Damping of Materials and Members in Structures

Home Search Collections Journals About Contact us My IOPscience
Damping of materials and members in structures
This content has been downloaded from IOPscience. Please scroll down to see the full text.
2011 J. Phys.: Conf. Ser. 268 012022
(http://iopscience.iop.org/1742-6596/268/1/012022)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 155.185.57.71
This content was downloaded on 02/02/2017 at 21:30
Please note that terms and conditions apply.
You may also be interested in:
Full-scale tests on steel frames under cyclic loading

M Ivanyi, P Ivanyi and M M Ivanyi
Mode switching and hysteresis in the edge tone

I Vaik and G Paál
Development of a resistive model for overhead line poles

Z Kvasznicza and G Elmer
Compensation and control of two-inputs systems with hysteresis

D Davino, A Giustiniani and C Visone
Attempts at a numerical realisation of stochastic differential equations containing Preisach

operator
S McCarthy and D Rachinskii
Novel MRE/CFRP sandwich structures for adaptive vibration control

J Kozlowska, A Boczkowska, A Czulak et al.
Dynamic characterization of a laminated composite magnetorheological fluid sandwich plate

R Manoharan, R Vasudevan and A K Jeevanantham
Magnetization curve modelling of soft magnetic alloys

I Mészáros
Mechano-electric optoisolator transducer with hysteresis

I M Ciuru, M Dimian and A Graur
5th International Workshop on Multi-Rate Processes and Hysteresis (MURPHYS 2010) IOP Publishing
Journal of Physics: Conference Series 268 (2011) 012022 doi:10.1088/1742-6596/268/1/012022
Damping of materials and members in structures
F Orban
Department of Mechanical Design. Pollack Mihaly Faculty of Engineering
University of Pécs, Pécs, Hungary
E-mail: [email protected]
Abstract. The state of a structure subject to oscillatory deformation can be described by the
combination of kinetic and potential energy. In the case of real structures there is also an
energy dissipative element as some of the energy is lost per deformation cycle. The energy
dissipation is caused by material damping which basically depends on three factors: amplitude
of stress, number of cycles and geometry. In the case of non-homogeneous stress distribution
the geometry of the structure influences the vibration damping. In this paper the influence of
the geometry will be investigated with special regard to the cross-section. The examinations
can be executed experimentally, theoretically and by the help of computer programs using
FEM. In most cases the main goal is to increase the damping of the structure.
1. Measures of damping
Damping of structures is a very complex phenomenon, which refers to two basic reasons a) material
damping b) friction damping at the connections. When a structure is subject to oscillatory
deformations the state of the structure can be described by the combination of kinetic and potential
energy. In the case of real structures some of this energy is lost per deformation cycles and this is
called material damping.
Damping is the conversion of mechanical energy of a vibrating structure into thermal energy. If we
want to quantify the level of damping in a structure the absorbed energy per cycle must be determined.
By plotting the force versus displacement for a given cycle of motion a hysteresis curve is generated
(figure 1).
At linear damping the hysteresis loop is an ellipses (figure 1a). In general metals have linear
damping in the case when the stress amplitude less than the fatigue limit. At nonlinear damping the
hysteresis loop is peaked when damping is a result of friction (figure 1b) One possibility to quantify
the level of damping is to determine the area captured within the hysteresis loop
D = ∫ F ⋅ dy , (1)
then the specific damping coefficient is:
D
ψ= , (2)
U
where U is the stored energy during loading.
For an unforced damped single degree of freedom (SDOF) system the general equation of motion
becomes:
Published under licence by IOP Publishing Ltd 1

y
m ⋅ &y& + k ⋅ y& + =0, (3)
c
where m is the mass, k is viscous damping constant and c is the spring constant.
Figure 1. Hysteresis loop a) linear damping b) non linear damping.
In the under-damped case when β < α

1 k , (4)
α2 = ; β =
m ⋅c 2m
&y& + 2 β ⋅ y& + α 2 ⋅ y = 0 . (5)

The solution is given by.
y (t ) = A ⋅ e − βt ⋅ sin (γt + ε 0 ) , (6)

where
γ = α2 −β2 , (7)
A1 e − βT
= = e βT , (8)
A2 e − β (t + T )
A1 1 A
δ = ln = ln i . (9)
A2 k Ai + k
The log decrement of transient response δ at linear damping is irrespective of time. There is a
simple interrelationship between the specific damping coefficient ψ and the log decrement of transient
response δ. It is possible to express the variable ψ as
U i +T Ui
dU
Ψ =− ∫ = ln . (10)
Ui U U i +T
The potential energy for a spring is
y2
U= , (11)
2c
2
Ui Ai2 A
ψ = ln = ln = 2ln i = 2δ . (12)
U i +T 2
Ai + T Ai + T
Figure 2. Transient response of a classically under-damped system.
The level of damping can be subjectively determined by noting the sharpness of resonant peak at f1
(figure 3). For a quantitative measure of damping the Half-Power Bandwidth Method can be
employed.
Figure 3. Compliance transfer function for a SDOF system.
The damping of the structure η can be determined from the ratio ∆f to f1. The resonant peak
value is Amax and in this case the loss factor can be expressed as
∆f
η= . (13)
f1
2. Material damping
Material damping depends on many factors. The most important of these factors are: type of materials,
stress amplitude, internal forces, the number of cycles, sizes of geometry, the quality of surfaces and
temperature. The factors were examined in Lazan book [1]. Damping depends mainly on the stress
amplitude as
D = J ⋅ σ an , (14)
3
where J and n are constants.

The value of n can be between 2 and 4 but generally n = 2.3 is used. Damping increases with the
number of cycles and finally a fatigue collapse happens. The values of J are very different for the
same material according different authors.
According to [2] J = 2.326 ⋅ 10 −8 for low carbon steel and therefore
D = 2.321⋅10−8 ⋅ σ 2.3 . (15)

At the uniaxial stressed state the stored energy can be calculated:
σ2
U= , (16)
2⋅ E
D
ψ= = 9.769 ⋅10−3 ⋅ σ 0.3 . (17)
U
According to Equation (17) the values computed for ψ are summarised in table 1.
Table 1. Specific damping coefficient values refers to σMAX.
σ N/mm2 20 40 60 80 100 120 140

ψ% 2.4 2.9 3.3 3.7 3.9 4.1 4.3
The relationship between measures of damping can be expressed:

ψ δ
η= = = 2ξ . (18)
2π π
The values of the viscous damping ratio ξ is listed in table 2 according to [3].
Table 2. Dynamic properties of materials under standard conditions.
Representative Damping Ratios

System Viscous Damping Ratio ξ
Metals (in elastic range) < 0.01
Continuous Metal Structures 0.02 to 0.04
Metal Structure with Joints 0.03 to 0.07
Aluminium / Steel Transmission Lines ≈ 0.0004
Small Diameter Piping Systems 0.01 to 0.02
Large Diameter Piping Systems 0.02 to 0.03
Auto Shock Absorbers ≈ 0.30
Rubber ≈ 0.05
Large Buildings during Earthquakes 0.01 to 0.05
4
3. Material damping in the structure

If the stress distribution is not homogeneous, the geometry of the structure influences the damping
properties. The damping, which belongs to the maximum stress amplitude is
Dam = J ⋅ σ am . (19)
The structural damping can be calculated as
Ds = Dam ⋅ V ⋅ α k ⋅ α h , (20)
where V is the volume of the structure, αk is the cross section factor, αh is the length factor.
The cross section factor for a rectangular cross section (figure 4a) is
1 b/2 1
αk = ⋅ 2 ∫ a ⋅ y n ⋅ dy = , (21)
A0 ⋅ yn max y =0 n +1
where A0 = a⋅b.
Figure 4. a) rectangular, b) I-section.
In the case of an I-section the cross section factor is:

⎡ h/2 ⎛n⎞ ⎤
n
2 n n tw ⋅ h
αk = ⎢t y ⋅ dy + b ⋅ t ⋅ ⎜ ⎟ ⎥ = 1 − ⋅ . (22)
n ⎢ w ∫ ⎝ 2 ⎠ ⎥⎦ n + 1 A0
⎛h⎞ ⎣ 0
A0 ⎜ ⎟
⎝2⎠
It can be seen that the cross-section factor for an I-section is twice as big as for the rectangular section.
The length factor is calculated as
n
1l⎛σ ⎞
α h = ∫ ⎜⎜ ax ⎟⎟ dx . (23)
l 0 ⎝ σ am ⎠
A cantilever beam with concentrated load is plotted in figure 5. If the load is concentrated on the
cantilever beam then the length factor is
l n
1 ⎛x⎞ 1
αh = ∫⎜ ⎟ dx =
l 0⎝ l ⎠ n +1
. (24)
5
Figure 5. Maximum stress distribution at cantilever beam.
A cantilever beam with distributed load can be seen in figure 6.
Figure 6. Maximum stress distribution at cantilever beam.
If the load is distributed load the length factor is

n
1 l ⎛ x2 ⎞
⎟ dx = 1 .
α h = ∫ ⎜⎜ ⎟
(25)
l ⎝ l2 ⎠ 2n + 1
A simply supported beam with concentrated load is plotted in figure 7.
Figure 7. Maximum stress distribution at simply supported beam.
6
If the load is concentrated the length factor for a simple beam can be determined as
n
1 l / 2⎛ 2x ⎞ 1
αh = 2 ∫ ⎜ ⎟ dx = . (26)
l 0 ⎝ l ⎠ n +1
Simply supported beam with concentrated load is plotted in figure 8.
Figure 8. Maximum stress distribution at simply supported beam.
In this case the length factor can be determined as

n
( )
n
1l⎡ 4 ⎤ 2⎛3⎞ 1
αh = ∫ ⎢ l x − x 2 ⎥ dx = ⎜ ⎟ + . (27)
l 0⎣l2 ⎦ 3⎝4⎠ 6
4. Damping analysis with FEM

The effect of the geometry on the damping will be analyzed by finite element method (FEM) in this
section. The GEOSTAR program controls the complete operations of the COSMOSM package and all
modules are accessed from there.
The FREQUENCY/BUCKLING option of the program calculates the frequencies without damping
The POST DYNAMIC analysis (ASTAR) is used to examine the damping and for this it is necessary
to define the damping ratio ξ. Load excitations may be applied as concentrated loads at specified
nodes or as distributed pressure applied to specified element faces. The response of the structure at
certain nodes may be calculated for displacement.
The log decrement of a transient response is calculated with the help of the displacement time
curve.
1 A
δ= ln i , (28)
k Ai + k
δ = β ⋅T , (29)
δ
β= = 2πξ ⋅ f1 . (30)
T
The calculated values β and T are suitable for measuring damping. The force-time diagram shown in
figure 9 will applied as an excitation. The examined cross-sections can be seen in figure 10.
Six models were examined according to table 3.
In the examined cases the cross-section area were the same and the maximum stresses were the
same for a given beam, in the case of the cantilever beam 75 MPa and for the simply supported beam
20 MPa.
7
Figure 9. Force-time diagram.
Figure 10. Dimensions of the cross-section.

Table 3. The investigated models
Model Type of beam Loading Cross-section Max stress

Force end of the
1. cantilever 75 MPa
beam
Force
2. cantilever
end of the beam I 75 MPa
3. cantilever distributed load 75 MPa
4. simply supported distributed load 20 MPa

5. simply supported distributed load I 20 MPa
Force
6. simply supported
end of the beam I 20 MPa
Computation of the damping is illustrated in figure 11, where the beam is fixed and the cross-
section is rectangular.
The damping results for a fixed beam with an I-section can be seen in figure 12.
The results are listed in table 4.
8
Figure 11. Damped vibration for a fixed beam with rectangular cross-section.
Figure 12. Damped vibration for a fixed beam with an I-section.
Table 4. Calculated results for β eigenvalue frequency.
Model β f1 [Hz ]
1. 2.08 16.06
2. 4.05 60.2
3. 1.77 16.06
4. 5.78 45.3
5. 20.67 170
6. 17.36 170
9
The cross section factor αk for I section (figure 4b) is twice as high as for the rectangular section
(figure 4a). According to the FEM calculations the results are similar. The effects of the load
distribution are smaller.
5. Measuring the damping in a laboratory

Different methods exist to measure the characteristics of damping. In one method the starting
displacement is applied and then the system is left free to move (figure 13). In this case a damped
vibration can be observed. To determine the changing of amplitudes the log decrement δ can be
calculated, however the changing of amplitudes and strain changing near the fixed point is the same.
Strain-gauge was used to measure the strains. Figure 14 illustrates the acceleration of beam end and
strain changing. From the measured values the value of β can be calculated. The first 10 amplitudes
according to measurements are listed in table 5.
Figure 13. Arrangement of measurement.

Table 5. Measurement results.
Strain Time Acceleration

(µ/m) (s) m s-2
355.68 0.44 -.0.72
307.44 0.54 -0.1
269.04 0.64 -0.089
220.08 0.74 -0.074
268.8 0.85 -0.086
232.08 0.95 -0.075
240.96 1.07 -0.075
250.56 1.17 -0.078
247.92 127 -0.078
233.76 1.37 -0.074
10
Figure 14. Acceleration of the beam end and strains changing near the fixed point
The model of the experiment can be seen in figure 15. The parameters for the model are l =550 mm,
a= 10 mm, mbeam = 0.434 kg, m = 0.5 kg.
l3
E = 2.1 ⋅ 10 5 MP, c = , (31)
3IE
1
α= . (32)
mred ⋅ c
The eigenvalue frequency is

α
f = = 10.55 Hz, T = 0.094 s , (33)
2π
1 A1 1 355.86
δ = ln = ln = 0.04669 (34)
k A10 9 233.76
and the results of the measurements:

1.37 − 0.44
T= = 0.103 sec , (35)
9
1
f = = 9.708 Hz . (36)
T
11
The calculated and measured values are very close to each other. In this case the damping ratio
becomes:
δ⋅f
ξ= = 0.007 . (37)
α
Figure 15. The model of the experiment.
6. Calculating structural damping

Let us determine the structural damping for the model shown in figure16.
Figure 16. Dimensions of the cantilever beam.
M MAX 100
σ MAX = = = 75 N/mm 2 . (38)
W 1.333
The maximum displacement at the end of the beam is:
F ⋅l3 100 ⋅ 1000 3

y MAX = = = 12.5 mm . (39)
3IE 3 ⋅ 1.333 ⋅ 10 4 ⋅ 2.1 ⋅ 10 5
The damping ratio according to the measurements is ξ =0.007. The loss factor is η = 0.014, the log
decrement is δ = 0.044 and the absorbed energy is
1
U =W = F ⋅ y MAX = 625 N mm . (40)
2
The structural damping result is
Ds = ψ ⋅ U = 2δ ⋅ U = 55 N mm . (41)
The damping of the structure can be increased by using sandwich beams.
12
7. Sandwich beams
Comparing the damping properties of this model the dimensions of this beam are selected according to
figure 17.
Figure 17. Sandwich beam.
Calculation of the maximum stress data is shown in figure 18.
Figure 18. Stress distribution of the sandwich beam.
To calculate the normal stresses the following expressions are applied:
13
F ⋅l τ ⋅b⋅l τ ⋅ b ⋅ l ⋅ h1
σ, = , σ" = 2 , σ "' = 2 , (42)
2W1 2 A1 4W1
σ MAX = 128.9 + 24.75 − 74.25 = 79.4 N/mm . (43)

The maximum displacement at the end of beam is
y MAX = 19.52 mm .
The loss factor can be calculated using the formula for sandwich beams developed by Ungar [4] (also
in Farkas [5])
η 2 ⋅ XY
η=
( )
1 + (2 + Y ) + (1 + Y ) 1 + η 22 ⋅ X 2
, (44)
where η2 is the loss factor of the rubber, X is the shear parameter and Y is the stiffness parameter of the
laminates.
⎛C ⋅L⎞
X = g0 ⎜ d ⎟, (45)
⎝ 2π ⎠
2Gd ⋅ b
g0 = , (46)
h2 ∑ Ai Ei
d 2 ⋅ b E1 ⋅ h1 ⋅ E3 ⋅ h3
Y= ⋅ . (47)
B f E1 ⋅ h1 + E3 ⋅ h3
For steel E1 = E 2 = 2.1 ⋅ 10 5 MPa , for rubber Gd = 4.5 MPa and η2 = 1.8, thus the loss factor of a
sandwich beam is η= 0.0738. The calculated absorbed energy is
1
U =W = F ⋅ y MAX = 536.8 N mm , (48)
2
ψ = 2π ⋅η = 0.4637 . (49)
The structural damping of a sandwich beam is:
Ds = ψ ⋅ U = 248.9 N mm . (50)
The dissipative energy is more than four times higher then it is in the steel beam case, which has
similar dimensions.
8. Other structures
Damping of different structures can be investigated similarly to the beams, however it is impossible to
determine the damping of a structure theoretically. A possible way to compare different structures and
determine which structure has better damping ability is to determine the value of β or T.
Figure 19 shows the first vibration mode shape in the case of the square plate stiffened by flat ribs.
The cover plate is simply supported around the contour.
14
Figure 19. The first vibration mode shape at stiffened plates.
The calculation has been evaluated by FEM. The first natural frequency is f1 = 49.27 Hz and the
time of period is T1 = 0.02 sec. Dimensions of the stiffened plates are: cover plate is 1410x1410x2
mm; flat ribs are 1410x90x3 mm. To examine a flat plate the dimensions are 1410x1410x2.5 mm, the
first natural frequency is f1 = 6.17 Hz and the time of period is T1 = 0.16 sec. The examined two
structures have the same volume. The stiffened plate has better damping ability according expression
(30).
References
[1] Lazan B J 1968 Damping of materials and members in structural mechanics (Oxford Pergamon
Press Oxford)
[2] Ponomarjov S D 1966 Static design of mechanical structures (in Hungarian) (Műszaki
Könyvkiaidó Budapest)
[3] Cremer L and Heckl M 1988 Structure borne sound (Springer Verlag New York)
[4] Ungar E E 1962 Loss factors of visco-elastically damped beam structures J Acoust. Soc. Amer
18 1082
[5] Farkas J 1984 Optimum design of metal structures (Budapest. Akadémiai Kiadó Chichester:
Ellis Horwood)
15

Damping of Materials and Members in Structures

Uploaded by

Copyright:

Available Formats

Damping of Materials and Members in Structures

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Damping of Materials and Members in Structures

Uploaded by

Copyright:

Available Formats

How does the paper quantify damping in structures?

How does the paper quantify damping in structures?

What are some factors that influence damping according to the paper?

What are some factors that influence damping according to the paper?

Home Search Collections Journals About Contact us My IOPscience

Damping of materials and members in structures

2011 J. Phys.: Conf. Ser. 268 012022

Please note that terms and conditions apply.

You may also be interested in:

Full-scale tests on steel frames under cyclic loading

Mode switching and hysteresis in the edge tone

Development of a resistive model for overhead line poles

Compensation and control of two-inputs systems with hysteresis

Attempts at a numerical realisation of stochastic differential equations containing Preisach

Novel MRE/CFRP sandwich structures for adaptive vibration control

Dynamic characterization of a laminated composite magnetorheological fluid sandwich plate

Magnetization curve modelling of soft magnetic alloys

Mechano-electric optoisolator transducer with hysteresis

Damping of materials and members in structures

Published under licence by IOP Publishing Ltd 1

Figure 1. Hysteresis loop a) linear damping b) non linear damping.

In the under-damped case when β < α

&y& + 2 β ⋅ y& + α 2 ⋅ y = 0 . (5)

y (t ) = A ⋅ e − βt ⋅ sin (γt + ε 0 ) , (6)

The potential energy for a spring is

Figure 2. Transient response of a classically under-damped system.

Figure 3. Compliance transfer function for a SDOF system.

where J and n are constants.

D = 2.321⋅10−8 ⋅ σ 2.3 . (15)

Table 1. Specific damping coefficient values refers to σMAX.

σ N/mm2 20 40 60 80 100 120 140

The relationship between measures of damping can be expressed:

Table 2. Dynamic properties of materials under standard conditions.

Representative Damping Ratios

3. Material damping in the structure

Figure 4. a) rectangular, b) I-section.

In the case of an I-section the cross section factor is:

Figure 5. Maximum stress distribution at cantilever beam.

A cantilever beam with distributed load can be seen in figure 6.

Figure 6. Maximum stress distribution at cantilever beam.

If the load is distributed load the length factor is

A simply supported beam with concentrated load is plotted in figure 7.

Figure 7. Maximum stress distribution at simply supported beam.

Figure 8. Maximum stress distribution at simply supported beam.

In this case the length factor can be determined as

4. Damping analysis with FEM

Figure 9. Force-time diagram.

Figure 10. Dimensions of the cross-section.

Model Type of beam Loading Cross-section Max stress

3. cantilever distributed load 75 MPa

4. simply supported distributed load 20 MPa

Figure 12. Damped vibration for a fixed beam with an I-section.

Table 4. Calculated results for β eigenvalue frequency.

5. Measuring the damping in a laboratory

Figure 13. Arrangement of measurement.

Strain Time Acceleration

The eigenvalue frequency is

and the results of the measurements:

Figure 15. The model of the experiment.