ModalMass PDF
ModalMass PDF
ModalMass PDF
Revision I
By Tom Irvine
Email: tom@vibrationdata.com
December 11, 2015
_______________________________________________________________________
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
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 systems generalized mass matrix m
is given by
T M
m
(2)
Let r be the influence vector which represents the displacements of the masses resulting
from static application of a unit ground displacement. The influence vector induces a
rigid body motion in all modes.
Define a coefficient vector L as
L T M r
(3)
Li
ii
m
(4)
m eff , i
Li 2
(5)
ii
m
i i = 1 for each index if the eigenvectors have been normalized with respect
Note that m
to the mass matrix.
i j , i j ) terms are zero regardless of the
Furthermore, the off-diagonal modal mass ( m
normalization and even if the physical mass matrix M has distributed mass. This is due to
the orthogonality of the eigenvectors. The off-diagonal modal mass terms do not appear
in equation (5), however. An example for a system with distributed mass is shown in
Appendix F.
Example
Consider the two-degree-of-freedom system shown in Figure 1, with the parameters
shown in Table 1.
x2
m2
k3
k2
x1
m1
k1
Figure 1.
Table 1. Parameters
Variable
Value
m1
2.0 kg
m2
k1
k2
1.0 kg
k3
1000 N/m
2000 N/m
3000 N/m
k3 x1 0
m1 0 x1 k1 k3
0 m x k
3 k 2 k3 x 2 0
2 2
(6)
2 0
M
kg
0
1
(7)
4000 3000
K
N/m
3000 5000
(8)
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)
(10)
(11)
f 1 4.78 Hz
(12)
(13)
(14)
f 2 12.4 Hz
(15)
0.6280 0.3251
0.4597 0.8881
(16)
The eigenvectors were previously normalized so that the generalized mass is the identity
matrix.
T M
m
(17)
(18)
(19)
1 0
m
0 1
(20)
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
1
(21)
(22)
0.6280 0.4597 2 0 1
L
0.3251 0.8881 0 1 1
(23)
0.6280 0.4597 2
L
0.3251 0.8881 1
1.7157
L
kg
0.2379
(24)
(25)
Li
ii
m
(26)
1.7157
0.2379
(27)
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
m eff , i
Li 2
(28)
ii
m
For mode 1,
m eff , 1
1.7157 kg2
(29)
1 kg
m eff , 1 2.944 kg
(30)
For mode 2,
m eff , 2
0.2379 kg2
1 kg
m eff , 2 0.056 kg
(31)
(32)
Note that
(33)
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.
6. T. Irvine, Longitudinal Vibration of a Rod via the Finite Element Method, Revision B,
Vibrationdata, 2008.
APPENDIX A
Equation of Motion, Isolated Avionics Component
y
x
z
m, J
kz1
kz2
kx1
kx2
ky2
ky1
kz3
kz4
kx3
ky3
kx4
ky4
y
x
z
a1
a2
C. G.
0
c1
c2
10
At least one
The mass and stiffness matrices are shown in upper triangular form due to symmetry.
m 0 0 0
m 0 0
m 0
M
Jx
0
0
0
0
Jy
0
0
0
0
0
J z
(A-1)
4k x
K =
4k y
0
4k z
2k x c1 c 2
2k y c1 c 2
4k z b
2
4k z b 2k y c12 c 2 2
2k z a 1 a 2
2k z a1 a 2 b
2k x c12 c 2 2 2k z a12 a 2 2
2k y a1 a 2
k y a1 a 2 c1 c 2
2k x c1 c 2 b
2
2
2
4k x b 2k y a1 a 2
4k x b
(A-2)
11
x
y
z
M K
x 0
y 0
z 0
0
0
0
(A-3)
The variables , and 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.
= 4.28 lbm
Jx
Jy
Jz
kx
= 80 lbf/in
ky
= 80 lbf/in
kz
= 80 lbf/in
a1
= 6.18 in
a2
= -2.68 in
= 3.85 in
c1
= 3. in
c2
= 3. in
12
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
r
0
0
0 0 0 0 0
1 0 0 0 0
0 1 0 0 0
0 0 1 0 0
0 0 0 1 0
0 0 0 0 1
(A-4)
(A-5)
Lij
(A-6)
ii
m
i i coefficient is 1 if the eigenvectors have been normalized with respect to the mass
Each m
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)
ii
m
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.
13
six_dof_iso.m
by Tom Irvine
ver 1.2
Email: tomirvine@aol.com
M (d^2x/dt^2) + K x = 0
m (lbm)
Jx (lbm in^2)
Jy (lbm in^2)
Jz (lbm in^2)
14
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
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
0.0570
0.2886
0.2980
1.5699
2.7318
15
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)
1.
2.
3.
4.
5.
6.
x
5.91
0
7.17
0
0
1.96
y
-6.81
0
6.23
0
0
-2.25
alpha
0
8.69
0
1.04
-3.69
0
beta
0
0.954
0
-2.26
1.61
0
theta
0
-0.744
0
-1.95
-2.3
0
-1.42
0
0
0
0
4.3
1.
2.
3.
4.
5.
6.
x
0.0656
0
0.0795
0
0
0.0217
y
-0.0755
0
0.0691
0
0
-0.025
z
0
0.0963
0
0.0115
-0.0409
0
alpha
beta
theta
0
-0.0693
-0.0769
0
0
0
-0.202
0
-0.238
0
0
0.21
0
0.111
0
-0.263
0.187
0
x
0.0043
0
0.00632
0
0
0.000471
y
0.00569
0
0.00477
0
0
0.000623
z
0
0.00928
0
0.000133
0.00168
0
alpha
0
0.0123
0
0.069
0.035
0
0.0111
0.0111
0.116
beta
theta
0
0.00592
0
0.0408
0.0566
0
0.0048
0
0
0
0
0.0439
0.103
16
0.0487
APPENDIX B
P
p( x , t ) o p( x ) f ( t )
L
(B-1)
1 L
p( x ) i ( x ) dx
L 0
(B-2)
where
i ( x )
17
APPENDIX C
Yn (x)
mass-normalized eigenvectors
m(x)
L
n m( x ) Yn ( x )dx
0
(C-1)
0 m( x ) Yn ( x )dx
m eff , n
L
2
0 m(x) Yn (x) dx
(C-2)
m( x ) Yn ( x )2 dx 1
(C-3)
Thus,
m eff , n n 2 m( x ) Yn ( x )dx
0
18
2
(C-4)
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 length
= is (mass/length)
Natural
Participation
Factor
EI
2
2 L
8
L
2
EI
EI
2
2 L
3
8
L
9 2
EI
EI
2
2 L
5
EI
EI
2
2 L
7
Frequency n
2
L2
4
2
L2
2
16
25
36
49
L2
2
L2
2
L2
2
L2
2
L2
8
25 2
8
49 2
95% of the total mass is accounted for using the first seven modes.
19
Natural
Frequency n
1.87510
L
4.69409
L
5
2L
7
2L
Participation
Factor
Effective Modal
Mass
EI
0.7830
0.6131 L
EI
0.4339
0.1883 L
EI
0.2544
0.06474 L
EI
0.1818
0.03306 L
90% of the total mass is accounted for using the first four modes.
20
APPENDIX E
Natural Frequency n
Participation
Factor
0.5 c / L
1.5 c / L
2.5 c / L
2
2L
2
2L
3
2
2L
5
Effective Modal
Mass
8
L
2
8
L
9 2
8
L
25 2
(E-1)
93% of the total mass is accounted for by using the first three modes.
21
APPENDIX F
E1
N1
E2
N2
E3
E4
N4
N3
N5
Figure F-1.
U(0) = 0
(Fixed end)
(F-1)
dU
0
dx x L
(Free end)
(F-2)
The natural frequencies and modes are determined using the finite element method in
Reference 6.
22
The resulting eigenvalue problem for the constrained system has the following mass and
stiffness matrices as calculated via Matlab script: rod_FEA.m.
Mass =
0.0016
0.0004
0
0
0.0004
0.0016
0.0004
0
0
0.0004
0.0016
0.0004
0
0
0.0004
0.0008
-0.6545
1.3090
-0.6545
0
0
-0.6545
1.3090
-0.6545
0
0
-0.6545
0.6545
Stiffness =
1.0e+006 *
1.3090
-0.6545
0
0
fn(Hz)
1029.9
3248.8
5901.6
8534.3
14.8349
11.3542
-6.1448
-16.0572
18.0062
-13.7813
-7.4584
19.4897
-9.1435
16.8950
-22.0744
23.8931
23
Let r be the influence vector which represents the displacements of the masses resulting from
static application of a unit ground displacement.
The influence vector for the sample problem is
r =
1
1
1
1
(F-3)
where
mass matrix
L =
0.0867
0.0233
0.0086
-0.0021
Li
ii
m
(F-4)
i i = 1 for each index since the eigenvectors have been previously normalized
Note that m
with respect to the mass matrix.
24
iLi
(F-5)
m eff , i
Li 2
(F-6)
ii
m
m eff , i L i 2
(F-7)
m eff =
0.0075
0.0005
0.0001
0.0000
The models total modal mass is 0.0081 lbf sec^2/in. This is equivalent to 3.14 lbm.
The true mass or the rod is 3.77 lbm.
Thus, the four-element model accounts for 83% of the true mass. This percentage can be
increased by using a larger number of elements with corresponding shorter lengths.
25
APPENDIX G
L1
L2
k1
k2
Figure G-1.
k 1 ( y - x - L1 )
)
k 2 ( y - x + L2 )
)
Figure G-2.
The free-body diagram is given in Figure G-2.
26
L1 L 2 .
is the mass
ki
zi
27
Sign Convention:
Translation: upward in vertical axis is positive.
Rotation: clockwise is positive.
Sum the forces in the vertical direction
F m x
(G-1)
m x k1 ( y x L1) k 2 ( y x L 2)
(G-2)
m x k1 (y x L1) k 2 (y x L 2) 0
(G-3)
(G-4)
m x k1 k 2 x (k1L1 k 2 L 2 ) (k1 k 2 ) y
(G-5)
M J
(G-6)
J k1L1 ( y x L1) k 2 L 2 ( y x L 2)
(G-7)
J k1L1 ( y x L1) k 2 L 2 ( y x L 2) 0
(G-8)
(G-9)
(G-10)
m 0 x k 1 k 2
0 J k L k L
1 1
2 2
k 1L1 k 2 L 2 x k 1 k 2
k 1L12 k 2 L 2 2 k 1L1 k 2 L 2
(G-11)
28
k 1 k 2
k 1L1 k 2 L 2
k 1L1 k 2 L 2 x k 1 k 2
k 1L12 k 2 L 2 2 k 1L1 k 2 L 2
(G-12)
k 1 k 2
k 1L1 k 2 L 2
k 1L1 k 2 L 2 r1 k 1 k 2
k 1L12 k 2 L 2 2 r2 k 1L1 k 2 L 2
r1 1
r 0
2
(G-13)
(G-14)
(G-15)
x=z+ y
(G-16)
m 0 z m y
0 J 0
k 1 k 2
k 1L1 k 2 L 2
k 1L1 k 2 L 2 z k 1 k 2
k 1L12 k 2 L 2 2 k 1L1 k 2 L 2
k 1L1 k 2 L 2 y
k 1L12 k 2 L 2 2 0
k 1 k 2
y
k 1L1 k 2 L 2
(G-17)
m 0 z k 1 k 2
0 J k L k L
1 1
2 2
k 1L1 k 2 L 2 z
m y
2
2
k 1L1 k 2 L 2
0
29
(G-18)
The equation is more formally
m 0 z k 1 k 2
0 J k L k L
1 1
2 2
k 1L1 k 2 L 2 z
m 0 r1
r y
2
2
0
J
k 1L1 k 2 L 2
2
(G-19)
using the
1
z
2
1 k 1 k 2
m 0
0 J
2 k 1L1 k 2 L 2
(G-20)
k 1L1 k 2 L 2
1
m 0 r1
0 J r y
k 1L12 k 2 L 2 2
2
2
(G-21)
T m0
1
0
T k 1 k 2
J
2
k 1L1 k 2 L 2
k 1L1 k 2 L 2
1
2
2
k 1L1 k 2 L 2
2
m 0 r1
T
y
0 J r2
(G-22)
30
1 12
1 0
0 1
2 0
m0
0 1
T
2
2 2
0 r 1
y
J r 2
(G-23)
m 0 r 1
T
0 J r 2
(G-24)
m 0 1
m
T
T
0 J 0
0
(G-25)
Example
Consider the system in Figure G-1. Assign the following values. The values are based on a
slender rod, aluminum, diameter =1 inch, total length=24 inch.
Table G-1. Parameters
Variable
Value
18.9 lbm
k1
20,000 lbf/in
k2
20,000 lbf/in
L1
8 in
L2
16 in
The following parameters were calculated for the sample system via a Matlab script.
The mass matrix is
31
m =
0.0490
0
0
2.3497
160000
6400000
Natural Frequencies =
133.8 Hz
267.9 Hz
Modes Shapes (column format) =
-4.4
0.1486
1.029
0.6352
Participation Factors =
0.2156
0.0504
Effective Modal Mass
0.0465
0.0025
The total modal mass is 0.0490 lbf sec^2/in, equivalent to18.9 lbm.
32
APPENDIX H
L
L1
L2
k1
k2
Figure H-1.
x1
k 1 ( y - x1 )
k 2 ( y x1 + L)
Figure H-2.
The free-body diagram is given in Figure H-2. Again, the displacement and rotation are
referenced to the left end.
33
Sign Convention:
Translation: upward in vertical axis is positive.
Rotation: clockwise is positive.
Sum the forces in the vertical direction
F m x
(H-1)
m x k1 ( y x1 ) k 2 ( y x1 L)
(H-2)
m x k1 (y x1 ) k 2 ( y x1 L) 0
(H-3)
m x k1y k1x1 k 2 y k 2 x1 k 2 L 0
(H-4)
m x k1 k 2 x1 k 2 L (k1 k 2 ) y
x x1 L1
(H5)
(H-6)
m x1 L1 k1 k 2 x1 k 2 L (k1 k 2 ) y
m x1 m L1 k1 k 2 x1 k 2 L (k1 k 2 ) y
(H-7)
(H-8)
M1 J1
(H-9)
J 1 k 2 L ( y - x1 + L ) mL1 x - x1
(H-10)
J 1 mL1 x - x1 k 2 L ( y - x1 + L ) 0
(H-11)
J 1 mL1 x - x1 k 2 Ly - k 2 Lx1 k 2 L2 0
(H-12)
J 1 mL1 x - x1 - k 2 Lx1 k 2 L2 k 2 Ly
(H-13)
34
x x1 L1
(H-14)
J 1 mL1 x1 L1 - x1 - k 2 Lx1 k 2 L2 k 2 Ly
J 1 mL1 x1 - k 2 Lx1 k 2 L2 k 2 Ly
(H-15)
(H-16)
mL1 x1 k 1 k 2
m
mL
J1 k 2 L
1
k 2 L x1 k 1 k 2
k 2 L2 k 2 L
(H-17)
Note that
2
J1 J mL1
mL1 x1 k 1 k 2
m
mL
J
mL
1
1 k 2 L
(H-18)
k 2 L x1 k 1 k 2
k 2 L2 k 2 L
(H-19)
k 1 k 2
k 2L
k 2 L x1 k 1 k 2
k 2 L2 k 2 L
(H-20)
k 1 k 2
k 2L
k 2 L r1 k 1 k 2
k 2 L2 r2 k 2 L
35
(H-21)
r 1 1
r 0
2
(H-22)
4.0527
0.6352
Participation Factors =
0.2155
-0.05039
Effective Modal Mass =
0.04642
0.002539
The total modal mass is 0.0490 lbf sec^2/in, equivalent to18.9 lbm.
36