Measurement 50 (2014) 1928

Contents lists available at ScienceDirect

Measurement
journal homepage: www.elsevier.com/locate/measurement

Optimization of surface roughness and ank wear using the

Taguchi method in milling of Hadeld steel with PVD and CVD
coated inserts
Turgay Kvak
Dzce University, Faculty of Technology, Department of Manufacturing Engineering, 81620 Dzce, Turkey

a r t i c l e

i n f o

Article history:
Received 2 March 2013
Received in revised form 4 November 2013
Accepted 12 December 2013
Available online 27 December 2013
Keywords:
Milling
PVD/CVD coating
Surface roughness
Flank wear
Taguchi method

a b s t r a c t
In this paper, the Taguchi method and regression analysis have been applied to evaluate
the machinability of Hadeld steel with PVD TiAlN- and CVD TiCN/Al2O3-coated carbide
inserts under dry milling conditions. Several experiments were conducted using the L18
(2  3  3) full-factorial design with a mixed orthogonal array on a CNC vertical machining
center. Analysis of variance (ANOVA) was used to determine the effects of the machining
parameters on surface roughness and ank wear. The cutting tool, cutting speed and feed
rate were selected as machining parameters. The analysis results revealed that the feed
rate was the dominant factor affecting surface roughness and cutting speed was the dominant factor affecting ank wear. Linear and quadratic regression analyses were applied to
predict the outcomes of the experiment. The predicted values and measured values were
very close to each other. Conrmation test results showed that the Taguchi method was
very successful in the optimization of machining parameters for minimum surface roughness and ank wear in the milling the Hadeld steel.
2013 Elsevier Ltd. All rights reserved.

1. Introduction
Hadeld steel (austenitic manganese steel) is a very
tough and non-magnetic alloy and having excellent wear
resistance with high strength and ductility [1]. In recent
years, due to its excellent wear resistance properties this
steel has been widely used in various engineering applications including excavators, mining equipment, pumping
equipment, railways, rolling mill parts for steel factories
and wear-resistant components of machining elements.
However, the machining of this steel is very difcult due
to its extreme hardness, low thermal conductivity and
strain hardening behavior. Consequently, greater cutting
forces, higher cutting temperatures and wearing are present making the control of dimension precision difcult
during machining [24].
Tel.: +90 0380 542 11 33; fax: +90 0380 542 11 34.
E-mail address: turgaykivak@duzce.edu.tr
0263-2241/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.measurement.2013.12.017

The technology of cutting tools is rapidly improving and

this development is necessary to improve the wear resistance and performance of machining on difcult-to-cut
materials. Improvement in performance can be achieved
by increasing the strength of the cutting tools in terms of
wear resistance [5]. Coating technology is one of most
important ways of increasing the machining performance
of cutting tools. In recent years, the tendency towards coating technology has increased with the increase of cutting
speed and feed rate. Cementite carbide cutting tools are
coated using two different methods: physical vapor deposition (PVD) and chemical vapor deposition (CVD). In the
last few years, the comparison of PVD and CVD coatings
from the point of machining performance has been the
subject of much research. In some studies, PVD coatings
produced better results from the point of machining performance [6,7]; whereas CVD coatings showed better performance in other studies [812]. For this reason,
specication of suitable cutting tools and coating types

20

T. Kvak / Measurement 50 (2014) 1928

becomes important for increasing machining performance

and decreasing production costs.
Surface quality is one of the most signicant indicators
of engineering materials; whereas surface roughness plays
an important role in the specication of the surface quality
of the produced parts. Good surface roughness provides
important improvements in the tribologic characteristics,
fatigue strength, corrosion resistance and aesthetic appearance of the product [1315]. In addition, the surface roughness affects several attributes of machined parts such as
friction, wear, and heat transmission [16]. Another point
which plays a key role in high-quality and economical production is tool wear. Surface roughness is related to tool
wear and is directly affected by it. There are a number of
parameters affecting surface roughness and tool wear such
as cutting tool material, coating material, cutting speed
and feed rate. Obtaining minimum surface roughness and
tool wear by the optimization of these parameters is very
important from the aspect of cost reduction. For this reason, in recent years, a number of statistical models have
been developed for the analysis and optimization of
machining parameters such as response surface methodology (RSM), regression techniques, analysis of variance
(ANOVA) and the Taguchi method. The Taguchi-based
optimization technique has produced a unique and powerful optimization discipline that differs from traditional
practices [17].
Mandal et al. [18] applied the Taguchi method and
regression analysis to assess machinability of AISI 4340
steel with newly-developed zirconia-toughened alumina
ceramic inserts. It has been observed that the depth of
cut is a maximum contribution to tool wear. Nalbant
et al. [19] used the Taguchi method to determine the optimal cutting parameters for surface roughness in turning.
The orthogonal array, the signal-to-noise ratio, and analysis of variance were employed to study the performance
characteristics in turning operations of AISI 1030 steel bars
using TiN-coated tools. Experimental results were provided to illustrate the effectiveness of this approach. Ghani
et al. [20] applied the Taguchi method to optimize cutting
parameters in end milling when machining hardened AISI
H13 steel with TiN-coated P10 carbide insert tools under
semi-nishing and nishing conditions of high-speed cutting. An orthogonal array, signal-to-noise ratio and Pareto
analysis of variance were employed to analyze the effect
of the milling parameters. The analysis of the results
showed that the optimal combination for a low-resultant
cutting force and a good surface nish was a high cutting
speed, low feed rate and low depth of cut. Fetecau and Stan
[16] investigated the turning of polytetrauoroethylene
composites using a polycrystalline diamond tool in order
to analyze the effects of the cutting parameters and insert
radius on the cutting force and surface roughness. The signal-to-noise ratio and the analysis of variance were applied
to the experimental data in order to determine the effect of
the process variables on the surface roughness and cutting
force, and predictive models were derived. Statistical results indicate that the cutting force and surface roughness
are signicantly inuenced by the feed rate. Yang et al. [21]
applied the designs-of-experiments (DOE) approach to
optimize the parameters of a computer numerical control

(CNC) in end milling for high-purity graphite under dry

machining. The feed rate was found to be the most significant factor affecting the groove difference and the roughness average in the end milling process for high-purity
graphite. Horng et al. [22] used the RSM to evaluate the
machinability of Hadeld steel in hard turning. The combined effects of four machining parameters, (cutting speed,
feed rate, depth of cut and tool corner radius), on the basis
of two performance characteristics (ank wear and surface
roughness), were investigated; the centered central composite design and the analysis of variance were employed.
Results showed that the ank wear was inuenced principally by the cutting speed and the interaction effect of feed
rate on the nose radius of the tool; the cutting speed and
the tool corner radius had a statistical signicance on the
surface roughness. Davim [23] applied the Taguchi method
to investigate the effect of cutting parameters (cutting
velocity and feed rate) and cutting time on drilling
metal-matrix composites. The analysis of the results
showed that the cutting time was the factor which had
the greatest inuence on the tool wear (50%), followed by
the feed rate (24%).
The main problem facing companies in the metal-cutting
industry is the need to increase manufacturing quality and
at the same time to decrease production costs. There are
many variables which affect the quality and production
costs of the product, including cutting parameters, tool
materials, tool geometry, coating technology, lubricants,
etc. Consequently, companies are forced to operate by using
the trial and error method. The optimization of controllable
variables can make a considerable contribution towards
solving the problem. At this point, the variables leading to
a nal solution are being optimized by using the Taguchi
method, thus considerably reducing the number of tests
needed. Therefore, by employing the trial and error method,
production costs will decrease signicantly and time loss
will be minimal. As a result, nowadays, this and similar
methods have become the focus of interest for both academics and companies, with the goal of increasing production
quality and operating with greater efciency.
In this study, the effects of machining parameters on the
surface roughness and ank wear in the milling of Hadeld
steel with PVD TiAlN- and CVD TiCN/Al2O3-coated carbide
inserts were investigated. Taguchis L18 array was used for
conducting the experiments. For the determination of optimal machining conditions (cutting tool, cutting speed and
feed rate) for minimum surface roughness and ank wear,
Taguchis signal-to-noise ratio was used. In addition, linear
and quadratic regression analyses were applied to predict
the measured value. Finally, the reliability of developed
models was tested by the conrmation experiments.

2. Experimental methods
2.1. Milling experiments
The milling experiments were carried out in dry cutting
conditions using a First MCV 350 model three-axis CNC
milling machine equipped with a maximum spindle speed
of 8000 rpm and a 7.5-kW drive motor. The experimental

21

T. Kvak / Measurement 50 (2014) 1928

set-up for the face milling tests is shown in Fig. 1. The

workpiece material used was Hadeld steel in the form
of a 100  50  14 mm block. The chemical composition
of Hadeld (GX120Mn14) steel is given in Table 1. The
milling tests were performed at three different cutting
speeds (90, 120, and 150 m/min) and three feed rates
(0.09, 0.12 and 0.15 mm/rev) while the depth of cut was
kept constant at 0.5 mm.
2.2. Cutting tools and coatings
The cutting experiments were conducted using two
types of cemented carbide tool inserts: PVD-coated and
CVD-coated tools, the properties of cutting tools and
coating materials are given in Table 2. Sandvik
Coromant inserts (R390-11 T3 04M-KM) mounted on a
(R390-016A16-11L) tool holder with a diameter 20 mm
and an overhang length of 40 mm was chosen for
experimental cuts. The milling experiments, only one insert was used, in order to eliminate the inuence of the
tool tip run-out on tool wear.
2.3. Surface roughness measurement
The average surface roughness (Ra) of the workpiece
was measured by a Taylor Hobson (Surtronic 25) portable
surface roughness tester; the cutoff length and evaluation
length were xed at 0.8 mm and 4 mm respectively. The

Spindle

surface roughness was measured parallel to the machined

surface from three different points; the average values of
the measurements were evaluated. Surface roughness
measurements are shown in Fig. 2.
2.4. Tool wear measurement
The ank wear of the worn tool was measured using a
professional handheld digital microscope (Dino-Lite,
AM413ZT) having, polarization and high magnication
capabilities up to 200. The ank wear result was recorded
after 3 min of milling by the Dino-Capture 2.0 software
supplied with the microscope. Each measurement was
conducted three times to avoid any possible mistake during the procedure.
3. Experimental design and optimization
3.1. The Taguchi method and design of experiments
The Taguchi method has been widely used in engineering analysis and is a powerful design. This method dramatically reduces the number of tests by using orthogonal
arrays and minimizes the effects of factors that cannot be
controlled. Furthermore, it provides a simple, efcient
and systematic approach to specifying the optimum cutting parameters in the manufacturing process [24,25].
The Taguchi method uses a loss function to calculate the
deviation between the experimental values and the desired values. This loss function is further converted into a
signalnoise (S/N) ratio (g) [25,26]. Normally, there are
three kinds of quality characteristics in the analysis of
the S/N ratio, namely the lower-the-better, the higherthe-better, and the nominal-the-best [27]. For each level

Ra3

Workpiece

Ra2

Ra1

Cutting
tool

Fig. 1. Experimental setup for face milling tests.

Surface roughness
measuring device

Table 1
Chemical composition of Hadeld steel (wt.%).
C

Si

Mn

Ni

Cr

Mo

Cu

1.22

0.521

13.5

0.133

1.19

0.038

0.088

0.003

0.018

Fig. 2. Surface roughness measurements.

Table 2
Properties of cutting tools and coating materials.
Coated materials

Material quality of ISO (Grade)

Coating thickness (lm)

Hardness (Hv)

PVD

TiAlN

C20 (GC1020)

1600

CVD

TiCN/Al2O3

C40 (GC3040)

1350

Coated method

22

T. Kvak / Measurement 50 (2014) 1928

of the process parameters, the S/N ratio is calculated based

on the S/N analysis. The goal of this study was to minimize
surface roughness and ank wear. Therefore the lowerthe-better quality characteristic was used as shown in Eq.
(1):

Lower  the  better minimize :

"
#
n
1X
g S=NS 10 log
y2i
n i1

where yi is the observed data at the ith experiment and n is

the number of observations of the experiment [18].
Cutting tools (PVD-, CVD-coated), cutting speed (V), and
feed rate (f) were selected as control factors and their levels were determined as shown in Table 3. The most suitable orthogonal array L18 (21  32) was selected to
determine the optimal cutting parameters and to analyze
the effects of machining parameters [28]. The L18 mixed
orthogonal array shown in Table 4 was used for conducting
the experiments.
4. Analysis and evaluation of experimental results
4.1. Analysis of the signal-to-noise (S/N) ratio
Surface roughness (Ra) and ank wear (Vb) were measured via the experimental design for each combination
of the control factors by using Taguchi techniques, optimization of the measured control factors were provided by
signal-to-noise (S/N) ratios. The lowest values of surface
roughness and tool wear are very important for quality
improvement of the product and lowering production

Table 3
Milling parameters and their levels.
Parameters

Symbol

Level 1

Level 2

Level 3

Cutting tools
Cutting speed (m/min)
Feed rate (mm/rev)

A
B
C

PVD
90
0.09

CVD
120
0.12

150
0.15

Table 4
Full factorial design with orthogonal array of Taguchi L18 (21  32).
Experiment no.

Factor A

Factor B

Factor C

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2

1
1
1
2
2
2
3
3
3
1
1
1
2
2
2
3
3
3

1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3

costs. For this reason, the lower-the-better equation

was used for the calculation of the S/N ratio. Table 5 shows
the values of the S/N ratios for observations of the surface
roughness and ank wear. At the end of the milling tests,
the average values of the surface roughness and ank wear
were calculated to be 0.57 lm and 0.23 mm respectively.
Similarly, average values of S/N ratio for surface roughness
and ank wear were calculated to be 5.026 dB and
13.033 dB respectively.
Analysis of the effect of each control factor (Ct, V, f) on
the surface roughness and ank wear was performed with
a S/N response table. The response tables of S/N for Ra
and Vb are shown in Table 6. This table, which is made
by using the Taguchi technique shows the optimal levels
of control factors for the optimal surface roughness and
ank wear values. The level values of control factors for
Ra and Vb given in Table 6 are shown in graph forms in
Figs. 3 and 4. Optimal machining parameters of the control
factors for minimizing the surface roughness and ank
wear can be easily determined from these graphs. The best
level for each control factor was found according to the
highest S/N ratio in the levels of that control factor. According to this, the levels and S/N ratios for the factors giving
the best Ra value were specied as factor A (Level 2,
S/N = 5.256), factor B (Level 3, S/N = 5.461) and factor C
(Level 1, S/N = 6.355). In other words, an optimum Ra value
was obtained with a CVD-coated tool (A2), at cutting speed
(B3) 150 m/min and at a feed rate (C1) 0.09 mm/rev
(Fig. 3). Similarly, the levels and S/N ratios for the factors
giving the best Vb were specied as factor A (Level 2,
S/N = 14.02), factor B (Level 1, S/N = 14.46) and factor C
(Level 1, S/N = 13.95). From the point of machining
conditions, the optimum Vb value was obtained with a
CVD-coated tool (A2), at a cutting speed of 90 m/min
(B3) and at a feed rate 0.09 mm/rev (C1) (Fig. 4).
4.2. Evaluation of experimental results
The changes in the surface roughness and ank wear
which were obtained as the result of experimental study
are seen in Figs. 5 and 6 respectively. Depending on the difference of the cutting tool, there was not much change in
the surface roughness values. However, in ank wear, the
CVD-coated tool displayed an advantage over the PVDcoated tool. This may be a result of the rapid deformation
of the sharp tool edge due to the lesser thickness of the
PVD-coating. The deformation hardening property of the
material also accelerated the deformation of the sharp
edges. However, in the CVD-coated tool, the low thermal
conductivity of the AI2O3-coated layer permitted less heat
ow into the tool substrate, thus ensuring improved wear
resistance.
In both of the cutting tools, surface roughness values
exhibited a decreasing tendency with increasing cutting
speed. An increase of cutting speed decreased the tool-chip
contact area and this also decreased the friction-providing
an increase in the surface quality. Increase in the cutting
speed of PVD- and CVD-coated tools has always been an
important parameter in the increase of ank wear. An increase in the cutting speed causes the thermal and
mechanical loads to increase by raising the temperatures

23

T. Kvak / Measurement 50 (2014) 1928

Table 5
The results of experiments and S/N ratios values.
Experiment
no.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

Control factors
A
Cutting tool
(Ct)

B
Cutting speed
(V)

C
Feed rate
(f)

PVD
PVD
PVD
PVD
PVD
PVD
PVD
PVD
PVD
CVD
CVD
CVD
CVD
CVD
CVD
CVD
CVD
CVD

90
90
90
120
120
120
150
150
150
90
90
90
120
120
120
150
150
150

0.09
0.12
0.15
0.09
0.12
0.15
0.09
0.12
0.15
0.09
0.12
0.15
0.09
0.12
0.15
0.09
0.12
0.15

Surface roughness, Ra
(lm)

S/N ratio for Ra

(dB)

Flank wear, Vb
(mm)

S/N ratio for Vb

(dB)

0.51
0.57
0.76
0.50
0.55
0.66
0.47
0.55
0.67
0.50
0.58
0.71
0.46
0.52
0.66
0.45
0.50
0.59

5.849
4.883
2.384
6.021
5.193
3.609
6.558
5.193
3.479
6.021
4.731
2.975
6.745
5.680
3.609
6.936
6.021
4.583

0.22
0.23
0.24
0.21
0.23
0.26
0.27
0.28
0.33
0.14
0.15
0.18
0.17
0.21
0.22
0.22
0.25
0.30

13.152
12.765
12.396
13.556
12.765
11.701
11.373
11.057
9.630
17.077
16.478
14.895
15.391
13.556
13.152
13.152
12.041
10.458

TRa (surface roughness total mean value) = 0.57 lm.

TRaS/N (surface roughness S/N ratio total mean value) = 5.026 dB.
TVb (ank wear total mean value) = 0.23 mm.
TVbS/N (ank wear S/N ratio total mean value) = 13.033 dB.

Table 6
S/N response table for Ra and Vb factor.
Levels

Control factors
Surface roughness (Ra)

Level 1
Level 2
Level 3
Delta

the most important characteristics of Hadeld steel is its

very rapid hardening during deformation [4]. This property
causes rapid wear of the cutting tools during machining
and shortens their tool life. Extremely high cutting speeds
and feed rates were observed to be effective in the increase
of tool wear. As a result, the CVD-coated tool gained an
advantage over the PVD-coated tool in obtaining lower
Ra and Vb values. The graphs showing the effects of the
control factors obtained with the Taguchi Method (Figs. 3
and 4) on the changes of Ra and Vb verify the results obtained from the experimental studies.

Flank wear (Vb)

4.796
5.256

0.459

4.475
5.143
5.461
0.988

6.355
5.283
3.440
2.915

12.04
14.02

1.98

14.46
13.35
11.28
3.18

13.95
13.11
12.04
1.91

Bold values shows the optimal levels of control factors.

in the cutting area, and this speeds up the deformation of

the cutting tools. It was observed that the most effective
parameter in the increase of surface roughness was feed
rate. Because the surface roughness is a function of feed
rate, an increasing feed rate caused a signicant increase
in the Ra values. Similarly, an increase in feed rate had
an important effect on the increase of ank wear. One of
A, [Ct]

4.3. ANOVA method

ANOVA is a statistical method which is used to determine the individual interactions of all of the control factors
in the test design. In this study, ANOVA was used to analyze the effects of cutting tools, cutting speeds and feed
B, [V (m/min)]

C, [f (mm/rev)]

6.5

Mean of SN ratios

6.0
5.5
5.0
4.5
4.0
3.5
PVD

CVD

90

120

150

0.09

0.12

Fig. 3. Effect of process parameters on average S/N ratio for Ra.

0.15

24

T. Kvak / Measurement 50 (2014) 1928

A, [Ct]

B, [V (m/min)]

C, [f (mm/rev)]

14.5

Mean of SN ratios

14.0
13.5
13.0
12.5
12.0
11.5
11.0
PVD

CVD

90

120

150

0.09

0.12

0.15

0.75
0.50
0.52
0.54
0.56
0.58

0.70
0.65
0.60
0.55
PVD

0.50
150
120
Cuttin
90
g Spee
d (m /m
in)

CVD

g
tin
ut

ols
To

Surface Roughness, Ra (m)

Surface Roughness, Ra (
m)

Fig. 4. Effect of process parameters on average S/N ratio for Vb.

0.75
0.50
0.55
0.60
0.65
0.70

0.70
0.65
0.60
0.55
PVD

0.50
0.15
0.12

Feed R
ate (m
m

0.09

CVD

ng
tti
Cu

o
To

ls

/rev)

Fig. 5. Effect of the cutting parameters on surface roughness.

Flank Wear, Vb (mm)

0.30
0.28
0.26
0.24
0.22
0.20
0.18

PVD

ls

0.16

150
120
Cuttin
g Spee
90
d (m /m
in)

CVD

Cu

ng
tti

o
To

0.16
0.18
0.20
0.22
0.24
0.26
0.28
0.30

0.32

Flank Wear, Vb (mm)

0.16
0.18
0.20
0.22
0.24
0.26
0.28
0.30

0.32

0.30
0.28
0.26
0.24
0.22
0.20
0.18

PVD

0.16

0.15

Fe e d R

0.12

ate (mm

/rev)

0.09

CVD

Cu

T
ng
tti

ls
oo

Fig. 6. Effect of the cutting parameters on ank wear.

rates on surface roughness and tool wear. The ANOVA results for the surface roughness and ank wear are shown
in Table 7. This analysis was carried out a 5% signicance
level and a 95% condence level. The signicance of control
factors in ANOVA is determined by comparing the F values
of each control factor. The last column of the table shows
the percentage value of each parameter contribution

which indicates the degree of inuence on the process performance. According to Table 7, the percent contributions
of the A, B and C factors on the surface roughness were
found to be 2.55%, 9.33% and 82.38% respectively. Thus,
the most important factor affecting the surface roughness
was feed rate (factor C, 82.38%). According to the ANOVA
results, the percent contributions of the A, B and C factors

25

T. Kvak / Measurement 50 (2014) 1928

Table 7
Results of ANOVA for surface roughness and ank wear.
Variance source

Degree of freedom (DoF)

Sum of squares (SS)

Mean square (MS)

F ratio

Contribution rate (%)

Ra
A
B
C
Error
Total

1
2
2
12
17

0.004050
0.014044
0.116578
0.005689
0.140361

0.004050
0.007022
0.058289
0.000474

8.54
14.81
122.95

2.55
9.33
82.38
5.74
100

Vb
A
B
C
Error
Total

1
2
2
12
17

0.0102722
0.0212333
0.0076
0.0029444
0.04205

0.0102722
0.0106167
0.0038000
0.0002454

41.86
43.27
15.49

23.84
49.33
16.91
9.92
100

Bold values shows the most inuential factor.

Here Ral and Vbl show the predictive equations of surface

roughness and ank wear respectively. In Fig. 7, the comparison of actual test results and predicted values which
were obtained by the linear regression model are given.
R2 values of the equations which were obtained by linear
regression model for Ra and Vb were found to be 92.27%
and 89.85% respectively.
The predictive equations for the quadratic regression of
surface roughness and ank wear are given below:

on ank wear were found to be 23.84%, 49.33% and 16.91%

respectively. This showed that the most effective factor on
ank wear was cutting speed (factor B, 49.33%). The percent of error was considerably low at 5.74% and 9.92% for
Ra and Vb respectively.
4.4. Regression analysis of surface roughness and ank wear
Regression analyses are used for the modeling and
analyzing of several variables where there is relationship
between a dependent variable and one or more independent variables [29]. In this study, the dependent variables
are surface roughness (Ra) and ank wear (Vb), whereas
the independent variables are cutting tool (Ct), cutting
speed (V) and feed rate (f). In obtaining predictive
equations for the surface roughness and tool wear, regression analysis was used. These predictive equations were
made for both linear and quadratic regression models.
The predictive equations which were obtained by the
linear regression model of surface roughness and ank
wear are given below.

Raq 0:674444 0:0766667Ct  0:00183333V  3:16667f

1:48148e  005V 2 37:037f 2  0:000555556CtV
 0:333333Ctf  0:0166667Vf
R  Sq 98:04%
R  Sqadj 96:29%

Vbq 0:691667  0:147778Ct  0:00555556V

 2:16667f 1:94444e  005V 2 5:55556f 2
0:000611111CtV 0:222222Ctf 0:0111111Vf
R  Sq 97:62%

Vbl 0:0366667  0:0477778Ct 0:00136111V 0:833333f

Sq 89:85%

R  Sqadj 87:67%

0.8

0.35

0.30

Actual Vb

0.7

Actual Ra

R  Sqadj 95:50%

Here Raq and Vbq show the predictive equations for surface
roughness and ank wear. In Fig. 8 test results and the
comparison of predicted values which were obtained by
the quadratic regression model are shown. As seen from
the gure, there is a very good relation between predicted
values and test results. The R2 values of the equations obtained by the quadratic regression model for Ra and Vb

Ral 0:358889  0:03Ct  0:00111111V 3:22222f

R  Sq 92:27% R  Sqadj 90:62%

0.6

0.5

0.25

0.20

0.15

0.4

0.4

0.5

0.6

Predicted Ra

0.7

0.8

0.10
0.10

0.15

0.20

0.25

Predicted Vb

Fig. 7. Comparison of the linear regression model with experimental results for Ra and Vb.

0.30

0.35

26

T. Kvak / Measurement 50 (2014) 1928

0.8

0.35

0.30

Actual Vb

Actual Ra

0.7

0.6

0.25

0.20

0.5
0.15

0.4
0.4

0.5

0.6

0.7

0.8

0.10
0.10

0.15

Predicted Ra

0.20

0.25

0.30

0.35

Predicted Vb

Fig. 8. Comparison of the quadratic regression model with experimental results for Ra and Vb.

were found to be 98.04% and 97.62% respectively. Hence,

more intensive predicted values were obtained by the quadratic regression model as compared to the linear regression model. As a result, the quadratic regression model
was shown to be successfully for the estimation of surface
roughness and ank wear.
4.5. Estimation of optimum surface roughness and ank wear
With the Taguchi optimization technique, a conrmation experiment was required to be conducted for validating of the optimized condition [18]. In the estimation of
optimum surface roughness and ank wear, Eqs. (6) and
(7) were used respectively.

average of all of the Ra and Vb values obtained from the

experimental study (Table 8). As a result of the calculations, it was estimated that Raopt = 0.433 lm and
Vbopt = 0.143 mm.
Whether the system had realized the optimization
accurately enough needed to be evaluated. For this purpose, the following equations were used in the specication of the condence interval (Cl) for estimated surface
and ank wear [30]:

CIRa;Vb

s


1
1
F a;1;fe V e

neff R

and

N
1 T dof

Raopt A2  T Ra B3  T Ra C 1  T Ra T Ra

neff

Vbopt A2  T Vb B1  T Vb C 1  T Vb T Vb

Here, Fa,1,fe is the F ratio at a 95% condence, a is the significance level, fe is the degrees-of-freedom of error, Ve is error
variance, neff is the effective number of replications, R is the
number of replications for conrmation experiments (Eq.
(8)). N is the total number of experiments, and Tdof is the
total main factor degrees of freedom (Eq. (9)).
F0.05,1,12 = 4.747 (from F test table), VeRa = 0.000474 and
VeVb = 0.0002454 (Table 6), R = 3 (Eq. (8)). N = 3, Tdof = 5
and neff = 3 (Eq. (9)).
By using the Eqs. (8) and (9) the condence intervals
were calculated as CIRa = 0.039 and CIVb = 0.028 The
estimated average optimal surface roughness and ank
wear with the condence interval at 95% condence is:

Here, (A2, B3, C1) and (A2, B1, C1) represent the optimum level average values of surface roughness (Raopt) and ank
wear (Vbopt) respectively (Table 3). TRa and TVb state the
Table 8
Mean response table for Ra and Vb factor.
Levels

Level 1
Level 2
Level 3
Delta

Control factors
Surface roughness (Ra)

Flank wear (Vb)

0.5822
0.5522

0.300

0.6050
0.5583
0.5388
0.0667

0.4817
0.5450
0.6750
0.1933

0.2522
0.2044

0.0478

0.1933
0.2167
0.2750
0.0817

0.2050
0.2250
0.2550
0.0500

Raopt  CIRa  < Raexp < Raopt CIRa ; i:e:; 0:433  0:039
< 0:45 < 0:433 0:039 0:394 < 0:45

Bold values shows the Ra and Vb values obtained from the experimental
study.

< 0:472

Table 9
Predicted values and conrmation test results by Taguchi method and regression equations.
Level

Ra (lm)
A2B3C1 (Optimum)
A1B1C3 (Random)
Vb (mm)
A2B1C1 (Optimum)
A1B1C1 (Random)

For Taguchi method

For linear regression equations

For quadratic regression equations

Exp.

Pred.

Error (%)

Exp.

Pred.

Error (%)

Exp.

Pred.

Error (%)

0.45
0.76

0.433
0.722

3.78
5.00

0.45
0.76

0.422
0.712

6.22
6.32

0.45
0.76

0.449
0.739

0.22
2.76

0.14
0.22

0.143
0.191

2.14
13.18

0.14
0.22

0.139
0.186

0.71
15.45

0.14
0.22

0.144
0.216

2.86
1.82

T. Kvak / Measurement 50 (2014) 1928

Vbopt  CIVb  < Vbexp < Vbopt CIVb ; i:e:; 0:143  0:028
< 0:14 < 0:143 0:028 0:115 < 0:14
< 0:171
The Raexp and Vbexp values, obtained from the experimental
study stayed within the condence interval limits. Thus,
the system optimization for surface roughness and ank
wear was achieved using the Taguchi method at a signicance level of 0.05.

27

All of these results showed that the Taguchi method

was a reliable methodology for the reduction of machining
time and manufacturing costs in the milling of Hadeld
steel. In the future, the results obtained can be used for
academic research as well as for industrial applications.
Further studies could consider other factors such as the
depth of cut, cutting tool geometry, coating materials
(TiN, TiN/TiAlN, TiCrN, etc.), cutting tool materials, chip
breaker, nose radius and lubricants, all of which affect
the surface roughness and tool wear.

4.6. Conrmation tests

Conrmation tests of the control factors were made for
the Taguchi method and regression equations at optimum
and random levels. In Table 9, the comparison of test
results and the predicted values obtained by using the
Taguchi method and regression equations (Eqs. (2)(5))
are given. The predicted values and the experimental values are very close to each other. For reliable statistical
analyses, error values must be less than 20% [29]. Although
the error percentages calculated in the ank wear are
higher than the ones in surface roughness, they are within
acceptable limits. Therefore, the results obtained from the
conrmation tests reect successful optimization.

5. Conclusion
In this study, the Taguchi method was used to determine optimal machining parameters in the milling of
Hadeld steel with PVD- and CVD-coated carbide inserts
under dry milling conditions. The experimental results
were evaluated using ANOVA. The following conclusions
may be drawn:
 The optimum levels of the control factors for minimizing the surface roughness and ank wear using S/N
rates were determined. The optimal conditions for surface roughness and ank wear were observed at A2B3C1
(i.e., cutting tool = CVD, cutting speed = 150 m/min and
feed rate = 0.09 mm/rev) and A2B1C1 (i.e., cutting
tool = CVD, cutting speed = 90 m/min and feed
rate = 0.09 mm/rev), respectively.
 CVD TiCN/Al2O3-coated carbide inserts exhibited better
performance than PVD TiAlN-coated carbide inserts and
could be recommended for use in the milling of
Hadeld steel.
 According to the results of statistical analyses, it was
found that the feed rate was the most signicant
parameter for surface roughness with a percentage contribution of 82.38% and that the cutting speed was the
most signicant parameter for ank wear with a percentage contribution of 49.33%.
 Developed quadratic regression models demonstrated a
very good relationship with high correlation coefcients (Ra = 0.962 and Vb = 0.955) between the measured and predicted values for surface roughness and
ank wear.
 According to the conrmation test results, measured
values were within the 95% condence interval.

References
[1] D. Canadinc, H. Sehitoglu, H.J. Maier, Y.I. Chumlyakov, Strain
hardening behavior of aluminum alloyed Hadeld steel single
crystals, Acta Materialia 53 (2005) 18311842.
[2] V.G. Gavriljuk, A.I. Tyshchenko, O.N. Razumov, Y.N. Petrov, B.D.
Shanina, H. Berns, Corrosion-resistant analogue of Hadeld steel,
Mater. Sci. Eng.: A 420 (2005) 4754.
[3] L. Niu, Y. Xu, H. Wu, Reinforcement of Hadeld steel matrix by
oriented high-chromium cast iron bars, Fatigue Fracture Eng. Mater.
Struct. 0 (2010) 17.
[4] J.T. Horng, N.M. Liu, K.T. Chiang, Investigating the machinability
evaluation of Hadeld steel in the hard turning with Al2O3/TiC mixed
ceramic tool based on the response surface methodology, J. Mater.
Processing Technol. 208 (2008) 532541.
[5] S. Thamizhmanii, M. Nagib, H. Sulaiman, Performance of deep
cryogenically treated and non-treated PVD inserts in milling, J.
Achievements Mater. Manuf. Eng. 49 (2) (2011) 460466.
[6] M. Ibrahim Sadik, T. Myrtveit, The Performance of PVD Coated Grade
in Milling of ADI 800, Int. J. Aerospace Mech. Eng. 5 (1) (2011) 5558.
[7] Q. Niu, M. Chen, W. Ming, Q. An, Evaluation of the performance of
coated carbide tools in face milling TC6 alloy under dry condition,
Int. J. Adv. Manuf. Technol. 64 (2013) 623631.
_ Toktas, G. Sur, The experimental
[8] M. Nalbant, H. Gkkaya, I.
investigation of the effects of uncoated, PVD- and CVD-coated
cemented carbide inserts and cutting parameters on surface
roughness in CNC turning and its prediction using articial neural
networks, Robotics Comput.-Integrated Manuf. 25 (2009) 211223.
[9] E.O. Ezugwu, C.I. Okele, Behavior of coated carbide tools in high
speed machining of a nickel base alloy, Tribology Trans. 45 (1) (2002)
122126.
[10] A. Jawaid, S. Koksal, S. Sharif, Wear behavior of PVD and CVD coated
carbide tools when face milling Inconel 718, Tribology Trans. 43 (2)
(2000) 325331.
[11] A. Jawaid, S. Sharif, S. Koksal, Evaluation of wear mechanisms of
coated carbide tools when face milling titanium alloy, J. Mater.
Processing Technol. 99 (2000) 266274.
[12] E.O. Ezugwu, Z.M. Wang, C.I. Okeke, Tool life and surface integrity
when machining Inconel 718 with PVD- and CVD-Coated tools,
Tribology Trans. 42 (2) (1999) 353360.
[13] I. Asiltrk, M. unkas, Modeling and prediction of surface roughness
in turning operations using articial neural network and multiple
regression method, Expert Syst. Appl. 38 (2011) 58265832.
[14] T. zel, Y. Karpat, Predictive modeling of surface roughness and tool
wear in hard turning using regression and neural networks, Int. J.
Machine Tools Manuf. 45 (2005) 467479.
[15] M.Y. Wang, H.Y. Chang, Experimental study of surface roughness in
slot end milling, Int. J. Machine Tools Manuf. 44 (2004) 5157.
[16] C. Fetecau, F. Stan, Study of cutting force and surface roughness in
the turning of polytetrauoroethylene composites with a
polycrystalline diamond tool, Measurement 45 (2012) 13671379.
[17] M.S. Phadke, Quality Engineering Using Robust Design, Prentice Hall,
Englewood Cliffs, New Jersey, 1989.
[18] N. Mandal, B. Doloi, B. Mondal, R. Das, Optimization of ank wear
using Zirconia Toughened Alumina (ZTA) cutting tool: taguchi
method and regression analysis, Measurement 44 (2011) 2149
2155.
[19] M. Nalbant, H. Gkkaya, G. Sur, Application of Taguchi method in the
optimization of cutting parameters for surface roughness in turning,
Mater. Des. 28 (2007) 13791385.
[20] J.A. Ghani, I.A. Choudhury, H.H. Hassan, Appplication of Taguchi
method in the optimization of end milling parameters, J. Mater.
Processing Technol. 145 (2004) 8492.

28

T. Kvak / Measurement 50 (2014) 1928

[21] Y.K. Yang, M.T. Chuang, S.S. Lin, Optimization of dry machining
parameters for high-purity graphite in end milling process via
design of experiments methods, J. Mater. Processing Technol. 209
(2009) 43954400.
[22] J.T. Horng, N.M. Liu, K.T. Chiang, Investigating the machinability
evaluation of Hadeld steel in the hard turning with Al2O3/TiC
mixed ceramic tool based on the response surface methodology, J.
Mater. Processing Technol. 208 (2008) 532554.
[23] J.P. Davim, Study of drilling metalmatrixcomposites based on the
Taguchi techniques, J. Mater. Processing Technol. 132 (2003) 3250
3254.
[24] K. Palanikumar, Experimental investigation and optimisation in
drilling of GFRP composites, Measurement 44 (2011) 21382148.
[25] I. Asilturk, H. Akkus, Determining the effect of cutting parameters on
surface roughness in hard turning using the Taguchi method,
Measurement 44 (2011) 16971704.

[26] O. Koksoy, Z.F. Muluk, Solution to the Taguchis problem with

correlated responses, gazi university, J. Sci. 17 (1) (2004) 5970.
[27] A. Gupta, H. Singh, A. Aggarwal, Taguchi-fuzzy multi output
optimization (MOO) in high speed CNC turning of AISI P-20 tool
steel, Expert Syst. Appl. 38 (2011) 68226828.
[28] K.R. Ranjit, A Primer on the Taguchi Method, Competitive
Manufacturing Series, Van Nostrand Reinhold, New York, 1990.
[29] M.H. Cetin, B. Ozcelik, E. Kuram, E. Demirbas, Evaluation of vegetable
based cutting uids with extreme pressure and cutting parameters
in turning of AISI 304L by Taguchi method, J. Cleaner Prod. 19 (2011)
20492056.
[30] A. Dvivedi, P. Kumar, Surface quality evaluation in ultrasonic drilling
through the Taguchi technique, Int. J. Adv. Manuf. Technol. 34 (2007)
131140.