International Journal of Machine Tools and Manufacture 124 (2018) 12–21

Contents lists available at ScienceDirect

International Journal of Machine Tools and Manufacture

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

Effects of pin thread on the in-process material flow behavior during friction
stir welding: A computational fluid dynamics study
Gaoqiang Chen a, Han Li a, Guoqing Wang b, Zhiqiang Guo c, Shuai Zhang a, Qilei Dai a,
Xibo Wang a, Gong Zhang a, Qingyu Shi a, *
a
State Key Laboratory of Tribology, Department of Mechanical Engineering, Tsinghua University, Beijing 100084, PR China
b
China Academy of Launch Vehicle Technology, Beijing 100076, PR China
c
SAIC Volkswagen Automotive CO. LTD, Shanghai 201805, PR China

A R T I C L E I N F O A B S T R A C T

Keywords: Pin thread is one of the most common geometrical features for the friction stir welding (FSW) tools. The main
Friction stir welding purpose of employing the pin thread is to improve the in-process material flow behaviors during FSW. However, it
Material flow has not been fully understood how exactly the pin thread influences the material flow because of the lack of in-
Numerical simulation process observation. In this study, we aim to analyze the effect of pin thread on the in-process material flow
Tool design during FSW of an Al-Mg-Zn alloy by using numerical simulation based on computational fluid dynamics (CFD). In
Pin thread
our numerical simulation, the transient rotation of the threaded pin is implemented explicitly via fully transient
control of the zone motion, and the mechanical interaction at the tool-workpiece interface is considered via the
recent developed shear-stress-based frictional boundary condition. The numerical simulation has been validated
by the experimental measured temperatures at 8 different locations, the distribution of marker materials and the
geometry of deformation zone in the weld. Based on the numerical simulation results, three effects of the pin
thread on the material flow have been elucidated. First, accelerated flow velocity and enhanced strain rate is
induced owing to the use of the pin thread, which is attributed to the fact that the interfacial sticking is preferable
inside the thread groove opening. Second, the pin thread has an effect to trap material in the high-velocity zone
inside the thread groove opening, which causes a many-circle flow pattern around the threaded pin. Third, the pin
thread contributes to a vertical pressure gradient, which is important for the in-process material transfer from the
top to the bottom. The approaches and concepts in this study can be applied for further fundamental investigation
of FSW and the computer aided design of the welding tools.

1. Introduction
Friction stir welding (FSW) [1] has been successfully applied in
fabricating many critical load-bearing structures, e.g. rocket fuel tanks
and high speed train bodies. The in-process material flow around the
welding tool is of vital importance in the scientific understanding of FSW
due to its critical role in governing many underlying processes [2], such
dynamic recrystallization [3] and heat generation [4,5], which are
closely related to the microstructure [6,7] and performance of the welds
[8,9]. Despite its critical importance, the material flow around the
welding tool is still one of the yet-least understood phenomenon in FSW.
For instance, the effect of the pin thread on the in-process material flow
has not been fully understood. The pin thread is one of the most common
and basic geometrical features for designing the pin geometry of the FSW
tools [10]. Employing thread or thread-like structures on the pin has been
considered to be unique to effectively improve the material flow around
the welding tool in order to avoid defects and form sound welds. Owing
to the important effect of the pin thread, further knowledge is required on
how exactly the pin thread influences the material flow behaviors during
FSW to support the welding tool design and process optimization. In this
study, we report our recent study to quantitatively elucidate the effect of
the pin thread on the in-process flow state close to the welding tool
during FSW by using three dimensional numerical simulation.
The importance to understand the effect of pin thread on the in-
process material flow in FSW has been well recognized in the previous
experiment-based researches. Several early studies have tried to reveal
the material flow field around a threaded pin during FSW by analyzing
the post-welding microstructural features in the welds. It was reported by

* Corresponding author.
E-mail address: shqy@tsinghua.edu.cn (Q. Shi).

https://doi.org/10.1016/j.ijmachtools.2017.09.002
Received 24 July 2017; Received in revised form 8 September 2017; Accepted 14 September 2017
Available online 15 September 2017
0890-6955/© 2017 Elsevier Ltd. All rights reserved.
G. Chen et al.
Table 1
Nominal chemical composition of the workpiece (AA7A52).
Elements Fe Si Mn Cu Mg
wt % 0.26 0.11 0.25 0.13 2.46
Guerra et al. [11] that the pin thread could contribute to a vertical flow
trend. A ‘stir effect’ of the pin thread is reported by Ouyang et al. [12],
which helps to deliver the material to form a sound weld. Besides,
Schmidt et al. [13] thought that the thread on the pin could introduce
some cyclic flow patterns in the friction stir welds. Actually, it is difficult
to understand the influence of the pin thread on the material flow,
because quite limited information on the in-process material flow could
be gained based on the post-welding analysis of microstructures.
Although some recent experiments have employed the X-ray radiography
[14] to in situ observe the material flow during FSW, but still very limited
information could be obtained due to the limited resolution in the current
in-situ experiment.
Due to the difficulties in experimental research, computational
methods have been developed to investigate the in-process material flow
during FSW [15]. In the literature, the computational methods for
studying the material flow during FSW could be divided into two cate-
gories, which are the Computational Solid Mechanics (CSM) approach
and the Computational fluid dynamics (CFD) approach. Schmidt et al.
[16], Buffa et al. [3,17] and Al-Badour et al. [18] have demonstrated that
different numerical strategies based on CSM were viable to predict the
in-process thermal-mechanical conditions, including the material flow
behavior, during FSW. However, few of the CSM-based simulations has
explicitly included the pin thread in their geometric model because of the
issues of computational cost and numerical robustness. These issues were
caused by the large element distortion induced by the rotation of a
threaded pin. In the CFD simulation work presented by Colegrave et al.
[19], it was demonstrated that the pin thread could be considered
explicitly in the three dimensional model. It was concluded that the
threaded pin contributed to a larger traversing force than the unthreaded
version. In later CFD analysis, Ji et al. [20] reported that the decrease of
width of thread groove could result in the increase of material flow ve-
locity in the vicinity of the tool. In CFD simulation by Yu et al. [21], a
transient model was proposed to include the transient rotation of the
threaded pin. They found that a swirling and upward flow field was
caused by the helical thread on the pin while the threads had only a small
effect on the temperature distribution. However, the flow velocity and
the strain rate were over-predicted in these CFD simulations due to the
use of a velocity-based fully sticking boundary condition [20,21].
Although the CFD simulation has been proved to be a very powerful tool
for analyzing the in-process material flow behaviors during FSW with a
threaded pin, more advanced boundary condition is still needed in the
CFD simulation for better understanding how exactly the pin thread in-
fluences the in-process material flow behaviors to support the engi-
neering design of FSW.
Our previous modeling work has shown that the CFD simulation with
Fig. 1. Illustration of the FSW experiment.
13
International Journal of Machine Tools and Manufacture 124 (2018) 12–21
Ni Cr Zn Ti Zr Al
0.05 0.17 4.71 0.06 0.08 Bal
a recent developed shear-stress-based boundary condition is more pre-
dictive for the deformation zone shape [22] and defects location [23] in
friction stir welds than those with a fully sticking boundary conditions.
Therefore, the use of the shear-stress-based boundary condition must
make CFD simulation a more useful tool for analyzing the material flow
field around a threaded pin. In this study, based on the shear-stress-based
boundary condition [22], three dimensional numerical simulation based
on the CFD approach is established to analyze the temperature field and
the material flow field during friction stir welding of an Al–Mg–Zn
alloy. Based on the numerical simulation, the predicted material flow
fields around a threaded pin and a smooth pin are predicted and
compared. Based on the comparison between the predictions on the
smooth and threaded pin, the effects of the pin thread on the in-process
material flow behavior are elucidated. Finally, the numerical predictions
on temperature and material flow are validated by the experimental
measurements.
2. FSW condition and experimentation
In this study, bead-on-plate FSW of AA7A52 plates was carried out.
The nominal chemical composition of the AA7A52 is given in Table 1.
The size of the workpiece was 150 mm (X)
90 mm (Y)
6 mm (Z). The
rotation rate of the welding tool was 800 rpm and the welding speed was
100 mm/min. A welding tool with threaded pin was used in the exper-
iment. The shoulder of the tool was 20 mm in diameter and had a concave
shape. The pin has a conical geometry, which was 8 mm near the
shoulder and 6 mm in diameter at the tip. A right-hand thread was
employed on the pin. In the welding process, a plunge depth of 0.5 mm
was employed, and the tool was tilt at 2 . Eight K-type thermal-couples
were pinned on the top surface of the workpiece to record the temper-
ature evolution during the welding process. The locations of the 8
thermal-couples are shown in Fig. 1, which were 15 mm, 20 mm, 25 mm
and 30 mm from the welding center line on the advancing side (AS) and
retreating side (RS). After welding, the cross section of the weld was
examined by optical microscope (OM). In two separate experiments,
0.1 mm-thick copper foil was used to visualize the material flow during
welding. Before welding, the copper foil was embedded at the location
2.7 mm from the welding centerline on the AS and RS, which is shown by
the colored line in Fig. 1. Owing to its good plasticity of pure copper, the
copper foil deforms and flows with the aluminum alloys during FSW. In
the experiments with marker material, the welding tool was pulled out of
the workpiece without dwelling at the end of tool travel period, so that
the state of marker material was frozen as the same as it was during
welding. After welding, the welds were examined by computer tomog-
raphy (CT) in transmission mode to examine the distribution of copper
G. Chen et al.
Fig. 2. Geometric model in the numerical simulation. (a) Top view of the model; CAE model of the welding tool with threaded pin (b) and smooth pin (c); (d) Tool-workpiece interface for
the welding cases with the threaded pin (d) and the smooth pin (e).
segments in the welds in the XY plane.
3. Numerical modeling approaches
The geometry of the numerical simulation model is illustrated in
Fig. 2 (a). The origin point is set to be located at the intersection point of
the tool axis and the top surface of the workpiece. The geometrical size of
the fluid domain is considered as the same as the size of the workpiece
used in the experiment. In the simulation model, the material is assumed
to flow into the fluid domain from the material inlet and flow out from
the material outlet. The thermal-mechanical interaction between the
welding tool and the workpiece is the most important part for prediction
of the temperature field and the material flow field during FSW. Proper
mathematic formulation has been employed to capture the heat input
and the friction at the tool-workpiece interface, which will be discussed
later in this section. In this study, both welding cases with the threaded
pin and the smooth pin are simulated. The tool shapes are shown in Fig. 2
(b) and (c) and the geometries of the tool-workpiece interfaces with
threaded and smooth pin are shown in in Fig. 2 (d) and (e) respectively.
The geometry of the tool-workpiece interface is determined by the
welding tool. It is worth noting that for the threaded pin, we use an extra
fluid volume to explicitly consider the thread groove. In Fig. 2 (d), the
extra fluid volume is shown in green, while the surrounding volume is
shown in yellow. Similar to the previous studies [21,24,25] in the liter-
ature, the extra fluid volume is taken as a rotating region, which is
geometrically connected to the surrounding fluid domain by the mesh
interface. The extra fluid volume is rotating as the same speed as the
welding tool during the calculation, while the surrounding volume is
stationary. In order to avoid mesh distortion caused by the transient zone
motion, the extra fluid volume and the surrounding volume are meshed
separately, as the mesh interface allows the use of non-conformal mesh
between the extra volume and the surrounding volume. The mesh
interface is implemented in the commercial CFD solver, ANSYS Fluent
[26]. In order to capture the transient rotation of the threaded tool in
each revolution, fine time step is necessary in the numerical calculation.
A time step of 5
104 s is used in this simulation. As such, 150 time
steps are required for each revolution of the welding tool which is
rotating at 800 rpm (0.075s for each revolution). In the whole calcula-
tion, a total period of 121.725s, which is 1623 revolutions of the welding
tool, is solved in order to get the temperature and material flow in the
14
International Journal of Machine Tools and Manufacture 124 (2018) 12–21
quasi-steady state. The numerical data of three revolutions ranging from
121.5s to 121.725s are recorded for analysis.
The general governing equations for the heat transfer and the mate-
rial flow for incompressible single-phase fluid are used in this work. The
conservation equations of mass and momentum are given as,
∂ρ
þ ∇⋅ðρ!
vÞ¼0 (1)
∂t
∂ρ!v
þ ∇⋅ðρ!
v!v Þ ¼ ∇p þ ∇⋅ðμð∇!
v þ ∇!
T
v ÞÞ (2)
∂t
where ρ is fluid density, μ is fluid viscosity, p is pressure, !
v is fluid ve-
locity and t is flow time. The energy conservation equation is given by,
∂ρH
þ ∇⋅ðρ!
v HÞ ¼ ∇⋅ðk∇TÞ þ SV (3)
∂t
where H is enthalpy, T is temperature in K, k is thermal conductivity and
SV is a spatial source term regarding the heat generation induced by
plastic deformation. The enthalpy H is defined in its integration form as,
T
H ¼ ∫ Tref CP dT (4)
where CP is the specific heat, T is temperature in K, and Tref is the
reference temperature, which is 300 K.
In the model, the incompressible single-phase non-Newtonian fluid is
adopted to represent the workpiece material in FSW, which allows the
fully-coupled calculation of simultaneous transfer of heat, mass and
momentum inside the workpiece. The density of the fluid is taken as
2:7
103 kg=m3 , which is typical for aluminum alloys [27]. The thermal
conductivity and the specific heat of the fluid are taken as a function of
temperature. The values of the thermal conductivity and the specific heat
of a similar Al–Mg–Zn alloys, AA7075 [27], are used in the present
simulation. The fluid viscosity is key to predict the momentum transfer
throughout the fluid zone. Owing to the solid-state-welding nature of
FSW, the material does not melt especially when the rotating rate is as
low as 800 rpm. As a result, the viscosity of the material is relatively high
and complex. In the current simulation, the fluid viscosity is defined to be
dependent on both temperature and strain rate. The formulation of the
viscosity is given based on the visco-plasticity theory [19],
G. Chen et al. International Journal of Machine Tools and Manufacture 124 (2018) 12–21

Fig. 5. Comparison of the predicted and experimental peak temperatures at

Fig. 3. Flow stress of AA7A52. The ultimate tensile strength (UTS) of AA7075 is obtained
different locations.
from the ASM handbooks [28].

σ τf ¼ μf ⋅σ n
μ¼
3_ε
where σ is the temperature-dependent flow stress and ε_ is the effective
strain rate. The material flows and deforms in a very limited volume
around the tool during FSW, which is bounded by the tool surface and the
surrounding non-deforming material which is at lower temperature.
Therefore, the dependence of flow stress on temperature is very important
for determining the geometry of the volume where high rate deformation
occurs. In this study, the flow stress as a function of temperature in the
simulation is calculated based on a nonlinear interpolation from the
experimental values. The flow stress of AA7A52 is tested by high-
temperature compression and tensile tests at room temperature. The
testing results are shown in Fig. 3 together with the typical tensile strength
values of AA7075 [28]. The flow stress curve used in the numerical model
is also plotted in Fig. 3. It is worth noting that the strengthening effect of
the strain rate on the flow stress is ignored, which can underestimate flow
stress in the high-rate deformation region. If this model is further used to
study the forces and torques carried by the welding tool, it is necessary to
(6)
(5)
where μf is the frictional coefficient (used as 0.25 [29]), σ n is the normal
stress at the interface. The normal stress is taken as a typical value of
50 MPa. As a result, the interfacial frictional shear stress is 12.5 MPa,
which is exerted at the tool-workpiece interface by our recent developed
boundary condition [22].
The frictional heat generation due to the interfacial slipping is taken
as a heat flux at the tool-workpiece interface, given by,
qf ¼ η⋅τf ⋅jj!
v rel jj (7)
where τf is the frictional shear stress and jj!v rel jj is the magnitude of the
interfacial relative/slipping velocity, η is the fraction of heat that flows
into the workpiece and is used as 0.7 [22]. The relative velocity ! v rel is
determined by the difference between the tool velocity and the in-process
material flow velocity, which is calculated by ! v rel ¼ ! v tool  !
v wp , where
! !
vtooland v are the interfacial tool velocity vector and the workpiece
wp

take the effect of strain rate on the flow stress into consideration. velocity vector at the tool-workpiece interface. In addition, the heat
In order to simulate the thermal-mechanical material behaviors dur- generation induced by the plastic deformation is also taken into account.
ing FSW, it is critical to have a proper model for the friction at the tool- This heat flux is taken as a volumetric heat flux, given as [4],
workpiece interface, because the interfacial friction is the driving force
qp ¼ κ⋅σ⋅_ε (8)
for the in-process material flow and also contribute much of the in-
process heat generation. Similar to the method in the CSM approaches,
where κ ¼ 0:6 [30] is the fraction of plastic work dissipated as heat, σ is
interfacial frictional shear stress is defined to represent the friction at the
the temperature-dependent flow stress and ε_ is the effective strain rate.
tool-workpiece interface in the current numerical simulation. The
The heat generation by plastic deformation is taken as the source term SV
magnitude of the frictional stress at the tool/workpiece interface is given
in the in equation (3).
based on the Coulomb's law of friction,

Fig. 4. The predicted temperature field. (a) Threaded pin; (b) Smooth pin.

15
G. Chen et al. International Journal of Machine Tools and Manufacture 124 (2018) 12–21

Fig. 6. Material flow field around the welding tool. (a)threaded tool; (b) smooth tool.

Table 2
Coordinates of the locations for probing the flow variables around the welding tool.

Point number X Position, mm Y Position, mm Z Position, mm

1 0.000 4.052 0.300

2 0.000 3.946 0.900
3 0.000 3.841 1.500
4 0.000 3.735 2.100
5 0.000 3.629 2.700
6 0.000 3.523 3.300
7 0.000 3.417 3.900
8 0.000 3.312 4.500
9 0.000 3.206 5.100
10 0.000 3.100 5.700

Fig. 8. Comparison of the flow variables between the threaded and the smooth pin (Point-
8 in Table 2). (a) Velocity; (b) strain rate.

4.2. Material flow field

Fig. 6 (a) and (b) show the predicted velocity field in the vicinity of
the welding tool for the threaded pin and the smooth pin. For the both
cases, it is found that the material flows at relatively high velocity in a
thin layer that is very close to the welding tool. In the both welding cases,
Fig. 7. Comparison of the average velocity near the threaded tool and smooth tool.
the high-velocity zone near the shoulder surface is well spread towards
the shoulder periphery, and the maximum width of the high-velocity
zone is similar to the diameter of the shoulder. This is because the tem-
4. Results & discussion
perature is relatively high near the shoulder, thus the flow stress is
lowered; as a result, the material is easy to flow when it is driven by the
4.1. Temperature field
interfacial friction. The range of the high-velocity zone shrink signifi-
cantly as the distance from the shoulder increases. Meanwhile, the ma-
Based on the numerical calculation, both the temperature field and
terial flow velocity is significantly accelerated by the pin thread at the
material flow field during the steady-state FSW of the 6 mm-thick
lower part of the pin. It could be seen from Fig. 6 (a) that the material
AA7A52 plate are predicted. Fig. 4 shows the predicted temperature
flows at high velocity inside the thread groove opening. The typical
distribution on the workpiece on the advancing side and the Y ¼ 0 plane.
material velocity inside the thread groove ranges from 220 mm/
It could be found that different welding tools with the threaded pin and
s~260 mm/s, which is very close to the tool velocity at the thread sur-
the smooth pin result in quite similar temperature fields. In other words,
face. It is implied that a nearly sticking state between the welding tool
the predicted temperature field is hardly influenced whether or not the
and the workpiece is present inside the thread groove openings. In
pin thread is adopted. The simulation results show that the predicted
comparison, for the welding case with the smooth pin, significant
peak temperature is 479.9  C for the threaded pin and 479.2  C for the
interfacial slipping occurs at the bottom part of the pin side, as the ma-
smooth pin. Besides, it can be seen from Fig. 4 that both the shape of the
terial velocity ranges from 0 to 50 mm/s, which is significantly lower
temperature contour lines and the space between them are similar over
than the tool velocity. It could be concluded that the material inside the
the workpiece area. Therefore, it could be known from the simulation
thread groove is preferable to present a nearly sticking state. Because
data that the pin thread has little influence on the temperature field
there is little difference in the temperature distribution between the
during FSW. In order to validate the predicted temperature field, the
threaded case and the smooth case, the presence of the sticking state
predicted and measured peak temperatures at the eight different dis-
inside the pin groove is considered to be induced by the geometric fac-
tances from the welding center line are compared in Fig. 5. Good
tors. In order to clarify the effect of pin thread on the material flow be-
agreement between the predicted and measured values could be found,
haviors, the flow velocity at the points in a straight line are probed during
which supports the validity of the numerical simulation.
the tool rotation. The coordinates of these points are listed in Table 2.

16
G. Chen et al.
Fig. 9. Contour of material flow velocity and strain rate at Z ¼ 4.5 mm. The time value used in this figure are relative time in three revolutions of the welding tool.
Fig. 7 displays the average material velocity at the points. It is clear that
the pin thread have a significant accelerating effect on the material flow
when the material flow is difficult at the bottom portion of the pin side.
We consider this accelerating effect is attributed to the presence of the
sticking state inside the thread groove opening.
Fig. 8 (a) and (b) show the variation of the material velocity and the
strain rate during the quasi-steady state welding period in three
sequential tool revolutions at point-8 in Table 2. The location of point-8 is
quite close to the welding tool surface. The time used as horizontal axis in
Fig. 8 are relative time in the three revolutions of the tool during welding.
It could be seen that both the material velocity and the strain rate around
the threaded pin is varying periodically during the tool revolution, while
both the flow variables for the smooth tool varies little. As shown in
Fig. 8(a), the minimum material velocity during the tool rotation near the
threaded pin is similar to the velocity near the smooth pin. After this low
velocity period, the material velocity near the threaded pin increases
gradually and then reaches a plateau. After the plateau, the velocity
decreases to the minimal value very quick. This variation repeats as the
welding tool rotates. Similar variation of strain rate during the tool
rotation could be found in Fig. 8(b). It could be found that, although the
17
International Journal of Machine Tools and Manufacture 124 (2018) 12–21
use of the pin thread results in accelerated flow velocity and enhanced
strain rate, but the accelerating/enhancing effect is not steady but
varying periodically as the welding tool rotates.
Fig. 9 (a) ~ (j) displays the variation of the flow velocity and the
strain rate around the lower portion of both the threaded pin and the
smooth pin during tool rotation. It is evident that there is a localized
region inside the pin thread groove opening, where the material is
flowing at high velocity and deforming with high strain rate. The present
of the localized high velocity and high-rate plastic deformation is
induced by the presence of the sticking state inside the pin groove
openings. Besides, this localized region seems to move with the thread as
shown in Fig. 9 (a) ~ (d) and (f) ~ (i), which explains the periodical
variation of the flow velocity and the strain rate. For the smooth pin,
much lower velocity and strain rate around the smooth pin at the bottom
region could be explained by the slipping state at the lower portion of the
pin side as shown in Fig. 9 (e) and (j). It is worth noting that the pin
thread induces significant increase on the strain rate with little change on
the temperature, which would cause the change of the Zener-Hollomon
 
parameter defined as Z ¼ ε_ ⋅exp RT Q
, where ε_ is strain rate, T is
G. Chen et al.
Fig. 10. Material flow field and deformation field around the welding tool. Strain rate
over the cross section for threaded tool (a) and smooth tool (b); (c) macrostructure on the
weld cross-section.
Fig. 11. Three dimensional flow path through the flow field close to the welding tool.
temperature and Q is activation energy of the recrystallization. There-
fore, employing the pin threaded could be an important approach to tune
the Zener-Hollomon parameter during FSW so as to control the weld
microstructure, as the Zener-Hollomon parameter has been reported to
control the grain size [31,32] and the texture [33] in hot deforma-
tion process.
18
International Journal of Machine Tools and Manufacture 124 (2018) 12–21
Fig. 12. Comparison of the flow path for the threaded/smooth pin. (a) Side view; (b)
Top view.
4.3. Plastic deformation field
The plastic deformation field around the threaded pin and the smooth
pin are shown in Fig. 10 (a) and (b) respectively. It could be seen that the
high rate plastic deformation zone is quite localized in the area that is
close to the welding tool. A high-rate deformation band, in which the
deformation rate is well beyond 50/s, is present. This deformation band
spread from the shoulder surface through the pin side and extend to the
pin bottom. In Fig. 10 (c), the predicted shape of the deformation band,
which is taken as the contour line of 50/s from the strain rate profile, is
compare to the weld macrostructure from our experiment. It is found that
both the predicted shapes of the deformation band in the crown part by
the model with/without the pin thread are in good agreement with the
experimental measured deformation zone shape on the cross-section of
the weld. Unfortunately, the strain rate decreases significantly in the
bottom part of the deformation zone. The deformation band is still able to
cover the whole thickness of the workpiece continuously for the case
with threaded pin, considering the fact that the localized high-rate
deformation zone is moving with the rotating tool. Besides, the geome-
try of the deformation band is also in good agreement with the mea-
surement, as shown in Fig. 10 (c), which validates the numerical
predictions on the material flow behaviors in the FSW. It could be seen in
Fig. 10 (c) that, for the welding case using a smooth pin, the deformation
band becomes discontinuous and does not reach the bottom part of the
workpiece, as shown in Fig. 10 (b). It could be implied that the pin thread
G. Chen et al.
Fig. 13. Flow paths of material around the tool on the retreating side (a ~ c) and advancing side (d ~ e). (a) and (d) are the experimental results; (b) and (e) are the predicted results; the
experimental/predicted results are compared in (c) and (f).
is important in determining the deformation zone shape in the vicinity of
the welding tool.
4.4. Material flow paths
It is known from the above discussion that the flow field around the
thread tool is periodically varying during FSW, while the flow field
around the smooth pin is fairly steady. Fig. 11 shows the material flow
paths for the threaded pin (a) and the smooth pin (b) based on the tracing
algorithm over a period of 10s. The material in front of the welding tool
flows through the flow field following the flow paths and then deposits in
the welds. It is found from the flow path for the threaded pin in Fig. 11(a)
that the material moves for many circles before deposition in the weld. In
comparison, the material flows for one and a half circles around the
smooth pin before depositing in the welds as shown in Fig. 11 (b). Three
time tags are placed in Fig. 11 (a) and (b) respectively. These time tags
show the times when the tracing begins, when the flow path is affected by
the welding tool, and when it deposits in the weld, for the both welding
cases using a threaded pin and a smooth pin. It could be found that the
duration of the material flow around the threaded pin is 4.77s, which is
much longer than that for the smooth pin (2.30s). The longer time for
flow around the pin would help to build up more plastic deformation in
the weld produced by the threaded pin.
Fig. 12 shows the material flow paths in side view (a) and top view
(b). By further examining the radius of the flow path of the threaded
pin, it is found that that the material is trapped in the high-velocity
zone inside the thread groove openings when it flows around the
pin. After the occurrence of the trapping, the material flows fast around
the pin. After flowing for many circles, it stops and deposits in the
weld. As we know, it is common phenomenon that the pin thread is
filled with material after welding. Here, our simulation reveals that the
pin thread opening is a volume within which the material can be stored
for short time during welding due to the trapping effect. In other
words, the trapped material is going to deposit in the weld, while there
will be new material being trapped into the thread groove. In this way,
the material in the thread groove will update itself during the weld-
ing process.
Reliable numerical results require sound experiments for
19
International Journal of Machine Tools and Manufacture 124 (2018) 12–21
verification. But actually there is much technical difficulty to probe
experimentally the in-process material flow during FSW to verify the
numerical simulation, as the material flow occurs in a very small vol-
ume surrounded by the welding tool and solid material. Here we use a
‘marker material’ approach in both experiments and numerical analysis
to visualize the material flow paths around the welding tool, as the
material in front of the pin can enter the volume where high rate flow
occurs as discussed above. For verification purpose, the flow paths of
the material on both RS and AS are compared to the experimental
measured distribution of the marker materials, which is shown in
Fig. 13. In the experiments, the copper foil was used as marker material
and embedded along the welding direction in the workpiece before
welding, which was placed 2.7 mm offset from the welding center
before FSW. Fig. 13 (a) and (d) shows respectively the distribution of
copper segments after FSW around the exit hole on RS and AS. Fig. 13
(b) and (e) shows the predicted flow paths originated from the longi-
tudinal plane on RS and AS, where the marker materials were placed in
the workpiece before welding. In Fig. 13 (c) and (f), the numerical
predicted flow paths are superimposed on the experimental results. It
could be found that both the numerical prediction and the experimental
agree quite well with each other. Both the experimental and predicted
results in Fig. 13 (c) shows that most of the marker materials on RS
before welding deposit on RS with similar distances from the welding
center after passing the welding tool, while small amount of the marker
material flows with the rotating tool and deposits on the downside of its
original locations. As shown in Fig. 13 (f), the marker material on the
AS flows with the tool for more circles than that on the RS; as a result,
the copper foil is broken into much finer segments in the weld. Besides,
we have noted an interesting fact from both predicted and experimental
flow paths that the material from the AS deposited mostly on AS despite
the complex flow around the pin. Although the distribution of the
copper segment becomes discrete and scattering in the experimental
measured photographs, but the location of the segments is indeed
valuable in reflecting many geometry features of the in-process material
flow paths during FSW. Based on the above discussion, the numerical
prediction on the material flow is verified by the experimental measured
distribution of the marker materials.
G. Chen et al.
Fig. 14. Pressure fluctuation and material flow in YZ-plane inside the thread groove.
4.5. Material trapped in the thread groove
There is a clear upward flow trend when it passes the smooth pin, as
shown by the black line in Fig. 12 (a). As a result, the deposition position
of the material is well above where it was before welding. The upward
flow trend when it flow around the pin could result in accumulation of
material at the top of the deformation zone, which may result in loss of
material via formation of excessive flash. In comparison, when the
threaded pin is adopted, the upward material flow is inhabited, and the
deposition position of the material is below the original position.
Therefore, the presence of pin thread is important for the in-process
material transfer from the top to the bottom especially when the FSW
is applied to weld thick plate.
Fig. 14 depicts the local distribution of the predicted pressure and the
in-plane flow state inside the pin thread groove. It is shown that there is a
significant vertical pressure gradient inside the pin groove opening,
which induces a pressure difference of ~0.9 MPa between the upper
surface and lower surface of the pin thread. This pressure gradient in-
duces a downward flow of the material that is trapped inside the pin
thread. Therefore, the fact that the upward material flow trend is
inhibited by the presence of pin thread as shown in Fig. 12 (a) could be
understood.
5. Conclusions
In this work, the temperature filed and material flow field during FSW
of an Al–Mg–Zn alloy with a threaded pin and a smooth pin is analyzed
and compared by using the CFD-based numerical simulation. A clearer
picture of the influence of the pin thread on the material flow in FSW is
shown by the quantitative analysis based on the simulation data.
Particularly, three effects of the pin thread has been elucidated. First,
accelerated flow velocity and enhanced strain rate is induced by the
threaded pin, which is attributed to the fact that the interfacial sticking is
preferable inside the thread groove. Second, the pin thread has an effect
to trap material in the high-velocity zone inside the thread groove
opening, which causes a many-circle flow pattern around the threaded
pin during the welding process. Third, the pin thread contributes to a
vertical pressure gradient, which is important for the in-process material
transfer from the top to the bottom.
It is demonstrated that the combination of fully transient control of
the zone motion and the shear-stress-based frictional boundary condition
could be successfully applied in numerical simulation of the thermal-
mechanical process during FSW with threaded pins. The approaches
20
International Journal of Machine Tools and Manufacture 124 (2018) 12–21
and concepts in this study can be applied for further investigation of FSW
and the computer aided design of the welding tools.
Acknowledgement
The research was supported by the National Natural Science Foun-
dation of China (grant no. 51375259 and grant no. 51705280) and the
National Science and Technology Major Project of the Ministry of Science
and Technology of China (no. 2012ZX04012-011).
Appendix A. Supplementary data
Supplementary data related to this article can be found at https://doi.
org/10.1016/j.ijmachtools.2017.09.002.
References
[1] R.S. Mishra, Z.Y. Ma, Friction stir welding and processing, Mater. Sci. Eng. R Rep.
50 (1–2) (2005) 1–78.
[2] L. Shi, C.S. Wu, H.J. Liu, The effect of the welding parameters and tool size on the
thermal process and tool torque in reverse dual-rotation friction stir welding, Int. J.
Mach. Tools Manuf. 91 (2015) 1–11.
[3] L. Fratini, G. Buffa, CDRX modelling in friction stir welding of aluminium alloys, Int.
J. Mach. Tools Manuf. 45 (10) (2005) 1188–1194.
[4] G.-q. Chen, Q.-y. Shi, Y.-j. Li, Y.-j. Sun, Q.-l. Dai, J.-y. Jia, Y.-c. Zhu, J.-j. Wu,
Computational fluid dynamics studies on heat generation during friction stir
welding of aluminum alloy, Comput. Mater. Sci. 79 (2013) 540–546.
[5] H. Su, C.S. Wu, A. Pittner, M. Rethmeier, Thermal energy generation and
distribution in friction stir welding of aluminum alloys, Energy 77 (0) (2014)
720–731.
[6] R.W. Fonda, J.F. Bingert, Texture variations in an aluminum friction stir weld, Scr.
Mater. 57 (11) (2007) 1052–1055.
[7] J. Chen, H. Fujii, Y. Sun, Y. Morisada, K. Kondoh, Optimization of mechanical
properties of fine-grained non-combustive magnesium alloy joint by asymmetrical
double-sided friction stir welding, J. Mater. Process. Technol. 242 (2017) 117–125.
[8] H. Badarinarayan, Q. Yang, S. Zhua, Effect of tool geometry on static strength of
friction stir spot-welded aluminum alloy, Int. J. Mach. Tools Manuf. 49 (2) (2009)
142–148.
[9] H.B. Chen, J.F. Wang, G.D. Zhen, S.B. Chen, T. Lin, Effects of initial oxide on
microstructural and mechanical properties of friction stir welded AA2219 alloy,
Mater. Des. 86 (2015) 49–54.
[10] R. Rai, A. De, H. Bhadeshia, T. DebRoy, Review: friction stir welding tools, Sci.
Technol. Weld. Join. 16 (4) (2011) 325–342.
[11] M. Guerra, C. Schmidt, J. McClure, L. Murr, A. Nunes, Flow patterns during friction
stir welding, Mater. Charact. 49 (2) (2002) 95–101.
[12] J. Ouyang, R. Kovacevic, Material flow and microstructure in the friction stir butt
welds of the same and dissimilar aluminum alloys, J. Mater. Eng. Perform. 11 (1)
(2002) 51–63.
[13] H.N.B. Schmidt, T. Dickerson, J.H. Hattel, Material flow in butt friction stir welds in
AA2024-T3, Acta Mater. 54 (4) (2006) 1199–1209.
G. Chen et al.
[14] Y. Morisada, H. Fujii, Y. Kawahito, K. Nakata, M. Tanaka, Three-dimensional
visualization of material flow during friction stir welding by two pairs of X-ray
transmission systems, Scr. Mater. 65 (2011) 1085–1088.
[15] X. He, F. Gu, A. Ball, A review of numerical analysis of friction stir welding, Prog.
Mater. Sci. 65 (2014) 1–66.
[16] H. Schmidt, J. Hattel, A local model for the thermomechanical conditions in friction
stir welding, Model. Simul. Mater. Sci. Eng. 13 (1) (2005) 77.
[17] G. Buffa, J. Hua, R. Shivpuri, L. Fratini, A continuum based fem model for friction
stir welding—model development, Mater. Sci. Eng. A 419 (1) (2006) 389–396.
[18] F. Al-Badour, N. Merah, A. Shuaib, A. Bazoune, Coupled Eulerian Lagrangian finite
element modeling of friction stir welding processes, J. Mater. Process. Technol. 213
(8) (2013) 1433–1439.
[19] P.A. Colegrove, H.R. Shercliff, 3-Dimensional CFD modelling of flow round a
threaded friction stir welding tool profile, J. Mater. Process. Technol. 169 (2)
(2005) 320–327.
[20] S. Ji, Q. Shi, L. Zhang, A. Zou, S. Gao, L. Zan, Numerical simulation of material flow
behavior of friction stir welding influenced by rotational tool geometry, Comput.
Mater. Sci. 63 (2012) 218–226.
[21] Z. Yu, W. Zhang, H. Choo, Z. Feng, Transient heat and material flow modeling of
friction stir processing of magnesium alloy using threaded tool, Metall. Mater.
Trans. A 43 (2) (2012) 724–737.
[22] G. Chen, Z. Feng, Y. Zhu, Q. Shi, An alternative frictional boundary condition for
computational fluid dynamics simulation of friction stir welding, J. Mater. Eng.
Perform. 25 (9) (2016) 4016–4023.
[23] Y. Zhu, G. Chen, Q. Chen, G. Zhang, Q. Shi, Simulation of material plastic flow
driven by non-uniform friction force during friction stir welding and related defect
prediction, Mater. Des. 108 (2016) 400–410.
21
International Journal of Machine Tools and Manufacture 124 (2018) 12–21
[24] H. Su, C.S. Wu, M. Bachmann, M. Rethmeier, Numerical modeling for the effect of
pin profiles on thermal and material flow characteristics in friction stir welding,
Mater. Des. 77 (0) (2015) 114–125.
[25] G. Chen, Q. Shi, Y. Fujiya, T. Horie, Simulation of metal flow during friction stir
welding based on the model of interactive force between tool and material,
J. Mater. Eng. Perform. 23 (4) (2014) 1321–1328.
[26] ANSYS, I., ANSYS® Fluent..
[27] K.C. Mills, Recommended Values of Thermophysical Properties for Selected
Commercial Alloys, Woodhead Publishing, 2002.
[28] J.R. Davis, Aluminum and aluminum alloys, in: J.R. Davis (Ed.), ASM Specialty
Handbook ®, ASM international, 1993.
[29] M. Assidi, L. Fourment, S. Guerdoux, T. Nelson, Friction model for friction stir
welding process simulation: calibrations from welding experiments, Int. J. Mach.
Tools Manuf. 50 (2) (2010) 143–155.
[30] R. Nandan, T. DebRoy, H. Bhadeshia, Recent advances in friction-stir
welding–process, weldment structure and properties, Prog. Mater. Sci. 53 (6)
(2008) 980–1023.
[31] K.V. Jata, S.L. Semiatin, Continuous dynamic recrystallization during friction stir
welding of high strength aluminum alloys, Scr. Mater. 43 (8) (2000) 743–749.
[32] Morisada, Y., T. Imaizumi, H. Fujii, Determination of strain rate in Friction Stir
Welding by three-dimensional visualization of material flow using X-ray
radiography. Scr. Mater., (0).
[33] Z. Yu, H. Choo, Z. Feng, S.C. Vogel, Influence of thermo-mechanical parameters on
texture and tensile behavior of friction stir processed Mg alloy, Scr. Mater. 63 (11)
(2010) 1112–1115.