See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/223282460

Prediction of fatigue life using modal analysis for grey and ductile cast iron

Article  in  International Journal of Fatigue · March 2007

DOI: 10.1016/j.ijfatigue.2006.05.004

CITATIONS READS

41 466

3 authors:

Ahmed Damir Ahmed Mohamed Elkhatib

National Research Council Canada Alexandria University
23 PUBLICATIONS   227 CITATIONS    13 PUBLICATIONS   93 CITATIONS   

SEE PROFILE SEE PROFILE

Galal Nassef
Alexandria University
49 PUBLICATIONS   90 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Turning of Titanium Alloy View project

Tool Condition Monitoring (TCM) View project

All content following this page was uploaded by Galal Nassef on 18 May 2020.

The user has requested enhancement of the downloaded file.

 International
Journalof
International Journal of Fatigue 29 (2007) 499–507
Fatigue
www.elsevier.com/locate/ijfatigue

Prediction of fatigue life using modal analysis for grey

and ductile cast iron
a,*
A.N. Damir , A. Elkhatib b, G. Nassef c

a
Production Engineering Department, Faculty of Engineering, Alexandria University, Egypt
b
Machinery Dynamics & Diagnostics, Production Engineering Department, Faculty of Engineering, Alexandria University, Egypt
c
Materials Science and Technology, Production Engineering Department, Faculty of Engineering, Alexandria University, Egypt

Received 15 August 2005; received in revised form 26 April 2006; accepted 6 May 2006
Available online 27 June 2006

Abstract

The aim of this work is to investigate the capability of experimental modal analysis, as a nondestructive tool, to characterize and
quantify fatigue behavior of materials. This is achieved by studying the response of modal parameters (damping ratio, natural frequency,
and FRF magnitude) to variations in material microstructure, as a main factor affecting fatigue life. This helps in correlating modal
parameters to fatigue behavior. Cast iron family represented by grey cast iron, ductile cast iron and austempered ductile iron (ADI)
is used in experiments as a case presenting considerable variations in microstructure. Modal testing was performed on specimens made
of the selected materials in order to extract the corresponding modal parameters. Rotating bending fatigue test was performed on stan-
dard fatigue specimens to correlate the modal parameters to the fatigue behavior. This enables the evaluation of the ability of modal
testing to predict the fatigue life of mechanical components.
Ó 2006 Elsevier Ltd. All rights reserved.

Keywords: Austempered ductile iron; Modal testing; Fatigue life

1. Introduction predictions of different components such as aircraft struc-

Recent technical demands for improving the perfor-
mance of engineering components have brought up the
need of proper estimation of components/system life to
avoid sudden or unexpected failure of equipment. The abil-
ity of any system to perform its required function without
failure remains a challenging concern for design engineers.
As considered as the main cause of failure in industrial
components, fatigue remains the main source of unex-
pected failure in mechanical components as the majority
of structures are subjected to cyclic, alternating stress. Con-
sequently, fatigue life can be satisfactorily considered as a
measure for the reliability of mechanical components.
Many research papers have been dealing with fatigue life
*
Corresponding author.
E-mail address: damiram@mcmaster.ca (A.N. Damir).
tural components, riveted lap joints, and welded joints
using different methods ranging from finite element analy-
sis to fracture mechanics theories that are mainly based on
destructive testing. However these methods are considered
approximate as they do not include all the factors affecting
fatigue life because they are not based on the actual com-
ponents and systems.
The demand for a nondestructive test that predicts the
fatigue life of individual components is growing by time
as it is expected to overcome the problem of variability
of results in destructive techniques. This leads to the intro-
duction of experimental modal analysis, as a non destruc-
tive tool, to help in determining the reliability of machine
components as it evaluates the structure integrity as it is
based on the theory of resonance testing [1]. A significant
amount of work has been done regarding the use of modal
parameters (i.e. natural frequency, modal damping, and
mode shapes) for damage detection and identification.

0142-1123/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijfatigue.2006.05.004
500 A.N. Damir et al. / International Journal of Fatigue 29 (2007) 499–507
The basic idea is that the modal parameters are, by defini-
tion, functions of the physical properties of components
(i.e. mass, stiffness or modulus of elasticity) and hence of
mechanical properties. This concept was used mainly in
the application of modal testing as an acceptance test for
castings [2,3]. Doebling et al. [4] presented an extensive
review of the technical literature concerning the detection,
location and characterization of structural damage by
examining changes in measured structural vibration
response and/or modal parameters. The ability of detecting
the presence of cracks in beams and determine its location
and size based on changes in natural frequencies and fre-
quency response function (FRF) amplitudes was studied
by Owalbi et al. [5]. They observed that the amplitude of
the FRF of the bending modes increase as the crack size
grow, this is accompanied by a decrease in natural frequen-
cies. The use of modal analysis for damage detection in alu-
minum cylinders was studied by Davis [6]. He declared that
the evaluation of natural frequencies and the curvature of
the FRF (mode shapes) can be used to determine the pres-
ence, location, and severity of the imparted damage. The
ability of modal testing to estimate the fatigue life of stan-
dard fatigue specimens made of brass was investigated by
Tobgy et al. [7]. Random notches were introduced in brass
specimens to achieve some variability in the fatigue life of
specimens. A correlation between fatigue life and corre-
sponding modal parameters of each specimen was
achieved. The feasibility of modal analysis as a nondestruc-
tive technique in detecting damage for health monitoring of
composite materials was investigated by Kessler et al. [8].
They performed experimental modal testing as well as a
finite element model on graphite/epoxy panels containing
representative damage modes. Strong correspondence
between the extent of damage (or local stiffness loss) and
reduction in natural frequency was found. Bedewi et al.
[9] developed an experimental method to predict the resid-
ual fatigue life of composite structures. Changes of modal
parameters, i.e. natural frequencies and damping ratios,
were investigated and correlated to the fatigue failure life
during the fatigue process for selected gaphite/epoxy com-
posite specimens to be used as nondestructive indicator of
fatigue life. The decrease in natural frequency and the
increase of damping ratio as a function of number of cycles
to failure were monitored to predict the fatigue failure life.
Table 1a
Chemical composition of ductile cast iron
C% Si% Mg% P%
3.424 2.462 0.0254 0.0342
Table 1b
Chemical composition of grey cast iron
C% Si% Mg% P%
3.893 1.605 0.0004 0.2165
The effect of fatigue damage on the dynamic natural fre-
quency over the entire fatigue life process for tensile-shear
spot-welded joints was studied by Shang et al. [10]. The
study aimed to obtain a physically based damage parame-
ter for spot-welded joints made from automotive annealed
steel. A remarkable decrease of the natural frequency of the
tested spot-welded was noticed after the half of the fatigue
life with more dramatic decrease at the end of the fatigue
life. Liu et al. [11] studied the effect of circumferential
cracks on coupled-response of a circular hollow section
beam both analytically and experimentally. They observed
that the natural frequencies shift to lower value with the
increase of crack severity (depth). Similar trend was
obtained by Kim [12] as he studied application of vibra-
tion-based techniques in on-line damage identification in
composite materials using frequency response function
(FRF). The peaks of the measured responses (i.e. the nat-
ural frequencies) shift to lower frequency as the debonding
extent increases and with the increase of the number of fati-
gue loading cycles.
The main concern of this work is to study the capabil-
ity of modal testing, as a non destructive test, to charac-
terize and quantify the fatigue behavior of mechanical
components. This can be achieved by studying the
response of modal parameters to microstructure character-
istics as a main factor affecting the fatigue life of compo-
nents. To elaborate the effect of change in microstructure
on the response of modal analysis, cast iron family pre-
sented by grey cast iron and ductile cast iron is used in
experiments as presenting two types of cast iron with
different microstructure. Besides, the change in phase
amount occurred in austempered ductile iron (ADI)
treated at different conditions is also considered; a condi-
tion presenting microstructure variation that leads to var-
iation in mechanical properties to be dynamically studied
[13].
2. Experimental work
In this paper, [14] two sets of experiments were carried
out to correlate modal parameters to fatigue. The experi-
ments were performed on specimens made of ductile and
grey cast iron with the following composition as shown
in Tables 1a and 1b.
S% Mn% Mo% Ni% Cu%
0.0032 0.327 0.0017 0.0026 0.0072
S% Mn% Cu% Ni% Fe%
0.1232 0.4169 0.0804 0.0396 93.6
 A.N. Damir et al. / International Journal of Fatigue 29 (2007) 499–507
Modal testing was performed on specimens of cylindri-
cal shape with defined dimensions to minimize the effect
of other parameters, mainly shape variation that can affect
the modal results. On the other hand, fatigue tests were
applied on standard fatigue specimens as shown in Figs.
1a and 1b.
First group of experiments were carried out on both
ductile and grey cast iron specimens. Fatigue tests were
performed on specimens after being dynamically tested to
express the variation of mechanical properties with the
morphology of graphite shape from flakes to nodules. This
set of experiments had the objective to study the ability of
modal testing to respond to variation in mechanical
properties.
Second group of experiments aimed to check the sensi-
tivity of modal parameters to variations in microstructure,
especially the variation in the amount of phases present in
the material. This set of experiments is performed on duc-
tile cast iron specimens, with same chemical composition,
treated with different austempering heat treatment condi-
tions as shown in Fig. 2. The specimens are first treated
with different heat treatment conditions in order to obtain
different amount of microstructure phases. They are first
austenitised at temperature of 900 °C for 2 h and then
quenched into lead bath for austempering stage at temper-
atures range from 325 °C to 400 °C for holding time of
1–3 h. This resulted in specimens of different amount of
microstructure phases (i.e. %retained austenite, graphite
count, and graphite size) that were proved to affect the
mechanical properties and specially the fatigue life of
ADI components [15]. Microstructure examination is con-
Fig. 1a. Cylindrical specimens used in experiments.
Fig. 1b. Fatigue standard specimen.
501
Austenitizing
900°C for 2 hrs
Quenching
in lead bath
Temperature
Austempering
325°C-400°C
for 1-3 hrs
Air cooling
Time
Fig. 2. Austempering heat treatment scheme.
ducted on different specimens to quantify the morphology
and proportions of the constituting phases obtained.
Experimental modal analysis was conducted on the same
specimens to evaluate the corresponding modal parameters
(damping ratio n, natural frequency, and magnitude).
Fatigue test was then conducted on the same specimens.
In this test all specimens were subjected to a constant rotat-
ing bending load producing a bending stress of approxi-
mately 500 MPa. The value of the bending load is set to
a level that ensures the fracture of all specimens with differ-
ent heat treatment conditions. This level is selected in
accordance to the S–N curve for different austempering
conditions. All specimens are tested at the same speed
(4000 rpm).
502 A.N. Damir et al. / International Journal of Fatigue 29 (2007) 499–507
The experimental procedure can be summarized in the
following steps:
(1) Austempering of ductile cast iron specimens with the
pre set treatment conditions.
(2) Machining of austempered specimens into cylindrical
shape with the mentioned dimensions.
(3) Subjecting cylindrical austempered specimens to
dynamic impact testing (Modal testing).
(4) Subjecting austempered specimens to microstructure
examination.
(5) Performing fatigue test on standard fatigue
specimens.
The test specimens were machined into cylindrical shape
with defined dimensions before being dynamically tested to
minimize the effect of other parameters, mainly shape var-
iation that can affect the modal results. Therefore, dynamic
Fig. 3. Example of frequency response function and extracted modal parameters.
Force transducer
(Excitation)
Channel 1
Specimen
Accelerometer
Cable
Fig. 4. Dynamic test of free supported specimen.
properties will be mainly the result of microstructure
variations.
Modal testing of the specimens was conducted using an
impact hammer as a source of excitation to excite tested
specimens. The impact hammer is equipped with a force
transducer to measure the exciting force signal. The vibra-
tion response of the specimen to the impulse excitation was
measured using an accelerometer.
Both the response of the specimens and the excitation
signal were measured and connected to a multi-channel
analyzer used for the analysis and conditioning of the sig-
nals. A PC equipped with B & K Pulse software is con-
nected to the multi-channel analyzer, which is used to
collect, analyze and display the signals in a user-friendly
interface. Frequency Response Function (FRF) is auto-
matically calculated and graphically presented through
the Pulse software as a result of dividing the response to
the exciting spectrum and modal parameters are extracted
Hammer Elastic holding
support
Accelerometer
(Response)
Channel 2
 A.N. Damir et al. / International Journal of Fatigue 29 (2007) 499–507
from the FRF for each austempered specimen. Fig. (3)
shows an example of the output of modal testing obtained
using pulse software.
The value of damping ratio is calculated from the mea-
sured FRF by means of the half power (3 dB) bandwidth
method. In this method, the damping ratio is set to be
equal to half the 3 dB bandwidth of the FRF peak. How-
ever, the half power method is not suitable for the cases
where the resonances are sharp and the peaks are too nar-
row for accurate measurements of the bandwidths and for
the case of highly coupled modes.
The coherence value at each mode frequency is indicated
from the coherence function curve to indicate the degree of
linearity and accuracy of obtained results. Measurements
with coherence values less than 0.995 were discarded
During dynamic test, the specimens are freely supported
to avoid the effect of ground support device on the
obtained results. The accelerometer is mechanically
mounted to the tested specimen using mechanical stud dur-
ing experiments. Fig. 4 shows the specimen freely sup-
ported when dynamically tested.
3. Results and discussion
3.1. Grey and ductile cast iron experiments
This set of experiments is considered as preliminary
experiments aimed to check the ability of modal parame-
ters to respond to variation of microstructure constituents
that affect the fatigue behavior of the component.
Experiments were performed on specimens of nodular
cast iron and grey cast iron.
The microstructure change in the selected materials is
expressed by the change of graphite shape that is nodular
in case of ductile cast iron and flakes type in grey cast iron.
This variation in graphite shape is known to greatly affect
the mechanical properties of cast iron and specially its fati-
gue resistance [16]. Graphite flakes in grey cast iron tend to
1.7
1.5
Damping ratio %
1.3
Grey cast iron
1.1
specimens
0.9
0.7
0.5
10
Fig. 5. Fatigue life vs damping ratio%.
503
decrease the fatigue life of grey cast iron components as
they act as crack initiators in iron matrix. However, nodu-
lar graphite improves the fatigue strength of the ductile cast
iron as they are considered like crack arresters.
Experimental study was performed to check the ability
of modal parameters to respond to the difference between
the two material types presented by the graphite morphol-
ogy, as well as within each family. As preliminary results, it
was noticed that the modal parameters were sensitive to the
change of microstructure between the two types of material
tested. Fig. 5 shows the scatter plot of damping ratio% with
the fatigue life of ductile and grey cast iron.
From the scatter plot, it is clear that the damping
ratio% is sensitive to the variation of microstructure mor-
phology between grey and ductile cast iron. Additionally,
for both ductile and grey cast iron, the fatigue life shows
an increasing trend with the increase of damping ratio as
shown in Fig. 5. Nodular specimens have higher damping
ratio and hence, higher fatigue life than grey cast iron
specimens. The increase of damping ratio with the fatigue
life is noticed for the two families and within each family
as a trend can be noticed within the ductile cast iron
specimens.
However, the natural frequency showed a greater
response to the material type than the fatigue life. Nodular
cast iron specimens have higher natural frequencies than
do grey cast iron specimens but a trend cannot be noticed
within each family. This can be explained by the depen-
dency of natural frequency on the modulus of elasticity,
which does not vary considerably within the same family
of material.
The FRF magnitude (Hd) increased with the decrease of
fatigue life between the two families. However, a trend
within the same family was not considerable.
Therefore, it can be concluded that modal parameters
can be used to detect the variation in microstructure con-
stituents and morphology that are crucial for mechanical
properties as they respond to the microstructure variation
Ductile iron
specimens
100 1000 10000
Life(log)
504 A.N. Damir et al. / International Journal of Fatigue 29 (2007) 499–507

between the two families. The damping ratio (n) is sensitive
to the type of material to be tested as well as to the varia-
tion in mechanical properties within the same material.
However, the natural frequency is mainly dependent on
material type, which can be explained by the dependency
of natural frequency on the material stiffness that is affected
by the matrix microstructure.
3.2. Austempered ductile iron
3.2.1. Modal parameters vs ADI microstructure phases
Scatter diagrams between the modal parameters and the
microstructure measures are shown in Figs. 6(a)–(c).
It was noticed from the plots of modal parameters with
microstructure parameters of ADI treated at different con-

(a) 1.8

1.6
Damping Ratio%

1.4

1.2

0.8
10 20 30 40 50 60
%R.A

(b) 1.8

1.6
Damping Ratio %

1.4

1.2

0.8
200 300 400 500 600 700
Gr.count/mm2

(c) 1.8

1.6
Damping Ratio %

1.4

1.2

0.8
5 7 9 11 13 15 17 19
Gr.size
Fig. 6. Plot of damping ratio% vs microstructure measures: (a) %RA; (b) graphite count/mm2; (c) graphite size.
 A.N. Damir et al. / International Journal of Fatigue 29 (2007) 499–507
ditions that the damping ratio (n) is most sensitive to
microstructure changes than other measured modal param-
eters. This can be attributed to the capacity of the material
to dissipate energy internally. Material damping is a term
that describes the complex physical effects that convert
kinetic and strain energy in a vibrating mechanical system,
consisting of a volume of macrocontinuous (solid) matter,
into heat. It is often associated with such mechanisms as
dislocation movements, plastic flow and slipping effect,
internal friction due to elastic deformation of the system
under torsion or bending, and microstructural defects. In
this work the damping ratio can represent the material
damping of tested specimens of same form and dimensions
(cylindrical specimens). Trends between n and microstruc-
ture are expected to be observed due to the role played
by microstructure (phases present and their amount) on
the possession of material damping characteristics.
A considerable influence of the percentage of retained
austenite phase on material damping is shown in
Fig. 6(a). The damping ratio shows an increasing trend
with the increase of % of retained austenite. It has been
proved that the presence of soft phase of retained austenite
would improve the fatigue characteristics of ADI by form-
ing barriers leading to a deviation of crack path. This
means that an increase in the amount of retained austenite
in ADI will lead to an increase in the material damping,
which would inhibit fatigue crack growth.
The scatter plot in Fig. 6(b) between damping ratio and
the graphite count shows that the damping ratio increases
with the increase of graphite count, which also improves
the fatigue behavior of ADI components. An increase in
the count of graphite nodules would result in an increase
in the graphite sites in the microstructure which enhances
the fatigue behavior by increasing the number of obstacles
to crack growth.
The damping ratio shows a decrease with the increase of
graphite size as shown in Fig. 6(c), which lowers the fatigue
strength of ADI specimens. This can be explained by the
effect of small size of graphite nodules that affect the den-
sity and distribution of graphite network. Larger graphite
site leads to smaller graphite count and therefore a decrease
in the surface area of damping element in the matrix (i.e.
graphite nodule).
The dominant natural frequency of specimens was
found to remain almost constant for all specimens treated
at different conditions. No considerable response to micro-
structure variations was noticed. This observation can be
explained by sensitivity of natural frequency by definition
to the modulus of elasticity, dimensions, and shape, which
are almost kept constant for the tested specimens. How-
ever, scatter in the frequency can be explained by the
expected variability in dimensions, shape and modulus of
elasticity due to different treatment conditions, but it
remains inconsiderable variations. Therefore, by definition
of natural frequency, a little response of natural frequency
with microstructure measures was expected. These results
are in accordance with the results obtained by Lin et al.
505
[17] that shows no significant effect of austempering on
the modulus of elasticity.
FRF magnitude (Hd) was found to decrease with the
increase of % retained austenite, graphite count, and
decrease of graphite size. However, the trend was not
significant.
3.2.2. Modal parameters vs fatigue life
The scatter plot of modal parameters vs fatigue life for
ADI specimens is shown in Fig. 7.
As stated previously, the aim of this set of experiments is
to correlate the modal parameters with the results of the
fatigue testing performed on ADI specimens treated at dif-
ferent conditions. The experimentally extracted modal
parameters (i.e. damping ratio% (n), natural frequency,
and magnitude) were studied against the fatigue life.
Fig. 7 shows the scatter plot of damping ratio% vs the
fatigue life. It can be noticed from the figure that an
increase in the damping ratio will lead to an increase of
the fatigue life. The increase in material damping will
probably improve the fatigue behavior of the mechanical
component due to the internal dissipation of applied
energy which reduces the resonance effect on the stressed
component. Higher system damping will lead to dissipation
of larger amount of stress energy that can cause failure
of the system if stored. Stored energy can lead to a
dramatic increase in the stress level within the system
leading to crack propagation and hence, to catastrophic
consequences.
The FRF magnitude was shown to decrease with the
increase of fatigue life, although the trend was not signifi-
cant. The opposite trends of damping ratio and Hd with
fatigue life is explained by the nature of relation between
the FRF magnitude and the damping ratio, where an
increase in the damping ratio of the system will lead to a
decrease in the magnitude of FRF. The natural frequency
had no noticeable trend with the variation of fatigue life.
This observation is in agreement with the response of fre-
quency to variation in mechanical properties shown in
experiments of ductile and grey cast iron. This observation
can be explained by sensitivity of natural frequency to the
2000000
1500000
Fatigue Life
1000000
500000
0
0.8 1 1.2 1.4 1.6 1. 8
Damping ratio%
Fig. 7. Fatigue life vs damping ratio%.
506 A.N. Damir et al. / International Journal of Fatigue 29 (2007) 499–507
modulus of elasticity, dimensions, and shape, which are
almost kept constant for the tested specimens. These results
are in accordance with the results obtained by Lin et al. [17]
that shows the non significant change in modulus of elastic-
ity for different austempered microstructure.
4. Statistical analysis
4.1. Damping ratio vs microstructure parameters
Multi regression analysis is applied to establish an
empirical formula that correlates the response, i.e. the
damping ratio, with the input variables that are expected
to affect the response. It aimed to estimate the damping
of austempered ductile iron specimen as a function of its
microstructure. T-test is applied on the input variables in
order to determine the most significant parameter(s) that
affect the response to avoid the effect of non significant
parameter on the proposed model.
It can be concluded from the T-test that the input vari-
able graphite size is not statistically significant in the effect
on the response. This can be explained by the inter relation
between graphite count and size that the later is a logic
result of the former. Therefore, it can be concluded that
the damping ratio (n) is mainly affected by the percentage
fraction of retained austenite (%RA) and the graphite
count per unit area.
According to the results of the T-test, the proposed
models will be a function of two inputs %RA and count/
mm2.
A linear model is proposed to correlate the material
damping to microstructure parameters using multivariate
regression
nð%Þ ¼ 0:73256 þ 0:008143 %RA þ 0:000327 count=mm2
Fig. 8 shows the response surface of damping ratio as a
function of % retained austenite and graphite count/mm2.
It can be observed from the surface plot that the damping
ratio increases linearly with the increase of fraction of car-
7
1.7
6
1.6
damping ratio %
1..5
1.4
1
1.3
1.2
1
1.1
15
25
35
1000
%R.A 1100
45 1200
1300
1400
55 1500
Fig. 8. Response surface for n% in terms of %RA and Gr. count/mm2.
bon-enriched retained austenite and the graphite count/
mm2 that indicates the increase of material damping. This
result is in accordance with the effect of microstructure on
fatigue behavior that shows an enhancement with the in-
crease of %RA and graphite count/mm2. This can be ex-
plained that an increase in the material damping will lead
to higher energy dissipation which means a reduction in
the stored stress energy that causes failure of components.
Different plot of residuals (errors) were analyzed to check
the conformity of the model.
The linear model is accepted concerning the acceptance
criteria. All terms of the proposed model are statistically
significant based on the P value in ANOVA analysis.
4.2. Fatigue life in terms of modal parameters
Multi regression analysis is applied to establish an
empirical formula that correlates the modal parameters
with the fatigue life. It aimed to estimate the fatigue life
of austempered ductile iron specimen as a function of its
modal parameters. T-test is applied on the input variables
in order to determine the most significant parameter(s) that
affect the response to avoid the effect of non significant
parameter on the proposed model.
From the COMPUTED T values, it can be concluded
that the Hd is not statistically significant in affecting the
fatigue life as it do not satisfy the criterion of T value. This
can be explained by the relation between the FRF magni-
tude (Hd) and the damping ratio (n) which concluded that
the damping ratio is enough to express the effect of both.
Therefore, it can be concluded that the fatigue life as an
output of the fatigue test performed at stress level equals
to 500 MPa can be correlated to the damping ratio and
the natural frequency. However, no significant trend was
found between the natural frequency and the fatigue life.
ð1Þ Additionally, it was noticed from the COMPUTED T
value test that accepted t value of n is higher than the value
of frequency, which means that the effect of n on the fatigue
life is predominant. Therefore, the multivariate regression
analysis can be reduced into simple regression analysis
between the fatigue life and damping ratio n.
2000000
1500000
Fatigue Life (N)
1000000
500000
700
800
900
0
0.8 1 1.2 1.4 1.6 1.8
Damping ratio %
Gr.couunt/mm2
Fig. 9. Experimental results and quadratic model for fatigue life in terms
of damping ratio%.
 A.N. Damir et al. / International Journal of Fatigue 29 (2007) 499–507
A quadratic model is proposed to describe the relation
between these variables as shown:
2
N ¼ 2308173  5294819 n% þ 2906551ðn%Þ
Fig. 9 shows the graphical plot of experimental results and
the proposed quadratic model of fatigue life of ADI spec-
imens as a function of the damping ratio%. From the figure
it can be noticed that the proposed model is in a good
agreement with experimental results. ADI specimens with
higher damping ratio% experience higher fatigue life.
The model is accepted regarding the statistical accep-
tance criteria. The empirical formula is valid for predicting
fatigue life of ADI specimens tested at stress level of
500 MPa.
5. Conclusion
The ability of modal testing to characterize fatigue
behavior of mechanical components has been investigated
in this work. The metallurgical characteristics of any com-
ponent, explained by the type of phases and their percent-
age, play a crucial role in determining the mechanical
properties of the part such as fatigue behavior.
Therefore, the ability of modal parameters to respond to
the change introduced in the microstructure would give an
indication of their ability to express the fatigue behavior of
mechanical components.
Experimental modal analysis has proved to show a great
potential for fatigue evaluation and characterization for
mechanical components. This conclusion is achieved by
studying the response of modal parameters to changes
occurred in microstructure in austempered ductile iron
(ADI) as a factor affecting fatigue life. The material effec-
tive damping ratio (n) shows the most noticeable response
to the changes in material hardness and fatigue life between
the ductile cast iron and grey cast iron and within each
family. A high damping ratio corresponds to low hardness
and higher fatigue life. For ADI specimens, the fatigue
behavior is improved by the increase of material damping
that is expressed by the increase of damping ratio. This
indicates the capability of modal testing, as non destructive
tool, to characterize and quantify the fatigue life of
materials.
Acknowledgement
The authors are greatly thankful for Professor. Mo-
hamed Nasr Damir for his noticeable support, valuable
View publication stats
507
instructions and priceless guidance throughout the period
of this work.
ð2Þ
References
[1] Bishop RED, Gladwell GML. An investigation into the theory of
resonance testing. Proc Roy Soc Phil Trans 1963;255:241.
[2] Macnaughtan MP. The cast iron brake disc developments for the 21st
century. MIRA new technology 2002. Available from: http://
www.atalink.co.uk/mira.htm.
[3] Quasar International, Inc. Using Quasar resonant inspection in a
production environment; 2000. Available from: http://www.quas-
arintl.com/.
[4] Doebling SW, Farrar CR, Prime MB, Shevitz DW, Damage
identification and health monitoring of structural and mechanical
systems from changes in their vibration characteristics: a literature
review, Los Alamos National Laboratory Report LA-13070-MS;
1996.
[5] Owolabi GM, Swamidas ASJ, Seshadri R. Crack detection in beams
using changes in frequencies and amplitudes of frequency response
functions. J Sound Vibrat 2003;265:1–22.
[6] Ivan C. Davis, Damage detection in aluminum cylinders using modal
analysis, M.Sc. thesis, Department of Mechanical Engineering,
Virginia Polytechnic Institute and State University; 2002.
[7] El Tobgy MS, The use of modal analysis as a non-destructive testing
tool for predicting fatigue life of components and structures, M.Sc.
Thesis, Production Engineering Department, Alexandria University;
April 2002.
[8] Kessler SS et al. Damage detection in composite materials using
frequency response methods. Composites: B 2002;33:87–95.
[9] Bedewi NE, Kung DN. Effect of fatigue loading on the modal
properties of composite structures and its utilization for prediction of
residual life. Compos Struct 1997;37(3/4):357–71.
[10] Shang D, Barkey ME, Wang Y, Lim T. Effect of fatigue damage on
the dynamic response frequency of spot-welded joints. Int J Fatigue
2003;25:311–6.
[11] Liu D, Gurgenci H, Veidt M. Crack detection in hollow section
structures through coupled response measurements. J Sound Vibrat
2003;261:17–29.
[12] Kim HY. Vibration-based damage identification using recon-
structed FRF in composite structures. J Sound Vibrat
2003;259(5):1131–46.
[13] Yescas-Gonzalez MA, Modeling the microstructure and mechanical
properties of austempered ductile irons, Ph.D. thesis, Department of
Materials Science and Metallurgy, University of Cambridge; 2001.
[14] Damir AN, Characterization of progressive fatigue using modal
analysis, M.Sc. thesis, Production Engineering Department, Alexan-
dria University; 2004.
[15] Bahmani M, Elliott R, Varahram N. The relationship between fatigue
strength and microstructure in an austempered Cu–Ni–Mn–Mo
alloyed ductile iron. J Mater Sci 1997;32:5383–8.
[16] Ductile iron Data, section 4: Austempered ductile iron. Available
from: http://www.ductile.org/didata/section4/4intro.htm; June 2001.
[17] Lin CK, Chang CW. Influence of heat treatment on fatigue
crack growth of austempered ductile iron. J Mater Sci
2002;37:709–16.