materials

Article
Effects of In-Process Ultrasonic Vibration on Weld Formation
and Grain Size of Wire and Arc Additive Manufactured Parts
Jun Zhang, Yanfeng Xing *, Jijun Zhang, Juyong Cao, Fuyong Yang and Xiaobing Zhang

School of Mechanical and Automobile Engineering, Shanghai University of Engineering Science,

Shanghai 201620, China; zj17355135811@163.com (J.Z.); 13761116082@163.com (J.Z.);
caojuyong86823@163.com (J.C.); yangfy09@163.com (F.Y.); xiaobingzh@hotmail.com (X.Z.)
* Correspondence: smsmsues@163.com; Tel.: +86-021-67791147

Abstract: Wire and arc additive manufacturing (WAAM) is a competitive technique, which enables
the fabrication of medium and large metallic components. However, due to the presence of coarse
columnar grains in the additively manufactured parts, the resultant mechanical properties will be
reduced, which limits the application of WAAM processes in the engineering fields. Grain refinement
and improved mechanical properties can be achieved by introducing ultrasonic vibration. Herein, we
applied ultrasonic vibration to the WAAM process and investigated the effects of wire feed speed,
welding speed, and ultrasonic amplitude on the weld formation and grain size during ultrasonic
vibration. Finally, a regression model between the average grain size and wire feed speed, welding
speed, and ultrasonic amplitude was established. The results showed that due to the difference in
heat input and cladding amount, wire feed speed, welding speed, and ultrasonic amplitude have a
significant influence on the weld width and reinforcement. Excessive ultrasonic amplitude could
cause the weld to crack during spreading. The average grain size increased with increasing wire feed
speed and decreasing welding speed. With increasing ultrasonic amplitude, the average grain size
exhibited a trend of decreasing first and then increasing. This would be helpful to manufacture parts
Citation: Zhang, J.; Xing, Y.; Zhang, of the required grain size in ultrasonic vibration-assisted WAAM fields.
J.; Cao, J.; Yang, F.; Zhang, X. Effects
of In-Process Ultrasonic Vibration on Keywords: wire and arc additive manufacturing; ultrasonic vibration; amplitude; cavitation; grain size
Weld Formation and Grain Size of
Wire and Arc Additive Manufactured
Parts. Materials 2022, 15, 5168.
https://doi.org/10.3390/ 1. Introduction
ma15155168
Wire and arc additive manufacturing (WAAM) is a layered manufacturing process
Academic Editor: Jana Bidulská using wire as a filler material and an electric arc as the heat source. Compared with other
Received: 24 June 2022
additive manufacturing processes, including laser and electron beam, WAAM seems to be
Accepted: 20 July 2022
suitable for fabricating medium and large parts due to its high deposition rate, excellent
Published: 26 July 2022
energy efficiency, open deposition environment, and low capital and feedstock costs [1,2].
In WAAM, the process of generating electric arc can be mainly divided into three categories,
Publisher’s Note: MDPI stays neutral
namely gas tungsten arc welding (GTAW), gas metal arc welding (GMAW), and plasma arc
with regard to jurisdictional claims in
welding (PAW) [3]. Of the three welding processes, the GMAW-based cold metal transfer
published maps and institutional affil-
(CMT) seems to be the most suitable for WAAM applications [4]. Compared with GMAW,
iations.
the CMT welding process produces less heat, resulting in less spatter and controlled metal
deposition rates [5]. Elrefaey et al. [6] found that joints deposited by the CMT process have
better mechanical characteristics than conventional welding processes and produced the
Copyright: © 2022 by the authors.
7075-T6 joints without spatter, cracks, and low porosity.
Licensee MDPI, Basel, Switzerland. Although WAAM has many advantages, it also has certain defects, such as oxidation,
This article is an open access article delamination, high residual stress, deformation, cracking, porosity, and surface finish,
distributed under the terms and etc. [5,7–9]. To improve forming quality and mechanical properties, a wide range of ancillary
conditions of the Creative Commons processes were proposed by researchers including heat treatment [10–13], inter-pass cold
Attribution (CC BY) license (https:// rolling [14–17], inter-pass cooling [18], molten pool oscillation [19], and ultrasonic vibration.
creativecommons.org/licenses/by/ For heat treatment, an appropriate post-process heat treatment process can reduce
4.0/). residual stress and improve mechanical properties, but an improper post-process heat

Materials 2022, 15, 5168. https://doi.org/10.3390/ma15155168 https://www.mdpi.com/journal/materials

Materials 2022, 15, 5168 2 of 17

treatment process would be counterproductive. For inter-pass cold rolling, it is only

suitable for some regular parts, such as thin-walled parts, owing to the restriction of the
cold rolling devices.
Cheng et al. [20] applied ultrasonic impact treatment and shot peening to three base
metal samples, respectively. The results showed that the peak compressive residual stress
exceeded the near-surface base metal material yield stress. Yuan et al. [21] obtained re-
fined grains by ultrasonically stirring the molten pool with a probe located behind the
arc. Tian et al. [22] found the introduction of synchronous ultrasonic vibration results in
increased weld penetration and reinforcement and decreased contact angle. Wang et al. [23]
found that as the ultrasonic frequency increases, the grain size becomes smaller, but results
in an increase in porosity. Chen et al. [24] pointed out that dendrite fragmentation and het-
erogeneous nucleation caused by cavitation and acoustic streaming lead to grain refinement.
Gorunov et al. [25] investigated the effect of frequency and power of ultrasonic-assisted
direct energy deposition on the microstructure of the deposited Ti-6Al-4V parts. They
found that ultrasonic vibration causes the transition from columnar grains to equiaxed
grains and the location of the equiaxed grain regions can be changed by changing the
frequency and power. Zhou et al. [26,27] developed a calculation model to predict the
impact stress field and depth of plastic deformation zone under ultrasonic impact treatment
and analyzed the stress and strain fields of ultrasonically impacted components by experi-
mental and numerical methods. Chen et al. [28] concluded that ultrasonic vibration refines
grains of the Cu-8Al-2Ni-2Fe-2Mn alloy. They found that parts with ultrasonic vibration
obtain smaller grain sizes and better material dispersion than parts without ultrasonic
vibration. Huu Phan et al. [29] pointed out that vibration frequency is the dominant factor
in the ultrasonic vibration-assisted EDM process, which can achieve a higher material
removal rate with better surface quality. Based on multi-objective optimization analysis,
Huu Phan et al. [30] concluded that the workpiece-related low-frequency vibration in
EDM leads to a significant improvement in surface quality. Ma et al. [31] presented that
reduced penetration and increased contact angle of a single bead induced by ultrasonic
vibration. The experimental results showed that the directional growth of coarse columnar
grains of Inconel 625 parts is inhibited and grains are refined under ultrasonic vibration.
Yuan et al. [32] realized the simultaneous coupling of ultrasonic vibration and the laser and
wire additive manufacturing process. The ultrasonic frequency and amplitude were fixed at
20 kHz and 16 µm, respectively. The epitaxial growth trend of the prior-β crystal is blocked
and refines the microstructure of the deposited Ti-6Al-4V alloy parts under the influence of
ultrasonic vibration. Then, they [33] applied the simultaneous ultrasonic-assisted laser and
wire additive manufacturing process to austenitic stainless steel, which also resulted in
grain refinement. Todaro et al. [34] pointed out that the introduction of ultrasonic vibration
increases many initial crystallites and reduces the temperature gradient in the bulk of the
molten pool, leading to the formation of fine grains. Sun et al. [35] applied ultrasonic
impact to the surface of deposited weld in time after deposition. The results showed that
inter-layer ultrasonic impact leads to a refined microstructure of low carbon steel parts, re-
duced local stress concentration, enhanced mechanical properties, and improved anisotropy.
Chen et al. [36] found that increased molten pool temperature induced by ultrasonic cav-
itation and convection results in a reduced temperature gradient, which increases the
nucleation rate and supercooling in favor of fine grain formation.
Based on the abovementioned literature on ultrasonic-assisted additive manufactur-
ing, it can be concluded that the introduction of ultrasonic vibration interrupts the growth
of columnar crystals and promotes the transformation of coarse columnar grains to fine
columnar grains or equiaxed grains, resulting in better mechanical properties. The ultra-
sonic vibration process is generally divided into three types. The first is the ultrasonic
vibration acting on the weld that is synchronously coupled with the additive manufacturing
process. The second is the fixed ultrasonic vibration acting on the substrate. The third is
the ultrasonic vibration after welding, which is called ultrasonic impact treatment. The
scope of application of the first ultrasonic vibration process is limited by synchronously
 sonic vibration acting on the weld that is synchronously coupled with the additive man-
ufacturing process. The second is the fixed ultrasonic vibration acting on the substrate.
The third is the ultrasonic vibration after welding, which is called ultrasonic impact
treatment. The scope of application of the first ultrasonic vibration process is limited by
Materials 2022, 15, 5168 synchronously coupled ultrasonic vibration equipment, so the structures of deposited 3 of 17
components are generally simple and regular. The third ultrasonic vibration process is
ultrasonic impact treatment after welding, so its influence on the weld microstructure is
not as strong
coupled as thevibration
ultrasonic first two ultrasonic
equipment, vibration processes.of
so the structures Therefore,
deposited this paper adopted
components are
the second ultrasonic vibration process.
generally simple and regular. The third ultrasonic vibration process is ultrasonic impact
At present,
treatment the research
after welding, on ultrasonic-assisted
so its influence additive manufacturing
on the weld microstructure generally
is not as strong as the
investigates
first the effect
two ultrasonic of ultrasonic
vibration processes. vibration
Therefore, on
thisthe microstructure
paper and mechanical
adopted the second ultrasonic
properties
vibration of additively manufactured parts under the condition of fixed ultrasonic fre-
process.
quency and amplitude,
At present, andon
the research there are a few studies
ultrasonic-assisted on themanufacturing
additive effect of ultrasonic frequency.
generally inves-
Therefore, this paper adopted the second ultrasonic vibration process to
tigates the effect of ultrasonic vibration on the microstructure and mechanical propertiesstudy the effect
of
of ultrasonic
additively amplitude,parts
manufactured wireunder
feed speed, and welding
the condition of fixedspeed on the
ultrasonic weld morphology
frequency and ampli-
and microstructure
tude, under
and there are a few in-process
studies on the ultrasonic vibration.
effect of ultrasonic The regression
frequency. model
Therefore, of av-
this paper
erage grain
adopted size versus
the second wire vibration
ultrasonic feed speed, welding
process speed,
to study theand ultrasonic
effect amplitude
of ultrasonic was
amplitude,
established.
wire The and
feed speed, effects of wire
welding feed on
speed speed, welding
the weld speed, and
morphology andultrasonic amplitude
microstructure under on
in-process
average grainultrasonic vibration.
size were analyzed. The regression model of average grain size versus wire
feed speed, welding speed, and ultrasonic amplitude was established. The effects of wire
2. Materials
feed and Methods
speed, welding speed, and ultrasonic amplitude on average grain size were analyzed.
2.1. Experimental Set-Up and Materials
2. Materials and Methods
In the experiment,
2.1. Experimental the WAAM
Set-Up and Materialssystem assisted with ultrasonic vibration consisted of
three parts, namely the robot control system, the welding system, and the ultrasonic vi-
In the experiment, the WAAM system assisted with ultrasonic vibration consisted
bration system as shown in Figure 1. The KUKA KR5 6-axis robot (Augsburg, Germany)
of three parts, namely the robot control system, the welding system, and the ultrasonic
was employed to control the movement of the welding torch. The heat source of the
vibration system as shown in Figure 1. The KUKA KR5 6-axis robot (Augsburg, Germany)
welding system was supplied by the TPS4000-CMT welding machine (Fronius, Wels,
was employed to control the movement of the welding torch. The heat source of the welding
Austria). A communication connection was established between the 6-axis robot and the
system was supplied by the TPS4000-CMT welding machine (Fronius, Wels, Austria). A
welding machine. Thus, the start and end of the welding and the feeding of the welding
communication connection was established between the 6-axis robot and the welding
wire can be controlled by the robot control system.
machine. Thus, the start and end of the welding and the feeding of the welding wire can be
controlled by the robot control system.

Figure 1. (a) Experiment platform of the wire and arc additive manufacturing (WAAM); (b) schematic
Figure 1. (a) Experiment platform of the wire and arc additive manufacturing (WAAM); (b) sche-
diagram of WAAM assisted with ultrasonic vibration.
matic diagram of WAAM assisted with ultrasonic vibration.
The welding wire used for deposition was ER4043 Al alloy wire with a diameter of 1.2 mm.
The substrate material was 6061 Al alloy with dimensions of 200 mm × 180 mm × 5 mm. The
chemical compositions of the welding wire and substrate material are shown in Table 1.
Before WAAM, the substrate was wiped with alcohol to clean its surface stains. The surface
was then sanded with sandpaper to remove the surface oxide film. The shielding gas
(high-purity argon) flow rate was set at 20 L/min.
 The welding wire used for deposition was ER4043 Al alloy wire with a diameter of
1.2 mm. The substrate material was 6061 Al alloy with dimensions of 200 mm × 180 mm ×
5 mm. The chemical compositions of the welding wire and substrate material are shown
in Table 1. Before WAAM, the substrate was wiped with alcohol to clean its surface
stains. The surface was then sanded with sandpaper to remove the surface oxide film.
Materials 2022, 15, 5168 4 of 17
The shielding gas (high-purity argon) flow rate was set at 20 L/min.

Table 1. Chemical compositions of the wire and substrate (wt. %).

Table 1. Chemical compositions of the wire and substrate (wt. %).
Material Si Fe Cu Mn Mg Zn Cr Ti Al
Material Si FeER4043 Cu
4.5–6 Mn
0.8 0.3 Mg 0.05 Zn
0.05 Cr
0.1 Ti 0.2 AlBal.
ER4043 4.5–6 0.8 0.3 0.05 0.05 0.1 0.2
0.04–0.3 Bal.
6061 Al 0.4–0.8 0.7 0.15–0.4 0.15 0.8–1.2 0.25 0.15 Bal.
6061 Al 0.4–0.8 0.7 0.15–0.4 0.15 0.8–1.2 0.25 0.04–0.35 5 0.15 Bal.

The ultrasonic vibration system consists of an ultrasonic generator, an ultrasonic

The ultrasonic vibration system consists of an ultrasonic generator, an ultrasonic
vibrator, and a clamp. The ultrasonic vibrator was perpendicular to the contact surface of
vibrator, and a clamp. The ultrasonic vibrator was perpendicular to the contact surface
the substrate through the clamp, which ensured the effective supply of ultrasonic vibra-
of the substrate through the clamp, which ensured the effective supply of ultrasonic
tion to the substrate. The ultrasonic generator can convert alternating current into a
vibration to the substrate. The ultrasonic generator can convert alternating current into a
high-frequency alternating current signal. Ultrasonic waves are formed when the output
high-frequency alternating current signal. Ultrasonic waves are formed when the output
frequencyexceeds
frequency exceeds20 20kHz.
kHz.
During the transmission
During the transmission of of ultrasonic
ultrasonic waves,
waves, there
there will
will be
be anan alternating
alternating cycle
cycle ofof
positive and negative pressures. When the ultrasonic waves are
positive and negative pressures. When the ultrasonic waves are in a positive phase, thein a positive phase, the
ultrasonic waves
ultrasonic waves will
will squeeze
squeeze thethe molecules
molecules of of the
the medium,
medium, compressing
compressing the the distance
distance
between the molecules to increase their density. When in a negative
between the molecules to increase their density. When in a negative phase, the molecular phase, the molecular
spacingbecomes
spacing becomeslarger
largerandandthethedensity
densityof ofthe
themedium
mediumdecreases.
decreases.The Theultrasonic
ultrasonicvibrator
vibrator
is generally composed of ultrasonic transducers, horns, and tool
is generally composed of ultrasonic transducers, horns, and tool heads. The ultrasonic heads. The ultrasonic
transducer can
transducer can convert
convert thethe high-frequency
high-frequency electrical
electrical signal
signal generated
generated by by the
the ultrasonic
ultrasonic
generator into mechanical power and transmit it with less
generator into mechanical power and transmit it with less mechanical loss. Themechanical loss. The horn
hornin-
creases the amplitude or velocity of mechanical vibration and focuses
increases the amplitude or velocity of mechanical vibration and focuses the ultrasonic the ultrasonic vi-
bration energy
vibration energyonto
ontoaasmaller
smallerarea.
area. The
Thefunction
function of of the
the tool
tool head
head is to directly
is to directly contact
contact the
the
surfaceof
surface ofthe
thesubstrate
substrateand andtransmit
transmitthe themechanical
mechanicalvibration
vibrationto tothe
themolten
moltenpool.
pool. Due
Due toto
the large energy generated by ultrasonic vibration, stainless steel, copper,
the large energy generated by ultrasonic vibration, stainless steel, copper, or titanium alloys or titanium
alloys
are are generally
generally used asused as the materials
the materials of the
of the tool head.tool
Thehead.
toolThe
headtool
usedhead used
in this in this ex-
experiment
isperiment
made ofisstainless
made ofsteel,
stainless
andsteel, and its
its shape is ashape
circleiswith
a circle with a diameter
a diameter of 10 mm, of 10 mm, as
as shown
shown
in Figurein2.Figure 2.

Figure2.2.Ultrasonic
Figure Ultrasonicgenerator
generatorand
andultrasonic
ultrasonicvibrator.
vibrator.

2.2. Experimental Method

In WAAM, the main welding parameters include wire feed speed, welding speed
(torch travel speed), welding current, welding voltage, and shielding gas flow rate. Wire
feed speed is directly related to welding current. The heat input increases with increasing
wire feed speed, resulting in deeper penetration. Welding speed is related to the diffusion
of heat. This will affect bead size and penetration. For ultrasonic vibration auxiliary
equipment, the variable parameters mainly include frequency, amplitude, and distance
between the ultrasonic vibrator and the beads, etc.
To study the effect of in-process ultrasonic vibration on wire and arc additive manufac-
tured parts, the single-bead multi-layer thin-walled samples were fabricated. The welding
 feed speed is directly related to welding current. The heat input increases with increasi
wire feed speed, resulting in deeper penetration. Welding speed is related to the diff
sion of heat. This will affect bead size and penetration. For ultrasonic vibration auxilia
equipment, the variable parameters mainly include frequency, amplitude, and distan
between the ultrasonic vibrator and the beads, etc.
Materials 2022, 15, 5168 5 of 17
To study the effect of in-process ultrasonic vibration on wire and arc additive ma
ufactured parts, the single-bead multi-layer thin-walled samples were fabricated. T
welding torch followed a reciprocating deposition path to counteract the bulge at t
torch followed start
a reciprocating deposition
of the arc and sag at thepath to the
end of counteract theinbulge
arc (shown at 3).
Figure theThe
start of the
stick out of the w
arc and sag at the
wasend
12 of theThe
mm. arc ultrasonic
(shown ingenerator
Figure 3).was
Theturned
stick out of the wire
on before wasand
welding 12 mm.
turned off af
The ultrasonic generator was turnedThe
each layer deposition. ondistance
before welding
between and turned off
the ultrasonic after each
vibrator layer
and the beads was
deposition. Thetodistance
105 mm. between the ultrasonic vibrator and the beads was set to 105 mm.

Figure 3. (a) Thin-walled parts fabricated with ultrasonic vibration; (b) thin-walled parts fabrica
Figure 3. (a) Thin-walled parts fabricated with ultrasonic vibration; (b) thin-walled parts fabricated
without ultrasonic vibration.
without ultrasonic vibration.
Since additive manufacturing is a layer-by-layer accumulation process; the study
Since additive manufacturing is a layer-by-layer accumulation process; the study of
the forming laws of single-bead single-layer parts is the basis for the deposition of t
the forming laws of single-bead single-layer parts is the basis for the deposition of the
single-bead multi-layer parts. Therefore, in the paper, the one-factor-at-a-time meth
single-bead multi-layer parts. Therefore, in the paper, the one-factor-at-a-time method
was adopted to study the effect of welding parameters on the formation of single-be
was adopted to study the effect of welding parameters on the formation of single-bead
single-layer parts under ultrasonic vibration. Excluding the unstable arc-starting po
single-layer parts under ultrasonic vibration. Excluding the unstable arc-starting point and
and arc-extinguishing point, the weld width and reinforcement of the stable part of ea
arc-extinguishing point, the weld width and reinforcement of the stable part of each weld
weld were measured three times by the vernier caliper. The studied welding paramete
were measured include
three times
wireby thespeed,
feed vernier caliper.speed,
welding The studied weldingamplitude.
and ultrasonic parametersBased
include
on the expe
wire feed speed, welding speed, and ultrasonic amplitude. Based on the experimental
mental results, the mechanism of ultrasonic vibration on the spreading motion of t
results, the mechanism
weld pool ofwas
ultrasonic
revealed.vibration on the spreading motion of the weld pool
was revealed. Then, the single-bead multi-layer thin-walled parts were fabricated with and wi
Then, the single-bead
out ultrasonicmulti-layer
vibration.thin-walled parts were
To further observe fabricated withofand
the microstructure thewithout
sample, the depo
ultrasonic vibration. To further observe
ited thin-walled the microstructure
parts were cut with a wireof the sample,
electric the deposited
discharge machine,thin-
then groun
walled parts were cut with
polished, a wireand
etched, electric discharge
finally cleanedmachine, thentoground,
with alcohol polished, etched,
obtain metallographic samples. T
and finally cleaned with alcohol
microstructure to obtain
was metallographic
observed using a Leicasamples. The microstructure
CM4M optical was Microsy
microscope (Leica
observed using tems,
a Leica CM4MGermany),
Wetzlar, optical microscope
and grain(Leica Microsystems,
size measurements Wetzlar,
were Germany),
performed using Image-P
and grain size measurements
Plus software. The were performed
BBD using Image-Pro
(Box–Behnken) method wasPlus usedsoftware. Thethe
to carry out BBD experimen
(Box–Behnken) design
methodofwas the used to carry
deposition of out the experimental
thin-walled parts anddesign of the deposition
the regression of
model of average gra
thin-walled parts and the regression model of average grain size versus wire feed speed,
size versus wire feed speed, welding speed, and ultrasonic amplitude was obtained.
welding speed, and ultrasonic amplitude was obtained.

2.3. Determination of Ultrasonic Vibration Frequency

Since the ultrasonic amplitude is the main parameter studied in this paper, it is
necessary to determine the appropriate ultrasonic frequency before the experiment. In
the experiment, the ultrasonic vibrator applied mechanical vibration to the surface of the
substrate, and the ultrasonic vibration was transmitted into the molten pool through the
substrate, which stirred the molten pool.
Two nonlinear effects of cavitation and acoustic streaming are caused by the intro-
duction of periodic positive and negative pressure and violent motion in the molten pool
by ultrasonic vibration. The reduction of pores in the molten pool and the reduction of
the average grain size are mainly attributed to ultrasonic cavitation. Cavitation results in
the creation of tiny bubbles or small cavities, which then undergo a series of processes
including inertial growth, pulsation, and collapse in the molten pool [37]. When the bubble
collapses, high temperature and pressure will be generated inside the bubble in a very
short time, and a strong shock wave will be generated instantly, causing the stirring and
Materials 2022, 15, 5168 6 of 17

mixing of the molten pool. Cavitation occurs when the acoustic pressure exceeds the
cavitation threshold.
The following Rayleigh–Plesset equation describes the nonlinear response of the
bubble in a fluid to a driving pressure field [38].
"  3k . #
.. 3 .2 1 R0 2σ 4µ R
RR + R = Pv − P∞ + Pg0 − − (1)
2 ρ R R R

where R is the radius of the bubble at time t, R0 is the initial bubble radius, ρ is the liquid
density, σ is the surface tension, µ is the dynamic viscosity of the liquid, k is the polytropic
constant for gas, Pv is the vapor pressure in the bubble, P∞ is the pressure at infinity, and
Pg0 is the gas pressure at the reference radius R0 .
In acoustic cavitation, owing to the effect of the acoustic field on the growth and
collapse of bubbles, the P∞ and Pg0 should be modified [39,40].
"   3k . #
..3 .2 1 2σ R0 2σ 4µ R
RR + R = Pv − ( P0 + P(t)) + P0 − Pv + − − (2)
2 ρ R R R R

where P0 is the static pressure and P(t) is the applied acoustic field.
Numerical simulation research based on Equation (2) found that lower ultrasonic
frequency promotes ultrasonic cavitation. Tian et al. [41] found that with the increase of
ultrasonic frequency, the period of cavitation bubbles becomes longer. Huang et al. [42]
pointed out that with the increase of ultrasonic frequency, the amplitude of cavitation
bubbles decreases, and the higher the frequency, the faster the attenuation of ultrasonic
energy. Therefore, more energy is required to provide the same intensity of cavitation.
Wei et al. [43] found that the radius of the cavitation bubbles increases to about 109, 80, 67,
32, 15, and 8 times the initial size during the ultrasonic frequency from 15, 20, 30, 50, and 100
to 200 kHz, respectively. The corresponding collapse period also increases. Experiments
showed that the cavitation bubbles are transient and they shrink and collapse within
1.2 periods when the ultrasonic frequency is lower than 20 kHz, while the cavitation bubbles
are stable and need more periods to shrink, expand, and collapse when the ultrasonic
frequency is higher than 30 kHz. This is because for high-frequency ultrasonic vibration,
the expansion time of the bubble is shortened, and the bubble radius does not have enough
time to reach the cavitation threshold, resulting in too little time available for the bubble to
collapse. It therefore periodically experiences collapse, expansion, oscillation, and collapse.
Therefore, the ultrasonic vibration frequency was selected as 20 kHz in this paper.

3. Results and Discussion

3.1. Effect of Ultrasonic Amplitude on Weld
To study the effect of ultrasonic amplitude on the forming of the single-bead single-layer
weld, the wire feed speed Vw and welding speed Vt were kept unchanged (Vw = 4.5 m/min,
Vt = 0.3 m/min), and only the ultrasonic amplitude was changed. The specific welding
parameter settings of the ultrasonic amplitude experiment are shown in Table 2.
As shown in Figure 4, the surface of sample 1 without ultrasonic vibration is covered
with the metallic luster, its weld width is 5 mm, and reinforcement is 4.2 mm. Compared
with sample 1, the surface of sample 2 obviously loses the metallic luster and capillary
waves appear (shown in the enlarged view of area A). Its weld width is 5.7 mm and
reinforcement is 4 mm. Capillary waves are generated by the violent oscillation of the
molten pool caused by ultrasonic vibration. However, the capillary wave is small in size
and will be remelted by the next deposition, so it does not affect the formed appearance [44].
A small number of tiny pores appear on the surface of sample 3, and a larger number of
pores gather on the surface at the arc-end point. Its weld width is 5.8 mm and reinforcement
is 4 mm. It can be seen in the enlarged view of area B that the capillary waves on the surface
of sample 3 are more obvious than that of sample 2, which indicates that the ultrasonic
Materials 2022, 15, 5168 7 of 17

vibration of sample 3 is more severe. The pores on the surface are generated because the
bubbles inside the molten pool escape upward under the action of ultrasonic vibration,
and the bubbles that do not fully escape before the solidification of the metal molten pool
form open pores. It can be concluded from samples 1, 2, and 3 that the ultrasonic vibration
makes the weld width wider and the reinforcement lower. The enlarged weld width is
attributed to the stirring and oscillation effect of ultrasonic vibration on the molten pool,
which accelerates the flow inside the molten pool. Due to the same wire feed speed and
welding speed, the deposition amount of the wire per unit time is equal. Thus, the increase
in the weld width will naturally reduce the reinforcement.
Table 2. Specific welding parameter settings of the ultrasonic amplitude experiment.

Shielding Gas
Wire Feed Welding Speed
S.NO. Amplitude (µm) Flow Rate
Speed (m/min) (m/min)
(L/min)
1 4.5 0.3 0 20
2 4.5 0.3 20 20
3 4.5 0.3 25 20
Materials 2022, 15, x FOR PEER REVIEW 4 4.5 0.3 30 208 of 19
5 4.5 0.3 35 20

Figure4.4.Effect
Figure Effectof
ofdifferent
different ultrasonic
ultrasonic amplitudes
amplitudeson
onweld:
weld:(a)(a)
sample 1; (b)
sample sample
1; (b) 2; (c)
sample 2; sample 3; 3;
(c) sample
(d) sample 4; (e) sample 5.
(d) sample 4; (e) sample 5.

Asshown
As shownin inFigure
Figure 4d,e,
4d,e, it
it can
can be
be seen
seenthat
thatthe
theincompletely
incompletelysolidified weld
solidified weldcracks
cracks
because of the excessive energy of ultrasonic vibration. Part of the cracked structure is
because of the excessive energy of ultrasonic vibration. Part of the cracked structure is
shown in Figure 5. It can be seen that the cracked structure has begun to take the shape of
shown in Figure 5. It can be seen that the cracked structure has begun to take the shape of
the weld, which indicates that the crack occurs in the spreading stage of the molten pool.
the weld, which indicates that the crack occurs in the spreading stage of the molten pool.
The greater the ultrasonic amplitude, the greater the energy generated by the ultrasonic
The greater the ultrasonic amplitude, the greater the energy generated by the ultrasonic
vibration. Thus, when the molten pool is not completely solidified, the weld will crack
vibration. Thus, when the molten pool is not completely solidified, the weld will crack when
when the tensile stress generated by ultrasonic vibration during the vibration period ex-
the tensile stress generated by ultrasonic vibration during the vibration period exceeds
ceeds the ultimate tensile strength that can be endured in this state. Therefore, it is nec-
the ultimate tensile strength that can be endured in this state. Therefore, it is necessary to
essary to select the appropriate ultrasonic amplitude to achieve good weld formation
select the appropriate ultrasonic amplitude to achieve good weld formation when applying
when applying ultrasonic vibration.
ultrasonic vibration.
 the weld, which indicates that the crack occurs in the spreading stage of the molten pool.
The greater the ultrasonic amplitude, the greater the energy generated by the ultrasonic
vibration. Thus, when the molten pool is not completely solidified, the weld will crack
when the tensile stress generated by ultrasonic vibration during the vibration period ex-
Materials 2022, 15, 5168
ceeds the ultimate tensile strength that can be endured in this state. Therefore, it is8 of
nec-
17
essary to select the appropriate ultrasonic amplitude to achieve good weld formation
when applying ultrasonic vibration.

Figure5.5.Part
Figure Partof
ofthe
thecollapsed
collapsedstructure.
structure.

3.2.
3.2.Effect
EffectofofWire
WireFeed
FeedSpeed
Speedand
andWelding
WeldingSpeed
SpeedononWeld
Weld
In
In the experiment, the wire feed speed was setto
the experiment, the wire feed speed was set to44m/min,
m/min, 55 m/min,
m/min, 66 m/min,
m/min,and
and7
7m/min
m/min respectively, and the welding speed was set to 0.3 m/min, 0.45
respectively, and the welding speed was set to 0.3 m/min, 0.45 m/min, and 0.6 m/min, and
0.6 m/min
m/min respectively.
respectively. Based
Based onon
thetheresults
resultsofofSection
Section3.1,
3.1, the
the ultrasonic
ultrasonic amplitude
amplitude was
was
chosen
chosen to be 20 μm. The specific welding parameter settings of the wire feedspeed
to be 20 µm. The specific welding parameter settings of the wire feed speedand
and
welding speed experiment are shown in
welding speed experiment are shown in Table 3. Table 3.
Table 3. Specific welding parameter settings of the wire feed speed and welding speed experiment.

Wire Feed Welding Shielding

Welding Welding Heat Input Amplitude
S. NO. Speed Speed Gas Flow
Voltage (V) Current (A) (J/mm) (µm)
(m/min) (m/min) Rate (L/min)
1 4 0.3 12 69 149.04 20 20
2 5 0.3 12.6 85 192.78 20 20
3 6 0.3 15.3 127 349.758 20 20
4 7 0.3 16.8 158 477.792 20 20
5 4 0.45 12 69 99.36 20 20
6 5 0.45 12.6 85 128.52 20 20
7 6 0.45 15.3 127 233.172 20 20
8 7 0.45 16.8 158 318.528 20 20
9 4 0.6 12 69 74.52 20 20
10 5 0.6 12.6 85 96.39 20 20
11 6 0.6 15.3 127 174.879 20 20
12 7 0.6 16.8 158 238.896 20 20

Figure 6 shows the weld macromorphology of the wire feed speed and welding speed
experiment. Figure 7 displays the effects of wire feed speed and welding speed on weld
width and reinforcement. As shown in Figures 6 and 7, with the increase in wire feed
speed, it can be clearly seen that the weld width increases, while the reinforcement first
increases slightly and then decreases. The analysis was carried out with the experiment
data at the welding speed of 0.45 m/min. When the wire feed speed (Vw ) was 4 m/min,
the average weld width was only 3.85 mm, and the reinforcement was 2.98 mm. When
Vw = 5 m/min, the average weld width was 5.17 mm, and the reinforcement was 3.45 mm.
When Vw = 6 m/min, the average weld width increased to 7.97 mm, and the reinforcement
decreased to 2.75 mm. The arc ending point of the weld was obviously depressed. When
Vw = 7 m/min, the weld width could reach 10.67 mm, and the reinforcement was only
2.57 mm. The obvious cracks appeared on the surface of the arc ending point depression.
 Materials 2022, 15, 5168 9 of 17

Two main parameters affect weld formation: one is the ratio of wire feed speed to welding
speed, which represents the amount of welding wire deposited per unit time and unit
Materials 2022, 15, x FOR PEER REVIEW 10 wire
length, and the other is heat input, which directly affects the melting of the welding of 19
and the substrate, and has a great influence on the flow of the molten pool. Therefore, we
analyze and explain the weld formation from these two aspects.

(a) Welding speed = 0.3 m/min (b) Welding speed = 0.45 m/min

Materials 2022, 15, x FOR PEER REVIEW 11 of 19

it is clear that (c)

metal deposition
Welding speed =amount
0.6 m/min is dominant among these two factors, so the re-
inforcement is reduced in the end.
Figure 6. Weld macromorphology of the wire feed speed and welding speed experiment: (a) sam-
Figure 6. Weld macromorphology of the wire feed speed and welding speed experiment: (a) samples
ples 1, 2, 3, and 4; (b) samples 5, 6, 7, and 8; (c) samples 9, 10, 11, and 12.
1, 2, 3, and 4; (b) samples 5, 6, 7, and 8; (c) samples 9, 10, 11, and 12.

The heat input can be calculated with the following equation:

𝜂𝑈𝐼
Heat input (J/mm) = (3)
𝑉𝑡
where η is the arc thermal efficiency, which was set to 0.9 in this experiment, U (V) is the
welding voltage, I (A) is the welding current, and Vt (m/min) is the welding speed. Wire
feed speed is directly related to welding current. As the wire feed speed increases, the
welding current and voltage increase, so the heat input increases. Higher heat input
means longer molten pool solidification time and reduced molten pool viscosity, result-
ing in better molten pool fluidity. Meanwhile, the ultrasonic vibration induces agitation
and oscillation of the molten pool, which contributes to the flow of the molten pool. Thus,
the weld width increases. For the reinforcement, when the welding speed is the same; the
increase in heat input from wire feed speed of 4 m/min to 5 m/min is small (shown in
Table 3), but the increased amount of wire is sufficient. Therefore, when the wire feed
speed increases from 4 m/min to 5 m/min, both the weld width and reinforcement are
increased.
It is worth noting that the slopes are not the same in Figure 7, which reflects the de-
Figure
gree Effects of wire
of7.7.increase feed speedAand welding
slopespeed on weld width (a) and reinforcement (b).
Figure Effects ofinwire
heat input.
feed speedlarge
and welding means
speeda on
large
weldincrease
width in(a)heat
and input, which
reinforcement (b).
results in better molten pool flow. Thus, a wider weld width and a lower reinforcement
The heat input can be calculated with the following equation:
areEffect
3.3. obtained. of Ultrasonic Vibration on the Average Grain Size of the Microstructure
As shown in Figure 7, it can be seen that with the increase ηU I in welding speed, both the
To study the effect of ultrasonic
Heat inputvibration
(
weld width and reinforcement decrease when keepingVthe J/mm ) =on the average grain size of the (3) micro-
t wire feed speed constant. For
structure,
weld width, thewhen
single-bead
the weldingmulti-layer thin-walled
speed increases, parts
and the wereof
amount fabricated (shownper
metal deposition in Figure
3).
unit length decreases. Meanwhile, according to Equation (3), the heat input per unitoscilla-
When the total number of deposited layers is the same, due to the stirring and
tion of the
length moltenresulting
decreases, pool caused by ultrasonic
in a weakened molten vibration, the Therefore,
pool fluidity. layer height theof thewidth
weld thin-walled
part with ultrasonic
is reduced. vibration
For reinforcement, is lower
reduced metal and its layeramount
deposition widthresults
is wider than rein-
in smaller that of the
forcement, while
thin-walled part reduced
withoutmolten pool fluidity
ultrasonic results
vibration. Thein microstructure
larger reinforcement. However, parts
of thin-walled
Materials 2022, 15, 5168 10 of 17

where η is the arc thermal efficiency, which was set to 0.9 in this experiment, U (V) is the
welding voltage, I (A) is the welding current, and Vt (m/min) is the welding speed. Wire
feed speed is directly related to welding current. As the wire feed speed increases, the
welding current and voltage increase, so the heat input increases. Higher heat input means
longer molten pool solidification time and reduced molten pool viscosity, resulting in better
molten pool fluidity. Meanwhile, the ultrasonic vibration induces agitation and oscillation
of the molten pool, which contributes to the flow of the molten pool. Thus, the weld width
increases. For the reinforcement, when the welding speed is the same; the increase in heat
input from wire feed speed of 4 m/min to 5 m/min is small (shown in Table 3), but the
increased amount of wire is sufficient. Therefore, when the wire feed speed increases from
4 m/min to 5 m/min, both the weld width and reinforcement are increased.
It is worth noting that the slopes are not the same in Figure 7, which reflects the degree
of increase in heat input. A large slope means a large increase in heat input, which results in
better molten pool flow. Thus, a wider weld width and a lower reinforcement are obtained.
As shown in Figure 7, it can be seen that with the increase in welding speed, both
the weld width and reinforcement decrease when keeping the wire feed speed constant.
For weld width, when the welding speed increases, and the amount of metal deposition
per unit length decreases. Meanwhile, according to Equation (3), the heat input per
unit length decreases, resulting in a weakened molten pool fluidity. Therefore, the weld
width is reduced. For reinforcement, reduced metal deposition amount results in smaller
reinforcement, while reduced molten pool fluidity results in larger reinforcement. However,
it is clear that metal deposition amount is dominant among these two factors, so the
reinforcement is reduced in the end.

3.3. Effect of Ultrasonic Vibration on the Average Grain Size of the Microstructure
To study the effect of ultrasonic vibration on the average grain size of the microstruc-
ture, the single-bead multi-layer thin-walled parts were fabricated (shown in Figure 3).
When the total number of deposited layers is the same, due to the stirring and oscillation
of the molten pool caused by ultrasonic vibration, the layer height of the thin-walled part
with ultrasonic vibration is lower and its layer width is wider than that of the thin-walled
part without ultrasonic vibration. The microstructure of thin-walled parts was analyzed,
and the regression model of wire feed speed, welding speed, ultrasonic amplitude, and
average grain size was established. A good prediction of the average grain size of WAAM
of aluminum alloy parts was achieved.
Based on the results of Section 3.1, when the ultrasonic amplitude was 30 µm, the weld
cracked under excessive ultrasonic energy. Therefore, the maximum ultrasonic amplitude
in this experiment was set to 28 µm. The shielding gas flow rate was 20 L/min. Each
sample was deposited with 6 layers, and the middle was taken to make a metallographic
sample. The grain size measurements were taken using Image-Pro Plus software.
The BBD (Box–Behnken) method was applied to obtain the average grain size regres-
sion model. Each factor took three levels and encoded them with (−1, 0, +1), where 0 is
the center point and –1 and +1 represent the upper and lower limits, respectively. The
coding generated by Design Export software is given in Table 4 based on the single-bead
single-layer experiment. The resulting experimental design matrix is shown in Table 5.
After 17 metallographic samples were prepared, each sample was observed and the average
grain size was measured. In order to ensure the validity of the test data, each sample was
intercepted at two different positions for observation.
Materials 2022, 15, 5168 11 of 17

Table 4. Coding for factor and level.

Wire Feed Speed Welding Speed Ultrasonic

Factor Level
(m/min) (m/min) Amplitude (µm)
−1 4.5 0.3 20
0 5 0.4 24
+1 5.5 0.5 28

Table 5. Experiment design matrix and the response.

Wire Feed Welding Speed Ultrasonic Average Grain

Exp. NO.
Speed (m/min) (m/min) Amplitude (µm) Size (µm)
1 4.5 0.4 28 13.3
2 5 0.4 24 11.4265
3 4.5 0.3 24 14.3
4 5 0.3 20 15.3
5 5.5 0.3 24 15.5
6 4.5 0.4 20 12.41
7 5.5 0.4 20 13.11
8 5 0.4 24 11.7
9 5 0.5 28 13.7
10 4.5 0.5 24 10.84
11 5.5 0.5 24 12.74
12 5 0.4 24 11.9
13 5 0.4 24 11.45
14 5 0.3 28 15.06
15 5 0.4 24 11.7
16 5.5 0.4 28 13.55
17 5 0.5 20 11.51

The microstructure of some samples is shown in Figure 8. Based on our previous

research, it can be concluded that there were a lot of columnar crystals in the microstructure
of the sample without ultrasonic vibration. Due to the excessively large temperature
gradient during the WAAM, a large number of dendrite structures existed, and a large
amount of Al-Si eutectic structure existed in the interstitial space of the dendrite structure.
The aluminum eutectic structure in the absence of ultrasonic vibration was mainly in the
form of long strips. Our previous research showed that the average grain size of the sample
without ultrasonic vibration was measured to be 30.2 µm, and the largest grain size could
reach 101.98 µm. As shown in Figure 8, it can be seen that the microstructure of the samples
with ultrasonic vibration was mainly equiaxed crystals, with a small number of columnar
crystals existing. Their average grain size is smaller than that of the samples without
ultrasonic vibration (shown in Table 5). It is worth noting that our previous studies have
shown that excessive ultrasonic amplitude can lead to larger grain size and cracks. It may
be because as the ultrasonic amplitude increases, the energy generated by the ultrasonic
waves in the medium is higher, and the increase in heat input causes the cooling rate inside
the molten pool to slow down, resulting in grain coarsening.
 Materials
Materials 2022,
2022, 15, 15, 5168
x FOR PEER REVIEW 12 of 13
17 of 19

Figure
Figure 8. Microstructureofofsome
8. Microstructure somesamples:
samples: (a)
(a) Exp.
Exp. 1;
1; (b)
(b)Exp.
Exp.5;5;(c)
(c)Exp.
Exp.7;7;(d) Exp.
(d) 8; 8;
Exp. (e)(e)
Exp. 10;10; (f)
Exp.
(f) Exp.
Exp. 13. 13.

The regression equation was obtained as follows:

The regression equation was obtained as follows:
Average grain size
𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑔𝑟𝑎𝑖𝑛 𝑠𝑖𝑧𝑒
= 126.275 − 17.2315A − 168.5505B − 3.2308C + 3.5AB (4)
− 0.05625AC + 1.51875BC= 126.275+−1.8194A
17.2315𝐴 2 +− 168.5505𝐵
125.485B − 3.2308𝐶2+ 3.5𝐴𝐵
2 + 0.062647C
(4)
− 0.05625𝐴𝐶 + 1.51875𝐵𝐶 + 1.8194𝐴2 + 125.485𝐵2 + 0.062647𝐶 2
The ANOVA results of the regression model are shown in Table 6. The model F-value
of 28.64 implies the model is significant. There is only a 0.01% chance that an F-value
The
this ANOVA
large results
could occur due of the regression
to noise. The p-values model are 0.05
less than shown in Table
indicate model 6.
termsThearemodel
significant.
F-value of 28.64 In this
implies A, B,model
case, the A2significant.
C, BC, is , B2 , and C2 are significant
There is onlymodel terms.
a 0.01% chanceValues
that an
greater than 0.1000 indicate the model terms are not significant. The
F-value this large could occur due to noise. The p-values less than 0.05 indicate model Lack of fit F-value
2
of 6.56
terms are issignificant.
not significant.
In thisThe coefficient
case, A, B, C,ofBC, A2, B2, and R
determination C2 of
are0.9736 meansmodel
significant that theterms.
regression model clarifies 97.36% of all deviations.
Values greater than 0.1000 indicate the model terms are not significant. The Lack of fit
Figure 9 shows a comparison of the average grain size predicted by2 Equation (4) with
F-value of 6.56 is not significant. The coefficient of determination R of 0.9736 means that
the actual measured value. When the predicted value is closer to the actual value, the
thepoints
regression
will bemodel
closer clarifies 97.36%
to the line. It canofbeall deviations.
seen from Figure 9 that the most of points are
distributed near the straight line, indicating that Equation (4) can well reflect the influence
Table 6. ANOVA
of each weldingresults of theon
parameter regression model.
the average grain size.
As shown in Figure 10, the average grain size keeps getting larger as wire feed speed
Source Sum of Squares Mean Square F-Value p-Value
increases and welding speed decreases. In the case of high wire feed speed and low
Model
welding speed, the amount 34.08of wire melted into 3.79the molten pool 28.64 0.0001
per unit length reaches the
A-wire feed
maximum, resulting in high heat input. At the same time, the low welding speed makes
the speed
2.05 2.05 15.51
welding heat source stay above the molten pool for too long, and the cooling speed is
0.0056
low. These two reasons lead to the growth of grains inside the weld under the influence of
B-welding speed 16.16 16.16 122.24 <0.0001
prolonged heat input. It can be clearly seen from Figure 11 that the interaction of ultrasonic
C-ultrasonic
amplitude and am- wire feed speed has no significant effect on the average grain size. As
1.34 1.34 10.17 0.0153
plitude
shown in Figure 12, the larger slope of the 3D response surface and the elliptical 2D contour
AB that the interaction
indicates 0.1225of ultrasonic 0.1225amplitude and welding 0.9267 0.3678
speed on the average
grainACsize is significant.0.0506
It can be seen that0.0506
with the increase of0.3830
ultrasonic amplitude, the
0.5556
average grain size shows a trend of decreasing first and then increasing.
BC 1.48 1.48 11.17 0.0124
A2 0.8711 0.8711 6.59 0.0372
B2 6.63 6.63 50.15 0.0002
C2 4.23 4.23 32.00 0.0008
Residual 0.9253 0.1322
 Materials 2022, 15, 5168 13 of 17

Table 6. ANOVA results of the regression model.

Source Sum of Squares Mean Square F-Value p-Value

Model 34.08 3.79 28.64 0.0001
A-wire feed speed 2.05 2.05 15.51 0.0056
B-welding speed 16.16 16.16 122.24 <0.0001
C-ultrasonic amplitude 1.34 1.34 10.17 0.0153
Materials 2022, 15, x FOR PEER REVIEW AB 0.1225 0.1225 0.9267 0.3678
14

AC 0.0506 0.0506 0.3830 0.5556

BC 1.48 1.48 11.17 0.0124
Figure 9A2shows a comparison0.8711 of the average
0.8711 grain 6.59
size predicted
0.0372by Equatio
with the actual
B2 measured value. 6.63When the predicted
6.63 value
50.15is closer0.0002
to the actual v
the points will
C2 be closer to the line.
4.23 It can be seen
4.23 from Figure
32.00 9 that the most of point
0.0008
distributedResidual
near the straight line,
0.9253 indicating0.1322
that Equation (4) can well reflect the i
ence of each welding
Lack of fit parameter on the average
0.77 0.26grain size.6.56 0.0504

Figure 9. Comparison of predicted grain size with actual grain size.

Figure 9. Comparison of predicted grain size with actual grain size.

As shown in Figure 10, the average grain size keeps getting larger as wire feed s
increases and welding speed decreases. In the case of high wire feed speed and
welding speed, the amount of wire melted into the molten pool per unit length rea
the maximum, resulting in high heat input. At the same time, the low welding s
makes the welding heat source stay above the molten pool for too long, and the coo
speed is low. These two reasons lead to the growth of grains inside the weld unde
influence of prolonged heat input. It can be clearly seen from Figure 11 that the int
tion of ultrasonic amplitude and wire feed speed has no significant effect on the ave
grain size. As shown in Figure 12, the larger slope of the 3D response surface and th
 Materials 2022, 15, x FOR PEER REVIEW 15 of 19
Materials
Materials 2022, 15, 5168 2022, 15, x FOR PEER REVIEW 15 of 19 14 of 17

Figure 10. (a) 3D response surface of the interaction between wire feed speed and welding speed;
10.2D(a)
Figure (b)
Figure 3D(a)response
10. 3D responsesurface
surface of the
theinteraction
interaction between
between wire
wire feed feed
speed
contour of the interaction between wire feed speed and welding speed.
speed
and and
welding welding speed;
speed;
(b) 2D contour of the interaction between wire feed speed and welding speed.
(b) 2D contour of the interaction between wire feed speed and welding speed.

Materials 2022, 15, x FOR PEER REVIEW 16 of 19

11. (a)11.3D
Figure Figure (a)response surface
3D response surfaceof
of the interaction
the interaction between
between wirespeed
wire feed feedand
speed and ultrasonic
ultrasonic am- ampli-
Figure 11. (a) 3D response surface of the interaction between wire feed speed and ultrasonic am-
plitude; (b) 2D contour of the interaction between wire feed speed and ultrasonic amplitude.
tude; (b) 2D contour
plitude; of theofinteraction
(b) 2D contour between
the interaction between wire feed
wire feed speed
speed and and ultrasonic
ultrasonic amplitude.
amplitude.

12. (a)12.
Figure Figure 3D(a)
response surface
3D response of of
surface thethe
interaction
interaction between welding
between welding speed
speed and and ultrasonic
ultrasonic am- amplitude;
plitude; (b) 2D contour of the interaction between welding speed and ultrasonic amplitude.
(b) 2D contour of the interaction between welding speed and ultrasonic amplitude.
4. Conclusions
In this study, the single-bead single-layer and single-bead multi-layer aluminum
alloy components were fabricated by in-process ultrasonic vibration-assisted WAAM.
The effects of wire feed speed, welding speed, and ultrasonic amplitude on the weld
Materials 2022, 15, 5168 15 of 17

Based on the above analysis, the lack of fit F-value of 6.56 and p-value of 0.0504 are
not very good. Therefore, we removed the insignificant terms to reduce the impact on the
model, resulting in the following simplified regression model:

Average grain size

= 126.025 − 17.1815A − 151.0505B − 3.51205C (5)
+ 1.51875BC + 1.8194A2 + 125.485B2 + 0.062647C2

4. Conclusions
In this study, the single-bead single-layer and single-bead multi-layer aluminum alloy
components were fabricated by in-process ultrasonic vibration-assisted WAAM. The effects
of wire feed speed, welding speed, and ultrasonic amplitude on the weld formation and
microstructure grain size were investigated. The following conclusions can be drawn:
1. With the increase of ultrasonic amplitude, the weld width became wider and the
reinforcement becomes lower. When the ultrasonic amplitude was too large and
exceeded the ultimate tensile strength that could be endured in the state, the weld
cracked without fully spreading.
2. Under ultrasonic vibration with an amplitude of 20 µm, increased welding speed re-
sulted in reduced weld width and reinforcement. With the increase of wire feed
speed, the weld width increased, and reinforcement first increased slightly and
then decreased.
3. Average grain size decreased with decreasing wire feed speed and increasing welding
speed. When the ultrasonic amplitude increased, the average grain size first decreased
and then increased.
4. The significant effect of the interaction between parameters on the average grain size
is as follows: welding speed and ultrasonic amplitude > wire feed speed and welding
speed > wire feed speed and ultrasonic amplitude.

Author Contributions: Conceptualization, J.Z. (Jun Zhang), Y.X. and J.Z. (Jijun Zhang); methodology,
J.Z. (Jun Zhang), Y.X. and J.Z. (Jijun Zhang); software, J.Z. (Jun Zhang) and J.Z. (Jijun Zhang); validation,
J.Z. (Jun Zhang) and J.Z. (Jijun Zhang); visualization J.Z. (Jun Zhang) and J.Z. (Jijun Zhang); formal
analysis, J.Z. (Jun Zhang); investigation, J.Z. (Jun Zhang); data curation, J.Z. (Jun Zhang); writing—
original draft, J.Z. (Jun Zhang); supervision, Y.X., J.C., F.Y. and X.Z.; funding acquisition, Y.X.; writing—
review & editing, Y.X.; project administration, J.C., F.Y. and X.Z.; All authors have read and agreed to the
published version of the manuscript.
Funding: This research was supported by the Natural Science Foundation of Shanghai (Grant No.
20ZR1422600).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The data used to support the findings of this study can be made
available upon reasonable request from the corresponding author.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Williams, S.W.; Martina, F.; Addison, A.C.; Ding, J.; Pardal, G.; Colegrove, P. Wire + Arc Additive Manufacturing. Mater. Sci.
Technol. 2016, 32, 641–647. [CrossRef]
2. Karunakaran, K.P.; Bernard, A.; Suryakumar, S.; Dembinski, L.; Taillandier, G. Rapid Manufacturing of Metallic Objects. Rapid
Prototyp. J. 2012, 18, 264–280. [CrossRef]
3. Liu, D.; Lee, B.; Babkin, A.; Chang, Y. Research Progress of Arc Additive Manufacture Technology. Materials 2021, 14, 1415.
[CrossRef] [PubMed]
4. Singh, S.; Sharma, S.k.; Rathod, D.W. A Review on Process Planning Strategies and Challenges of WAAM. Mater. Today Proc. 2021,
47, 6564–6575. [CrossRef]
Materials 2022, 15, 5168 16 of 17

5. Vimal, K.E.K.; Naveen Srinivas, M.; Rajak, S. Wire Arc Additive Manufacturing of Aluminium Alloys: A Review. Mater. Today
Proc. 2019, 41, 1139–1145. [CrossRef]
6. Elrefaey, A. Effectiveness of Cold Metal Transfer Process for Welding 7075 Aluminium Alloys. Sci. Technol. Weld. Join. 2015,
20, 280–285. [CrossRef]
7. Wu, B.; Pan, Z.; Ding, D.; Cuiuri, D.; Li, H.; Xu, J.; Norrish, J. A Review of the Wire Arc Additive Manufacturing of Metals:
Properties, Defects and Quality Improvement. J. Manuf. Process. 2018, 35, 127–139. [CrossRef]
8. Cunningham, C.R.; Flynn, J.M.; Shokrani, A.; Dhokia, V.; Newman, S.T. Invited Review Article: Strategies and Processes for High
Quality Wire Arc Additive Manufacturing. Addit. Manuf. 2018, 22, 672–686. [CrossRef]
9. Langelandsvik, G.; Akselsen, O.M.; Furu, T.; Furu, T. Review of Aluminum Alloy Development for Wire Arc. Materials 2021,
14, 5370. [CrossRef]
10. Baufeld, B.; Brandl, E.; Van Der Biest, O. Wire Based Additive Layer Manufacturing: Comparison of Microstructure and
Mechanical Properties of Ti–6Al–4V Components Fabricated by Laser-Beam Deposition and Shaped Metal Deposition. J. Mater.
Process. Technol. 2011, 211, 1146–1158. [CrossRef]
11. Brandl, E.; Baufeld, B.; Leyens, C.; Gault, R. Additive Manufactured Ti-6Al-4V Using Welding Wire: Comparison of Laser and Arc
Beam Deposition and Evaluation with Respect to Aerospace Material Specifications. Phys. Procedia 2010, 5, 595–606. [CrossRef]
12. Gu, J.; Ding, J.; Williams, S.W.; Gu, H.; Bai, J.; Zhai, Y.; Ma, P. The Strengthening Effect of Inter-Layer Cold Working and
Post-Deposition Heat Treatment on the Additively Manufactured Al–6.3Cu Alloy. Mater. Sci. Eng. A 2016, 651, 18–26. [CrossRef]
13. Ma, Y.; Cuiuri, D.; Li, H.; Pan, Z.; Shen, C. The Effect of Postproduction Heat Treatment on γ-TiAl Alloys Produced by the
GTAW-Based Additive Manufacturing Process. Mater. Sci. Eng. A 2016, 657, 86–95. [CrossRef]
14. Colegrove, P.A.; Coules, H.E.; Fairman, J.; Martina, F.; Kashoob, T.; Mamash, H.; Cozzolino, L.D. Microstructure and Residual
Stress Improvement in Wire and Arc Additively Manufactured Parts through High-Pressure Rolling. J. Mater. Process. Technol.
2013, 213, 1782–1791. [CrossRef]
15. Huang, J.; Ye, X.; Xu, Z. Effect of Cold Rolling on Microstructure and Mechanical Properties of AISI 301LN Metastable Austenitic
Stainless Steels. J. Iron Steel Res. Int. 2012, 19, 59–63. [CrossRef]
16. Martina, F.; Colegrove, P.A.; Williams, S.W.; Meyer, J. Microstructure of Interpass Rolled Wire + Arc Additive Manufacturing
Ti-6Al-4V Components. Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 2015, 46, 6103–6118. [CrossRef]
17. Martina, F.; Roy, M.J.; Szost, B.A.; Terzi, S.; Colegrove, P.A.; Williams, S.W.; Withers, P.J.; Meyer, J.; Hofmann, M. Residual Stress of
As-Deposited and Rolled Wire+arc Additive Manufacturing Ti–6Al–4V Components. Mater. Sci. Technol. 2016, 32, 1439–1448.
[CrossRef]
18. Wu, B.; Pan, Z.; Ding, D.; Cuiuri, D.; Li, H.; Fei, Z. The Effects of Forced Interpass Cooling on the Material Properties of Wire Arc
Additively Manufactured Ti6Al4V Alloy. J. Mater. Process. Technol. 2018, 258, 97–105. [CrossRef]
19. Colegrove, P.A.; Donoghue, J.; Martina, F.; Gu, J.; Prangnell, P.; Hönnige, J. Application of Bulk Deformation Methods for
Microstructural and Material Property Improvement and Residual Stress and Distortion Control in Additively Manufactured
Components. Scr. Mater. 2017, 135, 111–118. [CrossRef]
20. Cheng, X.; Fisher, J.W.; Prask, H.J.; Gnäupel-Herold, T.; Yen, B.T.; Roy, S. Residual Stress Modification by Post-Weld Treatment
and Its Beneficial Effect on Fatigue Strength of Welded Structures. Int. J. Fatigue 2003, 25, 1259–1269. [CrossRef]
21. Yuan, T.; Kou, S.; Luo, Z. Grain Refining by Ultrasonic Stirring of the Weld Pool. Acta Mater. 2016, 106, 144–154. [CrossRef]
22. Tian, Y.; Shen, J.; Hu, S.; Wang, Z.; Gou, J. Effects of Ultrasonic Vibration in the CMT Process on Welded Joints of Al Alloy. J.
Mater. Process. Technol. 2018, 259, 282–291. [CrossRef]
23. Wang, H.; Hu, Y.; Ning, F.; Cong, W. Ultrasonic Vibration-Assisted Laser Engineered Net Shaping of Inconel 718 Parts: Effects of
Ultrasonic Frequency on Microstructural and Mechanical Properties. J. Mater. Process. Technol. 2020, 276, 116395. [CrossRef]
24. Chen, C.; Fan, C.; Liu, Z.; Cai, X.; Lin, S.; Zhuo, Y. Microstructure Evolutions and Properties of Al–Cu Alloy Joint in the Pulsed
Power Ultrasonic-Assisted GMAW. Acta Metall. Sin. (Engl. Lett.) 2020, 33, 1397–1406. [CrossRef]
25. Gorunov, A.I. Additive Manufacturing of Ti6Al4V Parts Using Ultrasonic Assisted Direct Energy Deposition. J. Manuf. Process.
2020, 59, 545–556. [CrossRef]
26. Zhou, C.; Jiang, F.; Xu, D.; Guo, C.; Zhao, C.; Wang, Z.; Wang, J. A Calculation Model to Predict the Impact Stress Field and Depth
of Plastic Deformation Zone of Additive Manufactured Parts in the Process of Ultrasonic Impact Treatment. J. Mater. Process.
Technol. 2020, 280, 116599. [CrossRef]
27. Zhou, C.; Wang, J.; Guo, C.; Zhao, C.; Jiang, G.; Dong, T.; Jiang, F. Numerical Study of the Ultrasonic Impact on Additive
Manufactured Parts. Int. J. Mech. Sci. 2021, 197, 106334. [CrossRef]
28. Chen, W.; Chen, Y.; Zhang, T.; Wen, T.; Yin, Z.; Feng, X. Effect of Ultrasonic Vibration and Interpass Temperature on Microstructure
and Mechanical Properties of Cu-8Al-2Ni-2Fe-2Mn Alloy Fabricated by Wire Arc Additive Manufacturing. Metals 2020, 10, 215.
[CrossRef]
29. Huu Phan, N.; Muthuramalingam, T. Multi Criteria Decision Making of Vibration Assisted EDM Process Parameters on Machining
Silicon Steel Using Taguchi-DEAR Methodology. Silicon 2021, 13, 1879–1885. [CrossRef]
30. Huu Phan, N.; Muthuramalingam, T. Multi-Criteria Decision-Making of Vibration-Aided Machining for High Silicon-Carbon
Tool Steel with Taguchi–Topsis Approach. Silicon 2021, 13, 2771–2783. [CrossRef]
Materials 2022, 15, 5168 17 of 17

31. Ma, Q.; Chen, H.; Ren, N.; Zhang, Y.; Hu, L.; Meng, W.; Yin, X. Effects of Ultrasonic Vibration on Microstructure, Mechanical
Properties, and Fracture Mode of Inconel 625 Parts Fabricated by Cold Metal Transfer Arc Additive Manufacturing. J. Mater. Eng.
Perform. 2021, 30, 6808–6820. [CrossRef]
32. Yuan, D.; Shao, S.; Guo, C.; Jiang, F.; Wang, J. Grain Refining of Ti-6Al-4V Alloy Fabricated by Laser and Wire Additive
Manufacturing Assisted with Ultrasonic Vibration. Ultrason. Sonochem. 2021, 73, 105472. [CrossRef] [PubMed]
33. Yuan, D.; Sun, X.; Sun, L.; Zhang, Z.; Guo, C.; Wang, J.; Jiang, F. Improvement of the Grain Structure and Mechanical Properties of
Austenitic Stainless Steel Fabricated by Laser and Wire Additive Manufacturing Assisted with Ultrasonic Vibration. Mater. Sci.
Eng. A 2021, 813, 141177. [CrossRef]
34. Todaro, C.J.; Easton, M.A.; Qiu, D.; Brandt, M.; StJohn, D.H.; Qian, M. Grain Refinement of Stainless Steel in Ultrasound-Assisted
Additive Manufacturing. Addit. Manuf. 2021, 37, 101632. [CrossRef]
35. Sun, L.; Guo, C.; Huang, L.; Jiang, F.; Xu, K.; Huang, R. Effect and Mechanism of Inter-Layer Ultrasonic Impact Strengthening on
the Anisotropy of Low Carbon Steel Components Fabricated by Wire and Arc Additive Manufacturing. Mater. Sci. Eng. A 2022,
848, 143382. [CrossRef]
36. Chen, Y.; Xu, M.; Zhang, T.; Xie, J.; Wei, K.; Wang, S.; Yin, L.; He, P. Grain Refinement and Mechanical Properties Improvement of
Inconel 625 Alloy Fabricated by Ultrasonic-Assisted Wire and Arc Additive Manufacturing. J. Alloys Compd. 2022, 910, 164957.
[CrossRef]
37. Ning, F.; Cong, W. Ultrasonic Vibration-Assisted (UV-A) Manufacturing Processes: State of the Art and Future Perspectives. J.
Manuf. Process. 2020, 51, 174–190. [CrossRef]
38. Plesset, M.S.; Prosperetti, A. Bubble Dynamics and Cavitation. Annu. Rev. Fluid Mech. 1977, 9, 145–185. [CrossRef]
39. Leighton, T.G. The Rayleigh-Plesset Equation in Terms of Volume with Explicit Shear Losses. Ultrasonics 2008, 48, 85–90.
[CrossRef]
40. Du, J.; Chen, F. Cavitation Dynamics and Flow Aggressiveness in Ultrasonic Cavitation Erosion. Int. J. Mech. Sci. 2021, 204, 106545.
[CrossRef]
41. Tian, H.; Yang, C.; Liao, Z. Numerical Simulation of Cavitation Bubble Dynamics Based on Different Frame Rayleigh-Plesset
Equation. In Proceedings of the 2008 Asia Simulation Conference—7th International Conference on System Simulation and
Scientific Computing, Beijing, China, 10–12 October 2008; pp. 1312–1316. [CrossRef]
42. Huang, Y.; Wu, Y.; Yang, F. Numerical Simulation on the Dynamics for the Ultrasonic Cavitation Bubble in Chitosan Solution.
Appl. Mech. Mater. 2013, 275–277, 628–634. [CrossRef]
43. Wei, R.; Lv, X.; Yang, M. Numerical Simulation of Ultrasound-Induced Cavitation Bubbling in a Calcium Ferrite Melt. In TMS
Annual Meeting & Exhibition; The Minerals, Metals & Materials Series; Springer: Cham, Switzerland, 2018; Part F10; pp. 115–121.
[CrossRef]
44. Ji, F.; Qin, X.; Hu, Z.; Xiong, X.; Ni, M.; Wu, M. Influence of Ultrasonic Vibration on Molten Pool Behavior and Deposition Layer
Forming Morphology for Wire and Arc Additive Manufacturing. Int. Commun. Heat Mass Transf. 2022, 130, 105789. [CrossRef]