materials
Review
Interaction Mechanism of Arc, Keyhole, and Weld Pool in
Keyhole Plasma Arc Welding: A Review
Shinichi Tashiro
Citation: Tashiro, S. Interaction
Mechanism of Arc, Keyhole, and Weld
Pool in Keyhole Plasma Arc Welding:
A Review. Materials 2024, 17, 1348.
https://doi.org/10.3390/
ma17061348
Academic Editor: Murali Mohan
Cheepu
Received: 31 December 2023
Revised: 26 February 2024
Accepted: 12 March 2024
Published: 15 March 2024
Copyright: © 2024 by the author.
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 (https://
creativecommons.org/licenses/by/
4.0/).
Materials 2024, 17, 1348. https://doi.org/10.3390/ma17061348
Joining and Welding Research Institute, Osaka University, Osaka 567-0047, Japan;
tashiro.shinichi.jwri@osaka-u.ac.jp; Tel.: +81-6-6879-8666
Abstract: The Keyhole Plasma Arc Welding (KPAW) process utilizes arc plasma highly constricted
by a water-cooled cupper nozzle to produce great arc pressure for opening a keyhole in the weld
pool, achieving full penetration to the thick plate. However, advanced control of welding is known
to still be difficult due to the complexity of the process mechanism, in which thermal and dynamic
interactions among the arc, keyhole, and weld pool are critically important. In KPAW, two large
eddies are generally formed in the weld pool behind the keyhole by plasma shear force as the
dominant driving force. These govern the heat transport process in the weld pool and have a strong
influence on the weld pool formation process. The weld pool flow velocity is much faster than those
of other welding processes such as Tungsten Inert Gas (TIG) welding and Gas Metal Arc (GMA)
welding, enhancing the heat transport to lower the weld pool surface temperature. Since the strength
and direction of this shear force strongly depend on the keyhole shape, it is possible to control
the weld pool formation process by changing the keyhole shape by adjusting the torch design and
operating parameters. If the lower eddy is relatively stronger, the heat transport to the bottom side
increases and the penetration increases. However, burn-through is more likely to occur, and heat
transport to the top side decreases, causing undercut. In order to realize further sophistication of
KPAW, a deep theoretical understanding of the process mechanism is essential. In this article, the
recent progress in studies regarding the interaction mechanism of arc, keyhole, and weld pool in
KPAW is reviewed.
Keywords: plasma arc welding; arc plasma; keyhole; weld pool; interaction; driving force; heat
transport; mass transport
1. Introduction
Plasma Arc Welding (PAW) is a type of arc welding process utilizing a non-consumable
electrode [1]. PAW uses a tungsten electrode like Tungsten Inert Gas Welding (TIGW) to
generate an arc discharge between the electrode and the base metal. This arc is cooled by
a water-cooled copper nozzle attached to the downstream side of the tungsten electrode
and is constricted by the thermal pinch effect [2], so that an arc with a high current density
can be obtained. Accordingly, the plasma jet is strongly accelerated by the Lorentz force
and the plasma temperature is highly increased by Joule heating. The high current density
causes strong heat transfer from the arc to the base metal due to electron condensation
together with thermal conduction from the high-temperature plasma. Consequently, it is
possible to obtain a much higher arc pressure and heat flux to the base metal than TIGW,
easily forming deep and narrow penetration. Figure 1 shows a schematic diagram of PAW.
There are two types of welding modes in PAW, melt-in or conduction mode [3] and
keyhole mode [1], depending on the operating parameters. In the former mode, the current
value and the plasma gas flow rate are suppressed to generate an arc having a relatively
low arc pressure and heat flux, so that welding without forming a keyhole is performed.
The latter is called Keyhole PAW (KPAW). The base metal is partly melted by the high heat
flux from the arc, and the weld pool surface is gradually pushed down by the arc pressure
https://www.mdpi.com/journal/materials
Materials 2024, 17, x FOR PEER REVIEW 2 of 25

Materials 2024, 17, 1348 2 of 24

performed. The latter is called Keyhole PAW (KPAW). The base metal is partly melted by
the high heat flux from the arc, and the weld pool surface is gradually pushed down by
theform
to arc apressure to form
blind keyhole a blind
in the keyhole
base metal andin thefull
then base metal and
penetration canthen full penetration
be achieved to form can be
achieved to form an open keyhole. This mode is mainly used for square butt
an open keyhole. This mode is mainly used for square butt welding using the I-groove. welding
In using
the article,
this I-groove.
the In thismode
latter article, the latter mode is reviewed.
is reviewed.

Figure1.1.Schematic
Figure Schematicof of
plasma arc arc
plasma welding [4]. [4].
welding

When welding steel plates, KPAW enables to weld up to a plate thickness of about
When welding steel plates, KPAW enables to weld up to a plate thickness of about 6–
6–10 mm in a single welding pass [5]. In the case of Gas Metal Arc Welding (GMAW) of thick
plates, ain
10 mm V-aorsingle
Y-groovewelding pass [5].
is generally In thetocase
applied of Gas
the base Metal
metal, andArc Welding
several welding(GMAW)
passes of thick
plates,
are a V- or
required, Y-groove
because is generally
of the insufficientapplied
penetrationto the base of
ability metal,
GMAW and[6].
several
On thewelding
other passes
are required, because of the insufficient penetration ability
hand, KPAW generally uses an I-groove, which does not require operation beveling (groove of GMAW [6]. On the other
hand, KPAW
processing), andgenerally
the welding usescanan be I-groove,
completedwhich in a singledoeswelding
not require
pass. Thisoperation
enablesbeveling
high-efficiency
(groove processing), weldingand withthereduced
welding man-hours and decreases
can be completed heat input
in a single to thepass.
welding base This en
metal, suppressing thermal distortion. Other advantages
ables high-efficiency welding with reduced man-hours and decreases heat input are that the capital investment is to the
much cheaper than for laser welding and electron beam welding, the tolerance to joint gap
base metal, suppressing thermal distortion. Other advantages are that the capital invest
of the base metal is large, and also the electrode consumption is small. Due to the many
ment is much cheaper than for laser welding and electron beam welding, the tolerance to
excellent advantages described above, KPAW has a wide range of applications such as
joint gap steel
structural of the [7],base metal[8],
aerospace is large, and also the
and automobiles [9]. electrode consumption is small. Due to
the many
Much research has been carried out so far withKPAW
excellent advantages described above, the aimhas a wide range
of improving the of applications
quality
suchefficiency
and as structural
of KPAWsteel[10–12].
[7], aerospace
The weld [8],pool
andformation
automobiles [9].of KPAW is strongly
process
Muchby
influenced research
keyholehas been carried
behavior. In order outtosoguarantee
far with the full aim of improving
penetration, the quality and
it is generally
necessary to maintain an open keyhole state with an appropriate
efficiency of KPAW [10–12]. The weld pool formation process of KPAW is strongly keyhole diameter. The influ
keyhole diameter largely depends on the current and the plasma
enced by keyhole behavior. In order to guarantee full penetration, it is generally necessarygas flow rate. If these
are insufficient,
to maintain ana open
blind keyhole
keyhole isstate formed,
withmaking it difficult to
an appropriate obtain diameter.
keyhole full penetration.
The keyhole
On the other hand, if both are excessive, the keyhole diameter becomes large. However,
diameter largely depends on the current and the plasma gas flow rate. If these are insuf
the volume of the weld pool increases accordingly, and the width of the weld pool on the
ficient, surface
bottom a blindincreases.
keyholeAs is aformed,
result, the making
surface it difficult
tension to obtain
cannot supportfull penetration.
the weight of the On the
molten metal, thus leading to a burn-through. Therefore, in order to solve this problem, it the vol
other hand, if both are excessive, the keyhole diameter becomes large. However,
ume
is of the weld
indispensable topool increases
develop sensingaccordingly, and the for
and control methods width of the weld
the keyhole during pool on the bottom
welding.
surface increases.
Various types of Asdirect
a result, the surface
or indirect sensingtension
methods cannot support
for the keyholethehave
weight
beenofpro-
the molten
posed [10]. For example, the voltage of efflux plasma was measured
metal, thus leading to a burn-through. Therefore, in order to solve this problem, it is in with a probe installed
under the baseto
dispensable metal [13]. The
develop voltage
sensing and signal was methods
control found to appearfor theonly duringduring
keyhole the presence
welding.
of an Various
open keyhole and was enabled to be used for estimating
types of direct or indirect sensing methods for the keyhole have the keyhole diameter. The
been pro
voltage of plasma cloud above the top surface of the base metal also can be used for moni-
posed [10]. For example, the voltage of efflux plasma was measured with a probe installed
toring the keyhole state [14]. The plasma cloud is formed by the reflection of plasma jet in
under
the the base
keyhole; so, inmetal [13].keyhole
the blind The voltage signal
case, the plasmawascloudfound to appear
becomes onlyThe
denser. during
abovethe pres
ence of an open keyhole and was enabled to be used for estimating
methods are classified as indirect sensing. Those are convenient methods, but it is difficult the keyhole diameter
The
to voltage
obtain of plasma
accurate cloudonabove
information theshape,
the size, top surface of theof
and position base metalDirect
keyhole. also can be used for
sensing
monitoring
using a visionthe keyhole
sensor state [14].
can obtain moreThe plasma
accurate cloud is formed
information. by theshutter
An ultra-high reflection
speedof plasma
vision system
jet in the was used
keyhole; so,toinsimultaneously
the blind keyhole imagecase,the keyhole and the
the plasma weldbecomes
cloud pool fromdenser.
the The
above methods are classified as indirect sensing. Those are convenient methods, but it is
difficult to obtain accurate information on the size, shape, and position of keyhole. Direc
Materials 2024, 17, 1348 3 of 24

backside of the work-piece [15]. The camera system includes a strobe-illumination unit
(pulse laser), camera head, and system controller. By using this system, both the keyhole
and the weld pool are imaged clearly and simultaneously. Wu’s team developed a cost-
effective vision system for observing the keyhole from the backside of the work-piece, and
clear images of the keyhole are obtained [16]. This vision system can obtain the correlation
between the detailed keyhole behavior and welding parameters, so it might provide clues
to stabilize the keyhole. A pulse current waveform can be used together with the voltage
signal detection system of efflux plasma for controlling the keyhole behavior [17,18]. The
pulse parameters are adjusted in real-time referring to the voltage signal, maintaining the
“one keyhole one pulse” mode to achieve sound bead formation regardless of changes in
the plate thickness during welding [18].
In order to improve the quality and efficiency, variant novel processes of KPAW have
been developed [11]. Variable Polarity Plasma Arc Welding (VPPAW) is used to weld
aluminum, because it enables to clean an oxide layer on the aluminum surface utilizing
cathode spots during the Electrode Positive (EP) period [19]. A novel soft VPPAW has been
proposed for decreasing arc pressure acting on the keyhole by using a special nozzle with
three holes, suppressing welding defects such as undercut and porosity [20]. Gas focusing
PAW was designed for improving penetrability [21]. The focusing gas is used to focus arc
plasma which is flowing out of the nozzle, increasing the penetration depth. An ultrasonic
vibration-assisted KPAW process was developed [22]. The tungsten electrode connected
with a specially designed ultrasonic transducer directly transmitted ultrasonic vibration
into the plasma arc. During welding, an open keyhole can be produced with a lower
welding current and higher welding speed compared to the conventional KPAW. Plasma–
GMA (or Metal Inert Gas (MIG) using inert shielding gas for GMA) hybrid welding is
especially suitable for welding of thick plates [23,24]. PAW in the leading position produces
a keyhole and then GMAW in the trailing position fills the keyhole with metal droplets
transferred from the wire, achieving high-efficiency welding with low heat input. However,
the process mechanism of plasma–GMA hybrid welding is known to be very complex
especially due to an interaction between two arcs, therefore is difficult to control precisely.
Yu et al. tried to control the arc coupling phenomenon by applying an external magnetic
field for stabilizing metal transfer [25]. The result showed that the metal transfer mode
was significantly affected by the plasma current via the arc coupling. The arc coupling
duration enables it to be controlled by the external magnetic field. The influence of the
external magnetic field on the welded bead was also investigated [26]. It was found that
the mechanical properties of the welded bead were improved by the effect of the external
magnetic field. The transport of molten metal provided by GMAW in the weld pool was
then studied to discuss the element mixing process [27]. Recently, metal transfer and the
weld pool flow in VPPA-MIG hybrid welding of aluminum alloy plates were observed to
understand the process mechanism [28].
As described above, various approaches are attempted to improve the KPAW process.
However, a deep theoretical understanding of the process mechanism of KPAW is essential
to achieve further sophistication. In KPAW, the physical phenomena in the arc, keyhole,
and weld pool are strongly coupled via thermal and dynamic interactions. This article
reviews the recent progress in studies on the process mechanism of KPAW. Section 1 is
the introduction. Section 2 is concerned with plasma arc characteristics. In addition, the
effect of metal vapor evaporated form the weld pool is also presented. In Section 3, the
keyhole formation process is discussed. Section 4 presents the weld pool formation process.
Section 5 provides the summary and future work.

2. Plasma Arc Characteristics

First, the plasma arc characteristics of PAW are described. As mentioned in Section 1,
PAW generates a plasma arc via electric discharge between the tungsten electrode and
the base metal. The current density is increased by contracting the plasma arc with a
water-cooled copper nozzle mounted directly under the tungsten electrode. This increase in
Materials 2024, 17, 1348 4 of 24

current density enhances the Lorentz force to accelerate the plasma arc jet from the tungsten
electrode to the base metal. At the same time, strong Joule heating occurs and the plasma
arc becomes extremely high temperature, especially near the nozzle. The high heat flux
brought by the constricted plasma arc starts to melt the base metal, and the molten metal is
pushed down by the arc pressure to form a keyhole. As described later, it is considered that
the keyhole shape is determined mainly depending on the arc pressure field. When the
open keyhole is formed, the plasma arc jet flowing from the electrode side branches in the
keyhole. A part of this passes through the keyhole exit and flows down to the bottom side
to become efflux plasma, the rest flows outward along the keyhole wall on the top side [29].
Both impose a strong shear force on the surface of the weld pool and are considered to be
the main driving force of the weld pool convection in KPAW.
Therefore, in order to control the formation process of the keyhole and the weld pool, it
is essential to understand the influence mechanism of the torch design such as the electrode
shape, the water-cooled copper nozzle shape, and the electrode setback, as well as the
operating parameters such as the current and the composition and flow rate of the plasma
gas and shielding gas, on the plasma arc characteristics including the temperature field and
velocity field of the plasma arc, as well as the arc pressure, shear force, current density, and
heat flux applied to the base metal.

2.1. Plasma Temperature

The temperature field of the plasma arc governs the current path in the arc, so it is
the most important quantity to affect the heat source properties of PAW. The temperature
is mainly acquired by spectroscopic measurement. The measured temperature field is
very useful information for validating the numerical simulation results. When performing
numerical simulations of plasma arc characteristics, the Local Thermodynamic Equilibrium
(LTE) approximation [30] is generally applied for simplicity, but it is known that non-
equilibrium phenomena occur, for example, in low-temperature regions where the collision
frequency between electrons and heavy species decreases [31]. Since the current path in the
plasma arc is strongly affected by a change in electrical conductivity caused by the non-LTE
phenomena, it can be an important factor for improving the calculation accuracy.
Boselli et al. spectrally measured the temperature field of PAW in a low-current range
of 40 A by the relative line intensity method, and the temperature reached about 14,000 K on
the arc axis below the nozzle, which gradually decreased in the radial direction [32]. They
also calculated the temperature field using a non-equilibrium model, comparing the result
with that of the conventional LTE model and the temperature measurement results. The
excitation temperature of argon plasma measured around the outer edge of the plasma arc
was consistent with the electron temperature obtained by the non-equilibrium model and
was much higher than the LTE temperature, indicating that non-equilibrium phenomena
are strongly exhibited in this region.
Pan et al. proposed a fully coupled tungsten-plasma-anode model for the first time, to
study the heat transport and fluid flow during keyhole formation in KPAW [33]. For the
calculation, the current was set to 120 A, and the plasma arc was assumed to satisfy the LTE
approximation. The radial distribution of electron temperature at the mid-height of the arc
gap was measured by the Boltzmann diagram method. As a result, it reached 25,000 K on
the arc axis and dropped to almost room temperature at a radius of 3 mm. The measured
electron temperature distribution is in good agreement with the calculation result under
LTE approximation, indicating that the calculation assuming the LTE approximation is
appropriate for a relatively high current range.
Ishida et al. obtained the temperature field and metal vapor concentration field in
plasma–MIG hybrid welding by three-dimensional spectroscopic measurement [34] and
elucidated the coupling mechanism of the plasma arc and MIG arc [35]. When MIGW used
pulsed current, the arc coupling occurred only during the upslope of the pulse. It is shown
in Figure 2 that the plasma arc temperature at 220 A was about 19,000 K just below the
Materials 2024, 17, x FOR PEER REVIEW 5 of 25

Materials 2024, 17, 1348

pulsed current, the arc coupling occurred only during the upslope of the pulse.5 of It 24
is shown
in Figure 2 that the plasma arc temperature at 220 A was about 19,000 K just below the
nozzle, but this
nozzle, but thisincreased
increased when
when arcarc coupling
coupling occurred
occurred to connect
to connect the plasma
the plasma arc current
arc current
and
and MIGW currentpartly.
MIGW current partly.

Figure 2. Temperature field and metal vapor concentration field in plasma–MIG hybrid welding
Figure 2. Temperature field and metal vapor concentration field in plasma–MIG hybrid welding
obtained by three-dimensional spectroscopic measurement [35].
obtained by three-dimensional spectroscopic measurement [35].
2.2. Metal Vapor
2.2. Metal
As theVapor
temperature of the weld pool rises, metal vapor evaporates from the surface.
WhenAs thethe
metal vapor is mixed
temperature of thewith
weldthe pool
plasma, it has
rises, a great
metal effectevaporates
vapor on its thermodynamic
from the surface.
and transport properties and radiation loss [36]. Especially in a low-temperature
When the metal vapor is mixed with the plasma, it has a great effect on its thermodynamic range
of 10,000 K or less, the electrical conductivity of plasma increases by several orders of
and transport properties and radiation loss [36]. Especially in a low-temperature range of
magnitude even if metal vapor is mixed at only about 1%. In addition, the radiation loss
10,000 K or less, the electrical conductivity of plasma increases by several orders of mag-
increases remarkably as the mixing rate increases.
nitude Foreven if metal
example, vapor
in the case is
of mixed
TIGW of atstainless
only about steel1%. In addition,
SUS304 at 150 A, the radiation
the iron vapor loss in-
creases remarkably
concentration near theascenter
the mixing rate increases.
of the weld pool is about 7% when helium gas is used and
aboutFor0.2%example,
when argonin the
gascase of [37].
is used TIGW of stainless
Particularly, steel
in the SUS304
former case,at 150 shown
it was A, the that
iron vapor
the current density
concentration nearofthe
thecenter
plasmaofwasthegreatly
weld reduced,
pool is aboutdecreasing
7% whenthe penetration
helium gas depth.
is used and
aboutLi0.2%
et al.when
investigated
argon thegaseffect of metal
is used [37]. vapor in conduction
Particularly, in the PAW
formerwith argon
case, gas at
it was 40
shown that
A [3]. Due to the constriction effect of the arc, the iron vapor concentration
the current density of the plasma was greatly reduced, decreasing the penetration depth. near the center
of the weld pool reached about 9%. As a result, it was found that the thermal conduction
Li et al. investigated the effect of metal vapor in conduction PAW with argon gas at
from the arc to the base metal was reduced to about half due to the mixing of metal vapor.
40 A On [3].the
Due to the constriction effect of the arc, the iron vapor concentration near the
other hand, Jian et al. conducted a simulation of the keyhole formation process
center of theafter
immediately weldthepool reachedwhen
arc ignition about 9%. KPAW
argon As a result, it was found
was performed at 150that the thermal con-
A considering
duction from the arc to the base metal was reduced to about
the influence of metal vapor [38]. Inside the keyhole, the weld pool temperature half due to thewasmixing of
metal vapor. that there was almost no metal vapor. The temperature of the weld pool
low, indicating
On the
increased other
near the hand, Jian et
outer edge of al.
theconducted
keyhole, and a simulation of the
the metal vapor keyhole formation
concentration reachedprocess
0.2% maximum at that position.
immediately after the arc ignition when argon KPAW was performed at 150 A considering
Consequently,
the influence the plasma
of metal vapor arc inside
[38]. thethe
Inside keyhole
keyhole,is hardly affected
the weld by temperature
pool the metal vapor, was low,
and it is considered that this has little effect on the formation of the keyhole.
indicating that there was almost no metal vapor. The temperature of the weld pool in-
creased near the outer edge of the keyhole, and the metal vapor concentration reached
2.3. Arc Pressure
0.2% Inmaximum at that position.
variable polarity arc welding, the energy balance on the surface of the tungsten
Consequently, the depending
electrode changes greatly plasma arconinside the keyhole
the polarity. is hardlyNegative
In the Electrode affected(EN)
by the metal va-
phase,
por, and it tends
the surface is considered that
to be cooled bythis has little
thermionic effectemission,
electron on the formation of the keyhole.
and in the Electrode Positive
(EP) phase, it tends to be heated by electron condensation [39]. Since the electrode surface
2.3. Arc Pressure
In variable polarity arc welding, the energy balance on the surface of the tungsten
electrode changes greatly depending on the polarity. In the Electrode Negative (EN)
Materials 2024, 17, 1348 6 of 24

temperature governs thermionic electron emission characteristics, the current attachment is

thought to be greatly affected by the polarity.
Xu et al. proposed a VPPAW model incorporating a current attachment model consid-
ering the above energy balance and clarified the effect of energy balance of the electrode on
plasma arc characteristics and also the arc pressure on the base metal [40]. Due to energy
balance on the electrode surface, the electrode temperature was decreased over time by
thermionic electron emission at the early stage of the EN phase. As a result, the current
attachment was constricted to the tip of the electrode. It enhanced the Lorentz force acting
on the arc column in the EN phase, making the arc pressure higher.

2.4. Comparison with TIGW

The arc pressure, shear force, current density, and heat flux on the surface of the base
metal are considered to be particularly important as inputs from the plasma arc to the base
metal, affecting the formation processes of the keyhole and the weld pool. Here, these
values are compared with results under similar conditions of TIGW, which is a welding
method using a tungsten electrode as in PAW.
From the results of numerical simulations of argon TIGW at 150 A, the arc pressure
was found to be 400 Pa at the arc axis [41] and the shear force was increased up to 40 Pa at
a position slightly apart from the axis [42]. In the same condition, the current density and
heat flux were measured on the axis to be a maximum of 5 × 106 A/m2 and 5 × 107 W/m2 ,
respectively [43].
On the other hand, in PAW, the arc pressure at 150 A was measured by Xu et al. [44]
and Li et al. [21]. It was reported that the maximum values were 3.3 kPa and 3.5 kPa,
respectively. The shear force at 135 A was calculated to be about 500 Pa by Wu et al. [45].
The current density and heat flux at 120 A were calculated by Pan et al., as 8 × 106 A/m2
and 4 × 107 W/m2 [33]. Calculations at 170 A by Jian et al. showed 8 × 106 A/m2 and
1.2 × 108 W/m2 [46].
In summary, it was shown that the arc pressure and shear force of PAW reached about
10-times those of TIGW, and the current density and heat flux were higher than those of
TIGW. As mentioned above, these also vary greatly depending on the torch structure and
operating parameters.

2.5. Effect of Torch Design and Operating Parameters

Due to the difficulty to measure the temperature field and velocity field of the plasma
arc, there are few reports of experimental measurement results as described above, so
numerical simulation is mainly utilized for the investigation.
Schnick et al. predicted the plasma arc characteristics in PAW through numerical
investigation [47]. Here, the plasma arc is assumed to satisfy the LTE approximation.
The base metal is a water-cooled copper anode with a flat surface. The effects of design
parameters, gas composition, gas flow rate, and current on the temperature and velocity
fields of the plasma arc were clarified. Furthermore, the effects on the arc pressure and heat
flux on the base metal have also been elucidated.
Since the arc is constricted by the water-cooled copper nozzle, the Lorentz force works
strongly, and the maximum flow velocity of the argon plasma arc with a current of 100 A
reaches about 1500 m/s. This velocity is about 7.5-times the calculation result of TIGW at a
current of 150 A [48]. The arc pressure and shear force acting on the base metal increase as
the flow velocity of the plasma arc increases. In addition, the temperature of the plasma arc
reached a maximum of 25,000 K directly under the tungsten electrode due to an increase
in Joule heating. This is about 1.5-times the measurement result of TIGW at a current of
100 A [49].
Investigations also show that an increase in current raises the plasma temperature and
plasma flow velocity, while the plasma gas flow rate hardly affects the plasma temperature
but increases the plasma flow velocity. However, both increase the heat flux and the arc
pressure at the base metal. The nozzle diameter and cathode position are adjustable to
Materials 2024, 17, 1348 7 of 24

affect the plasma arc. The larger torch distance expands the plasma arc column gradually
due to the effect of the high viscosity of plasma, thus decreasing the maximum arc pressure.
On the other hand, the internal angle of the plasma nozzle is less effective. The effect of gas
composition was also discussed. The addition of helium to argon increases the viscosity
of plasma arc, giving a dominant effect on decreasing the arc pressure. The addition
of hydrogen raises the thermal conductivity, increasing the heat flux. According to the
demixing effect, the concentration of helium and hydrogen increases around the arc axis in
the vicinity of the base metal surface.

3. Keyhole Formation Process

Next, the keyhole formation process is described. Since the energy and force from
the plasma arc are brought to the base metal through the keyhole wall, these are strongly
affected by the shape and stability of the keyhole, so a deep understanding of the keyhole
formation process is required.
After the arc ignition, the surface of the base metal begins to melt due to the high heat
flux from the plasma arc. The surface of the weld pool is pushed down by the arc pressure,
and the molten metal is transported to the outside by the weld pool convection driven by
the shear stress due to the plasma flow and the Marangoni force caused by the temperature
gradient on the weld pool surface, thus forming a blind keyhole. It is considered that
the keyhole shape is mainly determined by the force balance between the arc pressure
acting to expand the keyhole and the surface tension and gravity acting to close it. As
the keyhole depth increases, the heat flux and arc pressure applied to the deepest part
gradually decrease, but if those are large enough, the keyhole penetrates to the bottom to
form an open keyhole.
In keyhole welding using a laser or electron beam, the weld pool temperature inside
the keyhole rises maximally up to around the boiling point. Therefore, the evaporation of
metal vapor is known to be significantly intensive, so the recoil pressure accompanying this
evaporation is the main driving force for the keyhole formation [50]. On the other hand, in
case of KPAW, since the surface temperature of weld pool is just above the melting point,
the evaporation of metal vapor is very weak [38], so it is considered that the recoil pressure
due to evaporation is negligibly small.

3.1. Keyhole Formation Models

In the study of the formation process of the keyhole and the weld pool around it, in
many cases, the heat transport process is investigated mainly by numerical simulation
supported by experimental observation. Since it is difficult to directly observe the dy-
namic keyhole phenomenon during welding, analysis by numerical simulation is effective.
Therefore, various types of simulation models have been proposed so far.
Investigations using simple arc heat source models [51,52] have been carried out since
the early stages of the research. Models were improved to follow dynamic changes in the
keyhole shape. In simulation using the arc heat source model, source terms described with
special functions are generally given to the energy conservation equation and momentum
conservation equation for considering the effects of heat flux, arc pressure, shear force,
etc., from the arc. In conjunction with this model, the Volume of Fluid (VOF) method
is often used to track the weld pool surface [53–56]. Some constants and coefficients of
functions used in the arc heat source model are empirically determined based on the
experimental results. This model has a low calculation cost and is highly convenient
from an engineering point of view. It is mainly suitable for predicting the dependence of
experimental parameters on the weld pool shape and penetration shape.
On the other hand, in recent years, in order to study the thermal and dynamic interac-
tions of the arc, keyhole, and weld pool in more detail, a coupled electrode-arc-keyhole-weld
pool model applying the VOF method has been developed [33,46,57,58]. This model does
not require various assumptions used in the arc heat source model and can comprehen-
sively clarify the effects of the torch design and operating parameters on the formation of
Materials 2024, 17, 1348 8 of 24

keyholes and weld pools through changes in arc characteristics. Therefore, compared with
the arc heat source model, it becomes possible to examine the keyhole formation process in
more detail. However, it is computationally very expensive and difficult to apply it to a
large-scale parametric study.
Furthermore, a simple coupled model [29,59] using a pre-fixed keyhole shape without
tracking the weld pool surface by the VOF method is proposed to reduce calculation costs.
A one-way coupling model [45] has also been proposed, in which the arc model and
the weld pool model are not coupled in a two-way, but the boundary conditions of the
weld pool are determined based on arc characteristics obtained by the arc model.

3.2. Driving Force for Keyhole Formtion

Wu et al. used the aforementioned model to conduct a detailed study of the keyhole
formation process [60]. In welding, arc pressure, shear force, Marangoni force, etc., which
are the driving forces for keyhole formation, work simultaneously. But, by giving these
forces individually as virtual forces, the dominant driving force for keyhole formation can
be examined. Consequently, it is found that the weld pool deformations driven by the
arc pressure and shear stress are suggested to be the two responsible mechanisms for the
formation of keyhole. The first one is especially dominant.
The process between full penetration of the weld pool and the formation of an open
keyhole is defined as the blasting penetration stage, where the molten bridge located at the
bottom of the blind keyhole rapidly breaks [54,61]. As a result of numerical simulation [45],
the pressure balance and also the mass conservation mechanisms are considered to strongly
affect the occurrence of blasting penetration. The decrease in the melting point and surface
tension as well as the deformation of the bottom surface are thought to facilitate the
blasting penetration. On the other hand, the increase in thermal conductivity restraints the
blasting penetration.

3.3. Energy Balance

The energy balance between the plasma arc and the weld pool through the keyhole
wall has also been studied by numerical simulation.
In KPAW, after the open keyhole formation by plasma arc, part of the plasma becomes
an efflux plasma that blows out below the base metal through the keyhole exit. Li et al.
estimated that about 10% of the plasma arc outflows from the keyhole exit [29], even though
this value is predicted to change depending on the torch design and operating parameter.
Wu et al. studied the energy balance in KPAW [62,63]. The energy transport among
the electrode, plasma arc, and base metal was numerically evaluated using the coupled
electrode-arc-keyhole-weld pool model applying the pre-fixed keyhole. Figure 3 shows the
energy balance in the KPAW process of steel plate [62]. Thermal efficiencies of the PAW
process with and without keyhole were also compared. In the case with a keyhole, both
the energy consumption by arc and the energy input increase. On the other hand, the total
energy transport from the arc to the weld pool slightly decreases, thus the thermal efficiency
decreases. The open keyhole formation suppresses the energy transport to the weld pool,
and 11.5% of the arc energy was lost by the efflux plasma. The calculated thermal efficiency
is only 60.7% [62]. The thermal efficiency of KPAW was found to be relatively lower than
those of GTAW and GMAW processes. Tanaka et al. presented the thermal efficiency of
82% in the GTAW process [41]. Joseph et al. experimentally obtained a thermal efficiency
of 68–72% in the pulsed GMAW process through liquid nitrogen calorimetry [64].

3.4. Relaationship between Keyhole Exit Deviation and Porosity Formation

Wu et al. clarified the effect of keyhole behaviors on the formation of porosity in
KPAW [65,66]. From the observed image of the keyhole exit, the dimensional parameters
as well as the keyhole exit deviation distance were obtained. Pulsed PAW was carried out
for bead-on-plate welding and closed butt joint welding. The result showed that in the case
of a large keyhole exit deviation distance, high reinforcement was formed in the front weld
 Materials 2024, 17, 1348 9 of 24

Materials 2024, 17, x FOR PEER REVIEW

surface to induce porosity in the solidified weld, while in the case of a small distance, a
sound weld was achieved. The keyhole exit deviation distance is suggested to be a critical
keyhole factor to determine the weld quality. In a keyhole with a large deviation distance,
the gas flow field was disturbed and much liquid metal was pushed toward the front pool
thermal efficiency ofthe68–72%
surface. Consequently, inwas
weld pool thedeformed
pulsed GMAW process
largely with through
an inside liquid
gas cavity. The nitroge
rimetry [64]. that the weld quality in KPAW highly depends on the keyhole behavior.
results indicate

Figure 3.3.The
Figure Theenergy balance
energy in thein
balance KPAW process [62].
the KPAW process [62].
3.5. Effect of Torch Design and Operating Parameters
3.4. Relaationship between
The keyhole shape Keyhole
after open Exitformation
keyhole Deviation
hasand Porosity
a strong effectFormation
on the energy and
force Wu
balance et between the plasma
al. clarified arc andofthekeyhole
the effect weld pool. It is suggested
behaviors that formation
on the this keyhole of por
shape strongly depends on the torch design and operating parameters as described above.
KPAW [65,66]. From the observed image of the keyhole exit, the dimensional para
Nguyen et al. studied the effects of plasma gas flow rate [67], current [68], and oxygen
as well as theinkeyhole
concentration shielding exit deviation
gas [69] in KPAWdistance wereIn
of a steel plate. obtained. Pulsed
addition, Xu PAW was carr
et al. clarified
for bead-on-plate
the effects of electrodewelding
setback [70]andandclosed butt joint
plate thickness [71] welding.
in VPPAW of The result alloys.
aluminum showed tha
Here,of
case thea results for KPAWexit
large keyhole of a deviation
steel plate are shown ashigh
distance, an example.
reinforcement was formed in th
Figure 4 shows the effect of plasma gas flow
weld surface to induce porosity in the solidified weld, whilerate on the keyhole shape [67].
in the caseIt can
of be
a small d
seen that at a flow rate of 0.7 L/min, the keyhole does not penetrate and becomes a blind
akeyhole,
soundand weld was achieved. The keyhole exit deviation distance is suggested to b
at a flow rate of 1.7 L/min or more, it becomes an open keyhole. Under the
ical keyhole
condition that factor
an opentokeyhole
determine the weld
is formed, as the quality. In a keyhole
flow rate increases, with a large
the diameter of the deviat
tance,
bottom the gasincreases
keyhole flow field was
and the disturbed
keyhole and much
wall behind the keyholeliquid metal
becomes wasvertical.
nearly pushed tow
These pool
front changes are considered
surface. to be due to
Consequently, thetheweld
remarkable
pool was increase in arc pressure
deformed largelydue to an ins
with
the increase in gas flow rate. It can be also seen that an undercut occurs at the weld toe
cavity. The results indicate that the weld quality in KPAW highly depends on the k
when the plasma gas flow rate is large.
behavior.

3.5. Effect of Torch Design and Operating Parameters

The keyhole shape after open keyhole formation has a strong effect on the
and force balance between the plasma arc and the weld pool. It is suggested that th
hole shape strongly depends on the torch design and operating parameters as de
above. Nguyen et al. studied the effects of plasma gas flow rate [67], current [6
oxygen concentration in shielding gas [69] in KPAW of a steel plate. In addition, X
clarified the effects of electrode setback [70] and plate thickness [71] in VPPAW of
num alloys. Here, the results for KPAW of a steel plate are shown as an example.
Figure 4 shows the effect of plasma gas flow rate on the keyhole shape [67]. I
seen that at a flow rate of 0.7 L/min, the keyhole does not penetrate and becomes
Materials2024,
Materials 2024,17,
17,1348
x FOR PEER REVIEW 10 of 24
10 of 25

Figure4.4.Weld
Figure Weldbead
beadappearances
appearancesand
andcross
crosssections
sectionsofofthe
theweld
weldbeads
beadsfor
fordifferent
differentplasma
plasmagas
gasflow
flow rates
rates [67]. [67].

Figure 55shows
Figure showsthe theeffect
effectof ofcurrent
currenton onthethekeyhole
keyholeshape shape[68].
[68].An
An open
openkeyhole
keyhole isis
formed at all current conditions. At 90 A, the diameter of the bottom keyhole isisslightly
formed at all current conditions. At 90 A, the diameter of the bottom keyhole slightly
smallwith
small withaavalue
valueof of1.6
1.6mm;
mm;but,but,atat120
120A, A,ititexpands
expandsto to2.82.8mm,
mm,andandatat150
150A,
A,ititreaches
reaches
5.2mm.
5.2 mm. At 150 A, A, the
the angle
angleofofthethekeyhole
keyholewall wallbehind
behindthe thekeyhole
keyholeisissmaller
smallerthan
thanthat at
that
at 120
120 A.A.InInaddition,
addition,the theweld
weldpool
poolwidth
widthincreases
increases withwith the current. When the the current
current isis
increased,
increased,thetheamount
amountof ofheat
heatinput
inputisisincreased,
increased,and andthetheelectromagnetic
electromagneticforce forceapplied
appliedto to
the
theplasma
plasmaarc arcisisstrengthened,
strengthened,so sothe
thearcarcpressure
pressureisisalsoalsoincreased.
increased.ItItcan
canbebeseen
seenthat
thatthe
the
result
resultatat90
90AAshows
showshumping
humpingonon the top
the topsurface
surface and
andthat at 150
that A shows
at 150 A showsburn-through
burn-through on
the bottom surface. A sound weld bead is obtained
on the bottom surface. A sound weld bead is obtained only at 120 A. only at 120 A.
Figure
Figure66shows
showsthe theeffect
effectof ofoxygen
oxygenconcentration
concentrationin inthetheshielding
shieldinggas
gasononthe
thekeyhole
keyhole
shape
shape [69]. Here,
Here,a asmall
small amount
amount of oxygen
of oxygen is mixed
is mixed in theinshielding
the shielding gas, assuming
gas, assuming shield-
shielding
ing failure.failure. The addition
The addition of 0.5%ofoxygen
0.5% oxygen to theshielding
to the argon argon shielding gas increases
gas increases the diameterthe
diameter of the bottom keyhole and decreases the angle of the
of the bottom keyhole and decreases the angle of the keyhole wall behind the keyhole. keyhole wall behind the
keyhole. This is probably
This is probably because because
the surfacethe tension
surface of tension of the
the weld weld
pool pool decreased
decreased dueabsorp-
due to the to the
absorption of oxygen into the
tion of oxygen into the weld pool. weld pool.
As
Asdescribed
described above,
above, thethe keyhole
keyhole shape
shapechanges
changes greatly
greatlywith withvariations
variations ininwelding
welding
conditions. This has
conditions. This hasaastrong
strongeffect
effectononthethe weld
weld pool
pool formation
formation process
process as seen
as seen in next
in the the
next section.
section.
 Materials 2024,
Materials 17, x FOR PEER REVIEW 11 of 11
25 of 24
2024,17,
Materials2024, 17, 1348
x FOR PEER REVIEW 11 of 25

Figure5.5.
Figure
Figure 5.Weld
Weld bead
Weldbead appearances
beadappearances
appearancesand
and
and cross sections
cross
cross ofofthe
sections
sections thethe
of weld
weld beads
weld fordifferent
beads
beads for different
for currents
different [68].[68].
currents
currents [68].

Figure6.6.
Figure
Figure 6.Weld
Weld bead
Weldbead appearances
beadappearances
appearancesand
and
and cross sections
cross
cross ofofthe
sections
sections thethe
of weld
weld beads
weld fordifferent
beads
beads for different
for shielding
different gas gas
shielding
shielding gas
compositions [69].
compositions[69].
compositions [69].
Materials 2024, 17, 1348 12 of 24

4. Weld Pool Formation Process

The weld pool formation process is described in this section. During the travel of
the torch in the welding direction, the molten metal on the front side of the keyhole is
transported through the sides of the keyhole toward the rear region to form the weld
pool. The heat transport in the weld pool is considered to be primarily due to weld
pool convection, especially in steel welding, driven by the buoyancy force, Lorentz force,
Marangoni force, and plasma shear force. Particularly, the shear force and Marangoni force
are thought to strongly depend on the keyhole shape, since those work on the surface of the
weld pool in the vicinity of the keyhole. The molten metal in the weld pool is supported by
the surface tension of the bottom surface of the weld pool, preventing burn-through. As
discussed in the previous sections, the arc, keyhole, and weld pool strongly interact together,
thus making KPAW a very complex process. Furthermore, the weld pool formation easily
becomes unstable, inducing various welding defects, for example, by disturbance or a
change in the condition in welding. Accordingly, to achieve high-quality welding, it is
necessary to deeply understand the weld pool formation processes.

4.1. Weld Pool Observation

Information on the flow field and temperature field is indispensable for understanding
the heat transport process in the weld pool, but due to the difficulty in measurement, there
are still few reports of these experimental measurement results.
Simultaneous measurement of both the keyhole behavior and its surrounding tem-
perature field of the weld pool is indispensable for understanding the thermo-physical
mechanism and also achieving the process control of KPAW. Zhang et al. used an infrared
camera and a coupled charge device (CCD) camera to observe the temperature profile
of the weld pool in the vicinity of the keyhole and the keyhole behavior from below the
base metal at the same time [72]. After image processing and calibration, the shapes and
sizes of keyhole and weld pool as well as their relative positions were determined with
temperature field.
Chen et al. observed the flow field on the surface of the weld pool employing a high-
speed camera equipped with a laser light system and also the temperature field using an
infrared thermal camera in VPPAW of aluminum alloys [73]. The formation mechanism of
the keyhole weld pool at different welding positions was studied, especially for clarifying
the effect of gravity on material flow, temperature field, and keyhole morphology. Based
on the result of both calculation and experiment, the effect of gravity on the weld pool
was found to be second only to shear force in all factors. Consequently, the heat and mass
transport in the low-velocity region of the VPPA keyhole weld pool are thought to be
readily influenced by the gravity.
In order to gain a deep understanding of the heat transport process in the weld pool, it
is necessary to measure the flow field of the entire weld pool including the inside. For the
first time, Nguyen et al. applied stereo synchronous imaging of tungsten tracer particles
employing two sets of X-ray transmission systems to measure the 3D flow field inside
the weld pool in KPAW [74]. The 2D convection on the surface of the weld pool was also
measured by observing the motion of zirconia tracer particles. Based on this comprehensive
measurement, the weld pool convection in the region behind the keyhole to the weld pool
end was totally visualized. For discussing the heat transport process in the weld pool, the
2D temperature field on the top and bottom surfaces of the weld pool was also obtained
by two-color pyrometry. A summary of the results of this comprehensive measurement is
provided below.
Figure 7 shows weld pool images and 2D temperature fields on the top and bottom
surfaces. On the top weld pool surface, the temperature was measured to be between
1718 K and 1795 K. The weld pool is divided into regions 1–3, according to the difference
in temperature. The temperature reached 1795 K in region 2. On the bottom weld pool
surface, the temperature was distributed between 1714 K and 1794 K. The temperature
monotonically decreased from 1794 K around the keyhole to 1714 K at the weld pool end.
Materials 2024, 17, x FOR PEER REVIEW 13 of 25
Materials 2024, 17, x FOR PEER REVIEW 13 of 25

Materials 2024, 17, 1348
surface, the temperature was distributed between 1714 K and 1794 K. The temperature
surface, the temperature
monotonically decreased was
fromdistributed between
1794 K around 1714 Kto
the keyhole and 1794
1714 K atK.the
The temperature
weld pool13end.
of 24
monotonically decreased from 1794 K around the keyhole to 1714 K at the weld pool end.

Figure 7. Observed image and temperature field on the surface of the weld pool. (a) Observed image;
Observedimage
fieldimage
Figure7.7.Observed
Figure
(b) temperature andtemperature
and
on top temperaturefield
(left figure) andfield onthe
on
bottom thesurface
surfaceof
(right ofthe
figure) theweld
weldpool.
surfacespool.(a)
[74]. (a)Observed
Observedimage;
image;
(b)temperature
(b) temperaturefield
field on
on top
top (left
(left figure)
figure) and
and bottom
bottom (right
(right figure)
figure) surfaces
surfaces [74].
[74].
Figure
Figure888shows
shows
shows2D2D convective
2Dconvective
convective patterns
patterns ononthethetop andand bottom weld poolpool
surfaces. On
the Figure
top weld pool surface, the patterns
definition of on
each toptop
theregion andis the
bottom
bottom
same weld
as
weld
pool
that in
surfaces.
surfaces.
Figure On
7. In
Ontop
the the weld
top weld pool
poolmetal surface,
surface, the definition
the definition of each
of each region region
is theis same
the same as that
astemperature.
that in Figure
in Figure 7. In7.
region
In region1, molten
1, molten flowed
metal from
flowed the
from keyhole
the keyholeto region 2
to region at a high It also
region
flowed 1,inward
molten metal
from theflowed
weld from
pool the keyhole
edge to the to region
center of the a2 high
2 at weld at a high
pool.
temperature.
temperature.
In region It3,also
the
It
also flowed inward from the weld pool edge to the center of the weld pool. In region 3,
flowed
molten inward
metalmetalfrom the
flowed in aweld
forwardpooldirection
edge to from
the center of the
the weld weld
pool endpool. In region
2 at a3,2high
the
the molten flowed in a forward direction from the weld pooltoend
region
to region at a
molten metal
temperature. flowed
In region in a forward
2, the 2,
molten direction
metalmetalfrom the
flows flows weld
from regions pool end to
1 and13andregion
were 2 at
merged. a high
Zir-
high temperature. In region the molten from regions 3 were merged.
temperature.
conia particles Inwere
region 2, thedue
rotated molten metal
to this flows from
convective regions
pattern. On 1the and 3 were merged. Zir-
Zirconia particles were rotated due to this convective pattern. Onbottom weld
the bottom pool pool
weld sur-
conia
face, particles
the molten were
metalrotated
flowed due to
backward this convective
to the weld pattern.
pool On
end.end. the
TheThebottom
maximum weld pool
velocities sur-
of
surface, the molten metal flowed backward to the weld pool maximum velocities
face,
0.53 the
m/sm/smolten
andand metal
0.810.81
m/sm/sflowed
werewere backward
seenseen
at X at to
= 2.0 the weld pool end. The maximum velocities of
of 0.53 X =mm on the
2.0 mm ontop
the surface
top surfaceand and
X = 1.0
X =mm on the
1.0 mm on
0.53
bottomm/ssurface,
and 0.81 m/s were seen at X = 2.0 mm on the top surface and X = 1.0 mm on the
respectively.
the bottom surface, respectively.
bottom surface, respectively.

Figure
Figure8.8.Convective
Convectivepattern
patternon
ontop
top(upper
(upperfigure)
figure)and
andbottom
bottom(lower
(lowerfigure)
figure)surfaces
surfacesof
ofthe
theweld
weld
Figure
pool 8. Convective pattern on top (upper figure) and bottom (lower figure) surfaces of the weld
[74].
pool [74].
pool [74].
Figure 9 shows 3D convective patterns in the weld pool. The result showed that two
eddy convective patterns in opposite directions were stably formed in the weld pool behind
the keyhole. The maximum velocities for the molten metal flows near the top and bottom
 Figure 9 shows 3D convective patterns in the weld pool. The result showed that two
eddy convective patterns in opposite directions were stably formed in the weld pool be-
Materials 2024, 17, 1348 hind the keyhole. The maximum velocities for the molten metal flows near the top and
14 of 24
bottom surfaces reached 0.35 and 0.30 m/s, respectively. The velocity of molten metal flow
around the middle height decreased to 0.15 m/s.
Thereached
surfaces dominant driving
0.35 and forcerespectively.
0.30 m/s, and heat transport process
The velocity considered
of molten based
metal flow aroundon the
above comprehensive measurement are
the middle height decreased to 0.15 m/s. discussed in Section 4.2.

Figure9.9.Three-dimensional
Figure Three-dimensional convective
convective patterns
patterns inside
inside the weld
the weld poolpool in KPAW
in KPAW of stainless
of stainless steel
steel [74].
[74].
The dominant driving force and heat transport process considered based on the above
comprehensive measurement
Xu et al. clarified the 3Dare discussed
keyhole in Section
detouring the4.2.
flow inside the weld pool in VPPAW
Xu et al. clarified the 3D keyhole detouring
of an aluminum alloy employing two sets of X-ray the flow inside the weld
transmission pool in
systems VPPAW
[75]. Further-
of an aluminum alloy employing two sets of X-ray transmission systems [75]. Furthermore,
more, measurement of the surface flow of the weld pool at different regions of the keyhole
measurement of the surface flow of the weld pool at different regions of the keyhole was
was also carried out by observing tracers. As a result, it was shown that a thin liquid metal
also carried out by observing tracers. As a result, it was shown that a thin liquid metal
layer was formed on the keyhole boundary because of the high thermal conductivity of
layer was formed on the keyhole boundary because of the high thermal conductivity of the
the aluminum
aluminum alloy.alloy. Moreover,
Moreover, themetal
the liquid liquidwasmetal
mainly wasformed
mainly formed
around thearound
keyholethe keyhole
bottom.
bottom. Figure 10 shows 3D convective patterns in the weld pool. The
Figure 10 shows 3D convective patterns in the weld pool. The liquid metal flowed upward liquid metal flowed
upward
and andthe
around around
keyhole.theThe
keyhole. The
molten molten
metal metal
at the frontatwall
theof
front
the wall of the
keyhole keyhole
around around
bottom
bottom detoured
detoured the keyhole thebackward.
keyhole backward.
After that, itAfter that, it was
was blanched intoblanched
upward and into upward and
downward
downward
flows, formingflows, forming
a vertical a vertical
separation separation
point. In addition, point. In addition,
comparing comparing
with the stainless with
steel the
stainless
case [74], steel casevelocity
the flow [74], theof flow
VPPAWvelocity of VPPAW
was found was found
to be much to be much smaller.
smaller.

Figure 10. Three-dimensional

Figure 10. Three-dimensionalkeyhole
keyholedetouring flow
detouring inside
flow the the
inside weld poolpool
weld in VPPAW of aluminum
in VPPAW of alumi-
num alloys
alloys [75]. [75].

4.2. Dominant Driving Force of Weld Pool Convection and Heat Transport Process in Weld Pool
Following the experimental results on KPAW of stainless steel [74] described in
Section 4.1, the dominant driving force of the weld pool convection in KPAW and its
influence on the heat transport in the weld pool are discussed. As in the TIG welding
Materials 2024, 17, x FOR PEER REVIEW 15 of 25

4.2. Dominant Driving Force of Weld Pool Convection and Heat Transport Process in Weld Pool
Materials 2024, 17, 1348 Following the experimental results on KPAW of stainless steel [74] described in15Sec- of 24
tion 4.1, the dominant driving force of the weld pool convection in KPAW and its influence
on the heat transport in the weld pool are discussed. As in the TIG welding case [41], the
case [41],
main the forces
driving main driving forces
of the weld of convection
pool the weld pool convection
in KPAW in KPAW Lorentz
are buoyancy, are buoyancy,
force,
Lorentz force,
Marangoni Marangoni
force, and shear force,
forceand shear force
especially in aespecially in astate
quasi-steady quasi-steady state
after stable afterkey-
open sta-
ble open
hole keyhole
formation. formation.
Figure Figureconvections
11 presents 11 presentsobtained
convections obtained
by the by the measurement
measurement and also the
and also the
estimated estimated
driving forcesdriving
acting forces
on eachacting on each convection.
convection.

Figure
Figure11.
11.Schematic
Schematicillustration of of
illustration thethe
main driving
main forces
driving andand
forces convective patterns
convective of theofweld
patterns pool
the weld
[74].
pool [74].

The buoyancy
The buoyancy is weakerweaker comparing
comparingwith withother
otherforces
forcesand
andacts inin
acts anan
upward
upward direction.
direc-
Therefore,
tion. Therefore,this force is considered
this force is consideredto be tonegligible for weld
be negligible pool pool
for weld formation.
formation.
TheLorentz
The Lorentzforce forceisisgenerally
generally in inan an inward
inward andand downward
downward direction
direction around around the
the sur-
surface of the weld pool, since the current
face of the weld pool, since the current expands conically. expands conically.
The shear
The shear force
force acting
acting onon the
the surface
surface of of the
the weld
weld pool
pool is
is composed
composed of of two
two forces.
forces. The
The
first one is caused by plasma flowing along the top weld pool surface. The directionthis
first one is caused by plasma flowing along the top weld pool surface. The direction of of
force
this is mainly
force is mainly upward
upward andandoutward,
outward, inducing
inducingthe the
clockwise
clockwiseconvection
convectionaround
around the the
top
surface.
top surface. TheThe second oneone
second is caused
is caused by plasma
by plasma flowing
flowingoutout
downward
downward through
through thethekeyhole
key-
exit. The direction of the force is downward, forming the counterclockwise
hole exit. The direction of the force is downward, forming the counterclockwise convec- convection
around
tion aroundthe bottom surface.
the bottom surface.
The Marangoni
The Marangoni force forceisisproduced
producedby bythethegradient
gradientof ofweld
weldpool
poolsurface
surfacetemperature.
temperature.
This force is generally directed from high temperature to
This force is generally directed from high temperature to low temperature in case low temperature in case ofof inert
inert
shielding gas. On the other hand, if an active gas like oxygen
shielding gas. On the other hand, if an active gas like oxygen is mixed a little with the is mixed a little with the
shielding gas,
shielding gas, asas expected
expected in in region
region 3, 3, the
the force
force direction
direction isis reversed
reversed [76].
[76].
In region 1, the molten metal flowed backward
In region 1, the molten metal flowed backward from behind the keyhole from behind the keyhole to to region
region 2
2 at a high temperature. The backward shear force was found
at a high temperature. The backward shear force was found to surpass the resultant Ma- to surpass the resultant
Marangoni and Lorentz forces in the forward direction, therefore considered the dominant
rangoni and Lorentz forces in the forward direction, therefore considered the dominant
force. The maximum velocity appeared in the backside of the keyhole. It decreased as
force. The maximum velocity appeared in the backside of the keyhole. It decreased as
region 2 was approached.
region 2 was approached.
In region 3, the molten metal flowed forward from the weld pool end to region 2 at a
In region 3, the molten metal flowed forward from the weld pool end to region 2 at a
high temperature. The forward Marangoni force and backward shear force are considered
high temperature. The forward Marangoni force and backward shear force are considered
to mainly act in this region, but the shear force is thought to be lower than the Marangoni
to mainly act in this region, but the shear force is thought to be lower than the Marangoni
force, because the velocity of plasma flow apart from the arc is largely decreased. Therefore,
the molten metal flowed in the forward direction.
In region 2, the molten metal flowing from regions 1 backward and 3 forward merged.
On the bottom surface, the molten metal flowed backward from the keyhole to the
weld pool end. The flow was accelerated by the downward shear force acting on the
 force, because the velocity of plasma flow apart from the arc is largely decreased. There-
fore, the molten metal flowed in the forward direction.
In region 2, the molten metal flowing from regions 1 backward and 3 forward
merged.
Materials 2024, 17, 1348
On the bottom surface, the molten metal flowed backward from the keyhole 16 toofthe
24
weld pool end. The flow was accelerated by the downward shear force acting on the key-
hole wall and backward Marangoni force on the bottom surface of the weld pool. Because
keyhole
of wall and by
the acceleration backward Marangoni
both forces, force onon
the flow velocity thethebottom
bottomsurface
surfaceof was the weldhigher
much pool.
Because of the acceleration
than that on the top surface. by both forces, the flow velocity on the bottom surface was
muchInside
higher than that on the top surface.
the weld pool, two eddies were seen behind the keyhole. The flow in the upper
eddyInside
was inthe weld pool,direction.
a clockwise two eddiesThis were seenwas
eddy behind the keyhole.
accelerated Theupward
by the flow in andthe upper
back-
eddy was in a clockwise direction. This eddy was accelerated by
ward shear forces and the forward Marangoni force. Because the shear force is thought tothe upward and backward
shear
be forces andlarger
significantly the forward
than the Marangoni
Marangoniforce. force,Because
this eddy theisshear force
mainly is thoughtbytothe
accelerated be
significantly
shear force. Thelarger
flow than thelower
in the Marangoni
eddy wasforce, this eddyaccelerated
essentially is mainly accelerated
by the downwardby the shear
shear
force. The flow in the lower eddy was essentially accelerated
force and the downward and backward Marangoni forces. Therefore, this eddy flowed by the downward shear force
in
and the downward and
a counterclockwise direction. backward Marangoni forces. Therefore, this eddy flowed in a
counterclockwise direction.
The formation process of the weld pool is also investigated in other arc welding pro-
The formation process of the weld pool is also investigated in other arc welding
cesses through the measurement of velocity field of weld pool convection by tracer parti-
processes through the measurement of velocity field of weld pool convection by tracer
cle observation. For instance, the velocity field of the weld pool surface in TIGW was meas-
particle observation. For instance, the velocity field of the weld pool surface in TIGW was
ured by Matsuda et al. [77]. They clarified that the velocity reached around 0.04 m/s
measured by Matsuda et al. [77]. They clarified that the velocity reached around 0.04 m/s
around the foot point of the arc axis. Zong et al. measured the velocity field in GMAW and
around the foot point of the arc axis. Zong et al. measured the velocity field in GMAW
found that the range of velocity was about 0.15–0.20 m/s around the arc and 0.04 m/s
and found that the range of velocity was about 0.15–0.20 m/s around the arc and 0.04 m/s
around the weld pool end [78].
around the weld pool end [78].
The weld pool formation process of KPAW is considered to be greatly different from
The weld pool formation process of KPAW is considered to be greatly different from
those of TIGW and GMAW. The shear force, Lorentz force, and Marangoni force are
those of TIGW and GMAW. The shear force, Lorentz force, and Marangoni force are known
known to play a major role as driving forces in TIGW and GMAW cases. In case of KPAW,
to play a major role as driving forces in TIGW and GMAW cases. In case of KPAW, the
the magnitudes
magnitudes of the ofLorentz
the Lorentz
forceforce and Marangoni
and Marangoni force
force are are predicted
predicted to havetoalmost
have similar
almost
similar levels with those of TIGW and GMAW in the same current
levels with those of TIGW and GMAW in the same current condition. On the other hand, condition. On the other
hand, the force
the shear shearisforce is considered
considered to be muchto belarger
muchcompared
larger compared
with those withofthose
TIGWofand TIGWGMAWand
GMAW due to the greater velocity of plasma flow [47], becoming
due to the greater velocity of plasma flow [47], becoming the dominant force. Consequently, the dominant force.
Consequently, the velocityflow
the velocity of convective of convective
surpasses flow
thosesurpasses
of TIGW those of TIGW and GMAW.
and GMAW.
Wu et al. quantitatively analyzed the weld
Wu et al. quantitatively analyzed the weld pool formation pool formation process
process through
through numerical
numeri-
simulation,
cal simulation, presenting
presenting thatthat
afterafter
openopen
keyhole
keyholeformation,
formation,two two
convective
convectiveeddies are pro-
eddies are
duced behind
produced behind thethekeyhole
keyhole as as
in in
Figure
Figure1212[60].
[60].InInthe
theweld
weldpool,
pool,the
the molten
molten metal was
metal was
transported
transported backward
backward on on the
the top
top surface
surface by the upper
by the upper eddy
eddy and
and backward
backward at at the
the bottom
bottom
surface by the lower eddy. The calculated weld pool convections
surface by the lower eddy. The calculated weld pool convections well agreed with the well agreed with the
above observation.
above observation.

Figure 12. Calculated temperature and fluid flow fields [60].

Wu et
Wu et al.
al.carried
carriedout
outa numerical
a numericalsimulation to investigate
simulation the heat
to investigate transport
the heat in a weld
transport in a
pool with an open keyhole in a quasi-steady state [62]. Figure 13 shows the
weld pool with an open keyhole in a quasi-steady state [62]. Figure 13 shows the 3D3D temperature
field of the weld pool. Although the weld pool temperature is found to increase around the
keyhole, the maximum temperature is only 1857 K, which is slightly higher than the melting
point of steel. The heat flux from arc to base metal in the KPAW process is large due to the
constricted arc, but the weld pool temperature is low. As presented by Nguyen et al. [74],
the maximum surface temperature of the weld pool was approximately 1800 K, which was
almost the same as that in the moving GTAW [79] and lower than that in the stationary
GTAW [80]. The reasons to cause this low temperature can be explained as follows.
 higher than the melting point of steel. The heat flux from arc to base metal in the KPAW
process is large due to the constricted arc, but the weld pool temperature is low. As pre-
sented by Nguyen et al. [74], the maximum surface temperature of the weld pool was
approximately 1800 K, which was almost the same as that in the moving GTAW [79] and
Materials 2024, 17, 1348 lower than that in the stationary GTAW [80]. The reasons to cause this low temperature
17 of 24
can be explained as follows.

Figure
Figure 13.
13. Three-dimensional
Three-dimensional temperature
temperature field
field in quasi-steady state
in quasi-steady state [62].
[62].

Heat transport in the weld weld pool

pool isis carried
carried out
out mainly
mainly byby convection
convection and
and conduction.
conduction.
The Peclet number is useful for discussingdiscussing the dominant
dominant heat transport mechanism. As
heat transport mechanism.
seen in Figures
Figures 99 and
and12,
12,two
twoconvective
convectiveeddies
eddieswerewereformed
formedbehind
behind the keyhole
the keyhole consisting
consist-
of the
ing upper
of the upper clockwise
clockwise eddy
eddyandandlower
lower counterclockwise
counterclockwiseeddy. eddy.The
Themolten
molten metal
metal was
transported backward
transported backward on on the
the top
top and
and bottom surfaces. In
bottom surfaces. In the
the above
above result,
result, aa maximum
maximum
velocity of 0.2884 m/s was seen in the weld pool. The calculated
velocity of 0.2884 m/s was seen in the weld pool. The calculated Peclet number from Peclet number from the
the
velocity was approximately 319. Wang et al. calculated a Peclet number
velocity was approximately 319. Wang et al. calculated a Peclet number of the weld pool of the weld pool
of about
of about 77
77 inin double
double electrodes
electrodes TIGWTIGW [81]. Generally, when
[81]. Generally, when the
the Peclet
Peclet number
number exceeds
exceeds
10, the effect of heat transport due to convection is considered to
10, the effect of heat transport due to convection is considered to be dominant. The be dominant. The result
result
indicates that the heat transport by fluid flow is the dominant mechanism
indicates that the heat transport by fluid flow is the dominant mechanism in the KPAW in the KPAW
weld pool.
weld pool. As
As aa result,
result, the
the energy
energy is is not
not accumulated
accumulated in in the
the weld
weldpool
poolaround
around the
thekeyhole,
keyhole,
leading to a lower weld pool temperature.
leading to a lower weld pool temperature.
4.3. Effect of Torch Design and Operating Parameters
4.3. Effect of Torch Design and Operating Parameters
In the above paper [74], only one experimental condition was performed. However,
In the above paper [74], only one experimental condition was performed. However,
it is considered that the driving force balance to govern the weld pool formation process
it is considered that the driving force balance to govern the weld pool formation process
significantly depends on the torch design and operating parameters. Because the shear
significantly depends on the torch design and operating parameters. Because the shear
force is strongly affected by the plasma flow field in the vicinity of the keyhole, the force
force is strongly affected by the plasma flow field in the vicinity of the keyhole, the force
balance is thought to greatly change depending on the keyhole shape. As shown in Figure 9,
balance
two large is eddies
thoughtare toproduced
greatly change
behind depending
the keyhole on in
thethis
keyhole shape. As
experimental shown inbecause
condition, Figure
9, two large eddies are produced behind the keyhole in this experimental
the plasma flow from the torch is separated to a downward flow outflowing to the bottom condition, be-
cause the plasma flow from the torch is separated to a downward
side through the keyhole exit and a horizontal and upward flow along the weld pool flow outflowing to the
bottom
surface side through
to form the keyhole
two opposite shear exit and aFor
forces. horizontal
example,and upward
in the flow
case of along
higher the weld
plasma gas
pool
flow surface to form two
rate or current, opposite
the arc pressure shear forces.
acting on For
the example,
surface ofinthe theweld
case pool
of higher plasma
is enhanced,
gas flowtorate
leading or current,
a wider keyholethe arc pressure
diameter actingkeyhole
and a steeper on the surface of the
wall. In this weld
case, the pool is en-
downward
hanced, leading to a wider keyhole diameter and a steeper keyhole
flow through the keyhole exit might be larger than the horizontal and upward flow. The wall. In this case, the
downward flow through the keyhole exit might be larger than the
formation of eddies is expected to be affected by an increase in shear force, making the horizontal and upward
flow.
lowerThe eddy formation
relatively of larger.
eddies is It expected
also causes to be affected by
a variation inantheincrease in shear in
heat transport force,
the mak-
weld
ing
pool.theNguyen
lower eddy et al.relatively
studied the larger. It of
effect also causesgas
plasma a variation
flow ratein[67],
the welding
heat transport
current in[68],
the
weld
and alsopool. theNguyen
shielding et al.
gasstudied the effect
composition [69],ofand
plasma
Xu etgas flow ratethe
al. clarified [67], welding
effects current
of electrode
[68],
setback and[70]also
andthe shielding
plate thicknessgas[71]composition
in VPPAW of [69], and Xu alloys.
aluminum et al. clarified the effects
Those studies used theof
same comprehensive experimental approach.
Figure 14 shows 3D convective patterns inside the weld pool for different plasma gas
flow rates [67]. The weld bead appearances and cross sections of the weld beads were
already presented in Figure 4. As the plasma gas rate is increased, a penetrated keyhole is
formed. The counterclockwise eddy inside the weld pool becomes large and then dominant.
The backward flow on the top surface is weakened, while the inward flow is induced by
the teardrop-shaped top weld pool profile. According to the dominant counterclockwise
eddy inside the weld pool, the molten metal temperature at the lateral sides of the top weld
pool is low.

Materials 2024, 17, 1348
As described above, convection is the dominant mechanism for the heat transport
inside the weld pool in KPAW. The strong counterclockwise eddy formed behind the key-
hole and the weak backward flow on the top surface of the weld pool produce the uneven
energy distribution between the top and bottom surfaces. The strong inward flow causes
the uneven energy distribution between the lateral sides and centerline of the top 18
weld
of 24
pool. All these reasons contribute to the undercut formation at the top surface.

Figure14.
Figure Three-dimensional convective
14.Three-dimensional convective patterns
patterns inside
inside the weld pool for different
different plasma gas
gas flow
rates:rates:
flow (A) 0.7
(A)L/min; (B) 1.7
0.7 L/min; (B) L/min; (C)(C)
1.7 L/min; 3.0 3.0
L/min [67].
L/min [67].

As described
Figure 15 shows above, convectionpatterns
3D convective is the dominant
inside themechanism
weld pool for for the heat transport
different currents
inside the weld pool in KPAW. The strong counterclockwise eddy
[68]. The weld bead appearances and cross sections of the weld beads have already been formed behind the
keyhole and the weak backward flow on the top surface of the weld
presented in Figure 5. In case of a low current of 90 A, an open keyhole was formed with pool produce the
auneven
smallerenergy
diameterdistribution between
on the bottom the top
surface and
than bottom
that surfaces.
on the The strong
top surface. This isinward
due toflow
the
causes the uneven energy distribution between the lateral sides and
insufficient penetration ability and is considered to produce a strong upward shear centerline of the top
force.
weld pool. the
Therefore, All upper
these reasons contribute
eddy grew up more to the undercutwith
compared formation at the
the lower topOn
one. surface.
the other
Figure 15 shows 3D convective patterns inside the weld pool for different
hand, in case of a high current of 150 A, an open keyhole was formed with a large diameter currents [68].
The weld bead appearances and cross sections of the weld beads
on the bottom surface. The lower eddy was enhanced due to a strong downward shear have already been pre-
force. This relative magnitude of the two eddies is thought to significantly affect the heata
sented in Figure 5. In case of a low current of 90 A, an open keyhole was formed with
smaller diameter on the bottom surface than that on the top surface. This is due to the
transport in the weld pool and also the welding defect occurrence. The stronger upper
insufficient penetration ability and is considered to produce a strong upward shear force.
eddy due to the insufficient heat input might cause an undercut and high reinforcement
Therefore, the upper eddy grew up more compared with the lower one. On the other hand,
on the top surface. On the contrary, the stronger lower eddy leads to deeper penetration
in case of a high current of 150 A, an open keyhole was formed with a large diameter on the
by the large heat transport in a downward direction. However, if the downward heat
bottom surface. The lower eddy was enhanced due to a strong downward shear force. This
transport is excessive, it leads to welding defects such as concaves on the top surface and
relative magnitude of the two eddies is thought to significantly affect the heat transport in
burn-throughs on the bottom surface. Consequently, it is considered that when the arc has
the weld pool and also the welding defect occurrence. The stronger upper eddy due to the
sufficient penetration ability to the base metal, the relative magnitude of the two eddies
insufficient heat input might cause an undercut and high reinforcement on the top surface.
should be suitably controlled for achieving the larger counterclockwise eddy and smaller
On the contrary, the stronger lower eddy leads to deeper penetration by the large heat
clockwise eddy, which enables achievement of a stable and high-quality welding prevent-
transport in a downward direction. However, if the downward heat transport is excessive,
ing welding defect occurrence, as in the optimal current of 120 A in this experiment.
it leads to welding defects such as concaves on the top surface and burn-throughs on the
bottom surface. Consequently, it is considered that when the arc has sufficient penetration
ability to the base metal, the relative magnitude of the two eddies should be suitably
controlled for achieving the larger counterclockwise eddy and smaller clockwise eddy,
Materials 2024, 17, x FOR PEER REVIEW
which enables achievement of a stable and high-quality welding preventing welding19defect of 25

occurrence, as in the optimal current of 120 A in this experiment.

different currents [68].

Figure 15. Three-dimensional convective patterns inside the weld pool for different

Figure 16 shows 3D convective patterns inside the weld pool for different shielding
gas compositions [69]. The weld bead appearances and cross sections of the weld beads
have already been presented in Figure 6. It was presented that in the case of pure Ar
shielding gas, the molten metal flowed upward behind the keyhole. In the case of Ar with
0.5% O2 shielding gas, the flow direction became downward behind the keyhole. The mol-
ten metal flow in the weld pool is varied depending on the keyhole shape, for example,
Materials 2024, 17, 1348 19 of 24
Figure 15. Three-dimensional convective patterns inside the weld pool for different currents [68].

Figure 16 shows 3D convective patterns inside the weld pool for different shielding
Figure 16 shows 3D convective patterns inside the weld pool for different shielding gas
gas compositions [69]. The weld bead appearances and cross sections of the weld beads
compositions [69]. The weld bead appearances and cross sections of the weld beads have
have already been presented in Figure 6. It was presented that in the case of pure Ar
already been presented in Figure 6. It was presented that in the case of pure Ar shielding
shielding gas, the molten metal flowed upward behind the keyhole. In the case of Ar with
gas, the molten metal flowed upward behind the keyhole. In the case of Ar with 0.5%
0.5% O2 shielding gas, the flow direction became downward behind the keyhole. The mol-
O2 shielding gas, the flow direction became downward behind the keyhole. The molten
ten metal flow in the weld pool is varied depending on the keyhole shape, for example,
metal flow in the weld pool is varied depending on the keyhole shape, for example, by
by the keyhole diameter and keyhole wall inclination angle. The keyhole diameter is large
the keyhole diameter and keyhole wall inclination angle. The keyhole diameter is large
on
onthe
thetop
topside
sideandandsmall
smallon onthe
the bottom
bottomside sidein
inthe
thepure
pureAr Arcase,
case,but
butititisis smaller
smallerononthe
the
top
topside
sideand
andlarger
larger onon the
the bottom
bottom side
side in
in the
the Ar
Ar mixed
mixed withwith 0.5%0.5% OO22case.
case.TheThemagnitude
magnitude
and
anddirection
directionofofthe
theshear
shearforce
forceare
are influenced
influenced by by this
this difference.
difference. As Asaaresult,
result,ititwas
wasfound
found
that
that the addition of a small amount of oxygen into Ar shielding gas assuming shielding
the addition of a small amount of oxygen into Ar shielding gas assuming shielding
failure
failurestrongly
stronglyaffected
affected the
theheat
heattransport
transportprocess.
process. ItItisisalso
alsoimplied
impliedthat
thatnarrow
narrowand anddeep
deep
penetration
penetration can be achieved by mixing oxygen slightly such as in the AA-TIG welding
can be achieved by mixing oxygen slightly such as in the AA-TIG welding
process
process [82].
[82].This
Thismechanism
mechanismisisconsidered
consideredto tobe
berelated
relatedprimarily
primarilyto tothe
thevariation
variationin inshear
shear
force because of different keyhole shape formation due to change in
force because of different keyhole shape formation due to change in the surface tension the surface tension
rather
ratherthan
thanthe
thevariation
variationin inthe
themagnitude
magnitudeand anddirection
direction of of the
the Marangoni
Marangoni force.force.

Figure16.
Figure Three-dimensionalconvective
16.Three-dimensional convectivepatterns
patterns inside
inside the
the weld
weld pool
pool forfor different
different shielding
shielding gasgas
compositions[69].
compositions [69].

4.4. Effect on Microstructure of Solidified Metal in Welded Joint

4.4. Effect on Microstructure of Solidified Metal in Welded Joint
As mentioned above, the heat and mass transport within the weld pool governs the
As mentioned
formation above, thefield
of the temperature heatin
and
themass
weldtransport
pool and within
therefore thehas
weld pool governs
a strong influencethe
on
formation of the temperature field in the weld pool and therefore has
the microstructure of solidified metal, mainly according to a temperature historya strong influence
during
on
thethe microstructure
cooling of solidified
process, thus affecting metal, mainly
mechanical according
properties of to
thea welded
temperature
joint. history dur-
ing theThere are many previous studies on the microstructure in KPAW. Forjoint.
cooling process, thus affecting mechanical properties of the welded example, mi-
There
crostructuresare when
many applying
previous studies
KPAW on the microstructure
to various materials havein KPAW. For example,
been reported, suchmi-
as
crostructures when applying KPAW to various materials have been reported,
duplex stainless steel [83,84], Inconel [85], aluminum alloy [86], and titanium alloy [87]. such as
Comparative studies with microstructures in other arc welding processes have also been
conducted. Gupta et al. compared the welded joints prepared by a single-pass KPAW and
multi-pass GATAW of M 250 (Maraging steel) with a thickness of 8 mm, showing improved
mechanical and metallurgical properties for KPAW in comparison to those for GTAW [88].
Welding parameters such as current waveform also affect the microstructure [85,87]. Kumar
et al. found that pulse current helps in grain refinement, which leads to a higher ultimate
tensile strength of the welded joints [87]. Large current and high peak current lead to more
heat and a longer cooling time, resulting in coarse grain. When the pulse rate is increased,
the tensile strength is also increased.
The above studies basically demonstrate the relationship between welding conditions
and microstructures experimentally. Although these results are very useful industrially,
its mechanism has not been fully elucidated academically. In particular, there are very
few studies on the formation process of the microstructure in welded joints based on
Materials 2024, 17, 1348 20 of 24

quantitative discussions of heat and mass transport within the weld pool. Yan et al.
investigated the gravity effect on the mechanical property of VPPAW of aluminum alloys
at vertical and horizontal positions [89]. Microstructures and residual stress in the welded
joint for vertical-up welding were symmetrical. This is in contrast to horizontal welding,
which tends to have asymmetric and inhomogeneous microstructures that lead to poor
mechanical properties. In the upper side of welded joints for horizontal welding, larger
grain sizes brought about lower hardness and lower tensile strength. The difference in the
mechanical properties between bilateral sides was suggested to be caused by gravity driving
the melting metal flow. After that, Liu et al. carried out the measurements of temperature
field and velocity field on the weld pool surface of VPPAW at different positions to discuss
the effect of heat and mass transport in the weld pool on microstructure and also mechanical
properties of the welded joint through the gravity force [90]. From the measurement, they
found that the asymmetric flow of molten metal under the influence of gravity is the main
factor leading to uneven temperature distribution and asymmetric grain distribution in the
weld pool. Furthermore, Xu et al. clarified the relationship between the microstructure of
the welded joint and the heat and mass transport in VPPAW of thick aluminum alloy plates
at a flat position by combining in situ three-dimensional X-ray imaging and multi-physics
modeling [71]. Figure 17 shows the microstructure and crystal size of the weld bead. It
was concluded that the large crystal size observed in the lower layer of the weld is partly
caused by heat treatment from the upper layer of the thick plate. An eddy with a high
Materials 2024, 17, x FOR PEER REVIEW 21 of 25
flow velocity to the rear of the weld pool destroys the crystal-growth process, and this is
considered to be one of the reasons for fine crystals appearing in the upper part of the weld.

Figure Microstructureand
Figure 17. Microstructure andcrystal
crystalsize
sizeofofthe
the weld
weld bead
bead in VPPAW
in VPPAW of aofthick
a thick aluminum
aluminum plateplate
[71].
[71].
In this way, it has been found that the formation of the microstructure is greatly
5. Summary
affected andflow
by the Futureof Work
the weld pool and the accompanying heat transport. Also, in
KPAW, a filler wire is often used when
In this article, the interaction the baseofmaterial
mechanism the arc, plate thickness
keyhole, is large.
and weld pool Especially
in KPAW
when the chemical composition of this filler wire is different from
was reviewed. In KPAW, two large eddies are constantly formed in the weld pool that of the basebehind
metal,
the keyhole by plasma shear force as the dominant driving force. These govern the pool
non-uniform element distribution tends to occur within the weld pool due to weld heat
flow [27], process
transport which isinalso
thethought
weld poolto affect
and havethe formation of the microstructure.
a strong influence on the weld poolElucidation of
formation
the influence
process. of thepool
The weld weld pool
flow formation
velocity process
is much on than
faster the formation of thewelding
those of other microstructure of
processes
the solidified metal in the welded joint is still insufficient and remains
such as TIG welding and GMA welding, enhancing the heat transport to lower the weld a future work.
pool surface temperature. Since the strength and direction of this shear force strongly de-
5. Summary and Future Work
pend on the keyhole shape, it is possible to control the weld pool formation process
In this
through article,the
changing thekeyhole
interaction mechanism
shape of the
by adjusting arc,
the keyhole,
torch designand
andweld pool inparam-
operating KPAW
was reviewed. In KPAW, two large eddies are constantly formed in the weld pool behind
eters. If the lower eddy is relatively stronger, the heat transport to the bottom side in-
the keyhole by plasma shear force as the dominant driving force. These govern the heat
creases and the penetration increases. However, burn-through is more likely to occur, and
heat transport to the top side decreases, causing undercut. When a small amount of oxy-
gen is added to the shielding gas, the surface tension of the keyhole decreases, making it
easier to form a keyhole.
In the above studies, most of them were carried out only for a flat position for sim-
plicity. However, KPAW can also be applied for welding of a pipe or flange that is per-
Materials 2024, 17, 1348 21 of 24

transport process in the weld pool and have a strong influence on the weld pool formation
process. The weld pool flow velocity is much faster than those of other welding processes
such as TIG welding and GMA welding, enhancing the heat transport to lower the weld
pool surface temperature. Since the strength and direction of this shear force strongly de-
pend on the keyhole shape, it is possible to control the weld pool formation process through
changing the keyhole shape by adjusting the torch design and operating parameters. If the
lower eddy is relatively stronger, the heat transport to the bottom side increases and the
penetration increases. However, burn-through is more likely to occur, and heat transport
to the top side decreases, causing undercut. When a small amount of oxygen is added to
the shielding gas, the surface tension of the keyhole decreases, making it easier to form
a keyhole.
In the above studies, most of them were carried out only for a flat position for simplicity.
However, KPAW can also be applied for welding of a pipe or flange that is performed in
various welding positions. Accordingly, the direction of gravity acting on the weld pool
largely changes during welding. It could critically affect the keyhole formation process,
so the interaction mechanism of the arc, keyhole, and weld pool is predicted to change
significantly. Furthermore, the effect of the interaction on various welding defects must be
investigated more in detail. For example, porosity occurrence is also required to be solved.
If gases such as hydrogen are absorbed into the weld pool during welding, those form
bubbles. When the bubbles solidify without being released from the weld pool due to flow
phenomena including buoyancy, a welding defect called porosity occurs. This is particularly
likely to be formed when welding aluminum alloys, causing stress concentration to induce
cracks. In order to suppress the porosity, it is necessary to promote the discharge of bubbles
out of the weld pool by suitably controlling heat and mass transport processes in weld
pool based on the understating of the interaction mechanism. Elucidation of the influence
of the weld pool formation process on the formation of the microstructure of solidified
metal in the welded joint is still insufficient. The above issues should also be addressed in
future works.

Funding: This work was supported by JSPS KAKENHI (Grant Number JP21K04710), and the Project
on Design & Engineering by Joint Inverse Innovation for Materials Architecture (DEJI2MA) from the
Ministry of Education, Culture, Sports, Science and Technology (MEXT).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Data are contained within the article.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Lucas, W. TIG and Plasma Welding, 3rd ed.; Abington Publishing: Cambridge, UK, 2008; pp. 80–108.
2. Murphy, A.B.; Lowke, J.J. Heat Transfer in Arc Welding. In Handbook of Thermal Science and Engineering; Springer: New York, NY,
USA, 2018; pp. 2657–2727.
3. Li, Z.; Wu, D.; Tashiro, S.; Tanaka, M.; Xin, J.; Wang, H.; Hua, X. Influence mechanism of metal vapor in plasma arc lap welding.
Weld. J. 2022, 101, 161s–171s. [CrossRef]
4. Nguyen, H.L.; Van Nguyen, A.; Duy, H.L.; Nguyen, T.-H.; Tashiro, S.; Tanaka, M. Relationship among Welding Defects with
Convection and Material Flow Dynamic Considering Principal Forces in Plasma Arc Welding. Metals 2021, 11, 1444. [CrossRef]
5. Sakuramoto, Y.; Wada, K.; Kanemaru, S.; Kamei, T. Application of Plasma Welding Process for Manufacturing of Stainless Steel
Cylindrical Vessel. In TAIYO NIPPON SANSO Technical Report; Taiyo Nippon Sanso: Tokyo, Japan, 2013; Volume 32, pp. 33–34.
(In Japanese)
6. Wang, L.L.; Lu, F.G.; Wang, H.P.; Murphy, A.B.; Tang, X.H. Effects of shielding gas composition on arc profile and molten pool
dynamics in gas metal arc welding of steels. J. Phys. D Appl. Phys. 2014, 47, 465202. [CrossRef]
7. Martikainen, J.K.; Moisio, T.J.I. Investigation of the effect of welding parameters on weld quality of plasma arc keyhole welding
of structural steels. Weld. J. 1993, 72, 329s–340s. Available online: https://app.aws.org/wj/supplement/WJ_1993_07_s329.pdf
(accessed on 11 March 2024).
Materials 2024, 17, 1348 22 of 24

8. Banas, C.M. Electron Beam, Laser Beam and Plasma Arc Welding Studies NASA Contractor Report NASA-CR-132386; United Aircraft
Research Laboratories: East Hartford, CT, USA, 1974.
9. Vilkas, E.P. Plasma arc welding of exhaust pipe system components. Weld. J. 1991, 70, 49–52.
10. Wu, C.S.; Wang, L.; Ren, W.J.; Zhang, X.Y. Plasma arc welding: Process, sensing, control and modeling. J. Manuf. Process. 2014, 16,
74–85. [CrossRef]
11. Liu, Z.M.; Cui, S.L.; Luo, Z.; Zhang, C.Z.; Wang, Z.M.; Zhang, Y.C. Plasma arc welding: Process variants and its recent
developments of sensing, controlling and modeling. J. Manuf. Process. 2016, 23, 315–327. [CrossRef]
12. Sahoo, A.; Tripathy, S. Development in plasma arc welding process: A review. Mater. Today Proc. 2021, 41, 363–368. [CrossRef]
13. Zhang, S.B.; Zhang, Y.M. Efflux plasma charge-based sensing and control of joint penetration during keyhole plasma arc welding.
Weld. J. 2001, 80, 157s–162s. Available online: http://files.aws.org/wj/supplement/WJ_2001_07_s157.pdf (accessed on 11
March 2024).
14. Zhang, Y.M.; Zhang, S.B.; Liu, Y.C. A plasma cloud charge sensor for pulse keyhole process control. Meas. Sci. Technol. 2001, 12,
1365–1370. [CrossRef]
15. Zhang, Y.M.; Zhang, S.B. Observation of the keyhole during plasma arc welding. Weld. J. 1999, 78, 53s–58s. Available online:
https://app.aws.org/wj/supplement/Zhang/ARTICLE4.pdf (accessed on 11 March 2024).
16. Liu, Z.M.; Wu, C.S.; Gao, J.Q. Vision-based observation of keyhole geometry in plasma arc welding. Int. J. Therm. Sci. 2013, 63,
38–45. [CrossRef]
17. Zhang, Y.M.; Liu, Y.C. Control of dynamic keyhole welding process. Automatica 2007, 43, 876–884. [CrossRef]
18. Wu, C.S.; Jia, C.B.; Chen, M.A. A control system for keyhole plasma arc welding of stainless steel plates with medium thickness.
Weld. J. 2010, 89, 225s–231s. Available online: http://files.aws.org/wj/supplement/WJ112010_225.pdf (accessed on 11 March
2024).
19. Nunes, A.C. Variable polarity plasma arc welding on the space shuttle external tank. Weld. J. 1984, 63, 27–35.
20. Zhang, Q.L.; Fan, C.L.; Lin, S.B.; Yang, C.L. Novel soft variable polarity plasma arc and its influence on keyhole in horizontal
welding of aluminium alloys. Sci. Technol. Weld. Join. 2014, 19, 493–499. [CrossRef]
21. Li, T.Q.; Yang, X.M.; Chen, L.; Zhang, Y.; Lei, Y.C.; Yan, J.C. Arc behaviour and weld formation in gas focusing plasma arc welding.
Sci. Technol. Weld. Join. 2020, 25, 329–335. [CrossRef]
22. Wu, C.S.; Zhao, C.Y.; Zhang, C.; Li, Y.F. Ultrasonic Vibration Assisted Keyholing Plasma Arc Welding. Weld. J. 2017, 96,
279s–286s. Available online: https://s3.amazonaws.com/WJ-www.aws.org/supplement/WJ_2017_08_s279.pdf (accessed on 11
March 2024).
23. Wang, W.; Yamane, S.; Wang, Q.; Shan, L.; Zhang, X.; Wei, Z.; Yan, Y.; Song, Y.; Numazawa, H.; Lu, J.; et al. Visual sensing and
quality control in plasma MIG welding. J. Manuf. Process. 2023, 86, 163–176. [CrossRef]
24. Skowrońska, B.; Szulc, J.; Bober, M.; Baranowski, B.; Chmielewski, T. Selected properties of RAMOR 500 steel welded joints by
hybrid PTA-MAG. J. Adv. Join. Process. 2022, 5, 100111. [CrossRef]
25. Yu, J.; Wang, B.; Zhang, H.; Wang, Q.; Wei, L.; Chen, P.; He, P.; Feng, J. Characteristics of Magnetic Field Assisting Plasma
GMAW-P. Weld. J. 2020, 99, 25s–38s. [CrossRef]
26. Yu, J.; Zhang, H.; Yang, X.; Teng, Y.; Wang, Q.; Wei, L.; He, P.; Zhang, W. Arc Characteristics and Welding Process of Magnetic
Field Assisting Plasma-GMAW-P. Weld. J. 2021, 100, 1s–12s. [CrossRef]
27. Wu, D.; Ishida, K.; Tashiro, S.; Nomura, K.; Hua, X.; Ma, N.; Tanaka, M. Dynamic keyhole behaviors and element mixing in
paraxial hybrid plasma-MIG welding with a gap. Int. J. Heat Mass Transf. 2023, 200, 123551. [CrossRef]
28. Yuji, T.; Tashiro, S.; Kinoshita, H.; Yasui, K.; Bouno, T.; Wu, Z.; Wu, D.; Poonthong, W.; Rahman, S.A.; Mamat, S.B.; et al. An
investigation on plasma-MIG hybrid welding process of thick plate aluminum. J. Adv. Join. Process. 2024, 9, 100188. [CrossRef]
29. Li, Y.; Feng, Y.; Li, Y.; Zhang, X.; Wu, C. Plasma arc and weld pool coupled modeling of transport phenomena in keyhole welding.
Int. J. Heat Mass Transf. 2016, 92, 628–638. [CrossRef]
30. Boulos, M.I.; Fauchais, P.; Pfender, E. Thermal Plasmas: Fundamentals and Applications; Springer: New York, NY, USA, 1994; pp. 4–5.
31. Trelles, J.P.; Chazelas, C.; Vardelle, A.; Heberlein, J.V.R. Arc Plasma Torch Modeling. J. Therm. Spray Technol. 2009, 18, 728–752.
[CrossRef]
32. Boselli, M.; Colombo, V.; Ghedini, E.; Gherardi, M.; Rotundo, F.; Sanibondi, P. Investigation of Thermal Nonequilibrium in a
Plasma Arc Welding Process: Modeling and Diagnostics. IEEE Trans. Plasma Sci. 2014, 42, 1237–1244. [CrossRef]
33. Pan, J.; Hu, S.; Yang, L.; Chen, S. Numerical analysis of the heat transfer and material flow during keyhole plasma arc welding
using a fully coupled tungsten-plasma-anode model. Acta Mater. 2016, 118, 221–229. [CrossRef]
34. Nomura, K.; Yoshii, K.; Toda, K.; Mimura, K.; Hirata, Y.; Asai, S. 3D measurement of temperature and metal vapor concentration
in MIG arc plasma using a multidirectional spectroscopic method. J. Phys. D Appl. Phys. 2017, 50, 425205. [CrossRef]
35. Ishida, K.; Tashiro, S.; Nomura, K.; Wu, D.; Tanaka, M. Elucidation of arc coupling mechanism in plasma-MIG hybrid welding
process through spectroscopic measurement of 3D distributions of plasma temperature and iron vapor concentration. J. Manuf.
Process. 2022, 77, 743–753. [CrossRef]
36. Murphy, A.B. The effect of metal vapour in arc welding. J. Phys. D Appl. Phys. 2010, 43, 434001. [CrossRef]
37. Tanaka, M.; Yamamoto, K.; Tashiro, S.; Nakata, K.; Yamamoto, E.; Yamazaki, K.; Suzuki, K.; Murphy, A.B.; Lowke, J.J. Time-
dependent calculations of molten pool formation and thermal plasma with metal vapour in gas tungsten arc welding. J. Phys. D
Appl. Phys. 2010, 43, 434009. [CrossRef]
Materials 2024, 17, 1348 23 of 24

38. Jian, X.; Wu, C.S. Influence of Fe vapour on weld pool behavior of plasma arc welding. Acta Metall. Sin. 2016, 52, 1467–1476.
Available online: https://www.ams.org.cn/EN/10.11900/0412.1961.2016.00008 (accessed on 11 March 2024).
39. Ogino, Y.; Hirata, Y.; Kawata, J.; Nomura, K. Numerical analysis of arc plasma and weld pool formation by a tandem TIG arc.
Weld. World 2013, 57, 411–423. [CrossRef]
40. Xu, B.; Tashiro, S.; Jiang, F.; Chen, S.; Tanaka, M. The effect of electrode energy balance on variable polarity plasma arc pressure.
Int. J. Heat Mass Transf. 2019, 145, 118715. [CrossRef]
41. Tanaka, M.; Lowke, J.J. Predictions of weld pool profiles using plasma physics. J. Phys. D Appl. Phys. 2007, 40, R1–R23. [CrossRef]
42. Tanaka, M.; Ushio, M.; Lowke, J.J. Numerical Analysis for Weld Formation Using a Free-Burning Helium Arc at Atmospheric
Pressure. JSME Int. J. Ser. B 2005, 48, 397–404. [CrossRef]
43. Ushio, M.; Tanaka, M.; Lowke, J.J. Anode Melting from Free-Burning Argon Arcs. IEEE Trans. Plasma Sci. 2004, 32, 108–117.
[CrossRef]
44. Xu, B.; Tashiro, S.; Jiang, F.; Chen, S.; Tanaka, M. Effect of Arc Pressure on the Digging Process in Variable Polarity Plasma Arc
Welding of A5052P Aluminum Alloy. Materials 2019, 12, 1071. [CrossRef] [PubMed]
45. Wu, D.; Tashiro, S.; Hua, X.; Tanaka, M. A novel electrode-arc-weld pool model for studying the keyhole formation in the keyhole
plasma arc welding process. J. Phys. D Appl. Phys. 2019, 52, 165203. [CrossRef]
46. Jian, X.; Wu, C.S. Numerical analysis of the coupled arc–weld pool–keyhole behaviors in stationary plasma arc welding. Int. J.
Heat Mass Transf. 2015, 84, 839–847. [CrossRef]
47. Schnick, M.; FÜssel, U.; Spille-Kohoff, A. Numerical Investigations of the Influence of Design Parameters, Gas Composition and
Electric Current in Plasma Arc Welding (PAW). Weld. World 2010, 54, R87–R96. [CrossRef]
48. Lowke, J.J.; Tanaka, M. LTE-diffusion approximation for arc calculations. J. Phys. D Appl. Phys. 2006, 39, 3634–3643. [CrossRef]
49. Hiraoka, K.; Shiwaku, T.; Ohji, T. Determining temperature distributions of gas tungsten arc (TIG) plasma by spectroscopic
methods. Weld. Int. 1997, 11, 688–696. [CrossRef]
50. Mackwooda, A.P.; Craferb, R.C. Thermal modelling of laser welding and related processes: A literature review. J. Phys. D Appl.
Phys. 1999, 32, 2902–2909. [CrossRef]
51. Wu, C.S.; Wang, H.G.; Zhang, Y.M. A New Heat Source Model for Keyhole Plasma Arc Welding in FEM Analysis of the
Temperature Profile. Weld. J. 2006, 85, 284s–291s. Available online: https://app.aws.org/wj/supplement/wj1206-284.pdf
(accessed on 11 March 2024).
52. Wu, C.S.; Hu, Q.X.; Gao, J.Q. An adaptive heat source model for finite-element analysis of keyhole plasma arc welding. Comput.
Mater. Sci. 2009, 46, 167–172. [CrossRef]
53. Fan, H.G.; Kovacevic, R. Keyhole formation and collapse in plasma arc welding. Opt. Laser Technol. 2005, 37, 99–115. [CrossRef]
54. Zhang, T.; Wu, C.S.; Feng, Y. Numerical analysis of heat transfer and fluid flow in keyhole plasma arc welding. Numer. Heat
Transf. Part A 2011, 60, 685–698. [CrossRef]
55. Li, T.Q.; Wu, C.S.; Chen, J. Transient variation of arc heat flux and pressure distribution on keyhole wall in PAW. Weld. World 2016,
60, 363–371. [CrossRef]
56. Li, T.Q.; Chen, L.; Zhang, Y.; Yang, X.M.; Lei, C. Metal flow of weld pool and keyhole evolution in gas focusing plasma arc
welding. Int. J. Heat Mass Transf. 2020, 150, 119296. [CrossRef]
57. Jian, X.; Wu, C.S.; Zhang, G.; Chen, J. A unified 3D model for an interaction mechanism of the plasma arc, weld pool and keyhole
in plasma arc welding. J. Phys. D Appl. Phys. 2015, 48, 465504. [CrossRef]
58. Li, Y.; Su, C.; Zhou, X.; Wu, C. A more precise unified model to describe comprehensive multiphysics and multiphase phenomena
in plasma arc welding. J. Manuf. Process. 2020, 59, 668–678. [CrossRef]
59. Li, Y.; Wang, L.; Wu, C. Simulation of keyhole plasma arc welding with electro-magneto-thermo-hydrodynamic interactions. Int.
J. Adv. Manuf. Technol. 2019, 101, 2497–2507. [CrossRef]
60. Wu, D.; Nguyen, A.V.; Tashiro, S.; Hua, X.; Tanaka, M. Elucidation of the weld pool convection and keyhole formation mechanism
in the keyhole plasma arc welding. Int. J. Heat Mass Transf. 2019, 131, 920–931. [CrossRef]
61. Chen, S.; Xu, B.; Jiang, F. Blasting type penetrating characteristic in variable polarity plasma arc welding of aluminum alloy of
type 5A06. Int. J. Heat Mass Transf. 2018, 118, 293–1306. [CrossRef]
62. Wu, D.; Tashiro, S.; Hua, X.; Tanaka, M. Analysis of the energy propagation in the keyhole plasma arc welding using a novel fully
coupled plasma arc-keyhole-weld pool model. Int. J. Heat Mass Transf. 2019, 141, 604–614. [CrossRef]
63. Wu, D.; Tashiro, S.; Hua, X.; Tanaka, M. Coupled mechanisms of the keyhole, energy transfer and compositional change associated
with the variable polarity plasma arc process. J. Phys. D Appl. Phys. 2021, 54, 115204. [CrossRef]
64. Joseph, A.; Harwig, D.; Farson, D.F.; Richardson, R. Measurement and calculation of arc power and heat transfer efficiency in
pulsed gas metal arc welding. Sci. Technol. Weld. Join. 2003, 8, 400–406. [CrossRef]
65. Wu, C.S.; Liu, Z.M. Dynamic variation of keyhole exit and its inclination in plasma arc welding. Weld. World 2015, 59, 365–371.
[CrossRef]
66. Liu, Z.M.; Wu, C.S.; Liu, Y.K.; Luo, Z. Keyhole Behaviors Influence Weld Defects in Plasma Arc Welding Process. Weld. J. 2015, 94,
281s–290s. Available online: https://app.aws.org/wj/supplement/WJ_2015_09_s281.pdf (accessed on 11 March 2024).
67. Nguyen, A.V.; Wu, D.; Tashiro, S.; Tanaka, M. Undercut Formation Mechanism in Keyhole Plasma Arc Welding. Weld. J. 2019, 98,
204s–212s. Available online: https://s3.amazonaws.com/WJ-www.aws.org/supplement/2019.98.018.pdf (accessed on 11 March
2024).
Materials 2024, 17, 1348 24 of 24

68. Nguyen, A.V.; Tashiro, S.; Ngo, M.H.; Bui, H.V.; Tanaka, M. Effect of the eddies formed inside a weld pool on welding defects
during plasma keyhole arc welding. J. Manuf. Process. 2020, 59, 649–657. [CrossRef]
69. Nguyen, A.V.; Tashiro, S.; Ngo, M.H.; Lee, A.H.; Bui, H.V.; Tanaka, M. Influence of shielding gas composition on molten metal
flow behavior during plasma keyhole arc welding process. J. Manuf. Process. 2020, 53, 431–437. [CrossRef]
70. Xu, B.; Chen, S.; Tashiro, S.; Jiang, F.; Tanaka, M. Physical mechanism of material flow in variable polarity plasma arc keyhole
welding revealed by in situ x-ray imaging. Phys. Fluids 2021, 33, 017121. [CrossRef]
71. Xu, B.; Tashiro, S.; Tanaka, M.; Jiang, F.; Chen, S. Physical mechanisms of fluid flow and joint inhomogeneity in variable-polarity
plasma arc welding of thick aluminum alloy plates. Phys. Fluids 2021, 33, 087103. [CrossRef]
72. Zhang, G.; Wu, C.S.; Liu, Z. Experimental observation of both keyhole and its surrounding thermal field in plasma arc welding.
Int. J. Heat Mass Transf. 2014, 70, 439–448. [CrossRef]
73. Chen, S.; Liu, J.; Jiang, F.; Xu, B.; Zhang, G.; Li, B. Gravity effects on temperature distribution and material flow in the keyhole
pool of VPPA Al welding. Int. J. Heat Mass Transf. 2022, 191, 122823. [CrossRef]
74. Nguyen, A.V.; Tashiro, S.; Bui, H.V.; Tanaka, M. Experimental investigation on the weld pool formation process in plasma keyhole
arc welding. J. Phys. D Appl. Phys. 2018, 51, 015204. [CrossRef]
75. Xu, B.; Chen, S.; Tashiro, S.; Jiang, F.; Nguyễn, V.A.; Tanaka, M. Material flow analyses of high-efficiency joint process in VPPA
keyhole flat welding by X-ray transmission system. J. Clean. Prod. 2020, 250, 119450. [CrossRef]
76. Lu, S.; Fujii, H.; Sugiyama, H.; Tanaka, M.; Nogi, K. Effects of oxygen additions to argon shielding gas on GTA weld shape. ISIJ
Int. 2003, 43, 1590–1595. [CrossRef]
77. Matsuda, S.; Tanahara, Y. Effect of magnetic field on flows and temperature distribution of molten pool in electromagnetic
controlled molten pool welding process. Q. J. Jpn. Weld. Soc. 2015, 84, 244–250. [CrossRef]
78. Zong, R.; Chen, J.; Wu, C.S.; Padhy, G.K. Influence of molten metal flow on undercutting formation in GMAW. Sci. Technol. Weld.
Join. 2016, 234, 169–176. [CrossRef]
79. Sato, T.; Okubo, A.; Oji, T.; Hirata, Y. Measurement of molten pool temperature distribution by UV thermal radiation. Weld. Int.
1998, 12, 627–634. [CrossRef]
80. Tanaka, M.; Terasaki, H.; Ushio, M.; Lowke, J.J. Numerical Study of a Free-burning Argon Arc with Anode Melting. Plasma Chem.
Plasma Process. 2003, 23, 585–606. [CrossRef]
81. Wang, X.; Fan, D.; Huang, J.; Huang, Y. Numerical simulation of arc plasma and weld pool in double electrodes tungsten inert
gas welding. Int. J. Heat Mass Transf. 2015, 85, 924–934. [CrossRef]
82. Fujii, H.; Sato, T.; Lu, S.; Nogi, K. Development of an advanced A-TIG (AA-TIG) welding method by control of Marangoni
convection. Mater. Sci. Eng. A 2008, 495, 296–303. [CrossRef]
83. Urena, A.; Otero, E.; Utrilla, M.V.; Munez, C.J. Weldability of a 2205 duplex stainless steel using plasma arc welding. J. Mater.
Process. Technol. 2007, 182, 624–631. [CrossRef]
84. Migiakis, K.; Daniolos, N.; Papadimitriou, G.D. Plasma Keyhole Welding of UNS S32760 Super Duplex Stainless Steel: Microstruc-
ture and Mechanical Properties. Mater. Manuf. Process. 2010, 25, 598–605. [CrossRef]
85. Prasad, K.S.; Rao, C.S.; Rao, D.N. Optimization of fusion zone grain size, hardness, and ultimate tensile strength of pulsed current
micro plasma arc welded Inconel 625 sheets using genetic algorithm. Int. J. Adv. Manuf. Technol. 2016, 85, 2287–2295. [CrossRef]
86. Cai, D.; Han, S.; Yan, D.; Luo, Z.; Wang, H.; Zhang, Y. Microstructure and fatigue behavior of variable polarity plasma arc welded
Al-6.2Mg alloys. J. Mater. Process. Technol. 2019, 269, 128–134. [CrossRef]
87. Kumar, A.S.; Rao, T.V.H.; Rao, V.V.S.K.; RamaKanth, R.T. Optimizing pulsed current micro plasma arc welding parameters
to maximize ultimate tensile strength of titanium (Ti-6Al-4V) alloy using Dragon fly algorithm. Mater. Today Proc. 2020, 27,
2086–2090. [CrossRef]
88. Gupta, R.; Reddy, R.; Mukherjee, M.K. Key-Hole Plasma Arc Welding Of 8 Mm Thick Maraging Steel—A Comparison with Multi-
Pass Gtaw. Weld. World 2012, 56, 69–75. [CrossRef]
89. Yan, Z.; Chen, S.; Jiang, F.; Zhang, W.; Huang, N. Study and optimization against the gravity effect on mechanical property of
VPPA horizontal welding of aluminum alloys. J. Manuf. Process. 2019, 46, 109–117. [CrossRef]
90. Liu, J.; Jiang, F.; Chen, S.; Wang, K.; Zhang, G.; Xu, B.; Cheng, W.; Ma, X. Cross-scale process quality control of variable polarity
plasma arc welding based on predefined temperature field. J. Mater. Res. Technol. 2023, 26, 5347–5359. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.