Davidson et al.

/ Usak University Journal of Material Sciences 2 (2012) 137 – 145

Usak University Journal of Material

Sciences
Journal homepage: http://uujms.usak.edu.tr

Research article

Analysis of tube end forming process using Taguchi design of

experiments
L. Venugopal, M.J. Davidson*, N. Selvaraj
Department of Mechanical Engineering, National Institute of Technology, Warangal A.P, India

Received 22 August 2012 Revised 07 December 2012 Accepted 07 December 2012

Abstract
In this study, the Taguchi method is used to find the optimum process parameters for maximum
expansion of tube ends. The various process parameters namely the punch/die cone angle, the expansion
ratio and the friction conditions are taken as the input process condition and the output; the maximum
radial displacement is critically analyzed. Radial expansion is the increase in diameter of the tube by the
die measured on the top most portion of the tube across its circumference. The optimal combination of the
process parameters is obtained through the signal to noise ratio (S/N) analysis and the analysis of
variance (ANOVA) methods. The parameters that affect the process are determined using Taguchi method
and the most significant process parameters and their percentage contribution was determined by using
ANOVA technique. Among all the three process parameters considered, it is found that the most significant
factor is the die cone angle (α) and this factor contributes 43.56% on the total output response value while
expansion ratio rp/r0 contributes 8.89% and the lubricant has contributed 38.59% on the total output. The
experimental results are in acceptance with the predicted values for the 95% confidence interval and it is
observed that the error is within the reasonable limit.
©2012 Usak University all rights reserved.

Key words: Tube end expansion, Taguchi analysis, ANOVA, AA6061 tubes

1. Introduction

End forming of tubes finds application in many fields such as power transmission
applications, fuel lines and exhaust pipes, hydraulic cylinders and heat exchangers etc.
End forming is defined as forming the end of tubular forms, either by inverting the tube
or by expanding it. A punch with a divergent end on one side will be inserted into the
tube. The tube will get the shape of the punch after the die is retracted back due to plastic
deformation. In this expansion process, formability is limited by the occurrence of local
buckling or ductile fracture. In metal forming processes, it is essential that the metal
deforms plastically without fracture.

Tube end forming process is governed by various process parameters namely punch/die
cone angle, ratio between punch radius and initial reference radius, friction, punch speed,
tube length and tube thickness. To develop the process for the tube end forming process,

*Corresponding author: Tel: +91 8985786887/2500

E-mail: [email protected]

©2012 Usak University all rights reserved. 137

 Davidson et al. / Usak University Journal of Material Sciences 2 (2012) 137 – 145

it is essential to know the process parameters that affect it and also the most significant
and non-significant process parameters that affect the desired design requirement. This
can be achieved by conducting large number of experiments. However, such a time
consuming experimental procedure can be substituted with “Taguchi method”, which is
based on design of experiments concept. Many researchers have worked on the
optimization techniques in recent years. Bing Li et al. [1] did multi-objective optimization
of hydro forming process parameters by using Taguchi method and finite element
method. They found that the internal pressure and the friction co-efficient have the
greatest influence on the free bulging of tube hydro forming process. Dae-Cheol et al. [2]
used Taguchi method and ANN to study the workability limited by the ductile fracture.
The workability of the material along the path of the process was determined using ANN
and Taguchi method was used to minimize the objective function.

Hsin-Te et al. [3] used design of experiments concepts to optimize the process
parameters of chemical-mechanical polishing process in wafer manufacturing. Material
removal rate and non-uniformity of surface profiles were selected as the target values,
and the process parameters such as platen speeds, carrier speeds, back side pressure,
slurry flow rate and a head down force were optimized. Jurkovic et al. [4] conducted
experiments to know the influence of forming process parameters with an objective to
minimize the load on the tool. The process parameters such as strain, die angle and
coefficient of friction were optimized by using both classical optimization techniques and
design of experiments concepts. The merits and demerits of both the classical and DOE
methods were highlighted. In this study, the authors calculated the maximum expansion
ratio of the tube and the fracture behavior of the tube. The corresponding changes in the
hardness values were investigated. Padmanabhan et al. [5] studied the influence of
process parameters on the deep drawing of stainless steel. Deep drawing process
parameters such as die radius, blank holder force and friction-coefficient were studied to
find the optimum condition for forming. It was found that the die radius followed by the
blank holder force and friction coefficient have the greatest influence on the deep
drawing of stainless steel blank. Rossella et al. [6] optimized aluminium foam
manufacturing process with Taguchi method. In this study, the relative density of
aluminium foams were minimized by optimally controlling three process parameters
namely silicon carbide content in powder mixture, the compaction pressure and the
forming temperature. Su-Hai et al. [7] applied the artificial neural networks (ANN) to
predict the forming load of magnesium alloy under hot extrusion. In this study, ANN
analysis was used to determine the die shapes for various extrusion ratios. Wei et al [8]
implemented multi-objective optimization to optimize the gear warm forming process
parameters.

Zhubin et al. [9] studied the formability and deformation behavior of AZ31B tube at
elevated temperature by tube bulging test. Almedia et al. [10] performed expansion and
reduction studies on AA 6060 tubes. They forced a tapered dedicated punch into the tube
end retracted it after achieving the desired shape. The variation of tube end formability
with the variation in tribological conditions of AA6060 tube was studied. The work
reported three different modes of deformation. Alves et al. [11] performed tube end
forming experiments on AA6060 alloy. External inversion of thin walled tubes was
performed by Rosa et al. [12] on AA 6060 alloy. It was reported that, during external
inversion, a smaller die radius lead to local buckling, whereas a larger die radius has led
to ductile damage fracture. Experiments were performed on samples with two different
mechanical and metallurgical conditions resulting from natural aging and annealing.
Schaeffer et al. [13] gave the brief description of each forming process such as expansion,
reduction, internal inversion and external inversion and focused on understanding the

138
 Davidson et al. / Usak University Journal of Material Sciences 2 (2012) 137 – 145

parameters that govern each process. The objective of the present work is to optimize the
process parameters involved in tube expansion process by using Taguchi and ANOVA
methods. The parameters that affect the process were determined using Taguchi method,
and the most significant process parameters and their percentage contribution were
determined by using ANOVA technique.

2. Plan of Experiments

2.1. Material Properties

The material used in the present work is AA6061 alloy. The major alloying elements are
Al-1Mg-0-0.6Si-0.7Fe-0.25Cu-0.2Cr. AA6061 alloy is the most extensively used metal in
the 6000 series aluminium alloys. It is a heat treatable alloy having good toughness
characteristics and excellent corrosion resistance behavior. AA6061 alloy has ultimate
tensile strength of 282.6 MPa, yield strength of 267 MPa and a percentage elongation of
7.58. Due to its excellent mechanical properties, the alloy is used in aircraft structures,
bicycle frames, automotive parts, land based pipelines and aircraft support structures.
Fig.1 shows the true stress-true strain curve obtained by means of compression test
conducted in the laboratory.

700
True stress (MPa)

600
500
400
300
200
100
0
0 0,5 1 1,5
True strain (mm)

Fig.1 True stress-true strain curve of AA6061 aluminium alloy

2.2. Equipment

The present experiment was carried out on a hydraulic press of 50 ton capacity. The
machine used in this work is shown in Fig.2.

Fig. 2 Hydraulic press

139
 Davidson et al. / Usak University Journal of Material Sciences 2 (2012) 137 – 145

The die was mounted on the ram. A constant ram speed of 0.5 mm/s was taken. Before
starting the test, the tube was adjusted in such a way that the axis of tube and punch were
aligned. During the forming process, the load was increased by rotating the pressure
valve manually. Punch displacements were measured using a LVDT. PC-based data
logging system was used to record and store the loads and displacements. The process
was continued until the complete punch moved inside the tube fully or a crack appeared
at the tube tip. In the similar way, experiments were done by using different lubricants
namely palm oil (vegetable oil), VG460 (mineral oil) and MOS 2 (solid lubricant). A tube
that was expanded by inserting the punch as described above is shown in Fig. 3. A die
with rp/r0=1.53, and α=30° was used in the above case.

Fig.3 Expansion of tube with the die of rp/r0=1.53, α=30°

2.3. Taguchi Method

Taguchi method was developed by Genichi Taguchi, a Japanese engineer. This process is
based on statistical design concepts. In the present study, Taguchi method has been used
to optimize the process parameters with the aim of achieving the maximum expansion of
tube end. The factors and levels considered in the experiments are given in Table 1.

Table 1
Control factors and levels
Factor
Factors Level 1 Level 2 Level 3
Designation
Die angle (α) A 15° 30° 45°
Expansion ratio rp/r0 B 1.39 1.53 1.67
Lubricant C Palm oil VG460 MOS2

The process parameters considered for the study are die cone angle (α), expansion ratio
rp/r0 and lubricant. Three levels have been considered for each parameter. The
experiments were designed using L9.standard orthogonal array as given in Table 2.

3. Results and Discussions

3.1. Taguchi Analysis

The experiments were conducted as per the experimental layout given in Table 2. A total
of nine experiments were done. Taguchi method uses the S/N (signal-to-noise) ratio. S/N
ratio is used to determine the most significant factor. There are three types of S/N ratio
criteria for optimization namely smaller the best, larger the better and nominal the best.

140
 Davidson et al. / Usak University Journal of Material Sciences 2 (2012) 137 – 145

Table 2
Layout of experimental design
Experiment No. Factor A Factor B Factor C Radial expansion
1 15° 1.39 Palm Oil 53.50
2 15° 1.53 VG 460 50.50
3 15° 1.67 MOS2 51.32
4 30° 1.39 VG460 51.80
5 30° 1.53 MOS2 53.40
6 30° 1.67 Palm 56.00
7 45° 1.39 MOS2 53.68
8 45° 1.53 Palm 54.82
9 45° 1.67 VG460 54.74

In the present study, to get the maximum expansion results, the quality parameter “larger
the better” has been chosen. The S/N ratio for “larger the best” a criterion is as follows:

[ ∑ ] (1)

where; Yi is the measured value of quality characteristic and ‘n’ is the number of
repetitions for the experimental combination. Signal to noise ratio values for different
levels and different factors are calculated from the above formula and given in Table 3.

Table 3
Response table for signal to noise ratios (Larger is better)
Level A B C
1 34.28 34.48 34.77
2 34.60 34.47 34.37
3 34.71 34.65 34.45
δ 0.43 0.18 0.40
Rank 1.00 3.00 2.00

Based on the delta values, the process parameters are ranked. From Table 3, it is clear
that the most significant factor is the die cone angle (α) followed by lubricant and
expansion ratio rp/r0. Fig. 4 gives the main effects plots for the S/N Ratio.

The graph is plotted between the mean of the S/N ratio values and the number of levels
for all the three factors. The optimum process parameters and levels are A3, B3 and C1. In
addition to the S/N ratio analysis, main effects of the process parameters on the mean
response are also analyzed. The mean response refers to the average value of the quality
characteristic for each factor at different levels. Thus, the average values of the
expansion, for each factor, at the three levels have been calculated and are plotted in Fig.
5.

Calculated values of means are given in Table 4. The mean response analysis also
indicates the same optimum level of the parameters (i.e. A3, B3 and C1) as obtained in
S/N ratio analysis.

141
 Davidson et al. / Usak University Journal of Material Sciences 2 (2012) 137 – 145

Table 4
Response table for means
Level A B C
1 51.77 52.99 54.77
2 53.73 52.91 52.35
3 54.41 54.02 52.80
δ 2.64 1.11 2.43
Rank 1.00 3.00 2.00

34,8 34,7
Mean of S/N

Mean of S/N
34,6 34,6

34,4 34,5
A B
34,2 34,4
0 2 4 0 2 4
Levels Levels

34,8
Means of S/N

34,6

34,4 C

34,2
0 2 4
Levels
Fig. 4 Main effects plot for S/N ratios

55 55
Mean of Means

Mean of Means

54
54
53
53
52 B
A
51 52
0 2 4 0 2 4
Levels Levels

55
Mean of Means

54

53 C

52
0 2 4
Levels

Fig. 5 Main effects plot for means

142
 Davidson et al. / Usak University Journal of Material Sciences 2 (2012) 137 – 145

3.2. Analysis of Variance (ANOVA)

The analysis of variance (ANOVA) is used to determine the most significant factor among
all the process parameters which are affecting the output quality characteristics using the
quantities such as degrees of freedom, sum of squares, variance, F-ratio and percent
contribution. Table 5 shows the computed results of the ANOVA with 95% confidence.
The F-ratio and the percent contributions of the various parameters as quantified under
the respective columns of Table 5 reveal that die cone angle (α) has the most significant
parameter affecting the tube expansion process. This factor contributes 43.56% on the
total value while expansion ratio rp/r0 contributes 8.89% and the lubricant has
contributed 38.59% on the total value.

Table 5
Analysis of variance (ANOVA) for expansion
Degree of Sum of Contribution
Source Variance F-Ratio
Freedom Squares Ratio (%)
A 2 11.274 5.637 4.87 43.565
B 2 2.301 1.151 0.99 8.891
C 2 9.988 4.994 4.31 38.590
% Error 2 2.315 1.158 8.945
Total 8 25.878 100.000

3.3. Prediction of Optimum Value

From the response values for S/N ratios and the response values from means, the
optimum control factors and levels have been determined as A 3, B3 and C1. The predicted
mean (M) of the response characteristic has been expressed as follows:

M= ̅ (̅̅̅̅ ̅) (̅̅̅̅ ̅) (̅̅̅̅ ̅) (2)

where; ̅ is average expansion of all the 9 experiments, ̅̅̅̅, ̅̅̅̅ and ̅̅̅̅ are the response
values of mean with process parameters at optimal levels. The calculated response values
are ̅ =53.306, ̅̅̅̅=54.41, ̅̅̅̅=54.02 and ̅̅̅̅=54.77, substituting all the values in Eq. 2, the
mean optimum value of the expansion has been predicted as M=56.588.

3.4. Confirmation Test

In order to confirm the validity of the model developed, a confirmation test was
conducted. The confirmation experiment is conducted in order to test the predicted
results with optimum levels of the process parameters. Table 6 shows the results of
confirmation experiments, and it is observed that the radial expansion of tubes obtained
from experiments is within the range of predicted 95% confidence interval and it is
observed that the error is within the reasonable limit.

Table 6
Results of confirmation experiments
Optimum Predicted values Experimental Error
Response
process levels (mm) values (mm) (%)
Radial expansion A3B3C1 56.588 58.160 2.700

143
 Davidson et al. / Usak University Journal of Material Sciences 2 (2012) 137 – 145

4. Conclusion

The effect of tube end forming process parameters on the expansion of tube ends is
experimentally investigated. Taguchi method has been used to optimize the process
parameters. The experiments were designed using L9 standard orthogonal array. The
process parameters considered for the study are die cone angle (α), expansion ratio rp/r0
and lubricant. The following conclusions are drawn from the present work:

 Among all the three parameters, the most significant factor is the die cone angle
(α), and this factor contributes to 43.56% of the total expansion value.
 The optimum control factors and levels have been determined as A3, B3 and C1.
 An optimized value for expansion of tube ends for 95% confidence interval was
predicted as 56.58 mm.
 The value obtained from the confirmation experiments is within the predicted
range which is 58.16.
 The predicted values obtained from the Taguchi method are in good agreement
with the experimental values. Thus, the developed model can be used to predict
the tube expansion values for the chosen input parameters successfully.

Acknowledgements

This topic of the research is supported by the Department of Science and Technology,
Government of India, under Young Scientist Scheme through the research grant
SR/FT/Et-071/2008 dated May 21, 2009 sanctioned to Dr M J Davidson. The authors
would like to thank DST for the funding.

References

1. Bing L, Nye TJ and Metzger Don R. Multi-objective optimization of forming

parameters for tube hydroforming process based on the Taguchi method.
International Journal of Advanced Manufacturing Technology, 2006; 28: 23 – 30.
2. Dae-Cheol K, Dong-Hwan K and Byung-Min K. Application of artificial neural
network and Taguchi method to preform design in metal forming considering
workability. International Journal of Machine Tools & Manufacture, 1999; 39:
771 – 785.
3. Hsin-Te L, Jie-Ren S and Yung-Kuang Y. Applications of Taguchi and design of
experiments methods in optimization of chemical mechanical polishing process
parameters. International Journal of Advanced Manufacturing Technology,
2008; 38: 674 – 682.
4. Jurkovic Z, Jurkovic M and Buljan S. Optimization of extrusion force prediction
model using different techniques. Journal of Achievements in Materials and
Manufacturing Engineering, 2006; 17: 353 – 356.
5. Padmanabhan R, Oliveira MC, Alves JL and Menezes LF. Influence of process
parameters on the deep drawing of stainless steel. Finite Elements in Analysis
and Design, 2007; 43: 1062 – 1067.
6. Rossella S, Filippis LAC, Ludovico AD and Boghetich G. Application of Taguchi
method for the multi objective optimization of aluminium foam manufacturing
parameters. International Journal of Material Forming, 2010; 3: 1 – 5.
7. Hsiang SH and Jer-Liang K. Applying ANN to predict the forming load and
mechanical property of magnesium alloy under hot extrusion. International
Journal of Advanced Manufacturing Technology, 2005; 26: 970 – 977.

144
 Davidson et al. / Usak University Journal of Material Sciences 2 (2012) 137 – 145

8. Wei F and Lin H. Multi-objective optimization of process parameters for the

helical gear precision forging by using Taguchi method. Journal of Mechanical
Science and Technology, 2011; 25: 519 – 1527.
9. Zhubin H, Yanli L, Jia W and Shijian Y. Study on the form ability and deformation
behavior of AZ31B tube at elevated temperature by tube bulging test. Journal of
Materials Engineering and Performance, 2010; 20(7): 1278 – 1284.
10. Almeida B, Alves PP, Rosa ML, Brito PAR and Martins PAF. Expansion and
reduction of thin-walled tubes using a die: Experimental and theoretical
investigation. International Journal of Machine tools & Manufacture, 2006; 46:
1643 – 1652.
11. Almeida B, Alves PP, Rosa ML, Brito PAR and Martins PAF. End forming of thin-
walled tubes. Journal of Material Processing Technology, 2006; 177: 183 – 187.
12. Schaeffer L and Brito AMG. FEM numerical simulation and experimental
investigation on end forming of thin-walled tubes using a die. Steel Research Int.,
2007; 78: 10 – 11.
13. Rosa PAR, Rodrigues JMC and Martins PAF. External inversion of thin-walled
tubes using a die: experimental and theoretical investigation. International
Journal of Machine tools & Manufacture, 2003; 43: 787 – 796.

145