j m a t e r r e s t e c h n o l .

2 0 1 9;8(2):1684–1693

Available online at www.sciencedirect.com

www.jmrt.com.br

Original Article

Influence of friction stir welding parameters on

metallurgical and mechanical properties of
dissimilar AA5454–AA7075 aluminum alloys

Mohamed M. Abd Elnabi a , Abou Bakr Elshalakany a,b,∗ , M.M. Abdel-Mottaleb b ,

T.A. Osman a , A. El Mokadem a
a Mechanical Design and Production Engineering Department, Cairo University, Giza, Egypt
b Production Engineering and Printing Technology Department, Akhbar El Yom Academy, Giza, Egypt

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

Article history: Friction Stir Welding (FSW) is a solid-state welding process used for welding similar and
Received 24 February 2018 dissimilar materials. FSW is especially suitable to join Al alloys sheets, and this technique
Accepted 3 October 2018 allows different material couples to be welded continuously. In this study dissimilar joints
Available online 18 January 2019 between aluminum alloy (AA5454) and aluminum alloy (AA7075) produced by friction stir
welding, to optimize these parameters and determine which of them is significant by using
Keywords: Taguchi L16 optimization method. Seven parameters at two levels were selected in this
Friction stir welding study. The selected parameters are tool rotational speed, traverse speed, pin profile (based
Taguchi method on taper angle), the ratio between shoulder diameter (D) and pin diameter (d) (D/d ratio),
Dissimilar metals tool tilt angle, plunge depth, and base metal location (weld location)). The ultimate tensile
Tensile strength strength (UTS) and ductility are considered as the mechanical properties of the dissimilar
Ductility joints. Then, mathematical models are built for ultimate tensile strength and ductility as
a function of significant parameters/interactions using response surface methodology. In
addition, the microstructures of the optimum joint and the weakest joint are studied using
optical microscopy. The results of this work showed that the rotational speed, traverse speed,
D/d ratio and plunge depth are significant parameters in determining UTS (mean, signal to
noise ratio (S/N)) at different confidence levels, but pin profile, location of base metal and
tool tilt angle are insignificant parameters at any confidence levels. The traverse speed has
the highest contribution to the process for UTS about 18.5% and 16.9% for S/N ratio and
mean, respectively. The accuracy of the models according to the UTS is 97.6% and 99.5% for
mean and S/N ratio, respectively. The maximum joint efficiency, compared to the strength
of the AA5454, is 85.3%.
© 2018 Brazilian Metallurgical, Materials and Mining Association. Published by Elsevier
Editora Ltda. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).

∗
Corresponding author at: Mechanical Design and Production Engineering Department, Cairo University, Giza, Egypt.
E-mail: eng bakr2011@yahoo.com (A.B. Elshalakany).
https://doi.org/10.1016/j.jmrt.2018.10.015
2238-7854/© 2018 Brazilian Metallurgical, Materials and Mining Association. Published by Elsevier Editora Ltda. This is an open access
article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 j m a t e r r e s t e c h n o l . 2 0 1 9;8(2):1684–1693 1685

plate, 1100–2000 rpm rotational speed, 22–68 mm/min traverse

1. Introduction speed and 4–6 KN axial force. [5,6,8,9,12]. Most of the studies
have been used a pin length smaller than the thickness of a
From the previous studies, it is clear that all the optimiza-
material to be welded about 0.1–0.5 mm. The ratios between
tion studies of FSW parameters investigate only two or three
shoulder diameter and pin diameter are between 3 and 4. At
variables with neglecting the effect of other variables. It is
3 mm and 6 mm thick plate, the diameters of the shoulder
relatively not effective on studying the process. The thesis
and the pin are 18 mm and 6 mm, respectively. The study of
is focused on studying most of FSW parameters together by
the interactions between the process parameters shows that
Taguchi method. The process a rotating FSW tool is plunged
the interaction variables have a little contribution to the pro-
between two clamped plates. The frictional heat causes a plas-
cess. For example [13], the interaction parameters included
ticized zone to form around the tool. The rotating tool moves
rotational speed/traverse speed, rotational speed/axial force
along the joint line. Then, a consolidated solid-phase joint is
and rotational speed/traverse speed have a contribution to the
formed [1,2]. FSW being a solid-state process eliminates many
process about 2.5%.
of the defects associated with fusion welding techniques such
The aim of this work is to study the effect of pro-
as shrinkage, solidification cracking and porosity. The weld-
cess parameters on the mechanical properties of dissimilar
ing temperature is approximately between 70 and 90% of the
joints between aluminum alloy (AA5454) and aluminum alloy
melting temperature of a material to be welded. The mech-
(AA7075) produced by friction stir welding, to optimize these
anism of the FSW is thermo-mechanical which needs three
parameters and to determine which of them is significant by
elements to complete the process [3]. These elements are heat
using Taguchi optimization method. Experimental work was
generation, plastic deformation and forging. The heat is pro-
carried out to produce friction stir welding joints between
duced by the friction action between the material and the tool
AA5454 and AA7075 at different levels of process parameters
of the FSW. The tool used in the process is a cylindrical tool
(tool rotational speed, traverse speed, pin profile (based on
consisting of a shoulder and a probe [4–6]. Then, the plasti-
taper angle), the ratio between shoulder diameter (D) and pin
cized metals are forged by the shoulder of the tool. The base
diameter (d) (D/d ratio), tool tilt angle, plunge depth, and base
materials must be fixed and clamped before welding to pro-
metal location (weld location)). The ultimate tensile strength
hibit the material from moving during the process. Formation
and ductility are considered as the mechanical properties of
of friction stir processing zone is influenced by material flow
the dissimilar joints.
behavior under the action of rotating tool [7,8]. However, mate-
rial flow behavior is mainly affected by tool geometry and
welding parameters. The process parameters are classified
as welding process parameters and tool design parameters. 2. Theory of experimental design
The welding process parameters are tool rotational speed, tra-
verse speed, dwell time, plunge depth, axial force, and tool Taguchi L16-orthogonal is employed for experiments. L16 (215 )
tilt angle. The tool design parameters are pin profile, shoulder has 16 rows corresponding to the number of tests with 15
diameter, the ratio between shoulder diameter and pin diame- columns at two levels. The plan of experiment is made of
ter. When welding dissimilar materials, there are another two 16 tests in which the first column is assigned to rotational
process parameters: tool offset and base metal location in a speed (rpm), the second column to traverse speed (mm/min),
joint (weld location) [9,10]. Recently, joining dissimilar mate- the fourth column to D/d ratio, the fifth column to pin profile
rials by the FSW method has become a very hot issue. The (based on taper angle), the eighth column to plunge depth,
dissimilar materials can be dissimilar aluminum alloy [11,12], twelfth column to tool tilt angle, eleventh column to loca-
or aluminum alloy to copper [13], aluminum alloy to stainless tion of base metal (lower metal (LM), AA5454- based on tool
steel and steel [14], etc. At the condition of aluminum alloys, rotation direction) and the remaining are assigned to the inter-
from previous studies, all the studies of the optimization of actions as shown in Fig. 1 [14]. Tests are replicated, resulting in
FSW parameters on similar or dissimilar joints of aluminum a total of 48 tests, to allow the analysis of variance and signal
alloys (by Taguchi methods or full factorial design) investi- to noise ratio calculation. Seven factors are used at two lev-
gate only two or three parameters. The results of the most els. Table 1 indicates the factors studied and the assignment
studies show that the rotational speed or the traverse speed of the corresponding levels. S/N ratios are calculated from the
or both are the significant parameters of FSW [4,5,8,10,11] measured values. In this study, the experimentally observed
but some studies show that the pin profile [7] or axial force
[5,8,10] is the significant parameter. The studies show that
the parameters of FSW may be significantly affected in the
process due to many reasons or conditions such as selected
parameters; levels of the parameters; the number of parame-
ters and joint conditions (material, similar or dissimilar joint,
material thickness). The optimal levels of the process param-
eters from the previous studies (aluminum alloy joint) can be
summarized as follows: With 3 mm thick plate, 1000–1225 rpm
– rotational speed and 17–35 mm/min – traverse speed. With
5 mm thick plate, 900–1600 rpm rotational speed, 28–160 mm
traverse speed and 2.5–7 KN axial force. With 6 mm thick Fig. 1 – Linear graph for L16-OA.
1686 j m a t e r r e s t e c h n o l . 2 0 1 9;8(2):1684–1693

Table 1 – Control variables and corresponding levels.

Control variables Level 1 Level 2

Rotational speed (rpm) 1000 1225

Traverse speed (mm/min) 17 21
Pin profile based on taper angle (◦ ) 0◦ Cylindrical profile 17◦ Tapered profile
D/d ratio 3 4
Tilt angle (◦ ) 1.5◦ 2◦
Location of base metal (LM) based on tool rotation direction (◦ ) −360◦ Retreating side +360◦ Advancing side
Plunge depth (mm) 0.1 0.25

Table 2 – Chemical composition of the base metals used in the second process.
Material Chemical composition (wt.%) Properties

Cr Cu Fe Mg Mn Si Ti Zn Al UTS (MPa) Ductility %

AA7075 0.28 1.59 0.16 1.61 0.05 0.27 0.09 5.18 90.65 402 10
AA5454 0.20 0.04 0.21 2.50 0.96 0.25 - 0.12 95.22 260 4

UTS values and ductility values is the higher the better [15–18].
The related equations are as follows:
  
S/N = −10 Log 1/n 1/y2i (1)

where n is the number of observations, and y is the observed

data.

Fig. 2 – Simple drawing of the FSW tools used in the first

3. Experimental procedures
process.

3.1. Base metals

The base metals used in this study are AA5454 and AA7075 of
similar plate thickness of 3.5 mm. The chemical composition
and the mechanical properties of the base metals are given in
Table 2.
3.2. FSW tools
The tool parameters selected in this study on the Taguchi array
L16 are pin profile and diameter ratio between shoulder diam-
eter (D) and pin diameter (d) (D/d ratio). Two levels of each
parameter are used which mean that the array of the tool
parameters on the L16-OA contains four different tools. The
pin profile was designed based on the taper angle of the pin (0◦ :
cylindrical pin, 0◦ < tapered pin). The D/d ratio is determined
by dividing the diameter of the shoulder to the diameter of the
pin at mid-length. The diameter ratios used in this array are 3
and 4. Simple drawing for the four tools is shown in Fig. 2 and
the dimensions are summarized in Table 3.
3.3. Friction stir welding procedures
The experimental joints were carried out by using a WMW
HECKERT vertical milling machine. A butt joint form was
used in the aluminum/aluminum joints. The direction of
the weld line is perpendicular to the rolling direction of the
plates [19,20]. The butt sides of the plates were cleaned and

Table 3 – Specifications of the FSW tools used in the first process.

Specifications Tool 1 (T1) Tool 2 (T2) Tool 3 (T3) Tool 4 (T4)

Shoulder
Diameter (mm) (Y)a ∅ 18 ∅ 16 ∅ 18 ∅ 16
Concavity angle (◦ ) 10◦

Pin
Taper angle (◦ ) 17◦ 17◦ 0Cylindrical 0Cylindrical
∅
At shoulder (A)a ∅7 (A)a ∅ 5 (X)a (X)a
At mid ∅6 ∅4 ∅6 ∅4
At end (B)a ∅5 (B)a ∅ 3
Length (mm) From shoulder surface 3.2

a
Drawing symbols.
 j m a t e r r e s t e c h n o l . 2 0 1 9;8(2):1684–1693 1687

Fig. 3 – (a) WMW ECKERT milling machine used in this work and (b) Experimental setup of FSW.

Fig. 4 – FSW configuration with: (a) position of clamps and (b) specimens location (dimensions in mm).

machined using the milling machine to produce a smooth sur-

face and to make suitable butt joints. The base metals were
clamped and supported by a steel backing plate. The plates
were fixed by two clamps. Positions and number of clamps are
selected after many experimental runs. Fig. 3(a) and (b) shows
Fig. 5 – Tensile test specimen according to ASTM B557.
the milling machine and the experimental setup, respectively.
The rotational and traverse speeds were calibrated and ver-
ified by using a digital tachometer. Before every experiment,
the values of the parameters corresponding to trial number are
adjusted. Then the rotating tool is plunged at the interface of perpendicular to the welding direction. A universal testing
the two plates, till the shoulder sinking into the metals by the machine was used to perform the tensile test. The tensile
value of the plunge depth. After that the traverse speed was specimens were cut by using a wire cut machine according
applied. Before the tool reaching the end of the work piece, it is to ASTM B557 [22,23] and the dimensions of the tensile speci-
released from the metals. Fig. 4 shows the clamping position mens are shown in Fig. 5.
and specimen location related to clamping position.
3.4.2. Metallography
3.4. The analysis of specimens Microstructure specimens were cut from a welded part in the
transverse direction of the weld. The specimens had a prepa-
3.4.1. Tensile test ration according to ASM-9 Metallography and Microstructures
Transverse tensile specimens are used to evaluate the [24–26]. The microstructure specimens are examined by using
strength of the dissimilar joints. The transverse direction is an Olympus optical microscope (OM).
1688 j m a t e r r e s t e c h n o l . 2 0 1 9;8(2):1684–1693

Table 4 – Experimental results of UTS and ductility according to L16-OA.

Control factors with interactions (L16OA) UTS (MPa) Ductility %

X1 X2 X1.X2 X3 X4 X5 X1.X5 X2.X5 X6 X7 X1.X7 X4.X6 X3.X6 R1 R2 R3 R1 R2 R3

1 1 1 1 1 1 1 1 1 1 1 1 1 1 208.12 203.92 197.97 9.6 9.1 8.75

2 1 1 1 1 1 2 2 2 2 2 2 2 2 199.72 195.28 185.25 8.4 8 6.25
3 1 1 1 2 2 1 1 1 1 2 2 2 2 205.67 185.45 194.22 9.6 7 8.4
4 1 1 1 2 2 2 2 2 2 1 1 1 1 169.64 199.93 217.13 6.5 8.5 9.25
5 1 2 2 1 1 1 1 2 2 1 1 2 2 198.87 218.68 210 9.5 11.75 9.50
6 1 2 2 1 1 2 2 1 1 2 2 1 1 206.12 208.08 202.25 8.75 9.4 9.25
7 1 2 2 2 2 1 1 2 2 2 2 1 1 209.14 224.18 212.69 8.25 11.2 8.125
8 1 2 2 2 2 2 2 1 1 1 1 2 2 182.11 182.19 181.25 6.5 6 7.125
9 2 1 2 1 2 1 2 1 2 1 2 1 2 180.48 205.31 208.61 7.125 7.75 8
10 2 1 2 1 2 2 1 2 1 2 1 2 1 205.23 230.66 199.84 9.6 9.8 8.25
11 2 1 2 2 1 1 2 1 2 2 1 2 1 204.57 198.29 197.68 9.4 7 8.50
12 2 1 2 2 1 2 1 2 1 1 2 1 2 203.51 190.10 186.88 8.2 7.5 7.50
13 2 2 1 1 2 1 2 2 1 1 2 2 1 218.88 218.31 214.60 11.2 9.5 8.25
14 2 2 1 1 2 2 1 1 2 2 1 1 2 212.24 206.53 207.96 10 8.75 8.50
15 2 2 1 2 1 1 2 2 1 2 1 1 2 199.64 207.22 200.41 9.6 8.25 8.25
16 2 2 1 2 1 2 1 1 2 1 2 2 1 214.50 208 212.0 8.25 8.00 8.00

affected by any other parameters or interaction variables

4. Results and discussions out of these screen parameters in mean and S/N ratio of
4.1. Experimental results analysis of L16-OA UTS under this process condition [30,31]. With these results
obtained from ANOVA tables about the process, it can build
Table 4 illustrates the results of the experiments for UTS and strong mathematical models [32–35].
ductility according to L16-OA. The results of S/N ratio and
mean for UTS and ductility are given in Table 5. ANOVA for L16-OA results of mean and S/N ratio shows
some results:

4.1.1. Analysis of variance for L16-OA

Table 6 gives a conclusion of the effect of the process
parameters, interactions, and its contribution on UTS and
ductility[27–29]. It can evaluate the efficiency of using L16-OA
to study seven parameters at two levels by evaluating the
degree of statistical errors obtained from ANOVA tables. For
UTS, the errors are 0.782% and 1.347% for S/N ratio and mean,
respectively. For ductility, the errors are 11.7% and 12.9% for
S/N ratio and mean, respectively. With these errors, the use
of L16-OA to study a process have seven parameters with
two levels is very efficient. It means that the process is not
A – For ultimate tensile strength
A.1 – Based on S/N ratio
• Rotational speed, traverse speed, D/d ratio and plunge depth
are significant parameters at 95% confidence level (C.L.). Pin
profile, location of base metal and tilt angle are insignificant
parameters at any confidence levels [36, 37].
• X1.X5 and X3.X6 are significant interactions at 95% con-
fidence level. Only X2.X5 is significant interaction at 90%

Table 5 – S/N ratio and mean values of UTS and ductility using L16-OA.
# Control factors with interactions (L16OA) UTS (MPa) Ductility %

X1 X2 X1.X2 X3 X4 X5 X1.X5 X2.X5 X6 X7 X1.X7 X4.X6 X3.X6 Mean S/N ratio Mean S/N ratio

1 1 1 1 1 1 1 1 1 1 1 1 1 1 203.337 46.1588 9.15 19.2097

2 1 1 1 1 1 2 2 2 2 2 2 2 2 193.417 45.7169 7.55 17.3363
3 1 1 1 2 2 1 1 1 1 2 2 2 2 195.113 45.7825 8.33 18.1977
4 1 1 1 2 2 2 2 2 2 1 1 1 1 195.567 45.6876 8.083 17.8518
5 1 2 2 1 1 1 1 2 2 1 1 2 2 209.183 46.3908 10.250 20.0872
6 1 2 2 1 1 2 2 1 1 2 2 1 1 205.483 46.2537 9.133 19.2002
7 1 2 2 2 2 1 1 2 2 2 2 1 1 215.337 46.6510 9.1917 18.9949
8 1 2 2 2 2 2 2 1 1 1 1 2 2 181.850 45.1942 6.541 16.2498
9 2 1 2 1 2 1 2 1 2 1 2 1 2 198.133 45.8836 7.6250 17.6135
10 2 1 2 1 2 2 1 2 1 2 1 2 1 211.910 46.4732 9.2167 19.2136
11 2 1 2 2 1 1 2 1 2 2 1 2 1 200.180 46.0253 8.300 18.1848
12 2 1 2 2 1 2 1 2 1 1 2 1 2 193.497 45.7159 7.7333 17.7445
13 2 2 1 1 2 1 2 2 1 1 2 2 1 217.263 46.7387 9.6500 19.4892
14 2 2 1 1 2 2 1 1 2 2 1 1 2 208.910 46.3974 9.083 19.1007
15 2 2 1 2 1 1 2 2 1 2 1 1 2 202.423 46.1216 8.700 18.7251
16 2 2 1 2 1 2 1 1 2 1 2 2 1 211.500 46.5041 8.0833 18.1491
 j m a t e r r e s t e c h n o l . 2 0 1 9;8(2):1684–1693 1689

Table 6 – Summary of the effect of the parameters and interactions on UTS and ductility.
UTS (MPa) Ductility

S/N ratio Mean S/N ratio Mean

Percentage of Significant at Percentage of Significant at Percentage of Significant at Percentage of Significant at

contribution % level % contribution % level % contribution % level % contribution % level %

X1 9.658 95 9.084 90 0.527 I 0.12 I

X2 18.577 95 16.934 95 9.534 90 10.308 95
X3 12.702 95 12.470 90 22.402 99 21.425 99
X4 0.014 I 0.11 I 0.164 I 0.660 I
X5 7.714 95 6.910 90 14.140 95 15.957 95
X6 1.577 I 2.088 I 0.223 I 0.040 I
X7 3.104 I 2.307 I 2.894 I 2.736 I
X1.X2 0.636 I 0.655 I 0.263 I 0.196 I
X1.X5 17.169 95 13.594 95 16.161 95 14.255 95
X1.X7 1.499 I 1.230 I 1.591 I 1.962 I
X2.X5 3.958 90 5.324 I 5.530 I 8.142 90
X3.X6 25.499 95 27.911 95 12.130 95 11.919 95
X4.X6 0.0045 I 0.023 I 1.038 I 0.287 I
Error 0.782 1.347 11.7 12.09

I, insignificant at any confidence levels (90%, 95%, 99%). X1, rotational speed (rpm); X2, traverse speed (mm/min); X3, D/d ratio; X4, pin profile
(based on taper angle); X5, plunge depth (mm); X6, tilt angle; X7, location of base metal (based on tool rotation direction).

confidence level. Interactions: X1.X2, X1.X7, and X4.X6 are
statistically insignificant at any confidence levels [38]
• The traverse speed and D/d ratio and the interactions of
X1.X5 and X3.X6 have the most contribution to the process
and totally contribute about 73.947% for S/N ratio to overall
contributions. The traverse speed has the highest contribu-
tion to the process about 18.577% as in references [21] 33%
and [4] 35% but the traverse speed in some previous pro-
cesses of joining aluminum alloys has the second highest
contribution to the process as in references [22] 28.3%, [8]
13.40%, [10] 29.8%, and [11] 32.24%. The D/d ratio has the
second highest contribution to the process about 12.702%.
Some of the previous processes of joining aluminum alloys
show that the rotational speed has the highest contri-
bution to the process instead of traverse speed and D/d
ratio as 67% [22], 68.13% [8], 53.55% [10] and 41.25% [11].
The change in contribution percentage of parameters to
the process depends on the levels of its parameters, the
other parameters applied in this process and the process
condition [23].
• It can ignore the effect of the insignificant parameters and
the insignificant interactions when performing the process
or develop a mathematical model for the response, espe-
cially if the parameters have a very low contribution to the
response. It can neglect the effects of pin profile, location of
base metal, tilt angle and the interactions of X1.X2, X1.X7
and X4.X6 which totally have a contribution of 6.8% to the
system output[41].
• The optimum levels of the process parameters which can
be used to obtain the optimum value are rotational speed
at level 2 (1225 rpm), traverse speed at level 2 (21 mm/min),
D/d ratio at level 1 (3), pin profile at level 1 (based on the taper
angle 0◦ (cylinder)), plunge depth at level 1 (0.1 mm), tilt
angle at level 2 (2◦ ) and location of base metal (lower metal,
based on the tool rotation direction) at level 2 (advancing
side 360◦ ).
A.2 – Based on mean (ANOM)
• Only traverse speed is significant parameters at 95% con-
fidence level. Rotational speed, D/d ratio and plunge depth
are significant parameters at 90% C.L. Pin profile, tilt angle
and location of base metal are insignificant parameters at
any C.L.s.
• Interactions: X1.X5 and X3.X6 are statistically significant at
95% C.L. Interactions: X1.X2, X1.X7, X2.X5 and X4.X6 are
statistically insignificant at any C.L.s.
• The traverse speed and D/d ratio and the interactions of
X1.X5 and X3.X6 have the most contribution to the pro-
cess and totally contribute about 71% for mean to overall
contributions. The traverse speed has the highest contri-
bution to the process about 16.943% as in Ref. [21] 34% but
the traverse speed in some previous processes of joining
aluminum alloy has the second highest contribution to the
process as in Refs. [22] 28.3%, [8] 13.70%, [10] 30.17%, and [11]
33.24%. The D/d ratio has the second highest contribution to
the process about 12.470%. Some of the previous processes
of joining aluminum alloys show that the rotational speed
has the highest contribution to the process instead of tra-
verse speed and D/d ratio as 67% [22], 68.13% [8], 52.60% [10]
and 41.30% [11].
• It can neglect the effect of the insignificant parameters
(pin profile, tilt angle and location of base metal) and the
insignificant interactions (X1.X2, X1.X7, X2.X5 and X4.X6)
which totally have a contribution of 11.7% on the response
when develops the model.
• The optimum levels which can be used to obtain the highest
UTS within the process condition are rotational speed at
level 2 (1225 rpm), traverse speed at level 2 (21 mm/min), D/d
ratio at level 1 (3), pin profile at level 2 (based on the taper
angle (17◦ )), plunge depth at level 1 (0.1 mm), tilt angle at
level 2 (2◦ ), location of base metal (lower metal, based on
the tool rotation direction) at level 2 (advancing side 360◦ ).
1690 j m a t e r r e s t e c h n o l . 2 0 1 9;8(2):1684–1693
• D/d ratio, plunge depth and traverse speed are significant
parameters at 99%, 95%, 90% confidence level, respectively.
Rotational speed, pin profile, tilt angle and the location of
base metal are insignificant parameters at any confidence
levels.
• Interactions: X1.X5 and X3.X6 are statistically significant at
95% C.L. Interactions: X1.X2, X1.X7, X2.X5 and X4.X6 are
statistically insignificant at any C.L.s.
• The traverse speed, D/d ratio, plunge depth and the statisti-
cal interactions between rotational speed and plunge depth
and between D/d ratio and tilt angle have the most contri-
bution to the process and totally contribute about 74.63%
for S/N ratio to overall contributions. The D/d ratio has the
highest contribution to the process about 22.402%.
• It can neglect the effect of the insignificant and the insignif-
icant interactions which totally have a contribution of
13.67% on the response when develops the model.
• The optimum levels of the process parameter which can
be used to achieve the optimum value are rotational level
at level 2 (1225 rpm), traverse speed at level 2 (21 mm/min),
D/d ratio at level 1 (3), pin profile at level 1 (based on the
taper angle 0◦ ), plunge depth at level 1 (0.1 mm), tilt angle
at level 1 (1.5◦ ), and the location of base metal (lower metal,
based on the tool rotation direction) at level 2 (advancing
side 360◦ ).
B.2 – Based on mean (ANOM)
• D/d ratio, plunge depth and traverse speed are significant
parameters at 99%, 95%, 95% confidence level, respectively.
Rotational speed, pin profile, tilt angle and the location of
base metal are insignificant parameters at any confidence
levels.
• Interactions: X1.X5, X2.X5 and X3.X6 are statistically signif-
icant at 95%, 90%, 95% C.L. Interactions: X1.X2, X1.X7 and
X4.X6 are statistically insignificant at any C.L.s.
• The traverse speed, D/d ratio, plunge depth and the statisti-
cal interactions between welding speeds and plunge depth
and between D/d ratio and tilt angle have the most contri-
bution to the process and totally contribute about 82.006%
for S/N ratio to overall contributions. The D/d ratio has the
highest contribution to the process about 21.425%.
• It can neglect the effect of the insignificant and the insignif-
icant interactions which totally have a contribution of 5.90%
on the response when develops the model.
• The optimum levels which can be used to obtain the
maximum ductility within the process condition are rota-
tional level at level 2 (1225 rpm), traverse speed at level
2 (21 mm/min), D/d ratio at level 1 (3), pin profile at level
1 (based on the taper angle 0◦ ), plunge depth at level 1
(0.1 mm), tilt angle at level 1 (1.5◦ ), and the location of base
metal (lower metal, based on the tool rotation direction) at
level 2 (advancing side 360◦ ) [43, 44].
4.2. Verification of experimental results
Once the optimal level of control parameters is selected,
the final step is to verify the improvement of quality
characteristics using the optimal level. The estimated optimal
level of parameters is calculated as:

Ypredicted = Ymean + [Yi − Ymean ] (2)
Yi is the mean (S/N ratio or mean) response at optimal level
of control parameters; Ymean is the total mean (S/N ratio or
mean) response.
The optimum results (or experiment) for UTS and ductil-
ity are not found in any experiment according to the selected
array. A comparison between the predicted values and the
experimental values of the UTS and ductility at the opti-
mum levels is given in Table 7. The deviation between the
experimental and predicted results for the optimum UTS
and ductility are very close according to the results of L16-
OA.
4.3. Metallographic analysis
The microstructure of the optimum joint which has the high-
est UTS within the process condition and the joint 8 (the
weakest joint obtained according to the L16-OA) are discussed
in this part. The joint efficiency of the optimum joint and
the weakest joint is 85.3% (221.3 MPa) and 69.9% (181.8 MPa),
respectively. The joint efficiency is compared to the strength
of the softer metal used in this dissimilar joint. The UTS of
the AA5454 and AA7075 is 260 MPa and 402 MPa, respectively.
The optimum joint was obtained by using the parameters
of 1225 rpm rotational speed, 21 mm/min traverse speed, 3
D/d ratio, tapered pin profile, 0.1 mm plunge depth, 2◦ tilt
angle, and AA5454 on advancing side. The weakest joint
was obtained by using the parameters of 1000 rpm rotational
speed, 21 mm/min traverse speed, 4 D/d ratio, tapered pin pro-
file, 0.25 mm plunge depth, 1.5◦ tilt angle, and AA5454 on the
retreating side. The amount of the generated heat in FSW
usually depends on two parameters: diameter of the shoul-
der and the combination of the welding speeds (RS + TS). It is
clear that the frictional heat generated during the welding of
the optimum joint is higher than the generated heat in the
joint 8. The microstructure of the base metal of the AA5454
and the AA7075 is shown in Fig. 6. The strain hardening effect
is observed in the microstructure of the AA5454 (5XXX strain
hardening alloy). This alloy is rolled twice from 14 mm to 7 mm
and from 7 mm to 3.5 mm. The grains of the AA7075 are very
large and have an elongated pancake shape due to the hot
rolling.
Fig. 7 shows the microstructure at the interface between
AA5454 and AA7075 of the optimum joint and the joint 8,
respectively. In the joint 8, there is little bonding without stir-
ring between AA5454 and AA7075 in all the interface regions
from the upper weld region to the bottom. At the inter-
face on the optimum joint, The microstructure of all the
regions from the upper (Fig. 7a), middle (Fig. 7b and c), and
to the bottom (Fig. 7d) shows good stirring and more con-
solidate between AA5454 and AA7075 which improves the
quality of the weld. In the AA7075 side. The grains in the
NZ become finer and equiaxed in both the joints. The size
of the grains decreases from the base metal through to the
NZ.
 j m a t e r r e s t e c h n o l . 2 0 1 9;8(2):1684–1693 1691

Table 7 – The optimum results of UTS and ductility based on mean and S/N ratio.
Optimal levels Experiment Prediction Deviation %

UTS (MPa) X12 , X22 , X31 , X42 , X51 , X62 , X72 222.5, 223.5, 217.1 = 221.03 217.263 1.75
Ductility % X12 , X22 , X31 , X41 , X51 , X61 , X72 9.25, 11.2, 12 = 10.833% 9.855 8.8

Xnm : n, parameter number, m: level number. X1, rotational speed (rpm); X2, traverse speed (mm/min); X3, D/d ratio; X4, pin profile (based on
taper angle); X5, plunge depth (mm); X6, tilt angle; X7 location of base metal (based on tool rotation direction).

Fig. 6 – Optical micrograph of base metals (a) AA7075; (b) AA5454.

Fig. 7 – Optical micrographs of the optimum joint (AA5454 to AA7075) at the interface in different regions: (a) upper; (b, c)
middle and (d) bottom.

4.4. Tensile properties 1550 rpm 1225 rpm 1000 rpm Al-Si base metal
Ultimate tensile strength

120
The results of the UTS of the 26 joints with a relation with 110
the welding speeds are given in Figs. 8 and 9 as shown in 100
(Mpa)

Figs. 8 and 9 the UTS results of the joints made by using the tool 90
(T1: D/d ratio = 3) and the tool (T2: D/d ratio = 4), respectively. All 80
70
the results of the UTS show a nonlinear effect under different
60
combinations of the welding speeds except the 1225 rpm rota- 50
tional speed with both the tools (T1, T2). At 1225 rpm rotational 15 25 35 45 55 65
Traverse speed (mm/min)
speed, The UTS decreased with increasing the traverse speed
from 17 mm/min to 67.5 mm/min with both the tools (T1, T2) Fig. 8 – The UTS results of the joints made in the first stage
and the best results were obtained between 17–21 mm/min of joining dissimilar aluminum alloys with tool (T1).
traverse speed.
1692 j m a t e r r e s t e c h n o l . 2 0 1 9;8(2):1684–1693

1550 rpm 1225 rpm 1000 rpm Al-Si base metal ductility were considered as the mechanical properties of the
Ultimate tensile strength (Mpa)

125 dissimilar joint. The results can be summarized as follows:

120

115 • For UTS: based on S/N ratio and mean, Rotational speed;
110
traverse speed; D/d ratio and plunge depth are significant
parameters at different confidence levels. Pin profile; loca-
105
tion of base metal and tilt angle are insignificant parameters
100
at any confidence levels. Based on standard deviation, tra-
95
verse speed and pin profile are significant parameters at
90
15 20 25 30 35 40 45 50 55 60 65 70
99% and 90% C.L., respectively. Rotational speed; D/d ratio;
plunge depth; tilt angle and location of base metal are not
Traverse speed (mm/min)
significant parameters at any C.L.s.
Fig. 9 – The UTS results of the joints made in the first stage • The optimal levels determined from the L16-OA to obtain
of joining dissimilar aluminum alloys with tool (T2). the highest UTS within process condition are 0.1 mm
plunge depth, 1225 rpm rotational speed, tapered pin pro-
file, 21 mm/min traverse speed, 2◦ tilt angle, 3 D/d ratio, and
AA5454 on the advancing side.
With the non-linear effect of the rotational speeds
• Based on the results of S/N ratio and mean of ductility:
(1000–1550 rpm) at the traverse speeds from 17 mm/min to
D/d ratio, plunge depth and traverse speed are signifi-
67.5 mm/min, the 42 mm/min is the transition zone of the
cant parameters at 99%, 95%, and 90% confidence level,
effect of the process parameters on the UTS at 1000 rpm and
respectively. Rotational speed, pin profile, tilt angle and the
1550 rpm with the tool (T1) and at 1550 rpm with the tool (T2)
location of base metal are insignificant parameters at any
but the 21 mm/min – traverse speed is the transition zone of
confidence levels.
the effect of the process parameter at 1000 rpm and the tool
• The optimal levels determined from the L16-OA to achieve
(T2). The best conditions lead to achieving the highest UTSs
the highest ductility within process condition are 0.1 mm
in this area are 17–21 mm/min traverse speeds with 1000 rpm
plunge depth, 1225 rpm rotational speed, cylindrical pin
rotational speed and the tools (T1, T2), 42–67.5 mm/min tra-
profile, 21 mm/min traverse speed, 1.5◦ tilt angle, 3 D/d ratio,
verse speeds with 1550 rpm rotational speed and tool (T2). To
and AA5454 on the advancing side.
optimize the process parameters, it must select at least two
• Six mathematical models were developed for UTS and duc-
levels of these parameters [37,42]. The joints made by using the
tility as a function of the significant parameters and the
tool having the 4 D/d ratio have higher UTS than those made
significant interactions. For UTS, the accuracy of the mod-
by using the tool having the 3 D/d ratio. The UTSs with the
els is 97.678%, 99.56%, and 58.670% for mean, S/N ratio, and
tool (T2) are approximately equal to the strength of the base
standard deviation, respectively. For ductility, the accuracy
metal of the Al–Si alloy at six parameter conditions (joints:
of the models is 93.56%, 97.320%, and 50% for mean, S/N
17, 19, 20, 23, 24, 25). Only one parameter condition with the
ratio, and standard deviation, respectively.
tool (T1) gives the same tensile strength of the Al–Si alloy
• The maximum joint efficiency obtained, compared to the
(the lower metal in this joint). The maximum UTS obtained
strength of the softer metal used in the dissimilar joint, is
is 125.94 MPa at the process parameters of 1225 rpm rota-
85.3%.
tional speed, 17 mm/min traverse speed, and the tool (T2). The
strength of the joints decreases with increasing the rotational
speed and with increasing the tool shoulder diameter. The
Conflict of interest
drop in the UTS compared to the strength of the lower metal
used in this joint depends on the efficiency of the combination
The authors declare that they have no conflict of interest.
of the process parameters related to the frictional heat gen-
erated. The joints 19, 23 and 24 have a higher UTS than the
UTS of Al–Si alloy about 3–5 MPa but it is normal because the references
UTS of all joints was only tested one time (one test specimen)
and then compared to the average UTS of Al–Si alloy (122, 118,
125.68 MPa). It is clear that the highest UTS obtained in this [1] Mishra RS, Ma ZY. Friction stir welding and processing. Mater
stage is approximately equal to the UTS of Al–Si alloy. Sci Eng R: Rep 2005;50(1–2):1–78.
[2] De Backer J, Ph.D. thesis Feedback control of robotic friction
stir welding. Belgium: University College KaHo Sint-Lieven in
5. Conclusions Gent; 2014.
[3] Essa GM, Zakria HM, Mahmoud TS, Khalifa TA.
Modeling and optimization of FSW parameters of AA5454 to Microstructure examination and microhardness of friction
stir welded joint of (AA7020-O) after PWHT. HBRC J
AA7075 joints are studied by using Taguchi methods. Taguchi
2018;14(1):22–8.
L16-OA was used to optimize the process parameters. Seven [4] Pradeep A. A review on friction stir welding of steel. Int J Eng
parameters at two levels are selected. These parameters are Res Dev 2012;3(11):75–91.
rotational speed, traverse speed, D/d ratio, pin profile (based on [5] Bahari MS, Jaffarullah MS, Mohamed Z. Heat analysis in
the taper angle), tool tilt angle, plunge depth, and the location friction stir welding using finite element method. Int J Mech
of base metal (based on the tool rotation direction). UTS and Eng 2018;5(4):174–88. ISSN: 1823-5514, eISSN: 2550-164X.
 j m a t e r r e s t e c h n o l . 2 0 1 9;8(2):1684–1693
[6] Singh K, Singh G, Singh H. Review on friction stir welding of
magnesium alloys. J Magnesium Alloys 2018;23(2):160–89.
[7] Selvam SK, Parameshwaran Pillai T. Analysis of heavy alloy
tool in friction stir welding. Int J Chem Tech Res
2013;5(3):1346–58.
[8] Thomas W, Lockyer SA, Kalee SW, Staines DG. Friction stir
welding – an update on recent developments. In: IMechE
Stressed Components in Aluminium Alloys. 2003.
[9] Rai1 R, De A, Bhadeshia HKDH, DebRoy T. Review: friction
stir welding tools. Sci Technol Weld Join 2011;16(4):325–42.
[10] Meilinger Á, Török I. The importance of friction stir welding
tool. Prod Process Syst 2013;6(1.):25–34.
[11] Rouhi S, Ashjari M, Mostafapour A. Experimental study on
the effects of friction stir welding parameters on the quality
and the mechanical properties of the AZ91 joints. Int J Eng
Dev Res 2015;3:234–9.
[12] Leonard AJ, Locker SA. Flaws in friction stir welds. In: Fourth
international symposium on friction stir welding. 2003. p.
14–6.
[13] Sahu PK, Pal S. Effect of FSW parameters on microstructure
and mechanical properties of AM20 welds. Mater Manuf
Process 2018;33:288–98.
[14] Montgomery DC. Introduction to statistical quality control.
6th ed; 2009. ISBN 978-0-470-16992-6.
[15] Taguchi G, Chowdhury S, Wu Y. Taguchi’s quality
engineering handbook; 2005. ISBN: 978-0-471-41334-9.
[16] Park SH. Robust design and analysis for quality engineering.
2nd ed. London: Chapman Hall; 1996.
[17] Ross PJ. Taguchi technique for quality engineering. 2nd ed.
McGraw-Hill, Donnelley& Sons Company; 1998. p. 1–327.
ISBN: 0-07-053958-8.
[18] Roy RK. Design of experiments using the Taguchi approach.
John Wiley & Sons, Inc.; 2011. p. 1–157. ISBN 0-471-36101-1.
[19] Camppbell FC. Lightweight materials; understanding the
basics. Material Park, OH: ASM International; 2012. p. 36–53.
[20] The Aluminum Association, Inc. Aluminum alloys: selection
and application. Washington, DC: The Aluminum
Association, Inc.; 1998.
[21] Mironov S, Onuma T, Sato YS, Yoneyama S, Kokawa H.
Tensile behavior of friction-stir welded AZ31 magnesium
alloy. Mater Sci Eng A 2017;679:272–81.
[22] Ram D, Mahendra G. Optimization of FSW process parameter
to achieve maximum tensile strength of aluminum alloy
AA6061. IRJET 2016;3(2):936–43.
[23] Ram, Kumar J. Parametric optimization for friction stir
welding of AL 6063 alloy using Taguchi technique. IJSTE
2016;6(4):604–10.
[24] Girish G, Deepandurai K. Taguchi optimization of FSW
parameters of 8011 aluminium alloy. IJSTE 2016;2(10):872–9.
[25] Srinivasa Rao T, Madhusudhan Reddy G, Koteswara Rao SR.
Microstructure and mechanical properties of friction stir
welded AA7075-T651 aluminum alloy thick plates. Trans
Nonferrous Met Soc China 2015;25:1770–8.
[26] Ramgopal V, Abdullah B, Mohammed A, Yaswanth Y.
Structural properties of similar and dissimilar aluminum
alloy joints by FSW. In: Proceedings of the ASME 2014
International Mechanical Engineering Congress and
Exposition. 2014.
[27] Sadeesh P, Venkatesh Kannan M, Rajkumar V, Avinash P,
Arivazhagan N, Devendranath Ramkumar K, Narayanan.
1693
Studies on friction stir welding of AA 2024 and AA 6061
dissimilar metals. Procedia Eng 2014;75:145–9.
[28] Dehghani K, Ghorbani R, Soltanipoor AR. Microstructural
evolution and mechanical properties during the friction stir
welding of 7075-O aluminum alloy. Int J Adv Manuf Technol
2014, http://dx.doi.org/10.1007/s00170-014-6574-0.
[29] Gupta SK, Pandey KN. Application of Taguchi method for
optimization of friction stir welding process parameters to
joining of Al alloy. Adv Mater Manuf Character
2013;3(1):253–8.
[30] Vagh AS, Pandya SN. Influence of process parameters on the
mechanical properties of friction stir welded AA 2014-T6
alloy using Taguchi orthogonal array. IJESET 2012;2(1):
51–8.
[31] Murali Krishna P, Ramanaiah N, Prasada Rao K. Optimization
of process parameters for friction Stir welding of dissimilar
aluminum alloys (AA2024-T6 and AA6351-T6) by using
Taguchi method. Int J Ind Eng Comput 2013;4:
71–80.
[32] Kumbhar NT, Bhanumurthy K. Friction stir welding of Al
5052 with Al 6061 alloys. Hindawi – J Metall 2012.
[33] Jayaraman M, Sivasubramanian R, Balasubramanian V,
Lakshminarayanan AK. Optimization of process parameters
for friction stir welding of cast aluminum alloy A319 by
Taguchi method. J Sci Ind Res 2009;68:36–43.
[34] Mostafa AM. Friction stir welding of 6061 aluminum alloys,
MSc Thesis. Giza, Egypt: Faculty of Engineering, Cairo
University; 2009.
[35] Lakshminarayanan AK, Bala Subramanian V. Process
parameters optimization for friction stir welding of RDE-40
aluminium alloy using Taguchi technique. Trans Nonferrous
Met Soc China 2008;18:548–54.
[36] Peel MJ, Steuwer A, Withers PJ, Dickerson T, Shi Q, Shercliff
H. Dissimilar friction stir welds in AA5083–AA6082. Part I:
Process Parameter effects on thermal history and weld
properties. Metall Mater Trans A 2006;37A:2183–93.
[37] Wang W, Deng D, Mao Z, Tong Y, Ran Y. Influence of tool
rotation rates on temperature profiles and mechanical
properties of friction stir welded AZ31 magnesium alloy. Int J
Adv Manuf Technol 2017;88(5–8):2191–200.
[38] Standard methods of tension testing wrought and cast
aluminum- and magnesium-alloy products, ASTM Standard
B557.
[41] Ditzel PJ, MSc Thesis Microstructure/property relationships
in aluminum friction stir welds. USA: The Ohio State
University; 1997.
[42] Palanivel R, Koshy Mathews P, Dinaharan I, Murugan N.
Mechanical and metallurgical properties of dissimilar
friction stir welded AA5083-H111 and AA6351-T6 aluminum
alloys. Trans Nonferrous Met Soc 2014;24:58–65.
[43] Abd Elnabi M, Osman T, El Mokadem A, Elshalakany A.
Mechanical Properties and Microstructure of Dissimilar
Friction Stir Welding of Pure Aluminum to Low Carbon Steel.
Adv. J. Grad. Res 2018;4(1):47–58.
[44] Elshalakany A, Shady Ali TA, Osman H, Megaid A, El
Mokadem. An experimental investigation of the formability
of low carbon steel tailor-welded blanks of different
thickness ratios. Int J Adv Manuf Technol 2017;88:
1459.