Content uploaded by Dezhi Li
Author content
All content in this area was uploaded by Dezhi Li on May 19, 2016
Content may be subject to copyright.
Page
1
of
13
10M
-
0249
Effect of setting velocity on self
-
piercing riveting process and joint
behaviour for automotive applications
L. Han
*
,
D. Li,
M. Thornton
Warwick Manufacturing Group, University of Warwick, Coventry, CV4 7AL, UK
M. Shergold
Jaguar Engineering C
entre, Jaguar Land Rover, Coventry, CV3 4LF, UK
* Corresponding author,
Tel: +44(0)2476575385
, Fax: +44(0)2476575
366
, E
-
mail:
li.han@warwick.ac.uk
Copyright © 2010 SAE International
ABSTRACT
T
he increased applica
tion of lightweight materials, such as aluminium has initiated many investigations into
new joining techniques for aluminium alloys. As a result, Self
-
piercing riveting (SPR) was introduced into the
automotive industry a
s the major production process to jo
in aluminium sheet body structures. Although both
hydraulic and servo types of SPR equipment are used by the industry, the
servo type is most commonly used in
a volume production environment. This type uses stored rotational inertia to set the rivet. The i
nitial rotational
velocity of the mass dictates the setting force and hence the tool is described as velocity
-
controlled. A study was
therefore conducted to examine the effect of setting velocity on the process including tooling and joint
performance. It
was found that the setting velocity would have a significant effect on tooling life. Over 80kN
force could be introduced into the tooling depending on selection of the setting velocity. The results also
suggested that the joint quality and strength are aff
ected by the setting velocity. An increase in the setting
velocity would lead to a decrease in the head height and an increase in the interlock. Too low velocity may
cause the rivet head to protrude; too high velocity may result in breakthrough failure. Bo
th cases would lead to
corrosion concerns. The examination also indicated that the higher the setting velocity the higher the shear
strength of the joints. In addition, photoelastic measurement using Deltavison software was also applied to this
investigati
on. The effect of setting velocity on the residual stress distribution of SPR joints is also discussed.
Key words:
Self
-
piercing riveting, setting velocity, joint behaviour
1.
INTRODUCTION
Today’s automotive industry is a challenging business. It is requ
ired not only to respond to environmental
concerns such as greenhouse gases and fuel economy, but also to meet customer expectations. Therefore, a need
for weight reduction has emerged and this in turn has led to the increased application of lightweight ma
terials,
such as aluminium and polymer composites. The use of aluminium alloys offers an opportunity for vehicle
weight reduction, which can lead to a reduction of fuel consumption and emissions without compromising
performance, comfort and safety [1, 2, 3
]. Aluminium alloys can offer high corrosion resistance, good
formability and good crashworthiness. In addition, the recyclability of aluminium alloys is also a considerable
attraction to manufacturers. However, the use of aluminium requires not only a dif
ferent approach in car design
Page
2
of
13
but also a different approach to manufacturing technology and in particular joining methods. As a result, self
-
piercing riveting (SPR) was introduced into the automotive industry
to join aluminium sheet body structures.
SPR a
s a
key production process has many advantages, such as no pre
-
drilled hole requirement, capability to
join a wide range of similar or dissimilar materials and combinations of materials, no fume emissions etc.
There
are currently two types of rivet equipme
nt available; the
electric servo
-
motor type and the hydraulic type. The
hydraulic type is force
-
controlled and uses oil pressure to firstly clamp the material and then to insert the rivet.
The pressures used by this process are high. When riveting thick st
ack
-
ups the force required to fully insert the
rivet can equal the amount of force generated by the hydraulic power pack. For automotive mass production,
one of the specifications is that the riveting tools have to be mounted on tool changers so that more
than one
tool can be used on a single robot. This basically rules out the large scale automated use of the hydraulic tool
due to concerns over running high pressure hydraulic hoses through a tool changer, which would be
continuously engaged and dis
-
engaged
. The electric servo
-
tool was therefore developed to solve these
problems for mass application. By switching to electrical power rather than hydraulic force, the tool can be
mounted on a tool changer allowing multiple tools per robot. This type of equipm
ent uses stored rotational
inertia to set the rivet. The initial rotational velocity of the mass dictates the setting force and hence the tool is
described as velocity
-
controlled. Although previous research [
4
,
5
] reported the force characteristics of the
SPR
process for hydraulic equipment, they either aimed at reducing the operational force [
4
], or focused on
developing a process monitoring system [
5
]. Setting velocity, as a key process parameter for servo equipment,
has not been fully investigated. Consi
dering the majority applications of SPR is velocity
-
controlled, it is
important to know the effects of this key parameter on the process and joint behavior. This paper therefore aims
to give a deep insight of the effect of setting velocity on quality of th
e joints, on tooling of the process and on
performance of such joints.
2.
EXPERIMENTAL PROCEDU
RE
2.1 Quality Criteria
Based on industry standard practice, a joint quality assessment criterion was created. All joints were assessed
for joint quality agains
t the primary measurements of; head height, interlock and remaining material thickness
of the bottom sheet, together with other secondary considerations as detailed schematically in Figure 1.
Samples were sectioned and then examined/measured using a micr
oscopy equipped with a4i image analysis
software. The head
-
height is the distance between the rivet head and the top riveted sheet, and the interlock is
the distance of the rivet shank flaring into the locked sheet/s. When the head is above the level of th
e top sheet,
a positive value for the head
-
height is given. A negative head
-
height implies that the head is below the level of
the top sheet. Both the head
-
height and the interlock distance are considered to be critical for a SPR joint.
Ideally the rivet h
ead should be flushed with the top sheet. In practice, if the rivet head sticks out too high above
the top sheet, moisture may ingress into the joint leading to corrosion concerns. Alternatively, it may be slightly
below the top sheet but may leave a visib
le dent on the surface leading to visual quality concerns. Therefore, top
and lower limits for head height were defined as +0.3mm to
-
0.5mm for quality assessment. The interlock is a
key feature of a SPR joint holding the riveted sheets together. Sufficien
t interlock is necessary and therefore a
minimum interlock value is required for quality assessment. Considering the different properties between
aluminium and High Strength Steel, different minimum interlock values have been specified; 0.2mm for HSS
and 0
.4mm for aluminium. In addition, fracture of the locked sheet (bottom sheet) must be avoided as it can
lead to severe corrosion [
6
]. A minimum remaining material thickness of 0.2mm is therefore introduced into the
quality assessment criteria.
Page
3
of
13
Figure
1
: quantitative measurements for a SPR joint
2.2 Materials and stacks
The materials used throughout
this project
were Aluminum AA5754 and HSLA (High Strength Low Alloy) 350
with various thicknesses. The AA5754 alloy sheets were pre
-
t
reated with a chromate free film
, and
coated with
a wax
-
based surface lubricant. The HSLA350 had Zinc plate and passive surface coating.
Table 1 and 2 present
the compositions and mechanical properties of AA5754 and HSLA350 respectively. These two material
s
formed various stacks that are presented in the format of (top/pierced sheet + middle sheet + bottom/locked
sheet). For example, (2.0mm AA5754 + 1.0mm HSLA350) nomenclature indicates that the top sheet is 2.0mm
AA5754 and the bottom sheet is 1.0mm HSLA35
0 with no middle sheet
.
Table
1
Compositions and Mechanical Properties of
aluminium
AA5754
MECHANICAL PROPERTIE
S
Young’s Modulus
(GPa)
Tensile strength (MPa)
Elongation
Hardness (H
V
)
70
250
25%
68
NOMINAL COMPOSITIONS
Si%
Fe%
Cu%
Mn%
Mg%
Al%
0
-
0.40
0
-
0.40
0
-
0.10
0
-
0.50
2.60
-
3.60
Balance
Table
2
Compositions and Mechanical Properties
of HSLA350 (EN 10268
-
H360LA
)
MECHANICAL PROPERTIE
S
Yield strength (MPa)
Max
.
Yield strength (MPa)
Elongation
390
480
26%
Chemical composition
C%Max
Si%Max
Mn%Max
P%M
ax
S%Max
Al%Min
Nb%Max
Ti%Max
0.10
0.50
1.20
0.025
0.025
0.015
0.090
0.15
Page
4
of
13
2.3 Selection of rivet and die
The selection of rivet and die depends on the sheet material to be joined and is a key factor for achieving good
quality joints. As the purpose of th
is excise is to understand the effect of setting velocity on various aspects of
the SPR joint, the selection process of rivet and die is not discussed in detail. Although the rivets and dies are
designated only by codes, the key parameters relating to the
discussion are given. All the rivets and dies were
supplied by the Henrob Ltd. and Henrob servo SPR gun was also used throughout the project. The velocity
presented through the paper is in nominal units as displayed on the Henrob controller. The velocity i
n mm/min
is proportional to the unit.
2.4
Die load profile
As the setting velocity is a component of setting force, it is understandable that the higher the setting velocity
the higher the setting force. But the setting force is different from the force actin
g on the die. In order to know
the force acting on the die, a KISTLER force transducer with a 120kN capacity was installed directly
underneath the die. A Pico data acquisition system was connected to the transducer to record the load profile
acting on the
die. It should be mentioned here that the die load profile is different from the force
-
displacement
curve, which shows the force against the distance travelled as discussed by previous research [Kim et al, King,].
2.5
Sample preparation and examination
All sa
mples were prepared by using special fixtures designed to maintain consistent dimensions for each
individual purpose. The sample dimensions for metallographic inspection and photoelastic measurement are
showed in Fig
.
2.
For metallographic inspection, a sp
ecially designed fixture was also used to ensure samples were sectioned at
the centre of the joints consistently. A Buehler Delta AbrasiMet cutting machine was used with very low
cutting speed in order to make the section surface smooth and even. The sect
ioned samples were then examined
microscopically and measured using a4i Analysis software.
The residual stress distribution around the rivet head and tail is not fully understood due to the complexity of
the process. It is assumed that the residual stress
generated during SPR process will affect the behaviour of such
joints, similar to the effect of residual stresses on the properties of a formed component. Therefore, this project
has for the first time obtained the residual stress distribution using photo
-
elastic techniques.
The photoelastic
technique requires the test sample to have
a special coating material, PS
-
1D, bonded to the surface. This is an
excellent high
-
sensitivity plastic, manufactured by the Photolastic Division of Vishay Measurements Group,
Inc, for accurate analysis in the elastic and elastic
-
plastic ranges of most metals [9]. The PS
-
1D sheet was cut to
40x40mm with a clearance hole punched in the potential rivet area. Then it was adhered to degreased aluminum
specimen at the overlap area a
nd left in a dry clean environment for 24 hours.
Page
5
of
13
Figure
2
: Sample dimensions for sectioning and Photoelastic Analysis
The dimensions for lap shear samples are shown in Fig. 3. At least five samples w
ere made and tested for each
setting velocity selected. A standard Instron tensile test machine with 30kN load capacity and a cross head
speed 10mm/min was used for all lap shear test.
Figure
3
: Sample dimensions for shear test
2.
RESULTS AND DISCUSSI
ON
3.1 Effect of setting velocity on joint quality
Fig
.
4
shows cross sections of the
joints in a (2.0mm AA5754 +1.0mm HSLA350)
stack,
made using C50d42
rivets and a
DZ0902050
die
at different velocities.
As the velocity was increa
sed incrementally
from 120 to 180
units, the characteristics of the sections
changed
in terms of
interlock, remaining material thickness, head
-
height
and gap existence. Comparing the two extremes of 120 and 180; at the minimum velocity 120, there is a visi
ble
gap at the faying surfaces between the two riveted sheets and an interlock of only 0.25mm. At the maximum
setting velocity of 180 the gap has been eliminated and the interlock increased to 0.33mm. The headheight is
40mm
40mm
15mm
20mm
R=12mm
100mm
PS
-
1D
Page
6
of
13
0.52mm at 120 velocity with a remaini
ng material thickness of 0.25mm. At the higher 180 velocity the head
height was reduced to 0.03mm with an associated reduction in remaining material thickness to 0.21mm.
For the sections (a) and (b), the head heights were both greater than 0.3mm and there
fore the joints failed the
quality criteria. The relationship between the velocity and head
-
height, interlock and remaining thickness is
shown in Fig. 5. In summary; t
he increase in the setting velocity led to a decrease in the head height and the
remainin
g thickness; but an increase in the interlock. The head
-
height had the highest variation rate.
(a) 120 (b) 140
(c) 160
(d) 180
Figure
4
:
Cross sections of the joints C50d42DZ0902050 (2+1) at different velocities
As the setting velocity reflects the force applied to the rivets, it follows that at high velocity the rivet
has more
dynamic energy available to pierce and deform the top sheet as well as to flare the rivet legs into the bottom
sheet. Again comparing the extreme velocities of 120 and 180 shown in Figures 4(a) and (d); at 120 velocity the
2.0 mm top sheet was red
uced to 1.58mm when measuring from underneath the rivet head to the bottom sheet,
whilst at 180 velocity this distance was 1.28mm. A further effect of high velocity is that the rivet would have
more energy to deform the bottom sheet by flaring into it. The
se factors contributed to the reduction in the head
-
height and remaining thickness, as well as the increase in the interlock that were obtained with the higher 180
velocity. It is worth mentioning that for this material combination, the bottom sheet is 1.0
mm HSLA350, it is
not easy to deform further following initial deformation which would inevitably lead to hardening of the
HSLA350. This may explain why the reduction in remaining material thickness is not as significant as in the
head
-
height, and the incr
ease in the interlock is minimal as the setting velocity increased.
Page
7
of
13
Figure
5
:
Effect of setting velocity on the joint head height
When the rivet and die combination was changed, the effect of setting velocity was exaggerated. Fi
g. 6 shows
two cross sections of the joints that have same
(2.0mm AA5754 +1.0mm HSLA350)
stack and C50d42 rivet as
shown in Fig. 5, but used a different die profile from the sections shown in Fig. 4. At 120 velocity, the joint
achieved reasonable head
-
heig
ht and interlock with no breakthrough. However, a relatively small change in
velocity from 120 to 140 produced a joint with too low head
-
height and with part of the bottom material
missing leading to a breakthrough failure. This suggested that even with t
he same rivet and die combination,
proper selection of setting velocity is required in order to achieve acceptable joints.
(a) 120 (b) 140
Figure
6
:
Cross sections
of the joints C50d42DZ11000 (2+1) at different velocities
3.2
Effect of setting velocity on die force profile
It is important to know the force acting on the die as die failure is a major concern in production. Figure 7
shows the die load profiles generated f
rom riveting a (2.0mmNG5754+2.0mmNG5754) stack with a
combination of rivet and die C50D42/DZ0902050 at various setting velocities. The peak load varied from about
40kN to over 80kN as the velocity changed from 80 to 180, confirming that increasing the sett
ing velocity leads
to an increase in the peak load value. In addition, the pattern of the load profile also changed when the setting
velocity changed. As the setting velocity reflects the setting force, it is understandable that the higher the setting
velo
city the higher the impact force on the die. The decaying wave pattern may be caused by matching the
Page
8
of
13
resonance of the particular C
-
frame. In summary for a given stack, the higher the setting force the higher the
force action on the die.
Figure
7
: Effect of setting velocity on die load profile
–
(2.0+2.0mm) AA5754
Figure
8
: Die load profile for different stacks withC50D42DZ09050 at velocity 140
Figure 8 shows the load profiles obtained from two stacks w
ith the only difference being that one has a bottom
layer of HSLA350 instead of AA5754 for the other. Although identical combination of the rivet and die as well
as setting velocity was applied to make the joints, the resulting peak force and force pattern
acting on the die are
different. As the setting velocity is the same for both stacks, the force acting on the rivets should be the same.
However, as the HSLA350 high strength steel requires more energy to deform than the lower strength AA5754
during the r
iveting process, the peak load on the die was reduced and a different load pattern was observed. It
follows that the same velocity applied to different material combinations leads to different impact force on the
die.
Page
9
of
13
3.3
Effect of setting velocity on joint l
ap shear behaviour
Figure 9 shows lap shear test results for the joints corresponding to the micro sections shown in Fig. 4. At
velocity 120, the average peak load was 4.01kN; whilst at velocity 180, the shear strength was 5.21kN. As the
setting velocity i
ncreased, the shear strength of the joints also increased. For this group of joints, pull
-
out of the
rivet from the bottom sheet, as shown in Fig. 10(a), was the only failure mode.
The shear strength, for a SPR
joint, depends on a combination of the sheet
material tensile strength, the bearing resistance and the interlock
strength of the joint [li].
The
pull
-
out
failure mode indicated that the interlock and bearing resistance of the
bottom sheet joints governed the joints shear strength. The increase in the
setting velocity led to an increase in
the interlock, and consequently to an increase in the shear strength of the joints.
F
rom
velocity
120 to 140, the
amount of free
-
play in the joint was the greatest and thereafter reduced.
The free
-
play is attributed
to the
existence of the gap between the two riveted sheets, which causes loosening of the joints and over protruded
rivet head at velocity 120. Both the gap and over protruded head facilitated pull
-
out of the rivet during lap shear
test leading to much low
er strength of the joints.
Figure
9
:
E
ffect
of setting v
elocity on
shear
strength of (2+1) joint
s
(a)
Pull
-
out failure
(b) Tearing of the bottom sheet
Figure
10
: Failure modes occurr
ed during lap shear test
Figure 11 shows peel and lap shear test results for the joints shown in Fig. 6. The joints made at velocity 120
had over 40% lower peel strength and slightly lower shear strength than the joints made at velocity 140.
T
he
load versu
s velocity trend shows the higher velocity the higher strength, but this increase in strength is actually
attributed to different failure mechanisms. For lap shear test,
the joints made at 120 velocity failed by pulling
Page
10
of
13
the rivet out of the bottom sheet,
a
s shown in Fig. 10a, indicating that the interlock and the bearing resistance of
the bottom sheet contributed to the shear strength of the joints; whilst the joint made at 140 velocity failed by
tearing of the bottom sheet (Fig. 10b) suggesting that the te
aring strength of the bottom sheet was sole
contributor of the shear strength of the joints. Similar to shear test geometry, the peel strength for a SPR joints
relies on the interlock. The failure modes that occurred during peel testing are presented in Fi
g.12. Although all
joints failed by pull
-
out of the rivet from the bottom sheets, the joints made at 140 velocity left an empty hole in
the locked sheet (Fig. 12a) indicating the full contact between the rivet and the locked sheet or a possible buckle
at t
he locked sheet, whilst the button in the locked sheet for the joints made at velocity 120 remained intact
(Fig.12b). This explained the higher peel strength of the joints made at velocity 140.
The small change in
velocity which altered the quality of the
joints led to a significant change in failure mechanism of the joints.
0
1000
2000
3000
4000
5000
6000
Mean Load (N)
120 velocity
140 velocity
Peel
Shear
Figure
11
: Peel and shear test results corresponding to the joints shown in Fig. 6
(a)
Pull
-
out failure
(b) Button intact
Figu
re
12
: Failure modes occurred during peel test
3.4
Effect of setting velocity on residual stress distribution
Fig
ure 13
shows residual stress distribution images obtained from the joints shown in Fig. 4 using photo
-
elastic
measurement
with Deltavision software. It can be seen that the fringe pattern differed as setting velocity varies
indicating the variation of residual stress distribution on the riveted sheets. Although multi
-
fringes occurred due
Page
1
1
of
13
to the effect of coating, the results
still showed that the increase in setting velocity led to an increase in
deformation of the transparency coatings, which reflected the deformation of the riveted sheet.
Figure
13
:
Residual stress distributio
n
images of the joints made at different velocities
Further analysis, using a line profile function in the Deltavision software, showed the maximum residual
stresses in a line extending from the edge of the rivet into the sheet at different setting velocit
ies, as shown in
Figure 14
. Although, the curves fluctuate, due to multi
-
fringe effects, the results are still predictable. For the
joints made at velocity 120, the maximum residual stress was about 12 camera units, which are proportional to
the stress val
ue, and occurred at about 3.2mm from the edge of the rivet head. In comparison, the maximum
residual stress for the joint made at velocity 180 was over 15 units and occurred at about 5.0mm from the edge
of the rivet head. In summary; the residual stress v
alue and the area of distribution increased as the velocity
increased.
120
1
4
0
1
6
0
1
8
0
Page
12
of
13
(a)
Velocity
120
(b)
Velocity
180
Figure
14
:
Stress distributions of the joints made at different velocities
4
C
ONCLUSIONS
As a key process parameter, setting vel
ocity reflects the riveting force and therefore controls the level of
deformation that occurs for both rivet and sheets during the riveting process. The experimental work reported
above showed noticeable effects of setting velocity on various aspects. Thes
e effects can be summarised as
below:
Setting velocity affects SPR joint quality. As the setting velocity increases, the head height and remaining
material thickness decreases; whilst the interlock increases. For s given stack, there is a range of minimum
and maximum velocity. Any velocity outside this range will lead to the head height, remaining material
thickness and interlock outside the quality criteria, causing failure of the joints.
Setting velocity influences SPR joint strength but this can vary de
pending on joint quality and consequently
failure mode. For rivet pull
-
out failure, the joint strength increases as the setting velocity increases. For
mixed failure modes of rivet pull
-
out and sheet failure, the joint strength may be determined by the she
et
material strength and therefore does not necessarily follow a consistent trend with the setting velocity.
Page
13
of
13
Setting velocity affects die force profile. For a given stack the higher the setting velocity, the higher the
force acting on the die. However, t
his effect can vary depending on the material properties and combinations
to be joined.
Setting velocity affects residual stress distribution. As the setting velocity increases, the peak value of
residual stress and the area of residual stress distribute
d increases.
Acknowledgments
We would like to thank Advantage West Midlands, Jaguar and Land Rover, Novelis for their support of the
advanced body joining research program.
References
1.
Polmear
.
I
.
J, “Light alloys”, Third edition, Edward Arnold, 1995.
2.
Lei
termann
,
W and Christlein
,
J
, “The 2
nd
generation Audi space frame of the A2: A trendsetting all
-
aluminium car body concept in a compact class car”, Seoul 2000 FISITA World automotive congress,
June 12
-
15, 2000, Seoul, Korea.
3.
Komatsu
,
Y
, Ban
,
K
, Ito
,
T,
Mu
raok
,
Y
, Yahabo
,
T
, Yasunaga
,
K
and Shlokawa
,
M
, “Application of all
aluminium automotive body for Honda NSX”, SAE 910548 (1991).
4.
Kim
,
Dae
-
Wook,
Xu
,
J,
Li
,
W, and
Blake
,
D,
“
Force characteristics of self
-
piercing riveting
”,
JEM525
,
IMechE 2006
,
Proc. IMech
E Vol. 220 Part B: J. Engineering Manufacture
.
5.
King
.
R. P
, “
Analysis and quality monitoring self
-
pierce riveting process
”, PhD dissertati
on, University
of Hertfordshire, 1997
.
6.
Han
,
L
,
Young
,
K
,
Hewitt
,
R and
Chrysanthou
,
A
, “Effect of breakthrough on the
behaviour of self
-
piercing riveted aluminium 5754
–
HSLA joints”, Transactions of Society of Automotive Engineers of
Japan, Vol.36 No.5 pp181
-
186, 2005.
