fatiga multiaxial
TRANSCRIPT
-
8/19/2019 fatiga multiaxial
1/13
Article
Multiaxial fatigue analysis of stranded-wire helical springs
Hossein Darban1, Mostafa Nosrati1 and
Faramarz Djavanroodi1,2
Abstract
In this paper, finite element method is implemented to model stranded-wire helical springs under differentloading conditions. Finite element results are coupled with multiaxial fatigue criteria such as Fatemi–Socie
and Kandil–Brown–Miller together with a uniaxial fatigue criterion, Coffin–Manson, to predict fatigue lifeof the stranded-wire helical springs. It is shown that due to damping effects between wires, stranded-wirehelical springs have longer fatigue life compared to their equivalent single-wire helical springs at a similarcondition. It is also demonstrated that fatigue life is longer for loadings with higher initial displacement of spring head. As practical examples, fatigue life of stranded-wire helical springs with 9 and 15 wires areestimated and compared. It is shown that the spring with 15 wires gives longer fatigue life. It is alsoobserved that Kandil–Brown–Miller and Fatemi–Socie criteria give the least and the highest fatigue lifeprediction, respectively.
Keywords
Stranded-wire helical spring, multiaxial fatigue criterion, frictional force, finite element method
Introduction
A stranded-wire helical (SWH) spring is constructed from twisting of several wires together. When
an axial load is applied at the end of a SWH spring, the strand will be subjected to a twisting
moment. In the case of tensile loading, if both of the strand and the coil have the same turn of helix,
the resulting twisting moment tends to tightly stick the wires together (Peng et al., 2012).
Consequently, if the helix of the strand be opposite in direction to the helix of the spring, the
twisting moment tends to unwind the strand adversely. The outstanding characteristic of the
SWH spring is the inherent tendency for damping of high velocity displacement of its coil due to
friction between wires (Min and Wang, 2007; Phillips and Costello, 1979). As a result, SWH springs
have longer fatigue life in comparison with single-wire helical springs (Min and Wang, 2007;
International Journal of Damage
Mechanics
0(0) 1–13
! The Author(s) 2014
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1056789514560914
ijd.sagepub.com
1Department of Mechanical Engineering, Iran University of Science & Technology, Tehran, Iran2Department of Mechanical Engineering, Prince Mohammad Bin Fahd University, Al Khobar, Saudi Arabia
Corresponding author:
Hossein Darban, Department of Mechanical Engineering, Iran University of Science & Technology, Tehran 16887, Iran.
Email: [email protected]
-
8/19/2019 fatiga multiaxial
2/13
Phillips and Costello, 1979). Hence, SWH springs are used frequently in many industrial cases (Clark,
1961; Costello and Phillips, 1979; Phillips and Costello, 1979). On the other hand, due to manufactur-
ing difficulties for SWH springs with more than three wires, the central wires must firstly be con-
structed and other wires have to be wound around this core. Moreover, the shape of this core depends
on number of wires (Clark, 1961). It is to be noted that, several works have been carried out to
calculate the geometric parameters of SWH springs (Wang et al., 2008, 2010; Zhang et al., 1999).Wang et al [7] proposed two mathematical models for determination of the twist angle and diameter of
the strands. Kunoh and Leech (1985) showed that for the case of small helix angle, the curvature plays
major role in the cross section shape of a strand and also in the contact position of adjacent wires.
During the past few decades, various fatigue criteria have been proposed to predict the fatigue life
of mechanical elements with assumption of uniaxial stress state. However, in many practical cases,
complicated geometries and loading types do not permit the researchers to analyze the problem in
the framework of uniaxial stress state. Therefore, various methods and criteria have been suggested
to describe and predict fatigue life of different materials under practical condition (Brown and
Miller, 1973; Eyercioglu et al., 1997; Fatemi and Socie, 1988; Glinka et al., 1995; Hanumanna
et al., 2001; Jan et al., 2012; Kandil et al., 1982; Kim and Kang, 2008; Lee et al., 2009;
McDiarmid, 1994; Macha and Sonsino, 1999; Navid Chakherlou and Abazadeh, 2011;Papadopoulos et al., 1997; Varvani-Farahani, 2000; Wang and Yao, 2004, 2006; You and Lee,
1996). These criteria use different approaches for prediction of fatigue life. From one aspect, multi-
axial fatigue criteria are categorized to stress, strain, or energy based (Kandil et al., 1982). As
another classification, some of them are formulated based on critical plane concept while others
are not (Kandil et al., 1982). The principles of critical plane theory are firstly proposed by Brown
et al. for multiaxial fatigue problems (Eyercioglu et al., 1997). The critical plane is a plane in which a
specific parameter in a fatigue criterion meets its maximum value and other parameters of the
criterion must be calculated in this plane (Chen et al., 1999; Del Llano-Vizcaya et al., 2006; Kim
and Park, 1999; Muralidharan and Manson, 1988; Pan et al., 1999; Wang and Brown, 1993). Del
Llano-Vizcaya et al. (2006) applied several multiaxial fatigue criteria such as Fatemi–Socie (FS)
criterion (Fatemi and Socie, 1988) and Wang–Brown criterion (1993) to a single-wire helical springand found good agreement between predicted fatigue lives and experimental results.
In the present study, the well-known multiaxial fatigue criteria such as, Fatemi–Socie (FS) and
Kandil–Brown–Miller (KBM) together with a uniaxial fatigue criterion, Coffin–Manson (CM), are
used to investigate fatigue life of SWH springs with 9 and 15 stranded wires with the complete and
comprehensive discussion presented for differences between the results. Comparison of the results
showed excellent advantages for application of a SWH spring against its equivalent single-wire
helical (ESWH) spring under severe harmonic loadings. Despite the fact that there is no experimen-
tal data about fatigue lives of SWH springs in the literature, the reliability of the present method is
discussed by comparing obtained results with experimental data on the fatigue lives of single-wire
helical springs (Del Llano-Vizcaya et al., 2006). The given numerical results as discussed in this
paper would be useful for interpretation of future experimental tests.
Problem definition
Because of extensive application of SWH springs with 9 and 15 wires in industrial equipments under
external harmonic loadings, their fatigue problems are fully studied and discussed in this work.
Figure 1 illustrates the initial configuration of the cross sections and wires diameter used in this
work. As shown in Figure 1, for the case of spring with 9 wires, 3 wires are placed at the center and
other wires are wounded around the core but for SWH spring with 15 wires, the core is constructed
2 International Journal of Damage Mechanics 0(0)
-
8/19/2019 fatiga multiaxial
3/13
of only 1 wire. Furthermore, it can be seen that the outer wires have larger diameter in comparison
with the inner ones for stability considerations. It should be noted that, constructing SWH springs
without the little gap between outer wires is very difficult and practically impossible. But when
applied loading reaches a specific value, the gap disappears and outer wires will be in contact
with each other. Not only all of the dimensions are derived from the actual SWH springs which
are frequently used in industry, but also some other effective parameters such as material properties,
boundary and loading conditions and contact between wires are modeled very similar to the prac-
tical situation. Geometric parameters of the springs with 9 and 15 wires are also given in Table 1.
The wires are made of steel CK101 with mechanical and strain-life properties that are given in
Table 2. In this table, 0 f and "0
f are axial fatigue strength and ductility coefficients while b and c are
axial fatigue strength and ductility exponents, respectively. Respectively, 0
f
, 0
f
, b0, and c0 have the
same definition as 0 f , "0
f , b, and c but for torsional loading.
The mesh patterns used for finite element analysis of SWH springs with 9 and 15 wires are shown
in Figure 2. Since the actual SWH springs are quite long which makes their modeling difficult and
lengthy process, only two coils of each spring have been modeled and other parameters such as
stiffness and displacement amplitude have been chosen accordingly. The critical points for the
springs are also illustrated in this figure. Critical point is a point that experiences highest equivalent
stress during a loading cycle. From Figure 2, it can be observed that the critical points are located
in the inner surface of the strands, which is consistent with the experimental reports
Figure 1. Initial configuration of the cross sections of SWH springs with (a) 9 and (b) 15 wires.
Table 1. Geometric parameters of the SWH springs with 9 and 15 wires.
Spring type
Number of
coils
Length
(mm)
Average coil
diameter (mm)
Strand
diameter
(mm)
Pitch of the
spring (mm)
Pitch of the
strand (mm)
9 wires 24 374 29.5 7.6 22 88
15 wires 40 797.2 29.5 6.5 35 155
Darban et al. 3
-
8/19/2019 fatiga multiaxial
4/13
(Del Llano-Vizcaya et al., 2006). Nearly 850,000 eight-node linear brick elements are used to obtain
appropriate results for SWH spring with 15 wires. The models are created using ABAQUS CAE and
dynamic explicit procedure has been selected for the analysis. In order to justify using this element
type and finite element procedure, it must be pointed out that the levels of equivalent stresses at
critical points for all models are much lower than the yield stress of steel CK101. For instance, as
shown in Figure 2(b), the equivalent stress at critical point is approximately half of the yield stress.
Assuming frictional contact between the wires, penalty approach of contact and friction is applied.
Friction coefficients for steel CK101 are equal to s¼ 0.78 and d ¼ 0.42. It is assumed that the
spring is placed inside of a cylindrical shield and therefore it is not permitted to buckle.
From Figure 2, it is seen that both the strand and the coils of the modeled springs have the same
turn of helix, which makes the springs suitable for tensile loading. It should be pointed out that the
Figure 2. Finite element models of SWH springs with (a) 15 and (b) 9 wires using 3D brick elements.
Table 2. Mechanical and strain-life properties of steel CK101 (Muralidharan and Manson, 1988).
Mechanical properties
Strain-life properties
(M method)
Yield stress (MPa) 1275.000 0f (MPa) 2983.000
Ultimate tensile strength (MPa) 1570.000 b¼ b0 0.120Reduction in area 0.200 "0f 0.309
Fracture strain 0.223 c ¼ c 0 0.600
E (Young’s modulus (GPa)) 205.000 0f (MPa) 1722.000
G (shear modulus (GPa)) 79.500 0f 0.535
4 International Journal of Damage Mechanics 0(0)
-
8/19/2019 fatiga multiaxial
5/13
springs are stretched before the loading. This puts the springs generally under tension during the
loading cycles. As one of the most important results of the present analysis, due to higher resistant
frictional force as a source of energy dissipation, the SWH spring with 15 wires experiences lower
stresses compared to the spring with 9 wires. In order to support this finding, the friction effect
between the wires is magnified by increasing the friction coefficient of wires. It is observed that
higher friction coefficient results in bigger difference between maximum stress in SWH springs with9 and 15 wires. This point confirms that the frictional force between wires is one of the main reasons
that SWH springs experience lower maximum stress when they are constructed with more wires.
It must be noted that, the fretting effect has not been considered in the modeling. Fretting occurs
when two bodies in contact undergo a small relative displacement, typically in the range of 10–30 mm
(Lee et al., 2004). This phenomenon can be affected by many parameters such as contact pressure,
displacement amplitude and coefficient of friction that make it difficult to model (Amiri et al., 2011;
Cruzado et al., 2013; Naidu and Sundara Raman, 2005; Shariyat, 2010).
Multiaxial fatigue criteria
The CM criterion, which has been widely used to estimate fatigue life of mechanical elements underuniaxial loading, is given as follows
"
2 ¼
0 f
E ð2N f Þ
b þ "0 f ð2N f Þc For axial loading
2 ¼
0 f
Gð2N f Þ
b0 þ 0 f ð2N f Þc0 For torsional loading
ð1Þ
where ", , and N f are the strain range in axial fatigue, the shear strain range in torsional fatigue
and the number of cycles to failure respectively. The other parameters in this equation are already
clarified. Although CM criterion is very frequent in use, it leads to incorrect fatigue life for mech-
anical elements under complicated harmonic loadings. For this reason, several multiaxial fatiguecriteria have been proposed (Brown and Miller, 1973; Eyercioglu et al., 1997; Fatemi and Socie,
1988; Glinka et al., 1995; Hanumanna et al., 2001; Jan et al., 2012; Kandil et al., 1982; Kim and
Kang, 2008; Lee et al., 2009; McDiarmid, 1994; Macha and Sonsino, 1999; Navid Chakherlou and
Abazadeh, 2011; Papadopoulos et al., 1997; Varvani-Farahani, 2000; Wang and Yao, 2004, 2006;
You and Lee, 1996). Among these criteria, those, which are based on critical plane concept, are more
attracted by the researches. Having suitable definitions for equivalent stress and strain factors for
simulation of complicated problems by their equivalent uniaxial problems is the main goal of these
criteria. This obviously helps to reduce the mathematical complication and the amount of numerical
analysis. Therefore, major differences between different multiaxial fatigue criteria are the definition
of the equivalent parameters that affects the location and orientation of the corresponding critical
planes.
The KBM criterion introduces critical plane as a plane in which the maximum shear strain range
max occurs. As it can be observed from equation (2), KBM criterion is strain based in which the
maximum shear strain range max and normal strain range "n play major role and have great
influences on the results (Kandil et al., 1982)
max
2 þ S "n ¼
0 f
E ð2N f Þ
b þ "0 f ð2N f Þc ð2Þ
Darban et al. 5
-
8/19/2019 fatiga multiaxial
6/13
where S is the material constant which has been suggested to be equal to 1 for KBM criterion.
Values of max and "n in this equation can be calculated from equations (3a) and (3b) as follows
(Varvani-Farahani, 2000)
max
2 ¼
"1 "3
2
1
"1 "3
2
2
ð3aÞ
"n
2 ¼
"1 þ "3
2
1
"1 þ "3
2
2
ð3bÞ
In the above relations, "1 and "3 are first and third principal strains, respectively. Moreover, 1and 2 show the cyclic loading limits. It is obvious that positive normal stress opens microcrack faces
and consequently leads to early fatigue failure. From equation (4), in the FS criterion maximum
normal stress is used instead of normal strain (Fatemi and Socie, 1988). As it can be seen, for
positive normal stresses, the left hand side of equation (4) increases and therefore the criterion
predicts smaller fatigue life. Furthermore, additional hardening and mean stress effects in the FS
criterion are considered through the normal stress. In addition, the FS criterion has the sameapproach as the KBM criterion in obtaining the critical plane. The FS criterion is given below
max
2 1 þ K
maxn y
¼
0 f
E ð2N f Þ
b þ "0 f ð2N f Þc ð4Þ
where maxn is the maximum normal stress, y is the yield strength and k is a material constant which
can be obtained from uniaxial and torsion fatigue tests, but for most of the materials, it can be
considered equal to 0.6 (Fatemi and Socie, 1988). On the other hand, since the fatigue tests are very
difficult and costly, Muralidharan and Manson (1988) have respectively presented M and MM
methods to approximately attain the material fatigue constants from monotonic tension test data.
Table 3 gives the relations between the strain-life and monotonic properties in which Rm and e f arethe ultimate tensile strength and the true fracture strain, respectively.
On reliability of the method
In this study, finite element method is firstly used to find the critical point. In this manner, a loading
cycle is divided to several increments, then critical points of increments are compared with each
other and finally, the largest equivalent stress is introduced as the critical point. At the next step,
stress information of the critical point is introduced to the computer to predict the fatigue life.
Table 3. The M and MM methods for calculation of the strain-life properties Muralidharan and Manson (1988).
Strain-life parameters
Constants
Axial case M method MM method
Constants
Torsional case
Fatigue strength coefficient 0f 1.9 Rm 0.623 Rm0.823 E0.168 0f ¼ 3
0.5 0f
Fatigue strength exponent b 0.12 0.09 b0¼ b
Fatigue ductility coefficient "0f 0.76 ef 0.6 0.0196 ef
0.155 (Rm E1)0.53 0f ¼ 3
0.5"0f
Fatigue ductility exponent c 0.6 0.56 c 0¼ c
6 International Journal of Damage Mechanics 0(0)
-
8/19/2019 fatiga multiaxial
7/13
It must be noted that, because of using the shear strain increment to find the critical plane in FS and
KBM criteria, the computer program calculates this parameter in all directions around the critical
point for 1 rises in and as shown in Figure 2. Hence, in this procedure, the shear strain range is
calculated and compared for 360 sequential planes around the critical point. In order to assess the
accuracy and applicability of the described method in fatigue problems, this method is used to
predict fatigue life of a single-wire helical spring with specific geometry and material parameterswhich is experimentally investigated by Del Llano-Vizcaya et al. (2006).
Figure 3 depicts the single-wire helical spring used as sample by Del Llano-Vizcaya et al. (2006) in
experimental tests. Number of coils, length, wire diameter, and outside coil diameter were N ¼ 9.5,
L¼ 153.6 mm, d ¼ 5.7 mm, and D¼ 44.4 mm, respectively. Wires of the spring were made of high
carbon steel AISI MB, with mechanical and strain-life properties which are tabulated in Table 4.
The tests were carried out under mean stress m¼ 254.9 MPa with variable stress amplitude a. The
strain-life properties obtained using M method, since this method provides better results than the
MM method (Del Llano-Vizcaya et al., 2006).
Figure 4 shows good agreement between the experimental results and those predicted by the
present method. It must be noted that despite the FS criterion, the CM criterion overestimates
the fatigue life of the spring due to the positive normal stress at the critical point. In addition,the KBM criterion gives conservative results for all loading conditions.
As an important outcome, deviation of the KBM criterion results from the real situations
increases at higher stress amplitudes so that for a¼ 141 MPa the deviatoric error is equal to 0.18
but for a¼ 148 MPa is equal to 0.37. It is also observed that the FS criterion gives the most reliable
results for most of the loading conditions.
Table 4. Mechanical and strain-life properties of high carbon steel AISI MB (Del Llano-Vizcaya et al.,
2006).
Mechanical properties
Strain-life properties
(M method)
Yield stress (MPa) 1350.000 0f (MPa) 3173.000
Ultimate tensile strength (MPa) 1670.000 b¼ b0 0.120
Reduction in area 0.325 "0f 0.434
Fracture strain 0.393 c ¼ c 0 0.600
Young’s modulus (GPa) 177.000 0f (MPa) 1832.000
Shear modulus (GPa) 68.600 0f 0.751
Figure 3. Geometric parameters of a single-wire helical spring.
Darban et al. 7
-
8/19/2019 fatiga multiaxial
8/13
Results and discussions
As a comprehensive and comparative study, Figures 5 and 6 depict the number of cycles to failure
with respect to the displacement amplitude for SWH springs with 9 and 15 wires subjected to initial
head displacement of 10 mm and 15 mm, respectively. Generally, it is seen in these figures that for
most of the loading conditions, the KBM criterion gives the lowest fatigue lives. As it is shown
previously, the SWH spring with 15 wires has lower stress magnitude and therefore would have
longer fatigue life in comparison with the SWH spring with 9 wires. Consequently, at similar loadingcondition, the spring with 15 wires has longer fatigue life than a spring with 9 wires. For example,
according to the FS criterion, fatigue life of a SWH spring with 15 wires subjected to initial residual
displacement and additional displacement amplitude of 10 mm is approximately equal to 3e5 cycles
but for a SWH spring with 9 wires the fatigue life is approximately equal to 1e4 cycles. Hence,
considering this fact that the frictional force between the wires decreases the stress intensity in SWH
springs, it is concluded that using more wires to construct SWH springs, increases resistance of the
springs against the fatigue failure.
Figures 5 and 6 illustrate that the FS criterion predicts higher fatigue life than CM criterion. This
is because, for FS criterion, sign of normal stress on the critical plane has significant influence on the
estimated fatigue life.
Also, it can be seen from Figures 5 and 6 that the initial head displacement of the spring exten-
sively changes spring fatigue life. For illustration of this fact, Figure 7 shows the influences of initial
head displacement on fatigue life of SWH spring with 15 wires for different displacement amplitudes.
It is shown that increase in the initial head displacement causes substantial rise in fatigue life of the
spring. For example, at displacement amplitude of 7.5 mm, fatigue life of the spring subjected to
15 mm of initial head displacement is 20 times greater than the spring with 10 mm of initial head
displacement. It is to be noted that the SWH springs in this study are designed to sustain tension and
therefore, are stretched before the loading. It is understood that for higher amount of initial dis-
placement, the spring experiences higher level of tension during the loading. It means that the wires
105
106
141
142
143
144
145
146
147
148
149
Number of cycles to failure
S t r e s s a m p l i t u d e ( M P a )
Experimental results [31]
FS criterion results
CM criterion results
KBM criterion results
Figure 4. Verification of the results for different criteria with previous experimental data.FS: Fatemi–Socie; CM: Coffin–Manson; KBM: Kandil–Brown–Miller.
8 International Journal of Damage Mechanics 0(0)
-
8/19/2019 fatiga multiaxial
9/13
are stuck together more firmly and consequently, more energy is dissipated due to friction between
the wires. This energy dissipation reduces the intensity of stress in the wires and as a result, the
spring gives longer fatigue life. However, the spring experiences both tension and compression
during a loading cycle in case the displacement amplitude is bigger than the initial displacement
of spring head. In this case, the wires are unwound during the compression and each of them
sustains a portion of the loading separately without any contact with the other wires. This reduces
the efficiency and fatigue life of the SWH spring. Hence, it seems that introducing initial head
displacement is beneficial for the fatigue lives of SWH springs. This finding can be considered for
installation of the SWH springs in practical applications.
102
103
104
105
106
107
0
2
4
6
8
10
12
14
Number of cycles to failure
D i s p l a c e m e n t a m p l i t u d e ( m m )
FS criterion results
CM criterion results
KBM criterion results
104
105
106
7.5
10
12.5
15
17.5
20
22.5
Number of c ycles to failure
D i s p l a c e m e n t a m p l
i t u
d e ( m m )
FS criterion results
CM criterion results
KBM criterion results
(a)(b)
Figure 5. Illustrative plots for fatigue life of SWH springs with (a) 9 and (b) 15 wires subjected to initial displace-
ment of 10 mm.FS: Fatemi–Socie; CM: Coffin–Manson; KBM: Kandil–Brown–Miller.
102
103
104
105
2
3
4
5
6
7
8
9
10
11
D
s p l a c e m
e n t a m p l
t u d e ( m m
)
Number of cycles to failure
FS criterion results
CM criterion results
KBM criterion results
103
104
105
106
107
108
6
8
10
12
14
16
18
20
Number of cycles to failure
D i s p l a c e m e n t a m p l i t u d e ( m m )
FS criterion results
CM criterion results
KBM criterion results
(a) (b)
Figure 6. Illustrative plots for fatigue life of SWH springs with (a) 9 and (b) 15 wires under initial displacement of
15 mm.
FS: Fatemi–Socie; CM: Coffin–Manson; KBM: Kandil–Brown–Miller.
Darban et al. 9
-
8/19/2019 fatiga multiaxial
10/13
Additionally, to show the advantages of the springs constructed from more than one wire over the
ordinary springs in fatigue analysis, an ESWH spring is defined. The wire diameter is chosen as a
dominant parameter for definition of an ESWH spring. The wire diameters of the corresponding
ESWH springs were calculated based on equality of its stiffness with that of a SWH springs. Other
geometric and material parameters for the ESWH spring are equal to those of SWH spring. Stiffness
of single-wire helical springs K can be calculated from equation (5) in which all parameters are
predetermined. Experimental work has been performed to obtain SWH springs stiffness. In this
manner, different tensile force is applied to the ends of SWH springs with 9 and 15 wires and the
corresponding deflections were measured. Figure 8 shows the obtained force–deflection plots for
SWH springs with 9 and 15 wires in which slope of each interpolated line indicates the stiffness of the
104
105
106
107
108
6
8
10
12
14
16
18
Number of cycles to failure (FS criterion)
D i s p l a c e m e n t a m p l i t u
d e ( m m )
initial displacement of 10 mm
initial displacement of 15 mm
Figure 7. Influence of initial displacement on the fatigue life of SWH springs with 15 wires.
FS: Fatemi–Socie.
0 50 100 150 200 250 300 3500
500
1000
1500
Deflection (mm)
F o r c e ( N )
y = 3.9893*x - 98.799
Experimental results
Interpolated line
0 100 200 300 400 500300
400
500
600
700
800
900
Deflection (mm)
F o r c e ( N )
y = 1.1341*x + 366.97
Experimental results
Interpolated line
(a) (b)
Figure 8. Experimental results of the stiffness test for the SWH springs with (a) 9 and (b) 15 wires.
10 International Journal of Damage Mechanics 0(0)
-
8/19/2019 fatiga multiaxial
11/13
corresponding spring. As it can be observed from this figure, stiffness of SWH spring with 9 and 15
wires are equal to 3.9893 N/mm and 1.1341 N/mm, respectively.
By substitution of K , G, D, and N into equation (5), the wire diameters of the corresponding
ESWH springs are obtained equal to 3.97 mm and 3.26 mm for SWH springs with 9 and 15 wires,
respectively.
K ¼ d 4G
8D3N ð5Þ
Predictably, under similar conditions, ESWHS springs show higher equivalent stresses at thecritical points so that it is about 2 and 6 times greater than the SWH springs with 9 and 15
wires, respectively.
Figure 9 shows the fatigue life for SWH spring made of 15 wires and its ESWH spring with respect
to the displacement amplitude under 3 mm of initial head displacement. This figure reveals advantage
of SWH spring in fatigue problems. Unlike single-wire springs, the fatigue life of SWH springs can be
infinite in some cases. As an illustrative example, for displacement amplitude of 1 mm, fatigue life of
ESWH spring is approximately 4e5 cycles but this value for SWH spring is 3.8e9 cycles.
Summary and conclusions
In the presented work, two multiaxial fatigue criteria together with a uniaxial strain-life fatigue
criterion are applied to the SWH springs and their ESWH springs. Loading conditions, geometry,
and material parameters are selected near to practical industrial situation. Finite element method is
employed to find the stress distribution and its related critical points in the spring’s body. Finite
element results are coupled with multiaxial fatigue criteria such as Fatemi–Socie (FS) and Kandil–
Brown–Miller (KBM) together with a uniaxial fatigue criterion, Coffin–Manson (CM) to predict
fatigue life of the SWH springs. Afterward, an ESWH spring is defined with consideration of the
stiffness equality. It is demonstrated that the friction force between the wires in SWH springs act as
resistance force and reduces the stress magnitudes.
109
1010
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
D i s p l a c e m e n t a m p l i t u d e ( m m )
Number of c cles to failure
FS criterion results
CM criterion results
KBM criterion results
105
106
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Number of cycles to failure
D i s p l a c e m e n t a m p l i t u d e ( m m )
FS criterion results
CM criterion results
KBM criterion results
(a) (b)
Figure 9. Fatigue life of SWH springs with (a) 15 wires and (b) their ESWH springs subjected to 3 mm of initial
displacement.
FS: Fatemi–Socie; CM: Coffin–Manson; KBM: Kandil–Brown–Miller.
Darban et al. 11
-
8/19/2019 fatiga multiaxial
12/13
It is shown that longer spring fatigue life is achieved with increasing the number of strands and
this enhancement in fatigue lives of the SWH springs is directly related to the frictional force
between the wires. As another finding, it is concluded that fatigue life is increased for higher initial
head displacement. This result can be applied for installation of the SWH springs in practical
applications. It is also shown that the KBM criterion gives more conservative results in comparison
with other criteria. It is also observed that the KBM criterion is not suitable for single-wire springsunder severe loading. In addition, it is seen that the sign of normal stress plays a significant role in
FS criterion so that for positive normal stresses, the predicted fatigue life reduces while the trend
inverses for negative normal stresses. Comparing the results, fatigue life predicted by FS criterion is
closer to real situations.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit
sectors.
Conflict of interest
None declared.
References
Amiri M, Naderi M and Khonsari MM (2011) An experimental approach to evaluate the critical damage.
International Journal of Damage Mechanics 20: 89–112.
Brown MW and Miller KJ (1973) A theory for fatigue under multiaxial stress–strain conditions. Proceedings of
Institution of Mechanical Engineers 187: 745–755.
Chen X, Xu S and Huang D (1999) Critical plane–strain energy density criterion of multiaxial low-cycle fatigue
life under non-proportional loading. Fracture of Engineering Materials and Structures 22: 679–686.
Clark HH (1961) Stranded-wire Helical Springs, Spring Design and Application, 1st ed. New York: McGraw–
Hill, pp. 92–96.
Costello GA and Phillips JW (1979) Static response of stranded wire helical springs. International Journal of Mechanical Sciences 21: 171–178.
Cruzado A, Leen SB, Urchegui MA, et al. (2013) Finite element simulation on fretting wear and fatigue in thin
steel wires. International Journal of Fatigue 55: 7–21.
Del Llano-Vizcaya L, Rubio-Gonza ´ lez C, Mesmacque G, et al. (2006) Multiaxial fatigue and failure analysis of
helical compression springs. Engineering Failure Analysis 13: 1303–1313.
Eyercioglu O, Walton D and Dean TA (1997) Comparative bending fatigue strength of precision forged spur
gears. Proceedings of Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science
211: 293–299.
Fatemi A and Socie DF (1988) A critical plane approach to multiaxial fatigue damage including out-of-phase
loading. Fracture of Engineering Materials and Structures 11: 149–165.
Glinka G, Shen G and Plumtree A (1995) A multiaxial fatigue strain energy density parameter related to the
critical plane. Fracture of Engineering Materials and Structures 18: 37–46.Hanumanna D, Narayanan S and Krishnamurthy S (2001) Bending fatigue testing of gear teeth under random
loading. Proceedings of Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering
Science 215: 773–784.
Jan MM, Gaenser HP and Eichlsede W (2012) Prediction of the low cycle fatigue regime of the S–N curve with
application to an aluminium alloy. Proceedings of Institution of Mechanical Engineers, Part C: Journal of
Mechanical Engineering Science 226: 1198–1209.
Kandil FA, Brown MW and Miller KJ (1982) Biaxial Low-cycle Fatigue Fracture of 316 Stainless Steel at
Elevated Temperatures, 1st ed. London: The Metals Society, pp. 203–210.
12 International Journal of Damage Mechanics 0(0)
-
8/19/2019 fatiga multiaxial
13/13
Kim KS and Park JC (1999) Shear strain based multiaxial fatigue parameters applied to variable amplitude
loading. International Journal of Fatigue 21: 475–483.
Kim MH and Kang SW (2008) Testing and analysis of fatigue behaviour in edge details: A comparative study
using hot spot and structural stresses. Proceedings of Institution of Mechanical Engineers, Part C: Journal of
Mechanical Engineering Science 222: 2351–2363.
Kunoh T and Leech CM (1985) Curvature effects on contact position of wire strands. International Journal of
Mechanical Sciences 27: 465–470.
Lee H, Jin O and Mall S (2004) Fretting fatigue behavior of Ti–6Al–4V with dissimilar mating materials.
International Journal of Fatigue 26: 393–402.
Li J, Zhang ZP, Sun Q, et al. (2009) A new multiaxial fatigue damage model for various metallic materials
under the combination of tension and torsion loadings. International Journal of Fatigue 31: 776–781.
McDiarmid DL (1994) A shear stress based critical–plane criterion of multiaxial fatigue failure for design and
life prediction. Fracture of Engineering Materials and Structures 17: 1475–1484.
Macha E and Sonsino CM (1999) Energy criteria of multiaxial fatigue failure. Fatigue and Fracture of
Engineering Materials and Structures 22: 1053–1070.
Min JJ and Wang SL (2007) Analysis on dynamic calculation of stranded-wire helical spring. Chinese Journal of
Mechanical Engineering 43: 199–203.
Muralidharan U and Manson SS (1988) A modified universal slops equation for estimation of fatigue. Journal
of Engineering Materials and Technology 110: 55–58.Naidu NK and Sundara Raman SG (2005) Effect of contact pressure on fretting fatigue behavior of Al–Mg–Si
alloy AA6061. International Journal of Fatigue 27: 283–291.
Navid Chakherlou T and Abazadeh B (2011) Estimation of fatigue life for plates including pre-treated fastener
holes using different multiaxial fatigue criteria. International Journal of Fatigue 33: 343–353.
Pan WF, Hung CY and Chen LL (1999) Fatigue life estimation under multiaxial loading. International Journal
of Fatigue 21: 3–10.
Papadopoulos IV, Davoli P, Gorla C, et al. (1997) A comparative study of multiaxial high-cycle fatigue criteria
for metals. International Journal of Fatigue 19: 219–235.
Peng Y, Wang S, Zhou J, et al. (2012) Structural design, numerical simulation and control system of a machine
tool for stranded wire helical springs. Journal of Manufacturing Systems 31: 34–41.
Phillips JW and Costello GA (1979) General axial response of stranded wire helical springs. International
Journal of Non-Linear Mechanics 14: 247–257.Shariyat M (2010) New multiaxial HCF criteria based on instantaneous fatigue damage tracing in components
with complicated geometries and random non-proportional loading conditions. International Journal of
Damage Mechanics 19: 659–690.
Varvani-Farahani A (2000) A new energy-critical plane parameter for fatigue life assessment of various metallic
materials subjected to in-phase and out-of-phase multiaxial fatigue loading conditions. International Journal
of Fatigue 22: 295–305.
Wang CH and Brown MW (1993) A path-independent parameter for fatigue under proportional and non-
proportional loading. Fracture of Engineering Materials and Structures 16: 1285–1298.
Wang S, Lei S, Zhou J, et al. (2010) Mathematical model for determination of strand twist angle and diameter
in stranded-wire helical springs. Journal of Mechanical Science and Technology 24: 1203–1210.
Wang SL, Zhou J and Kang L (2008) Dynamic tension of stranded-wire helical spring during reeling. Chinese
Journal of Mechanical Engineering 44: 36–42.Wang YY and Yao WX (2004) Evaluation and comparison of several multiaxial fatigue criteria. International
Journal of Fatigue 26: 17–25.
Wang YY and Yao WX (2006) A multiaxial fatigue criterion for various metallic materials under proportional
and non-proportional loading. International Journal of Fatigue 28: 401–408.
You BR and Lee SB (1996) A critical review on multiaxial fatigue assessments of metals. International Journal
of Fatigue 18: 235–244.
Zhang YH, Liu HH and Wang DC (1999) Spring Handbook, 1st ed. Beijing: Mechanical industry, pp. 312–329.
Darban et al. 13