Professional Documents
Culture Documents
3
Paper ID JTE100121
Available online at: www.astm.org
Danilo Cambiaghi,1 Andrea Magalini,2 Giorgio Ramorino,3 Theonis Ricc,4 and David Vetturi5
ABSTRACT: This paper develops a methodology to test and evaluate the dynamic properties of elastomeric compounds and of rubber to metal
devices by means of an electrodynamic shaker. The dynamic tensile and shear moduli of elastomeric materials and the dynamic stiffness of rubber
components are evaluated under different conditions of excitation frequency, temperature, dynamic deformation, and static load.
KEYWORDS: rubbers, dynamic behavior, nonlinearity
Introduction
Associated with automotive industrial applications, vibrations are
considered unwanted dynamic motions. They often lead to material
fatigue, decreased component durability, structural damage and
failure, deterioration of systems performance, and increased noise
level. The traditional approach to reduce these motions consists in
installing vibration isolators between the source of disturbance and
the component to be protected. These isolators should also provide
sufficient damping to decrease vibration levels at resonance.
The ability of elastomers to provide damping makes these materials a perfect choice for applications in vibration isolators; so
rubber to metal devices are increasingly adopted for several specialized industrial applications and, specifically, they are widespread in the automotive field engine mounts, shock absorbers,
torsional dampers, etc.. As a consequence of the viscoelastic nature of rubbers, their dynamic and thermal behavior is significantly
dependent on frequency and temperature. Further, due to the high
content of carbon black, usually incorporated together with other
additives within the compounds, to optimize the mechanical performances of the rubber components, the dynamic behavior of these
materials is markedly nonlinear. Moreover elastomers properties
depend on the amount of the applied static load static prestrain or
equally preload.
This complex behavior has to be considered to estimate accurately the dynamic performances of damping rubber to metal devices; so a thorough knowledge and characterization of the dynamic properties of elastomeric compounds is needed in designing
such components.
Manuscript received September 21, 2005; accepted for publication January
12, 2006; published online March 2006.
1
Full Professor of Machine Design, Dipartimento di Ingegneria Meccanica,
Universit degli Studi di Brescia, via Branze 38, 25123 Brescia, Italy.
2
Research Engineer, Dipartimento di Ingegneria Meccanica, Universit degli
Studi di Brescia, via Branze 38, 25123 Brescia, Italy.
3
Research Engineer, Dipartimento di Chimica e Fisica per lIngegneria e per
i Materiali, Universit degli Studi di Brescia, via Valotti 9, 25123 Brescia, Italy.
4
Full Professor of Materials Science and Technology, Dipartimento di
Chimica e Fisica per lIngegneria e per i Materiali, Universit degli Studi di
Brescia, via Valotti 9, 25123 Brescia, Italy.
5
Associate Professor of Mechanical and Thermal Measurements, Dipartimento di Ingegneria Meccanica, Universit degli Studi di Brescia, via Branze
38, 25123 Brescia, Italy.
Moreover in the industrial practice, although a lot of information about rubber is available, when a specific rubber compound is
considered, it is usually difficult to find useful data for mechanical
design.
It is furthermore to be noted that, after obtaining the required
data by testing on materials, such data can hardly be handled by
usual computing tools. Especially the commercial FEM-FEA finite elements modeling and analysis technologies can give only
partially satisfactory results if they are applied to materials which
are strongly nonlinear, hysteresis affected, and temperature sensitive. Although such problems can be managed at academic level, it
seems that at industry level they will prevent a practical efficient
use of FEA for rubber items modeling. So, a research program,
aimed at acquiring the missing knowledge and developing the
means needed for the design of the aforementioned mechanical
items, was started. It may be basically thought of as constituted by
the three following different but closely related stages:
Experimentation: to be conducted both on materials and
components.
Modeling: performed by discrete particle based and concentrated parameters models.
Simulation: numerically conducted on the developed
models, applied on real geometry components, involving
materials data pointed out during the experimental
characterization.
The final aim of the research is to provide a practical method
which could be used to optimize the design of rubber to metal components by numerical simulations, making use of experimental data
describing the dynamic performances of the elastomeric compounds.
This contribution is intended to focus on the adopted experimental and data evaluation procedures, showing results and proposing some hints for discussion. The description of the developed
modeling and simulation approach is given in 1.
The testing method described in the present work, inspired by
developments in the field of dynamic characterization of shock absorbers 2, is based on the use of an electrodynamic shaker and the
methodology described is a refinement of procedures previously
proposed by the authors since 1997 3,4. This testing technique
adopts a configuration which was standardized by ISO in 1997
5,6 and refined by standards in 2002 7. However, the present
Copyright 2006 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
FIG. 2Block diagram of the resilient elements testing configuration as proposed by ISO 10846.
FIG. 1Storage full line and loss modulus dashed line of a vulcanized filled
rubber SBR containing 70 phr carbon black as a function of temperature,
measured by a commercial DMTA for a strain amplitude of 0.7 % at a frequency
of 10 Hz measures by the authors.
approach extends beyond the content of these documents, providing proper procedures for the characterization of the high frequency 10 Hz dynamic properties of soft materials as well as
components at varying conditions for frequency, dynamic deformation amplitude, and static preload.
For a linear viscoelastic material, E and E depend only on temperature and frequency and are independent from the level of deformation. By contrast, in the case of filler reinforced elastomers, due
to the high content of filler, the dynamic behavior of such materials
appears markedly nonlinear. This implies that E and E show a
very strong deformation amplitude dependence, a phenomenon
which often is referred to as the Payne effect 10,11. This effect
typically consists of a decrease of the storage modulus by increasing the strain amplitude, associated to an increase or a maximum of
the loss modulus. The strain amplitude dependence of the dynamic
moduli shows that the loss modulus reaches a maximum at about
the same strain amplitude of the inflection point in the storage
modulus and then falls to lower values as the strain amplitude increases. These effects, typically observed in the range of strain amplitude between 0.01 % and some 10 %, strongly increase by increasing the filler loading.
In order to quantify the dynamic behavior of filled rubbers, it is
necessary to produce data with varying frequency, temperature and
strain amplitude within the ranges of interest. Moreover also static
load static prestrain has to be considered as a quantity influencing
the materials properties 12,13.
E
E
Typical results of E and E versus temperature for filled rubbers, measured by a commercial DMTA dynamic mechanical thermal analyzer produced by Polymer Lab Ltd UK, are shown in
Fig. 1.
The determination of storage and loss modulus in wide ranges
of frequency master curves is usually obtained by empirical
frequency-temperature reduction of experimental data, acquired at
different temperatures and in limited ranges of frequency 8. From
data such as those of Fig. 1 obtained at different frequencies within
the range between 0.3 Hz and 30 Hz, the authors determine dynamic moduli master curves extending over several decades of frequency i.e., between 104 and 1010 Hz at the reference temperature
of 25 C 9.
Methods
Testing Equipment and Configuration
International standards relate to laboratory testing of vibroacoustic
transfer properties of resilient elements vibration isolators such as
the aforementioned vibration damping devices for automotive applications. The testing configuration proposed by these documents
can be outlined by a block diagram Fig. 2. According to Fig. 2, the
system consists in three blocks, representing a vibration source, a
group of n tested resilient elements, and a receiving structure, respectively. It is assumed that a point of contact contact area which
vibrates as the surface of a rigid body between the vibration source
and the resilient system and between the resilient system and the
receiving structure, exists.
A vector containing three orthogonal forces and three orthogonal moments F and a vector containing three orthogonal displacements and three orthogonal rotations u are associated to each contact point. The tested resilient system operates by a certain
unknown transfer function, which may be described by a suitable
stiffness or transmissibility matrix 5, giving the relationship between force-displacement input F1 , u1 and output F2 , u2
couples. Each one of the vectors contains 6 n elements. This general approach allows individuating two different kinds of testing
techniques. The first one is the so-called direct method 6; in this
approach either the input displacement, velocity or acceleration,
and the blocking output forces dynamic forces on the output side
of the tested resilient device which results in zero displacement output are measured. The second one is the so-called indirect method
7; it consists of a measurement of a resilient element vibration
transmissibility for displacement, velocity, or acceleration, with
the output element loaded by a known mass. Reference is made
Data Analysis
The dynamic properties of a resilient system are described by a
complex matrix given in terms of transmissibility or stiffness. If a
unidirectional vibration is considered, for a single degree of freedom model applied to the analyzed system, this matrix is simply
reduced to a single complex number. The complex values K* and
H* denote stiffness and transmissibility, respectively. These quantities assume proper values in different conditions of frequency, temperature, dynamic strain amplitude, and static load.
The time history of the vibrating table is described by the function xt, where t identifies the time variable. The vibration of the
suspended mass is denoted by a function yt. If a sinusoidal wave
is considered as exciting signal for the vibrating table and under the
hypothesis that the tested system is linear and stationary, also the
vibration of the suspended mass can be described by a sinusoidal
function in the time domain; specifically both the sinusoidal waves
x and y have the same frequency . In these conditions, each
quantity of interest can be conventionally handled as a complex
number having a module which is equal to the sinusoid amplitude
and the same phase of the sinusoidal function. As well as displacements x and y also velocities and accelerations are described by
sinusoidal functions in the time domain and they can be handled as
complex quantities.
Imposing the dynamic equilibrium for the suspended mass, the
following equation can be written, meaning that viscoelastic forces
have to equal inertial actions:
m Y + K*Y X = 0
H* =
K*
Y
= *
X K m 2
m 2 H*
H* 1
Thus, in fixed conditions of frequency, dynamic strain amplitude, temperature, and static load, if the transmissibility function
H* is experimentally measured, the complex stiffness K* is estimated by Eq 6 for the tested system.
For the test equipment described in the previous section, accelerations concerned with the vibrating table and the suspended mass
are acquired in the time domain. These acceleration time histories
are computed by means of the least squares method or by the Fourier transform in order to associate to each signal a proper complex
number according to the signal spectrum. At a certain frequency
the ratio of the two founded complex accelerations Y / X gives
the system transmissibility H*. So at each frequency the complex stiffness K* is calculated by Eq 6. This complex number
relates to the complex stiffness of the whole tested resilient system;
if this is constituted by n resilient components, the complex stiffness to be assigned to each one of these is K* / n. Starting from
complex values of stiffness, the complex moduli can be estimated
using the following Eqs 7 and 8:
E* =
1
K* l
n
A 1 + 2 s2
K* l
n
A
G* =
d
4l
Results
Typical results for transmissibility H* in terms of modulus H and
phase angle evaluated by ES electrodynamic shaker, for the
shear test system, are shown in Fig. 6 as a function of frequency.
Values of the resonance frequency r and the transmissibility
amplitude at the peak were also measured for different dynamic
strain amplitudes. It appears that a strain increase basically produces a decrease of the resonance frequency r and of the value
of transmissibility at the peak.
These findings are shown in Fig. 7, where the resonance frequency and the value of transmissibility at the peak are plotted as
functions of the dynamic strain amplitude, indicating the intrinsic
nonlinear behavior of the elastomeric material constituting the testing system.
Data of stiffness K* in terms of modulus K and loss phase
= a tan 2K / K, measured by ES, are shown in Figs. 8 and 9a,
versus frequency and prestrain, respectively. It can be noted that
rubber shows its typical dynamic stiffening effect. Data have to be
considered as reliable only at frequencies higher than 100 Hz, due
to the presence of resonance phenomena in the range 50 80 Hz;
specifically, at frequencies lower than the resonance frequency, the
H* module amount is almost 1, thus a computation of the K* value
FIG. 8Stiffness dashed line and loss phase full line as a function of frequency for the same testing device as in Fig. 6, measured by ES for a strain
amplitude amount of 0.2 % at room temperature.
Concluding Remarks
FIG. 9a Stiffness full symbols and loss phase open symbols and b storage modulus E concerning the testing system containing prestrained rubber
samples same materials as in Fig. 6 as a function of prestrain; measurements
obtained by ES for strain amplitude 0.2 %, at 25 C and 160 Hz.
Acknowledgments
The authors are grateful to CFGomma SpA Brescia, Italy and to
Pirelli SpA Milan, Italy for their support.
References
1 Magalini, A., Vetturi, D., Ramorino, G., Saleri, A., and
Vigan, F., Dynamic Analysis of Rubber to Metal Devices
Using an Unconventional Numerical Modeling for Elastomers, Constitutive Models for Rubber III, J. Busfield and
A. Muhr, eds., Balkema, Lisse, The Netherlands, 2003.
2 Da Forno, R., Debei, S., and Saggin, B., Caratterizzazione
Dinamica di Smorzatori di Vibrazione, Proceedings of II
Congresso Nazionale di Misure Meccaniche e Termiche,
Bressanone, Italy, June 1921, Universit degli Studi di Padova, Padova, Italy1995.
3 Vetturi, D., Prove Dinamiche per la Caratterizzazione delle
Propriet Meccaniche di una Mescola per Pneumatici Automobilistici, Proceedings of III Congresso Nazionale di Misure Meccaniche e Termiche, Ancona, Italy, July 12, Universit degli Studi di Ancona, Ancona, Italy, 1997.
4 Cambiaghi, D., Vetturi, D., Ramorino, G., and Ricc, T., Caratterizzazione delle Propriet Dinamiche e Termiche di Materiali Elastomerici: Sviluppo di Metodologie Sperimentali Innovative, Proceedings of XV AIMETA National Congress of
the Italian Association for Theoretical and Applied Mechanics, Taormina, Italy, September 2629, Associazione Italiana
di Meccanica Teorica e Applicata AIMETA, Taormina, Italy
2001.
5 ISO, ISO 10846-1:1997, Acoustics and Vibration
Laboratory Measurement of Vibro-Acoustic Transfer Properties of Resilient Elements. Part 1: Principles and Guidelines,
ISO, 1997.
6 ISO, ISO 10846-2:1997, Acoustics and Vibration
Laboratory Measurement of Vibro-Acoustic Transfer Properties of Resilient Elements. Part 2: Dynamic Stiffness of Elastic Supports for Translatory MotionDirect Method, ISO,
1997.
7 ISO, ISO 10846-3:2002, Acoustics and Vibration
Laboratory Measurement of Vibro-Acoustic Transfer Properties of Resilient Elements. Part 3: Indirect Method for Determination of the Dynamic Stiffness of Resilient Supports for