Journal of Materials Processing Technology 241 (2017) 8692

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

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

Research Paper

Probing torque, traverse force and tool durability in friction stir

welding of aluminum alloys
V. Buchibabu a , G.M. Reddy b , A. De a,
a
Indian Institute of Technology Bombay, India
b
Defence Metallurgical Research Laboratory, Hyderabad, India

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

Article history: A coupled theoretical and experimental study was undertaken to examine the impact of plate thickness,
Received 14 June 2016 alloy strength and processing conditions on the susceptibility of tool failure in friction stir welding of
Received in revised form AA7075 and AA7039. A three-dimensional steady-state heat transfer model was used to compute the
22 September 2016
temperature eld in tool and workpiece alloy. The torque and traverse force, and the mechanical stresses
Accepted 7 November 2016
on the tool were estimated analytically and validated with the corresponding experimentally measured
Available online 9 November 2016
results for a range of processing conditions. The propensity of the tool to failure was assessed by a dura-
bility factor, which was estimated as a ratio of tool material strength at the computed peak temperature
Keywords:
Friction stir welding
and the resultant mechanical stress on the tool probe. The results showed that rise in welding speed,
AA7075 alloy strength and plate thickness could lead to increased mechanical stresses on tool probe, reduced
AA7039 durability factor and enhanced susceptibility of tool failure. The overall methodology was used success-
Torque fully to provide the underlying basis for two reported practical examples of tool failures in friction stir
Traverse force welding.
Tool durability 2016 Elsevier B.V. All rights reserved.

1. Introduction of the temperature and mechanical stresses experienced by the tool

AA7xxx series of alloys exhibits excellent combinations of low
density, good tensile and impact strength, and corrosion resistance
properties. Threadgill et al. (2009) showed friction stir welding
(FSW) as an efcient recourse to fusion arc welding of aluminum
alloys, as FSW avoided melting of workpiece alloy during welding.
Buchibabu et al. (2016) studied FSW of AA7039 alloy and reported
the weld joint properties to be close to that of the base material.
Cam and Mistikoglu, (2014) documented several applications of
AA7xxx series of alloys in critical structures to achieve high strength
to weight ratio and the important role of FSW to provide sound
joints in these alloys. Although FSW is increasingly used for the
joining of AA7xxx series of alloys in preference to fusion weld-
ing, the premature failure of the FSW tool especially for welding of
thicker plates and stronger alloys is a major concern (Bhadeshia and
Debroy, 2009). Rai et al. (2011) also demonstrated the severity of
forces on the FSW tool with increase in plate thickness and strength
of workpiece alloy. Arora et al. (2012) opined the tool probe as the
structurally weakest section that could fail due to high tempera-
ture and mechanical stresses during welding. A critical evaluation
Corresponding author.
E-mail address: amit@iitb.ac.in (A. De).
is therefore ever signicant for FSW of thicker plates and stronger
alloys.
Sorensen and Stahl (2007) reported a signicant increase in
tool traverse force with increase in the plate thickness in FSW of
AA6061. Trimble et al. (2012) observed a considerable increase in
tool traverse force with rise in welding speed during FSW of 5 mm
thick AA2024. Leitoa et al. (2012) noted a substantial increase
in tool torque with increase in both plate thickness and welding
speed. Astarita et al. (2014) showed signicant increase in tool tra-
verse force and axial force with increase in welding speed in FSW
of 5 mm thick AA2024. Manvatkar et al. (2012) used articial neu-
ral network based models to assess increase in tool traverse force
and torque with rise in welding speed in FSW of AA7075. Nielsen
(2009) and Leitoa et al. (2012) found that increase in welding
speed led to rise in tool traverse force and premature tool failure
during FSW of respectively of 9.5 mm thick AA7075 and 5 mm thick
AA5083. DebRoy et al. (2012) showed that the failure of tool probe
in FSW was more likely due to mechanical stresses arising out of
a combined bending and torsion. These studies clearly indicated a
need to examine the mechanical stresses on the tool as a precur-
sor to its unexpected failure during welding of thicker and stronger
aluminum alloys.
Prado et al. (2003) found signicant increase in rate of tool
probe wear with rise in welding speed in FSW of AA6061-

http://dx.doi.org/10.1016/j.jmatprotec.2016.11.008
0924-0136/ 2016 Elsevier B.V. All rights reserved.
 V. Buchibabu et al. / Journal of Materials Processing Technology 241 (2017) 8692 87

T6 + 20Vol%Al2 O3 metal matrix composite, where the tool was Table 1

Chemical composition (in wt%) of workpiece and tool materials (Woolman and
made with H13 steel. Liu et al. (2005) showed a substantial distor-
Mottram, 1969).
tion of the tool probe of WC-Co alloy during FSW of AC4A + 30%SiC
metal matrix composite. Gan et al. (2007) noted a considerable tool Workpiece Zn Mg Cu Cr Fe Mn Al
wear and mushrooming of the tool probe during FSW of L80 steel. AA7075 6.01 2.51 1.42 0.21 0.08 0.01 Bal.
According to these authors, the tool material yield strength should AA7039 3.62 4.29 0.1 0.01 0.28 0.08 Bal.
be higher than 400 MPa at 1000 C and hence, a typical tool of W-
25%Re alloy could only survive for FSW of L80 steels. Miyazawa et al. Tool C Cr Mo Mn Si Fe

(2012) suggested typical Iridium (Ir) based alloy as tool material to EN40 0.30.5 2.53.5 0.71.2 0.40.8 0.10.35 Bal.
avoid premature tool failure in FSW of 304 stainless steels. Although
these studies manifested that harder tool materials could delay the
failure of tool during welding of stronger alloys, no efforts were
made to examine the stresses and temperature that were likely to
be experienced by these tools during actual welding.
Several researchers attempted to compute the peak tempera-
ture, traverse force and torque experienced by the FSW tool as
these were difcult to monitor in real-time due to the inherent
nature of the process. Frigaard et al. (2001) and Khandkar et al.
(2003) computed the peak temperature in FSW of Al-Zn and Al-
Cu alloys, respectively. Su et al. (2015) showed that the tool probe
prole had a fair effect on the peak temperature in FSW of AA2024.
Colegrove and Shercliff (2005) calculated the tool traverse force
Fig. 1. Schematic illustration indicating thermocouple locations for on-line temper-
and torque along with the peak temperature in FSW of AA7075. ature monitoring in FSW. All dimensions are in mm.
Trimble et al. (2012) estimated the tool traverse force and torque
in FSW of AA2024 but did not analyse the mechanical stresses on
taram (1969) as shown in Table 1. The tool dimensions, and the
the tool. DebRoy et al. (2012) and De et al. (2014) reported coupled
range of process conditions and plate thicknesses are presented in
numerical-analytical models to obtain peak temperature, traverse
Tables 2 and 3, respectively. The range of process conditions were
force, torque and stresses on the tool during FSW of several alloys.
selected based on several trial experiments to obtain defect-free
Most of these studies were limited to the plate thicknesses lesser
joints. The tool was tilted at an angle of 2 with respect to the ver-
than 6 mm. A systematic evaluation of the mechanical stresses on
tical axis in the welding direction. Four K-type thermocouples of
the tool probe during FSW of thicker plates and stronger alloys is
1 mm diameter were used to monitor the thermal cycles at selected
therefore essential to provide a practical guideline for the appro-
locations as indicated by TC1, TC2, TC3 and TC4 in Fig. 1. The torque
priate selection of welding variables.
and traverse force were monitored using the in-built dynamometer
A coupled theoretical and experimental study was therefore
inside the friction stir welding machine at a simultaneous sampling
undertaken here to evaluate the torque, traverse force and mechan-
rate of 30 Hz (Ramanjaneyulu et al., 2013). For a welding condition,
ical stresses on the tool during FSW of two aluminum alloys
the time-averaged values of torque and traverse force were esti-
of different strengths and plate thicknesses. A three-dimensional
mated considering their respective transient records over a length
steady-state conduction heat transfer analysis was carried out con-
of 100 mm in the middle of the weld.
sidering the tool and workpiece as the solution domain. The tool
torque and traverse force were estimated considering the mechan-
3. Theoretical modeling
ical strength of the workpiece alloy at the computed temperature
eld. The computed thermal cycles at selected locations, and the
A steady-state three-dimensional heat conduction analysis was
estimated tool torques and traverse forces were tested with the
carried out considering the governing equation as
corresponding measured results for a wide range of welding con-      
ditions. A set of mechanics based relations were used to estimate T T T T
the components of stresses from the computed torque and tra- k + k + k + Q = CU1 (1)
x x y y z z x
verse force and the resultant maximum shear stress on the tool
probe. The propensity of tool failure due to mechanical stress was where , k, C, U1 and T referred to the density, thermal conduc-
assessed by a durability factor considering only the static loading tivity, specic heat, the welding speed and temperature variable,
on the tool (DebRoy et al., 2012). The durability factor presumed respectively. The term Q embodied the rate of heat generation per
the shear yield strength of the tool material as the limiting stress unit volume and was estimated following Arora et al. (2012) and
to failure and was estimated as a ratio of the same to the maximum De et al. (2014) as
resultant shear stress experienced by the probe for a FSW condi-  
Q = h [m (1 )y + f PN ](r U1 sin ) Ai /Vi (2)
tion. The sensitivity of the tool durability factor to plate thickness
and the strength of workpiece alloy was examined for a wide range where m and h referred to the fractional mechanical work due to
of welding conditions. sticking friction converted to heat and the fraction of total heat
transferred to workpiece, respectively; r and referred to the
radial distance of a point from the tool axis and its orientation
2. Experimental procedure with the welding direction; and PN , , and y depicted axial pres-
sure, angular speed of tool and the shear yield stress of workpiece
AA7075-T6 and AA7039-T6 plates were friction stir welded sep- alloy, respectively. The values of h and m were considered as 0.7
arately in butt joint congurations using a displacement controlled and 0.4, respectively (Mehta et al., 2013, 2014). The term (Ai /Vi )
FSW machine. These two alloys were chosen due to their con- depicted the ratio of tool-workpiece contact area to the shear vol-
siderable difference in room temperature mechanical strengths. ume in workpiece alloy adjacent to the tool (Nandan et al., 2006;
The chemical compositions of workpiece and tool material were Mehta et al., 2014). In Eq. (2), and f represented the local varia-
considered from ASM Handbook (1992) and Woolman and Mot- tions in fractional sliding and the coefcient of friction respectively
88 V. Buchibabu et al. / Journal of Materials Processing Technology 241 (2017) 8692

Table 2
Tool dimensions (in mm) considered.

Tool feature Plate thickness Shoulder diameter Probe diameter Probe length Thread pitch

at root at tip

Size (mm) 9.5 24 12 6 9.35 1.6

6.3 24 12 8 6.15 1.6

Table 3
Combination of tool rotational and welding speeds.

Aluminium Alloy Thickness (mm) Tool rotational speed, N (rpm) Welding speed, v (mm/s)

AA7075 9.5 355, 450, 560 1, 2, 3, 4*, 5*

AA7039 9.5 355, 450, 560 1, 2, 3
AA7075 6.3 355, 450 1, 2, 3

*performed at a tool rotational speed of 355 rpm only.

along the tool workpiece interface and were estimated following direction. The tool durability factor was estimated as P /max where
Mehta et al. (2013) as P was the shear yield strength of the tool material. The value ofP
for EN40 tool material was estimated as P = (930.24 0.33TP )/ 3
= 0.026 + C exp(r/1.87); (3)
  following Woolman and Mottaram (1969), where TP was computed
f = 0.51 exp rw (4) peak temperature experienced by the tool for a FSW condition. It is
therefore prudent that the susceptibility of the tool probe to failure
The value of the tting constant C in Eq. (3) was considered as
due to mechanical stress will reduce with increase in its durability
0.5 and 0.3 for AA7075 and AA7039 alloys, respectively.
factor for a welding condition.
A symmetric analysis was carried out considering the plane of
The numerical model was developed using the nite element
symmetry along the original weld joint interface. A lumped heat
based commercial analysis software ABAQUS (6.8EF-1 version). The
transfer coefcient as h0 (T T0 )0.25 was applied to account for
solution domain was discretized with three-dimensional eight
heat loss from the bottom surface where h0 = 90 W/m2 K. A con-
node brick elements DCC3D8 in ABAQUS with the temperature as
stant heat transfer coefcient of 30 W/m2 K was used for the rest of
the nodal degree of freedom. The user-dened subroutines DFLUX,
the surfaces. The shape of tool probe was presumed to be straight
UFILM and UMASFLD were used respectively to apply non-uniform
cylindrical and the tool tilt angle was neglected for simplicity in
heat ux along the tool workpiece interface, convective heat
numerical modeling.
transfer co-efcient at the workpiece bottom surface and weld-
The tool torque (M) and traverse force (F) were estimated analyt-
ing speed. The properties of the workpiece alloys were considered
ically in terms of tool geometry and welding conditions following
from the ASM Handbook (1992) and Simthells (1983) as shown in
Arora et al. (2012) as
Table 4.
RS L
   
M= r (1 )y + f PN (2 rdr) + r (1 )y + f
y 4. Results and discussion
RP 0
4.1. Computed temperature eld and thermal cycles
RP
  Fig. 2(a) and (b) shows the computed temperature distribution
(2 rdl) + r (1 )y + f PN (2 rdr) (5)
during FSW of 9.5 mm thick AA7075 alloy plates for two differ-
0 ent welding speeds of 1 and 3 mm/s respectively at a constant
tool rotational speed of 450 rpm. The region heated above 650 K,
RS L RP which is nearly 0.85 times the solidus temperature of AA7075,
   
F= f PN (2 rdr) +
y dA + f PN (2 rdr) (6)
RP 0 0

where L, Rs , Rp and y referred to the probe length, tool shoulder

diameter, probe diameter at its root, and temperature dependent
yield strength of the workpiece alloy in front of the tool. It is
noteworthy that Eqs. (5) and (6) embody contributions from tool
shoulder, and both probe vertical and bottom surfaces respectively
by the rst, and second and third terms. The computed traverse
force and torque on the tool probe were used to estimate the com-
ponents of normal and shear stress due to bending, and shear stress
due to torsion, respectively. The maximum shear stress (max ) on
the tool probe was estimated in terms of those components of
stresses as



2 B
 2  2
max = + B + T cos + T sin (7)
2
Fig. 2. Computed temperature eld in FSW of (a, b) 9.5 mm thick AA7075, (c) 6.3 mm
where B referred to normal stress due to bending, and B and T
thick AA7075 and (d) 9.5 mm thick AA7039 for a constant tool rotational speed of
were shear stresses due to bending and torsion, respectively, and 450 rpm. The corresponding welding speeds are (a) 1, (b) 3, (c, d) 1 mm/s. (DS and
was the angle subtended between B and T in a counter-clockwise DP refer respectively to relative positions of shoulder and probe diameters).
 V. Buchibabu et al. / Journal of Materials Processing Technology 241 (2017) 8692 89

Table 4
Thermophysical properties of aluminum alloys (ASM Handbook, 1992; Simthells, 1983). The temperature variable T varies as 300 K T 900 K.

AA7075-T6 AA7039-T6 AA5083 AA2219

Density, kg/m3 2800 2750 2680 2770

Solidus Temp., K 749 775 645 645
Th. conductivity, W/m K 74.07 + 0.25 T4.621e 83.86 + 0.20 T3.413e 111.026 + 0.09217 T 75.83 + 0.102 T + 5.136e
5 T2 7 T2 4 T2
Specic heat, J/kg K 850.42 + 1.21 T4.396e 859.02 + 1.21 T4.13e 805.38 + 0.39 T + 9.979e 937.84 0.67 T
4 T2 4 T2 5 T2 0.00148 T2
Yield strength, MPa 11.76 + 8206.07 3.22 + 4846.31 10.34 + 139.94/ 9.97 + 219.50/
exp(T/107.60) exp(T/118.48) [1 + exp{(T 553.12)/48.37}] [1 + exp{(T 540.4)/55.69}]

is shown in red color and considered as the softened region to
experience the traction by the tool probe. The size of the high tem-
perature region was slightly wider underneath the tool shoulder
and reduced towards the probe bottom that was attributed to larger
contact area and greater rate of frictional heat generation along
the shoulder workpiece interface in comparison to that around
the probe workpiece interface. A comparison of Fig. 2(a) and (b)
showed a decrease in size of the high temperature region with rise
in welding speed from 1 to 3 mm/s that was attributed to lower
amount of heat generation rate per unit length of weld.
Fig. 2(c) shows the computed temperature eld in FSW of
6.3 mm thick AA7075 alloy. A comparison of Fig. 2(a) and (c) showed
an increase in size of the high temperature region in the latter case
due to smaller plate thickness and resulting decrease in rate of
conduction heat loss through the plate material. Fig. 2(d) shows
the computed temperature eld in FSW of 9.5 mm thick AA7039
alloy. For the same processing conditions and plate thickness, the
thermal proles for FSW of AA7075 and AA7039 were compara-
ble, as observed in Fig. 2(a) and (d), due to similar thermophysical
properties of these two alloys.
Fig. 3 shows a comparison of the numerically computed and the
corresponding measured thermal cycles at the selected monitor-
ing locations (refer Fig. 1) in FSW of AA7075 and AA7039 alloys. A
comparison of Fig. 3(a) and (b) showed a decrease in the peak tem-
perature and increase in the cooling rate as the welding speed was
increased from 1 to 3 mm/s for a particular tool rotational speed.
This was attributed to the reduced rate of heat generation per unit
length of weld with rise in welding speed. Fig. 3(a) and (c) show a
slight increase in peak temperature with decrease in plate thickness
from 9.5 to 6.3 mm for a specic welding condition. A comparison
of Fig. 3(a) and (d) depicted that the nature of the thermal cycles
was nearly similar in FSW of both AA7075 and AA7039 alloys for a
given welding condition and plate thickness. Fig. 3 also depicted a
fair agreement between the numerically computed and measured
thermal cycles.
4.2. Measured and estimated torque and traverse force
Fig. 4 showed a comparison of the experimentally measured and
the corresponding estimated tool torques during FSW of 9.5 mm
thick AA7075 alloy plates and of AA7039 plates as function of tool
rotational speed and welding speed. The torque increased with
increase in the welding speed and decrease in tool rotational speed.
Increase in tool rotational speed resulted in greater rate of heat
generation and softening of the workpiece alloy. In contrast, rise
in welding speed led to decrease in heat generated per unit length
of the weld. As a result, the workpiece alloy was not softened ade-
quately and the tool torque was increased. Pew et al. (2007) and
Mehta et al. (2013) also reported a similar phenomena in FSW of
both Al-Cu and Al-Zn alloys. The error bars in Fig. 4 depicted a mea-

Fig. 3. Comparison of numerically computed and experimentally measured thermal cycles in FSW of (a, b) 9.5 mm thick AA7075, (c) 6.3 mm thick AA7075 and (d) 9.5 mm
thick AA7039 plates for a tool rotational speed of 450 rpm. The corresponding welding speeds are (a) 1, (b) 3, and (c, d) 1 mm/s.
90 V. Buchibabu et al. / Journal of Materials Processing Technology 241 (2017) 8692
Fig. 4. Experimentally measured and the corresponding analytically estimated tool torques in FSW of 9.5 mm thick (a, b, c) AA7075 and (d, e, f) AA7039 plates.
sure of the variability in torque measurement over the length of the
weld. Overall, Fig. 4 showed a sound agreement of the estimated
torques with the corresponding measured results.
A comparison of the experimentally measured and the corre-
sponding estimated tool traverse forces during FSW of 9.5 mm thick
AA7075 and of AA7039 alloy plates were shown in Fig. 5. Both
experimentally measured and estimated results showed signicant
increase in tool traverse force with increase in welding speed. The
heat generated per unit length of weld reduced with increase in
welding speed resulting in inadequate softening of workpiece alloy.
As a result, the tool was required to pass through the colder and
harder workpiece alloy leading to increase in the traverse force. In
contrast, the traverse force reduced gently with increase in rota-
tional speed that was attributed to greater rate of heat generation
and softening of workpiece alloy. A comparison of Fig. 5(ac) with
(df) showed higher tool traverse force for AA7075 compared to
AA7039 alloys for the same process conditions. This was attributed
to greater mechanical strength of AA7075 alloy compared to that
of AA7039 (ASM Handbook, 1992; Simthells, 1983). Overall, a fair
agreement between the experimentally measured and the esti-
mated tool traverse forces were noted in Fig. 5 for a wide range
of process conditions.
Fig. 4(a) showed that the tool torque increased approximately
from 60 (1.6) to 83 (1.3) Nm in FSW of 9.5 mm thick AA7075
alloy as the welding speed was increased from 1 to 3 mm/s at a con-
stant tool rotational speed of 355 rpm. The corresponding increase
in tool traverse force was from 3.5 (0.3) to 5.1 (0.4) kN as shown
in Fig. 5(a). A set of additional experiments was carried out fur-
ther to examine the effect of welding speed on both tool torque
and traverse force. As the welding speed was increased to 4 and
5 mm/s at a constant tool rotational speed of 355 rpm, the tool
torque and traverse force increased to around 99 (2.3) and 110
(1.8) Nm, and 5.9 (0.3) and 6.8 (0.2) kN, respectively. The cor-
responding estimated values were around 87 and 94 Nm, and 5.4
and 6.0 kN. Although sound weld joints were obtained for the com-
plete range of welding speeds from 1 to 5 mm/s at the tool rotational
speed of 355 rpm, the welds at the highest speed of 5 mm/s showed
discontinuities in repeated trials.
Fig. 6(a) and (b) showed the experimentally measured and cor-
responding analytically estimated tool torques and traverse forces
in FSW of 6.3 mm thick AA7075 alloy as function of welding speed
at a constant tool rotational speed of 355 rpm. A comparison of Figs.
6(a) and 4(a) indicated a decrease in the tool torque with smaller
plate thickness. Likewise, Figs. 6(b) and 5(a) depicted a consider-
able decrease in the tool traverse force during FSW of 6.3 mm thick
AA7075 alloy compared with that observed during FSW of 9.5 mm
thick AA7075. Smaller tool torque and traverse force, especially
the latter, during FSW of lower plate thickness was attributed to
reduced rate of heat diffusion through the workpiece and greater
softening of the alloy through the plate thickness.
The inuence of process conditions on tool torque and traverse
force was rarely reported for FSW of AA7075 and AA7039 alloys.
Colegrove and Shercliff (2005) reported typical traverse force of
around 3 kN in FSW of 6.35 mm thick AA7075 at a tool rotational
speed of 250 rpm and welding speed of 5 mm/s. Sorensen and Stahl
(2007) observed an increase in tool traverse force from 2.1 to 3.2 kN
with rise in probe length from 1.7 to 5.6 mm in FSW of AA6061.
In FSW of 4.8 mm thick AA2024 alloy at a tool rotational speed of
450 rpm, Trimble et al. (2012) found the torque and traverse force to
vary from 10 to 55 Nm and from 1 to 6 kN, respectively for welding
speeds ranging from 1.0 to 8.3 mm/s. The order of magnitude of
tool torque and traverse force and their nature of variation with
the process conditions reported in Figs. 46 are comparable with
that reported in published literature.
4.3. Estimation of tool durability
The estimated values of the tool durability factor in FSW of
9.5 mm thick AA7075 and AA7039 alloys are plotted in Fig. 7 for
the range of process conditions and tool material considered here.
Increase in welding speed led to decrease in the tool durabil-
ity factor as shown in Fig. 7 that was attributed to decrease in
heat generated per unit length of the weld and softening of the
workpiece alloy, and consequent rise in torque, traverse force, and
the resulting max on the tool probe. For example, the computed
peak temperature reduced from 690 to 647 K and max on tool
probe increased from 252.2 to 404.4 MPa as the welding speed was
increased from 1 to 3 mm/s for a tool rotational speed of 355 rpm
in FSW of 9.5 mm thick AA7075 alloy. The shear yield strength (p )
of the tool material corresponding to the computed peak temper-
 V. Buchibabu et al. / Journal of Materials Processing Technology 241 (2017) 8692
Fig. 5. Experimentally measured and the corresponding analytically estimated tool traverse force in FSW of 9.5 mm thick (a, b, c) AA7075 and (d, e, f) AA7039 plates.
Fig. 6. Experimentally measured and the corresponding analytically estimated (a)
tool torque and (b) traverse force in FSW of 6.3 mm thick AA7075 alloy as function
of welding speed for a constant tool rotational speed of 355 rpm.
atures of 690 and 647 K were found as 698.4 and 711.9 MPa. As a
result, the tool durability factor reduced from 2.8 to 1.8 as shown
in Fig. 7(a) for increase in welding speed from 1 to 3 mm/s at a tool
rotational speed of 355 rpm.
Fig. 7 also showed an increase in the durability factor with rise in
tool rotational speed that was attributed to greater rate of heat gen-
eration and softening of workpiece alloy at faster rotational speeds.
For example, the computed peak temperature increased from 647
to 711 K and the resultant maximum stress (max ) on the tool probe
reduced from 404.4 to 219.1 MPa as the tool rotational speed was
increased from 355 to 560 rpm at a welding speed of 3 mm/s in FSW
of 9.5 mm thick AA7075 alloy. The shear yield strength (p ) of the
tool material at the computed peak temperatures of 647 and 711 K
were 711.9 and 691.8 MPa, respectively. As a result, the tool dura-
bility factor increased from 1.8 to 2.8 as shown in Fig. 7(a) for rise
in tool rotational speed from 355 to 560 rpm at a welding speed of
3 mm/s. For a given FSW condition, the tool durability was higher
for AA7039 alloy compared to AA7075 as shown in Fig. 7(a) and
(b). This was attributed to lesser mechanical strength of AA7039
alloy that resulted in lower traverse force as observed in Fig. 5. As a
result, the resultant mechanical stress on the tool reduced and the
tool durability increased for FSW of AA7039 in comparison to that
for AA7075.
The tool durability factor was also found sensitive to the plate
thickness. For example, the estimated tool durability factors were
91
Fig. 7. Estimated values of tool durability factors as function of welding speed and
tool rotational speed in FSW of 9.5 mm thick (a) AA7075 and (b) AA7039 alloys.
found to reduce from 7.8 to 5.9 in FSW of 6.3 mm thick AA7075
alloy for an increase in welding speed from 1 to 3 mm/s at a tool
rotational speed of 355 rpm. These tool durability values are much
higher compared to the same estimated in FSW of 9.5 mm thick
AA7075 that are shown in Fig. 7(a).
Leitoa et al. (2012) and Doude et al. (2015) reported tool probe
failure in FSW of AA5083 and in AA2219 alloys at tool rotational
speeds and welding speeds of (350 rpm, 5 mm/s), and (150 rpm,
1.67 mm/s), respectively. An evaluation of the tool durability fac-
tors for these two cases were undertaken to examine the usefulness
of the methodology presented in this work. The thermo-physical
materials properties of AA5083 and AA2219 alloys were considered
following Table 4 for the numerical and analytical calculations. For
FSW of AA5083, the computed peak temperature and the analyti-
cally estimated tool traverse force and torque were found as 601 K,
4.9 kN and 71 Nm, respectively. The consequent max on the tool
probe and the tool durability factor were estimated as 426 MPa and
0.94, respectively. In other words, max on the tool probe was fairly
higher than the shear yield strength of the tool material at the peak
temperature during welding leading to an early failure of tool probe
as observed by Leitoa et al. (2012). Similarly, the computed peak
temperature and the analytically estimated tool traverse force and
torque were found as 456 K, 8.3 kN and 66 Nm, respectively for the
other example of FSW of AA2219 alloy. The corresponding max on
the tool probe and the tool durability factor were found as 478 MPa
92 V. Buchibabu et al. / Journal of Materials Processing Technology 241 (2017) 8692
and 0.45, respectively. An early tool failure in this case was also
unavoidable as also reported by Doude et al. (2015). Evidently, the
unexpected tool failure in both of these cases were avoidable by
selection of appropriate welding conditions through a systematic
approach as presented here.
In summary, a three-dimensional steady-state heat conduction
model and mechanics based analytical relations were used to com-
pute peak temperature, and tool torque and traverse force in FSW
process. The computed values of thermal cycles, torque and tra-
verse force were tested for FSW of 9.5 mm thick AA7039 and,
6.3 mm and 9.5 mm thick AA7075 alloys for a wide range of con-
ditions. The estimated tool torque and traverse force were used to
compute components of stresses due to bending and torsion, and
the maximum resultant shear stress on the tool probe. A ratio of
the shear yield strength of the tool material to the maximum shear
stress on the tool probe was conceived as a tool durability factor
that showed signicant reduction with increase in welding speed,
plate thickness and strength of workpiece alloy. The tool durability
factors were found lesser than unity for typical FSW cases reported
in literature that encountered premature tool failure.
5. Conclusions
(1) The susceptibility of tool failure during FSW of 6.3 and 9.5 mm
thick AA7075 and of 9.5 mm thick AA7039 alloys was exam-
ined in terms of a durability factor that was estimated as the
ratio of the shear strength of the tool material to the resultant
maximum shear stress on the tool probe.
(2) An increase in both welding speed and plate thickness resulted
in lower peak temperature, and higher torque, traverse force
and resultant maximum shear stress on the tool. As a result,
the tool durability factor reduced signicantly with increase in
welding speed and plate thickness.
(3) For a given welding condition, increase in workpiece alloy
strength led to higher traverse force and resultant maximum
shear stress, and lower tool durability. Thus the durability factor
in FSW of AA7075 alloy remained consistently lower compared
to that for AA7039.
(4) Tool durability factor lesser than unity depicts a higher resul-
tant stress on the tool probe than shear strength of the tool
material. The failure of the tool in such a case is imminent as
shown here for two distinct cases of tool fracture reported in
independent literature.
References
ASM Handbook, 1992. Properties and selection: non-ferrous alloys and special
purposes. Materials, Vol. 2.
Arora, A., Mehta, M., De, A., Debroy, T., 2012. Load bearing capacity of tool pin
during friction stir welding. Int. J. Adv. Manuf. Technol. 61, 911920.
Astarita, A., Squillace, A., Carrino, L., 2014. Experimental study on the forces acting
on the tool in the friction-stir welding of AA2024 T3 Sheets. J. Mater. Eng.
Perform. 23, 37543761.
Bhadeshia, H.K.D.H., Debroy, T., 2009. Critical assessment: friction stir welding of
steels. Sci. Technol. Weld. Join. 14, 193196.
Buchibabu, V., Reddy, G.M., Kulkarni, D., De, A., 2016. Friction stir welding of a
thick Al-Zn-Mg alloy plate. J. Mater. Eng. Perform. 25, 11631171.
Cam, G., Mistikoglu, S., 2014. Recent developments in friction stir welding of
Al-alloys. J. Mater. Eng. Perform. 23, 19361952.
Colegrove, P.A., Shercliff, H.R., 2005. 3-Dimensional CFD modelling of ow round a
threaded friction stir welding tool prole. J. Mater. Process. Technol. 169,
320327.
De, A., Bhadeshia, H.K.D.H., DebRoy, T., 2014. Friction stir welding of mild steel:
tool durability and steel microstructure. Mater. Sci. Technol. 30, 10501056.
DebRoy, T., De, A., Bhadeshia, H.K.D.H., Manvatkar, V.D., Arora, A., 2012. Tool
durability maps for friction stir welding of an aluminum alloy. Proc. R. Soc. A
468, 35523570.
Doude, H., Schneider, J., Patton, B., Stafford, S., Waters, T., Varner, C., 2015.
Optimizing weld quality of a friction stir welded Aluminum Alloy. J. Mater.
Process. Technol. 222, 188196.
Frigaard, ., Grong, ., Midling, .T., 2001. A process model for friction stir welding
of age hardening aluminum alloys. Metall. Mater. Trans. A 32A, 11891200.
Gan, W., Li, Z.T., Khurana, S., 2007. Tool materials selection for friction stir welding
of L80 steel. Sci. Technol. Weld. Join. 12, 610613.
Khandkar, M.Z.H., Khan, J.A., Reynolds, A.P., 2003. Prediction of temperature
distribution and thermal history during friction stir welding: input torque
based model. Sci. Technol. Weld. Join. 8, 165174.
Leitoa, C., Louro, R., Rodriguesa, D.M., 2012. Using torque sensitivity analysis in
accessing Friction Stir Welding/Processing conditions. J. Mater. Process.
Technol. 212, 20512057.
Liu, H.J., Feng, J.C., Fujji, H., Nogi, K., 2005. Wear characteristics of a WC-Co tool in
friction stir welding of AC4A + 30vol%SiCp composite. J. Mach. Tool. Manuf. 45,
16351639.
Manvatkar, V.D., Arora, A., De, A., Debroy, T., 2012. Neural network models of peak
temperature, torque, traverse force, bending stress and maximum shear stress
during friction stir welding. Sci. Technol. Weld. Join. 17, 460466.
Mehta, M., Chatterjee, K., De, A., 2013. Monitoring of torque and traverse force in
friction stir welding from input electrical signatures of driving motors. Sci.
Technol. Weld. Join. 18, 191197.
Mehta, M., De, A., Debroy, T., 2014. Material adhesion and stresses on friction stir
welding tool pins. Sci. Technol. Weld. Join. 19, 534540.
Miyazawa, T., Iwamoto, Y., Maruko, T., Fujii, H., 2012. Friction stir welding of 304
stain less steel using Ir based alloy tool. Sci. Technol. Weld. Join. 17, 207212.
Nandan, R., Roy, G.G., Debroy, T., 2006. Numerical simulation of three-dimensional
heat transfer and plastic ow during friction stir welding. Metall. Mater. Trans.
A 37, 12471259.
Nielsen, B., 2009. Developing Response Surfaces Based on Tool Geometry for a
Convex Scrolled Shoulder Step Spiral (CS4) Friction Stir Processing Tool Used
to Weld AL 7075. M.S. Thesis. Brigham Young University, Utah, USA.
Pew, J.W., Nelson, T.W., Sorensen, C.D., 2007. Torque based weld power model for
friction stir welding. Sci. Technol. Weld. Join. 12, 341347.
Prado, R.A., Murr, L.E., Soto, K.F., McClure, J.C., 2003. Self-optimization in tool wear
for friction-stir welding of Al 6061 + 20% Al2 O3 MMC. Mater. Sci. Eng. A 349,
156165.
Rai, R., De, A., Bhadeshia, H.K.D.H., DebRoy, T., 2011. Review: friction stir welding
tools. Sci. Technol. Weld. Join. 16, 325342.
Ramanjaneyulu, K., Madhusudhan Reddy, G., Venugopal Rao, A., Markandeya, R.,
2013. Structure-property correlation of AA2014 friction stir welds: role of tool
pin prole. J. Mater. Eng. Perform. 22, 22242240.
Simthells, C.J., 1983. Smithells Metals Reference Book, sixth edition. Butterworths,
London, 22-14-22-15.
Sorensen, C.D., Stahl, A.L., 2007. Experimental measurements of load distributions
on friction stir weld pin tools. Metall. Mater. Trans. B 38, 451459.
Su, H., Wu, C.S., Bachmann, M., Rethmeier, M., 2015. Numerical modeling for the
effect of pin proles on the thermal and material ow characterstics in friction
stir welding. Mater. Des. 77, 114125.
Threadgill, P.L., Leonard, A.J., Shercliff, H.R., Withers, P.J., 2009. Friction stir welding
of aluminium alloys. Int. Mater. Rev. 54, 4993.
Trimble, D., Monaghan, J., Odonnell, G.E., 2012. Force generation during friction
stir welding of AA2024-T3. CIRP Ann. Manuf. Technol. 61, 912.
Woolman, J., Mottram, R.A., 1969. The mechanical and physical properties of the
British standard En Steels (B.S. 970-1955: Volume 3 (En40 to En 363), 1st ed.
Pergamon press, Oxford, pp. 129.