Materials Science and Engineering A 527 (2010) 6413–6421

Contents lists available at ScienceDirect

Materials Science and Engineering A

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

Effect of clamping force on fretting fatigue behaviour of bolted assemblies:

Case of couple steel–aluminium
A. Benhamena a,f,∗ , A. Talha b , N. Benseddiq c,d , A. Amrouche d,e , G. Mesmacque c,d , M. Benguediab f
a
Laboratoire LPQ3 M BP 763, University of Mascara, Algeria
b
Haut Ecole d’Ingénieur (HEI), Lille, France
c
Univ Lille Nord de France, University Lille1, F-59000 Lille, France
d
LML, F-59650 Villeneuve d’Ascq, France
e
UA (Université d’Artois), France
f
Department of Mechanical Engineering, University of Sidi Bel Abbes, Algeria

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

Article history: One of the principal failures of bolting assemblies is the fretting. The combination of vibration and service
Received 5 February 2010 load can introduce the damage of fretting fatigue on the contact surfaces of the assembly which lead at a
Received in revised form 26 June 2010 nucleation of crack and then the structure fracture. Contact pressure or clamping force and displacement
Accepted 28 June 2010
at the interface or relative slip are two important factors which control the fretting wear and the fret-
ting fatigue. This paper describes the effect of clamping force (tightening torque) on the fretting fatigue
behaviour of bolted assemblies. Both fretting fatigue experiments and simulation with FEM were carried
Keywords:
out. With the increase of contact force, the sites of cracks initiation changed from the edge of the cen-
Clamping force
Torque
tral hole to the free edge of the contact zone and the fretting fatigue life increase dramatically. A good
Fretting fatigue correlation was found between the FEM simulations and the experimental results.
Crack initiation site © 2010 Elsevier B.V. All rights reserved.
Adhesive wear
Abrasive wear

1. Introduction on structural integrity, the phenomenon of fretting has been exten-

Bolted joints in mechanical structures transmit a more impor-
tant effort in various applications. According to Valtinat et al. [1]
bolted joints have higher tensile and fatigue strengths than welded
joints. The prediction of fracture and the reliability of such assembly
in various practical applications are primordial given their impact
on the economic plan and security. Fretting is caused by the oscil-
lating movement with small amplitude that may occur between
contacting surfaces subjected to vibration. The oscillations cause
sliding to occur in a small region at the edge of contact, while the
center of contact remains stuck together. Fretting causes wear and
very high local stresses to occur at or near the edge of contact
[2], which in turn, can result in the nucleation of cracks and the
reduction of the fatigue life endurance. In high speed train vehi-
cles, the fretting fatigue can be a serious problem and is one of
the costliest sources of in-service damage according to Guo et al.
[3]. This damage is related to cyclic loading and relative displace-
ment at interface in such assembly. Because of its dramatic impact
∗ Corresponding author at: Laboratoire LPQ3 M BP 763, University of Mascara,
Route de Mamounia Mascara, Mascara, Algeria. Fax: +213 045 804 162.
E-mail address: noureddine.benseddiq@univ-lille1.fr (A. Benhamena).
sively studied by various methods [4–10], analytical, numerical and
experimental works have been carried out in order to achieve a bet-
ter understanding of the effect of some local mechanical parameters
such as the material behaviour, the contact pressure, the friction
coefficient, the amplitude of relative displacement and tangential
stress at the interface in order to predict the lifetime of assembly
under fretting. The effect of contact force on the fretting fatigue life
has received considerable attention from many investigations by
several authors such as Nakazawa et al. [11], Switek [12], Anton
et al. [13], Lee et al. [14] and Mall et al. [15]. All theses authors
showed that the increase of contact force lead to decrease of fretting
fatigue life. In contrast, Waterhouse [16] shows that, at low con-
tact force, fretting fatigue life at low cyclic stress levels decreased
with an increase in contact force whereas at high contact force life
increased with an increase in contact force. Fernando [17] affirmed
the observation of later author at lower contact force under fretting.
In the same context, an extensive examination of this subject can
be found in references [18–22] in order to investigate the effect of
relative displacement on the fretting fatigue life. All these authors
observe that the increase of relative displacement reduce signifi-
cantly the life time of component. In addition, a more detailed study
was made by Vingsbo and Soderber [23] in order to identify cor-
rectly the effect of this parameter (partial slip) under fretting. The

0921-5093/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.msea.2010.06.080
6414 A. Benhamena et al. / Materials Science and Engineering A 527 (2010) 6413–6421
Fig. 1. Specimen geometry (all dimensions in mm).
references cited above show that considerable work has been done
to study the problems of fretting fatigue in various industrial appli-
cations. However, relatively few studies have focus only the effect
of clamping force on the fatigue behaviour of bolted plates [3,24,25]
and of bolted assemblies [26,27]. In service and under cyclic load-
ing, the relative displacement at the interface is inevitable for bolted
assemblies, in addition the effect of clamping force on the stress
concentration close to the hole, the frictional stresses and the fric-
tion coefficient are not fully understood and represent a very active
research field. The contact force and the amplitude of relative dis-
placement are two key variables witch to control the fretting fatigue
life in bolted assemblies. This fact has lead the authors to the focus
of the current study, which involves in first part, on experimental
analysis to investigate the effect of clamping force on the fretting
fatigue life of bolted assembly, the sites of crack initiation, the fret-
ting marks were examined by SEM/EDS. Numerical studies were
carried out in order to identify the fields of compressive strain due
the clamping force, the stresses at the interface in contact zone and
to valid the FEM model.
2. Experimental and numerical model
2.1. Experimental procedure
The experimental part of this work is subdivided in three steps:
tension test, fatigue test and microscopic examination (SEM and
EDS).
2.1.1. Materials and specimens
Aluminium alloy A6xxxx from 6.0 mm thickness and high
strength low alloy steel (HSLA355) of 3.0 mm thickness were
used to produce (6.0 mm A6xxxx + 3.0 mm HSLA355) joints for this
investigation (Fig. 1). Tensile specimens were machined with the
dimensions and geometry according to the ASTM B557-06 designa-
Fig. 2. Schematic drawing of central part of the specimen used in fretting fatigue tests and the strain gauges.
Table 1
Mechanical properties of A6xxxx aluminium alloy and HSLA355 at room
temperature.
Alloy  y (MPa)  R (MPa) E (GPa) Elongation
A6xxxx 196 315 71 12%
HSLA355 525 585 200 15%
Fig. 3. Schematic drawing of the Fretting fatigue setup.
tion. The quasi-static tension test was carried out under a crosshead
speed of 0.02 mm s−1 . Tensile strain of specimens was measured
with an Instron® extensometer model 2620-601. The mechanical
properties of the A6xxxx and the HSLA355 sheet materials are listed
in Table 1 [28,29].
2.1.2. Fretting fatigue test
Fretting fatigue test method is shown in Fig. 2. Fretting fatigue
occurs when a component subjected to cyclic loading is in contact
with other component under a normal loading. Surface damage is
induced by the microscopic motion at the contact surface between
these two components which in turn reduces the fatigue life as
compared to plain fatigue without fretting. In present work, the
fretting pad was used to calibrate the fastening torque (clamp-
ing force) in the bolt during the fretting fatigue tests. The fixture,
specially designed, includes steel bolts with a nominal diameter of
10 mm (M10 × 1.5), stainless steel nut, stainless steel washer and a
pad, which is used for fixing strain gauges more easily, in order
to control the clamping force. The pad was manufactured from
steel with circular contact surface at the bottom (Fig. 3) were stuck
ten strain gauges pasted in different positions; CEA-06-062UW-
120 strain gauges (SG) are used to measure the strains. At the pad,
 A. Benhamena et al. / Materials Science and Engineering A 527 (2010) 6413–6421
Fig. 4. (a) Geometrical and finite-element mesh of assembly, (b) modeling of contact surface, (c) modeling of clamping force and (c) mesh near the contact (damaged surface:
position 1 and position 2).
two strain gauges (SG1, SG2) have been glued on axial directions
every 180◦ in order to measure the compressive axial strain due to
the clamping force (fc ) and eight strain gauges (SG3 to SG10) have
been glued on perpendicular plan to axial directions every 45◦ in
order to establish a mean value for the strain due to the cyclic load-
ing (Sc ) during the fretting fatigue test, The gauge locations (SG)
are shown in Fig. 2. The pad was placed between the washer and
the aluminium plate in order to control the crack initiation at con-
tact surface (position 1, see Fig. 3). Software was developed in the
environment (LabVIEU) to detect indirectly the crack initiation in
a hidden area (contact zone) by the treatment of gauge signals, full
details of the test setup are given elsewhere [3]. Fretting-fatigue
tests were conducted using constant amplitude loading using a
sinusoidal waveform in tension–tension mode. The ratio of the min-
imum load and the maximum load or R ratio was 0.1 and the test
frequency was 20 Hz in all the tests. The level of maximum load
is 42 KN, which are approximately 35% of the average peak load
observed in static tension tests of aluminium plate.
Contact pressure or clamping force and displacement at the
interface or relative slip are important factors that control fretting
wear and fretting fatigue. Four values of clamping force (tightening
torque, T = 2, 4, 6, and 8 daN m) was taken in order to determine
the effect of this parameter (tightening torque) and to localize
the zone where the fretting mechanism occur in bolted assem-
blies during the fretting fatigue test. It is known that the strain
is proportional to the clamping force. So, after fretting fatigue test,
the strain was measured using a digital data acquisition system
(LabVIEU) and the Wheatstone bridge theory. Finally, examination
6415
of the wear scars on the joining surfaces between the two plates
(steel–aluminium, position 2) and the interfaces between alu-
minium plate and pad (position 1) were investigated by scanning
electron microscopy/energy dispersive X-ray (SEM/EDX) analysis
in order to characterize the type of fracture damage in contact
surface and to obtain insight into the failure mechanisms.
2.2. Finite element model
Fretting tests may be carried out in the partial slip regime. As
highlighted by Hills and Nowell [30], it is very difficult to achieve
a well controlled experiment using an external actuator due to the
low displacement amplitudes. Although the relative displacement
cannot be easily measured and its value is required, an analysis of
field stress at the process zone (contact zone) in partial slip regime.
Finite element analysis (FEA) is an important tool to design prac-
tical mechanical joints, such as the bolted assemblies. According
to the dimensions of the structure, a three dimensional model was
generated using the commercial software ANSYS® (ANSYS 11 [31])
in order to determine and to perform the analyses of stress field
at contact zone. The bolted assembly shown in Fig. 2 is symmet-
ric in the longitudinal direction (y direction). So, a half model with
symmetry conditions was used in the finite element model FEM in
order to reduce the calculation time (Fig. 4a). A three-dimensional
brick elements (SOLID45) is used for modelling of bolted assem-
bly, this element is defined by eight nodes and each having three
degrees of freedom. In addition, a surface-to-surface contact ele-
ment, which consists of contact elements (CONTAC173) and target
6416 A. Benhamena et al. / Materials Science and Engineering A 527 (2010) 6413–6421

surface elements (TARGE169), is used on the interfaces between all

connected parts of bolted assembly in order to simulate numeri-
cally the contact problems. Friction between the contact surfaces
at the connection is modeled using the classical Coulomb model,
where the friction coefficient was set at 0.2. Various mesh schemes
are tried to achieve convergence. The optimized model has 14,752
nodes, 17,020 elements and 190 contact elements. A two-step non-
linear analysis was performed, in order to simulate the clamping
force numerically, in the first step, the preload (clamping force)
was modeled as a uniform negative pressure applied on the screw
and a positive pressure applied over a ring of 10.mm inner radius
and 16.mm outer radius, this ring represents the action of the nut
(Fig. 4c). This phase represent the joining of bolted assembly. After
this operation, there is a relaxation of the tensile strain due to the
deformability of the pad and the plates under the clamping force
Fig. 5. Clamping force versus tightening torque.
action, on attaining equilibrium and in the second step, this assem-
bly was subjected to cyclic loading that generated a multi-axial
stress fields at contact zone. The bolt shank had a clearance of 1.mm the tightening torque. We note a fairly good agreement between
due to the dimensional tolerance of the boltholes in the two plates the theoretical and experimental results.
(steel and aluminium) and pad. In practice the plates and pad are These results were confirmed by the numerical analysis pre-
tightened by a torque applied to the bolts; this action was simu- sented in Fig. 6. This latter shows the compressive strain axial
lated by transferring the torque as axial strain in the bolt shank during the simulation of fretting fatigue test under maximal cyclic
during the assembly process. Although during cyclic loading there loading equal to 70 MPa and a tightening torque value of 8 daN m.
is a possibility that the two plates or pad surface in boltholes could We also notice in this case that the compressive strain is strongly
contact the surfaces of the bolt shank, the clearance was consid- concentrated at the interface between the pad and the aluminium
ered large enough for this action not to occur in order to simulate plate. This phenomenon can be explained by the fact of strain
the fretting fatigue at position 1 and at position 2. The theory of incompatibility at the interface and the frictional coefficient effect.
incremental plasticity is introduced to modelling the material non- Indeed, under this condition and in position at the pad, the
linearity. The iterative method of Newton–Raphson is used as an compressive strain measured by the strain gauges (SG1, SG2) is
approach to solve nonlinear equations by finite elements. 0.001195. At the same position, the compressive strain is calculated
numerically and its value equal to 0.001273.
3. Results and discussions According to the experimental, theoretical and numerical anal-
ysis, it can be seen that the result gave a good correlation between
3.1. Validation of the finite element model various methods, thus establishing confidence in the results of the
finite element modelling for bolted assemblies.
The amount of tightening torque T required to achieve a set
amount of preload fc (clamping force) in a bolt depends upon thread 3.2. Effect of the torque tightening
pitch, the friction coefficient in threading, bolt diameter, the friction
coefficient between the nut and bolt and the mean collar diameter. The experimental results of fretting fatigue show that the tight-
For the threaded fasteners according to the ISO normalisation, there ening torque has a significant effect on the lifetime in fatigue
is a relationship between the tightening torque and the clamping of the bolted plates (Fig. 7). With increasing torque the life-
force expressed by the following time in fretting fatigue (Nr) decreases abruptly from 1,688,388
cycles for a torque T = 3 daN m to 1,280,820 cycles for T = 4 daN m.
T = (0.161p + 0.58t d + h rm )fc (1)
When the contact pressure increases specifically in range of the
where p is the pitch, d is the major diameter, fc is the clampin force, torque between 4 and 6 daN m lifetime does not vary much. For
t is the friction coefficient between the nut and bolt, h is the higher torque value (T > 6 daN m) lifetime decreases rapidly. In our
coefficient of collar friction and dm is the mean collar diameter
(dm = 1.25d and dm = 2rm ).
Although the torque–preload relationship can be calculated the-
oretically (Eq. (1)). But in practice it is not possible to accurately
measure the bolt tension in a joint in service. So, it is better to mea-
sure both parameters (T, fc ) and hence calibrate the bolt tension
experimentally. To ensure that a desired preload (clamping force)
has been achieved with a bolt, it is more practical to use a torque
wrench to apply the load to the bolt through the nut (see Fig. 2).
The axial strain measured by the strain gauges (SG1, SG2) were
recorded for each applied torque level. The theoretical expression
for the relation between the clamping force and the axial strain is
2 2
de − di
fc = EP εm (2)
4
where di and de are the internal and external diameters of the pad
(see Fig. 2), respectively. Ep is the yuong’s modulus of the pad. εa is
the compressive axial strain mesuerd by SG1 and SG2 .
The theoretical (Eq. (1)) and experimental results are illustrated Fig. 6. Compressive strain at the pad under maximal cyclic loading 70 MPa and
in Fig. 5; this presents the clamping force variation according to torque 8 daN m.
 A. Benhamena et al. / Materials Science and Engineering A 527 (2010) 6413–6421 6417

Fig. 7. Lifetime versus tightening torque.

example, for a torque of 8 daN m the lifetime is 712.446 cycles

only.
The main reason for this variable behaviour is due to tribological
changes in the contact area, occurring because of the variation in
contact force depending on the amplitude of cyclic loading.

3.3. Inspection of damage

Many contact problems can be solved with the help of the Hertz
theory [32] in order to obtain the surface stress distribution and
subsequent bulk stress and strain states. However, when a contact
does not fulfil the Hertz assumptions, the analytical solution may
differ significantly from the real one. This problem has been solved
by Cattaneo [33] and Mindlin [34] for the case of the spherical con-
tact submitted to stick-slip behaviour. This situation happens when
Fig. 8. (a) The distribution of stress  y in the direction of cyclic loading and the
the tangential load (Q) is lower than the product between the nor-
experimental observation. (b) Status of the contact area (position 1) (T = 8 daN m
mal load (p) and the friction coefficient () (Q < p). Globally the and Fmax = 40 KN).
contact is sticking (adhesion), but to respect the Coulomb friction
law over the entire contact interface, a slipping annulus appears at
the periphery of the contact zone. According to these authors, the relative to the sliding direction (direction of cyclic loading). The
contact zone is subdivided in two zones (adhesion and slip zones). numerical analysis (Fig. 8b) shows a transition from the sticking
A numerical calculation has been conducted to simulate the effect contact to the sliding contact characterized by an angle  = 45◦ . For
of the torque on the stress and the strain field. The numerical results a high torque corresponding to the high contact pressure promotes
show that the stress field is tensile near the contact zone (BC) and stress concentration near the contact zone. Therefore, the presence
compressive near the edge of the hole (BT) (Fig. 8a). The damage is of high gradient of pressure can lead to the initiation and crack
greater near the contact zone (BC) than the edge of the hole (BT) growth. Fig. 9 shows the evolution of the status of the contact area
where the damage is lower. Generally, the cracks appear on the (no-contact, sticking and sliding) versus the maximum cyclic load-
edge of contact (BC) in a direction at an angle between 40◦ and 50◦ , ing. The cracking phenomenon of the fretting fatigue occurs in the

Fig. 9. Status of the contact area (position 2) versus to the cyclic loading.
6418 A. Benhamena et al. / Materials Science and Engineering A 527 (2010) 6413–6421

Fig. 10. (a) The distribution of stress  y in the direction of cyclic loading. (b) Status of the contact area (T = 8 daN m and Fmax = 12 KN).

transitional phase Stick-Slip which is considered as the most criti-
cal phase. The tightening torque in this example is T = 8 daN m. We
observe in the contact area (R ≤ 12.5 mm) than the sticking zone
decreases with increasing levels of cyclic loading and disappears
completely from 66% of full load. The transition from sticking to
sliding is the most critical phase in fretting fatigue. Generally, the
cracking occurs in this area.
The distribution of the stress  y under cyclic loading of 12 kN and
a torque of 8 daN m is illustrated in Fig. 10a. The stresses that trig-
ger the initiation of the crack are higher around the contact area,
especially near the contact interface of the steel and aluminium
plate (position 2). The comparison between the experimental
results and the status of the contact (Fig. 10b) leads to the same
conclusion.
In all analysis realized in the following study, the inspection of
damage and the degradation of contact surface has been done for
a cycle number in fretting fatigue of 800000 cycles. In this work,
we can see the appearance of adhesion and slip zones only at lower
values of torque (Fig. 11h) while for higher value; the slip zone is
disappeared completely and appears only the adhesion zone. This
behaviour can be explained by the fact that the decrease of torque
value (clamping force) leads a decreasing of the adhesion zone and
therefore an increasing of the slip zone, which favoured the par-
tial sliding at the interface of bolted assemblies. We note that the
 A. Benhamena et al. / Materials Science and Engineering A 527 (2010) 6413–6421
Fig. 11. Crack propagation at various tightening torque under maximal cyclic loading 70 MPa.
initiation and cracks propagation taken a position at the interface
between the adhesion and slip zones (Fig. 11c–h). One can thus
conclude that the position of crack initiation is related directly on
the change of adhesion zone size. This latter is varied as function of
clamping force level. In other words, for higher torque value, a large
zone of adhesion leading to a slip zone very narrow, the initiation
and crack propagation occurs at the border contact in the interface
or outer edge of the contact region. However, it can be observed that
for the range of means value of torque (T < 6 daN m), which lead to
reduce the size of adhesion zone and to raise the size of slip zone,
consequently, the initiation and crack growth is located within the
contact zone at the interface between the adhesion and slip zones.
This phenomenon can be explained by the fact of change of net pres-
sure acting on each zone (adhesion and slip zones) in the contact
zone. We note also that, whatever the torque value, the frictional
marks take an elliptical shape parallel to the cyclic loading direc-
tion and its size is more pronounce for higher torque value. This
behaviour is in good agreement with experimental observations of
Wagle and Kato [35].
3.4. Wear mechanisms
To better understand the torque effect on the wear mechanism
of such bolted assembly (steel–aluminium) under cyclic loading, a
thorough examination of damage surface at the interface between
6419
the plates (position 2) using SEM-EDS. Because the degradation
and crack initiation was more pronounced in this position for
each torque level according to experimental result (Fig. 11) and
the experimental observations of Wagle and Kato [35]. Fig. 12a
shows the overall topography of aluminium alloy, before the fret-
ting fatigue test (in state virgin). The analysis by EDS (Fig. 12b)
indicates a strong presence of Al, is a typical in this kind of alu-
minium alloy. The SEM micrograph of Fig. 12c (from area ‘B’ of
Fig. 11d) shows the presence of wear debris of the damaged sur-
face of aluminium for a torque value equal to 6 daN m. The damaged
areas is covered by with narrow grooves (typically in the range of
5–50 ␮m wide), wear debris and a number of micro cracks, were
produced at a position and orientation different relative to slid-
ing direction (cyclic loading direction). According to this figure, we
can be noted that the wear mechanism occurred in this case by
adhesion. Comparison of EDS analyses of worn surface (Fig. 12d)
and that of unworn surface (without fretting, Fig. 12b), shows the
loss of aluminium (Al) from damaged surface with increasing of
the torque. So, the intensity of Al peak decreased and that of Fe and
Si peaks increased with increased in torque value (6 daN m). It is
likely that this loss of Al occurs under the action of harder particles
components of Fe and Si of steel specimen under fretting condition.
From the above observations, we showed that the torque value is a
very important effect on the size of adhesion zone (Fig. 11). At very
high torque (8 daN m), we can be note that the morphology of dam-
6420 A. Benhamena et al. / Materials Science and Engineering A 527 (2010) 6413–6421
Fig. 12. SEM micrographs and EDS analysis of aluminium plate under fretting at different tightening torques and without fretting.
aged area (Fig. 12e) is very different compared to a torque value of
6 daN m (Fig. 12c). This means that the wear mechanism is related
to the torque value. In other words, the increasing of torque leads
to an increasing of the frictional stresses and to reduce the relative
slip at the interface of bolted assembly. The interaction of these two
factors (frictional stresses and relative slip) produces a degrada-
tion of contact surfaces during the fretting fatigue test. In the first
step, the degradation of contact surfaces begins by wear mecha-
nism of adhesion, after a some number of fatigue cycles, another
wear mechanism dominates, is the abrasion wear. EDS analysis in
Fig. 12f revealed that intensity of Si peak decreased and that of
Fe peak increased with increase in torque (8 daN m) in compari-
son to that a torque value of 6 daN m (Fig. 12d). This indicates that
the brittle material (HSLA355), generated during fatigue fretting,
probably caused the wear of softer material (A6xxxx). This wear is
observed especially in adhesion zone where the contact stress is
higher. From the experimental observations of our work, it is clear
that the torque magnitude has a determinant effect on the mecha-
nism of degradation of contact surfaces under fretting. This shows
that, at lower torque value the adhesive wear dominates while the
abrasive wear has been dominated at higher torque. We recall that
the abrasive wear is more severe than the adhesive wear [36]. In
conclusion, under cyclic loading at the vicinity of interface, there is
a repeated stress along the sliding zone. So, the damage and micro-
crack are initiated in this zone. Once initiated, the cracks propagate
and intersect with other microcrack, leading to a network of inter-
secting cracks, which produces particle pull-out. These particles are
removed during subsequent motion, resulting in fatigue wear.
This behaviour is in agreement with the experimental obser-
vations of Wagle and Kato [35]. These authors found that for a
bolted assembly of A2024-T3 aluminium alloy, the failure mech-
anism take place at the interface between the plates (position 2)
and is related to the tightening torque magnitude. They found also,
there are three modes of failure of bolted assembly under cyclic
loading. In a range of smaller torques, the mechanical fatigue failure
occurred at the bolt hole edge. For means values of the tightening
torque, the fretting wear failure occurred near the bolt hole due to
the gross sliding of the specimen surfaces while at higher tighten-
 A. Benhamena et al. / Materials Science and Engineering A 527 (2010) 6413–6421
ing torque, the fretting fatigue failure occurred and cracks initiated
at contact border.
4. Conclusions
The aim of this study is to analyze the effect of tightening torque
on bolted assemblies under fretting fatigue condition, from these
experimental and numerical results, we can deduce the following
conclusions:
- The developed finite element modelling approach to simulate
clamping force (tightening torque) in the bolt of such bolted
assembly was validated against the experimental results.
- The wear mechanism and contact surfaces degradation depends
on the magnitude of tightening torque.
- The adhesive wear dominates at tightening torque and fretting
fatigue cyclic number lower while the abrasive wear dominates
for tightening torque and fretting fatigue cyclic number higher.
- The size of the adhesion and slip zones on contact zone is related
at the magnitude of tightening torque.
References
[1] G. Valtinat, I. Hadrych, H. Huhn, Published on Conference: Connections in Steel
Structures IV, AISC and ECCS, Roanoke/USA, 2000.
[2] A. Zografos, D. Dini, A.V. Olver, Tribol. Int. 42 (11–12) (2009) 1663–1675.
[3] R. Guo, R.C. Duan, G. Mesmacque, L. Zhang, A. Amrouche, R. Guo, Mater. Sci.
Eng. A (483–484) (2008) 398–401.
[4] Z.R. Zhou, K. Nakazawa, M.H. Zhu, N. Maruyama, Ph. Kapsac, L. Vincent, Tribol.
Int. 39 (2006) 1068–1073.
[5] C. Vallellano, J. Dominguez, A. Navarro, Eng. Fail. Anal. 11 (2004) 727–736.
[6] A. Vallellano, C. Navarro, J. Dominguez, Eng. Fail. Anal. 16 (2009) 2646–2657.
[7] H. Tang, D. Cao, H. Yao, M. Xie, R. Duan, Eng. Fail. Anal. 16 (2009) 2004–2008.
[8] M. Buciumeanu, A.S. Miranda, A.C.M. Pinho, F.S. Silva, Eng. Fail. Anal. 14 (2007)
810–821.
6421
[9] S. Sathiya Naarayan, D.V.T.G. Pavan Kumar, S. Chandra, Eng. Fail. Anal. 16 (2009)
2255–2273.
[10] S. Griza, F. Bertoni, G. Zanon, A. Reguly, T.R. Strohaecker, Eng. Fail. Anal. 16
(2009) 1542–1548.
[11] K. Nakazawa, N. Maruyama, T. Hanawa, Tribol. Int. 36 (2003) 79–85.
[12] W. Switek, ASTM STP 1367 (2000) 154–166.
[13] D.L. Anton, M.J. Lutain, L.H. Favrow, D. Logan, B. Annigeri, ASTM STP 1367 (2001)
119–140.
[14] S.K. Lee, K. Nakazawa, M. Sumita, N. Maruyama, ASTM STP 1367 (2001)
199–212.
[15] S. Mall, V.K. Jain, S.A. Namjoshi, C.D. Lykins, ASTM STP 1425 (2003) 338–352.
[16] R.B. Waterhouse, ASTM STP 1367 (2000) 3–18.
[17] U.S. Fernando, G.H. Farrahi, M.W. Brown, in: R.B. Waterhouse, T.C. Lindley (Eds.),
Fretting Fatigue. ESIS 18, Mechanical Engineering Publications, London, 1994,
pp. 183–195.
[18] Y. Ochi, Y. Kido, T. Akiyama, T. Mastumura, ASTM STP 1425 (2003) 220–234.
[19] H. Lee, S. Mall, Tribol. Int. 39 (2006) 1213–1219.
[20] O. Jin, S. Mall, Wear 253 (2002) 585–596.
[21] O. Jin, S. Mall, Int. J. Fatigue 24 (2002) 1243–1253.
[22] K. Nakazawa, M. Sumita, N. Maruyama, Fatigue Fract. Eng. Mater. Struct. 17
(1994) 751–759.
[23] O. Vingsbo, S. Soderber, Wear 126 (1988) 131–147.
[24] A. Aragon, J.M. Alegre, F. Gutiérrez-Solana, Eng. Fail. Anal. (2006) 271–281.
[25] T.N. Chakherlou, R.H. Oskouei, J. Vogwell, Eng. Fail. Anal. 15 (2008) 563–574.
[26] C.Y. Park, A.F. Grandt Jr., Eng. Fail. Anal. 13 (2006) 136–154.
[27] T.N. Chakherlou, M. Mirzajanzadeh, J. Vogwell, Eng. Fail. Anal. 16 (2009)
2388–2399.
[28] A. Catimel, F. Savanne, Durabilité en fatigue de structures assemblées par riv-
etage, Rapport interne, Lille Mechanics Laboratory (LML), 2008.
[29] L. Gouin, G. Laudrin, Traitement curatif de fissures de fatigue par perçage et
expansion à froid: application à des structures ferroviaires, Rapport interne,
Lille Mechanics Laboratory (LML), 2008.
[30] D.A. Hills, D. Nowell, Mechanics of Fretting Fatigue, Kluwer Academic publish-
ers, Dordrecht, 1994.
[31] ANSYS v11.0, Ansys Inc. Documentation, ANSYS Elements Reference, (2007).
[32] H. Hertz, J. Reine Angew. Math. 92 (1882) 156–171.
[33] C. Cattaneo, Accad. Nazionale. Rend. Accad. Lincei. 6 (27) (1938), 342–348,
434–436, 474–478.
[34] R.D. Mindlin, ASME J. Appl. Mech. 16 (1949) 259–268.
[35] S. Wagle, H. Kato, Int. J. Fatigue 31 (2009) 1378–1385.
[36] P. Groche, G. Nitzsche, A. Elsen, CIRP Ann. Manuf. Technol. 57 (2008) 295–298.