International Journal of Heat and Mass Transfer 125 (2018) 32–42

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Effects of tool tilt angle on the in-process heat transfer and mass transfer
during friction stir welding
Shuai Zhang, Qingyu Shi, Qu Liu, Ruishan Xie, Gong Zhang, Gaoqiang Chen ⇑
State Key Laboratory of Tribology, Department of Mechanical Engineering, Tsinghua University, Beijing, China

a r t i c l e i n f o a b s t r a c t

Article history: Tool tilt angle in friction stir welding (FSW) is an important factor that influences the joint quality, but the
Received 28 December 2017 underlying mechanism has not been fully understood due to the lack of in-process observation. In this
Received in revised form 10 March 2018 study, we present a CFD model, in which a geometrical model and an incomplete contact boundary con-
Accepted 15 April 2018
dition are proposed, to investigate the effects of tool tilt angle on the in-process heat transfer and mass
transfer during FSW. Three effects of tool tilt angle on heat transfer and mass transfer have been con-
cluded based on the simulation results. First, a higher temperature is generated by the tilt welding tool
Keywords:
on the advancing side (AS), which is attributed to the incomplete contact at the shoulder/workpiece
Friction stir welding
Computational analysis
interface. Second, the tilt welding tool generates a higher frictional force at the tool/workpiece interface,
Heat transfer which significantly improves the interfacial material flow velocity behind the tool. Third, the tilt welding
Mass transfer tool generates a stronger stirring action to the material in the vicinity of the welding tool, which is ben-
eficial for the mixing of materials and the formation of friction stir welds. The approaches and concepts in
this study can be used for the optimization of the application of tool tilt angle in FSW.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction
Friction stir welding [1,2] (FSW) is a kind of solid state welding
technology and widely applied in joining of similar/dissimilar
alloys [3,4]. It is a thermal-mechanical coupled process, in which
the microstructure and properties of the welds is closely related
to the in-process heat transfer and mass transfer. FSW process is
very sophisticated and affected by many factors including the
welding parameters [4,5], tool tilt angle [6,7], tool shape [8–10]
and contact states [11,12] at the welding tool/workpiece interface.
Among those factors, tool tilt angle can obviously affect the heat
transfer and mass transfer in the stir zone and accordingly controls
the integrity and the properties of the welds [13]. The performance
of the welds prepared at different tool tilt angle is quite different
[7,13]. However, the underlying mechanism remains unclear. Thus,
it is essential to investigate the effects of tool tilt angle on heat
transfer and mass transfer during FSW for the sake of welds quality
control and optimization of the process parameters.
Tool tilt angle is defined basing on the positional relationship
between the welding tool and the workpiece, where zero denotes
the case that the welding tool is perpendicular to the workpiece
while a non-zero value denotes the case that the welding tool is tilt
⇑ Corresponding author.
E-mail address: cheng3@aliyun.com (G. Chen).
to the workpiece [6,7,14–16]. In the past decade, many efforts have
been made to study the effects of tool tilt angle on the welds integ-
rity and properties. Early experiments conducted by Chen [13]
showed that the welds integrity was closely linked to the tool tilt
angle. At a smaller tool tilt angle, material flow behind the pin
was insufficient, which resulted in the void defects. At a larger tool
tilt angle, void defects might be also generated due to the weld
flash generated on the RS and no sufficient material filled up the
cavity in the stir zone. Barlas and Ozsarac [17] found that root flaw
could be effectively avoided by properly selecting the tool tilt
angle. Root flaw would be removed at the tool tilt angle of 2° when
the other welding parameters were constants. Zhang et al. [15]
reported that the tunnel defect could be eliminated by increasing
the tool tilt angle from 0° to 3°, which was attributed to the
increase in heat generation and downward forging force at the lar-
ger tool tilt angle. On the other hand, welds properties especially
dissimilar welds properties were also closely related to the tool tilt
angle. Mehta and Badheka [7] investigated the effects of tool tilt
angle on the mechanical properties of dissimilar copper-
aluminum friction stir welds. The macro hardness increased as
the tool tilt angle increased from 0° to 4° and the maximum tensile
strength was obtained at the tool tilt angle of 4°. Dehghani [18]
reported that the tensile strength of the dissimilar aluminum-
mild steel friction stir welds decreased with the increase of tool tilt
angle. This was because the thickness of Al5Fe2 intermetallic

https://doi.org/10.1016/j.ijheatmasstransfer.2018.04.067
0017-9310/Ó 2018 Elsevier Ltd. All rights reserved.
 S. Zhang et al. / International Journal of Heat and Mass Transfer 125 (2018) 32–42 33

compounds (IMCs) increased as the tool tilt angle increased, which

was detrimental to the tensile strength of the welds. Kimapong and
Watanabe [19] studied the effects of tool tilt angle on the shear
strength of the welds. They reported that the thickness of IMCs
increased as the tool tilt angle increased, which lead to the reduc-
tion of shear strength of the welds. Although lots of experiments
had been conducted to investigate the effects of tool tilt angle on
the FSW process, the mechanisms were still unclear.
Comparing with experimental approaches, numerical simula-
tion can obtain full-field and full-process information such as
material flow and temperature distribution [20,21], which makes
it become an effective method to carry out the analysis of the
effects of tool tilt angle on FSW process. Computational fluid
dynamic (CFD) is an important method in numerical simulation
of FSW process [22] and many studies on FSW have been made
by using this method. Colegrove et al. [23,24] developed a three-
dimensional CFD model to study the temperature distribution
and material flow behavior in FSW of aluminum alloys. Tool tilt
angle was taken into account in the geometric models, but the
effects of tool tilt angle on the FSW process were not elaborated
in these papers. Nandan et al. [25] established a CFD model to
obtain the 3D temperature distribution and plastic flow fields dur-
ing FSW and concluded that the plastic flow significantly affected
heat transfer within the workpiece. Su et al. [26] proposed a
method to gain the values of friction coefficient and contact vari-
able for the optimization of the process parameters and the distri-
bution features of thermal energy density in the vicinity of the tool Fig. 1. The experimental setup and the schematic diagram of temperature
measurements.
were elucidated. In recent researches, CFD simulation [22] has
been widely and successfully applied in investigating the issues
of FSW.
However, till now, very limited work in the literature considers from the welding center line in the top surface on AS and RS, as
the tool tilt angle in simulation of FSW, as including the tool tilt shown in Fig. 1(b). The thermocouples were mechanically pinned
angle is still of technical difficulty for the CFD simulation. One of in the holes that were 1 mm underneath the top surface of the
the difficulties is that considering the influence of tool tilt angle workpiece.
will make the model excessively complex regarding model estab-
lishment, solving equations and simulation results. Thus, tool tilt
angle is generally not considered in previous models for the sake 3. Description of the CFD model
of predigestion. In fact, tool tilt angle can influence the in-
process heat transfer and mass transfer and thus affect the A CFD model was established in this study to investigate the
microstructure and performance of the welds. Therefore, the in- effects of tool tilt angle on heat transfer and mass transfer during
process heat transfer and mass transfer at proximity of the welding FSW process. Eulerian framework was used in this model. As the
tool are not expressed precisely by using the simplified CFD model. tool tilt angle was taken into account in the CFD model, there
In this study, a three-dimensional CFD model is developed to was a big difference in the geometric model and boundary condi-
investigate the effects of tool tilt angle on heat transfer and mass tion. First, the tilt welding tool had a cutting effect on the work-
transfer during FSW. Especially, tool tilt angle is taken into account piece, which would produce a thinner joint. As a result, in the
in the geometrical model. The incomplete contact boundary condi- geometric model, the welds surface behind the welding tool was
tion at the tilt welding tool/workpiece interface is considered. The lower than the top surface of the workpiece. Second, the contact
effects of tool tilt angle on the in-process heat transfer and mass area between the shoulder and the workpiece would decrease
transfer in the vicinity of the welding tool are analyzed based on when the welding tool had a non-zero tilt angle with the top sur-
the simulation results. face of the workpiece, which resulted in an incomplete contact
boundary condition at the shoulder/workpiece interface.
In our simulation, ANSYS Fluent [27] was adopted for solving
2. Experiment the governing equations. The workpiece was taken as an incom-
pressible non-Newtonian viscous fluid. The heat capacity and ther-
Friction stir butt joining of AA2024-T4 was conducted in this mal conductivity were temperature dependent. The density of the
study. The dimensions of each workpiece were 300
75
5 mm workpiece was set as a constant, which was 2785 kg/m3 [28].
(length
width
thickness), and they were fixed by the fixtures,
as shown in Fig. 1(a). The shoulder was flat and the pin was 3.1. Geometric model
unthreaded. The shoulder diameter was 16 mm, the top and the
root diameter of the pin were 6 mm and 4 mm respectively. The The geometric model and mesh schemes at the welding tool/-
length of the pin was 4.8 mm. The rotational speed and the weld- workpiece interface for the tool tilt angle of 2.5° was shown in
ing speed were 800 rpm and 20 mm/min, respectively. The tool tilt Fig. 2. In the geometric model, the coordinate origin was set as
angles were set as 2.5° and 0°, while the other welding parameters the intersection point of the welding tool axis and the top surface
were constants. of the workpiece. When the welding process was carried out with a
The temperature history was recorded by eight K-type non-zero tool tilt angle, the welding tool would be tilt to the rear
(CH1-CH8) thermocouples located 10, 14, 20, and 28 mm away side. The shoulder surface would have 2.5° with the top surface
34 S. Zhang et al. / International Journal of Heat and Mass Transfer 125 (2018) 32–42

(a) Geometric model

(b)Mesh schemes at the tool/workpiece interface

Fig. 2. Geometric model and mesh schemes at the interface for the tool tilt angle of 2.5°.

of the workpiece, which resulted in a thinner joint (h1 < h2, h1 was 3.2. Incomplete contact boundary condition
the thickness of the weld joint and h2 was the thickness of the
workpiece.), as shown in Fig. 2(a). Thus, in the geometric model, As the welding process was carried out, a portion of the soft-
the surface (the blue1 area in Fig. 2(a)) behind the welding tool ened metal material would flow into the gap between the front half
should be lower than the top surface of the workpiece. However, of the shoulder and the workpiece. However, the amount of the
there was no penetration to the workpiece behind the tool due to material was insufficient to fill up the gap, which lead to the occur-
the depth that the tilt pin inserted into the workpiece was still rence of the incomplete contact state at the shoulder/workpiece
shorter than the thickness of the workpiece. Geometric model varied
with the tool tilt angle.
Large amount of heat was generated within the deformation
zone during FSW process. The gradients of heat flux and veloc-
ity were very large in the deformation zone and the smaller
mesh size (corresponding to the higher spatial resolution) at
proximity of the welding tool was necessary to obtain the
more accurate simulation results. Thus, in this study, a finer
mesh was applied in the adjacent area of the tool/workpiece
boundary as shown in Fig. 2(b), while a coarser mesh was
applied in the other area. The maximum and minimum volume
of hexahedral cells in the model were 2.90
1010 m3 and
1.32
1013 m3 respectively.

1
For interpretation of color in Fig. 2, the reader is referred to the web version of
this article. Fig. 3. Incomplete contact in the experiment of FSW.
 S. Zhang et al. / International Journal of Heat and Mass Transfer 125 (2018) 32–42 35

Fig. 4. The characterization of the incomplete contact boundary condition.

interface, as shown in Fig. 3. In order to characterize the incom- Thus, any point Q(x, y, z) in the contact area must be met as
plete contact boundary condition at the tilt welding tool/work- follows:
piece interface, two assumptions were proposed in this study.
ðx  X 0 Þ2 þ ðy  Y 0 Þ2 þ ðz  Z 0 Þ2 6 R2 ð5Þ
(1) In the dwell stage: The trajectory of the material in the gap Frictional tangential force was applied on the welding tool/-
between the front half of the shoulder and the workpiece workpiece interface to represent the interaction between the weld-
was round when the softened material passed through the ing tool and workpiece, which was given by Eq. (6) [11].
tilt welding tool, as shown by circle O1 in Fig. 4 (X-axis
was the welding direction).
(2) In the welding stage: As the welding tool moved forwards, 1100
the circle O1 would turn b counterclockwise under the dri- 194

Thermal conductivity (W/(m K))

ven effect of the tilt welding tool. 192
1050
Heat capacity (J/(K kg))

190
Where R0 was the radius of the shoulder, O was the center of the
shoulder, h was the difference between R and R0. In this study, a 188
1000
was in plane xz and b was in the plane of the shoulder surface. b 186
was a constant that was related to the welding parameters includ- 184
ing the rotational speed, the welding speed and the tool tilt angle. 950
182
In this study, h equaled to 6 mm and b equaled to 45°. Before iden-
180
tifying the contact area, the center of circle O1 (X 0 ; Y 0 ; Z 0 ) had to be 900
178
determined firstly. The solution process was shown as follows:
Heat capacity 176
2
R2 þ h 850 Thermal conductivity
R¼ 0 ð1Þ 174
2h
300 400 500 600 700 800
X 0 ¼ ðR  hÞ  cos b  cos a ð2Þ Temperature (K)

Y 0 ¼ ðR  hÞ  sin b ð3Þ Fig. 6. Material properties [30].

Z 0 ¼ ðR  hÞ  cos b  sin a ð4Þ

200
9000 strain rate-0.01s-1
175
strain rate-1s-1
8500
150 strain rate-100s-1
Flow stress(Mpa)

8000
125
Axial Force(N)

7500
100
7000
75
6500
50
6000
25
5500
0
5000 550 600 650 700 750 800
0 1 2 3 4 5
Tool tilt angle(°) Temperature (K)
Fig. 5. Average axial force at different tool tilt angle [7]. Fig. 7. Flow stress varied with the temperature.
36 S. Zhang et al. / International Journal of Heat and Mass Transfer 125 (2018) 32–42

(a) 2.5° tool tilt angle (b) 0° tool tilt angle

Fig. 8. Predicted temperature field at different tool tilt angle.

Table 1
The difference of peak temperature on the AS and RS.
Temperature difference of AS
and RS (2.5°)
Distance from the welding center line (mm) Measured (K)
10 35
14 15
20 13
28 11
s ¼ uf  rn
where uf [29] was the sliding friction coefficient, rn was the normal
pressure on the welding tool/workpiece interface. According to the
method used in the reference [26], the normal pressure on the
interface was the quotient of axial force and shoulder area. During
the welding process, the axial force would increase with the
increase of the tool tilt angle [6,7], which indicated that the normal
pressure on the interface would increase with the increase of the
tool tilt angle. Thus, in this study, we applied a higher normal pres-
sure on the interface for the 2.5° tool tilt angle CFD model. We ref-
erenced to the recent study [7] to determine the normal pressure
for the case of 2.5° tool tilt angle, as shown in Fig. 5. It was found
from Fig. 5 that the axial force would increase by 25–30% when
the tool tilt angle increased from 0° to 2.5°. Therefore, in this study,
the normal pressure in case of the 2.5° tool tilt angle was 26% higher
than that of the 0° tool tilt angle.
3.3. Material properties
Predicted (K)
29
In this study, the workpiece was assumed to be in-compressible
18 non-Newtonian fluid, the heat capacity and thermal conductivity
12 were temperature dependent [30], as shown in Fig. 6.
7 The effective flow stress of the workpiece was temperature and
strain rate dependent and given by Eq. (7) [31]:
"

1=n #
ð6Þ 1 e_Q
r ¼ rp sinh exp ð7Þ
A RT
where A, rp and n were material constants, and they were deter-
mined by the values of the scientific literature [32]. R was the ideal
gas constant, Q was the thermal deformation activation energy, e_
was the effective strain rate, T was the temperature in Kelvin. There
was a great error between the calculated results and the experimen-
tal data at high temperature. The material became liquid at the high
temperature close to the solidus of AA2024-T4 and a substantial
drop in flow stress was expected, which was not captured by Eq.
(7). Thus, the method [9,33] was introduced in this study to modify
the flow stress at high temperatures. The aim of the method was to
provide realistic representation of material behaviour in absence of
test data for temperatures close to solidus, without introducing

700 700

600 600
Peak temperature(K)
Peak temperature(K)

500 500

400 400

300 300

200 Advancing side -predicted 200 Retreating side -predicted

Advancing side -measured Retreating side -measured
100 100

0 0
10 15 20 25 30 10 15 20 25 30
Distance from welding center line (mm) Distance from welding center line (mm)
(a) Advancing side (b) Retreating side
Fig. 9. Comparison of the predicted and measured peak temperatures at different locations.
 S. Zhang et al. / International Journal of Heat and Mass Transfer 125 (2018) 32–42 37

further adjustable parameters into the constitutive model. The mod-
ified flow stress at high temperature was shown in Fig. 7.
3.4. Heat generation
Both the frictional heat and plastic deformation heat were con-
sidered in this CFD model. The frictional heat was calculated by Eq.
(8):
qf ¼ u  s  v rel
where s was the frictional tangential force; v rel was the relative
velocity; u was the ratio of the heat absorbed by the workpiece
[34,35].
Plastic deformation heat was given by Eq. (9). Where d ¼ 0:6
[36] was the conversion efficiency of the mechanical work to ther-
mal energy, r was the flow stress, e_ was the effective strain rate.
qp ¼ d  r  e_
4. Results and discussion
4.1. Effect of tool tilt angle on the temperature field
Fig. 8 shows the predicted temperature field for the tool tilt
angle of 2.5° and 0°. It can be seen clearly that different tool tilt
angle generates quite different temperature field. In other words,
the heat transfer is significantly influenced by the tool tilt angle.
In the case of the tool tilt angle of 0°, the temperature field is
almost symmetric about the welding tool axis and there is no dif-
ference in peak temperature on the AS and RS, which is consistent
with the conclusions of other research works [23–26,29]. However,
an asymmetric temperature field is generated by the tilt welding
tool, in which the peak temperature on the AS is apparently higher
than that on RS. This is because the incomplete contact state occurs
ð8Þ at the shoulder/workpiece interface and the contact area decreases
on the RS, thus the heat generation decreases on the RS. It can also
be seen from Table 1 that both the predicted and measured peak
temperature on the AS is higher than that on the RS at the same
distance from the welding center line. The difference of the peak
temperature on the AS and RS gradually decreases with the dis-
tance away from the welding center line. Those are common in
the FSW experiment and have been already reported in the litera-
ð9Þ ture [31], but hardly have been achieved in numerical simulation
of FSW.
Predicted and measured peak temperature at different distance
from the welding center line in the vicinity of the tilt welding tool
are compared in Fig. 9 to validate the accuracy of the CFD model. It
is found that all the predicted peak temperature correspond with
the measured peak temperature, which supports the validity of
the numerical simulation. In addition, the peak temperature at
10 mm away from the welding center line on AS and RS are

(a) 2.5°tool tilt angle (b) 0°tool tilt angle

Fig. 10. Predicted velocity field for different tool tilt angle.

2.0
0°
1.8 2.5°
1.6
1.4 Accelerated
Velocity(mm/s)

material velocity
1.2
1.0
0.8
0.6
0.4
0.2
Front RS Rear RS Rear AS Front AS
0.0
0 40 80 120 160 200 240 280 320 360
Angle (°)
(a) Data extraction path (b) Velocity profile for different tool tilt angle
Fig. 11. Accelerated material velocity.
38 S. Zhang et al. / International Journal of Heat and Mass Transfer 125 (2018) 32–42

588 K and 553 K respectively, which are very high but still lower
than the melting point, indicating that the material is in solid state
during FSW process.
4.2. Effect of tool tilt angle on material velocity
Fig. 10 shows the predicted material velocity field at the prox-
imity of the welding tool for the tool tilt angle of 2.5° and 0°. It
can be seen from Fig. 10 that material flows at a relatively high
velocity in the thin layer close to the inner parts of shoulder, while
lower material velocity can be seen in the thin layer near the
shoulder periphery and lower parts of pin side. This is because sig-
nificant sliding occurs at these parts. On the other hand, the mate-
rial velocity field is different for the different tool tilt angle. In the
case of the tool tilt angle of 0°, the material velocity field is almost
symmetric about the welding tool axis, and material flows at a rel-
atively low no more than 5 mm/s velocity in the thin layer close to
the bottom of pin side, which indicates the poor liquidity. Different

(a) 2.5°tool tilt angle (b) 0°tool tilt angle

Fig. 12. Frictional force at the tool/workpiece interface.

(a) 2.5° tool tilt angle, z=-1.8mm (b) 0° tool tilt angle, z=-1.8mm

(c) 2.5° tool tilt angle, z=-2.5mm (d) 0° tool tilt angle, z=-2.5mm
Fig. 13. Top views of the predicted material flow paths in different horizontal planes located at 1.8 mm and 2.5 mm away from the top surface of the workpiece.
 S. Zhang et al. / International Journal of Heat and Mass Transfer 125 (2018) 32–42 39

from that, an asymmetric material velocity field is generated by
the tilt welding tool. In addition, material velocity in the thin
layer close to the lower parts of the pin side ranges from 0 to
50 mm/s, which is apparent higher than that of the tool tilt
angle of 0°.
Interfacial material velocity data along the circumference at the
lower parts of the pin side (shown in Fig. 11(a)) is extracted to
quantitatively investigate the effect of tool tilt angle on the mate-
rial flow, as shown in Fig. 11(b). It can be seen clearly that the tilt
welding tool has a significant accelerating effect on the material
flow when the material flow is difficult at the rear side of the pin
side. This is because the tilt welding tool generates a higher fric-
tional force at the tool/workpiece interface behind the tool, as
shown in Fig. 12. It is known that material ahead of the welding
tool flows through the welding tool and then deposits in the welds.
Material flow driven by the frictional force at the tool/workpiece

(a) 2.5° tool tilt angle, z=-1.8mm (b) 0° tool tilt angle, z=-1.8mm

(c) 2.5° tool tilt angle, z=-2.5mm (d) 0° tool tilt angle, z=-2.5mm
Fig. 14. Side views of the predicted material flow paths in different horizontal planes located at 1.8 mm and 2.5 mm away from the top surface of the workpiece.

(a) 2.5°tool tilt angle (b) 0°tool tilt angle

Fig. 15. Predicted strain rate field.
40 S. Zhang et al. / International Journal of Heat and Mass Transfer 125 (2018) 32–42

interface. A higher frictional force can effectively improve the
interfacial material flow velocity. High interfacial material velocity
is good for the formation of the friction stir welds, while low inter-
facial material velocity may result in defects [20] such as voids and
tunnels. Thus, the accelerating effect on the material flow gener-
ated by the tilt welding tool is good for formation of the welds
and achieving a sound weld.
4.3. Effect of tool tilt angle on material flow path
Figs. 13 and 14 shows the top views and side views of the pre-
dicted material flow paths for the tool tilt angle of 2.5° and 0°.
Material ahead of the welding tool moves for a certain distance
before depositing in the welds, in which the temperature, velocity
and strain rate vary along the flow path. It can be seen from Figs. 13
and 14 that material flow paths are quite different for the different
tool tilt angle. Fig. 13 shows the top views of the predicted material
14
0°
2.5°
12
10
Strain rate(s-1)
flow paths in different horizontal planes located at 1.8 mm and 2.5
mm away from the top surface of the workpiece. It is found that
material in front of the tilt welding tool moves in a more compli-
cated mode when it flows through the welding tool and then
deposits in the weld. The noticeable material flow in the horizontal
direction can also be seen in Fig. 13. In comparison, in the case of
the tool tilt angle of 0°, material ahead of the welding tool moves
in a simple mode when it flows through the welding tool and
material flow in the horizontal direction is not obvious. Addition-
ally, material ahead of the tilt welding tool usually moves for a
longer distance before depositing in the welds compared with
the case of the tool tilt angle of 0°.
Fig. 14 shows the side views of the predicted material flow
paths in different horizontal planes located at 1.8 mm and 2.5
mm away from the top surface of the workpiece. It is seen clearly
from Fig. 14 that material flow in the vertical direction driven by
the tilt welding tool is noticeable and crossover can also be seen
among the material flow paths. In the horizontal planes located
at 1.8 mm and 2.5 mm away from the top surface of the workpiece,
some material ahead of the tilt welding tool flows towards the top
surface of the workpiece, others flows towards the bottom surface
of the workpiece and then deposits in the rear side of the pin. The
enhanced material flow towards the bottom surface of the work-
piece driven by the tilt welding tool can effectively fill up the cavity

behind the tool, which is beneficial to gain a sound weld. In con-

8 trast to this case, material flow in the vertical direction driven by
the perpendicular welding tool is not obvious and material ahead
6
of the welding tool almost flows in the same horizontal plane when
4 it flows through the welding tool.

2
4.4. Plastic deformation field
Front RS Rear RS Rear AS Front AS
0
0 40 80 120 160 200 240 280 320 360 4.4.1. Enhanced strain rate
Angle (°) The strain rate is considered to be a key parameter in the FSW
process to define the stirring action caused by the welding tool
Fig. 16. Enhanced strain rate at the bottom of the pin.
[37]. Fig. 15 shows the strain rate field in cross section of the welds
for the tool tilt angle of 2.5° and 0°. The larger size of the iso-strain
rate profile indicates that the tilt welding tool has a stronger stir-
ring action effect on the material in the vicinity of the tool. Strain
rate in the thin layer close to the lower parts of the pin side (shown
in Fig. 11(a)) is extracted to quantitatively study the stirring action

Table 2
TMAZ widths and variance for different strain values.

Strain L1 (mm) L2 (mm) L3 (mm) Variance

Predicted 0° 0.6 14.4 7.2 4.4 0.87
2.0 13.6 6.0 4.0 0.01
4.0 12.0 5.6 3.6 0.88
2.5° 0.6 15.6 8.8 6.0 2.89
2.0 14.4 7.2 4.4 0.09
4.0 14.0 6.4 3.2 0.42
Experiment 0° 13.57 5.95 3.81
measured 2.5° 14.55 6.82 4.09
Fig. 17. Reverse particles tracing methodology.

(a) 2.5° tool tilt angle (b) 0° tool tilt angle

Fig. 18. Macrostructure of the welds.
 S. Zhang et al. / International Journal of Heat and Mass Transfer 125 (2018) 32–42
(a) 2.5° tool tilt angle
Fig. 19. Comparison of the predicted and experimental measured TMAZ.
effect caused by different tool tilt angle, as shown in Fig. 16. It can
be seen clearly from Fig. 16 that the tilt welding tool can appar-
ently strengthen the stirring action to the material on the advanc-
ing side of the rear of the welding tool.
4.4.2. The characterization of deformation zone
It is known that the welds typically comprises three distinct
microstructural zones including the weld nugget zone (WNZ),
thermal-mechanically affected zone (TMAZ) and heat affected zone
(HAZ) [38]. TMAZ usually experiences much lower plastic strains
(eeff  0  5) [39] and no dynamic recrystallization occurs in this
zone. In other words, it is possible to characterize the shape and
size of the TMAZ by using an iso-strain line.
To refine the value of strain that can be used to determine the
TMAZ, reverse particles tracing methodology is used to calculate
the strain field of the workpiece, in which the Backward Euler time
integration scheme is applied [40]. As shown in Fig. 17, 500 tracing
particles are uniformly distributed in a 16
5 mm ZY plane (cross-
section plane) behind the welding tool. These tracing particles
started from the initial plane and moves along the positive X-
axis direction. When these particles moves to the front of the weld-
ing tool and the Z-axial and Y-axial coordinates of these particles
become constant. Then, an observation plane is placed ahead the
tool to analyze the strain field in the cross section. Strain can be
obtained through integrating the strain rate tensor with time along
the particle flow path. While the strain rate is computed from the
local velocity gradients [41].
Three lines are used, shown in Fig. 18, to compare the predicted
and measured TMAZ width; these lines are located 0, 2.5 and 5 mm
from the top surface of the workpiece for L1, L2 and L3 respectively.
A number of different strain values is used to define the limit of the
TMAZ, as shown in Table 2. It can be seen that a strain of 2 shows a
consistently good corresponding with the experimental values in
terms of the shape and size of TMAZ using this strain value. In
addition, it can also be seen from Figs. 18 and 19 that the tilt weld-
ing tool produces a larger deformation zone, which is attributed to
the stronger stirring action caused by the tilt welding tool. This
suggests that the approach proposed in this study can be used to
characterize the shape and size of TMAZ, which can also be used
to support the validity of the CFD model.
5. Conclusion
In this study, a CFD model is established to investigate the effect
of tool tilt angle on the in-process heat transfer and material trans-
fer during FSW. Based on the simulation results, the conclusions
can be summarized as following:
(1) A higher temperature is generated by the tilt welding tool on
the advancing side, which is attributed to the occurrence of
the incomplete contact state at the shoulder/workpiece
interface.
(2) The tilt welding tool generates a higher frictional force at the
tool/workpiece interface behind the tool, which significantly
improves material flow velocity behind the tool.
41
(b) 0° tool tilt angle
(3) The tilt welding tool generates a stronger stirring action to
the material in the vicinity of the welding tool, which is ben-
eficial for the mixing of material and the formation of the
friction stir welds.
(4) The predicted temperature profile and deformation zone
correspond with the experimental results.
Conflict of interest
Authors declare that there is no conflict of interest.
Acknowledgements
This work was supported by the National Natural Science Foun-
dation of China (Grant No. 51705280 and Grant No. 51375259),
Special Program for Applied Research on Super Computation of
the NSFC-Guangdong Joint Fund (the second phase, Grant No.
U1501501) and the Tsinghua National Laboratory for Information
Science and Technology.
References
[1] W.M. Thomas, Friction Stir Butt Welding, 1991.
[2] R.S. Mishra, Z.Y. Ma, Friction stir welding and processing, Mater. Sci. Eng.: R:
Rep. 50 (1) (2005) 1–78.
[3] X. Liu, S. Lan, J. Ni, Analysis of process parameters effects on friction stir
welding of dissimilar aluminum alloy to advanced high strength steel, Mater.
Des. 59 (2014) 50–62.
[4] Y.G. Kim, H. Fujii, T. Tsumura, et al., Effect of welding parameters on
microstructure in the stir zone of FSW joints of aluminum die casting alloy,
Mater. Lett. 60 (29) (2006) 3830–3837.
[5] S.R. Ren, Z.Y. Ma, L.Q. Chen, Effect of welding parameters on tensile properties
and fracture behavior of friction stir welded Al–Mg–Si alloy, Scripta Mater. 56
(1) (2007) 69–72.
[6] E. Majid, A.D. Hamed, H. Morteza, Investigations of tool tilt angle on properties
friction stir welding of A441 AlSl to AA1100 aluminum, J. Eng. Manuf. 23 (7)
(2016) 1234–1241.
[7] K.P. Mehta, V.J. Badheka, Effect of tilt angle on properties of dissimilar friction
stir welding copper to aluminum, Mater. Manuf. Process. 31 (2016) 255–263.
[8] G. Chen, H. Li, G. Wang, et al., Effects of pin thread on the in-process material
flow behavior during friction stir welding: a computational fluid dynamics
study, Int. J. Mach. Tools Manuf. 124 (2018) 12–21.
[9] H. Su, C.S. Wu, M. Bachmann, et al., Numerical modeling for the effect of pin
profiles on thermal and material flow characteristics in friction stir welding,
Mater. Des. 77 (2015) 114–125.
[10] Z. Yu, W. Zhang, H. Choo, et al., 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.
[11] G. Chen, Z. Feng, Y. Zhu, et al., An alternative frictional boundary condition for
computational fluid dynamics simulation of friction stir welding, J. Mater. Eng.
Perform. 25 (9) (2016) 4016–4023.
[12] H. Wang, P.A. Colegrove, J.F. dos Santos, Numerical investigation of the tool
contact condition during friction stir welding of aerospace aluminum alloy,
Comput. Mater. Sci. 71 (2013) 101–108.
[13] H.B. Chen, K. Yan, T. Lin, et al., The investigation of typical welding defects for
5456 aluminum alloy friction stir welds, Mater. Sci. Eng. A 433 (1) (2006) 64–
69.
[14] K.R. Seighalani, M.K.B. Givi, A.M. Nasiri, et al., Investigations on the effects of
the tool material, geometry, and tilt angle on friction stir welding of pure
titanium, J. Mater. Eng. Perform. 19 (7) (2010) 955–962.
[15] G.F. Zhang, W. Su, J. Zhang, et al., Visual observation of effect of tilting tool on
forging action during FSW of aluminium sheet, Sci. Technol. Weld. Joining 16
(1) (2011) 87–91.
42 S. Zhang et al. / International Journal of Heat and Mass Transfer 125 (2018) 32–42
[16] A. Arici, S. Selale, Effects of tool tilt angle on tensile strength and fracture
locations of friction stir welding of polyethylene, Sci. Technol. Weld. Joining 12
(6) (2007) 536–539.
[17] Z. Barlas, U. Ozsarac, Effects of FSW parameters on joint properties of AlMg3
alloy, Weld. J. 91 (1) (2012) 16S–22S.
[18] M. Dehghani, A. Amadeh, S.A.A. Akbari Mousavi, Investigations on the effects
of friction stir welding parameters on intermetallic and defect formation in
joining aluminum alloy to mild steel, Mater. Des. 49 (2013) 433–441, https://
doi.org/10.1016/j.matdes.2013.01.013.
[19] K. Kimapong, T. Watanabe, Effect of welding process parameters on
mechanical property of FSW lap joint between aluminum alloy and steel,
Mater Trans 46 (2005) 2211–2217.
[20] Y. Zhu, G. Chen, Q. Chen, et al., 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] Y. Xiao, H. Zhan, Y. Gu, et al., Modeling heat transfer during friction stir
welding using a meshless particle method, Int. J. Heat Mass Transf. 104 (2017)
288–300.
[22] X. He, F. Gu, A. Ball, A review of numerical analysis of friction stir welding,
Prog. Mater Sci. 65 (2014) 1–66.
[23] P.A. Colegrove, H.R. Shercliff, Experimental and numerical analysis of
aluminium alloy 7075–T7351 friction stir welds, Sci. Technol. Weld. Joining
8 (5) (2003) 360–368.
[24] 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.
[25] R. Nandan, G.G. Roy, T. Debroy, Numerical simulation of three-dimensional
heat transfer and plastic flow during friction stir welding, Metall. Mater. Trans.
A 37 (4) (2006) 1247–1259.
[26] H. Su, C.S. Wu, A. Pittner, et al., Thermal energy generation and
distribution in friction stir welding of aluminum alloys, Energy 77 (2014)
720–731.
[27] ANSYS Inc, ANSYS Ò Fluent 15.0.
[28] K.C. Mills, Recommended Values of Thermophysical Properties for Selected
Commercial Alloys, Woodhead Publishing, 2002.
[29] G. Chen, Q. Ma, S. Zhang, et al., Computational fluid dynamics simulation of
friction stir welding: a comparative study on different frictional boundary
conditions, J. Mater. Sci. Technol. (2017).
[30] J.R. Davis, Aluminum and Aluminum Alloys, ASM International, 1993.
[31] T. Sheppard, D.S. Wright, Determination of flow stress: part 1 constitutive
equation for aluminium alloys at elevated temperatures, Met. Technol. (2013).
[32] K.E. Tello, A.P. Gerlich, P.F. Mendez, Constants for hot deformation constitutive
models for recent experimental data, Sci. Technol. Weld. Joining 15 (3) (2010)
260–266.
[33] P.A. Colegrove, H.R. Shercliff, CFD modelling of friction stir welding of thick
plate 7449 aluminium alloy, Sci. Technol. Weld. Joining 11 (4) (2006) 429–441.
[34] A. Bastier, M.H. Maitournam, K.D. Van, et al., Steady state thermomechanical
modelling of friction stir welding, Sci. Technol. Weld. Joining 11 (3) (2006)
278–288.
[35] P. Carlone et al., Influence of process parameters on microstructure and
mechanical properties in AA2024-T3 friction stir welding, Metallogr.
Microstruct. Anal. 2 (2013) 213–222.
[36] 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.
[37] H. Pashazadeh, J. Teimournezhad, A. Masoumi, Numerical investigation on the
mechanical, thermal, metallurgical and material flow characteristics in friction
stir welding of copper sheets with experimental verification, Mater. Des. 55
(2014) 619–632.
[38] K.A.A. Hassan, P.B. Prangnell, A.F. Norman, et al., Effect of welding parameters
on nugget zone microstructure and properties in high strength aluminium
alloy friction stir welds, Sci. Technol. Weld. Joining 8 (4) (2003) 257–268.
[39] P.B. Prangnell, C.P. Heason, Grain structure formation during friction stir
welding observed by the ‘stop action technique’, Acta Mater. 53 (11) (2005)
3179–3192.
[40] N. Dialami, M. Chiumenti, M. Cervera, et al., Material flow visualization in
friction stir welding via particle tracing, Int. J. Mater. Form. 8 (2) (2015) 167–
181.
[41] A. Arora, Z. Zhang, A. De, et al., Strains and strain rates during friction stir
welding, Scripta Mater. 61 (9) (2009) 863–866.