metals

Article
Effect of Heat Input on Weld Formation and Tensile
Properties in Keyhole Mode TIG Welding Process
Zhenyu Fei 1,2 , Zengxi Pan 1,2 , Dominic Cuiuri 1,2 , Huijun Li 1,2 , Bintao Wu 1 , Donghong Ding 3, *
and Lihong Su 1,2
1 School of Mechanical, Materials, Mechatronic and Biomedical Engineering, University of Wollongong,
Northfield Avenue, Wollongong, NSW 2522, Australia; zf996@uowmail.edu.au (Z.F.);
zengxi@uow.edu.au (Z.P.); dominic@uow.edu.au (D.C.); huijun@uow.edu.au (H.L.);
bw677@uowmail.edu.au (B.W.); lihongsu@uow.edu.au (L.S.)
2 Defence Materials Technology Centre, 24 Wakefield Street, Hawthorn, VIC 3122, Australia
3 School of Mechatronic Engineering, Foshan University, 33 Guangyun Road, Foshan 528225, China
* Correspondence: dd443@uowmail.edu.au; Tel.: +86-13054466797




Received: 20 October 2019; Accepted: 4 December 2019; Published: 7 December 2019


Abstract: Keyhole mode Tungsten Inert Gas (K-TIG) welding is a novel advanced deep penetration
welding technology which provides an alternative to high power density welding in terms of
achieving keyhole mode welding. In order to facilitate welding procedure optimisation in this newly
developed welding technology, the relationship among welding parameters, weld formation and
tensile properties during the K-TIG welding was investigated in detail. Results show that except for
travel speed, the heat input level also plays an important role in forming undercut defect by changing
the plasma jet trajectory inside keyhole channel, leading to the formation of hump in the weld centre
and exacerbation of undercut formation. Both undercut defect and root side fusion boundary can act
as a stress concentration point, which affects the fracture mode and tensile properties considerably.
The research results provide a practical guidance of process parameter optimisation and quality
assurance for the K-TIG welding process.

Keywords: K-TIG; heat input; weld formation; tensile properties; welding procedure optimisation

1. Introduction
Tungsten Inert Gas welding, also known as TIG or GTAW, is a clean and widely used welding
technology for metals. Nevertheless, a main disadvantage of this welding process is the limited
penetration ability. It may require multipasses to complete the weld for medium thickness materials,
which results in very low productivity. In order to overcome this drawback, K-TIG was developed as a
new TIG variant to increase the penetration capability and production efficiency. In K-TIG welding, a
free burning arc is applied as the heat source to melt the weldment by forming a cylindrical keyhole [1],
which provides a nice alternative to high power density welding. This type of keyhole welding is
characterized by its dependence on arc force, as opposed to ablation pressure in laser-based and
electron beam-based welding, and a combination of stagnation and recoil pressure in plasma arc
welding (PAW), as reported by Jarvis and Ahmed [2]. Compared with standard melt-in mode TIG
welding, K-TIG process possesses higher energy density, process efficiency and greater penetration
ability. Although heat density of K-TIG is far lower compared to high power density welding, it is
easier to operate and more cost-effective, as stated by Liu et al. [3]. In addition, Liu et al. [4] reported
that K-TIG was able to achieve continuous open keyhole and was more stable than PAW. It was also
demonstrated by Liu et al. [5] that K-TIG had much wider operating window compared with PAW
owing to the fact that it has much smaller arc pressure/arc current ratio.

Metals 2019, 9, 1327; doi:10.3390/met9121327 www.mdpi.com/journal/metals

Metals 2019, 9, 1327 2 of 15

To date, K-TIG has found applications in industries for joining medium thickness materials, say
titanium [6], zirconium [7], stainless steel [8], low carbon steel [9] and dissimilar metal welding [10].
It was shown in these studies that K-TIG can complete the medium thickness weld in a single pass
with a V-shaped morphology. It was also shown that K-TIG produced a welded joint with smaller
fusion zone and larger heat-affected zone compared with conventional TIG welding process. In
addition, it has been demonstrated that K-TIG has the potential to increase the productivity while
maintaining the mechanical properties at a level similar to the joints produced by conventional TIG
welding. Recently, it was reported by Cui et al. [11] that K-TIG can even be used in underwater
condition and can produce a duplex stainless-steel weld that meets underwater welding standards. In
addition, because of the fact that the material properties of some welded joints produced by K-TIG
are not satisfactory, several optimisation technologies have been introduced into the K-TIG welding
process. For example, Fei et al. [12] introduced filler materials into K-TIG to improve the weld
microstructure and mechanical properties of armour steel weld. Fei et al. [13] also applied specially
designed interlayer to tailor the microstructure and solidification sequence of K-TIG-welded armour
steel joint. Post-weld heat treatment and ultrahigh frequency were utilised by Xie et al. [14] to improve
both the mechanical properties and corrosion resistance of 430 ferritic stainless steel welded joint. On
the other hand, since K-TIG has a narrower operating window for carbon steel because of its higher
thermal conductivity, several studies have focused on the keyhole behaviour and stability improvement,
such as the application of back purging unit [15], one pulse one keyhole technique [16] and high
frequency pulse current waveform [17]. Furthermore, in order to develop on-line control strategy for
the K-TIG welding process, several penetration monitoring systems have been developed, such as
arc voltage-based oscillation frequency sensing [18] and combination of acoustics and vision-based
sensing [19].
Overall, the current research directions of the K-TIG welding technology can be divided into four
parts, namely metallurgical qualification of K-TIG welded joints, optimisation technology for joint
microstructure and properties, process dynamics and stability improvement as well as development
of on-line monitoring system. Up to date, there has been little research on the relationship among
process parameters, weld formation and mechanical properties in the K-TIG welding process. More
importantly, unlike conventional TIG welding for which the weld formation can be partially controlled
or compensated by the operators, the K-TIG operates in a completely automated mode, with the weld
formation being entirely dependent on the welding parameter combination. Although in some studies,
small amount of welding trials were conducted to test the influence of welding parameters in the
K-TIG welding, the main purpose was to find a parameter combination to achieve full penetration,
as did by Cui et al. [8] and Feng et al. [20]. The mechanism for defect formation and the correlation
between weld formation and mechanical properties in the K-TIG welding process have not been
addressed. To a practical view, the mechanical properties, especially tensile properties, are easily
affected by weld formation which is a function of process parameters used during welding. Figuring
out the relationship among welding parameters, weld formation, and mechanical properties is helpful
to the process optimisation and is crucial for the avoidance of defect formation and premature failure
during service.
In this study, K-TIG welding was performed on 6.2 mm thickness high hardness armour (HHA)
plates using different welding parameter combination, aiming at evaluating the effect of heat input on
weld formation and tensile properties. The mechanism by which various behaviours occurred has
been discussed. The results would provide a practical guidance of parameter optimisation and quality
assurance for the K-TIG welding process.

2. Materials and Methods

The material used for investigation is 500 grade armour steel, also known as HHA. Detailed
chemical composition and mechanical properties of the base plate are listed in Tables 1 and 2,
respectively. The chemical composition was measured by atomic emission spectroscopy. The base
Metals 2019, 9, 1327 3 of 15

Metals 2019,
metal shows9, x tempered
FOR PEER REVIEW 3 of as
martensite microstructure, as shown in Figure 1. It is worth noting here that 15

this paper is intended to analyse the relationship among welding parameters, weld formation and
as this paper is intended to analyse the relationship among welding parameters, weld formation and
tensile properties, the material-dependent aspect will not be discussed in detail.
tensile properties, the material-dependent aspect will not be discussed in detail.
Table 1. Chemical composition of HHA (wt%).
Table 1. Chemical composition of HHA (wt%).
C Si Mn P S Ni Cr Mo B Fe
C Si Mn P S Ni Cr Mo B Fe
0.27 0.3 0.3 0.014 0.0025 0.19 1.05 0.25 0.0012 Bal.
0.27 0.3 0.3 0.014 0.0025 0.19 1.05 0.25 0.0012 Bal.

Table 2. Mechanical properties of HHA.

Table 2. Mechanical properties of HHA.
Materials Ultimate Tensile Strength (MPa) Elongation (%) Hardness (HV)
Materials Ultimate tensile strength (MPa) Elongation (%) Hardness (HV)
HHA 1775 14 495
HHA 1775 14 495

Figure 1.
Figure 1. Microstructure of the
Microstructure of the HHA
HHA base
base metal.
metal.

depicted in
As depicted inFigure
Figure2,2,the
the experimental
experimental equipment
equipment designed
designed to conduct
to conduct the welding
the welding trialstrials
was
was composed
composed of a K-TIG
of a K-TIG powerpower
supply,supply, a data acquisition
a data acquisition device
device and and a high-speed
a high-speed camera. Thecamera. The
welding
weldingwas
output output was provided
provided by apower
by a special specialsource
powerwith source with capability
capability up to 1000 up A.
to A1000 A. A high-speed
high-speed camera
camera
in in combination
combination withtransparency
with a 10% a 10% transparency neutral
neutral filter wasfilter wasbeside
placed placedthe
beside the worktable
worktable in order to in
order tothe
record record the arc behaviour.
arc behaviour. An inert Angas inert
back gas
purgeback purge
unit was unit was designed
designed to provide toshielding
provide shielding
from the
from the atmospheric
atmospheric contamination
contamination on the rooton theofroot
side the side
weldofpool.
the weld pool. shielding
The argon The argongas shielding gas was
was transferred
transferred
into the backinto the back
purging unitpurging
from the unit from the
left-hand sideleft-hand
hose holeside hose
and hole and
allowed to beallowed
ejectedto be ejected
through the
throughonthe
switch theswitch on the
right-hand sideright-hand side to avoidofaccumulation
to avoid accumulation pressure inside of the
pressure inside
unit. The the capturing
image unit. The
image
was capturingafter
conducted wasthe
conducted after thestable.
process reached process reached
During stable.the
welding, During welding,
torch and the torch
high-speed and
camera
high-speed
were camera were
kept stationary whilekept
thestationary
worktablewhile
moved theatworktable
the pre-setmoved
speed.at the pre-set speed.
To test the influence of process parameters on weld formation and tensile properties, seven
butt-joint welding experiments were performed on 6.2 mm thickness HHA plates by using control
variable method. Process window identification tests were conducted to find the appropriate parameter
combination leading to full penetration with the absence of both incomplete penetration because of
insufficient heat input and weld pool collapse because of excessive heat input. Detailed parameter
combination is depicted in Table 3. The heat input was calculated by the product of welding current
and voltage divided by the travel speed plus a thermal coefficient of 0.6. The arc length was measured
by the stacked thickness measurement device before welding. The dimension of the plates used in this
study is 250 mm × 75 mm × 6.2 mm. The welding current and travel speed were chosen as variables.
The fixed parameters for all experimental groups are depicted in Table 4. After welding, both transverse
and longitudinal cross sections of the welds were cut from the weldments. The longitudinal cross
section was cut in the arc extinguishment area in a bid to capture the in-situ profile during welding.
The surface morphology and cross section macrographs were captured by Nikon digital camera and
Figure 2. Schematic of experimental system.
Metals 2019, 9, 1327 4 of 15

Figure 1. Microstructure of the HHA base metal.

Leica M205A stereomicroscope, respectively. The mounted samples were etched with 2 wt% nital
for both
As macro
depictedandinmicrostructure observation. equipment
Figure 2, the experimental Microhardness test was
designed undertaken
to conduct using Struers
the welding trials
DuraScan-70 automatic hardness tester 2 mm below the surface with 0.5 mm
was composed of a K-TIG power supply, a data acquisition device and a high-speed camera. Theinterval under a load of
1welding
kg for 10 s dwell time. The preparation of tensile samples and the execution
output was provided by a special power source with capability up to 1000 A. A high-speedof tensile tests were
in accordance with ASTM
camera in combination E8/E8M-16a
with guideline. The
a 10% transparency tensile
neutral samples
filter were cut
was placed perpendicular
beside to the
the worktable in
welding
order todirection
record the using Wire Electrical
arc behaviour. Discharge
An inert gas backMachining.
purge unitTensile
was tests were to
designed performed in as-cut
provide shielding
condition without any finishing
from the atmospheric on theon
contamination surface of tensile
the root side ofsamples
the weld with 1 mm/min
pool. The argon crosshead speed
shielding gas and
was
25 mm gauge length (Figure 3). The thickness of all tensile test specimens is 6.2 mm.
transferred into the back purging unit from the left-hand side hose hole and allowed to be ejected A CCD camera
was used the
through to measure
switch on thethe
total elongationside
right-hand by capturing the movementof
to avoid accumulation ofpressure
two white spotsthe
inside with 25-mm
unit. The
interval on the tensile specimens. In order to ensure the reproducibility of tensile
image capturing was conducted after the process reached stable. During welding, the torch and results, three samples
were tested and
high-speed the average
camera were kept value was considered.
stationary while the worktable moved at the pre-set speed.

Figure2.2. Schematic
Figure Schematic of
of experimental
experimentalsystem.
system.

Table 3. Welding parameter combination.

Welding Current Arc Voltage Travel Speed Heat Input

Test No.
(A) (V) (cm/min) (kJ/cm)
1 450 16.41 34.2 7.78
2 465 16.52 34.2 8.1
3 480 16.69 34.2 8.42
4 510 17.11 34.2 9.18
5 450 16.59 30 8.96
6 450 16.44 26 10.22
7 450 16.57 24 11.16

Table 4. Fixed welding parameters (all experiments).

Process Parameters Details

Electrode material Lanthanated tungsten
Electrode diameter 6.4 mm
Electrode tip angle 45 degree
Shielding gas 99.95% Ar
Shielding gas flow rate 20 l/min
Back purging gas 99.95% Ar
Back purging gas flow rate 10 l/min
Post flow shielding time 10 s
Arc length 3 mm
Metals 2019, 9, 1327 5 of 15
Metals 2019, 9, x FOR PEER REVIEW 5 of 15

MetalsMetals
2019, 2019,
9, x FOR
9, x PEER REVIEW
FOR PEER REVIEW 5 of 15
5 of 15
MetalsMetals
2019, 2019, 9, x PEER
9, x FOR FOR PEER REVIEW
REVIEW 5 of 15
5 of 15
MetalsMetals
2019, 2019, 9, x PEER
9, x FOR FOR PEER REVIEW
REVIEW 5 of 15
5 of 15
MetalsMetals
2019, 2019,
9, x FOR
9, x PEER REVIEW
FOR PEER REVIEW 5 of 15
5 of 15

Figure3.3.Dimension
Figure Dimensionof
oftensile
tensilespecimen.
specimen.
3. Results and Discussion
3. Results and Discussion
Figure 3. Dimension
Figure of tensile
3. Dimension specimen.
of tensile specimen.
3.1. Surface Formation Figure Figure 3. Dimension
3. Dimension of tensile
of tensile specimen.
specimen.
Figure Figure 3. Dimension
3. Dimension of tensile
of tensile specimen.
specimen.
3.1. Surface Formation
3. Results and and
3. Results Discussion
Discussion Figure 3. Dimension
Figure of tensile
3. Dimension specimen.
of tensile specimen.
Tables 5 3.and 3.6Results
Results and and
illustrate Discussion
both
Discussion the face side and root side surface formation for the weldments with
3. Results
3. Results and Discussion
and Discussion
different
Tables 3. Results
parameters.
3.1.
53.1.
andSurface
3. and
It can
Results
3.1. Discussion
Formation
Surface and beDiscussion
Formation observed
both the that face when the travel speed was fixed at 34.2for cm/min, undercut
3.1.6Surface
Surface illustrateFormation
Formation side and root side surface formation the weldments
was experienced
with different 3.1. for
3.1. all
Surface
Surface the weldments,
Formation
Formation as marked by red arrows and shown in Table 5. When the
3.1.parameters.
Surface
Tables
3.1. Formation
Surface5 and
Tables
It
Formation
5 and
can
6 illustratebe observed
6 illustratebothboth the face that
sidewhen
the face and and
side
the
root travel
side
root surface
side
speed
surface
was for
formation
formation
fixedthe at 34.2 cm/min,
for weldments
the weldments
travel speed
undercut was
was with reduced
Tables Tables
experienced
different 5 to 30
and 5 and
6 cm/min,
6 illustrate
illustrate
for
parameters. all the both
It can the bothundercut
the
weldments,
be the
face
observed face
side asdefect
side
and and
root
marked disappeared,
root
side byside
surface
redsurface together
formation
formation
arrows for
and with for
the
shown smooth
the weldments
weldments surface5.
incm/min,
Table
with
Tables different
Tables
5 and 6parameters.
5 and 6 illustrate
illustrate It
both can
both
the be the
face sidethat
observed
face side
and when
thatandwhen
root the
sidetravel
root the
side
surface speed
travel
surface was
speed fixed
was
formation
formation at
forfixedfor
the 34.2
at cm/min,
the34.2weldments
weldments
formation with with different
different parameters.
parameters. It can It be
can be
observedobserved that that
when when
the the
travel travel
speed speed
was was
fixed fixed
at at
34.2 34.2
cm/min,cm/min,
When the on both
travel
undercut sides,
Tables
speed
undercut
withwith was 5 and
Tables
different
different as
was 5 6shown
and illustrate
experienced
was 6
reduced
experienced
parameters.
parameters. in
illustrate
for
It can Table
both
to
all
for
It be
the
both
30
the
all 6. A
face
the
cm/min, further
side
face
weldments,
the
canobserved
and
side
weldments,
be observed the
as
that that decrease
root
and
marked
as
whenwhen
side
root
undercut
marked by in
surface
side travel
defect
red
by
the travel
the travel arrows
red
speed speed
formation
surface formationfor
disappeared,
and
arrows
speed
was was led
the
shown
and
fixedfixed for to
shown
at 34.2 inwavy
weldments
the weldments
together
Table
at 34.2 in
cm/min, ripples
5.
Table
cm/min, with
5.
with
When undercut
undercut
withthe was
different
different
travelwas experienced
experienced
parameters.
parameters.
speed was Itforcan for
all
It
reduced bethe
can all the
be
to weldments,
weldments,
observedobserved
30 that
cm/min, as
when
that
theas
marked marked
the
when
undercut by red
travel
the by red
arrows
speed
travel
defect arrows
and and
wasdisappeared,
speed fixed
was
disappeared, atshown
shown
fixed 34.2in34.2
at
together inwith
Table
cm/min, Table
5. 5.
cm/min,
along the face
smooth surface sideWhenweld
formation
undercut
undercut the
was was travel
edges,
on speed
as
both
experienced
experienced was
highlighted
sides,
for reduced
forthe
all byto
astoweldments,
all shown
the 30
red cm/min,
circles,
in as
weldments, Table the undercut
although
6.undercut
as marked
marked A byfurther defect
undercut
by arrows
red decrease
red arrows and and was
shown in not together
travel
shown speed
in Table
in Table with
experienced.
5. 5. led
When When
undercut the
undercutwasthewas
travel travel
speed
experienced speed
wasfor
experienced was
reduced
allreduced
for the tocm/min,
30as
weldments,
alltothe 30 cm/min,
as the the
undercut
marked by defect
red defect
arrows disappeared,
disappeared,
and and shown together
together
inspeed
Table with5.with
Also, even
to wavy
smooth
When
irregular
ripples
surface
smooth
Whenthe
along the
travel
bead the
formation
surface
travel
speed
face
face waswason
formation
speed
side wasboth
on
reduced
produced
weld
sides,
both
reduced
edges, 30
ontoweldments,
sides, shown
cm/min,
the
as rootin
30ashighlighted
shown
cm/min,the as
Table
inthe
side
marked
6.
Table A 6.
undercut
undercut
when
by
by
further
Athe
defect
red
red
further
defectarrows
decrease
travel
circles,
decreasein
disappeared,
disappeared,speed
although
shown
travel
in travel
was
inspeed
together
together Table
led
with
reduced
undercut
5.
led
with to
When
to smooth
smooth the
When
wavy surface
the
ripplessurface
travel travel
along formation
formation
speed was
speed
the on
face on
both
reduced
was both
sides,
reduced
side sides,
to weld
weld as
30to as
shown
cm/min,
edges,30 shownin
cm/min,
as the in
Table Table
6.
undercut
the
highlighted A 6.
undercut
by A
furtherfurther
defect
red decrease
decrease
disappeared,
defect
circles, in
disappeared,
although in
travel travel
speed
together
together
undercut speed
led
with
was ledwas
with
smooth to wavy
smoothsurface ripples
surface
formationalong
formation the
on both face
on both side
sides,sides, edges,
as shown
as shown as highlighted
in Table
in Table 6. A 6. by red
A further
further circles, although
decrease
decrease in travel
in travel undercut
speed speed was
led led
24 cm/min in
not experienced. Test
to
smoothto
wavy 7
smoothsample.
wavy
ripples
Also,
surface
not experienced.
not
ripples
even
experienced.
along
formation
surface Also, along
the
irregular
formation
even
Also,
the
face
on face
side
bead
both
on
irregular
even
side
weld
sides,
both
irregular
weld
face
bead edges,
was
asbead
sides, edges,
shown
face as
asface
was as
inwas
shown highlighted
highlighted
producedTable 6. on
inproduced
produced Table by
A by
red
the
further
on6. A red
circles,
root
onfurther
the root
the
circles,
decrease
side
root
although
although
side when
inwhen
decrease
when
side travel
the undercut
undercut
thespeed
in travel
travel
the
was
travel was
speed
ledspeed
speed
speed
travel led
to wavy
to wavy ripples
ripples along
along the face
the face side side
weldweld edges,
edges, as highlighted
as highlighted bycircles,
by red red circles, although
although undercut
undercut was was
not not experienced.
experienced. Also,Also,
even even irregular
irregular beadbead
face face
was was produced
produced on on
by the the
root root
side side
when when
the the travel
travel speedspeed
was reducedtotowavy
was
not to
24
was
not
ripples
wavy to along
ripples
cm/min
reduced
reduced
experienced. 24in the
along
Test
cm/min
to
experienced. 24
Also,
face
the
cm/min
Also,
even even
side
face
Test
in weld
7side
7inirregular
sample.
Test edges,
7weld
sample.
irregular sample.
beadbead
face
as highlighted
edges,
face
was was produced
produced
red
as highlighted on
on the
circles,
byroot
redroot
the side
although
circles,
side
when when
undercut
although was
theundercut
travel
the travel speed was
speed
not was reduced
wasexperienced.
reduced to 24
Tabletocm/min
24
Also, cm/min
even
Weld
5.Also, in in Test
Test
irregular
surface 7 sample.
7 sample.
bead face face
formation was produced
with on the
different onroot
arc side side
current. when the travel speed
was not
was experienced.
reduced
reduced to 24tocm/min
24 cm/mineven
in inirregular
Test Test bead
7 sample.
7 sample. was produced the root when the travel speed
was was
reduced to 24tocm/min
reduced inTable
24 cm/min
Table Test
5. in 7 sample.
Test
Weld
5. 7 sample.
surface
Weld formation
surface with with
formation different arc current.
different arc current.
Defect Table Fixed
5. Table
Weld 5.surface
Table 5. Weld
Weld
parameters: formation
surface
surface
Speedformationwith
= formation
34.2 withdifferent
with
different
cm/min, Arc arccurrent.
different
arc
length current.
arc=current.
3 mm Variable
Test Table Table
5. Weld5. Weld surface
surface formation
formation with
withcm/min,different
different arc= current.
arc current.
Type Fixed parameters:
Fixed parameters: Speed =
Speed 34.2
= cm/min,
34.2 Arc length
Arc length3 mm= 3 mm
No. Table
FixedFixed 5.
TableWeld5. surface
Weld
parameters:
parameters: Speedformation
surface
Speed
= 34.2 with
formation
= 34.2 different
with
cm/min,
cm/min, arc
different
Arc
Arc length current.
arc
length current.
= mm
= 3= mm 3 mm
Fixed parameters:
Face
Fixed
Face
side
Face
Fixed
Speed
side
side
parameters:
parameters:
= Speed
Speed
34.2 cm/min,
= 34.2= 34.2
Arc
cm/min,
cm/min,
length
Root
Root
Arc
Arc length
side
side
length
Root
3side
= 3 mm= 3 mm
Current
Face
Fixed Face
side
Fixed side
parameters:
parameters: Speed = 34.2
Speed = cm/min,
34.2 Arc length
cm/min, Root
Arc = 3 mm
Root
side
length side
= 3 mm
Face side FaceFace
side side Root
Root Rootside
side side
FaceFace
side side RootRoot
side side
1 Undercut 450 A

2 Undercut 465 A

3 Undercut 480 A

4 Undercut 510 A
Metals 2019, 9, 1327 6 of 15

Metals
Metals 2019,
2019, 9,
9, xx FOR
FOR PEER
PEER REVIEW
REVIEW 66 of
of 15
15
Metals 2019, 9,Table
x FOR PEER REVIEW
6. Weld surface formation with different travel speed. 6 of 15
Metals
Metals 2019,
2019, 9,
9, xx FOR
FOR PEER
PEER REVIEW
REVIEW 66 of
of 15
15
Table
Table 6.
6. Weld
Weld surface
surface formation
formation with
with different
different travel
travel speed.
speed.
Test Defect FixedTable
Table
6. Weld surface
parameters:
Table 6.
6. Weld
Weld surface
surface
formation
Arc length = 3 mm,
formation
formation
withWelding
with
different current
different
travel speed.
different travel
with Welding travel = 450 A
speed.
speed. Variable
Fixed parameters: Arc length = 3 mm, current = 450 A
No. Type Fixed parameters: Arc length = 3 mm, Welding current
Fixed Face
FixedFace
side
parameters:
side
parameters: Arc
Arc length
length == 33 mm,
mm, Welding
Welding
Root side=== 450
side
current
Root
current 450 A
450 A
A
Speed
Face
Face side
side Root
Root side
side
Face side Root side

5 N/A 30cm/min

6 Ripple 26 cm/min

7 Ripple 24 mm/min

It is known that high welding speed is easy to cause undercut formation. This is because the
It
It ismetal
known that high much welding speed is easy to cause undercut formation. This is because the
molten It is
is known
known that
that high
experiences high welding
welding speed
stronger is
is easy
speed backwards easy to to cause
momentum
cause undercut
undercut formation.
at higher
formation. travel This
This is
is because
speed. The mass
because the
the
molten
molten
flow ratemetal
metal
is the experiences
experiences
highest along much
much the stronger
stronger
centreline backwards
backwards
of the weld, momentum
momentum
which tendsat
at higher
higher
to draw travel
travel
molten speed.
speed.
metal The
The mass
mass
towards
It is known moltenthatmetal highexperiences
welding much speedstronger is easybackwards to causemomentum undercutatformation. higher travelThis speed.isThe becausemass the
flow
flow rate
rate is the highest along the centreline lackof the weld, on which tends to draw molten metal towards
molten metal thethe rate is
flowcentreline
experiences is the
theof highest
the flow
highest
much along
along
strongerand the
the centreline
induces
centreline
backwards of
ofofthe
the weld,
meltweld,
momentum which
the side
which tends
walls
tends
atwalls to
to draw
higher on
draw the molten
face side
molten
travel metal
metal
speed. towards
[21]. Thus,mass
towards
The
the centreline
centreline
undercut of
of the
was consistentlythe flow
flow and
and
observed induces
induces when lack
lack highof
of melt
melt
travelon onspeedthe
the side
side
was walls
used. on
on the
the
Considering face
face side
side [21].
that [21].
the TestsThus,
Thus, 4
the centreline
flow rate is theundercut
highest along of the flow
the centreline and induces of thelack of
weld, melt on
which the side
tends walls
to draw on the face
molten side [21].
metal Thus,
towards
undercut was
was consistently observed when high travel speed
travel was used.
the Considering that the Tests 44
and
undercut5 sampleswas consistently
possess similar
consistently observed
heat input
observed when
when andhigh
high travel
travel speed
different speed was
was used.
speed, used. Considering
presence
Considering that
that the
of undercut the Tests
in Test
Tests 4
the centreline and
andofmay
4and 55 samples
the flowpossess
5 samples
further
samples
and induces
possess
indicate
possess
similar
similar
that travel
similar
heat
lack
heat input
inputof and
input and
heat speed
melt
plays
and
different
aon
different
different partthe travel
travel
in side
the
travel
speed,walls
speed,
formation
speed,
the
the presence
on
the presence
the face
of undercut.
presence
of
of undercut
of undercut
Inside
undercut addition,in
[21].
in Test
in Test
TestasThus,
undercut was 44reported
may further
4consistently
may
may further by Tomsic
further
indicate
observed
indicate
indicate andthat
that
Jackson
that
travel
when
travel
travel
speed
high
speed
[22],
speedhighplays heat aaainput
travel
plays
plays
part
speed
part
part
in
in
in the
makes the formation
was
the used.
formation
the keyhole
formation
of
of undercut.
Considering
undercut.
ofwiden,
undercut. moreIn In addition,
that
addition,
Inmolten
addition,the
metal as
Tests
as
as 4
reported
reported by Tomsic and Jackson [22], high heat input makes the keyhole widen, more molten metal
possessby
displaced
and 5 samplesdisplaced
reported byby Tomsic
the keyhole
similar
by
Tomsic
the
and
and Jackson
heat
keyhole
flows
input
Jackson
flows
[22],
backand
[22], tohigh
the heat
heat input
trailing
different
high region
travel
input makes
makes for
speed,athe
the keyhole
longer the
keyhole widen,
distance
presence
widen, and more
more molten
aofuniform
moltenbead
undercut metal
metal is Test
in
displaced
hard
displaced by
to obtain. the
theThis
bythat keyhole
keyhole could flows
flows be back
back
the reason
back
to
to the trailing
to the for the region
theatrailing
trailing region
formation
region
for
for
for
aaof longer
wavydistance
a longer
longer ripples and
distance
distance and
on the
and
aa uniform
a uniform
face side
uniform
bead
bead
bead andis
is
is as
4 may furtherhard indicate
hard to
to obtain.
obtain.
travel
This
This could
could
speedbe
be the
the
plays
reason
reason
part
for
for the
the
in the
formation
formation
formationof
of wavy
wavy
of undercut.
ripples
ripples on
on the
the
In
face
face
addition,
side
side and
and
irregular bead face on the root side. As the mitigation
hard to obtain. This could be the reason for the formation of wavy ripples on the face side and of undercut defect was observed with
irregular
reported by Tomsic
irregular and bead Jacksonface on[22],the root
highside. heatAs toinputthe mitigation
makes
that thethe of undercut
keyhole defect
playswas
widen, more observed molten inwith
themetal
irregular bead
increasing arc
bead face
face on
current, on itthe
the root
root side.
is reasonable side. As As the
the mitigation
believe mitigation of
of undercut
heat input also
undercut defect
defect was
wascertain observed
observedrole with
with
displaced by theincreasing
keyhole arc current,
flows it is reasonable to believe that the heat input also plays certain role in the
increasing
formation
increasing arc
of current,back
current,
arcundercut it is
itinisthe to K-TIG
the trailing
reasonable
reasonable to
weldingbelieve
to believe region that for
that
process. the
theThe aheat
heat longer
input
effect
input ofdistance
also inputand
plays
heatplays
also certain
on the
certain a uniform
role in thebead
in
undercut
role the
formation
formation of
of undercut
undercut in
in the
the K-TIG
K-TIG welding
welding process.
process. The effect of heat input on the undercut
is hard to obtain.formation
formation This is
of discussed
couldundercut be ininfollowing
the reason
the K-TIG section.
for the
welding process. The
formation Theofeffect wavy
effect of heat
heat input
of ripples input on on
onthe the
thefaceundercut
undercut side and
formation
formation is
is discussed in following section.
formation
irregular bead 3.2.
faceWeld is discussed
the discussed
on Longitudinalroot side.
in
in following
following
As the
section.
section.
mitigation of undercut defect was observed with increasing
Cross-Section
arc current, it 3.2.
is
3.2. Weld
reasonable
Weld
3.2. Weld
Longitudinal
Longitudinal to Cross-Section
believe
Cross-Section that the heat input also plays certain role in the formation of
The Longitudinal
weld longitudinal Cross-Section cross-section and arc behaviour are shown in Figure 4. After arc was
undercut in theextinguished,
K-TIGThe
The welding
weld process.
longitudinal The effectand
cross-section of heatarc input
behaviour onare theshown
undercut
inside in
formation
theFigure 4. After is
arc discussed
was
The weldweld the longitudinal
weld poolcross-section
longitudinal was suddenly
cross-section and arc
arc behaviour
and frozen. The profile
behaviour are
are shown
shown in
in Figure
keyhole
Figure 4.
4. After
channel
After arc
arc waswas
was
extinguished,
in following section.
extinguished,
retained. Whenthe the weld
weld
lower heatpool
pool input was
was suddenly
suddenly
used, asfrozen.frozen.
in Test The The profile
profile inside
inside the
arcthe keyhole
keyhole channel
waschannel was
was
extinguished, the weld pool waswas suddenly frozen. 1 sample,
The profilestrong inside plasma
the keyhole experienced
channel was
retained.
retained. When
Whenside,lowerlower heat
heat input
input was
was used,
used, as
as in Test 11 sample,
relativestrong to thearc plasma was experienced
on the face
retained. When together
lower heat withinput verywas high
used, as in
deflection in Test
Testangle1 sample,
sample, strong
strong arc plasma
horizontal
arc plasma line, was
was experienced
as shown in
experienced
on
3.2. Weld Longitudinal
on the face side,
Cross-Section together with very high deflection angle relative to the horizontal line, asdecreased
shown in
on the face side, together with very high deflection angle relative to the horizontal line, as
Figurethe face
4a. side,
As thetogether
heat with
input very high
increased, deflection
the amount angle of relative
arc to
plasma the horizontal
on the face line,
side as shown
shown in in
Figure
Figure
significantly,4a.
4a. As
As the
the
along heat
heat
with input
input
the increased,
increased,
increase in thethe
the amount
amount
slope of of
of
keyhole arc
arc plasma
plasma
leading on
on
wall, the
the
as face
face
shown side
side in decreased
decreased
Figure 4b.
Figure 4a. As the heat input increased, the amount of arc plasma on the face side decreased
The weldsignificantly, is more,along
longitudinal
significantly, with
thecross-section
along with the increase
angleand ofin the
arc slope
behaviour of keyhole areleading shown wall, as shown
in Figure 4.in Figurewith
After 4b. was
arc
What
significantly, along with the
deflection the increase
increase in
in the
the slope
plasma slopecloud of keyhole
of relative
keyhole to leading wall,
wall, as
the horizontal
leading as shown
line decreased
shown in
in Figure
Figure 4b. 4b.
What
extinguished, the
What weld
increasing is
is more,
pool
more,
heat the the
thewas
input.deflection
suddenly
deflection angle
angle of
frozen.
of plasma
plasma cloud
Thecloud
cloudprofilerelative
relative inside to
to the horizontal
thehorizontal
the keyhole line
horizontal line
channel
line decreased
decreased was with with
retained.
with
What is more, deflection angle of plasma relative to the decreased
increasing
increasing
As can heat
be input.
heat input.
seen from the schematic in Figure 4, when lower heatwas inputexperienced
was used, there was
When lower heat input
increasing
As can
was
heat
be
used,
input.
seen from
asthe in schematic
Test 1 sample, in Figure
strong
4, when
arclower plasma heat input was used, there
onwas the face
As
insufficient
As can be
heat
can high
be seenseen from
deposited the
on
from the schematic schematic
the lower in
half Figure
of
in Figure the 4, when
keyhole
4, when lower
leading
lower heat heat
wall. input
Thus,
input wasthe
was used, used,
keyhole there was
leading
there was
side, together with
insufficientvery heat deflection angle relative to the horizontal line, as shown in Figure 4a. As
insufficient
wall exhibited
insufficient heat
heat a deposited
deposited
severely
depositedcurved
on
on
on the
theshape.
lower Only
the lower
lower half
half of
half of of
the
the
small keyhole
amount leading
the keyhole
keyhole leading
of the arc
leading
wall.
wall.
wall.
Thus,
Thus,
plasma
Thus,can
the
the keyhole
ejectedleading
thebekeyhole
keyhole leading
via the
leading
the heat input wall
increased,
wall exit, the
exhibited a amount
severely ofofarc
curved plasma
shape. Only onsmall the face
amount side
of decreased
the arc plasma significantly,
can be ejected along
via the with
wall exhibited
keyhole exhibited aa severely
leaving majority
severely curved
curved shape.
them
shape. Only
guided small
small amount
Only towards the face
amount of the
the arc
of side. Asplasma
arc the heat
plasma can
input
can be
be wasejected
ejected via
via the
increased, the
the increase inkeyhole
the
keyhole
more slope
heatexit,
exit,
was leaving
of keyhole
leaving
deposited majority
majority leading
on of
of
the them
them
keyhole guided
wall,
guided as
bottom towards
shown
towards region.
keyhole exit, leaving majority of them guided towards the face side. As the heat input was increased,
the
in
the Theface
Figure
face side.
side.
bottom As
4b.
As the
What
the
region heat
heat
(A input
is more,
input
zone) was
was
of theincreased,
the deflection
increased,
keyhole
more
more heat
heat was
was indeposited
deposited 4aon on the
the keyhole
keyhole bottom
bottom region.
liquid The
region. The bottom
bottom region
region (A
(A zone)
zone) of
of the keyhole
angle of plasma leading
more cloud heatwall
relative
was Figure
depositedto the onwas themelted
horizontal
keyhole away
line
bottom into
decreased
region. and
The with displaced
bottom increasing to the
region (Arear
heat
zone) partof the
input.the keyhole
to form
keyhole the
leading
leading
weld, wall
wall
which in
in
makes Figure
Figurethe 4a4a
keyholewas
was melted
melted
exit deviate away
away into
into
less liquid
liquid
from theand and
welding displaced
displaced
torch.to to the
the rear
rear
Therefore, part
part
larger to
to form
form
amount the
the
of
leading
As can beweld, seen wall
from in Figure
the the 4a
schematic was melted
indeviate away
Figure into
4,from liquid
when and displaced
lower heat to the
input rear part
was used, to form therethe was
weld,
the
weld, arcwhich
which
plasma
which
makes
makes
makes could the
thebekeyhole
keyhole
easily ejected
keyhole
exit
exit
exit deviatethrough
deviate
less
less
lessthe from
from
the
the
keyhole welding
the welding
torch.
torch.
exit, leaving
welding
Therefore,
very smalllarger
torch. Therefore,
Therefore, larger
amount
larger
amount
amount
of them
amount
of
of
of
insufficient heat
the
the deposited
arc plasma on
could the
be lower
easily half
ejected of the
through keyhole
the keyhole leadingexit, wall.
leaving Thus,
very small the keyhole
amount of leading
them
the arc
guided arc plasma
towardscould
plasma the face
could be
be easily
side. In
easily ejected
addition,
ejected through
through the
the keyhole
the increase keyhole in heat exit,
exit,inputleaving
leaving very
further small
small amount
very decreased amount of
of them
the slope them of
wall exhibitedguided
a severely
guided
keyhole
guided
towards
towards
rear edge,
towards curvedthe
the face
face
thewhich
side.
faceshape.
side.
allowed
side.
In
In addition,
InOnly
addition,
the plasma
addition, small the
the
thecloud
increase
amount
increase
increase to be in
in ofheat
heattheinput
indeflected
heat arc
input
input at further
plasma
further
afurther decreased
can
decreased
very small
decreased be the
angle. the
the
slope
ejectedslope via
slope of
of the
of
keyhole
keyhole rear edge, which allowed the plasma cloud to be deflected at a very small angle.
keyhole rear
keyhole exit, leaving edge,
majority
rear edge, which of them
which allowed
allowedguided the
the plasma
towards
plasma cloud
cloud theto be
be deflected
to face side. As
deflected at
at aathevery small
veryheat
smallinput angle.
angle. was increased,
more heat was deposited on the keyhole bottom region. The bottom region (A zone) of the keyhole
leading wall in Figure 4a was melted away into liquid and displaced to the rear part to form the weld,
which makes the keyhole exit deviate less from the welding torch. Therefore, larger amount of the arc
plasma could be easily ejected through the keyhole exit, leaving very small amount of them guided
towards the face side. In addition, the increase in heat input further decreased the slope of keyhole
rear edge, which allowed the plasma cloud to be deflected at a very small angle.
 Metals 2019, 9, 1327 7 of 15

Metals 2019, 9, x FOR PEER REVIEW 7 of 15

(a) (b)
Figure 4. Weld
Figure longitudinal
4. Weld cross
longitudinal section
cross andand
section arcarc
shape; (a) (a)
shape; Test 1; (b)
Test Test
1; (b) 7. 7.
Test

3.3.3.3. Weld
Weld Transverse
Transverse Cross-Section
Cross-Section
The
The weld
weld macrographsare
macrographs aredepicted
depictedininFigure
Figure 5.
5. ItIt can
can be
be seen
seen that
thatboth
boththe
theface
faceside
sideand root
and rootside
width increased with increasing heat input. More importantly, there existed slight
side width increased with increasing heat input. More importantly, there existed slight hump in hump in Tests 1 and
2 samples, as shown in Figure 5a,b. The hump decreased and even disappeared
Tests 1 and 2 samples, as shown in Figure 5a,b. The hump decreased and even disappeared with with increasing heat
input, asheat
increasing shown in Figure
input, 5c,d. in
as shown This is consistent
Figure with
5c,d. This the mitigation
is consistent withor elimination
the mitigationoforundercut defect.
elimination
As mentioned
of undercut defect. above, the heat input is also believed to be a controlling factor in the formation of
undercut defect. Also, the variation trend of heat input, hump and undercut formation are consistent.
Thus, it is necessary to consider the interaction among them. As mentioned above, when a high heat
input was used, the keyhole front wall was less curved, majority of the plasma gas jet spurted out
via the keyhole exit and very small amount of them was reflected by the keyhole front wall towards
the rear keyhole wall, as indicated in Figure 6a. Therefore, the face side weld formation was smooth
with the absence of humping and evident undercut. While if the heat input was reduced, the keyhole
wall on the front side bent against the travel direction and the distance between the welding torch and
the keyhole exit would increase, as demonstrated previously in Figure 4a. In this case, the plasma
Metals 2019, 9, x FOR PEER REVIEW 8 of 15
Metals 2019, 9, 1327 8 of 15
(a) (b)
Figure 4.dramatically
gas flow field changed Weld longitudinal cross section
as depicted and arc
in Figure 6b. shape;
Almost(a)all
Test
of1;the
(b)plasma
Test 7. jet was guided
towards the lower region of the keyhole wall on the backside. As the welding current used in K-TIG
3.3. Weldprocess
welding Transverse
wasCross-Section
relatively high, the pressure associated with the plasma jet flow could overwhelm
the surface
The weld macrographsfrom
tension coming are the weld pool
depicted on the5.backside.
in Figure It can be Therefore, the trailing
seen that both the faceweld
side pool was
and root
severely deformed. Certain amount of the weld pool was pushed towards
side width increased with increasing heat input. More importantly, there existed slight hump in the face side, and hump
appeared
TestsFigurein5.the
1 and weld centre.
2Macrographs
samples, Thistypical
asofshown
four kind ofwelds
hump
in Figure would
5a,b.
with The reduce
hump
different heatthe amount
decreased
input; of 1liquid
and
(a) Test even metal on
(b)the
disappeared
(7.78 kJ/cm); Testedges
with
of the weld
increasing pool on
heat input,
2 (8.1 kJ/cm); the face side,
(c) Testas4 shown which may
in Figure
(9.18 kJ/cm); facilitate
5c,d.
(d) Test the
ThiskJ/cm).
7 (11.16 formation of undercut. This implies
is consistent with the mitigation or elimination that the
heat input level
of undercut is also an important parameter in determining the undercut formation.
defect.
As mentioned above, the heat input is also believed to be a controlling factor in the formation of
undercut defect. Also, the variation trend of heat input, hump and undercut formation are
consistent. Thus, it is necessary to consider the interaction among them. As mentioned above, when
a high heat input was used, the keyhole front wall was less curved, majority of the plasma gas jet
spurted out via the keyhole exit and very small amount of them was reflected by the keyhole front
wall towards the rear keyhole wall, as indicated in Figure 6a. Therefore, the face side weld formation
was smooth
Metals withPEER
2019, 9, x FOR absence of humping and evident undercut. While if the heat input8 of
theREVIEW was
15
reduced, the keyhole wall on the front side bent against the travel direction and the distance
between the welding torch and the keyhole exit would increase, as demonstrated previously in
Figure 4a. In this case, the plasma gas flow field changed dramatically as depicted in Figure 6b.
Almost all of the plasma jet was guided towards the lower region of the keyhole wall on the
backside. As the welding current used in K-TIG welding process was relatively high, the pressure
associated with the plasma jet flow could overwhelm the surface tension coming from the weld pool
on the backside. Therefore, the trailing weld pool was severely deformed. Certain amount of
the weld
Figure
Figure 5.5. Macrographs
Macrographs of of four
four typical welds with
typical welds with different
differentheat
heatinput;
input;(a)
(a)Test
Test11(7.78
(7.78kJ/cm);
kJ/cm);(b)
(b)Test
Test2
2(8.1
(8.1kJ/cm);
kJ/cm); (c) Test 4 (9.18 kJ/cm); (d) Test 7 (11.16 kJ/cm).
(c) Test 4 (9.18 kJ/cm); (d) Test 7 (11.16 kJ/cm).

As mentioned above, the heat input is also believed to be a controlling factor in the formation of
undercut defect. Also, the variation trend of heat input, hump and undercut formation are
consistent. Thus, it is necessary to consider the interaction among them. As mentioned above, when
a high heat input was used, the keyhole front wall was less curved, majority of the plasma gas jet
spurted out via the keyhole exit and very small amount of them was reflected by the keyhole front
wall towards the rear keyhole wall, as indicated in Figure 6a. Therefore, the face side weld formation
was smooth with the absence of humping and evident undercut. While if the heat input was
reduced, the keyhole wall on the front side bent against the travel direction and the distance
between the welding torch and the keyhole exit would increase, as demonstrated previously in
Figure 4a. In this case, the plasma gas flow field changed dramatically as depicted in Figure 6b.
Almost all of the plasma jet was guided towards the lower region of the keyhole wall on the
backside. As the welding current used in K-TIG welding process was relatively high, the pressure
associated with the plasma jet flow could overwhelm the surface tension coming from the weld pool
on the backside. Therefore, the trailing weld pool was severely deformed. Certain amount of the
weld pool was pushed towards the face side, and (a)hump appeared in the weld centre. This kind of
hump would reduce the amount of liquid metal on the edges of the weld pool on the face side, which
Figure 6. Cont.
may facilitate the formation of undercut. This implies that the heat input level is also an important
parameter in determining the undercut formation.
Metals 2019, 9, 1327 9 of 15
Metals 2019, 9, x FOR PEER REVIEW 9 of 15

(b)
Figure 6. Schematic of plasma gas flow inside keyhole; (a) high heat input; (b) low heat input.
input.

3.4. Weld Microstructure

The weld microstructure with different heat input is depicted in Figure 7. The weld metal in
Test 1 consists of martensite and small amount of bainite, whereas only bainite with larger grain size
is observed in the weld metal of Test 7. The same is true for the microstructure in the coarse grain
heat-affected zone (CGHAZ) where a combination of martensite with bainite is present in Test 1 and
only bainitic microstructure is found in Test 7. When it comes to fine grain heat-affected zone (FGHAZ),
the microstructure consists predominantly of martensite. However, it is evident that the fraction of
bainite in FGHAZ of Test 7 is higher than that of Test 1.
It is known that microstructure is closely linked to welding thermal cycle and hardenability.
During welding thermal cycle, CGHAZ and FGHAZ were heated to austenite phase region. Upon
cooling, the reformed austenite would transform to martensite, bainite or ferrite, depending on the
cooling rate and hardenability of the base metal. Since the base metal contains large amount of carbon
and chromium, the hardenability is relatively high, which favours the formation of martensite and/or
bainite instead of ferrite. As the heat input increases, the cooling rate decreases, which favours the
formation of bainite rather than martensite. The same is true for the weld metal as it still experiences
phase transformation from austenite to various transformation products upon cooling, although it was
melted during welding thermal cycle. That is why higher fraction of martensite was formed in the
weld and HAZ with lower heat input.

3.5. Hardness
The hardness distribution of various welds is depicted in Figure 8. Lower hardness is observed
in the weld metal region. The hardness increases when moving from the weld metal to the CGHAZ
and FGHAZ. After that, the hardness decreases again to the lowest point when it comes to the
over-tempered region. Finally, the hardness would increase continuously until it reaches the hardness
value of the base metal (495 Hv). The hardness variation trend is very similar to that found in the
armour steel weld produced with conventional fusion welding, as shown by Reddy et al. [23]. In
addition, the overall hardness decreases with increasing heat input. Once the heat input reaches
10.22 kJ/cm and beyond, the overall hardness decrease becomes evident.
Test 1 consists of martensite and small amount of bainite, whereas only bainite with larger grain size
is observed in the weld metal of Test 7. The same is true for the microstructure in the coarse grain
heat-affected zone (CGHAZ) where a combination of martensite with bainite is present in Test 1 and
only bainitic microstructure is found in Test 7. When it comes to fine grain heat-affected zone
Metals 2019, 9,the
(FGHAZ), 1327microstructure consists predominantly of martensite. However, it is evident that
10 ofthe
15
fraction of bainite in FGHAZ of Test 7 is higher than that of Test 1.

Metals 2019, 9, x FOR PEER REVIEW 11 of 15

3.5. Hardness
The hardness distribution of various welds is depicted in Figure 8. Lower hardness is observed
in the weld metal region. The hardness increases when moving from the weld metal to the CGHAZ
and FGHAZ. After that, the hardness decreases again to the lowest point when it comes to the
over-tempered region. Finally, the hardness would increase continuously until it reaches the
hardness value of the base metal (495 Hv). The hardness variation trend is very similar to that found
(a) (b)
in the armour steel weld produced with conventional fusion welding, as shown by Reddy et al. [23].
In addition, Microstructure
Figure 7.theMicrostructure ininweld
overall hardness metal
metaland
decreases
weld andheat-affected
with zone
increasing
heat-affected (HAZ);
heat
zone (a)(a)
input.
(HAZ); Test
Once 1 (HI
Test the
1 (HI = 7.78
=heat
7.78 kJ/cm);
input (b)
reaches
kJ/cm);
Test
(b) 7 (HI
Test 7 =
(HI =
11.16 kJ/cm)
11.16 Note:
kJ/cm) M
Note: and
M B
and represent
B martensite
represent and
martensite
10.22 kJ/cm and beyond, the overall hardness decrease becomes evident. bainite
and respectively.
bainite respectively.

It is known that microstructure is closely linked to welding thermal cycle and hardenability.
During welding thermal cycle, CGHAZ and FGHAZ were heated to austenite phase region. Upon
cooling, the reformed austenite would transform to martensite, bainite or ferrite, depending on the
cooling rate and hardenability of the base metal. Since the base metal contains large amount of
carbon and chromium, the hardenability is relatively high, which favours the formation of
martensite and/or bainite instead of ferrite. As the heat input increases, the cooling rate decreases,
which favours the formation of bainite rather than martensite. The same is true for the weld metal as
it still experiences phase transformation from austenite to various transformation products upon
cooling, although it was melted during welding thermal cycle. That is why higher fraction of
martensite was formed in the weld and HAZ with lower heat input.

Figure 8. Hardness
Figure distribution
8. Hardness distributionacross weldwith
across weld withvarying
varying heat
heat input
input.

The variation
The variation trend
trend of
of hardness
hardness across
across the
the weld
weld isis in
in good
good agreement
agreement with
with the
the microstructure
microstructure
change. As martensite and smaller grain size favours higher hardness, the increase
change. As martensite and smaller grain size favours higher hardness, the increase in hardness in hardnessfrom
from
weld metal to
weld toFGHAZ
FGHAZisisunexpected
unexpected because
because of of
thethe
increased
increased fraction of martensite
fraction and decreased
of martensite grain
and decreased
size from
grain size weld
from metal
weld to FGHAZ,
metal as shown
to FGHAZ, as inshown
Figurein7. Figure
In the over-tempered region, theregion,
7. In the over-tempered martensite
the
martensite was heavily tempered, which led to the reduction in dislocation density, precipitation of
carbides and reduction in solid solution strengthening in the matrix. Thus, the hardness in the
over-tempered region decreased dramatically compared with the HHA base metal (495Hv).
When the heat input was increased, the fraction of martensite decreased in both the weld metal
and HAZ, as shown in Figure 7, because of the slower cooling rate induced from higher heat input.
Metals 2019, 9, 1327 11 of 15

was heavily tempered, which led to the reduction in dislocation density, precipitation of carbides and
reduction in solid solution strengthening in the matrix. Thus, the hardness in the over-tempered region
decreased dramatically compared with the HHA base metal (495Hv).
When the heat input was increased, the fraction of martensite decreased in both the weld
metal and HAZ, as shown in Figure 7, because of the slower cooling rate induced from higher heat
input. Furthermore, the softening behaviour in the over-tempered region became severer because of
prolonged holding time in this region with increasing heat input, leading to more significant reduction
in dislocation density, precipitation of carbides and reduction in solid solution strengthening. That is
why the overall hardness of the welded joints decreased with increasing heat input.

3.6. Tensile Properties

The tensile results of the welded joints with different heat input are shown in Figure 9, while
the fracture location of each welded joint is shown in Figure 10. Two tensile samples for each setting
were used to present the fracture location on both face and root side, with the dotted yellow lines
representing the weld metal region on both face and root sides. Figure 9b shows that the elongation
of Test 1 (6.47%) and Test 3 (7.65%) is much lower than the other four samples although the ultimate
tensile strength (UTS) remains at a high level. The fracture behaviour of these two samples is depicted
in Figure 10a,b respectively. They both fractured from the face side undercut. Once the heat input was
increased to more than 8.96 kJ/cm, as is the case in Tests 4–7 samples, the elongation reaches more
than 10%. Although the Tests 4–7 tensile samples all fractured in the weld metal, the Tests 4 and 5
samples fractured in the weld metal from the root side fusion boundary (Figure 10c,d), while the
fracture location of Tests 6 and 7 samples were randomly located in the weld metal (Figure 10e,f).12Itofis15
Metals 2019, 9, x FOR PEER REVIEW
worth noting here that the wavy ripples seem to have no evident effect on the tensile properties in
Tests
10 6 and 7 samples.

Figure
Figure 9. 9. Tensile
Tensile properties;
properties; (a)(a) tensile
tensile curves
curves forfor welded
welded joints
joints with
with different
different heat
heat input;
input; (b)(b) tensile
tensile
results for welded joints with different heat
results for welded joints with different heat input.input.

It is well accepted that the tensile strength is closely linked to the overall hardness of the welded
joint. With the increase in heat input from 8.96 kJ/cm to 11.16 kJ/cm, the overall hardness decreases,
as shown in Figure 8, which is probably a result of the change in cooling rate and microstructure. In
addition, higher heat input also tends to produce wider weld metal and HAZ, which could increase
the width of reduced hardness region, as is the case in this study. Both these two factors can contribute
to the decrease in tensile strength with increasing heat input. The significant reduction in elongation in
Tests 1 and 3 is a result of undercut defect, which is a well-known stress concentration point. This can
also be confirmed by fractography, as shown in Figure 11a,c and e. Note that the left-hand side on
the fracture surface is the face side of the tensile sample, whereas the right-hand side is the root side
of the tensile sample. The fracture surface in Test 1 sample consists predominantly of delamination
both across two planes and on a single plane, as indicated by yellow and red arrows, respectively in
Metals 2019, 9, 1327 12 of 15

Figure 11a, as well as small amount of radial zone on the left. This means that after quick fracture
initiation and propagation, the tensile sample experienced delamination fracture mode along the
vertical direction down to the root side, as shown in Figure 11b. The recovery of elongation in Tests 4–7
samples
Metals 2019,indicates thatREVIEW
9, x FOR PEER the mitigation or elimination of undercut effect is an effective way to improve 12 of 15
the tensile properties by reducing or eliminating stress concentration. However, it is worth mentioning
10
that although the tensile properties of Test 4 were not affected by slight undercut in this study, it is still
highly desirable to eliminate it in real fabrication. Even in the absence of undercut, the fracture still
presents two different routes. As can be seen in Figure 11c, after crack initiated from the root side fusion
boundary (dark grey area on the right-hand side), quick crack propagation occurred until reaching
the face side, as evidenced by the completely radial zone in the fracture surface. The schematic is
shown in Figure 11d. This indicates that the fusion boundary on the root side can also act as a stress
concentration point in the absence of appreciable backside reinforcement, although it is less harmful
than undercut. Thus, it is recommended that post-weld surface treatment be carried out for the K-TIG
welded joint if the workpiece root side is accessible. With the formation of appreciable reinforcement
on the root side, as is the case in Test 6, evident fibrous zone is present in the centre region, as indicated
by the dotted yellow circle in Figure 11e. This is a result of plastic deformation occurring before crack
initiated. In addition, there is larger amount of shear lip formed on the fracture surface, indicating
that this sample followed a regular fracture route in which the stress condition changed to shear from
tension and resulted in a fracture around 45 degree relative to the maximum stress direction after crack
Figure 9. Tensile properties; (a) tensile curves for welded joints with different heat input; (b) tensile
reached the edge. The schematic of fracture path is shown in Figure 11f.
results for welded joints with different heat input.

Figure 10. Fracture

Figure 10. Fracturebehaviour
behaviourofofwelded
welded joints;
joints; (a) (a)
TestTest 1=
1 (HI (HI = 7.78
7.78 kJ/cm);
kJ/cm); (b)3 Test
(b) Test (HI =3 8.42
(HI kJ/cm);
= 8.42
kJ/cm);
(c) Test (c) Test
5 (HI = 58.96
(HIkJ/cm);
= 8.96 kJ/cm);
(d) Test(d) Test=49.18
4 (HI (HI kJ/cm);
= 9.18 kJ/cm);
(e) Test(e) Test=6 10.22
6 (HI (HI = kJ/cm);
10.22 kJ/cm);
(f) Test(f)7 Test
(HI =7
(HI = 11.16
11.16 kJ/cm).kJ/cm).

It is well accepted that the tensile strength is closely linked to the overall hardness of the welded
joint. With the increase in heat input from 8.96 kJ/cm to 11.16 kJ/cm, the overall hardness decreases,
as shown in Figure 8, which is probably a result of the change in cooling rate and microstructure. In
addition, higher heat input also tends to produce wider weld metal and HAZ, which could increase
the width of reduced hardness region, as is the case in this study. Both these two factors can
as is the case in Test 6, evident fibrous zone is present in the centre region, as indicated by the dotted
yellow circle in Figure 11e. This is a result of plastic deformation occurring before crack initiated. In
addition, there is larger amount of shear lip formed on the fracture surface, indicating that this
sample followed a regular fracture route in which the stress condition changed to shear from tension
Metals 2019, 9, 1327
and resulted in a fracture around 45 degree relative to the maximum stress direction after13crackof 15

reached the edge. The schematic of fracture path is shown in Figure 11f.

Figure 11. Schematic of fracture mechanism; (a,c,e): Fractography of Test 1 (HI = 7.78 kJ/cm), Test 5 (HI
= 8.96 kJ/cm) and Test 6 (HI = 10.22 kJ/cm) respectively. (b,d,f) The corresponding schematic of fracture
mechanism for these three samples.

It is worth noting that the much-reduced elongation would make the prediction of service life
ineffective and lead to premature failure. In addition, as the stress concentration in the root side fusion
boundary may also lead to premature failure in highly stressed or fatigue conditions, it is highly
desirable and necessary to eliminate any stress concentration and avoid the formation of undercut
defect in the K-TIG welded joint. It is suggested that the heat input level be carefully monitored during
the welding process through either arc behaviour or other feedback signals described in the literature,
together with appropriate use of welding speed. The results presented in this study can be useful in
the real fabrication process and could be used as a guidance of parameter optimisation and quality
assurance for the K-TIG welding process.

4. Conclusions
1. Undercuts were easily produced when the travel speed was fixed at 34.2 cm/min, while wavy
ripples were experienced when the heat input was too high. The formation of hump in the weld
centre is a result of low heat input, which changes the profile of weld longitudinal cross section
and plasma trajectory inside the keyhole channel and leads to the exacerbation of undercut
formation. This indicates that both travel speed and heat input are the contributing factors for
undercut formation.
2. The tensile properties of the K-TIG welded joint were closely correlated to the weld formation.
The undercut defect decreased the elongation considerably by imposing stress concentration
and inducing delamination fracture mode. In addition, although the tensile properties were not
appreciably affected by the height of root side reinforcement, stress concentration tended to occur
Metals 2019, 9, 1327 14 of 15

near the root side fusion boundary in the absence of appreciable reinforcement, which would
also affect the regular fracture route. The wavy ripples produced with high heat input had no
evident effect on the tensile properties. It is required that a combination of appropriate heat input
level and travel speed be selected, leading to the absence of undercut defect, either appreciable
root side reinforcement or a combination of post-weld surface treatment with the absence of
appreciable root side reinforcement. It has been demonstrated that the welding parameters
should be carefully controlled to avoid defect formation and maintain tensile properties in the
K-TIG-welded joint.

Author Contributions: Conceptualization, Z.F., B.W. and D.D.; methodology, Z.F. and L.S.; software, D.D. and
L.S.; validation, Z.P. and D.C.; formal analysis, Z.F., B.W. and D.C.; investigation, Z.F. and D.C.; resources, Z.P. and
H.L.; data curation, D.D., H.L. and Z.F.; writing—original draft preparation, Z.F.; writing—review and editing,
Z.P., H.L. and D.C.; supervision, D.C., Z.P. and H.L.; project administration, L.S. and H.L.; funding acquisition, Z.P
and H.L.
Funding: This research has been conducted with the support of the Australian Government Research Training
Program Scholarship. This paper includes research that was supported by DMTC Limited (Australia). The authors
have prepared this paper in accordance with the intellectual property rights granted to partners from the original
DMTC project. This work was partially supported by NSFC under grant No. 51805085.
Conflicts of Interest: The authors declare that they have no conflict of interest.

References
1. Jarvis, B.L. Keyhole Gas Tungsten Arc Welding: A New Process Variant. Ph.D. Thesis, University of
Wollongong, Wollongong, Australia, 2001.
2. Jarvis, B.L.; Ahmed, N.U. Development of keyhole mode gas tungsten arc welding process. Sci. Technol.
Weld. Join. 2000, 5, 1–7. [CrossRef]
3. Liu, Z.M.; Fang, Y.X.; Cui, S.L.; Yi, S.; Qiu, J.Y.; Jiang, Q.; Liu, W.D.; Luo, Z. Sustaining the open keyhole in
slow-falling current edge during K-TIG process: Principle and parameters. Int. J. Heat. Mass. Trans. 2017,
112, 255–266. [CrossRef]
4. Liu, Z.M.; Fang, Y.X.; Cui, S.L.; Luo, Z.; Liu, W.D.; Liu, Z.Y.; Jiang, Q.; Yi, S. Stable keyhole welding process
with K-TIG. J. Mater. Process. Technol. 2016, 238, 65–72. [CrossRef]
5. Liu, Z.M.; Fang, Y.X.; Cui, S.L.; Yi, S.; Qiu, J.Y.; Jiang, Q.; Liu, W.D.; Luo, Z. Keyhole thermal behavior in
GTAW welding process. Int. J. Therm. Sci. 2017, 114, 352–362. [CrossRef]
6. Lathabai, S.; Jarvis, B.L.; Barton, K.J. Comparison of keyhole and conventional gas tungsten arc welds in
commercially pure titanium. Mater. Sci. Eng. A 2001, 299, 81–93. [CrossRef]
7. Lathabai, S.; Jarvis, B.L.; Barton, K.J. Keyhole gas tungsten arc welding of commercially pure zirconium. Sci.
Technol. Weld. Join. 2008, 13, 573–581. [CrossRef]
8. Cui, S.W.; Shi, Y.H.; Sun, K.; Gu, S.Y. Microstructure evolution and mechanical properties of keyhole deep
penetration TIG welds of S32101 duplex stainless steel. Mater. Sci. Eng. A 2018, 709, 214–222. [CrossRef]
9. Fei, Z.Y.; Pan, Z.X.; Cuiuri, D.; Li, H.J.; Wu, B.T.; Ding, D.H.; Su, L.H.; Gazder, A.A. Investigation into the
viability of K-TIG for joining armour grade quenched and tempered steel. J. Manuf. Process. 2018, 32, 482–493.
[CrossRef]
10. Fei, Z.Y.; Pan, Z.X.; Cuiuri, D.; Li, H.J.; Van Duin, S.; Yu, Z.P. Microstructural characterization and mechanical
properties of K-TIG welded SAF2205/AISI316L dissimilar joint. J. Manuf. Process. 2019, 45, 340–355.
[CrossRef]
11. Cui, S.W.; Xian, Z.Y.; Shi, Y.H.; Liao, B.Y.; Zhu, T. Microstructure and Impact Toughness of Local-Dry Keyhole
Tungsten Inert Gas Welded Joints. Materials 2019, 12, 1638. [CrossRef]
12. Fei, Z.Y.; Pan, Z.X.; Cuiuri, D.; Li, H.J.; Wu, B.T.; Su, L.H. Improving the weld microstructure and material
properties of K-TIG welded armour steel joint using filler material. Int. J. Adv. Manuf. Technol. 2019, 100,
1931–1944. [CrossRef]
13. Fei, Z.Y.; Pan, Z.X.; Cuiuri, D.; Li, H.J.; Gazder, A.A. A Combination of Keyhole GTAW with a Trapezoidal
Interlayer: A New Insight into Armour Steel Welding. Materials 2019, 12, 3571. [CrossRef]
14. Xie, Y.; Cai, Y.C.; Zhang, X.; Luo, Z. Characterization of keyhole gas tungsten arc welded AISI 430 steel and
joint performance optimization. Int. J. Adv. Manuf. Technol. 2018, 99, 347–361. [CrossRef]
Metals 2019, 9, 1327 15 of 15

15. Liu, Z.M.; Fang, Y.X.; Qiu, J.Y.; Feng, M.N.; Luo, Z.; Yuan, J.R. Stabilization of weld pool through jet flow argon
gas backing in C-Mn steel keyhole TIG welding. J. Mater. Process. Technol. 2017, 250, 132–143. [CrossRef]
16. Cui, S.L.; Liu, Z.M.; Fang, Y.Q.; Luo, Z.; Manladan, S.M.; Yi, S. Keyhole process in K-TIG welding on 4 mm
thick 304 stainless steel. J. Mater. Process. Technol. 2017, 243, 217–228. [CrossRef]
17. Fang, Y.X.; Liu, Z.M.; Cui, S.L.; Zhang, Y.; Qiu, J.Y.; Luo, Z. Improving Q345 weld microstructure and
mechanical properties with high frequency current arc in keyhole mode TIG welding. J. Mater. Process.
Technol. 2017, 250, 280–288. [CrossRef]
18. Cui, Y.X.; Shi, Y.H.; Hong, X.B. Analysis of the frequency features of arc voltage and its application to the
recognition of welding penetration in K-TIG welding. J. Manuf. Process. 2019, 46, 225–233. [CrossRef]
19. Zhu, T.; Shi, Y.H.; Cui, S.W.; Cui, Y.X. Recognition of Weld Penetration During K-TIG Welding Based on
Acoustic and Visual Sensing. Sens. Imaging 2019, 20, 3. [CrossRef]
20. Feng, Y.Q.; Luo, Z.; Liu, Z.M.; Li, Y.; Luo, Y.C.; Huang, Y.X. Keyhole gas tungsten arc welding of AISI 316L
stainless steel. Mater. Des. 2015, 85, 24–33. [CrossRef]
21. Eriksson, I.; Powell, J.; Kaplan, A.F.H. Measurements of fluid flow on keyhole front during laser welding.
Sci. Technol. Weld. Join. 2011, 16, 636–641. [CrossRef]
22. Tomsic, M.J.; Jackson, C.E. Energy distribution in keyhole mode plasma arc welds. Weld. J. 1973, 110s–115s.
23. Reddy, G.M.; Mohandas, T.; Papukutty, K.K. Effect of welding process on the ballistic performance of
high-strength low-alloy steel weldments. J. Mater. Process. Technol. 1998, 74, 27–35. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).