Optics & Laser Technology 64 (2014) 172183

Contents lists available at ScienceDirect

Optics & Laser Technology

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

A study of narrow gap laser welding for thick plates using

the multi-layer and multi-pass method
Ruoyang Li, Tianjiao Wang, Chunming Wang n, Fei Yan, Xinyu Shao, Xiyuan Hu, Jianmin Li
School of Materials Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, China

art ic l e i nf o

a b s t r a c t

Article history:
Received 4 November 2013
Received in revised form
22 February 2014
Accepted 18 April 2014
Available online 11 June 2014

This paper details a new method that combines laser autogenous welding, laser wire lling welding and
hybrid laser-GMAW welding to weld 30 mm thick plate using a multi-layer, multi-pass process. A Y
shaped groove was used to create the joint. Research was also performed to optimize the groove size and
the processing parameters. Laser autogenous welding is rst used to create the backing weld. The lower,
narrowest part of the groove is then welded using laser wire lling welding. Finally, the upper part of the
groove is welded using laser-GMAW hybrid welding. Additionally, the wire feeding and droplet transfer
behaviors are observed by high speed photography. The two main conclusions from this work are: the
wire is often biased towards the side walls, resulting in a lack of fusion at the joint and the creation of
other defects for larger groove sizes. Additionally, this results in the droplet transfer behavior becoming
unstable, leading to a poor weld appearance for smaller groove sizes.
& 2014 Elsevier Ltd. All rights reserved.

Keywords:
Thick plate
Narrow gap
High speed camera

1. Introduction
Thick plate welding technology is very important to the
shipbuilding, pipeline, nuclear, and submarine manufacturing
elds. The equipment for these processes is quite large, and for
plate thicknesses of 20 mm, workpieces only can be jointed piece
by piece using traditional methods. However, many problems arise
as a result of the large groove area that is necessary to join two
pieces, such as the residual stress and residual deformation that
results from the large restraints required and poor weld joint
mechanical properties because of a lack of plasticity [1]. Presently,
laser welding refers mainly to single-run autogenous operations in
which the weld joint is formed by solidifying the base metal
without the addition of any other material. This method has many
disadvantages, including the precise t-up requirements prior to
welding and the limited weld range for a certain laser output
power even if high efciency lasers are used and small welding
deformations are observed. After 40 years, the most popular two
lasers, YAG and CO2, are still the two primary lasers that are used
to join most sheet, (o 5 mm), medium and heavy plate
(5
20 mm) thicknesses. Even for the most advanced ber laser,
the single-run penetration depth that can be achieved is less than
25 mm [2]. Katayama et al. [3] achieved deep-penetration weld
beads to depths of 70 mm in Type 304 stainless steel using 10 kW

n
Corresponding author at: School of Materials Science and Engineering,
Huazhong University of Science and Technology, Wuhan 430074, China.
Tel: 86 13871541964; fax: 86 27 87543894.
E-mail addresses: cmwang@mail.hust.edu.cn, 619326146@qq.com (C. Wang).

http://dx.doi.org/10.1016/j.optlastec.2014.04.015
0030-3992/& 2014 Elsevier Ltd. All rights reserved.

and 16 kW high power disk lasers, a welding speed of 0.3 m/min

and at pressures of 0.1 kPa. Unfortunately, the cost to control the
atmosphere around the workpiece is unrealistic for industrial
production of large and thick plates. Overall, the best method for
thick plate ( 420 mm) welding is likely through the use of multipass non-autogenous welding.
Presently, few studies have been conducted on the use of laser
non-autogenous welding technologies. Hybrid YAG-MIG welding
processes have been used to successfully weld 8 mm thick aluminum alloy plate in two passes at the Huazhong University of
Science and Technology [4]. The Beijing University of Technology
[5] successfully laser welded a 20 mm thick aluminum alloy plate
with a narrow gap using ller wire in six passes with a 3.5 kW CO2
laser. The results from this project have not yet been reported.
Wang Baiping [6] et al. reported welding 16 mm thick stainless
steel plates using laser wire lling welding in three passes.
Meanwhile, a number of foreign researchers have made contributions to this eld. Osaka University welded 24.5 mm thick plate
successfully using a 10 kW ber laser, at a welding speed of 0.3 m/
min [7]. Karhu Miikka [8] et al. reported welding thick AISI 316 L
austenitic stainless steel using a 3 kW hybrid YAG laser-GMAW by
employing multiple-layers and multiple passes. Hot cracking was
observed in this weld. Hayashi, Tomotaka et al. [9] successfully
welded a 22 mm thick steel plate with a square groove and a 4 mm
gap using a high power CO2 laser-MIG hybrid welding technique
[9]. Choi and Hae Woon [10] welded 15 mm thick A572 Gr50 steel
with a 3.2 mm gap using a hybrid welding process in six passes.
MHI has made a great deal of progress in this eld with the help of
a self-produced super-power (13.5 kW) YAG laser and a specially

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

designed external ray, which permitted a 41 mm thick plate to be

laser-GMAW hybrid welded in six passes [11].
In recent years, the Corus in England, the FORCE in Denmark,
and the Fraunhofer ILT in Germany have performed research on
thick plate hybrid laser arc welding to achieve penetration depths
down to 30 mm. This work has been funded by the European Coal
and Steel Research Foundation [12]. Although western researchers,
especially Europeans, have already made many signicant contributions to this eld, thick plate laser non-autogenous welding
has not been widely adopted for practical applications. The
existing studies suffer from difculties related to process control
stability, technological adaptability, joint appearance and internal
defects (lack of fusion and porosity). Therefore, the advantages of
these three methods, laser autogenous welding, laser wire lling
welding and laser-arc hybrid welding, are a good place to start to
weld thick plate for practical applications and for different
thicknesses and materials.

173

Table 1
Elemental breakdown of Q235 steel.
element

Mn

Si

Cr

Ni

Cu

content/%

0.16

0.61

0.20

0.023

0.019

o 0.30

o 0.30

o 0.30

Table 2
Elemental breakdown of H08Mn2SiA.
element

Mn

Si

content/%

0.06
0.09

1.80
1.95

0.70
0.85

r 0.020

r 0.015

2. Experimental details.
2.1. Experimental materials and device.

Fig. 1. High speed camera optical system diagram.

The material used in this study was Q235 plate steel with
dimensions of 150 mm x 75 mm x X mm (where X 16, 18 and 30).
A 1.2 mm diameter solid ller wire was used (H08Mn2SiA).
The values X 16, 18, and 30 represent the groove total height
for the laser wire lling welding, hybrid laser-GMAW welding and
the new welding processes.
The chemical composition of the base metal and the ller wire
are given in Tables 1 and 2.
The laser equipment consisted of an IPG Photonics 4 kW solidstate Yb-ber laser system (YLR-4000) mounted to an ABB
IRB4400 special welding robot. The optics consisted of a 250 mm
focal length and a focal diameter of 300 m. A Fronius MAG arc
welding machine was used that consisted of a Fronius TPS4000
inverter power supply, a special push and pull wire feeder and a
Fronius VR2000 MIG/MAG welding torch. A CMOS high speed
camera (Photofocus, Switzerland) was used to observe the welding
process at acquisition rates of up to 10000 frames per second. To
suppress the light bloom associated with the welding process,
several optical lenses were employed that matched the wavelength (808 nm) of the laser source. As shown in Figs. 1 and 2, the
high speed camera optical lens combinations included band
attenuators to lower the intensity of the light to protect the sensor,
a narrow-band lter determined by the backlight conditions and
an outer layer glass applied to prevent the inner optical lens from
being damaged.

Fig. 2. Image of the high speed camera.

2.2. Experimental methods.

Fig. 3. Narrow gap air supply device.

Prior to welding, the grooved surface was polished and cleaned

with acetone. Every welding seam was initially rubbed with
different types of sandpaper or ground with a grinding wheel,
and cleaned by acetone. Laser autogenous welding was used rst
to form the backing weld. Pure Ar shielding gas was owed into
the groove bottom to protect the backing weld. A special shielding
gas supply nozzle, capable of entering the narrow gap groove was
designed for this purpose, as shown in Fig. 3
The parameters for the laser autogenous backing welding were:
a laser power 4 kW, a welding speed 1.0 m/min, and a defocusing length 0 mm.
Laser wire lling welding was used to ll the lower portion of
the groove. The wire led at a distance of 0.35 mm
0.65 mm and at
an angle of 301 with respect to the surface. The defocusing length
was kept at 0 mm and the same shielding gas nozzle was used.

Fig. 4 shows the Y shaped groove used. Two different angles

( 101/61) were used to study how the groove size affects the
wire feeding process and the weld shape during laser wire lling
welding in a narrow gap.
According to previous experimental results, the most appropriate parameters for the welding process are a speed of 0.5 m/min
and a wire diameter of 1.2 mm to ll the groove gap. The melt wire
was employed to ll the groove gap entirely. The wire feeding speed
is shown in eq. (1) and is determined based on the law of the
conservation of mass:
VF

V W  AG
;
AF

174

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

where VF is wire feeding speed (m/min), AG is the gap crosssectional area (mm2), VW is the welding speed (m/min), and AF is
wire cross-sectional area (mm2).
Using this equation, we can select reasonable wire feeding
speeds to ll the gap effectively based on the different groove sizes
for similar beam powers and welding speeds. Meanwhile, the
effects of different groove sizes on the weld quality for identical
wire feeding rates were studied to adjust the wire feed percentages based on the different passes. We fed wire into a 101 groove
at percentages of 40%, 30%, and 30%. The feed percentages for the
61 groove were 50% and 50%.
The constant and variable parameters used during multi-pass
laser welding with ller wire are shown in Tables 3 and 4
The distance between the laser beam and wire is maintained in
a range from  0.35
0.65 mm to form a liquid bridge transition
during the preparation period [13]. The weld bead quality and the
groove fusion play an important role in the later welding processes
during multi-pass welding. Thus the weld bead surface position
should be re-measured and the laser beam focal point position and
wire feed position should be adjusted after each pass. As shown in
Fig. 5, the mechanism for wire feeding and the stability of the
welding process was monitored by the high speed camera using
an acquisition rate of 2000 frames per second to analyze the
experimental results and optimize the welding parameters.
Laser-GMAW hybrid welding was used to weld the relatively
wider groove upper part. The arc led at a distance of 0
5 mm, at
an angle of 551from the surface and the wire length past this point
was between 15
30 mm.
According to J. B. Greens research [14], the downward force on
the melted metal can be shown by the following formula when the
melting metals lower surface end was bigger than the upper
surface end:
F I 2  log

Rb
;
Ra

downward axial force on the melted metal may decrease, leading

to solidication of the melted metal as it is transferred to the
welding pool.
As the droplet at the welding wire tip grows, the distance to the
welding pool grows at a smaller rate than the distance to the side
walls. The droplet grows until it is transferred to the welding pool.
In this way, it is easy to form an arc column using the side walls as
the cathode, leading to melted metal being transferred to the side
walls after the distance between them decreases.
According to the aforementioned research, Rb2 o Rb1 may be an
important factor that promotes smooth droplet transfer. A groove
with three different angles ( 401, 601, and 901), as shown in
Fig. 6, was designed to meet the requirements and monitored
using the high speed camera. The technical parameters are shown
in Table 5.

Table 4
Multi-pass laser welding with ller wire variable process parameters.
Vf (m/min)

angle

Vf (the rst pass)

Vf (the second pass)
Vf (the third pass)

101

61

4.5
3.4
3.4

3.3
3.3

where F is the axial force on the melted metal, I is current, Rb is the

lower part of the ball shaped melted metal (the current ows from
this area to the radius of the melted metal spherical cap, J. B - the
green shaded area denes the effective area), and Ra is wire radius.
The effective arc areas for at plate welding and narrow gap
welding are shown in Figs. 3 and 4, respectively.
The gap welding pool was relative narrow. Therefore, the
spherical cap radius of the effective area is Rb2 oRb1, and the
Fig. 5. The high speed camera monitor device in the narrow gap laser wire lling
welding.

16mm

18mm

6mm

6mm

Fig. 4. Schematic of the groove angle used for multi-pass laser welding using
a ller wire.

Fig. 6. Laser arc hybrid welding groove size.

Table 3
Multi-pass laser welding with ller wire constant process parameters.
Laser
power PkW
4

Welding speed
Vwm/min
0.5

Gas ow m3/h
0.1

Defocusing
amount (mm)
0

Wire feed style

Wire feed angle

Wire leads

451

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

3. Experimental procedure and analysis

3.1. The inuence of the narrow gap on the wire feeding mechanism
for laser welding using ller wire
3.1.1. Experimental results and analysis
The laser welding operations were monitored using the high
speed camera. The following phenomena were identied and
described below.
(1) The wire melted into groove smoothly for most of the welding
processes.
The high speed images of the wire melting into the groove and
showing the behavior of the multi-layer welding process for
Table 5
Laser-arc hybrid welding parameters.
Laser power

4 KW

current
Laser arc distance
Defocusing amount
gap
Welding direction
Welding speed
Groove angle
Shielding gas
Gas ow
Stick out
Arc length
Arc mode

120 A
3 mm
 2 mm
0.5 mm
Arc leads
0.7 m/min
901, 601, 401
Ar 20% CO2
1.1 m3/h
20 mm
2 mm
Continues mode

175

the 61 and 101 grooves are shown in Figs. 7 and 8. The wire was
fed into the groove, as shown in the top portion of the still
image and the plasma is the white spray surrounding the wire.
The welding pool was beneath the tip of the wire. The wire
was fed into the welding pool smoothly so that dynamic
droplet transfers could be observed in the high speed photos.
(2) A wire deection phenomenon was observed during the
second laser welding pass using the ller wire with a 101
groove.
The high speed images from the second laser pass using wire
lling welding are shown in Figs. 9 and 10. It is easy to
recognize the two sides of the groove from the weld bead
position. In Fig. 9, we can see the wire deecting during the
second laser welding pass in the 101 groove with the only
limits being the two sides of the groove. In Fig. 10, the welding
pool was biased to one side of the groove. The weld bead was
also biased to the same side during the third laser welding
pass with the 101 groove.
(3) The wire melts into the welding pool smoothly and steadily for
the 61 groove when monitored with the high speed camera.
Additionally, the nal appearance of the weld was generally
good. However, the wire deection phenomenon appeared
during the second and third passes when using the 101 groove.
This was mainly because the weld bead surface became
uneven and the wire feeding position during the later welding
pass was greatly affected. This affected the ow and accuracy
of wire feeding process and changed the laser-arc distance.
Although the wire deection phenomenon had almost no
impact on the wire liquid bridge transfer mode, the wire
feeding speed could no longer be timely adjusted, resulting in

Fig. 7. Wire feeding high speed images during laser welding for the 101 groove size with the ller wire. (a) The rst pass, (b) the second pass and (c) the third pass.

Fig. 8. Wire feeding high speed images during laser welding for the 61groove size with the ller wire. (a) The rst pass and (b) the second pass.

176

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

Fig. 9. High speed photos during the second laser welding pass using ller wire in a 101 groove.

Fig. 11. Micrograph of the weld cross section using a 101groove.

Fig. 10. The third pass welding pool biased towards one side of the groove.

the wire continuing to enter the groove and worsen the wire
deection problem.
(4) The wider the gap was, the easier it was to cause the wire
deection phenomenon during narrow gap multi-pass welding. The wire deection phenomenon did not occur during the
rst pass because the smaller size of the gap limited the wire
deection to some extent. The wire deection phenomenon
was not observed during the third pass for the 101 groove
because the frozen weld bead formed by the previous pass
blocked the fed wire. Additionally the plasma created by the
wire deection phenomenon made observations difcult.
However, we can still conclude that the wider gap increased
the possibility that the wire feeding may become unstable.

3.1.2. Test results and analysis

An 18 mm thick mild steel plate was welded by laser autogenous backing welding and laser welding with ller wire. After
detecting the weld bead in the different grooves using ultrasonic

testing and analyzing the weld bead cross-sectional appearance,

several items were apparent.
(1) Ultrasonic attenuation detection
We performed ultrasonic testing along the longitudinal axis of the
weld and identied defects beneath the surface at approximately
2.2
4.2 mm and 7.5
9 mm for the 101 groove and occasional
defects at a depth of 2 mm and 4
6 mm for the 61 groove.
(2) Weld bead cross section macro-appearance analysis.
Metallographic images of the multi-pass laser wire lling
welds with the 101 and 61 grooves are shown in Figs. 11 and
12, respectively. It was found that a lack of fusion problems
existed at the upper part of the second and third passes for the
101 groove (see Fig. 11). Additionally, the same defect was
occasionally found at the upper part of the second pass for the
61 groove (see Fig. 12). From the ultrasonic testing results, we
can conclude that it was very possible that the lack of fusion
defects at a depth of approximately 2
4 mm existed along
most of the longitudinal axis of the weld.
The defects from a lack of fusion on the side wall, caused by the
ller wire deection, occurred at the upper part of the second pass
weld bead for the 101 groove. The laser energy went rst into

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

melting the wire was then absorbed by the side walls when the
wire shifted to one side. Accordingly, problems associated with the
lack of fusion were identied on the side wall.
The melted metal did not spread completely to the right side of
the weld top, as shown in Fig. 11. Meanwhile, even if the wire was
fed into groove steadily, it still caused lack of fusion problems on
the side wall when the groove gap increased to a certain extent.
We therefore may conclude that a gap that is too large causes the
wire deection phenomenon, which were the two major reasons

177

for problems associated with the lack of fusion on the side walls.
Slag and porosity defects were observed at two adjacent weld
heating overlap regions and at the fusion center zone. It was
difcult to remove the oxide slag (SiO2 and MnO) and the Si and
Mn present in the base metal and the welding wire. The highmelting temperature slag was difcult to remelt and rose to
surface, causing slag to be present at the bottom of the weld and
causing porosity defects during the following pass. These defects
were found to always occur at the two adjacent weld overlapping
region.
(3) Lack of fusion during laser autogenous backing welding.
The penetration depth for the 101 groove was deeper than for
the 61 groove, as shown in Figs. 11 and 12. The smaller the
groove angle was, the shallower the penetration depth was
during laser autogenous backing welding. This was mainly
because the side walls of the smaller groove gap absorbed
more laser energy and decreased the amount of energy that
reached the groove bottom.

3.1.3. Welding quality optimization schemes

Fig. 12. Micrograph of the weld cross section using a 61groove.

(1) Optimization schemes to decrease the of lack fusion

According to the analysis, the root cause of the lack of fusion
was the laser energy not being absorbed adequately by the

Table 6
Droplet transfer behavior on a at plate.

Flat plate

0ms

8ms

16ms

17ms

18ms

19ms

0ms

39ms

46ms

47ms

48ms

welding

Droplet transfer average frequency:74/sec

Table 7
Droplet transfer behavior for the 901 groove angle.

Groove
Angle
(

Droplet transfer average frequency 26/sec

49ms

178

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

Table 8
Droplet transfer behavior for the 601 groove angle.

0ms

41ms

46ms

47ms

49ms

50ms

Groove
angle
(

56ms

57ms

58ms

6ms

9ms

11ms

22ms

55ms

70ms

75ms

81ms

Droplet transfer average frequency 12/sec

Table 9
Droplet transfer behavior for the 401 groove angle.

0ms

Groove
angle
(

Droplet transfer average frequency -/sec

groove side walls, making it difcult to form a bond between

the side walls and the fused wire. The defects associated with
the lack of fusion are almost eliminated for the 61 groove,
meaning the side walls absorbed the laser energy more
efciently and that the wire was feed into the groove more
evenly, improved the welding quality and decreasing the
number of weld defects. In addition, decreasing the groove
size can also reduce the wire lling content and improve
welding efciency.

Therefore, the most important conclusion was that the groove

size should be kept less than 2.5 mm for laser wire lled
welding when attempting to weld a 30 mm thick plate.
(2) Optimization schemes to decrease slag
The groove gap was too narrow to clean the weld bead entirely
after each pass, leading to the formation of slag defects. A
1.2 mm thick cleaning plate wrapped in a disposable cotton
swab was designed to clean the weld bead and groove side
walls after each welding pass.

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

3.2. The droplet transfer mechanism in laser-GMAW hybrid welding

under narrow gap conditions
3.2.1. Experimental results and analysis

179

was relatively faster, resulting in a smaller droplet and a more

stable welding process. Compared to the at plate hybrid
welding process, the droplet grew bigger if slightly shaken
before transfer and the transfer frequency decreased to

(1) The inuence of the groove angle on laser-GMAW hybrid

welding droplet transfer.
Droplet transfer high speed images from the laser-GMAW
hybrid welding using different groove angles and average
droplet transfer frequencies are shown in Table 69.
The key hole, in the left of the images, was formed by the laser
so that welding wires could be fed into groove from the right
side. The wire tips are observed in the droplet transfer necking
down images. Each set of images shows a complete droplet
transfer cycle. From the high speed images, we can see that the
arc was shorter, the droplet diameter was no more than the
wire diameter, and the transfer mode was incomplete during a
short circuit transfer demonstrating a spray-like feature for at
plate conditions. However, the droplet grew bigger and the
transfer mode changed to a drop-like shape transfer during
shaking under narrow gap conditions.
The droplet transfer behavior during laser-GMAW hybrid
welding of the at plate is shown in Table 6. The calculated
droplet transfer frequency was 74 times/s. The transfer speed

Fig. 14. The weld appearance.

Fig. 15. The weld cross section showing two hybrid welding passes.

Fig. 13. The tendency curve between transfer frequency and groove angle.

Fig. 16. The groove geometry and sizes.

Table 10
The inuence of the groove angle on the weld appearance.

Laser power=4kw current=80A welding speed=0.7m/min

DLA=4mm defocusing distance=-2mm gap=0.5mm
Groove angle()
40

60

90

Flat plate

2.6mm

2mm

1.5mm

Weld cross
section
appearance

Surface
appearance
Unfused
depth

180

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

26 times/s for the 901 groove angle conditions, as shown in

Table 7.
The droplets shaking to the groove side walls and then falling
into the side walls or welding pool to complete transfer
behavior with calculated at a transfer frequency 12 times/s
for the 601 groove shown in Table 8. When the groove angle
decreased to 401, the droplet transferred continuously to a
single side wall, as shown in Table 9. Although the key hole
was blocked by a frozen droplet, we still see the droplet
transfer became unstable, causing the transfer position and
continuity to become uncertain. This resulted in the melted
metal freezing on the both sides of weld bead groove.
As a result, the groove angle has a great effect on the droplet
transfer behavior for a Y shaped groove under narrow gap
welding conditions. As the groove angle decreased, the droplet
transfer frequency decreased. When the groove angle reached
its limitation, it caused the droplet to transfer to the groove
side walls, rendering the system unstable. Fig. 13 shows the
tendency curve between the transfer frequency and the
groove angle.
(2) The inuence of the groove angle on the weld appearance
during laser-GMAW hybrid welding
To study how the groove angle affects the weld quality, a
deeper analysis was conducted based on the weld appearance
and cross-sectional macro-morphology. A close relationship
was observed between the weld appearance and the droplet
transfer mode for the same conditions. From the high speed
images, we can see that even if the droplets grew bigger and
the transfer frequency decreased, as for at plate welding, the
transfer process was still stable and the weld appearance was
satisfactory for the 901 groove angle conditions. When groove
angle decreased to 601, the weld longitudinal surface uctuated due to the droplets shaking before the transfer and
resulting in an uncertain transfer cycle. The droplets transferred to a single side wall for the 401 groove angle conditions.
The sh-scale pattern on the weld surface was uneven, the
weld surface was irregular, and a number of droplets froze
without any pattern on the both sides of groove.
Therefore, we conclude that the unstable droplet transfer stage
was the main reason for the poor weld appearance resulting
from narrow gap laser-GMAW hybrid welding.
Additionally, the laser-GMAW hybrid welding penetration
decreased as the groove angle decreased. As shown in
Table 10, the fusion zone in the weld cross section increased
from 1.5 mm to 2.6 mm when groove angle decreased from

901 to 401. This result conrmed the fact that narrower side
walls absorbed more laser energy leading to a lack of fusion at
the root face during laser autogenous backing welding.

3.2.2 Weld metal re-melting during multi-layer, multi-pass

hybrid laser-GMAW welding of a 18 mm thick plate
As shown in Fig. 14 and 15, there were no obvious weld defects,
such as porosity or slag. However, two main conclusions were
reached. The side wall erosion phenomenon was considered to be
a serious issue, showing that the weld width was much wider than
the groove gap, which meant the welding heat input was much
larger. The second hybrid welding pass remelted most of the rst
pass, which wasted a signicant amount of energy and reduced
the weld quality. During the narrow gap multi-pass hybrid laser
welding process, the laser energy was mainly used to heat and
melt the former weld, whereas the side wall temeperature was
lower, which caused dendrities to rapidly grow from the fusion
line to the weld seam center due to supercooling. The narrow gap
weld with a taller channel and narrower sizes could not dissipate
enough heat, which caused the crystal to be unable to maintain a
consistent orientation and grew larger. Accordingly, the former
weld should not be repeatedly remelted to prevent the crystal
from growing. As a result, the laser power was decreased to 1 kW.

Fig. 18. The 30 mm thick plate laser non-autogenous weld bead appearance.

Fig. 19. The cross-section of the weld bead produced from laser non-autogenous
welding of the 30 mm thick plate.

Fig. 17. The wire feeding nozzle.

Table 11
The 30 mm thick plate multi-layer, multi-pass laser welding parameters.
Welding pass
number

Laser
power (kW)

Welding
speed (m/min)

Wire feeding
speed (m/min)

1
2
3
4
5
6
7

4
4
4
1
1
1
1

1
0.5
0.5
0.4
0.4
0.4
0.4

4.5
4.5
7
7
7
7.5

current (A)

Laser arc
distance (mm)

Defocusing
distance (mm)

230
230
230
240

3
3
3
3

0
0
0
0

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

3.3. 30 mm thick plate multi-layer, multi-pass laser

non-autogenous welding
3.3.1. Experimental design
To reduce the wire consumption during the welding of large
thicknesses, a new welding method was theorized that combined
three welding methods: laser autogenous welding to rst form a
backing weld, laser wire lling welding to ll parts of groove, and
nally laser-GMAW hybrid welding to ll the remaining volume in
the groove. The protective gas devices and weld cleaning devices
described above were used for this process as well.
3.3.2. Groove and wire feeding device design
A 5 mm root face was left at the bottom of the groove with the
platform of each side measuring 0.7 mm. It has been shown that a

Fig. 20. micro photo of magnied slag 500 times.

181

5 mm tall root face can be fully penetrated using a 4 kW laser and

a 1 m/min welding speed with a narrow gap and a platform at the
groove bottom. Moreover, the platform was useful to create a high
quality backing weld that could be used to suppress the second
pass weld defects. The total width of the two sides of the platform
using a butt joint was D 2n0.7 1.4 mm, which was designed to
allow a 1.2 mm diameter welding wire to enter into groove bottom
smoothly.
For the laser welding using ller wire portion of the operation,
the height of the groove was 12 mm, which was determined by the
stiffness of the wire. The groove angle was 21, and the wire was
required to remain straight on the groove bottom to assure an
accurate relative position between the laser and wire, which
guaranteed the weld quality when the wire feeding nozzle was
too large to enter the groove. The angle was to address two items:
the groove size that affected wire consumption and a groove that
was excessively large resulted in a lack of fusion during welding
under narrow gap conditions. Accordingly, the groove angle
should be as small as possible without affecting the amount of
laser energy being deposited into the groove. According to the
results, a groove top width smaller than 2.5 mm was appropriate.
The ability of the device used for laser welding with the ller
wire to smoothly enter into the groove and the reduced wire
consumption are the two main factors the laser-GMAW hybrid
welding groove was designed for, as shown in Fig. 16.
As shown in Fig. 17, a specially designed wire feeding nozzle
made of high strength copper was fabricated to be able to enter
into the upper part of the groove smoothly. The inner diameter
size was determined by the wire diameter to ensure that the wire
is fed-in smoothly.
The optimized process parameters to a weld 30 mm thick plate,
shown in Table 11, were designed based on the experimental
results from welding a 16 mm thick mild steel using a narrow gap
with laser wire lled welding and from an 18 mm thick narrow
gap laser-GMAW hybrid welding operation.

Fig. 21. The HAZ microstructural morphology produced from laser autogenous welding: (a) the total HAZ, (b) the fusion zone, (c) the ne grained zone, (d) the overheat
zone, and (e) the incomplete recrystallization zone.

182

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

Fig. 22. The HAZ microstructural morphology produced from laser welding with a ller wire: (a) the total HAZ, (b) the fusion zone, (c) the overheat zone, and (d) the
incomplete recrystallization zone.

Fig. 23. The HAZ microstructural morphology produced from laser-GMAW hybrid welding: (a) the total HAZ, (b) the fusion zone, (c) the overheat zone, and (d) the
incomplete recrystallization zone.

R. Li et al. / Optics & Laser Technology 64 (2014) 172183

The wire consumption was calculated by formula (1). LaserGMAW hybrid welding laser power was decreased to 1 kW based
on the analysis in section 3.2.2.
3.3.3. Experimental result and analysis
The 30 mm thick plate was successfully welded. The weld
appearance and cross-section are shown as Figs. 18 and 19. It is
clear that the weld was very even and its appearance was
satisfactory.
(1) Weld defects and analysis.
Weld defects were found at the junction between rst laser
autogenous welding pass and the second laser wire lling
welding pass when the weld cross-section was examined.
When the cross-section was examined at 500x, as shown in
Fig. 20, holes that resulted from the platform at the bottom of
the groove being too narrow are observed, making it difcult
to clean out the residue completely after welding.
(2) The weld heat affected zone (HAZ) microstructural morphology is shown in Figs. 2123.
The HAZ produced using laser autogenous backing welding
was the narrowest, followed by the laser welding with ller
wire HAZ and the HAZ produced using laser-GMAW hybrid
welding. The local ne crystalline structure only occurred in
the HAZ from the laser autogenous backing welding. The HAZs
produced from laser welding with ller wire and laser-GMAW
exhibited a ne crystalline structure that was mixed between
the overheat zone structure and the incomplete recrystallization structure. These two zones were therefore wider, which
was the result of being repeatedly heated.
4. Conclusion
(1) During narrow gap, multi-pass laser wire lling welding, the
wire was biased to the side walls of the groove if the gap was
increased past a certain point.
(2) At the top of the weld top, where the groove gap was largest,
side wall lack of fusion defects were observed for the narrow
gap multi-pass laser welding with ller wire operation.
(3) Slag and porosity defects occurred often and their position
positions adhered to some rules during narrow gap, multi-pass
laser welding with ller wire operation. Most of the slag and
porosity was caused by the incompletely cleaned scale that
was left by the previous welding passes.
(4) A small distance between the side walls and a narrower
welding pool resulted in changes to the electromagnetic
contraction and the plasma jet forces. This small distance also

183

decreased the droplet transfer frequency or made the droplets

adhere to the side walls during narrow gap, multi-pass, laserGMAW hybrid welding.
(5) The 30 mm thick plate was satisfactorily welded using one
pass of laser autogenous backing welding, two passes of laser
wire lling welding, and four passes of laser-GMAW hybrid
welding.

Acknowledgements
This work was supported by the National Natural Science
Foundation of China (Grant No. 51375191) and the National
Program on Key Basic Research Project (973 Program) of China
(Grant No. 2014CB46703).
References
[1] Cunyin Hu, Fuju Zhang. Narrow gap welding technology and economical
characters analysis. Welding Technology 2001;30(2):478.
[2] Xiao Rongshi Kang Li, Wuxiong Yang. Chen Kai. The new technology of highstrength aluminum alloy thick plate narrow gap laser welding. New Technology& New Process 2006;5:635.
[3] Katayama S. Deep penetration welding with high power disk lasers in low
vacuum. 30th International Congress on Applications of Lasers and ElectroOptics 2011:66978.
[4] Tianjiao Wang. Research on Narrow Gap Non-autogenous Laser Multi-pass
Welding. Master Dissertation. Wu Han: Huazhong University of Science and
Technology library; 2013.
[5] Arata, Y, Maruo, H, Miyamoto, I, Nishio, R. High power CO2 laser welding of thick
plate-Multi pass welding with ller wire. Transactions of JWRI, 15(2):27-34.
[6] Baiping Wang, Yong Zhao, Jian Huang. Investigation on Microstructure of
Thick Plate Stainless Steel Joint Welded by Multi-Pass Laser Welding with
Filler Wire. Chinese Journal of Lasers 2013;40(2):15.
[7] Roman JM, Kechemair D, Ricaud JP. CO2 laser welding of very large thickness
materials with wire ller. Welding International 1994;8(5):3769.
[8] Ishide, T, Hashimoto, Y, Akada, T, Nagashima, T, Hamada, S. The latest YAG laser
welding system-development of hybrid YAG laser welding technology. Proceedings of the Laser Materials Processing Conference, ICALEO '97. Laser
Institute of America. 1997, 49-156.
[9] Hayashi T, Katayama S. High power CO2 laser MIG hybrid welding process for
increased gap tolerance-Hybrid weldability of thick steel plates with square
groove. Quarterly Journal of the Japan Welding Society 2003;21(4):52231.
[10] Choi Hae Woon. Hybrid welding process for sheet metal and narrow gap ll
pass. Transactions of Korean Society of Mechanical Engineers 2008;32
(11):97883.
[11] The communication lecture information of Japanese experts in 2008.
[12] Webster S, Kristen JK, Petring D. Joining of thick section steels using hybrid
laser welding. Ironmaking and Steelmaking 2008;7:496505.
[13] Yangchun Yu. The wire melted into behavior and process research in laser
welding with ller wire. Doctoral dissertation. Wu Han: Huazhong University
of Science and Technology library; 2010.
[14] Greene, W.J. An Analysis of Transfer in Gas Shielded Welding Arcs, A.I.E.E.
Winter General Meeting, New York (1960) (T.A.I.E.E.).