

Modal Participation Factor

Consider a beam of length L loaded by a distributed force p(x,t).

Consider that the loading per unit length is separable in the form

P
p( x , t )  o p( x ) f ( t ) (B-1)
L

The modal participation factor  i for mode i is defined as

1 L
L 0
i  p( x ) i ( x ) dx (B-2)

where

i ( x ) is the normal mode shape for mode i

16
 APPENDIX C

The following convention appears to be more useful than that given in Appendix B.

Modal Participation Factor for a Beam

Let

Yn ( x ) = mass-normalized eigenvectors
m(x) = mass per length

The participation factor is

L
n   m( x ) Yn ( x )dx (C-1)
0

The effective modal mass is

2
 L 
 0 m( x ) Yn ( x )dx 
m eff , n    (C-2)
0 m(x) Yn (x) dx
L 2

The eigenvectors should be normalized such that

m( x ) Yn ( x )2 dx  1
L
0 (C-3)

Thus,
2
 L 
m eff , n  n 2    m( x ) Yn ( x )dx  (C-4)
 0 

17
 APPENDIX D

Effective Modal Mass Values for Bernoulli-Euler Beams

The results are calculated using formulas from Reference 4. The variables are

E = is the modulus of elasticity

I = is the area moment of inertia
L = is the length
 = is (mass/length)

Table D-1. Bending Vibration, Beam Simply-Supported at Both Ends

Natural Participation
Mode Effective Modal Mass
Frequency n Factor

2 EI 2 8
L
1 2 L
L2   2

2 EI
2 4 0 0
L2 

2 EI 2 8
L
3 9 2 L
L2  3 9 2

2 EI
4 16 0 0
L2 

2 EI 2 8
L
5 25 2 L
L2  5 25 2

2 EI
6 36 0 0
L2 

2 EI 2 8
L
7 49 2 L
L2  7 49 2

95% of the total mass is accounted for using the first seven modes.

18
 Table D-2. Bending Vibration, Fixed-Free Beam
Natural Participation Effective Modal
Mode
Frequency n Factor Mass

2
1 1.87510  EI
0.7830 L 0.6131 L
 L  
 

2
2  4.69409  EI
0.4339 L 0.1883 L
 L  
 

2
3  5  EI
0.2544 L 0.06474 L
 2L  
 

2
4  7  EI
0.1818 L 0.03306 L
 2L  
 

90% of the total mass is accounted for using the first four modes.

19
 APPENDIX E

Rod, Longitudinal Vibration

The results are taken from Reference 5.

Table E-1. Longitudinal Vibration of a Rod, Fixed-Free

Participation Effective Modal
Mode Natural Frequency n
Factor Mass
2 8
1 0.5  c / L 2L L
 2
2 8
2 1.5  c / L 2L L
3 9 2
2 8
3 2.5  c / L 2L L
5 25 2

The longitudinal wave speed c is

E
c (E-1)


93% of the total mass is accounted for by using the first three modes.

20