materials

Article
Effect of Filling Rate on Underwater Wet Welding
Process and Weld Appearance
Xin Zhang 1,2 , Ning Guo 1,2,3, *, Changsheng Xu 1,2 , Haoran Kan 1,2 , Yanbo Tan 1,2 and
Hao Chen 1,2
1 State Key Laboratory of Advanced Welding and Joining, Harbin Institute of Technology, Harbin 150000,
China; zxin299@126.com (X.Z.); xchshah@163.com (C.X.); 18678673126@163.com (H.K.);
b17863066695@163.com (Y.T.); chh523@126.com (H.C.)
2 Shandong Provincial Key Laboratory of Special Welding Technology, Harbin Institute of Technology at
Weihai, Weihai 264209, China
3 Shandong Institute of Shipbuilding Technology, Weihai 264209, China
* Correspondence: gn21c@hitwh.edu.cn; Tel./Fax: +86-631-5677156




Received: 27 December 2019; Accepted: 24 February 2020; Published: 27 February 2020


Abstract: Real-time electric signal, matter transfer mode and welding pool behavior were obtained to
investigate the effect of wires’ filling rate on arc stability and joints’ appearance during underwater
wet flux-cored arc welding (FCAW). The electric signal results showed that arc stability first decreased
and then increased rapidly because the raise of filling rate affected the number of charged particles
and the electrical conductivity of welding arc atmosphere. Two typical transfer modes, globular
repelled transfer mode and surface tension transfer mode, were observed in this study. The ratio of
surface tension transition could be increased by adding wires’ filling rate. Meanwhile, the geometry
of molten pool was changed and the distance between droplets to welding pool reduced as the filling
rate increased. The fusion line became more regular and the radius of curvature increased under the
effect of bubbles in the molten pool. As the filling rate improving, more slags on the welds surface
were acquired and the welds were much flatter and smoother.

Keywords: underwater wet welding; welding process; filling rate; droplet transfer mode

1. Introduction
Underwater welding is a rapidly developing key marine technology used in the repair and
maintenance of ocean platforms, marine vessels and submarine pipeline [1,2]. There are three methods
of underwater welding specific to welding environment: underwater wet welding (UWW), underwater
local-dry welding and underwater dry welding [3]. As the most widespread underwater welding
technology, UWW technology has developed comprehensively since ever invented by virtue of its
maintainability, operability and consumability.
In the past century, the most common filling materials in UWW technology was covered from
electrodes to flux-cored wires [4]. By adjusting the component and the relative content of electrode
coating and core metal, high quality welded joints could be obtained in various commonly used
structural steels and stainless steels [5]. However, there were several technical defects that restricted its
development. The welding process could not achieve automation because the electrode was required
to be replaced artificially and frequently. Project lifecycle was extended by the low efficiency and the
welder was placed in a danger environment with high water pressure. The quality of the welding
joint was greatly affected by the welder’s ability and lead to associated uncertainty. Therefore, a novel
method, flux cored arc welding (FCAW) has become a new focus of research in recent decades.
The work efficiency and welded joints’ performance has drastically improved with underwater wet
FCAW. The joints’ properties and appearance can be controlled by adjusting the powder composition

Materials 2020, 13, 1061; doi:10.3390/ma13051061 www.mdpi.com/journal/materials

Materials 2020, 13, 1061 2 of 15

and relative content of the compound and metal particles in the flux cored wires. CaF2 plays a
significant role in dehydrogenation and appearance. Zhang et al. investigated the influence of CaF2
content on the welding performance and properties in UWW. The addition of CaF2 from 0% to
65% significantly improved the weld formation and the depth-width ratio descended from 33.5% to
15.7% [6]. Guo et al. believed that the appropriate content of Ni could improve the tensile strength
of E40 steel joint, but excess Ni could cause the decrease of impact toughness [7]. Li et al. studied
the effect of alumino-thermic addition on underwater wet welding process stability [8]. The thermite
addition would improve the stability of the wet welding process while there was a lot of numerous
tiny spatters on the surface of weld at larger content of thermite. Liu et al. investigated the effects of
Mo, Ti and B on microstructure and mechanical properties of UWW joints indicating that although an
excessive content of Mo, Ti and B in the deposited metal would induce an improvement of the weld
strength, it would also lead to a sharp deterioration of the weld plasticity and toughness [9].
A more suitable welding process was more conducive to obtain high quality welded joints. Guo
et al. observed four fundamental transfer modes in underwater FCAW: the globular repelled transfer
mode, the surface tension transfer mode, the explosive short-circuit transfer mode and the “submerged
arc transfer mode” pointing out that the previous two were the main transfer modes [10]. Jia et al.
proposed the impeding effect of arc bubbles on droplet transfer process [11]. Chen et al. studied the
behaviors and characteristics of arc bubble by an in-situ X-ray method. A three-stage model of the
effect of bubble behaviors on droplet transfer was presented [12].
The content of powder has an enormous influence on the joints’ appearance, properties and
welding process. Holding a uniform filling rate of the wires was a necessary condition. However, there
were few literatures that indicated the influence of the powder filling rate on welding process and
weld joints in underwater FCAW. Most of the flux-cored tubular wires used in underwater welding
was manufactured in four steps: pressing groove, filling powder, squeezing into circular and reducing
the diameter. Filling rate, represented by η, means the mass fraction of powder added in the groove in
the filling powder step. It could be defined as the weight of the powder divided by the total weight of
the powder and the coating. It could be derived as equation 1:

η = m1/(m1 + m2) (1)

where, m1 is the weight of powder per unit length, m2 is the weight of metal strap outside per
unit length.
Wires with larger filling rate means more powder has melted and transferred into the molten pool
during underwater FCAW. Thus, the welding beam will have more alloying elements and a thicker
coating slag. On the other hand, the diameter of wires has typically been constant. A larger filling rate
means a thinner metal strap outside as a contrast. In UWW process, the current was mainly passed by
the outer metal. A thinner outer mater would reduce the effective conductive area for electronic and
increase the electrical resistance. The heat quantities of elongation would change.
As stated above, filling rate cannot be ignored in relation to effect of powder components. In this
study, a mature flux cored wire powder formula was selected and four welding wires with different
filling rates were designed to evaluate the effect of filling rate addition on underwater wet welding
process stability and weld appearance.

2. Materials and Experiments

A special self-protecting flux cored wire for underwater wet welding was employed in this
experiment. The dimension of the pure nickel strap (99.99%) outside is 8 mm × 0.3 mm. The mixed
powder inside is composed of slag-forming constituents, gas-forming constituents, arc stabilizer and
thermite, which will produce an Al2 O3 (about 20 Wt.%) − CaF2 (about 70 Wt.%) slag system. The main
components of powder are shown in Table 1. In order to accurately analyze the effect of filling rate, the
electrode wire with 1.6mm in diameter was filled with filling rate of 15, 20, 24 and 28% respectively.
Materials 2020, 13, 1061 3 of 15
Materials 2020, 13, x FOR PEER REVIEW 3 of 15

powder
The filling, process
formation rolling sealing and
of wires diameter
included reduction,
four as shownpressing
steps: U-groove in Figureformation,
1. The basepowder
metal used in
filling,
this investigation
rolling sealing andwas full-austenite
diameter 304
reduction, asLshown
stainless steel with
in Figure a dimension
1. The base metalof 200in
used mmthis× investigation
150 mm × 10
mm.full-austenite 304 L stainless steel with a dimension of 200 mm × 150 mm × 10 mm.
was

Table 1. Flux
Table 1. Flux composition
composition used
used in
in experiments
experiments (wt.%).
(wt.%).

CaF2 Al
CaF2 Al Mn LiF Mn LiF
60 2060 20 10 10 10 10

Figure 1. Flux-cored wires formation process.

process.

A SAF-FRODIGI@WAVE500
SAF-FRODIGI@WAVE500 welding welding source
source was
was used
used and
and the
the automatic
automatic welding
welding process was
operated underwater
underwater in 0.5m depth. Direct current electrode positive (DCEP) was used and welding
parameters
parameters are
are exhibited
exhibited in
in Table
Table2.2.

2. Welding
Table 2.
Table Welding parameters
parameters during
during underwater
underwater wet
wet welding
welding

Welding Voltage
Welding Voltage Welding
WeldingSpeed
Speed Wire
Wire Extension WireFeeding
Extension Wire Feeding SpeedWelding
Speed Welding Current
Current
20 V20 V 2.52.5mm/s
mm/s 15 mm
15 mm 2 2m/min
m/min 200A
200 A

To precisely
To precisely observe
observe the
the droplet
droplet transfer
transfer behavior,
behavior, an an X-ray
X-ray in-situ
in-situ imaging
imaging system
system waswas utilized
to capture the droplet transfer process image as shown in Figure 2. It consisted of a micro-focused
droplet transfer process image as shown in Figure 2. It consisted of a micro-focused X-
ray tube,
X-ray tube,ananimage
imageintensifier
intensifierand
andaahigh-speed
high-speedcamera.
camera.The Thehigh-speed
high-speedcamera camera has
has aa frame rate up
5000 frames
to 5000 frames per
per second.
second. The whole system was wrapped wrapped in in aa lead
lead chamber
chamber toto isolate
isolate X-rays.
X-rays. The
The
welding current and arc voltage were recorded by a high-speed data acquisition
welding current and arc voltage were recorded by a high-speed data acquisition system with a 10 kHz system with a 10
kHz sampling
sampling rate
rate in in order
order to directly
to directly and timely
and timely reflectreflect the welding
the welding stability.
stability.
Materials 2020, 13, 1061 4 of 15
Materials 2020, 13, x FOR PEER REVIEW 4 of 15

Figure2.2.Experimental
Figure Experimentalequipment
equipmentschematic.
schematic.

3.3.Results
Resultsand
andDiscussion
Discussion
In
In underwater wet FCAW,
underwater wet FCAW, larger
largerfilling
fillingrate
rateofof wires
wires related
related to thinner
to thinner outer
outer metals
metals and
and more
more
powder powder particle
particle inside.inside. Wiresdifferent
Wires with with different fillingcould
filling rates ratescause
couldchanges
cause changes in the
in the electrical
electrical characteristic of arc, the droplet transfer mode, the welding stability and even
characteristic of arc, the droplet transfer mode, the welding stability and even the welded joints’the welded
joints’ appearance.
appearance.

3.1.
3.1.The
TheVariation
VariationofofElectrical
ElectricalSignal
SignalIndexes
Indexes
Figure
Figure3 3shows
showsthe change
the change of average
of average welding
welding current and and
current arc voltage as filling
arc voltage rate growing
as filling from
rate growing
15% to 28%. The average welding current of wires with different filling
from 15% to 28%. The average welding current of wires with different filling rate decreased from rate decreased from 129A
to 97Ato
129A as97A
the as
filling rate increased
the filling from 15%
rate increased fromto15%24%toand24%then
andgrown to 127A
then grown at 28%.
to 127A The error
at 28%. bar
The error
showed
bar showed an opposite tend. In welding process, too much short circuits or arc extinctions werethe
an opposite tend. In welding process, too much short circuits or arc extinctions were the
fundamental cause of welding arc stability decline. When a short circuit occurred,
fundamental cause of welding arc stability decline. When a short circuit occurred, smaller arc voltage smaller arc voltage
and
andlarger
largercurrent
currentoccurred.
occurred.When When anan arcarc
extinction
extinction occurred,
occurred, a small welding
a small welding current and and
current a large arc
a large
voltage occurred. In this study, the standard deviation was calculated based
arc voltage occurred. In this study, the standard deviation was calculated based on the measured on the measured welding
current
welding and arc voltage.
current and arcThe variation
voltage. Thecoefficient
variation of arc voltage
coefficient (welding
of arc voltage current)
(welding wascurrent)
the ratiowasof the
the
standard deviation and the mean value of the measured arc voltage (welding
ratio of the standard deviation and the mean value of the measured arc voltage (welding current). Its current). Its reciprocal
could morecould
reciprocal precisely
more reflect arc stability
precisely reflect arcby the reciprocal
stability by theof reciprocal
variation coefficient
of variation of coefficient
arc voltageofand arc
welding
voltage current.
and welding A larger reciprocal
current. A larger of variation
reciprocalcoefficient indicated
of variation a smaller
coefficient indicatedstandard deviation
a smaller of
standard
arc voltage (welding current). This signified that there were less moments in
deviation of arc voltage (welding current). This signified that there were less moments in the welding the welding process that
short
process circuit
thator arc extinction
short circuit or arc occurred.
extinction As occurred.
shown in As Figure
shown 4, the
in reciprocal
Figure 4, the of current
reciprocalandofvoltage
current
variation
and voltage variation coefficient decreased as the filling rate adding from 15% to 24%increased.
coefficient decreased as the filling rate adding from 15% to 24% and then sharply and then
Itsharply
is indicated that the
increased. It isarcindicated
stability that
was the deteriorated with
arc stability more
was powder inwith
deteriorated wires. However,
more powderit in became
wires.
more stable as the filling rate increased.
However, it became more stable as the filling rate increased.
Materials 2020, 13, 1061 5 of 15
Materials 2020, 13, x FOR PEER REVIEW 5 of 15
Materials 2020, 13, x FOR PEER REVIEW 5 of 15

Figure 3. Effect of filling rate on average welding current.

Figure 3. Effect of filling rate on average welding
welding current.
current.

Figure4.4.Reciprocal
Figure Reciprocalof
ofvariation
variationcoefficient
coefficientatatdifferent
differentfilling
fillingrate.
rate.
Figure 4. Reciprocal of variation coefficient at different filling rate.
In
Inthis
this study,
study, the
the probability
probability of of the
the arc
arcvoltage
voltagedistribution
distributionwas wascharacterized
characterized as as Figure
Figure 5.
5. As
As In this
illustrated study, the probability of the arc voltage distribution was characterized as Figure 5. As
illustrated byby Chen[13],
Chen [13],there
therewere
werethree
threeregions
regionsin inthe
thearc
arc voltage
voltage distribution
distribution probability
probability diagram:
diagram:
illustrated
the by Chen [13],inthere were three regionsregion
in the arc voltage distribution probability diagram:
the short-circuit region in left; the arc extinction region in right; the large humpininthe
short-circuit region left; the arc extinction in right; the large hump thecenter
centerwhich
which
the short-circuit
represented region in left; theofarc extinction region in right; the large hump in the center which
representedthe thestable
stablecombustion
combustion welding
of welding arc.
arc.AnAn intense
intense and narrow
and narrow large hump
large hump were
were indicated a
indicated
represented
stable weldingthearc,
stable
and combustion
vice versa. of welding
When the arc. Anrate
filling intense
was and
28%, narrow
the curvelarge hump
was were indicated
significantly lower
a stable welding arc, and vice versa. When the filling rate was 28%, the curve was significantly lower
a stable
than welding arc, and vice versa. When the filling rate was 28%, the curve was significantly lower
thanothers
othersininthe
theshort-circuit
short-circuitregion
regionandandarc
arcextinction
extinctionregion
regionwhile
whilethe
themiddle
middlehump
humpwaswasalsoalsomore
more
than others
concentrated in the
and short-circuit
higher, which region and
indicated arc
a extinction
more stable region
welding while
arc. the middle hump was also more
concentrated and higher, which indicated a more stable welding arc.
concentrated and higher, which indicated a more stable welding arc.
Materials 2020, 13, 1061 6 of 15
Materials 2020, 13, x FOR PEER REVIEW 6 of 15

Figure
Figure 5. 5.
TheThe effect
effect ofof filling
filling rate
rate onon probability
probability of of voltage
voltage distribution.
distribution.

3.2. Metal
3.2. Transfer
Metal Process
Transfer Process
The
The droplet
droplet transfer
transfermodemode waswasclosely related
closely to thetostability
related of welding
the stability processprocess
of welding [14]. In [14].
this study,
In this
two transfer
study, twomodetransfer weremodeobserved in every filling
were observed in everyratefilling
by X-ray ratein-situ
by X-ray imaging
in-situsystem
imaging in the welding
system in the
process. Taking the droplet transfer process with the filling rate of
welding process. Taking the droplet transfer process with the filling rate of 28%, as an example, the 28%, as an example, the typical
globular
typicalrepelled
globulartransfer
repelled process
transfer and the electrical
process and the signal were signal
electrical depictedwerein detail,
depicted as shown in Figure
in detail, as shown 6.
Atin9.456
Figure s, the
6. Atwire 9.456melted
s, the and
wireamelted
small droplet
and a small formed dropletin the tip. As
formed the tip.
in the wireAscontinued to melt,
the wire continued
the
to melt, the droplet continued to grow and sway up and down at the tip of the wire, causing thetoarc
droplet continued to grow and sway up and down at the tip of the wire, causing the arc voltage
constantly
voltage tochange.
constantly At 9.568
change.s, theAtdroplet
9.568 s,left thethe tip ofleft
droplet wiretheandtip entered
of wire into arc bubble.
and entered intoAt arc9.629 s,
bubble.
the droplet drip into the welding pool. The electrical signal photograph
At 9.629 s, the droplet drip into the welding pool. The electrical signal photograph demonstrated that demonstrated that during the
globular
duringrepelled
the globular transfer process,
repelled the welding
transfer process, current and arc current
the welding weldingand were fluctuated
arc welding in a tiny
were range.
fluctuated
in aFigure 7 presented the surface tension transfer process and electrical signal when the filling rate
tiny range.
was 28%. Figure 7surface
The presented tension transfer
the surface process
tension could process
transfer be divided andinto two steps:
electrical signalrepulsion
when the step fillingand
rate
transfer step. From 34.007 s to 34.090 s, a repulsion step occurs
was 28%. The surface tension transfer process could be divided into two steps: repulsion step and in surface tension transfer process
under the effect
transfer of the 34.007
step. From repulsive s toforce.
34.090 However, the droplet
s, a repulsion did notinleave
step occurs the end
surface of the
tension wire until
transfer it
process
had contacted with the molten pool at 34.097 s and then entered the pool
under the effect of the repulsive force. However, the droplet did not leave the end of the wire until it under the action of surface
tension at 34.103 with
had contacted s. The theelectrical
molten pool signalatwas 34.097the ssame
and thenas theentered
repulsion the transfer
pool under mode theinaction
the repulsion
of surface
step. As long
tension as thes.droplet
at 34.103 contacted
The electrical the was
signal molten the pool
samesurface, the welding
as the repulsion current
transfer modesharply
in theincreased
repulsion
tostep.
275 AAsand the arc voltage harshly decreased to 0 V, like an arc extinction
long as the droplet contacted the molten pool surface, the welding current sharply increased process.
Figure
to 275 A and8 shows
the arcthe effectharshly
voltage of wires’ filling rate
decreased to 0on V,thelikemetal
an arctransfer
extinctionmode. The percentage of
process.
surfaceFigure
tension 8 shows the effect of wires’ filling rate on the metal transfer mode.were
transfer mode were always above 50% in this experiment, which The higher
percentage thanof
globular repelled transfer mode. About 51.43% of the droplet
surface tension transfer mode were always above 50% in this experiment, which were higher than transfer were processed in surface
tension
globular transfer mode,
repelled while mode.
transfer it rose to 86.84%
About as theof
51.43% filling rate growing
the droplet to 24%.
transfer wereThen the percentage
processed in surface
oftension
surface transfer
tension transfer mode dropped slightly to 83.87% when the filling
mode, while it rose to 86.84% as the filling rate growing to 24%. Then the percentage rate was 28%. It could be
considered
of surfacethat as long
tension as the mode
transfer filling dropped
rate increases to ato
slightly certain
83.87% value
when (in the
thisfilling
study, rate
24%), wasthe28%.
proportion
It could
ofbe
different droplet
considered transition
that as longmodesas thetendsfillingtorate
be stable
increasesand onlyto a fluctuates
certain value within (ina this
certain small
study, range.
24%), the
Figure 9 shows the effect of filling rate on probability of droplet growth
proportion of different droplet transition modes tends to be stable and only fluctuates within a certain time. As the filling rate was
15%,
smalla considerable
range. Figure number
9 shows of droplets
the effectwith a formation
of filling rate ontime more than
probability of 0.2 s wasgrowth
droplet generated.time.When As the
the droplet filling rate was 20%, the number of droplets whose
filling rate was 15%, a considerable number of droplets with a formation time more than 0.2 s wasgrowing time below 0.2 s increased.
However,
generated. continuing
When thetodroplet increase filling
filling rate,
rate was the droplets’
20%, the numbergrowth of time did not
droplets whose decrease
growing significantly.
time below
The
0.2reason would be
s increased. given in continuing
However, Section 3.4. to increase filling rate, the droplets’ growth time did not
decrease significantly. The reason would be given in Section 3.4.
Materials 2020, 13, 1061 7 of 15
Materials 2020, 13, x FOR PEER REVIEW 7 of 15

Figure 6. Typical globular repelled transfer mode and electric signal (28% filling rate).
Figure 6. Typical globular repelled transfer mode and electric signal (28% filling rate).

Figure 7.
Figure Typical surface
7. Typical surface tension
tension transfer
transfer mode
mode and
and electric
electric signal
signal (28%
(28% filling
filling rate).
rate).
Materials 2020, 13, 1061 8 of 15
Materials 2020, 13, x FOR PEER REVIEW 8 of 15
Materials 2020, 13, x FOR PEER REVIEW 8 of 15

Figure 8. The effect of filling rate on the relative proportion of metal transfer mode.
Figure8.8.The
Figure Theeffect
effectofoffilling
fillingrate
rateononthe
therelative
relativeproportion
proportionofofmetal
metaltransfer mode.
transfer mode.

Figure 9. The effect of

Figure 9. of filling
filling rate
rate on
on probability
probability of
of droplet
droplet growth
growthtime.
time.
Figure 9. The effect of filling rate on probability of droplet growth time.
Additionally,
Additionally,threethreespatter
spattertypes
typeswere
were observed
observed inin every
every welding
welding process
process asas shown
shown in Figure 10.
in Figure 10.
Additionally,
According to Fu et three
al. [15], spatter
the threetypes were
kinds of observed
spatter indroplet
were every welding
repelled process
spatter, as shown
weld pool inoscillation
Figure 10.
According to Fu et al. [15], the three kinds of spatter were droplet repelled spatter, weld pool
According
spatter to Fu et spatter.
andspatter
explosive al. [15],The
the droplet
three kinds of spatter
repelled spatter were
has droplet repelled
aspatter
significant spatter,
theweld pool
oscillation and explosive spatter. The droplet repelled has ainfluence on influence
significant weldingon
oscillation
appearance spatter and explosive spatter. The droplet repelled spatter has a significant influence on
the weldingbecause
appearance this because
kind of this
spatters
kindwith large diameter
of spatters with largewould fall would
diameter on bothfall
sides of the
on both weld
sides of
the
metalwelding
andmetal appearance
difficult because
todifficult
remove.toThe this kind
explosive of spatters with large diameter would fall on both sides of
the weld and remove. Thespatter alsospatter
explosive has an also
obviously
has aneffect on weld
obviously appearance
effect on weld
the
by weld metal
causing the and todifficult to remove. The explosive spatter also has an obviously effect on weld
appearance bypool
causing vibrate
the poolviolently.
to vibrate violently.
appearance by causing the pool to vibrate violently.
Materials 2020, 13, 1061 9 of 15
Materials 2020,
Materials 2020, 13,
13, xx FOR
FOR PEER
PEER REVIEW
REVIEW 99 of
of 15
15

Figure 10.
Figure 10. Spatters
Spatters during
Spattersduring metal
duringmetal transfer
metaltransfer process
transferprocess atat
process
at 15% filling
15%
15% rate.
filling
filling (a) Droplet
rate.
rate. (a) Droplet
(a) repelled
Droplet spatter.
repelled
repelled (b)
spatter.
spatter. (b)
Weld
(b)
Weld pool
Weld
pool oscillation
pool spatter.
oscillation
oscillation (c) Explosive
spatter.
spatter. (c) Explosive
(c) spatter.
Explosive spatter.
spatter.
3.3. Weld Appearances
3.3. Weld
Weld Appearances
3.3. Appearances
Figure
Figure 11 11 shows
11 shows
shows the the appearance
the appearance
appearance of of welded
of welded joints
joints acquired
acquired by by wires
wires with
with different
different filling
filling rate.
rate.
Figure welded joints acquired by wires with different filling rate.
When
When thethe filling
the filling rate
filling rate were
rate were
were 15%15%
15% andand 20%,
and 20%, on
20%, on the
on the weld
the weld surface
weld surface
surface waswas covered
was covered almost
covered almost
almost nono slag
no slag as
slag as presented
as presented
presented
When
by
by Figure
Figure 11a,b.
11a,b. As
As the
the filling
filling rate
rate growing,
growing, there
there were
were more
more and
and more
more slags were
slags were formed
formed as as
given in
given
by Figure 11a,b. As the filling rate growing, there were more and more slags were formed as given
Figure
in Figure11b,c,e.
Figure 11b,c,e. At
11b,c,e. Atthe same
At the
the same time,
same time, the
time, thequantity
the quantity of
quantity ofspatters
of spattersincreased.
spatters increased. The
increased. The welded
The welded beams
welded beams were
beams were more
were moreand
more
in
more
and straight
more while while
straight the surface
the more glabrous
surface more and smoother
glabrous and after disengaging
smoother after slags, as displayed
disengaging slags, as
and more straight while the surface more glabrous and smoother after disengaging slags, as
in Figure
displayed in11a–d,f.
in Figure According
Figure 11a–d,f. to these
11a–d,f. According figures,
According to the
to these weld
these figures,appearance
figures, thethe weld was substantially
weld appearance
appearance was improved
was substantiallywith
substantially
displayed
increasing
improved the filling
with rate. It
increasing could
the be rate.
filling attributed
It couldto the
be protectiontoofthe
attributed theprotection
more and of morethe slags.
more The
and
improved with increasing the filling rate. It could be attributed to the protection of the more and
existence
more of
slags. molten
The slags
existence reduced
of molten the fluctuation
slags reduced amplitude
the of molten
fluctuation pool
amplitude surface.
of When
molten the
pool bubbles
surface.
more slags. The existence of molten slags reduced the fluctuation amplitude of molten pool surface.
in the molten
When pool in
the bubbles
bubbles broken,
the moltenit could
molten poolalso effectively
broken, it could
couldinhibit the metal to
also effectively
effectively fly out
inhibit theofmetal
the molten
metal to fly pool.
fly out
out of
When the in the pool broken, it also inhibit the to of
However,
the molten as the
pool. stability
However, of the
as arc
the first decreased
stability of the and
arc then
first increased,
decreased the
and arc
then force of
increased,the droplet
the arc was
force
the molten pool. However, as the stability of the arc first decreased and then increased, the arc force
not stable,
of the
the andwas
droplet the droplet
not stable, repelled
stable, and the spatter
the increased.
droplet repelled spatter
spatter increased.
increased.
of droplet was not and droplet repelled

Figure 11.
Figure
Figure 11. The
11. The effect
The effectofof
effect offilling
fillingrate
filling rate
rate on
onon weld
weld
weld appearance
appearance
appearance with
with different
different
with filling
filling
different rate
rate rate
filling wires.
wires. (a) 15%;
(a) 15%;
wires. (a) 15%; (b)
(b) 20%;
(b)
20%;
(c) (c)
24% 24%
(with (with
slag); slag);
(d) 24% (d) 24% (without
(without slag); slag);
(e) 28%(e) 28%
(with (with
slag); slag);
(f) 28%(f) 28% (without
(without
20%; (c) 24% (with slag); (d) 24% (without slag); (e) 28% (with slag); (f) 28% (without slag). slag). slag).

The metallographic sample was taken according to the position shown by the red dotted line in
The metallographic sample was taken according to the position shown by the red dotted line in
Figure
Figure 11.
11. After
After polishing
polishingand
andcorrosion,
corrosion,thethecross
crosssections shown
sections shown ininFigure
Figure12 12
were obtained.
were Figure
obtained. 12
Figure
Figure 11. After polishing and corrosion, the cross sections shown in Figure 12 were obtained. Figure
shows the
12 shows cross
shows the sections
the cross of welding
cross sections
sections of joints
of welding produced
welding joints at different
joints produced
produced at filling
at different rate after
different filling etching.
filling rate Obviously,
rate after
after etching.
etching.
12
the outline became
Obviously, the much
outline more regular.
became much The number
more regular. of defects,
The number suchof as blowholes
defects, such and
as slag inclusion,
blowholes and
Obviously, the outline became much more regular. The number of defects, such as blowholes and
slag inclusion,
slag inclusion, also
also decreased
decreased asas the
the filling
filling rate
rate increased
increased from
from 15%
15% to to 28%.
28%. The
The fusion
fusion line
line in
in Figure
Figure
12a was irregular arc while the others were regular with a larger and larger radius, as shown in Figure
12b–d.
Aspect ratio was an important parameter of weld formation. The ratio of weld depth to weld
width could reflect the characteristics of weld formation to a certain extent. Figure 13 illustrated the
weld penetration
Materials 2020, 13, 1061depth, the aspect ratio and the reinforcement by calculation. The weld aspect 10ratio
of 15
varied from 26.5% to 15.8% as filling rate increased from 15% to 28%. Since the weld width has just
Materials 2020, 13, x FOR PEER REVIEW 10 of 15
little fluctuation, the reduction of weld dilution rate was mainly caused by weld penetration depth.
also decreased
The weld as the filling
penetration depthrate increased
decreased as from
filling15%
ratetoincreased
28%. Thewhich
fusionwas
line the
in Figure
result 12a
of awas irregular
combination
12a was irregular arc while the others were regular with a larger and larger radius, as shown in Figure
arc while
of factors. the others were regular with a larger and larger radius, as shown in Figure 12b–d.
12b–d.
Aspect ratio was an important parameter of weld formation. The ratio of weld depth to weld
width could reflect the characteristics of weld formation to a certain extent. Figure 13 illustrated the
weld penetration depth, the aspect ratio and the reinforcement by calculation. The weld aspect ratio
varied from 26.5% to 15.8% as filling rate increased from 15% to 28%. Since the weld width has just
little fluctuation, the reduction of weld dilution rate was mainly caused by weld penetration depth.
The weld penetration depth decreased as filling rate increased which was the result of a combination
of factors.

Figure 12. The cross sections of welding joints with different filling rate. (a) 15% filling rate; (b) 20%
filling rate;
filling rate; (c) 24% filling rate; (d)
(d) 28%
28% filling
filling rate.
rate.

Aspect ratio was an important parameter of weld formation. The ratio of weld depth to weld
width could reflect the characteristics of weld formation to a certain extent. Figure 13 illustrated the
weld penetration depth, the aspect ratio and the reinforcement by calculation. The weld aspect ratio
varied from 26.5% to 15.8% as filling rate increased from 15% to 28%. Since the weld width has just
little fluctuation, the reduction of weld dilution rate was mainly caused by weld penetration depth.
Figure 12. The cross sections of welding joints with different filling rate. (a) 15% filling rate; (b) 20%
The weld penetration depth decreased as filling rate increased which was the result of a combination
filling rate; (c) 24% filling rate; (d) 28% filling rate.
of factors.

Figure 13. The effect of filling rate on weld aspect ratio, weld penetration and reinforcement.

3.4. Influence Mechanism Analysis of Thermite

Three aspects, electrical signal indexes, metal transfer process and weld appearances, were used
to evaluate the effect of wire’s filling rate on UWW process. The mechanism analysis would be
discussed in the following three aspects including arc electrical properties, the forces acting on the
droplets in the molten electrode tip and molten pool behavior.
Figure 13. The
Figure 13. The effect
effect of
of filling
filling rate
rate on
on weld
weld aspect
aspect ratio,
ratio, weld
weld penetration
penetration and
and reinforcement.
reinforcement.
3.4. Influence Mechanism Analysis of Thermite
3.4. Influence Mechanism Analysis of Thermite
Three aspects, electrical signal indexes, metal transfer process and weld appearances, were used to
Three aspects, electrical signal indexes, metal transfer process and weld appearances, were used
evaluate the effect of wire’s filling rate on UWW process. The mechanism analysis would be discussed
to evaluate the effect of wire’s filling rate on UWW process. The mechanism analysis would be
discussed in the following three aspects including arc electrical properties, the forces acting on the
droplets in the molten electrode tip and molten pool behavior.
Materials 2020, 13, 1061 11 of 15

in the following three aspects including arc electrical properties, the forces acting on the droplets in the
molten electrode tip and molten pool behavior.
Firstly, the effect of filling rate on the welding arc stability was mainly the change of electric
conductivity of arc plasma atmosphere. As shown in Table 1, CaF2 , as an important dehydrogenation
composition [16], accounted for 60% of the mass of the powder. Under the high temperature of arc,
CaF2 molecular would decompose into F atom and Ca atom (Formula 2). The F atom was very easily
to absorb free electrons in the arc atmosphere and formed F− (Formula 3). The F- could combine with
H+ and the conductivity of the arc atmosphere decreased (Formula 4) [16].

CaF2 → Ca + 2F (2)

F + e− → F− (3)

F− + H+ → HF (4)

When the filling rate increased from 15% to 24%, the consumption of charged particles continued
to strengthen. The number of e- in the welding arc atmosphere continued to decrease. The conductivity
of the arc atmosphere decreased, which caused the decrease of arc stability. However, continuing
to increase the filling rate brought in enough CaF2 , which made the reaction in Formula 3 reached
saturation. Free electrons were no longer scarce, and excess ions, such as F− and H+ , could enhance
electrical conductivity in arc atmosphere. Finally, the arc stability was improved.
Secondly, in both droplet transfer modes, repulsion step was the first step. The stress state of the
droplet before the transition had great influence on the droplet transfer mode. The schematic of forces
analysis for globular repelled step was shown in Figure 14. According to Guo et al. [17], a droplet was
mainly subjected to six forces during welding in water based on the metal transfer static force balance
theory, which were electromagnetic force (Fe ), surface tension (Fs ), gravity (G), plasma drag force (F1 ),
the vaporization force (Fa ) and the gas flow drag force (FL ). When welding in water, the bubbles were
small and the drag force (FL ) acted transitorily. The droplet transferred into the molten pool under
static-force balance theory (SFBT) [18]. The plasma drag force (F1 ) and gravity (G) were detaching
force and the others were holding force.
Materials 2020, 13, x FOR PEER REVIEW 12 of 15

Schematic of
Figure 14. Schematic of forces
forces analysis for globular repelled step.

FL existed
The only in of
flow behavior underwater wet
molten pool haswelding. It came on
direct influence from
thethe pushing
final effect ofHowever,
weld forming. arc bubbles on
there
droplet when it periodically floated up. Chen et al. pointed out that the formula of flow
were few researches on the behavior of UWW pool. In this study, the molten pool was observed using drag force
acting
the on droplets
X-ray [12]: system and the schematic morphology of molten pool were displayed in
in-situ imaging
1 2
FL = Cd ρThe
Figure 15 (the moment before the droplet transition). g Sv molten pool surface was extremely close(5) to
2
the tip of wires. Violent fluctuation was caused by the bubbles burst and the ever-changing arc force.
As the filling rate of wires adding, the shape of welded metal drastically changed. As shown in Figure
15a, the molten pool was full of bubbles dispersedly with a small diameter about 2–3 mm, which
increased the volume of the pool. The front of the lack-of-slag molten pool sank down a little bit
under the pressure of arc force and liquid vaporization thrust. From Figure 15b,c, the bubbles in the
pool gradually diminished since the dehydrogenation of powder and were much more concentrated
Materials 2020, 13, x FOR PEER REVIEW 12 of 15

Materials 2020, 13, 1061 12 of 15

where Cd is the flow drag coefficient, ρg is the density of the gas inside the bubble, S is the contact area
between bubble and droplet, ν is the expansion rate or the rising velocity of the bubble.
A larger droplet has a larger contact area(S) with arc bubble. Therefore, it was subjected to a larger
FL which caused it to leave the wire at a larger angle, resulted in higher percentage of globular repelled
transfer mode. The droplet growth time could reflect its size. The longer the droplet grew, the larger
the volume, the larger the area of contact with the bubble (S) and the larger FL it subjected. According
to Figure 9, when the filling rate was 15%, the quantity of droplet with larger volume was the most.
Continuing to increase the filling rate, the droplet’s volume decreased and then substantially remained
almost constant. The ratio of globular transfer mode would first reduce and then remained. However,
it is different from Figure 8, where the ratio of globular transfer mode continued to decrease. It could
be believed that droplet transition form was not only related to droplet stress, but also related to the
Figure 14. Schematic of forces analysis for globular repelled step.
shape of molten pool.
The
The flow
flow behavior
behavior of of molten
molten poolpool has direct influence
has direct influence on on the final weld
the final forming. However,
weld forming. However, there
there
were
were fewfew researches
researches onon the
the behavior
behavior of of UWW
UWW pool.
pool. InIn this
this study,
study, the molten pool
the molten was observed
pool was observed using
using
the
the X-ray in-situ imaging system and the schematic morphology of molten pool were displayed
X-ray in-situ imaging system and the schematic morphology of molten pool were displayed in in
Figure 15 (the moment before the droplet transition). The molten pool surface
Figure 15 (the moment before the droplet transition). The molten pool surface was extremely close to was extremely close
to
thethe
tiptip of wires.
of wires. Violent
Violent fluctuation
fluctuation waswas caused
caused by the bybubbles
the bubbles
burstburst
and theand the ever-changing
ever-changing arc
arc force.
force. As the filling rate of wires adding, the shape of welded metal drastically
As the filling rate of wires adding, the shape of welded metal drastically changed. As shown in Figure changed. As shown
in
15a,Figure 15a, thepool
the molten molten
waspoolfull was full of bubbles
of bubbles dispersedly
dispersedly with adiameter
with a small small diameter
about 2–3 about
mm, 2–3 mm,
which
which
increased the volume of the pool. The front of the lack-of-slag molten pool sank down a little bit
increased the volume of the pool. The front of the lack-of-slag molten pool sank down a little bit
under the pressure of arc force and liquid vaporization thrust. From Figure
under the pressure of arc force and liquid vaporization thrust. From Figure 15b,c, the bubbles in the15b,c, the bubbles in the
pool
pool gradually
graduallydiminished
diminishedsince sincethe
thedehydrogenation
dehydrogenationofofpowder powder andandwere
weremuch
much more
moreconcentrated
concentrated in
the back half as the filling rate increased. The molten pool no longer fluctuated
in the back half as the filling rate increased. The molten pool no longer fluctuated violently under the violently under the
stronger
stronger protection
protection ofof molten
molten slag. The first
slag. The first half
half directly
directly below
below thethe welding
welding wirewire gradually
gradually became
became
lower
lower and the elongation increased from 9.27mm to 11.35mm. When the filling rate was
and the elongation increased from 9.27mm to 11.35mm. When the filling rate was higher
higher than
than
15%,
15%, thetheprobability
probabilityofof thethe
droplet contacting
droplet the pool
contacting increased
the pool although
increased the droplet
although size was
the droplet similar.
size was
Therefore, the proportion of surface tension transition increased.
similar. Therefore, the proportion of surface tension transition increased.

Figure 15.
Figure Schematic diagram
15. Schematic diagram of
of molten
molten pool
pool at
at different
different filling
filling rate.
rate.

As
As the filling rate
the filling gradually increased,
rate gradually increased, the
the liquid
liquid metal
metal in the molten
in the molten pool
pool gradually
gradually moved
moved back
back
and
and the
the part
part directly
directly below
below the
the welding
welding wire
wire gradually
gradually decreased
decreased because
because the
the viscosity
viscosity of
of the
the molten
molten
pool
pool decreased. It can be seen from Table 1 that the main components in the molten pool wereNiNiand
decreased. It can be seen from Table 1 that the main components in the molten pool were a
and
small amount of CaF
a small amount of CaF , Li, Al, Mn, Fe, etc. These components would vaporize in large quantities
2 2, Li, Al, Mn, Fe, etc. These components would vaporize in large quantities under
the high temperature of the molten pool and the welding arc, which would take away lots of heat and
cause the molten pool to cool down. When the filling rate was 15%, due to the lack of slag protection,
Materials 2020, 13, x FOR PEER REVIEW 13 of 15

Materials
under the 2020, 13, 1061
high temperature 13 of of
of the molten pool and the welding arc, which would take away lots 15
heat and cause the molten pool to cool down. When the filling rate was 15%, due to the lack of slag
protection, the elements’ gasification in the molten pool was serious, taking away a mass of heat.
the elements’ gasification in the molten pool was serious, taking away a mass of heat. Moreover, more
Moreover, more slag separated the molten metal from the water environment, greatly delaying the
slag separated the molten metal from the water environment, greatly delaying the heat conduction of
heat conduction of the molten pool. The following exothermic reactions would occur with the Al
the molten pool. The following exothermic reactions would occur with the Al element in the wire:
element in the wire:

2Al++Fe
2Al Fe2 2OO33 →
→ Al O33 ++2Fe
Al22O 2Fe (6)
(6)

2Al++Fe
2Al FeAl22O44 ++2Fe
→ FeAl
Fe3 3OO44 → 2Fe (7)
(7)

2Al++4FeO
2Al 4FeO→
→FeAl O44 +
FeAl22O 3Fe
+ 3Fe (8)
(8)
With
With thetheincrease
increaseofoffilling rate,
filling thethe
rate, increase of Al
increase ofelement
Al element would also increase
would the heat
also increase theofheat
the molten
of the
pool. However, the weld beam was a kind of Ni-based alloy. The viscosity
molten pool. However, the weld beam was a kind of Ni-based alloy. The viscosity of the Ni-based of the Ni-based liquid alloy
was negatively correlated with temperature [19,20]. With the filling rate increased,
liquid alloy was negatively correlated with temperature [19,20]. With the filling rate increased, the the temperature of
the molten pool increased and the viscosity decreased. The liquid metal
temperature of the molten pool increased and the viscosity decreased. The liquid metal in the pool in the pool was pushed back
by
wasthe arc force
pushed back ofbythethewelding
arc forceandofthethegas gasification
welding and the thrust. As a result,thrust.
gas gasification the height
As a of the first
result, half of
the height
the molten pool dropped. At same time, the dry elongation of the wire
of the first half of the molten pool dropped. At same time, the dry elongation of the wire was was increasing. As shown in
Figure 15, the shape of first half of molten pool under the welding wire
increasing. As shown in Figure 15, the shape of first half of molten pool under the welding wire changed from convex to flat or
even
changedconcave.
from convex to flat or even concave.
Finally,
Finally, based
based on on thethe particularity
particularity of of FCAW,
FCAW, aa conjecture
conjecture on on the
the effect
effect ofof filling
filling rate
rate onon fusion
fusion
line shape were
line shape wereproposed
proposedininthis thispaper:
paper:AsAs cancan
be be
seenseen in Figure
in Figure 15, with
15, with the increase
the increase of filling
of filling rate,
rate, the molten pool moved back and the size of the inner bubble gradually
the molten pool moved back and the size of the inner bubble gradually increased. The larger bubbles increased. The larger
bubbles
in the backin the
halfback half continued
continued to existthe
to exist during during the formation
formation and rupture and processes.
rupture processes.
Under the Under
effectthe
of
effect of bubbles, there were more liquid metal at the edge of the bubble
bubbles, there were more liquid metal at the edge of the bubble and less liquid metal in the central and less liquid metal in the
central which contacted
which contacted with the withbase
the base
metalmetal
beneathbeneath
the the molten
molten pool.
pool. WithWith thethescouring
scouringofofmore more high
high
temperature
temperature liquidliquidmetal,
metal, thethe
amount
amount of melting base metal
of melting beneath
base metal the edge
beneath theof edge
moltenofpool increased
molten pool
while
increased while the counterpart beneath the center of molten pool reduced as shown in Figure the
the counterpart beneath the center of molten pool reduced as shown in Figure 16. Then, 16.
radius
Then, theof the fusion
radius linefusion
of the decreased. With the increase
line decreased. With the of increase
filling rate, this phenomenon
of filling was much more
rate, this phenomenon was
obvious.
much more However,
obvious. it isHowever,
worth mentioning
it is worth thatmentioning
the weld penetration
that the weldwas mainly determined
penetration by the
was mainly
welding parameters. The filling rate had little effect on weld penetration
determined by the welding parameters. The filling rate had little effect on weld penetration but had but had significant effect on
the shape of fusion line.
significant effect on the shape of fusion line.

Figure 16. Schematic diagram of fusion line shape change.

4. Conclusions
Based on the presented results it was concluded that:

1. with the increase of the filling rate of flux-cored wire, the arc stability first decreased and reached
the minimum value when the filling rate was 24%. Then the welding arc became more stable
Materials 2020, 13, 1061 14 of 15

rapidly as the filling rate growing. It was due to the effect of CaF2 and other components in the
powder on the number of charged particles in the arc.
2. Two droplet transfer modes, globular repelled transfer mode and surface tension transfer mode,
were observed in this study. With the filling rate increased from 15% to 28%, the proportion of
surface tension transfer mode harshly increased from 51.43% to 83.87%. The increase of molten
pool temperature and slag coverage reduced the viscosity of molten pool and affected the shape
of molten pool. The wiggle space of the droplet on the tip of the wire was reduced. Resulted in
surface tension transfer mode increased.
3. Slag on the weld surface increased and the welds were much flatter and smoother as the filling
rate increased. The ratio of depth to width decreased gradually while the reinforcement increased.
The fusion line became more regular and the radius of curvature increased under the action of
bubbles in the molten pool.

Author Contributions: X.Z.: Writing—Original draft, Investigation, Data curation, Formal analysis; N.G.:
Methodology, Supervision, Conceptualization; C.X.: Investigation, Formal analysis; H.K.: Visualization; Y.T.: Data
curation; H.C.: Software. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by the Fundamental Research Funds for the Central Universities grant number
HIT.NSRIF.201602, HIT.NSRIF.201704, HIT. MKSTISP.201617 and the Shandong Provincial Key Research and
Development Plan grant number 2017CXGC0922, 2018GGX103003.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Tomków, J.; Rogalski, G.; Fydrych, D.; Łabanowski, J. Advantages of the Application of the Temper Bead
Welding Technique During Wet Welding. Materials 2019, 12, 915. [CrossRef]
2. Yin, Y.; Yang, X.; Cui, L. Investigation on welding parameters and bonding characteristics of underwater wet
friction taper plug welding for pipeline steel. Int. J. Adv. Manuf. Technol. 2015, 81, 851–861. [CrossRef]
3. Guo, N.; Cheng, Q.; Zhang, X.; Fu, Y.; Huang, L. Microstructure and Mechanical Properties of Underwater
Laser Welding of Titanium Alloy. Materials 2019, 12, 2703. [CrossRef] [PubMed]
4. Fydrych, D.; Świerczyńska, A.; Rogalski, G.; Łabanowski, J. Temper Bead Welding of S420G2+M Steel in
Water Environment. Adv. Mater. Sci. 2016, 16, 5–16. [CrossRef]
5. Gao, W.; Wang, D.; Cheng, F.; Di, X.; Deng, C.; Xu, W. Microstructural and mechanical performance of
underwater wet welded S355 steel. J. Mater. Process. Technol. 2016, 238, 333–340. [CrossRef]
6. Zhang, X.; Guo, N.; Xu, C.; Du, Y.; Chen, B.; Feng, J. Influence of CaF2 on microstructural characteristics and
mechanical properties of 304 stainless steel underwater wet welding using flux-cored wire. J. Manuf. Process.
2019, 45, 138–146. [CrossRef]
7. Guo, N.; Liu, D.; Guo, W.; Li, H.; Feng, J. Effect of Ni on microstructure and mechanical properties of
underwater wet welding joint. Mater. Des. 2015, 77, 25–31. [CrossRef]
8. Li, H.; Liu, D.; Guo, N.; Chen, H.; Du, Y.; Feng, J. The effect of alumino-thermic addition on underwater wet
welding process stability. J. Mater. Process. Technol. 2017, 245, 149–156. [CrossRef]
9. Liu, D.; Guo, N.; Xu, C.; Li, H.; Yang, K.; Feng, J. Effects of Mo, Ti and B on Microstructure and Mechanical
Properties of Underwater Wet Welding Joints. J. Mater. Eng. Perform. 2017, 26, 2350–2358. [CrossRef]
10. Ning, G.; Xu, C. Classification of mater transfer mode in underwater wet welding. Weld. Res. 2016, 95,
133–140.
11. Yang, Q.; Han, Y.; Jia, C.; Wu, J.; Dong, S.; Wu, C. Impeding effect of bubbles on metal transfer in underwater
wet FCAW. J. Manuf. Process. 2019, 45, 682–689. [CrossRef]
12. Chen, H.; Guo, N.; Huang, L.; Zhang, X.; Feng, J.; Wang, G. Effects of arc bubble behaviors and characteristics
on droplet transfer in underwater wet welding using in-situ imaging method. Mater. Des. 2019, 170, 107696.
[CrossRef]
13. Chen, H.; Guo, N.; Shi, X.; Du, Y.; Feng, J.; Wang, G. Effect of water flow on the arc stability and metal transfer
in underwater flux-cored wet welding. J. Manuf. Process. 2018, 31, 103–115. [CrossRef]
14. Guo, N.; Du, Y.; Maksimov, S.; Feng, J.; Yin, Z.; Krazhanovskyi, D.; Fu, Y. Study of metal transfer control in
underwater wet FCAW using pulsed wire feed method. Weld. World 2017, 62, 87–94. [CrossRef]
Materials 2020, 13, 1061 15 of 15

15. Fu, Y.; Guo, N.; Du, Y.; Chen, H.; Xu, C.; Feng, J. Effect of metal transfer mode on spatter and arc stability in
underwater flux-cored wire wet welding. J. Manuf. Process. 2018, 35, 161–168. [CrossRef]
16. Plessis, J.; Toit, M. Reducing Diffusible Hydrogen Contents of Shielded Metal Arc Welds Through Addition
of Flux-Oxidizing Ingredients. J. Mater. Eng. Perfrom. 2008, 17, 50–56. [CrossRef]
17. Guo, N.; Wang, M.; Du, Y.; Guo, W.; Feng, J. Metal transfer in underwater flux-cored wire wet welding at
shallow water depth. Mater. Lett. 2015, 144, 90–92. [CrossRef]
18. Graat, L.; Waszink, J. Experimental Investigation of the Forces Acting on a Drop of Weld Metal. Weld. J. 1983,
62, 109s–116s.
19. Shi, R.; Avey, T.; Luo, A. A CALPHAD (CALculation of PHAse Diagrams)-based viscosity model for
Al-Ni-Fe-Co melt system. J. Mol. Liq. 2019, 291, 111271. [CrossRef]
20. Liu, Y.; Lv, X.; Bai, C.; Lai, P.; Wang, J. Viscosity evaluation of Fe-Ni-Co ternary alloy from the measured
binary systems. J. Ind. Eng. Chem. 2015, 30, 106–111. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).