Anda di halaman 1dari 12

Materials and Design 62 (2014) 205216

Contents lists available at ScienceDirect

Materials and Design


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

Microstructural and mechanical characterisation of laser-welded lap


joints with linear and circular beads in thin low carbon steel sheets
D. Benasciutti a,, A. Lanzutti b, G. Rupil c, E. Fraenkel Haeberle c
a

Dipartimento di Ingegneria Elettrica Gestionale Meccanica (DIEGM), Universit di Udine, via delle Scienze 208, 33100 Udine, Italy
Dipartimento di Chimica Fisica e Ambiente, Universit di Udine, via del Cotonicio 108, 33100 Udine, Italy
c
F.I.S.A. Fabbrica Italiana Sedili Autoferroviari S.r.l., via G. De Simon, 33010 Osoppo, Udine, Italy
b

a r t i c l e

i n f o

Article history:
Received 24 December 2013
Accepted 5 May 2014
Available online 13 May 2014
Keywords:
Laser-welded lap joint
Low carbon steel
Microstructure
Micro-hardness
Static strength

a b s t r a c t
This paper presents a microstructural and mechanical characterisation of laser-welded lap joints in low
carbon steel thin sheets. Different combinations of steel types (DC05, S355MC) and thickness values are
used to assemble welded specimens with linear and circular weld bead. Metallurgical observations and
micro-hardness tests are used to characterise the weld microstructure. Mechanical response in tensile
test is then used to evaluate the static strength, rotation angle of weld bead and failure mode of welded
specimens. Lap-joints with circular weld showed a lower rotation angle compared to linear welds. The
fracture in all tested specimens occurred at the base metal, far away from the weld. A simplied mechanical model is nally proposed to derive theoretical formulae for estimating the tensile strength of welded
joints as a function of material properties and weld geometry. The analytical results are in good agreement with experimental ndings and they estimate an increased strength for circular welds, compared
to linear weld with same lateral width. A design chart is also derived to allow a design of laser-welded
joints with virtually equal strength of base metal and weld zone.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
Laser welding is becoming an attractive and economically
advantageous joining technique in several engineering elds. The
main advantages are low process cost, high welding speed and
concentrated heat power, which gives reduced distortions and
narrow weld bead with limited microstructure changes in the
heat-affected zone. Laser welding has largely been applied, for
example, in the automotive industry to manufacture parts of car
such as doors, front and side panels, side beams and wheel arches
[1]. Applications of laser welding are partly documented also in
railway industry, as possible replacement of resistance spot welding to increase surface quality of welded assembly on rail vehicle
side panels [2].
A lap joint is a geometry commonly adopted in various welded
assemblies and its characteristics have been investigated in the
literature. Experimental studies of laser-welded lap joints, with
various combinations of metals and alloys, have been focused on
microstructural and metallurgical characteristics [35], as well as
on mechanical strength under static [613] and fatigue loadings
[5,14,15]. For example, Sokolov et al. [3] presented an experimental
Corresponding author. Tel.: +39 (0) 432 558048; fax: +39 (0) 432 558251.
E-mail address: denis.benasciutti@uniud.it (D. Benasciutti).
http://dx.doi.org/10.1016/j.matdes.2014.05.005
0261-3069/ 2014 Elsevier Ltd. All rights reserved.

study on laser welds in S355 steel with high thickness values


(20 and 25 mm). Microstructure and hardness proles were used
to characterise welded joints and to suggest preferable welding
parameters. Hardness trends were use to identify microstructural
changes in the joint. They found an increase of hardness in welded
region, compared to base metal. An increase in welding speed also
induces an overall increase in hardness and a reduction in the
width of heat-affected zone. Yilbas et al. [4] investigated laser
welds of low carbon steel plates with numerical simulations and
experimental tests. Finite element analyses were used to compute
the temperature distribution and the residual stresses in weld
zones, which were in good agreement with measurements. The
metallurgical and morphological modications in the weld were
also examined. Farabi et al. [5] characterised the microstructure
and mechanical properties of laser joints in dual phase steel
DP600. They found a considerable increase of hardness in the
fusion zone, due to the large amount of martensitic structure promoted by rapid cooling during welding. On the other hand, a softer
zone was observed at the outer heat-affected zone (HAZ), due to
tempering of the pre-existing martensite. This soft zone was the
position where all specimens fractured in tensile tests. In fatigue
tests, welded joints showed a slightly lower fatigue limit compared
to base metal. At high stress amplitudes, however, they showed
comparable fatigue strength within the experimental scatter.

206

D. Benasciutti et al. / Materials and Design 62 (2014) 205216

The strength of welded joints has been investigated by theoretical and numerical approaches, as well as by experimental studies.
For example, Ha and Huh [6] carried out experimental tests to
formulate an analytical failure criterion for laser weld under combined normal and shear loading. Their study pointed out also the
change in failure mechanism (from base metal to interfacial failure), as a function of the loading angle. Ono et al. [7] presented
an experimental characterisation of static and fatigue strength of
laser-welded lap joints in thin steel sheet. They also proposed a
simple analytical model to estimate the strength and fracture position, based on weld joint properties. Miyazaki and Furusako [8,9]
developed a similar model to estimate static failure evaluation,
which also included the failure in the portion adjacent to the weld.
Numerical simulations with nite elements were also
attempted to evaluate the weld strength under tensile loading.
For example, Terasaki and Kitamura [10] adopted an elasticplastic
nite element model to check whether the equivalent plastic strain
can be used to estimate the static tensile strength of lap joints.
Comparison with experimental results showed that the equivalent
plastic strain is only suitable to estimate base metal failure, while
weld shear failure has to be assessed by conventional failure theories that assume a constant shear stress. Pan and co-workers
[11,12,15] proposed a non-homogeneous elasticplastic model to
simulate the mechanical response and failure mode of lap joints
in high strength low alloy (HSLA) steel. Results from nite
elements analysis were in good agreement with experimental
observations. The main limitation of nite elements modelling,
however, is that numerical simulation can continue until numerical instability occurs. Therefore, the simulated loaddisplacement
curves are monotonically increasing, without any maximum load
that can be used to dene the joint strength.
In the context of previous literature studies, this work presents
the preliminary results of a research project aimed to characterise
the microstructure and the mechanical strength of laser-welded
lap joints in steel sheets, to assess their possible application in railway industry. The interest is focused on laser welds made of thin
low carbon steel sheets with different thickness and chemical composition. A deep drawing DC05 steel and S355MC high strength
structural steel, with thickness values in the range of 1.0
1.5 mm, are considered. Two different weld geometries (linear
and circular) are also compared, see Fig. 1. The linear weld has a
straight weld bead, perpendicular to specimen longitudinal axis.
A circular weld, instead, has a weld bead that forms a circumference located at the specimen centre.
Metallurgical analyses, micro-hardness measurements and
mechanical tensile tests were carried out to characterise the microstructure and mechanical properties of laser-welded lap joints
with different weld geometry. The proposed experimental characterisation suggests that laser welding of dissimilar low carbon thin
steel sheets could be a suitable joining technique for structural
applications in railway industry.
2. Materials and experimental procedure
Microstructure and mechanical properties of laser-welded lap
joints are investigated. Microstructure is characterised by metallurgical analysis and micro-hardness measurements. Mechanical
behaviour is studied by tension tests, which provide the tensile
strength, longitudinal deformation and rotation angle of the weld
bead, as well as typical failure mechanisms.
2.1. Base materials and laser-welded lap joints
Two types of steel (DC05, S355MC) are used in welded
joints. The DC05 material (EN 10130:2006 [16]) is a cold rolled,

(a)

linear weld

(b)

circular weld

Fig. 1. Geometry of laser-welded lap joint with (a) linear and (b) circular weld
geometry.

non-ageing low carbon steel especially suited for deep drawing and
other demanding forming applications. The S355MC steel (material
n. 1.0976) (EN 10149-2:2013 [17]) is a thermo-mechanically rolled
steel with high yield stress and high impact strength properties.
Table 1 lists the nominal mechanical properties and chemical
compositions of DC05 and S355 steels used in this study.
The laser-welded lap joints were obtained by welding two overlapping thin sheets with a bre laser. The welding parameters are
summarised in Table 2. No post-weld heat treatment was applied
after welding.
Several preliminary tests were performed to arrive at the optimised welding parameters given in Table 2. In fact, previous
parameters [18] gave unsatisfactory welded specimens, characterised by incomplete penetration, inhomogeneous microstructure
and insufcient mechanical strength compared to base metal
(e.g. failure occurred at weld bead). On the other hand, it is not
the aim of the present work to further investigate the correlation
between welding parameters and weld properties.
Fig. 1 shows the geometry and Fig. 2 a top-view of the welded
specimens considered in this study. Different combinations of
sheet thickness and metal types were used (see Table 3): a thickness of 1.0 and 1.2 mm for DC05 steel, 1.5 mm thickness for
S355MC steel.
Specimens were shaped by laser cutting after welding. The
overlapped welded sheets prior to laser cutting had a rectangular
geometry, with same width and length as nal welded specimens:
length 231 mm, width 30 mm (linear weld) and 35 mm (circular
weld). Two different weld geometries (linear and circular) were
considered, both positioned at the overlap centre. The central
straight portion of the specimen with reduced cross section has a
length of 100 mm, while the width is 15 mm (for linear weld)
and 20 mm (for circular weld). The linear weld has a length equal
to the sheet width and is positioned transversely to the longitudinal specimen axis. Instead, the diameter of circular weld (15 mm)
is lower than sheet width, which assures that the weld is
completely inside the metal sheet and it is not cut during specimen
shaping by laser cutting. The overlap length is 100 mm. Two doublers, 40 mm long, were positioned at both ends of the welded

207

D. Benasciutti et al. / Materials and Design 62 (2014) 205216


Table 1
Chemical composition (wt%) and mechanical properties of DC05 and S355MC steels used in laser-welded lap joints.
Material

Yield strength, ry (MPa)

Ultimate tensile strength, ru (MPa)

Elongation, A (%)

Mn

Si

Al

Ti, V

DC05
S355 MC

163
369

283
457

42
30

0.056
0.096

0.057
0.326

0.269

0.040

0.007
0.006

0.011
0.005

0.022

0.060

Table 2
Welding parameters used in the present study.
Laser power (kW)
Welding speed (m/mm)
Feeding bre diameter (lm)
Collimation (mm)
Focal length (mm)
Focal point position (mm)
Shielding gas

2
2
100
110
200
190
Nitrogen

water. In linear welds, the cutting direction was perpendicular to


the weld. In circular weld, the cutting direction was aligned to
the diameter of the weld line, along the specimen longitudinal axis.
The metallurgical samples cut from welded joints were embedded into a hard epoxy resin and then grounded and polished, using
a diamond polishing suspension, in order to obtain a mirror like
surface. The samples were then etched with Nital 2 etchant for
30 s in order to highlight the microstructure of material.
2.3. Micro-hardness of laser-welded joints
Micro-hardness tests were carried out by using a Vickers indenter, with an applied load of 200 g for 15 s. The indentations were
positioned at regular intervals of 200 lm in the transverse direction across the weld zone, from the fusion zone up to the base
metal. Micro-hardness values were measured at the centre of both
the upper and lower sheets of laser-welded lap joints.
2.4. Tensile tests

Fig. 2. Top views of laser-welded lap joint with (a) linear and (b) circular weld
geometry.

Table 3
Combination of sheet thickness and steel type in laser-welded lap joints analysed in
this study.
Specimen code

Thickness
(steel type)

Thickness
(steel type)

Weld geometry

1.5_1.0_L
1.2_1.0_L
1.5_1.2_C
1.5_1.0_C

1.5 mm
1.2 mm
1.5 mm
1.5 mm

1.0 mm
1.0 mm
1.2 mm
1.0 mm

Linear
Linear
Circular
Circular

(S355MC)
(DC05)
(S355MC)
(S355MC)

(DC05)
(DC05)
(DC05)
(DC05)

Tensile tests of base metal sheets were performed according to


the standard EN ISO 6892-1 [19] for tension testing of metallic
materials. Steel sheets were shaped by laser cutting into dog-bone
specimens, with a width of 15 mm and a gauge length1 of 100 mm,
see Fig. 3.
The specimens of base metal sheets were intentionally designed
to have identical dimensions of lap joints with linear weld, see
Fig. 1(a), to allow a comparison of loaddisplacement curves. As
for laser-welded specimens, two thickness values were considered
depending on steel type: 1.2 mm for DC05 steel, 1.5 mm for S355
steel. Tensile tests of base metal were carried out using a
servo-hydraulic MTS testing machine, with a 10 kN load cell. The
cross-head speed was 0.1 mm/min, in the elastic range, and
1 mm/min in the plastic range of the tensile curve. In the elastic
range the test were performed in strain control mode, using an
MTS extensometer (model MTS 636.11F-24), while in the plastic
range in displacement control (LVDT sensor of the instrument).
Tensile tests of laser-welded lap joints were carried out on a
MTS810 machine, under displacement control mode, with a speed
of 1 mm/min. Three tensile tests were replicated for each welded
sample conguration listed in Table 3.
3. Results and discussion
3.1. Metallurgical observations

specimen, to align the xtures and thus to eliminate the initial


bending in tensile loading, due to the non-aligned grips. Lap joints
with comparable dimensions were investigated in similar studies
in the literature [7,8,12].

2.2. Weld microstructure


The weld microstructure was characterised by metallurgical
analyses of samples cut from welded specimens. To this end,
laser-welded lap joints were sectioned close to the weld zone by
an abrasive cut-off machine. The region of the cut was chilled by

Figs. 46 show examples of optical micrographs of the etched


cross section for welded joints with both linear and circular weld
geometry.
Different zones with different microstructures characterise the
welded specimen: base metal (BM), heat-affected zone (HAZ) and
fusion zone (FZ). All examined specimens showed a fully
penetrated weld bead, with an average width of about 1 mm (i.e.
1
According to the standard EN ISO 6892-1 Metallic materials tensile testing
[19], the gauge length is the length of the portion of the test specimen on which
elongation is measured at any moment during the test.

208

D. Benasciutti et al. / Materials and Design 62 (2014) 205216

Fig. 3. Specimen geometry used in tensile test of base metals. Thickness is 1.2 mm
for DC05 steel and 1.5 mm for S355MC steel.

comparable to the thickness of joined sheets). The HAZ has an


average width ranging from 300 to 400 lm. In our specimens, a
larger width of FZ is generally observed in the sheet closer to the
laser source (top part in the gures), due to a focalisation phenomenon of laser welding. Instead, the HAZ has almost the same width
in both upper and lower sheet.
For laser joints with linear welds, the FZ of 1.5 mm sheets
(S355MC steel) has a mixed-microstructure formed by low-carbon
bainitic structures, surrounded by pro-eutectoid a-ferrite. Instead,
the FZ of joints in DC05 steel is predominantly formed by ferritic
structures, with some grains of low-carbon bainite. In all welded
joints examined, the HAZ is characterised by large a-ferrite
columnar-shaped grains. For 1.5 mm sheet, very small low-carbon
bainitic zones are occasionally observed.
Due to their particular geometry, the metallurgical samples cut
from laser joints with circular welds have two distinct welded
regions, see Fig. 6(a). The microstructure of HAZ is very similar to
that described for linear welds. Instead, the FZ has a bainitic
structure that is softer compared to that of linear weld (see
micro-hardness values commented later on). This particular
behaviour could be correlated to an excess of ferrite, if compared
to the microstructure observed for linear welds. The softer
microstructure of circular welds could be explained by a slower
cooling rate in weld material, promoted by the particular circular
geometry of the weld. In fact, during welding the heat remains
entrapped inside the circular region and it then gives higher temperatures for longer time, with a reduced cooling rate on the weld
material.
3.2. Micro-hardness tests
The micro-hardness proles are plotted in Fig. 7 (for linear
weld) and Fig. 8 (for circular weld), with values for both the upper
and lower sheet in welded specimen.

Hardness values are shown to generally depend on steel type, as


well as on the position in the weld region as a result of different
microstructures. The hardness of BM ranged in values 7480 HV
for DC05 steel and 112125 HV for S355MC steel, respectively.
The hardness of BM is generally independent of sheet thickness,
although for DC05 steel the 1.2 mm sheet seems to have slightly
higher values compared to 1.0 mm sheet, see top/bottom values
for 1.2_1.0_L welded joint in Fig. 7.
In all joint types, an increase of micro-hardness is observed in
FZ, due to harder microstructures that are promoted by melting
and subsequent rapid cooling during welding, and which have
already been observed in previous metallurgical examinations. In
some analysed specimens, the micro-hardness appears to decay
towards the centre of FZ, while it increases at the edges. Examples
are specimens 1.5_1.0_L (top sheet), 1.5_1.2_C (top sheet) and
1.5_1.0_C (left weld in top sheet). This hardness drop indicates a
softening behaviour, which has also been observed in other studies
[24,20] where it was attributed to a self-annealing effect occurring at the centre of weld zone due to a reduced cooling compared
to base metal, while higher cooling rates characterise the edges of
weld zone. This remark could explain why this hardness decay has
mostly been observed in the top sheet that is closer to laser source,
where higher temperatures are observed [4,21].
The width of hardness prole in FZ is generally higher in the top
sheet, compared to bottom one, due to a larger FZ caused by the
focalisation of laser beam, as mentioned above. The difference is
more evident in circular welds (see Fig. 8), although a slight difference is also observed in linear weld (Fig. 7). This particular trend
can be correlated to the different width of weld zone, which is
inuenced by the position and thickness of welded sheets. In particular, a larger FZ characterises the top sheet directly exposed to
laser source (see also micrographs in Figs. 46). Furthermore, a larger thickness for top sheets also promotes a faster cooling during
welding and thus harder microstructures. On the other hand, an
inuence of steel type (S355MC always in top sheet) cannot be
excluded at all.
A rapid decrease of hardness occurs within the HAZ around the
FZ. This can clearly be observed for both linear and circular welds.
However, in circular weld joints the metal enclosed by the circular
weld line tends to have a slightly lower micro-hardness compared
to base metal, see Fig. 8(a) and (b). This decrease, although not very
pronounced, can easily be detected, especially for the upper sheet
in S355MC steel that is directly exposed to the laser source. This
lower hardness values, observed in the HAZ enclosed by the weld
circle, characterise the softer microstructures that are induced
by the lower cooling rate promoted by the circular weld geometry,
as discussed in the previous section.

Fig. 4. (a) Optical micrograph of the etched cross section of a 1.5 mm/1.0 mm laser joint with linear weld (top sheet: 1.5 mm, S355MC steel; bottom sheet: 1.0 mm, DC05
steel) and (b) microstructure of FZ, HAZ and BM.

D. Benasciutti et al. / Materials and Design 62 (2014) 205216

209

Fig. 5. (a) Optical micrograph of the etched cross section of a 1.2 mm/1.0 mm laser joint with linear weld (top sheet: 1.2 mm, DC05 steel; bottom sheet: 1.0 mm, DC05 steel)
and (b) microstructure of FZ, HAZ and BM.

Fig. 6. (a) Optical micrograph of the etched cross section of a 1.5 mm/1.0 mm laser joint with circular weld (top sheet: 1.5 mm, S355MC steel; bottom sheet: 1.0 mm, DC05
steel) and (b) microstructure of FZ, HAZ and BM.

3.3. Tensile tests: strength, deformation and failure mode


Typical stressstrain curves for base metal steels are shown in
Fig. 9. Both steels exhibit high-ductility behaviour, with DC05 steel

having the largest elongation at failure (experimental values of 45%


for DC05 and 24% for S355MC steel, respectively). Conversely, the
S355MC steel shows, as expected, the highest strength compared
to DC05 material.

210

D. Benasciutti et al. / Materials and Design 62 (2014) 205216

450

250
1.2_1.0_L (top)

400

1.2_1.0_L (bot)
1.5_1.0_L (top)

200

bottom (DC05)

1.5_1.0_L (bot)

175
150
125
100

S355MC steel

350

stress (MPa)

Micro-hardness, HV200

225

top (S355MC)

300
250
200

DC05 steel

150
100

top (DC05)

50
75
0
0.00

bottom (DC05)

50
-2

-1.5

-1

-0.5

0.5

1.5

0.10

0.20

distance from the weld center (mm)

0.30

0.40

0.50

strain (-)

Fig. 9. Engineering stressstrain curves for DC05 and S355MC base metals.

Fig. 7. Micro-hardness values for laser-welded specimens with linear weld


geometry.

7
1.2 mm (DC05)/1.0 mm (DC05) - linear weld

200

6
5

1.5_1.2_C (bot)

force (kN)

Micro-hardness, HV200

1.5_1.2_C (top)

175
150

top (S355MC)

125

4
B

75

0
bottom (DC05)

50
-10

-5

DC05 (t=1.2 mm)

sample A
sample B
sample C
DC05 (t=1.2 mm)
DC05 (t=1.0 mm, L=50 mm)

2
100

10

(a)
5

20

(a)
30

40

50

displacement (mm)

10

distance from the center of circular weld (mm)

1.5 mm (S355)/1.0 mm (DC05) - linear weld

6
200

175

force (kN)

Micro-hardness, HV200

1.5_1.0_C (top)
1.5_1.0_C (bot)

150
top (S355MC)

4
B

DC05 (t=1.2 mm)

sample A
sample B
sample C
DC05 (t=1.2 mm)
DC05 (t=1.0 mm, L=50 mm)

125
1
100

bottom (DC05)

75

10

20

(b)
30

40

50

displacement (mm)
(b)

50
-10

-5

10

distance from the center of circular weld (mm)


Fig. 8. Micro-hardness values for laser-welded specimens with circular weld
geometry.

The loaddisplacement curves of laser-welded lap joints are


shown in Fig. 10 (for linear weld) and Fig. 11 (for circular weld).
All plots also show for comparison the tensile test curve for
DC05 steel, together with a tensile curve (dashed line) normalised
to thickness and width of welded sheets, as explained below. A
summary of experimental results is also given in Table 4. For each
tested specimen, the table gives the maximum load, the deforma-

Fig. 10. Loaddisplacement curves for welded specimens with linear weld: (a)
1.2 mm/1.0 mm sheets and (b) 1.5 mm/1.0 mm sheets. The tensile curve (grey
continuous) of DC05 steel and the scaled tensile curve Fnor/DLnor (grey dashed) are
also shown.

tion and rotation angle at fracture. The maximum load per unit
sheet width is also calculated, to allow a comparison between linear and circular welded specimens that have different width Ws of
metal sheet.
In all welded joints, the fracture occurred at base material, far
away from the weld. This conrms that a full penetrated bead,
characterised by an increased micro-hardness and thus by an
increased yield strength compared to base metal, is the strongest
part of the welded specimen. As expected, welded specimens in
two steels S355MC/DC05 failed at the weakest metal (DC05), while

D. Benasciutti et al. / Materials and Design 62 (2014) 205216

which reduces the active gauge length to approximately one half


of total sheet length. The sheet in the welded joint then behaves
with a reduced gauge length, compared to the sheet used in base
metal tensile test. Therefore, for a comparison with base metal it
is necessary to introduce a scaled displacement in Eq. (1) to
account for the different gauge length in welded specimens. For
example, by scaling the elongations to a reference length of
50 mm (i.e., half of gauge length 100 mm), the base metal curves
shifts on the left, more close to the curves of welded joints
(although a better agreement is observed only for circular weld).
It can be concluded that the overlap conguration of lap joints
reduces the deformation capability of the specimen under shear
loading, compared to a homogeneous metal sheet with identical
width and length.
The Fnor/DLnor curve dened in Eq. (1) is plotted as a dashed line
in Figs. 10 and 11, while the last column in Table 4 lists the
maximum tensile load scaled to unit width (to allow a comparison
with theoretical estimations in Table 5). A quite good agreement is
generally observed between the normalised tensile curve and the
experimental results.
During the tests, the weld bead rotates as the applied displacement increases, due to the offset of the applied load. The rotation
angle is inuenced by the weld geometry and by the thickness of
welded sheets as well. An average angle at failure of 30 was
observed for 1.2 mm/1.0 mm specimens, while a slightly lower
value of 16.3 was registered for 1.5 mm/1.0 mm specimens, indicating that a larger sheet thickness gives a higher rotation stiffness
and then a reduced rotation of weld bead. Other studies in literature report rotation angles for linear welds of about 4550 for
completely failed specimens [12], while angles less than 25 are
reported in [7]. Instead, experimental tests conrmed that specimens with circular welds give much smaller rotations (of about
34), see Table 4.

9
1.5 mm (S355)/1.2 mm (DC05) - circular weld

8
7

force (kN)

5
4
DC05 (t=1.2 mm)

sample A
sample B
sample C
DC05 (t=1.2 mm)
DC05 (t=1.2 mm, L=50 mm)

2
1
0

10

20

(a)
30

40

50

displacement (mm)
7

1.5 mm (S355)/1.0 mm (DC05) - circular weld

6
C

force (kN)

5
A

4
B

DC05 (t=1.2 mm)

3
sample A
sample B
sample C
DC05 (t=1.2 mm)
DC05 (t=1.0 mm, L=50 mm)

2
1
0

10

20

(b)
30

40

50

displacement (mm)
Fig. 11. Loaddisplacement curves for welded specimens with circular weld: (a)
1.5 mm/1.2 mm sheets and (b) 1.5 mm/1.0 mm sheets. The tensile curve (grey
continuous) of DC05 steel and the scaled tensile curve Fnor/DLnor (grey dashed) are
also shown.

specimens in DC05 sheets failed in the thinnest sheet (1.0 mm).


Some samples of fractured specimens are given in Figs. 12 and 13.
Experimental results then conrmed that the strength of tested
welded specimens is controlled by base metal. However, a direct
comparison with the tensile curve for base metals is not possible,
because of different thickness and width used in welded specimens, compare Figs. 1 and 3.
A comparison could be attempted, instead, by scaling the values
F/DL of the tensile curve for DC05 base metal to thickness and
width of welded specimen (the same normal stress is thus
obtained):

F nor



W 1 t1
F
W 0 t0

DLnor


50
DL
100

211

In the above expression, W0 = 15 mm and t0 = 1.2 mm are, respectively, the width and thickness of base metal sheet. Instead, width
W1 and thickness t1 refers to welded specimens and they depends
on weld geometry (see second and third column in Table 4): in
linear welds W1 = 15 mm and t1 = 1.0 mm, while in circular welds
W1 = 20 mm and t1 = 1.2 mm (for 1.5_1.2_C) and t1 = 1.0 mm (for
1.5_1.0_C). Note that in linear welds only the thickness t1 is
changed, while in circular welds both thickness t1 and width W1
are different compared to base metal specimen.
As shown in Figs. 10 and 11, all tested welded joints showed a
high ductility at failure, with values, however, that are lower than
base metal. Similarly to strength, also the elongation of welded
specimens cannot be directly compared to base metal. This can
be explained by the overlap length of 100 mm in welded joints,

4. Theoretical estimation of tensile strength of laser-welded lap


joints
The determination of tensile strength and failure criterion for
laser-welded lap joints is a critical issue. International design codes
do not give any theoretical approach to assess the mechanical
strength of laser-welded lap joints in thin sheets. For example, Landolfo et al. [13] proposed to adopt the indications of Eurocode 3
Part 1-3 [22] for seam and spot welds to estimate the load bearing
capacity of laser welded connections.
On the other side, interesting experimental and numerical
studies on the mechanical behaviour of laser-welded lap joints
were also proposed in the literature. For example, Ha and Huh
[6] formulated an analytical failure criterion for laser weld under
combined normal and shear loading based on experimental results.
However, even though the stress failure surface has been dened
analytically, the actual distribution of normal and shear stress on
a given welded joint, as a function of the applied load and joint
stiffness, would remain somehow to be determined.
An alternative approach to evaluate the mechanical response of
laser welds is nite elements modelling. Typically, a two-dimensional elasticplastic model, with explicit representation of the
weld bead, has been used to simulate the weld joint behaviour
[1012]. The numerical model can simulate the rotation angle of
weld bead and the loaddisplacement curve, as well as the plastic
strain distribution and the expected failure locations. However, the
approach seems not fully capable to identify a unique failure
criterion to estimate the failure load of welded joint. For example,
Terasaki and Kitamura [10] concluded that the equivalent plastic
strain calculated by nite elements analysis is only suitable to predict base metal strength, while weld shear strength must be
assessed by conventional failure theories. Similarly, the numerical

212

D. Benasciutti et al. / Materials and Design 62 (2014) 205216

Table 4
Experimental results of tensile tests on lap joints with linear and circular weld bead.

a
b

Specimen
code

Thickness, t
(mm)a

Width, Ws
(mm)

Sample

1.5_1.0_L

1.0

15

1.2_1.0_L

1.0

1.5_1.2_C

1.5_1.0_C

Experimental data
Maximum load
(kN)

Deformation at fracture
(mm)

Rotation angle at fracture


(deg)

Max. load per unit width


(N/mm)b

A
B
C

4.41
4.13
4.60

28.7
28.7
27.9

12
21
16

294
275
307

15

A
B
C

4.51
4.50
4.10

30.4
30.5
30.5

33
32
32

301
300
273

1.2

20

A
B
C

7.37
7.82
7.78

22.1
22.7
20.2

5
4
4

368
391
389

1.0

20

A
B
C

5.43
5.34
6.23

24.7
22.7
23.6

3
4
3

271
267
312

Thickness values of the sheet where fracture occurred.


The values are normalised to sheet width Ws given in the third column.

analyses in [11] identify the initiation sites of ductile fracture as


observed experimentally, even if they do not quantify the weld
limit load (the nite elements simulation can continue until
numerical instability takes place, so that the simulated load
displacement curves are monotonically not decreasing without a
maximum load). Otherwise, one should use a more sophisticated
approach to estimate the failure position in the welded joint, as
for example that based on J-integral calculations proposed by Asim
et al. [12]. However, the fracture toughness that denes the critical
J-value is a quantity not generally available, nor easy to be estimated experimentally.
Last, but not least, it goes without saying that the application of
a two-dimensional elasticplastic model to analyse complex threedimensional structures containing many welds would not be
straightforward, other than computationally quite demanding.
Instead, models with shell elements, without explicit weld bead
modelling, have been adopted with thin sheets (an example is
proposed in [23] for fatigue analysis). For such shell models,
however, a suitable criterion (based, for example, on a simplied
theoretical model) should be given to estimate the weld strength.
The question then arises whether an elasticplastic nite elements modelling could be appropriate, or whether a simplied
theoretical model could be equally effective for estimating the
mechanical strength of laser-welded lap joint conguration under
tensile loading.
In the context of an easy design approach, the above considerations
seem to justify the use of a simplied theoretical model to estimate
the tensile strength of lap joints, as a function of weld geometry and
mechanical properties, as suggested in literature [7,8].
Fig. 14 shows the deformed conguration of a lap joint under
tensile loading and the stress state at each part of the joint. The
base metal sheets, far away from the weld, can reasonably be considered to be under a uniaxial tensile stress, just like a sheet in the
tensile test. The offset position of welded sheets gives, instead, a
load misalignment that produces a bending moment, when the
specimen is subjected to remote tensile loading. This bending
moment gives a bending stress distribution in the sheet portion
adjacent to the weld (heat-affected zone). In addition, the weld
bead rotates of an angle h when increasing the tensile load F. As a
result, the weld bead is subjected to a combined normal and shear
stress, which are the dominant stress components irrespective of a
more complex stress eld that can exist around the weld [6].
The tensile strength of the welded specimen is then controlled
by the lowest strength value among base metal, weld material
and heat-affected zone.

The ultimate (or limit) load for failure at base metal sheet under
tensile loading is:

F BM ru;b W s t

2
2

where ru,b (N/mm ) is the ultimate tensile strength of base metal,


Ws (mm) is the width and t (mm) the thickness of base metal sheet.
The ultimate load of weld bead can be estimated by considering
the normal and shear stress produced by the cross-tension force
Fsin h and shear tension force Fcos h on the weld under a rotation
angle h. In a joint with linear weld, the normal and shear stresses
on the weld due to an applied load F are:

F sin h
rw;lin W
N=mm2
w tw
F cos h
sw;lin W w tw N=mm2

where the weld has thickness tw (mm) and length Ww (mm), while h
is the rotation angle under the applied force F. The thickness tw is
the weld size (weld width), which depends on laser welding parameters; it can be measured, for example, by metallurgical observations on welded joint samples. Stress components can be
assumed, with reasonable accuracy, as uniformly distributed over
the weld, as conrmed by nite element simulations [6]. The
strength condition against weld fracture is:

req r2w;lin 3s2w;lin

F p
HV w
sin 2 h 3 cos 2 h 6 ru;w 9:8
W w tw
3
4

where ru,w (N/mm2) is the ultimate tensile strength of weld metal.


Since ru,w is difcult to be determined experimentally (especially
for weld beads lower than 1 mm), expression (4) correlates ru,w
to the hardness HVw (kg) of weld metal, as often suggested in the
literature [710,24]). In expression (4), the equivalent von Mises
stress req is used to combine the normal rw and shear sw stresses
on the weld. The condition req = ru,w in Eq. (4) gives the limit load
for weld failure in linear weld:

HV w W w tw
F w;lin 3:267 p
sin 2 h 3 cos 2 h

Similarly, in a joint with circular weld the normal and shear


stresses (for simplicity assumed, uniformly distributed over the
weld) are expressed as:

h
rw;cir Fpsin
N=mm2
dtw
h
sw;cir Fpcos
N=mm2
dt w

213

D. Benasciutti et al. / Materials and Design 62 (2014) 205216

b
haz
w,lin
w,lin

Fig. 14. Stress state at each part in a lap joint with linear weld. The weld has
thickness tw and width Ww (orthogonal to load direction).

Fig. 12. Typical failures in tensile test of welded samples with linear weld.
Thickness values: (a) 1.5 mm/1.0 mm and (b) 1.2 mm/1.0 mm. Arrows indicate the
location of failure.

Fig. 13. Typical failures in tensile test of welded samples with circular weld.
Thickness values: (a) 1.5 mm/1.2 mm and (b) 1.5 mm/1.0 mm. Arrows indicate the
location of failure.

where d (mm) is the diameter of the weld line and tw (mm) the
thickness of weld bead. After calculating the equivalent stress req,
the limiting condition req = ru,w gives the following ultimate load
for failure at the circular weld:

HV w pdtw
F w;cir 3:267 p
sin 2 h 3 cos 2 h

A circular weld has a larger weld area compared to linear weld


with the same thickness tw. Then, as shown by expressions (5) and
(7), it is then expected to have a higher strength (about three times
greater) than linear weld with equal lateral width (d = Ww). In fact,

by taking d = Ww, the following strength ratio Fw,cir/Fw,lin = p is


obtained, which conrms the advantage provided by a circular
weld. In addition, as conrmed by the experimental results, the
circular welds also have a higher stiffness that gives a smaller rotation angle h compared to linear welds.
The two Eqs. (3) and (6) compute the weld stresses as a function
of the rotation angle h. As h increases during loading, the weld
stress changes from pure shear (for h = 0) to pure normal stress
(for h = 90), which tries to open the weld joint (peeling stress).
The shear stress is predominant for small h. In this study, where
lap joints were subjected to tensile loading, the condition h = 90
was not reached, because fracture occurred at much smaller
angles.
As shown by Eqs. (5) and (7), the weld strength depends on the
rotation angle h of the non-load carrying sheets. For a given applied
load F, the angle h depends on overall deformation of the welded
specimen, which is a function of its geometrical and mechanical
properties (the system is statically indeterminate). Large displacements compared to sheet thickness have also to be considered and
they make the analysis even more complex. A simple analytical
expression that correlates h to F is generally not known, as it is
rather difcult to be derived analytically. An approximate best-t
regression equation is proposed in [8] for lap joint with linear
weld:

h 45; 000

F 1:5
0:5
W 1:5
w tw t

r2:75
u;b

The experimental results in this study show a maximum rotation angle at failure of about 30 for linear welds and 34 for circular welds, see Table 4. Other studies in literature report rotation
angles of about 4550 for completely failed specimens [12], while
angles less than 25 are reported in [7]. Instead, larger angles (of
about 38.8) are estimated by Eq. (8) for linear weld. Fig. 15 shows
a typical trend of angle h calculated by Eq. (8) for a linear and
circular weld. In circular welds the angle h in Eq. (8) has been calculated by taking tw = d, to account for the much wider overlap
region where metal sheets are joined together, compared to linear
weld. Accordingly, a much lower rotation angle is estimated for
circular weld, as it also observed experimentally.

Table 5
Tensile strength (per unit width of metal sheet) at each joint portion, estimated by the theoretical model described in Section 4.
Specimen code

1.5_1.0_L
1.2_1.0_L
1.5_1.2_C
1.5_1.0_C

Weld hardness, HVw

190
175
160
160

Ultimate tensile strength (MPa)

Strength of different joint portions (N/mm)a


b

Strength of lap joint (N/mm)a


c

BM

HAZ

WM

BM

HAZ

WM

WM (small h)

283
283
283
283

452
427
403
403

621
572
523
523

283
283
340
283

292
290
482
402

372
340
712
711

358
330
711
708

BM = base metal, HAZ = heat-affected zone, WM = weld metal.


a
Values are divided by the width Ws of base metal sheet.
b
Calculated by a numerical procedure using Eq. (5) (linear weld) and Eq. (7) (circular weld), with h given by Eq. (8).
c
Calculated by the approximated analytical expressions in Eq. (9), assuming small h.

283
283
340
283

D. Benasciutti et al. / Materials and Design 62 (2014) 205216

50

100

150

200

250

300

tensile load per unit width (N/mm)


Fig. 15. Trend of angle h vs. applied load (normalised to sheet width), for a linear
weld (1.5_1.0_L) and circular weld (1.5_1.0_C), as calculated by Eq. (8).

Eqs. (5) and (7), combined with Eq. (8), require a numerical
solution to compute the strength Fw of weld bead. However, the
assumption of a small h allows Eqs. (5) and (7) to be simplied
to the following closed-form expressions:

F 0w;lin 3:267 HV wpW3w tw 1:886HV w W w t w

F 0w;cir 3:267 HV wpp3dtw 1:886HV w pdt w

with an error of about 4% for h 20 and 18% for h 45 compared to Eqs. (5) and (7), respectively. This means that expressions
(9) are on the safe side, because they estimate a lower strength
compared to the exact solutions (5) and (7).
The ultimate load FHAZ for failure at the heat-affected zone
under bending moment can be estimated by the modelling
approach proposed in [8,9], which has been shown to provide
results in reasonable agreement with experiments. The approach
is briey summarised in Appendix A.
The maximum bending stress rhaz,max at the outermost surface
in the metal sheet:

rhaz;max


n
t=2 d
E
6 ru;haz
Ri t=2  d

10

is compared to the ultimate tensile strength ru,haz (N/mm2) of the


metal in HAZ. In the above expression (10), Ri is the curvature radius
of the inner surface, t is the sheet thickness and d is the distance
between the neutral plane and the thickness centre, see Fig. 16.
The mechanical properties of HAZ are inhomogeneous, as they gradually change from those of weld metal to base metal when moving
away from weld zone. The tensile strength of HAZ is then estimated
as an average ru,haz = (ru,b + ru,w)/2 of the strengths of base and
weld metals [8,9]. On the other hand, as suggested in [25] (see
Par. 9.3.2 in Chapter 9), the strength ru,haz may further be increased

Rm

Ri
Re

r
F

rlin

F 0w;lin

F BM

(r)
haz,max

Fig. 16. Denition of parameters in the HAZ subjected to bending stress.

HV w
p
3HV BM


 
 
Ww
tw
tw
c1 c2;lin
Ws
t
t

11

while for a circular weld is:

rcir

F 0w;cir

F BM

HV w
p
3HV BM



 
 
tw
tw
c1 c2;cir
Ws
t
t

pd

12

Symbols HVw, HVBM are, respectively, the hardness values for


weld and base metal, while t, tw indicate the thickness for base
metal sheet and weld zone. The hardness ratio c1 is a coefcient
that depends on properties of welded joint materials and on the
hardness increment due to laser welding. A hardness ratio HVw/
HVBM of about 2.22.6 is commonly observed in welded joints,
which gives a coefcient c1 = 1.271.5. Instead, coefcients c2,lin,
c2,cir are function of welded joint conguration (weld geometry
and dimension).

2.5
bisector

tw,lim

2
1.5

base metal
failure

1
0.5
line

weld
metal

n=

10

rli

20

w
ith

30

by a factor of about 1.5 taking the full yielding of the cross-sectional


area in bending into account.
The maximum bending stress rhaz,max in Eq. (10) is a monotonic
function of the applied load F. This relationship, however, is
dened only implicitly through the quantities Ri and d, which both
depend on F and h through Eqs. (14) and (15) in Appendix A. As a
result, the limit load Fhaz for fracture at HAZ can only be evaluated
by numerical calculation, because no closed-form expression can
be derived.
Table 5 lists the ultimate load (per unit width Ws) of welded
specimens with linear and circular weld, estimated by the theoretical model described above. Separate values for base metal, weld
bead and heat-affected zone are also indicated. The strength of
weld bead is estimated by Eqs. (5) and (7) as a function of h, as well
as by the approximate expressions (9).
The accuracy of the analytical model can be assessed by comparing the experimental and analytical strengths given in the last
column of Tables 4 and 5, respectively. As can be seen, in all
welded specimens the analytical model correctly estimates that
fracture would occur in base metal, as it was observed
experimentally.
The results in Table 5 could also be used to compare the difference in strength between base metal and weld metal as a function
of weld bead geometry (see the columns labelled BM and WM). As
can be seen, the larger difference observed in circular welds means
a higher safety margin for fracture at weld metal, compared to base
metal failure.
The analytical expressions previously derived could provide
useful guidelines in the design of welded lap joints, based on the
strength ratio of weld and base metal. For linear weld the ratio is:

lin
e

rotation angle (deg)

40

weld thickness, t w (mm)

214

LINEAR
WELD

weld
failure

base metal
failure

weld failure

with

r cir=

CIRCULAR
WELD

sheet thickness, t (mm)


Fig. 17. Relationship between sheet thickness t and weld thickness tw. The limit
weld thickness is tw,lim = 2 mm. Markers are the experimental points (d linear
welds; h circular welds). Failure regions are also indicated.

D. Benasciutti et al. / Materials and Design 62 (2014) 205216

The above expressions (11) and (12) do not consider the


strength of HAZ, as no analytical closed-form expression is available. On the other hand, for the welded specimens under study
the strength of HAZ is always greater than base metal, see Table 5.
In Eqs. (11) and (12), a strength ratio rlin = 1 (or rcir = 1) would
indicate that the welded specimen has virtually an equal strength
for weld and base metals, while whatever value but unity would
suggest that fracture prevalently occurs at weld metal (for r < 1)
or at base metal (for r > 1).
Expressions (11) and (12) suggest that to design a joint with
equal strength for weld and base metal, the weld thickness tw
has to increase linearly with sheet thickness t. Fig. 17 shows an
example of design chart, which is calibrated on the characteristics
of the laser welds analysed in this study. For any joint conguration, the symbol t in the x-axis of Fig. 17 indicates the thickness
of DC05 steel sheet, where fracture occurred (in the graph, the line
for S355MC steel sheet is not shown for more clarity). Based on the
hardness values in Figs. 7 and 8, a coefcient c1 = 1.44 (linear weld)
and c1 = 1.27 (circular weld) is calculated.
Due to the higher strength of weld metal compared to base
metal, the weld thickness tw can be lower than any t sheet thickness (the inclined lines are below the bisector). Furthermore, the
design chart conrms the higher strength of circular welds compared to linear welds. In fact, at a given weld thickness tw, circular
welds can tolerate a higher sheet thickness t. Conversely, at any
given sheet thickness t, the circular weld geometry requires a
lower weld thickness tw to attain the same base metal strength.
The weld thickness tw, however, cannot increase indenitely
and becomes saturated with increasing plate thickness, as laser
welding gives weld beads approximately of several millimetres
even for thick plates. For example, Sokolov et al. [3] report an average weld width below 2 mm for butt-joints in thick plates of
20 mm. This effect has also been emphasised by Ono et al. [7],
who indicated a saturated thickness of about 1.2 mm above which
fracture always occurred in weld metal. The limit weld thickness
clearly depends on the material type and weld geometry, thus a
unique value cannot be given for all joint congurations. In
Fig. 17, a limit weld thickness tw,lim = 2 mm is assumed, which
could be used to nd a limit sheet thickness above which weld failure would occur: t = 2.9 mm for linear weld, t = 7.2 mm for circular
weld (not visible in the gure, because of x-scale axis).
Finally, Fig. 17 also indicates the regions where different types
of fracture would occur, according to the proposed model. As can
be seen, the markers of the welded joints tested experimentally
correctly fall within the region of base metal failure, which actually
conrm the reliability of the proposed theoretical design approach.

5. Conclusions
This paper presented a microstructural and mechanical characterisation of laser-welded lap joints in thin low carbon steel.
Welded specimens with linear and circular weld bead were studied. Characterisation was based on metallurgical observations,
micro-hardness measurements and tensile tests, which gave information on the ultimate tensile strength, rotation angle of the weld
bead and fracture mode of welded specimens. The main ndings of
this study can be summarised as follows:
 Micro-hardness proles in all tested joints conrmed an increment of hardness in the weld zone and heat-affected zone.
 All tested specimens fractured in the base metal, far away from
the weld zone. This means that the overall joint strength is controlled by base metal strength, which is consistent with a fullpenetrated weld zone characterised by higher micro-hardness
(and hence higher static strength) compared to base metal.

215

 In the experiments, lap joints with circular weld showed a


lower rotation angle under loading, compared to linear welds.
A maximum rotation angle at failure of about 30 for linear
welds and 34 for circular welds was observed.
 A simple theoretical model was proposed to estimate the
strength of base metal, heat-affected zone and weld zone in
laser-welded lap joints, as a function of material properties
and weld geometry. For the weld bead, the analytical results
estimate an increased strength (of about three times) of circular
welds, compared to linear weld having an equal lateral width.
The proposed model is also able to estimate the overall strength
of the welded specimen and to locate the nal fracture. Analytical estimations are shown to agree with experimental results. A
design chart is nally derived, which allows a proportion of
laser-welded lap joints with virtually equal strength of base
metal and weld zone.
Acknowledgements
The nancial support of this research by F.I.S.A. Fabbrica Italiana Sedili Autoferroviari S.r.l., under contract CV12_045 FISA SRL,
is greatly acknowledged.
Appendix A. Bending stress in lap joint
This Appendix briey summarises the main results of the theoretical model proposed in [8,9] to estimate the static strength of
HAZ under the bending moment produced by the load eccentricity.
The model assumes a bending deformation with constant curvature radius. The circumferential strain at a plane at a distance r
from the curvature centre is:

er

r  Re
Re

13

where Re = Ri + t/2  d is the curvature radius of the neutral plane


(where the strain is zero), Ri is the curvature radius of the innermost
surface, d is the distance between the neutral plane and the geometrical centre of the sheet, t is the sheet thickness (see Fig. 16). The
curvature radius is found to be related to the rotation angle h
[8,9] as:

Ri

t cos h  2t  t2w sin h


1  cos h

14

while parameter d can be obtained by solving the following non-linear equation:

n1 )

n (
n1 

W wE
1
t
t 

F
d
 d  
n 1 Ri t=2  d
2
2

15

The expressions (8), (14), and (15) can be used to compute the rotation angle h, curvature radius Ri and distance d for each value of the
applied force F. Both h and d are monotonically increasing functions
of F, while Ri is a decreasing function of F (when F = 0, it is d = 0 and
Ri = 1).
The bending stress at position r is calculated as r(r) = Ee(r)n,
where E is the modulus of elasticity (Youngs modulus) and n is a
work hardening exponent. They are estimated as E = ru,haz(e/n)n,
n = ln (1 + uel,haz), based on the uniform elongation uel,haz (which
is the average between the uniform elongation of base metal and
weld metal) and tensile strength ru,haz (e is the base of natural
logarithm).
The maximum bending stress occurs at the outermost surface of
the sheet (where r = Ri + t):

rhaz;max E

n
t=2 a
Ri t=2  a

16

216

D. Benasciutti et al. / Materials and Design 62 (2014) 205216

The maximum bending stress is a monotonic function of F, which


however is dened implicitly through Eqs. (8), (14), (15), and
(16). The joint is expected to fail in HAZ when the stress rhaz,max
exceeds the tensile strength ru,haz of this region.

References
[1] Graudenz M, Baur M. Applications of laser welding in the automotive industry.
In:
Katayama
S,
editor.
Handbook
of
laser
welding
technologies. Cambridge: Woodhead Publishing Limited; 2013. p. 57595.
[2] Wang H. Applications of laser welding in the railway industry. In: Katayama S,
editor. Handbook of laser welding technologies. Cambridge: Woodhead
Publishing Limited; 2013. p. 57595.
[3] Sokolov M, Salminen A, Kuznetsov M, Tsibulskiy I. Laser welding and weld
hardness analysis of thick section S355 structural steel. Mater Des
2011;32(10):512731.
[4] Yilbas BS, Arif AFM, Abdul Aleem BJ. Laser welding of low carbon steel and
thermal stress analysis. Opt Laser Technol 2010;42:7608.
[5] Farabi N, Chen DL, Li J, Zhou Y, Dong SJ. Microstructure and mechanical
properties of laser welded DP600 steel joints. Mater Sci Eng A-Struct Mater
Prop Microstruct Process 2010;527:121522.
[6] Ha J, Huh H. Failure characterization of laser welds under combined loading
conditions. Int J Mech Sci 2013;69:4058.
[7] Ono M, Kabasawa M, Omura M. Static and fatigue strength of laser-welded lap
joints in thin steel sheet. Weld Int 1997;11(6):4627.
[8] Miyazaki Y, Furusako S. Tensile shear strength of laser welded lap joints.
Nippon Steel Tech Rep 2007;95:2834.
[9] Furusako S, Miyazaki Y, Hashimoto K, Kobayashi J. Establishment of a model
predicting tensile shear strength and fracture portion of laser-welded lap
joints. In: Proc SPIE 4831, rst international symposium on high-power laser
macroprocessing; 2003. vol. 4831, no. 197.
[10] Terasaki T, Kitamura T. Prediction of static fracture strength of laser-welded
lap joints by numerical analysis. Weld Int 2004;18(7):52430.
[11] Lee J, Asim K, Pan J. Modeling of failure mode of laser welds in lap-shear
specimens of HSLA steel sheets. Eng Fract Mech 2011;78(2):37496.

[12] Asim K, Lee J, Pan J. Failure mode of laser welds in lap-shear specimens of high
strength low alloy (HSLA) steel sheets. Fatigue Fract Eng Mater Struct
2011;35:21937.
[13] Landolfo R, Mammana O, Portioli F, Di Lorenzo G, Guerrieri MR. Experimental
investigation on laser welded connections for built-up cold-formed steel
beams. J Constr Steel Res 2009;65(1):196208.
[14] Hsu C, Albright CE. Fatigue analysis of laser welded lap joints. Eng Fract Mech
1991;39(3):57580.
[15] Asim K, Sripichai K, Pan J. Fatigue behavior of laser welds in lap-shear
specimens of high strength low alloy steel sheets. Int J Fatigue
2014;61:28396.
[16] EN 10130:2006. Cold rolled low carbon steel at products for cold forming
technical delivery conditions, CEN European Committee for Standardization;
2006.
[17] EN 10149-2:2013. Hot rolled at products made of high yield strength steels
for cold forming Part 2: technical delivery conditions for
thermomechanically rolled steels, CEN European Committee for
Standardization; 2013.
[18] Rupil G. Comportamento statico di un giunto a saldatura laser, Masters Degree
Thesis, University of Udine; 2012 [in Italian].
[19] EN ISO 6892-1:2009. Metallic materials tensile testing Part 1: method of
test at room temperature (ISO 6892-1:2009), CEN European Committee for
Standardization; 2009.
[20] Yilbas BS, Akhtar S, Shuja SZ. Laser forming and welding processes. Springer;
2013.
[21] GuoMing Han, Jian Zhao, JianQang Li. Dynamic simulation of the temperature
eld of stainless steel laser welding. Mater Des 2007;28(1):2405.
[22] EN 1993-1-3:2006. Eurocode 3: design of steel structures Part 13: general
rules supplementary rules for cold-formed thin gauge members and
sheeting, CEN European Committee for Standardization; 2006.
[23] Henrysson H-F. Finite element based fatigue life predictions of laser overlap
welds in automotive structures. In: Proc of International Conference on
Fracture (ICF11), Turin, Italy, October; 2005.
[24] Pavlina EJ, Van Tyne CJ. Correlation of yield strength and tensile strength with
hardness for steels. J Mater Eng Perform 2008;17(6):88893.
[25] Radaj D, Sonsino CM, Fricke W. Fatigue assessment of welded joints by local
approaches. 2nd ed. Woodhead Publishing; 2006.