Effect of process parameter on mechanical properties of friction stir welded aluminium alloy joints using factorial design

Gill Ankur, Chahal Gurpreet Singh, International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 771
ISSN: 2454-132X
Impact factor: 4.295
(Volume 3, Issue 6)
Available online at www.ijariit.com
Effect of Process Parameter on Mechanical Properties of Friction
Stir Welded Aluminium Alloy Joints Using Factorial Design
Ankur Gill
Assistant Professor
Department of Mechanical Engineering,
Swami Vivekananda Institute of Engineering & Technology,
Chandigarh
ankurgill6@gmail.com
Gurpreet Singh Chahal
Assistant Municipal Engineer
MCD, Malerkotla
preetchahal26@gmail.com
Abstract: In this paper effect of process parameter on mechanical properties of friction stir welded aluminum alloy joints using
factorial design is calculated. In experimentation, Two Level of Factorial Design is found to be a more effective tool to
investigate the interaction effects of parameters on the required response. Various models have been proposed in this context
out of that 95% are at an adequate confidence level. In an analysis of stir welding, tensile strength is found to get an increase
with a decrease in rotational speed of tool while decreases with increase in the welding speed.At the low rotational speed of tool
high welding speed, there is a small increasing response to Tensile Strength. But at high rpm, and high welding speed, there is
a significant increase in Tensile Strength. At a low value of welding speed and high rotational speed, there is a sharp decrease
in Tensile Strength. At high welding speed, an increase in rpm, there is a little decrease in Tensile Strength. Prediction of the
Tensile Strength at any combination of the two parameters welding speed and rotational speed can be done. With increase
shoulder diameter, there is a decrease in Impact Strength. Impact Strength increases with the increase in tool rotational speed.
Keywords: Friction-stir Welding, Polynomials, Milling, Tensile Strength, Rotational Speed, Hardness.
I. INTRODUCTION
Friction-stir welding (FSW) is a solid-state joining process (the metal is not melted) that uses a third body tool to join two facing
surfaces. Heat is generated between the tool and material which leads to a very soft region near the FSW tool. It then mechanically
intermixes the two pieces of metal at the place of the join, then the softened metal (due to the elevated temperature) can be joined
using mechanical pressure (which is applied by the tool), much like joining clay, or dough. It is primarily used on aluminium, and
most often on extruded aluminum (non-heat treatable alloys), and on structures which need superior weld strength without a post
weld heat treatment. A constantly rotated non consumable cylindrical-shouldered tool with a profiled probe is transversely fed at a
constant rate into a butt joint between two clamped pieces of butted material. The probe is slightly shorter than the weld depth
required, with the tool shoulder riding atop the work surface. Frictional heat is generated between the wear-resistant welding
components and the work pieces. This heat, along with that generated by the mechanical mixing process and the adiabatic heat
within the material, causes the stirred materials to soften without melting. As the pin is moved forward, a special profile on its
leading face forces plasticised material to the rear where clamping force assists in a forged consolidation of the weld. This process
of the tool traversing along the weld line in a plasticised tubular shaft of metal results in severe solid state deformation involving
dynamic recrystallization of the base material. The solid-state nature of the FSW process, combined with its unusual tool and
asymmetric nature, results in a highly characteristic microstructure. There are two tool speeds to be considered in friction-stir
welding; how fast the tool rotates and how quickly it traverses the interface. These two parameters have considerable importance
and must be chosen with care to ensure a successful and efficient welding cycle. The relationship between the welding speeds and
the heat input during welding is complex but, in general, it can be said that increasing the rotation speed or decreasing the traverse
speed will result in a hotter weld.

Fig: Schematic Diagram of FSW Process
II. METHODOLOGY
Following flow chart will be followed for the completion of research problem:
III. DESIGN OF EXPERIMENT
The design of the experiment is a procedure of selecting the number of trials and conditions running them, essential and sufficient
for solving a problem that has been set with the required precision. The use of design of experiment makes the behaviour of
investigator purposeful, organized and appreciably facilitates an increase in productivity of his work and reliability of results
obtained.
A two level factorial design of (23
= 8) eight trials, which is a standard statistical tool to investigate the effects of a
number of parameters on the required response, was selected for determining the effect of three independent direct welding
parameters. The commonly employed method of varying one parameter at a time, though popular, does not give any information
about interaction among parameters. The selecting of two level factorial design also helped in reducing experimental runs to the
minimum possible.

The design of the experiment is a procedure of selecting the number of trials and conditions running them, essential and sufficient
for solving a problem that has been set with the required precision. The use of design of experiment makes the behavior of
investigator purposeful, organized and appreciably facilitates an increase in productivity of his work and reliability of results
obtained.
A two level factorial design of (23
= 8) eight trials, which is a standard statistical tool to investigate the effects of a
number of parameters on the required response, was selected for determining the effect of three independent direct welding
parameters. Tool rotation speed, welding speed, and shoulder diameter were identified as critical welding variables for carrying
out the experiment. The working range covering the lowest and the highest level of the direct welding parameters were carefully
selected by carrying out the trial runs so as to maintain defect free friction stir welding. The direct and indirect parameters except
under consideration were kept constant. The upper level was coded as (+1) and lower level as (-1) or simply (+) and (-) for the
factors with a continuous determination region, this can always be done with the aid of transformation with in the variation
interval.
Xj = (Xjn - Xjo )/ Jj.........................................................(1)
Where Xj, Xjn, and Xjo are the coded, natural and basic value of the parameter respectively. Jj and j are the variation and number of
parameters respectively.
Table: Showing the Units, Symbols used and Limits of Welding Parameters
Parameters Units Symbols Lower limit Upper limit
Tool Rotational Speed RPM N 1200 1500
Welding Speed Mm/min S 40 60
Shoulder Diameter Mm D 18 21
The design matrix developed to conduct the eight trials runs of 23
fractional factorial designs as given in Table. The signs under
the columns 1, 2, 3 were arranged in standard Yate’s order.
Table: Showing Design Matrix
Sl. No N
1
S
2
D
3
1 + + +
2 - + +
3 + - +
4 - - +
5 + + -
6 - + -
7 + - -
8 - - -
IV. EXPERIMENTATION
Vertical Milling Machine was used to carry out the experiments. The material of the tool was High Carbon High Chrome steel.
The metal plates taken for investigation was aluminium having dimensions of 150 x 100 x 6mm. Two friction stir welding tools
with shoulder diameters (D) 18 mm and 21 mm, 5.8 mm long (h) and pin diameter were 6 mm were used. The welding was
carried out by using a properly designed clamping fixture that allows fixing the two plates. The iron plates composing the fixture
were finished at the grinding machine in order to assure a uniform pressure distribution on the two fixed specimens. The complete
set of eight trials was repeated twice for the sake of determining the ‘variance of optimization parameter’ and once for ‘variance
of adequacy’ for this model. The experiments were performed in a random order in order to avoid any systematic error. The
responses for the three set of experiments are given in Table. The Design of experiment software Design Expert was used to
develop to Design of Experiment.

Table: Showing Experimental Results
S. No.
Run
No.
Tool
Rotational
speed
Welding
speed
mm/mim
Shoulder
diameter
(mm)
Tensile
strength
(MPa)
Micro
hardness Hv
Impact
toughness (J)
1 5 -1 -1 -1 209 70 17
2 14 -1 -1 -1 210 71 17.5
3 20 -1 -1 -1 209 70 17
4 19 1 -1 -1 217 78 20
5 22 1 -1 -1 217 79 21
6 4 1 -1 -1 215 78 21
7 10 -1 1 -1 190 60 15
8 21 -1 1 -1 192 61 15
9 2 -1 1 -1 195 62 16
10 18 1 1 -1 215 75 17
11 6 1 1 -1 212 74 18
12 8 1 1 -1 216 75 18
13 15 -1 -1 1 212 70 14
14 7 -1 -1 1 214 69 15
15 1 -1 -1 1 213 69 14
16 16 1 -1 1 220 80 18
17 24 1 -1 1 221 79 18
18 12 1 -1 1 219 80 19
19 9 -1 1 1 186 60 13
20 23 -1 1 1 183 59 14
21 11 -1 1 1 188 61 13
22 3 1 1 1 215 68 15
23 13 1 1 1 217 68 15.5
24 17 1 1 1 217 69 15
The models of the type Y = f (N, S, D) could be developed to facilitate the prediction of a response within the specified
dimensional tolerance for a particular set of direct process parameters. Assuming a linear relationship in the first instance and
taking into account all the possible two factor interaction and confounded interactions, it could be written as:
Y=b0+b1N+b2S+b3D+ b12NS+b13ND+b23SD…………………... (2)
Models were developed by the method of regression. Adequacy of the model and significance was tested by the analysis of
variance technique and student’s‘t’-test respectively.
The regression coefficients of the selected model were calculated using Equation -3. This is based on the method of least squares.
bj = ∑NiXjiYi/N, j = 0,1,…..,k ………….(3)
Where,
Xji = value of a factor or interaction in coded form
Yi =Average value of response parameter
N= No.of observations
k= Number of coefficients of the model

Table: Coefficients of Model
Sl. No Coefficient of Regression Due to
1 b0 Combined Effect of all Parameters
2 b1 Tool Rotational Speed
3 b2 Welding Speed
4 b3 Shoulder Diameter
5 b12 Interaction of N & S
6 b13 Interaction of N & D
7 b23 Interaction of S & D
Table: Design Matrix for Calculating the Value of Coefficients
b0 b1
N
b2
S
b3
D
b12
NS
b13
ND
b23
SD
+1 +1 +1 +1 +1 +1 +1
+1 -1 +1 +1 -1 -1 +1
+1 +1 -1 +1 -1 +1 -1
+1 -1 -1 +1 +1 -1 -1
+1 +1 +1 -1 +1 -1 -1
+1 -1 +1 -1 -1 +1 -1
+1 +1 -1 -1 -1 -1 +1
+1 -1 -1 -1 +1 +1 +1
Developed models could be obtained by putting the values of the regression coefficients obtained from equation (4) in the selected
model,
Y = b0 + b1N + b2S + b3D + b12NS + b13ND + b23SD …………………... (4)
The regression coefficients of the selected model were calculated using Equation -3. This is based on the method of least squares.
bj = ∑NiXjiYi/N, j = 0,1,…..,k ………….(3)
Where,
Xji = value of a factor or interaction in coded form
Yi =Average value of response parameter
N= No.of observations
k= Number of coefficients of the model
The adequacy of the model was determined by the analysis of variance technique. The regression coefficients were determined
by the method of least square, from which the F-ratio for the polynomial was found. The variance of the response and the
adequacies were calculated. The ‘F’- the ratio of the model were compared with the corresponding ‘F’- ratio from the standard
table and it was found that the model is adequate within 95% level of confidence, thus justifying the use of assumed polynomials.
After calculating the coefficients of the model, it must be tested for its fitness. Thus adequacy of the model was tested using
analysis of variance technique. For this, the variance of optimization parameter (S2
y) was determined. For two repetitions of each
trial the formula is:
S2
y = 2∑N
i=1(∆Y2
)/N ……………………………. (5)
Where ∆Y2
= (Yiq – Ym)2
Ym = Arithmetical mean of repetitions (response in the repetitions)
Yiq = Value of response in a repetition trial
N = Number of observations
i= No. of trials
q=No. of repetitions

Further, the variance of adequacy, also called residual variance was determined by using following formula:
S2
ad = ∑N
i=1 ∆ (Yp – Ym)2
/f………………………..(6)
Where
S2
ad = Variance of adequacy
Ym = Observed or Measured response
Yp = Estimated/Predicated value of response (obtained from model)
f = N-(K+1) (Degree of freedom)
∑N
i=1 = residual sum of square
k = Number of independently controllable variables.
The ratio of the variance of adequacy to the variance of optimization parameter gave Fisher ratio:
F = S2
ad/ S2
y…………….………………………… (7)
F-values, thus, obtained and denoted as (Fm) were compared from the table value as (Ft). It was found that the model was adequate
at 95% level of significance thus justifying the use of the assumed polynomial.
The final model could be obtained by dropping statistically in-significant terms from the developed models.
The vertical milling machine selected to install in Mechanical Workshop, Department of Mechanical Engineering, Gita College of
Engineering and Technology, Kurukshetra. The specifications of the vertical milling machine in presented in Table.
Table: Showing Milling Machine Specifications
S. No Specifications Units
1 Size of the Table 1500×400mm
2 No. of Speeds 10
3 Spindle Speed range Low(140,250,360 and 500)
High(1200,1950,2500 and 2800)
4 Motor 5 HP
Table 7: Tool Specifications
Shoulder Diameter: 18 mm and 21 mm
Shoulder Height: 30 mm
The height of Pin: 5.8 mm.
Pin Diameter: 6 mm
Table 8: Chemical Composition of Base Material
Si Fe Cu Mn Mg Al
0.57 0.35 022 0.12 1.1 Bal
Table 9: Mechanical Properties of Base Material
Ultimate Tensile Strength (MPa) Elongation Hardness Hv
280 20 100

V. RESULTS
The design expert software 6.0 was used to analyze the results. The final mathematical models for tensile strength after
dropping insignificant co-efficient is given below:
T= 208+8.33N -6.25S +0.33D+ 4.83NS+1.08ND-1.50SD
This mathematical model can be used to predict the effects of the parameters on the tensile strength by substituting the
values of respective factors in coded form. In the mathematical model for tensile strength, the effects of interaction of two
parameters can be explained well.
Table: 10 ANOVA for Selected Factorial Model for Tensile Strength
Source
Sum of
Squares
DF Mean Square F Value
p-value Prob
> F
Significant
Model 3249.67 6 541.61 127.59 < 0.0001 Yes
N-tool
rotational
speed
1666.67 1 1666.7 392.61 < 0.0001 Yes
S-Welding
speed
937.50 1 937.0 220.84 < 0.0001 Yes
D-shoulder
Diameter
2.67 1 267 0.63 0.0809 No
NS 560.67 1 560.67 132.07 < 0.0001 Yes
ND 28.17 1 28.17 6.64 0.0196 Yes
SD 54.00 1 54.00 12.72 0.0024 Yes
Residual 72.17 17 4.25
Lack of Fit 28.17 1 28.17 10.24 0.56 No
Pure Error 44.00 16 2.75
Cor Total 3321.3 23
The "Pred R-Squared" of 0.9567 is in reasonable agreement with the "Adj R-Squared" of 0.9706; i.e. the difference is less than 0.2.
"Adeq Precision" measures the signal to noise ratio. A ratio greater than 4 is desirable. Your ratio of 30.855 indicates an adequate
signal. This model can be used to navigate the design space. Scatter diagrams, which show the predicted and the observed values
of responses, were also drawn so as to test the validity of these models. A good agreement was found to exist between the actual
and the predicted responses of percentage elongation as shown in Fig. 4.1 and Fig 4.2 respectively. The Model F-value of 127.59
implies the model is significant. There is only a 0.01% chance that an F-value this large could occur due to noise. Values of
"Prob > F" less than 0.0500 indicate model terms are significant. In this case, N, S, D, NS, ND are significant model terms. Values
greater than 0.1000 indicate the model terms are not significant. If there are many insignificant model terms (not counting those
required to support hierarchy), the model reduction may improve your model. The "Lack of Fit F-value" of 10.24 implies the Lack
of Fit is not significant. There is only a 0.56% chance that a "Lack of Fit F-value" this large could occur due to noise. Significant
lack of fit is bad. The design expert software 6.0 was used to analyse the results. The final mathematical models for tensile
strength after dropping insignificant co-efficient is given below:
T= 208+8.33N -6.25S +0.33D+ 4.83NS+1.08ND-1.50SD
Fig 4.1 Normal plot of Residuals for Tensile Strength (MPa)

Fig.4.2 Predicted v/s Actual Plot for Tensile Strength (MPa)
Figure 4.3 Effect of Tool Rotational Speed on Tensile Strength
Fig.4.4 Effect of Welding Speed on Tensile Strength

Fig.4.5 Interaction effect of Rotational Speed and Welding Speed on Tensile Strength
Fig.4.6 3D Surface Plot for Influence of Interaction of rotational Speed and Welding Speed on Tensile Strength
Fig.4.7 3D Surface Plot for Influence of Interaction of Rotational Speed and Shoulder Diameter on Tensile Strength

Table: Showing ANOVA for Selected Factorial Model Micro Hardness
Source
Sum of
Squares
DF
Mean
Square
F Value
p-value
Prob > F Significant
Model 1083.92 6 180.65 102.23 < 0.0001
N-tool
rotational
speed
610.04 1 610.04 345.21 < 0.0001
Yes
S-Welding
speed
425.04 1 425.04 240.52 < 0.0001
Yes
D-Shoulder
diameter
18.38 1 18.38 10.40 0.0050
Yes
NS 5.04 1 5.04 2.85 0.1095
No
ND 3.38 1 3.38 1.91 0.1849
No
SD 22.04 1 22.04 12.47 0.0026
Yes
Residual 30.04 17 1.77
Lack of Fit 22.04 1 22.04 44.08 < 0.0001
Yes
Pure Error 8.00 16 0.50
Core Total 1113.96 23
VI. CONCLUSIONS
1. Two Level Factorial Design is found to be an effective tool to investigate the interaction effects of parameters on the
required response.
2. Proposed models are adequate at 95% confidence level, thus justifying the use of assumed polynomials.
3. Tensile Strength increase with a decrease in rotational speed of the tool.
4. Tensile Strength decrease with increase in welding speed.
5. At the low rotational speed of tool high welding speed, there is the small increasing effect on Tensile Strength. But at
high rpm, and high welding speed, there is a significant increase in Tensile Strength.
6. At low welding speed and high rotational speed, there is a sharp decrease in Tensile Strength. At high welding speed, as
increase the rpm, there is a little decrease in Tensile Strength.
7. Prediction of the Tensile Strength at any combination of the two parameters welding speed and rotational speed can be
done.
8. With increase shoulder diameter, there is a decrease in Impact Strength.
9. Impact Strength increases with the increase in tool rotational speed.
REFERENCES
1. Muthukumaran S. and Mukherjee s. K. (2007), “ First mode offers compactness in friction stir welding process”, welding
–productivity and quality (WPQ-2007), pp. 481-487
2. R. Nandan, T. DebRoy, H.K.D.H. Bhadeshia Recent advances in friction-stir welding – Process, weldment structure and
properties Progress in Materials Science 53 (2008) 980–1023
3. Schmidt H. N. Dickerson T. L. and Hattel J. H. (2006) “Material flow in butt friction stir welds in AA2024-T3” acta
Materialia, 54, pp. 1199-1209.
4. Schmidt H. N. Dickerson T. L. and Hattel J. H. (2006) “Material flow in butt friction stir welds in AA2024-T3”, Acta
Materialia, vol. 54, 2006, pp. 1199-1209.
5. Tomotake Hirata, Taizo Oguri, Hideki Hagino, Tsutomu Tanaka, (2007) “Influence of friction stir welding parameters on
grain size and formability in 5083 aluminum alloy” Materials Science and Engineering A 456, pp 344–349.

6. William D. Lockwood, Borislav Tomaz, A.P. Reynolds,” Mechanical response of friction stir welded AA2024:
experiment and modeling”, Materials Science and Engineering A323 (2002) 348–353
7. Yingchun Chena, Huijie Liu, Jicai Feng (2006) “Friction stir welding characteristics of different heat-treated-state 2219
aluminum alloy plates” Materials Science and Engineering A420, pp 21–25
8. Z. Zhang, H.W. Zhang Numerical studies on the effect of transverse speed in friction stir welding Materials and Design
30 (2009) 900–907
9. A. K. Lakshminarayanan, V. Balasubramanian “Process parameters optimization for friction stir welding of RDE-40
aluminium alloy using Taguchi technique”, Trans. Tech Non-ferrous Soc. Of china, 18, pp. 548-554
10. A. K. Lakshminarayanan, V. Balasubramanian 2009 “Comparison of RSM with ANN in predicting the tensile strength
of friction stir welded AA7039 aluminium alloy joints”, Trans. Tech Non-ferrous Soc. Of china Journal of Material
Science Technology, 19 p. 9-18
11. A. Scialpi, L.A.C. De Filippis, P. Cavaliere, (2007) “Influence of shoulder geometry on microstructure and mechanical
properties of friction stir welded 6082 aluminium alloy” Materials and Design 28 pp. 1124–1129
12. A. Simar, Y. Br´echet, B. De Meester , A. Denquin, T. Pardoen, “Microstructure, local and global mechanical properties
of friction stir welds in aluminium alloy 6005A-T6”, Materials Science and Engineering A 486 (2008) 85–95

Effect of process parameter on mechanical properties of friction stir welded aluminium alloy joints using factorial design

More Related Content

What's hot (20)

Similar to Effect of process parameter on mechanical properties of friction stir welded aluminium alloy joints using factorial design (20)

Recently uploaded (20)

Effect of process parameter on mechanical properties of friction stir welded aluminium alloy joints using factorial design