Materials and Design xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier.com/locate/matdes

Fatigue life estimation of double lap simple bolted and hybrid

(bolted/bonded) joints using several multiaxial fatigue criteria
F. Esmaeili a,⇑, M. Zehsaz b, T.N. Chakherlou b, S. Barzegar b
a
Department of Mechanical Engineering, College of Engineering, Tabriz Branch, Islamic Azad University, Tabriz, Iran
b
Faculty of Mechanical Engineering, University of Tabriz, P.O. Box 51666-14766, Tabriz, Iran

a r t i c l e i n f o a b s t r a c t

Article history: In this investigation, the effects of torque tightening on the fatigue life of double lap simple bolted and
Received 16 June 2014 hybrid (bolted/bonded) joints have been studied via experimental and multi-axial fatigue analysis. To do
Accepted 3 November 2014 so, two kinds of joints, i.e. simple double lap and hybrid joints were studied. For each kind of the joints, three
Available online xxxx
sets of specimens were prepared and subjected to tightening torques of 1, 2.5 and 5 Nm and then fatigue
tests were carried out at different cyclic longitudinal load levels. A nonlinear finite element ANSYS code
Keywords: was used to obtain stress and strain distribution in the joint plates due to torque tightening of the bolt
Fatigue
and longitudinal applied loads. Fatigue lives of the specimens were predicted with six different multi-axial
Torque tightening
Hybrid joint
fatigue criteria by means of local stress and strain distribution obtained from the finite element analysis.
Multi-axial Experimental tests revealed that the hybrid joints have higher fatigue life in comparison with the simple
Critical plane bolted joints. In addition, multi-axial fatigue analysis and experimental results revealed that the fatigue life
Finite element analysis of specimens was improved by increasing the clamping force resulting from the compressive stresses
created around the hole. Furthermore, the investigation revealed the positive role of clamping force
resulting from torque tightening on the fatigue life of both simple and hybrid joints.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction An alternative joining method is adhesive bonding, which is

important for the aerospace industry. In order to perceive the
Removable joints such as bolts are often used in aerospace advantages of adhesive bonding in fatigue, two fundamental differ-
industry for connecting various components. However, the pres- ences are important between the two kinds of lap joints. First, in a
ence of geometrical discontinuity as a result of hole drilling in mechanical joint, the overlapping areas are attached to one
bolted joints leads to stress concentration and thus decreases the another at discrete points only, i.e. by the fasteners. Obviously,
fatigue strength of structure under cyclic loading [1,2]. Therefore, severe stress concentrations should occur. However, if the connec-
it is of great importance to reduce the effect of the stress concen- tion is made continuously in the full overlapping area by adhesive
tration and attain enhanced fatigue life [3]. bonding, these stress concentrations do not occur. Because, the
When a nut and bolt are used to join mechanical members adhesive bonded joints do not require holes, and the stress concen-
together, the nut is tightened by applying a torque and a force is trations that exist near the holes of a bolted joint, are avoided
exerted on the nut and then the bolt and nut are pulled towards when using adhesive bonding. Therefore, the stress distributions
each other. This force creates tension in the bolt, which clamps in the joint are relatively uniform compared with those in the
the assembled parts of the joint together. Preload is the technical mechanical joint. Secondly, metallic direct contact between the
term for the tension caused by torque tightening the nut that holds two sheets does not happen in adhesively bonded joints [8].
the assembled part together [4–7]. Previous works revealed that In order to employ the advantages and avoid the disadvantages
the clamping force can reduce the stress concentration at the of adhesive bonding, and mechanical joints, and therefore, to
bolted hole region, and therefore improve the strength of the joint obtain high performance joints, a combination of mechanical joints
considerably [6,7]. Esmaeili et al. [6,7] investigated the effects of (riveted, bolted etc.) and an adhesive, namely hybrid joints, are
bolt clamping force on the fatigue strength of bolted plates. used [8]. Hybrid joints are used in many engineering application
such as aerospace, automotive, and naval industries because of
their better performance compared to simple joints such as
⇑ Corresponding author. Tel.: +98 411 3392492.
adhesively bonded, welded and mechanically fastened joints [9].
E-mail address: [email protected] (F. Esmaeili).

http://dx.doi.org/10.1016/j.matdes.2014.11.003
0261-3069/Ó 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
2 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx

A number of researchers [10–13] have investigated the hybrid Energy criteria have been used in conjunction with the critical
joints. Chan and Vedhgiri [10] conducted experiments with com- plane approach, as proposed by Liu [30], Glinka et al. [28], and
posite joints as well as a parametric study using finite element Liu [30] who calculated the virtual strain energy in the critical
analysis to study the stacking sequence effect on joint strength. plane by the use of crack initiation modes and stress–strain Mohr’s
Solmaz and Topkaya [13] have performed a progressive failure circles. Varani-Farahani (VF) [29] proposed critical plane based
analysis on adhesively, riveted, and hybrid bonded double-lap energy parameters that are weighted by axial and shear fatigue
joints. In addition, there are several reference books dealing with properties of the material.
simple adhesive joints and hybrid joints such as those of da Silva A number of researchers have studied the performance of sim-
et al. [14], and more recently that of da Silva et al. [15]. In a study, ple bolted and hybrid joints under static and fatigue loading condi-
Esmaeili et al. [6] investigated the fatigue strength of simple dou- tions, the effect of tightening torque on the fatigue strength of
ble lap bolted and hybrid (bolted-bonded) joints by means of vol- hybrid joints are not so well documented. Consequently, because
umetric method. According to this study, it was found that, the of complex stress distribution in simple bolted and hybrid joints
predicted fatigue life using the volumetric approach, have very due to the tightening torque of bolt and longitudinal remote load,
good agreement with the experimental results. there is still the need for perfect and reliable assessment of multi-
Pirondi et al. [16] compared hybrid weld-bonded, rivet-bonded, axial fatigue criteria to give a clear description of their efficiency.
clinch-bonded, and simple joints under various conditions, via Hence, the current research has sought to improve the existing
experimental analysis. The effect of the material, geometrical fac- body of knowledge about the performance of these multiaxial fati-
tors, and environment on static strength and energy absorption gue criteria in simple bolted and hybrid joints in general and the
were evaluated through the analysis of variance. effect of torque tightening on the fatigue strength of joints in
It is obvious that most of machines and structures, such as air- particular.
craft structures, in their service lives are subjected to multi-axial
stresses in which two or three principal stresses vary with time; 2. Experimental procedures
i.e., the corresponding principal stresses are out-of-phase or the
principal directions change during a cycle of loading. Therefore, The specimens employed in this investigation were made from
multi-axial fatigue analysis becomes an important tool for estimat- 2024-T3 aluminium alloy with thickness of 2 mm. Table 1 lists the
ing the fatigue strength of these components. The origin of multi- mechanical properties of the aluminium alloy obtained from ten-
axiality in stress is dependent on various parameters such as kind sion (static) tests according to the ASTM: E-8M.
of loading, complex geometry of the mechanical parts and residual Two different kinds of joints i.e. simple double lap and hybrid
stresses or pre-stresses. In order to estimate fatigue strength of the (bolted/bonded) joints were prepared. Configurations and dimen-
components, many multi-axial fatigue criteria have been proposed sions of test specimens were designed and manufactured to con-
in the literature for metals. form as nearly as practical to ASTM: E466-07. Test specimens’
Despite the existence of so many multi-axial fatigue criteria in configurations and dimensions for both kinds of joints are illus-
the literature, engineers are frequently faced some problems in trated schematically in Fig. 1.
application of these criteria to engineering design. One of the most The hybrid joints were manufactured using the structural two
important problems is that most of the multi-axial fatigue criteria component epoxy adhesive, namely Loctite 3421 [31], prepared
can only estimate good fatigue results for proportional loading. The by mechanical mixing of the resin and hardener in equal amount
second problem is the application of these criteria for complex by weight. This adhesive was selected due to its high strength
multi-axial loadings. In order to estimate the fatigue strength of and long working life.
mechanical components under multi-axial loading numerous In order to obtain the tensile stress–strain curve of the adhesive,
strength criteria have been proposed during the last decades. Most several dog-bone specimens were prepared according to ASTM:
of these multi-axial fatigue criteria have been discussed for several D638-10. The adhesives were injected into a mold, and left to be
materials and loading conditions by a number of authors such as cured at room temperature for 24 h. Finally, the prepared speci-
Brown and Miller [17], You and Lee [18], and Wang and Yao [19]. mens were tested on a 100 kN Zwick/Z100 static testing machine
From technical point of view, multi-axial criteria may be cate- with a crosshead speed of 5 mm/min. The engineering
gorized into three main groups [20], namely stress based criteria, stress–strain curve of the adhesive is shown in Fig. 2.
strain based criteria and energy based approaches. As for the stress To eliminate any possible surface scratches, the surface of the
criteria proposed by Susmel and Lazzarin [21] and Crossland [22], plates (specimens) were polished using different grinding (sand)
these are based only on stress and are suitable for high cycle papers with grits of 400, 600 and 1000 at first.
fatigue when the deformation is elastic or the plastic strain is To prepare the specimens, fastener holes with a diameter of
small. The strain criteria such as those of Kandil–Brown–Miller 5 mm, were drilled and reamed in the joint plates. A hex head
(KBM) [23], Fatemi–Socie [24], Li-Zhang [25] and Wang–Brown M5 (class 10.9) steel bolt was used for the mechanical fastening
[26], are appropriate for cases in which there is significant and suitable types of steel washers (DIN 125-A-5.3-st) and nuts
plasticity. The energy based multi-axial criteria such as the (DIN 934-M5-8) were used to prepare the joint as illustrated in
Smith–Watson–Topper (SWT) model [27], Glinka et al. [28] and Fig. 1. Finally, the nut was tightened by using a torque-wrench
Varani-Farahani [29] include both stress and strain terms. up to required amounts of torques, i.e. 1 Nm, 2.5 Nm and 5 Nm.
Multi-axial fatigue criteria, from another point of view, can be
categorized as those criteria which use the critical plane concept
and those do not use this concept. In the multi-axial fatigue criteria Table 1
based on critical plane, initially a material plane on which a com- Mechanical properties of 2024-T3 aluminium alloy.

bination of some stress components has the maximum value is Value

determined. Then, fatigue parameters as combinations of the max- Young’s modulus (GPa) 72
imum shear strain or stress and normal strain or stress on the crit- Yield stress (MPa) 315
ical plane are calculated [24]. The criteria will be checked then on Tensile strength (MPa) 550
this plane. Like to classical models, critical plane models can be Poisson’s ratio 0.33
Elongation (%) 18
stress-, strain- or energy-based.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx 3

(a) Simple bolted joint

(b) Hybrid (bonded- bolted) joint

Fig. 1. Configurations and dimensions of the joints.

40 The prepared bonded joints were left in ambient temperature

for 72 h, in accordance with the adhesive manufacturer suggestion.
In the second step of the preparation of the hybrid joints, the
30
double lap bonded specimens with the adhesive were bolt tight-
Stress (MPa)

ened using the same amounts of torques as the simple bolted

20 joints, i.e. 1 Nm, 2.5 Nm and 5 Nm.

10 2.1. Clamping force measurement

In order to measure the clamping force or bolt pretension

0 resulting from the torque tightening, at different applied torques,
0 0.02 0.04 0.06
a special experimental method was designed using a steel bush
Strain (mm/mm)
that was placed between the nut and the plate. At the bush outer
Fig. 2. Engineering stress–strain curve of Loctite 3421 adhesive. surface, two strain gauges were stuck to measure the compressive
axial strain and so the stress in the bush using Hooke’s stress–
strain law. Having the bush cross-sectional area at hand and the
As mentioned earlier, aluminium alloy 2024-T3 sheets were axial stress, the axial force in the bush and therefore the clamp
used as the adherend to prepare hybrid joints in this investigation. force was determined. The used method and the bush dimensions
The preparation of the hybrid joints has been implemented in are shown in Fig. 3. As shown in Fig. 3b the size of the hole in the
two main steps. Firstly, a double lap bonded joint was constructed. bush was selected to be 5 mm, in order to pass the bolt shank
In order to obtain high strength joint, the joint plates were cleaned through the hole of the bush. And outer diameter of the bush
with acetone and then they were dried, prior to applying the adhe- was carefully selected and calculated to avoid plastic deformation,
sive layer. In order to achieve a constant thickness of adhesive so that the Hooke’s stress–strain law can be used for clamping
layer, 0.5 mm thick sheets were used between the adherends. measurement. The cross sectional area and the sizes of inner and

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
4 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx

(b)

(a) (c)

Fig. 3. (a) Measuring clamping force with the load cell, (b) dimensions of the bush (mm), and (c) the placement of the strain gauges on it.

outer diameters of the bush are almost the same as those in the
bolt washer (which is the standard one). The length of the bush 8000
was selected long to have uniform strain in the middle part (where 7000 Simple Bolted Joint

Bolt Clamping Force (N)

we have stuck the strain gages). Hybrid Joint Fcl = 987.8 T
6000
To calibrate the applied torque and clamping force, torques
5000
were applied in 1 Nm increments from 1 to 7 Nm to the nut using
a torque wrench, and then the axial strains were recorded for each 4000 Fcl = 841.43 T
value of the torques. This test was repeated three times for each 3000
case to obtain the mean value of compressive strains (em), and
2000
determine the corresponding clamping forces using Eq. (1). The
relation between the applied tightening torque and the mean value 1000
of compressive strains for both kinds of joints, are given in Fig. 4. 0
The elastic modulus for the bush material (Ebush) was also experi- 0 1 2 3 4 5 6 7 8
mentally determined in order to obtain the accurate values for Tightening Torque (Nm)
the mean axial clamping force. Fig. 5. Relation between the applied tightening torque and resulting clamping
force.
p
F cl ¼ Ebush Abush em ¼ 204; 188  ð92  52 Þem
4
¼ 89:8  105 em ðNÞ ð1Þ only for measuring the clamping force resulting from the torque
tightening, and it is the same for both kinds of the joints. In addition,
where Abush is the area of the bush cross section. The relation the bush was not used during the fatigue test process. Therefore, it
between the calculated clamping forces and the applied torques for has not any effect on the stress and stress distribution for the
the specimen is shown in Fig. 5. As the figure shows, there is a linear specimens under application of longitudinal cyclic loads during the
relationship between the clamping force and the applied torque. This fatigue tests.
confirms that the bush material is still in its elastic region, even
under the maximum applied torque. Additionally, it is important to
2.2. Fatigue tests
mention that the designed experiment in this investigation was used
In the next step, eighteen specimens were prepared for each
kind of joint. A total number of 36 specimens were tested in this
90
investigation to study the effect of tightening torque on the fatigue
lives of both kinds of joints.
The Mean Value of Compressive

Simple Bolted Joint

80
Hybrid Joint
Fatigue tests were carried out using constant amplitude and
70 y = 11.062x sinusoidal cyclic loads in a servo-hydraulic 250 kN Zwick/Roell
R² = 0.9934
60 fatigue testing machine with a frequency of 10 Hz and stress ratio
Strain×10 -5

(load ratio) of 0.1. Fatigue tests were performed on the simple dou-
50
y = 9.3667x ble lap bolted specimens with three different amounts of tighten-
40 R² = 0.9999
ing torques, i.e. 1 Nm, 2.5 Nm and 5 Nm which created clamping
30 forces equal to Fcl = 976, 2440 and 4880 N respectively, according
20 to the linear equation obtained from Fig. 5. In each case, six fatigue
tests were performed with different maximum remote longitudinal
10
loads to final failure (full separation of specimens).
0 In addition, in the case of double lap hybrid specimens, fatigue
0 1 2 3 4 5 6 7 8
tests were performed with the same tightening torques used in the
Tightening Torque (Nm)
simple bolted specimens, i.e. 1 Nm, 2.5 Nm and 5 Nm which
Fig. 4. Relation between the applied tightening torque and the mean value of created clamping forces equal to Fcl = 840, 2100 and 4200 N respec-
compressive strains. tively, according to the linear equation obtained from Fig. 5.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx 5

2.3. Finite element analysis 600

500
As mentioned earlier, in order to estimate the fatigue strength

Stress (MPa)
of the double lap bolted joints, the distribution of stress and strain 400
should be obtained around the bolt hole, and finally, the number of
300
cycles to failure can be calculated from selected multi-axial fatigue
criteria. 200
In order to obtain the stress distribution in the joint plates for
both kinds of the joints, three-dimensional finite element models 100
were simulated by ANSYS 9.0 general finite element code [32]. 0
All of the adherends and adhesive layer were meshed with eight- 0 0.05 0.1 0.15 0.2
node hexahedral structural solid elements Solid45. The finite Strain (mm/mm)
element mesh of the double lap hybrid joint specimens is pre-
Fig. 7. True stress–strain curve of 2024-T3 aluminium alloy.
sented in Fig. 6. together with its corresponding loading and
boundary conditions. The nodes located at the left edge of the FE
(finite element) model were considered to have all their degrees bolt a linear elastic material relation was assumed with a Young’s
of freedom constrained. Only one quarter of the specimen was modulus of 207 GPa and a Poisson’s ratio of 0.30 as it was observed
modelled, due to double symmetry (with respect to X–Z and X–Y that the bolt material remained in the elastic region when it was
Cartesian planes) and symmetric displacement boundary condition subjected to the maximum applied torque (8 Nm) [33].
was applied to the corresponding planes as shown in the Fig. 6. Numerical analyses were carried out in two main steps includ-
In order to transfer the pressure between the contacting sur- ing the application of the clamping force which was followed by a
faces of the bolt head (or nut) and the plate, flexible-to-flexible longitudinal load to the aluminium alloy. In the first step of load-
contact state was used. Each contact pair consisted of target ele- ing, axial displacement was applied to the bottom face of the bolt
ment and contact element. TARGET 170 was used as a target ele- shank to simulate the clamping force. This process was completed
ment and CONTACT 174 was used as a contact element. The for three initial clamping forces resulting from the different
friction effect between the surfaces of the washer (bolt head) and amounts of tightening torques for both kinds of joints using a trial
Al-alloy plates was included in the FE model using Elastic Coulomb and error method. It should be noted that, the application of a lon-
model with friction coefficient of l = 0.29 which was obtained gitudinal tensile load decreases the thickness of plate due to Pois-
from experimental tests based on the sliding of the washer under son’s effect causing the clamped material to release from the
its own weight on the sloped surface from the Al-alloy plate. Also compression, and thereby a reduction in the bolt tension. So to
based on similar experiments, the friction coefficient was found to apply a displacement boundary condition instead of applying force
be l = 0.4 for the contact between the plates in simple bolted on the bolt shank is mandatory in order to have an accurate simu-
specimens. lation and consider the bolt pre-tension relaxation during applying
To represent the aluminium alloy 2024-T3 stress–strain behav- tensile load to the far end of the plate.
iour, an elastic–plastic multi-linear kinematic hardening material Then the solution was restarted for each state and in the second
model with Von Mises criterion was used. This behaviour of the step, the value of maximum force in each cyclic loading (as in
material was obtained from simple tensile tests and shown in experimental test) was applied to the end of the plate in the model
Fig. 7. The elastic modulus and Poisson’s ratio were measured to as a longitudinal static load. Afterwards, in order to simulate the
be E = 72 GPa and m = 0.33 respectively. Also, for the adhesive layer unloading at the second half cycle of the cyclic loading, the value
the multi-linear isotropic material model was used and the Pois- of minimum force according to the experimental tests in each cyc-
son’s ratio was considered equal to 0.35. Furthermore, for the steel lic loading was applied to the end of the plate. This longitudinal

Fig. 6. Three dimensional finite element model with loading and boundary conditions.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
6 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx

loading and unloading, was performed for two cycles to investigate seen, the most compressive stresses are observed at the edge of
the probable variation of the stress and strain distributions during the hole which increased from 20 to 101 MPa when the clamp-
cyclic loading. ing force increased from 976 to 4880 N in simple bolted joints.
According to the results of the first load step solution of the In the second load step, according to the experimental test
finite element analysis, some beneficial compressive stresses were process, a tensile remote stress was applied to the FE models to
observed near the hole of the plates. The compressive stress con- simulate the tensile loading of the specimens. Therefore, a tensile
tours around the bolt hole of the main plate, created due to 1, (remote) stress equal to the applied cyclic load range was applied
2.5 and 5 Nm tightening torques are shown in Fig. 8. As it can be on the right end of the main plate while the displacement of the

Hybrid Joint Simple Joint

(a) Clamped by T=1 Nm

Hybrid Joint Simple Joint

(b) Clamped by T=2.5 Nm.

Hybrid Joint Simple Joint

(c) Clamped by T=5 Nm.

Fig. 8. Distribution of resultant compressive stress rz in MPa due to tightening torque of (a) 1 Nm, (b) 2.5 Nm and (c) 5 Nm for simple and hybrid joints.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx 7

left end of the connector plates was constrained. The resultant lon- In order to calculate the required parameters to employ multi-
gitudinal stress distribution in the main plate is shown in Fig. 9 for axial fatigue criteria, the data obtained from the non-linear finite
different clamping forces under the application of a maximum element analysis results, were used. The required numerical data
remote stress equals to Smax = 168 MPa. As expected, the maximum such as principal stresses and strains were recorded from the
stress values occurred at the edge of the hole which decreased nodes of the region around the hole (at exactly hole edge and away
from 402 to 378 MPa for simple bolted specimens and 278 to from it), which have the maximum tendency for crack initiation
252 MPa for hybrid specimens when the tightening torque and propagation. The variations of principal stresses for the typical
increased from 1 to 5 Nm. critical node, at the hole edge in the smallest cross section area for

Hybrid Joint Simple Joint

(a) Clamped by T=1 Nm

Hybrid Joint Simple Joint

(b) Clamped by T=2.5 Nm

Hybrid Joint Simple Joint

(c) Clamped by T=5 Nm

Fig. 9. Distribution of longitudinal normal stress rx in MPa due to tightening torque (a) T = 1 Nm, (b) T = 2.5 Nm and (c) T = 5 Nm and applied remote longitudinal tensile
stress of 168 MPa.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
8 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx

both kinds of joints are shown in Fig. 10 with different amounts of Table 2
tightening torques under the application of a maximum remote Aluminium 2024-T3 mechanical properties used in multiaxial fatigue criteria.

stress equals to Smax = 168 MPa. As the figure shows, the state of ry Yield stress 315 (MPa)
stresses in this location is multi-axial. t Poisson ratio 0.3
E Young’s modulus 72 (GPa)
r0f Axial fatigue strength 1110 (MPa)
b Axial fatigue strength exponent 0.11
3. Multi-axial fatigue criteria
e0f Axial fatigue ductility coefficient 0.17
c Axial fatigue ductility exponent 0.5
In this research, to estimate the fatigue life of double lap bolted c0f Shear fatigue ductility coefficient 0.364
joints subjected to different tightening torques, six multi-axial fati- s0f Shear fatigue strength 640 (MPa)
gue criteria, i.e. SWT, Glinka, KBM, FS, Crossland and VF were con- b0 Shear fatigue strength exponent 0.11
sidered. These criteria are discussed briefly in the following c0 Shear fatigue ductility exponent 0.5
paragraphs. Table 2 lists the corresponding material properties
used in these criteria.

300 500
S1-1Nm
S1-1Nm
400 S2-1Nm
250 S2-1Nm
S3-1Nm S3-1Nm
300
200

Stress (MPa)
Stress (MPa)

200
150
100
100
0
First Cycle Second Cycle
50
-100
0 -200 First Cycle Second Cycle
100 200 300 400 500
-50 -300
Hybrid Joint Simple Joint
(a) Clamped by T=1 Nm
300
500
S1-2.5Nm
250 S1-2.5Nm
S2-2.5Nm 400
S2-2.5Nm
200 S3-2.5Nm 300 S3-2.5Nm
Stress (MPa)
Stress (MPa)

150 200
100
100
First Cycle 0
Second Cycle
50
-100
0 -200
First Cycle Second Cycle
100 200 300 400 500
-50 -300
Hybrid Joint Simple Joint
(b) Clamped by T=2.5 Nm
300
S1-5Nm 400
250 S2-5Nm S1-5Nm
S3-5Nm 300 S2-5Nm
200
S3-5Nm
Stress (MPa)

150 200
Stress (MPa)

100
100
50 First Cycle Second Cycle
0
0
200 300 400 500 -100
-50100
First Cycle Second Cycle
-100 -200

Hybrid Joint Simple Joint

(c) Clamped by T=5 Nm
Fig. 10. Variations of principal stresses for the typical critical node with different tightening torques (a) T = 1 Nm, (b) T = 2.5 Nm and (c) T = 5 Nm and applied remote
longitudinal tensile stress of 168 MPa.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx 9

3.1. Smith–Watson–Topper (SWT) 3.4. Fatemi and Socie (FS)

The SWT [27] multi-axial fatigue model is expressed as Fatemi–Socie [24] fatigue model is expressed as:
 
De1 ðr0f Þ
2 Dcmax rn;max s0f _ _
r max
¼ ð2N f Þ2b þ r0f e0f ð2Nf Þbþc ð2Þ 1þk ¼ ð2Nf Þb þ c0f ð2Nf Þc ð9Þ
n
2 E 2 ry 2G

where rmax and De1 are the maximum normal stress and the max- where Dcmax is the maximum shear strain and rn,max is the maxi-
n
imum principal strain range at the critical plane. In this paper, the mum normal on the plane where Dcmax occurs. Also k = 0.3 is the
maximum value of the product, rmax De1
in any node was used. To Fatemi–Socie constant and ry is the tensile yield strength which
n 2
do so, rn and De1 were calculated during cyclic loading in any
max can be obtained from uniaxial and torsional tests. The maximum
node of the FE models and consequently the maximum amount of normal stress term in the Fatemi–Socie parameters enable this cri-
the product of these two parameters was employed in Eq. (2) to cal- terion to take account the mean stresses effect.
culate the predicted fatigue life.
3.5. Crossland
3.2. Kandil, Brown and Miller (KBM)
The Crossland criterion [22] is a stress based multi-axial fatigue
criterion which uses the second invariant of deviatoric stress ten-
The general form of KBM’s [23] parameter is expressed as:
sor and maximum hydrostatic stress in its equation.
Dcmax r0f qffiffiffiffiffiffiffi
þ Sk Den ¼ ð2Nf Þb þ e0f ð2Nf Þc ð3Þ J 2;a þ krH; max ¼ r0f ð2Nf Þb ð10Þ
2 E
qffiffiffiffiffiffiffi 1 1=2
The critical plane of this parameter is the plane of maximum shear J 2;a ¼ pffiffiffi ½ðDr1  Dr2 Þ2 þ ðDr2  Dr3 Þ2 þ ðDr1  Dr3 Þ2  ð11Þ
strain, where Dcmax is the maximum shear strain range and Den is 2 6
the corresponding normal strain range at the critical plane. Also, where J2,a is the amplitude of second invariant of deviatoric stress
Sk is a material dependent constant which is chosen so that the tensor and rH,max is the maximum value of the hydrostatic stress
equation gives the same fatigue life as for uniaxial stresses. In this (sum of the mean and amplitude values of the hydrostatic stress).
work, for Al 2024-T3 Sk = 1 gave the best fit. These values can be In addition, Dr1, Dr2 and Dr3 are the range of the first, second
determined with principal stresses and strains obtained from the and third principal stresses, respectively. Also k is a material depen-
finite element analysis, and therefore, using Eqs. (4) and (5) for dent constant which was assumed to be 1 for this research.
nodes around the bolt hole. In Eqs. (4) and (5), e1, and e3, are the first
and third principal strains respectively. In addition, h1 and h2 in 3.6. Jahed–Varani method (JV)
these equations are indicating loading and unloading of a cycle.
These parameters were determined for every node around the hole The Jahed–Varvani (JV) [34] energy-based model, that considers
(in a zone with a radial distance roughly twice the radius of the the sum of plastic and positive elastic strain energy densities, is
hole) and the maximum value of the left hand side of Eq. (2) was expressed as:
used to predict the fatigue life of the specimens. " #
0 
1 ð1 þ rm n =rf Þ Dc Ds
Dc e1  e3  e  e 
1 3 ðDrn :Den Þ þ :
¼  ð4Þ r0f e0f s0f c0f 2 2
2 2 h1 2 h2 max

Den e1 þ e3  e þ e 
1 3
r0f b 0 c
s0f b0 0 c0
¼  ð5Þ ¼ ð2Nf Þ þ ef ð2N f Þ þ ð2Nf Þ þ cf ð2N f Þ ð12Þ
2 2 h1 2 h2 E G
where the normal mean stress rmn acting on the critical plane is
given in:
3.3. Glinka
1
Glinka et al. [28] proposed a fatigue parameter by using the
rmn ¼ ðrmax
n þ rmin
n Þ ð13Þ
2
summation of elastic and plastic energy densities on the critical
In Eq. (12) Dc and Den are calculated from Eqs. (6) and (7), and rn is
shear plane.
the normal stress that was assumed equal to the first principal
2 stress. Eq. (12) was solved for data of the critical node in which
Dc Ds Den Drn ðr0f Þ Ee0f
: þ : ¼ ð2Nf Þ2b þ ð2Nf Þbþc ð6Þ the left hand side of the equation has achieved its maximum value.
2 2 2 2 2E 2
where Dc, Ds, Den and Drn are the range of shear strain, shear 4. Results and discussion
stress range, normal strain range and normal stress range on the
critical plane respectively. These values can be determined with A series of experimental tests were fulfilled to investigate the
principal stresses and strains obtained from the finite element anal- effects of torque tightening on the fatigue life of double shear lap
ysis, and therefore, using Eqs. 4, 5, 7 and 8 for nodes around the bolt joints subjected to a longitudinal load. Three values of tightening
hole. The maximum value of the left hand side of Eq. (5) among the torques were chosen to be applied to the bolt for tightening the
calculated values of the nodes can be used to predict the fatigue life joint. To do so, the corresponding clamping force, i.e. Fcl = 976,
of specimens. 2440 and 4880 N for simple bolted joints, and Fcl = 840, 2100 and
r  r  r  r  4200 N for hybrid joints were applied on the plates. Therefore, a
1 3 1 3
Ds ¼  ð7Þ displacement boundary condition in the Z direction was applied
2 h1 2 h2
r þ r  r þ r  on the lower face of the bolt shank to achieve the desired clamping
1 3 1 3
Dr n ¼  ð8Þ forces equal to the experimental test results. The magnitude of the
2 h1 2 h2
required displacement was found after a few trial and error
where r1, and r3 are the biggest and smallest principal stress values processes to achieve the desired reaction forces or the clamping
respectively. forces resulting from the tightening torques.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
10 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx

In a general trend, increasing the tightening torque improves with tightening torque of T = 1 Nm subjected to a maximum remote
the fatigue life. In addition, a finite element analysis was used to longitudinal load equal to 7.2 kN is shown in Fig. 12.
obtain the stress distribution in the joint plates due to clamping Furthermore, in the current investigation the finite element
force and longitudinal applied loads in order to have a detailed simulation results were used to estimate the fatigue lives based
clarification of the joint’s performance. For better illustration and on selected multi-axial fatigue criteria and also for comparison
comparison purpose, the effect of tightening torque on the fatigue with the experimentally obtained fatigue lives. The predicted fati-
strength of specimens, the resulting fatigue lives were displayed in gue lives and the experimental lives are plotted together in
a semi-log S–N diagram in Fig. 11 for both kinds of joints in accor- Figs. 13–15, for specimens tightened with different amounts of
dance with ASTM 468-11. tightening torques equal to 1, 2.5 and 5 Nm respectively. In
The results shown in Fig. 11 indicate that increasing the tight- Figs. 13–15, the horizontal axis (x-axis) is the experimental life
ening torque, leads to a significant enhancement in the fatigue and the vertical axis (y-axis) is the predicted life. Both lives are
strength as expected. The following can be pointed out according in log scale. The solid line indicates that the predicted results are
to the obtained results. Numerical simulations and experimental equal to the experimental tests results. Two other bounds are also
results revealed that the fatigue life of joints were improved by plotted with dashed lines. The inner bound is in accordance with
increasing the tightening torque due to compressive stresses which the life factor of 2. The outer bound is in accordance with the life
appeared around the hole. The improvement in fatigue life can be factor of 10.
attributed to the way in which the joint transmits the applied load. From Fig. 13, it can be seen that for specimens with tightening
As the tightening torque is increased, a large part of the load is torques equals to 1 Nm, the Crossland, SWT, and FS criteria agree
transmitted by friction (at the plate faces). The remaining part of with the experimental tests results very well. According to
the load is transmitted by the bearing (through the bolt hole sur- Fig. 13, nearly all of the points fall in the range of life factor 2. Also,
face). In this way, the stress concentrations caused by the bearing it can be seen that the prediction made by Glinka criterion for high
of the bolt on the edges of the holes are significantly reduced, cycle fatigue data mostly fall in the range of life factor 2. On the
which is beneficial with respect to any failure at the edge of the other hand, the low cycle fatigue lives were overestimated. The
bolt hole. In other words, as the clamping force value increases, KBM criterion overestimates all range of the lives from low to high
the maximum tensile stress value at the edge of the hole is lives for all kinds of specimens. Fig. 13 also shows that the low
decreased. This is due to higher clamping force value, which has
a significant effect in reducing the maximum stress at the edge
of the hole and at same time developing friction force on contact
surfaces. They reduce the severity of the stress concentration at
the edge of the hole.
The improvement in fatigue life can also be related to compres- Fretting
sion around the hole caused by the compression of the plates by
the bolt pretension. As mentioned earlier, such negative stresses
can reduce the total amount of resultant stresses that cause fatigue
crack initiation and propagation in the plate due to the applied ten-
sile external loads.
In addition, in the hybrid joints, the stress concentration around
the hole is reduced significantly. As a result, the local stress at the
edge of the hole is lower, since some portion of the total load is
transmitted by the adhesive layer.
Finally, the comparison of the obtained results from the experi-
mental tests, confirms that the hybrid joints have better fatigue
performance than simple joints for all levels of the tightening tor-
que. In addition, it should be noted that, in the hybrid joints, direct
metallic contact between the plates does not occur (due to the adhe- (a) simple bolted
sive layer presence), therefore, the possible fretting fatigue (that
causes an early fatigue crack initiation) is eliminated. The typical
fractured specimens for simple and hybrid joints under fatigue tests
Maximum Applied Stress (MPa)

240
Hybrid Joint: Clamped by T=1N.m
220 Hybrid Joint: Clamped by T=2.5 N.m
Hybrid Joint: Clamped by T=5N.m
200 Simple Bolted Joint: Clamped by T=2.5 N.m
180 Simple Bolted Joint: Clamped by T=1 N.m
Simple Boltedd Joint: Clamped by T=5 N.m
160
140
120
100
80
60
1.0E+04 1.0E+05 1.0E+06 1.0E+07
Fatigue Lives (Cycles) (b) hybrid
Fig. 11. S–N curve obtained from experimental fatigue tests for both kinds of the Fig. 12. Typical fractured specimens under fatigue tests with tightening torque of
joints. T = 1 Nm that subjected to maximum remote longitudinal load equals to 7.2 kN.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx 11

1.E+08 1.E+08
SWT SWT
Crossland Crossland
1.E+07 KBM 1.E+07 KBM
FS FS
Predicted Fague Life

Predicted Fague Life

Glinka Glinka
1.E+06 VF 1.E+06
VF

1.E+05 1.E+05

1.E+04 1.E+04

1.E+03 1.E+03
1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08
Experimental Fague Life Experimental Fague Life
(a) Simple Joint (b) Hybrid Joint
Fig. 13. Comparison between the predicted and experimental fatigue lives for specimens with tightening torque equals to 1 Nm.

1.E+08 1.E+08
SWT SWT
Crossland Crossland
1.E+07 KBM 1.E+07 KBM
FS FS
Predicted Fague Life

Predicted Fague Life

Glinka Glinka
1.E+06 VF 1.E+06 VF

1.E+05 1.E+05

1.E+04 1.E+04

1.E+03 1.E+03
1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08
Experimental Fague Life Experimental Fague Life
(a) Simple Joint (b) Hybrid Joint
Fig. 14. Comparison between the predicted and experimental fatigue lives for specimens with tightening torque equals to 2.5 Nm.

1.E+08 1.E+08
SWT SWT
Crossland Crossland
1.E+07 KBM 1.E+07 KBM
FS FS
Predicted Fague Life

Predicted Fague Life

Glinka Glinka
1.E+06 VF 1.E+06 VF

1.E+05 1.E+05

1.E+04 1.E+04

1.E+03 1.E+03
1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08
Experimental Fague Life Experimental Fague Life
(a) Simple Joint (b) Hybrid Joint
Fig. 15. Comparison between the predicted and experimental fatigue lives for specimens with tightening torque equals to 5 Nm.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
12 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx

Table 3
Mean absolute values of errors for employed criteria (in percentage).

Criterion Simple joints Hybrid joints

T = 1 Nm T = 2.5 Nm T = 5 Nm T = 1 Nm T = 2.5 Nm T = 5 Nm
SWT 17 50 60 23 37 72
Glinka 34 35 65 40 42 72
KBM 61 68 70 35 63 75
FS 21 37 30 47 37 41
Crossland 11 17 21 12 16 16
VF 54 78 78 45 71 89

cycle fatigue lives are underestimated and the high cycle fatigue near the bolt hole in the bolted and hybrid joints is complex as a
lives are overestimated when using the VF criterion. result of tightening torque of the bolt and longitudinal remote load.
The comparison between the predicted fatigue lives using Consequently, the obtained results in the current study may not be
multi-axial fatigue criteria and experimental fatigue lives for both at the same accuracy level of estimated fatigue lives presented in
kinds of joints with a tightening torque equals to 2.5 Nm are pre- the references such as [19].
sented in Fig. 14. In these specimens, the predicted results using In order to compare more quantitatively the results of multi-
the Crossland criterion fall into the range of life factor 2. Therefore axial fatigue criteria with experimental test results an error index
this criterion predicts the fatigue life very accurately. Also the was used. The error index and mean absolute values of errors are
results of the FS criterion for half of the points are within the range defined as:
of life factor 2 so it predicts fatigue life fairly accurately. However
 
the SWT, Glinka and KBM models do not predict the fatigue life as N exp
accurately as Crossland and FS criteria for this specimen. In addi-
E ¼ log ð14Þ
Nest
tion, as it can be seen in Fig. 14, the VF criterion underestimates
1X n
and overestimates the fatigue lives for low and high cycle fatigue, E¼ jEi j ð15Þ
n i¼1
respectively.
The predicted fatigue lives using multi-axial fatigue criteria ver-
sus experimental fatigue lives for specimens with a tightening tor- where Nexp is the experimental life and Nest is the predicted life.
que equals to 5 Nm are illustrated in Fig. 15. In these specimens, the The mean absolute values of errors in the different employed
predicted results using the Crossland and FS criteria except for two multi-axial fatigue criteria are presented in in Table 3.
points fall into the range of life factor 2. Therefore, these criteria are The error index shows that the Crossland’s criterion has the best
able to estimate fatigue lives with sufficient level of accuracy for accuracy from low to high lives for all kinds of specimens. Because,
these specimens. As the figure shows, the other four criteria i.e. the Crossland criterion takes into account, directly, the mean stress
SWT, Glinka, VF, and KBM overestimate the fatigue lives. effect resulting from the applied tightening torques during
It is worth mentioning that, from a design point of view, the estimating the fatigue life of the investigated specimens. It must
Crossland multi-axial fatigue criterion is better than the other five be noted that, according to Crossland criterion, considering the
criteria, because it does not overestimate the life for all types of compressive mean stress leads to increase the fatigue life of the
specimens. specimens while tensile mean stress reduces the fatigue life.
As mentioned in the introduction, among the technical litera- According to the results obtained from the numerical analysis
tures which are available in the field of the multiaxial fatigue cri- section, the bolt tightening torque, introduces the clamping com-
teria, many investigations can be found regarding to the pressive stresses around the hole of the main plate, which could
evaluation of the performance of multiaxial fatigue criteria for var- be beneficial to delay the crack initiation and propagation. As the
ious materials, notched and unnotched components, different torque is increased the pretension is also increased, so the bolt will
loading conditions and stress states by several authors. According be stretched more and the member material will be placed under
to the reference [29], Varvani-Farahani used a new energy-critical great compression. This compressive pre-stress will be beneficial
plane parameter for fatigue life estimation of various metallic to prevent fatigue cracks initiation and growing (by reducing stress
materials subjected to in-phase and out-of-phase multiaxial fati- intensity factor) from the bolt hole as explained in Ref. [3]. It can,
gue loading conditions. In another investigation conducted by therefore, be found that the fatigue life improvement of specimens,
Wang and Yao [19], the capabilities of several multiaxial fatigue can be estimated successfully using the Crossland multi-axial fati-
criteria have been evaluated for six different materials under vari- gue criteria, due to considering the mean stress effect in this crite-
ous loading conditions. In a separate study Chakherlou and Aba- ria. This is in complete agreement with available results of
zadeh [35], estimated the fatigue lives of aluminium 7075-T6 previously conducted investigation such as those in Refs. [7,22,35].
plate specimens including a pre-treated fastener hole using the Also, Table 3 shows that the KBM’s and VF’s criteria are the most
seven different multiaxial fatigue criteria. Nevertheless, in these inaccurate among the employed multi-axial fatigue criteria inves-
studies, loading conditions were in-phase or standard out-of-phase tigated in this research.
tension–torsion where the local stress and strain could be easily
calculated for the specimens. Moreover, these criteria were vali-
dated for standard specimens like dog bone cylinder or thin walled 5. Conclusions
tubular. In these studies, multiaxial fatigue criteria were evaluated
in the simple and straight states of loading (and local stresses) with In this study, the effects of torque tightening on the fatigue
standard specimens. Consequently, it is difficult to perform a com- strength of double lap simple bolted and hybrid joints were inves-
prehensive comparison on the existing multiaxial fatigue criteria in tigated via experimental and multi-axial fatigue analyses.
the literature and estimated fatigue lives of the double lap simple Based on the obtained results, the following conclusions can be
bolted and hybrid joints. Because, the stress and strain distribution drawn from the experimental and multi-axial fatigue analyses:

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003
 F. Esmaeili et al. / Materials and Design xxx (2014) xxx–xxx 13

 According to the obtained results from the test of the two differ- [11] Kelly G. Quasi-static strength and fatigue life of hybrid (bonded/bolted)
composite single-lap joints. Compos Struct 2006;72:119–29.
ent specimens consisting of simple bolted joints and hybrid
[12] Barut A, Madenci E. Analysis of bolted–bonded composite single-lap joints
joints, the later joints have longer fatigue lives compared to under combined in-plane and transverse loading. Compos Struct
the simple bolted joints. 2009;88:579–94.
 Increasing the tightening torque or clamping force on the joint [13] Solmaz MY, Topkaya T. Progressive failure analysis in adhesively, riveted, and
hybrid bonded double-lap joints. J Adhes 2013;89:822–36.
leads to an enhancement of the fatigue strength of double lap [14] da Silva LFM, Öchsner A, Adams RD. Handbook of adhesion
bolted joints. technology. Springer; 2011.
 The improvement in fatigue life can be attributed to the way in [15] da Silva LFM, Pirondi A, Öchsner A. Hybrid adhesive joints. Springer; 2011.
[16] Pirondi A, Moroni F. Clinch-bonded and rivet-bonded hybrid joints:
which the joint transmits the applied load. As the tightening application of damage models for simulation of forming and failure. J Adhes
torque is increased, a large part of the load is transmitted by Sci Technol 2009;23:1547–74.
friction. The remaining part of the load is transmitted by bear- [17] Brown MW, Miller KJ. Two decades of progress in the assessment of multiaxial
low-cycle fatigue life. In: Amzallag C, Lewis B, Rabbe P, editors. Low-cycle
ing. The improvement in fatigue life can also be related to com- fatigue and life prediction, ASTM STP 770. American Society for Testing and
pression around the hole caused by the compression of the Materials; 1982. p. 482–99.
plates by the bolt pretension. Such negative stresses can reduce [18] You B-R, Lee S-B. A critical review on multiaxial fatigue assessments of metals.
Int J Fatigue 1996;18:235–44.
the total amount of resultant stresses that cause fatigue crack [19] Wang Y-Y, Yao W-X. Evaluation and comparison of several multiaxial fatigue
initiation and propagation in the plate due to the applied tensile criteria. Int J Fatigue 2004;26:17–25.
external loads. Furthermore, in the hybrid joints, direct metallic [20] Shamsaei N, Fatemi A. Effect of hardness on multiaxial fatigue behaviour and
some simple approximations for steels. Fatigue Fract Eng Mater Struct
contact between the plates does not occur; therefore, the possi-
2009;32:631–46;
ble fretting fatigue is eliminated. Susmel L, Lazzarin P. Fatigue Fract. Eng. Mater. Struct. 2002;25:63–78.
 Among the applied criteria, the Crossland’s criterion has the [21] Susmel L, Lazzarin P. A bi-parametric Wöhler curve for high cycle multiaxial
best accuracy over low to high lives for all types of the fatigue assessment. Fatigue Fract Eng Mater Struct 2002;25:63–78.
[22] Crossland B. Effect of large hydrostatic pressures on the torsional fatigue
specimens. strength of an alloy steel. In: Proceedings of the international conference on
 The KBM’s and VF’s criteria show rather inaccurate prediction of fatigue of metals. London: Institution of Mechanical Engineers; 1956. p.
fatigue life among the employed multi-axial fatigue criteria for 138–49.
[23] Kandil FA, Brown MW, Miller KJ. Biaxial low-cycle fatigue fracture of 316
double lap bolted joints investigated in this paper. stainless steel at elevated temperatures. In: Book 280. London: The Metals,
Society; 1982. p. 203–10.
[24] Fatemi A, Socie DF. A critical plane approach to multiaxial fatigue damage
including out-of-phase loading. Fatigue Fract Eng Mater Struct 1988;11:
149–65.
References [25] Li J, Zhang Z-p, Sun Q, Li C-w, Qiao Y-j. A new multiaxial fatigue damage model
for various metallic materials under the combination of tension and torsion
[1] Chakherlou TN, Mirzajanzadeh M, Abazadeh B, Saeedi K. An investigation loadings. Int J Fatigue 2009;31:776–81.
about interference fit effect on improving fatigue life of a holed single plate in [26] Wang CH, Brown MW. A path-independent parameter for fatigue under
joints. Eur J Mech A Solids 2010;29:675–82. proportional and non-proportional loading. Fatigue Fract Eng Mater Struct
[2] Alegre JM, Gutiérrez-Solana F, Aragón A. A finite element simulation 1993;16:1285–97.
methodology of the fatigue behaviour of punched and drilled plate [27] Smith RN, Watson P, Topper TH. A stress strain function for the fatigue of
components. Eng Fail Anal 2004;11:737–50. metal. J Mater 1970;5:767–78.
[3] Chakherlou TN, Abazadeh B, Vogwell J. The effect of bolt clamping force on the [28] Glinka G, Shen G, Plumtree A. A multiaxial fatigue strain energy density
fracture strength and the stress intensity factor of a plate containing a fastener parameter related to the critical plane. Fatigue Fract Eng Mater Struct
hole with edge cracks. Eng Fail Anal 2009;16:242–53. 1995;18:37–46.
[4] Oskouei RH, Chakherlou TN. Reduction in clamping force due to applied [29] Varvani-Farahani A. A new energy-critical plane parameter for fatigue life
longitudinal load to aerospace structural bolted plates. Aerosp Sci Technol assessment of various metallic materials subjected to in-phase and out-of-
2009;13:325–30. phase multiaxial fatigue loading conditions. Int J Fatigue 2000;22:295–305.
[5] Shankar K, Dhamari R. Fatigue behaviour of aluminium alloy 7075 bolted joints [30] Liu KC. A method based on virtual strain-energy parameters for multiaxial
treated with oily film corrosion compounds. Mater Des 2002;23:209–16. fatigue. In: McDowell DL, Ellis R, editors. Advances in Multiaxial Fatigue, ASTM
[6] Esmaeili F, Chakherlou TN, Zehsaz M, Hasanifard S. Investigating the effect of STP 1191. American Society for Testing and Materials; 1993. p. 67–84.
clamping force on the fatigue life of bolted plates using volumetric approach. J [31] NN. Technical Data Sheet, Product 3421, Loctite Corp. Dublin; 2000.
Mech Sci Technol 2013;27:3657–64. [32] Swanson Analysis Systems Inc., ANSYS, Release 9; 2004.
[7] Esmaeili F, Chakherlou TN, Zehsaz M. Prediction of fatigue life in aircraft [33] Oskouei RH. An investigation into bolt clamping effects on distributions of
double lap bolted joints using several multiaxial fatigue criteria. Mater Des stresses and strains near fastener hole and its effect on fatigue life. MSc thesis,
2014;59:430–8. University of Tabriz, Chapter 3; 2005.
[8] Fu M, Mallick PK. Fatigue of hybrid (adhesive/bolted) joints in SRIM [34] Jahed H, Varvani-Farahani A. Upper and lower fatigue life limits model using
composites. Int J Adhes Adhes 2001;21:145–59. energy-based fatigue properties. Int J Fatigue 2006;28:467–73.
[9] Gómez S, Oñoro J, Pecharromán J. A simple mechanical model of a structural [35] Chakherlou TN, Abazadeh B. Estimation of fatigue life for plates including pre-
hybrid adhesive/riveted single lap joint. Int J Adhes Adhes 2007;27:263–7. treated fastener holes using different multiaxial fatigue criteria. Int J Fatigue
[10] Chan WS, Vedhagiri S. Analysis of composite bonded/bolted joints used in 2011;33:343–53.
repairing. J Compos Mater 2001;35:1045–61.

Please cite this article in press as: Esmaeili F et al. Fatigue life estimation of double lap simple bolted and hybrid (bolted/bonded) joints using several mul-
tiaxial fatigue criteria. J Mater Design (2014), http://dx.doi.org/10.1016/j.matdes.2014.11.003