1-s2.0-s0263224113005903-main
TRANSCRIPT
-
7/25/2019 1-s2.0-S0263224113005903-main
1/11
Optimization of surface roughness, cutting force and tool wear
of nitrogen alloyed duplex stainless steel in a dry turning
process using Taguchi method
D. Philip Selvaraj a,, P. Chandramohan b, M. Mohanraj c
a School of Mechanical Sciences, Karunya University, Coimbatore 641114, Indiab Department of Mechanical Engineering, Professional Group of Institutions, Palladam 641662, Indiac Department of Mechanical Engineering, Hindusthan College of Engineering and Technology, Coimbatore 641032, India
a r t i c l e i n f o
Article history:
Received 17 January 2013
Received in revised form 24 May 2013
Accepted 25 November 2013
Available online 5 December 2013
Keywords:
Duplex stainless steel
Dry turning
Taguchi method
S/Nratio
ANOVA
Optimization
a b s t r a c t
In this work, the dry turning parameters of two different grades of nitrogen alloyed duplex
stainless steel are optimized by using Taguchi method. The turning operations were carried
out with TiC and TiCN coated carbide cutting tool inserts. The experiments were conducted
at three different cutting speeds (80, 100 and 120 m/min) with three different feed rates
(0.04, 0.08 and 0.12 mm/rev) and a constant depth of cut (0.5 mm). The cutting parameters
are optimized using signal to noise ratio and the analysis of variance. The effects of cutting
speed and feed rate on surface roughness, cutting force and tool wear were analyzed. The
results revealed that the feed rate is the more significant parameter influencing the surface
roughness and cutting force. The cutting speed was identified as the more significant
parameter influencing the tool wear. Tool wear was analyzed using scanning electron
microscope image. The confirmation tests are carried out at optimum cutting conditions.The results at optimum cutting condition are predicted using estimated signal to noise
ratio equation. The predicted results are found to be closer to experimental results within
8% deviations.
2013 Elsevier Ltd. All rights reserved.
1. Introduction
In the past few decades, the applications of stainless
steel materials have been increased enormously in various
engineering fields. The combination of good corrosion
resistance, wide range of strength levels, good formabilityand aesthetically pleasing appearance have made stainless
steels as a good choice for wide range of applications. But,
their machinability is more difficult compared to other al-
loy steels due to low thermal conductivity, high built-up
edge (BUE) formation tendency and high deformation
hardening. Duplex stainless steel (DSS) combines the ben-
efits of both ferritic stainless steel (FSS) and austenitic
stainless steel (ASS) by proper balancing of ferrite and aus-
tenite. The duplex structure improves stress-corrosion
cracking resistance, compared to ASSs, and improves the
toughness and ductility compared to FSSs [1]. Modern
DSS grades tend to be difficult to machine, by virtue of
their higher austenite and nitrogen contents. The use ofDSSs has been increased due to their high strength, higher
pitting corrosion resistance equivalent and stress corrosion
resistance [2]. DSSs are extensively being used in many
industrial sectors like desalination, chemical tankers, pres-
sure vessels, storage tanks, machinery in the pulp and pa-
per industry, and also in civil engineering applications.
They have higher contents of chromium and lower con-
tents of nickel and molybdenum and they are excellent
engineering materials[3].
Agrawal et al. [4] have been studied the machining
characteristics of cast ASSs with reference to cutting force
0263-2241/$ - see front matter 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.measurement.2013.11.037
Corresponding author. Tel.: +91 9994650780; fax: +91 4222615615.
E-mail addresses: [email protected], philipselvaraj@karunya.
edu(D. P. Selvaraj).
Measurement 49 (2014) 205215
Contents lists available at ScienceDirect
Measurement
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c at e / m e a s u r e m e n t
http://dx.doi.org/10.1016/j.measurement.2013.11.037mailto:[email protected]:[email protected]:[email protected]:[email protected]://dx.doi.org/10.1016/j.measurement.2013.11.037http://www.sciencedirect.com/science/journal/02632241http://www.elsevier.com/locate/measurementhttp://www.elsevier.com/locate/measurementhttp://www.sciencedirect.com/science/journal/02632241http://dx.doi.org/10.1016/j.measurement.2013.11.037mailto:[email protected]:[email protected]:[email protected]://dx.doi.org/10.1016/j.measurement.2013.11.037http://-/?-http://-/?-http://crossmark.crossref.org/dialog/?doi=10.1016/j.measurement.2013.11.037&domain=pdf -
7/25/2019 1-s2.0-S0263224113005903-main
2/11
requirement, tool rake-face wear and chip characteristics.
It has been reported that the composition of stainless steel
work piece material influences the machinability. The wear
mechanisms when turning X5CrMnN18 ASS materials
were catastrophic failure at tool nose due to high cutting
forces and sharp edge chipping. The addition of nitrogen
to ASS increases the strength and decreases the machin-
ability [5]. In powder metallurgy produced DSSs, the
machining difficulties are increased due to the presence
of more hard oxide particles, high strength and work hard-
ening rate[6]. The surface roughness values were found to
decrease with increasing cutting speed when turning AISI
304 ASS. This can be attributed to the presence of BUE at
lower cutting speeds. The poor performance of the tool
was due to higher influence of the heat on the cutting tool
and less efficient heat dissipation at the lower cutting
speeds[7]. Similarly, Ciftci[8]investigated the machining
characteristics of ASS using chemical vapor deposition
coated carbide cutting tools. His results reported that cut-
ting speed is highly influencing the surface roughness val-
ues. In another work, Senthikumar et al.[9]evaluated the
tool life of alumina based ceramic cutting tool for machin-
ing hardened martensitic stainless steel (MSS). It has been
reported that the flank wear affects the tool life at lower
cutting speed, whereas, crater wear or notch wear affects
the tool life at higher cutting speed. Noordin et al. [10]
recommended higher insert radius, low feed rate and low
depth of cut to obtain better surface finish under dry turn-
ing operation. The flank wear rate of Cubic Boron Nitride
(CBN) tool was more compared to Polycrystalline Cubic
Boron Nitride (PCBN) tool while machining MSS due to
more abrasion and diffusion [11]. Krolczyk et al. [12]
developed a mathematical model using response surface
method (RSM) to predict the surface roughness of DSS in
dry turning. The cutting parameters considered were cut-
ting speed, feed rate and depth of cut. They found that
the feed rate is the main influencing factor on the surface
roughness. Bouzid Sai et al. [13]investigated the residual
stresses, microstructure, surface roughness and micro
hardness of carbon steels and DSS materials during milling
operations. They found that a high value cutting speed
with less feed rate has improved the quality of the machin-
ing surface. Depth of cut has less influence on the surface
characteristics.
Muthukrishnan and Davim[14]optimized the machin-
ing parameters of Al/SiC metal matrix composites using
ANOVA and ANN analysis and reported that feed rate has
high physical influence on the surface roughness. Similarly,
Palanikumar [15] used Taguchi method to optimize the
drilling parameters of glass fiber-reinforced plastics com-
posites. It has been reported that feed rate was the more
influential parameter than spindle speed. In another work,
Mandal et al.[16]applied Taguchi method and regression
analysis to assess the machinability of AISI 4340 steel with
newly developed Zirconia Toughened Alumina ceramic in-
serts. Their results reported that the main contributing fac-
tors for the tool flank wear are depth of cut and the cutting
speed. The feed rate has less influence on the flank wear.
Similarly, Asiltrk and Akkus [17] conducted dry turning
tests on hardened AISI 4140 steel (51 HRC) with coated
carbide cutting tools. They used Taguchi method to opti-
mize the cutting parameters. Their results reported that
the feed rate has the more significant effect on surface
roughness (Ra- roughness average and Rz-average maxi-
mum height of the profile). The cited literatures confirmed
that limited investigations have been carried out on the
machining characteristics of nitrogen alloyed DSS. Hence,
an attempt has been made in this work to optimize the cut-
ting parameters to minimize the surface roughness, cutting
force and tool wear during dry turning operations of nitro-
gen alloyed DSS.
2. Taguchi method
Taguchi method is widely used for optimizing indus-
trial/production processes. The Taguchi design optimiza-
tion method can be divided into three stages: (a) system
design, (b) parameter design and (c) tolerance design.
Among the three stages, the parameter design stage is con-
sidered to be the important stage [1820]. The steps fol-
lowed in the Taguchi parameter design are: selecting the
proper orthogonal array (OA); running experiments basedon the OA; analyzing data; identifying the optimum condi-
tion; and conducting confirmation runs [21]. Many
researchers have been used Taguchi method to optimize
the various machining operations like turning, end milling,
drilling, etc. in various alloys[2228].
3. Experimental details and data analysis
The experiments are designed using Taguchis design of
experiment method. This research work was carried out at
Centre for Research in Design and Manufacturing engineer-
ing (CRDM), Karunya University, Coimbatore, India. The
experimental data are analyzed by using the signal to noiseratio (S/Nratio) and the analysis of variance (ANOVA). The
S/Nratio analysis is used to find out the optimum machin-
ing conditions. The ANOVA analysis is used to find the per-
centage contribution of the cutting speed and feed rate on
surface roughness, cutting force and tool wear.
3.1. Work piece material
The work piece materials selected for investigation are
the cast DSS ASTM A 995 grade 5A and grade 4A with the
compositions as shown inTable 1. The mechanical proper-
ties of the material investigated are given in Table 2. The
diameter and length of the work piece used in the experi-
mentation are 80 mm and 300 mm, respectively. One end
of the work piece is held in a chuck and other end is sup-
ported with a tailstock. Generally the length-to-diameter
ratio used is 3:16:1, while machining with the tail stock
to prevent deflection. The length-to-diameter ratio of the
specimen used in this work is 3.75:1, which falls within
the acceptable range.
3.2. Experimental procedure
The turning tests are conducted on a medium duty Kir-
loskar Turn master-35 Lathe with a variable speed be-
tween 100 and 1500 rpm and a power rating of 2.2 kW.
206 D. P. Selvaraj et al./ Measurement 49 (2014) 205215
http://-/?- -
7/25/2019 1-s2.0-S0263224113005903-main
3/11
The cutting tools used are carbide inserts (Taegu Tec make)
coated with TiC and TiCN with a specification of SNMG
120408 MT TT5100. The inserts are clamped on a pin and
hole type tool holder (Taegu Tec make) with a specification
of PSBNR 2525M12. The surface roughness is measured
using a TR-100 surface roughness tester. Cutting force ismeasured using Kistler piezoelectric dynamometer (model
9257B). Cutting force measurements are interfaced with a
computer using data acquisition system. The tool wear is
observed using scanning electron microscope (SEM) JEOL
JSM-6390 model. The schematic diagram of the experi-
mental set-up is depicted inFig. 1. Dry machining is more
popular in manufacturing as a means of reducing overhead
costs and protecting the environment[29]. It has great sig-
nificance for the factors of both economics and environ-
ment [30]. Hence, the experiments were carried out
under dry condition (without using cutting fluid).
3.3. Experimental plan
The cutting parameters for turning operation are cut-
ting speed, feed rate and depth of cut. The influence of cut-
ting speed and feed rate are more significant compared to
depth of cut[12,13,27]. Hence, cutting speed and feed rate
are selected as the main cutting parameters in this study.
Depth of cut was maintained at constant value in the turn-
ing operation [31] and in the milling operation[32]. For
turning DSS material using carbide cutting tool, the cutting
speed range is 80120 m/min as per the standard pub-
lished by International Molybdenum Association (IMOA
1999). The feed rate range is selected based on the avail-ability of the range of feed in the machine utilized for
experimentation. The depth of cut is taken constant as
0.5 mm based on the investigation reported by Thakur
et al. [31]. The experiments are planned using the Tagu-
chis OA. The machining tests were conducted according
to a 3-level and 2-factor L9OA. The experiments were con-
ducted at three different cutting speeds (80, 100 and
120 m/min) with three different feed rates (0.04, 0.08
and 0.12 mm/rev) and a constant depth of cut (0.5 mm).
The cutting parameters and their levels are indicated in
Table 3. The experimental layout for the L9 OA is shown
inTable 4.
3.4. Analysis of the S/N ratio
TheS/N ratio is the ratio of the mean to the standard
deviation. It is used to measure the quality characteristic
deviating from the desired value. TheS/Nratio (g) is given
by the following equation discussed by Yang and Tarng
[21].
g 10logM:S:D 1
Here, M.S.D is the mean square deviation for the output
characteristic. To obtain optimal cutting performance,
the-lower-the-better quality characteristic for surface
roughness, cutting force and tool wear must be taken.
The M.S.D. for the-lower-the-better quality characteristic
of surface roughness, cutting force and tool wear can be gi-
ven by the following equations discussed by Yang and
Tarng[21].
M:S:D 1
m
Xm
i1
S2i
2
M:S:D 1
m
Xm
i1
F2i
3
M:S:D 1m
Xm
i1
T2i
4
Here,m is the number of tests, Si,FiandTiare the values of
the surface roughness, cutting force and tool wear, respec-
tively for theith test. The estimatedS/Nratio (g) has been
used to predict and verify the quality characteristic at the
optimal level. The estimated S/Nratio g at the optimal level
Table 1
Chemical composition of ASTM A 995 grade 5A and 4A DSS (wt%).
Alloy C Si Mn S P Cr Ni Mo Cu N Fe
5A 0.028 0.67 0.87 0.005 0.028 25.10 6.63 4.16 0.17 Bal.
4A 0.028 0.65 0.71 0.006 0.027 22.16 5.66 3.33 0.14 0.24 Bal.
Table 2Mechanical properties of ASTM A 995 grade 5A and 4A DSS.
Alloy Tensile
strength (MPa)
Yield strength
(MPa)
Elongation
(%)
Hardness
(BHN)
5A 741 546 32.2 223
4A 732 595 30.2 212
Tool
Work Piece TailstockChuck
KistlerDynamometer
Chargeamplifier
Computer
Fig. 1. Schematic diagram of experimental set-up.
Table 3
Cutting parameters and their levels.
Symbol Cutting parameters Level 1 Level 2 Level 3
V Cutting speed (m/min) 80 100 120
F Feed rate (mm/rev) 0.04 0.08 0.12
D. P. Selvaraj et al. / Measurement 49 (2014) 205215 207
http://-/?- -
7/25/2019 1-s2.0-S0263224113005903-main
4/11
of the design parameters can be calculated by the follow-
ing equation discussed by Yang and Tarng[21].
g gm
Xo
i1
gi
gm
5
Here, gm is the total mean S/Nratio, gi is the meanS/N
ratio at the optimal level, and o is the number of the main
design parameters that affect the quality characteristic.
3.5. Analysis of the variance
ANOVA is carried out to identify the design parameters
that significantly affect the response. The total sum of the
squared deviations (SST) is calculated by using the follow-
ing equation discussed by Yang and Tarng[21].
SSTXn
i1gi gm
2
6
Here, n is the number of experiments, gi is the mean S/N
ratio for the ith experiment and gm is the total mean S/N
ratio. The two sources of the SST are: the sum of the
squared deviations (SSd) due to each design parameter
and the sum of the squared error (SSe).
4. Results and discussion
The experimental results of surface roughness, cutting
force and tool wear with their corresponding S/N ratio
are shown inTables 57, respectively for 5A and 4A Grade
DSS. The meanS/Nratio for cutting speed at level 1 is cal-
culated by averaging theS/Nratios for the experiments 1
3. The mean S/Nratio for feed rate at level 1 is calculated
by averaging the S/N ratios for the experiments 1, 4 and
7. Similarly, the mean S/Nratio for cutting speed and feed
rate at levels 2 and 3 are calculated.
4.1. S/N ratio and ANOVA results
The S/Nresponse table for surface roughness, cutting
force and tool wear of 5A and 4A grade DSS are shown in
Tables 8 and 9, respectively. The S/Nresponse graph for
surface roughness of 5A and 4A grade DSS is depicted in
Fig. 2. In the S/N response graph, V1 stands for cutting
speed at level 1 (80 m/min), V2 stands for cutting speed
at level 2 (100 m/min),V3 stands for cutting speed at level
3 (120 m/min),F1stands for feed rate at level 1 (0.04 mm/
rev),F2stands for feed rate at level 2 (0.08 mm/rev) andF3stands for feed rate at level 3 (0.12 mm/rev). The greater
S/Nratio corresponds to the smaller variance of the output
characteristic around the desired value. From Fig. 2, the
higherS/Nratio for surface roughness of 5A and 4A grade
are obtained at cutting speed level 2 and feed rate level
Table 4
Experimental layout using an L9 orthogonal
array.
Experimental
number
Cutting parameter
level
Cutting
speed (A)
Feed
rate (B)
1 1 1
2 1 2
3 1 3
4 2 1
5 2 2
6 2 3
7 3 1
8 3 2
9 3 3
Table 5
Experimental results for surface roughness and S/Nratio of 5A and 4A grade
DSS.
S.
no.
Cutting
Speed
(m/min)
Feed rate
(mm/rev)
Surface
roughness Ra
(lm)
S/Nratio (dB)
5A
Grade
4A
Grade
5A
Grade
4A
Grade
1 80 0.04 1.20 0.60 1.58 4.44
2 80 0.08 1.25 0.68 1.94 3.35
3 80 0.12 1.32 0.85 2.41 1.41
4 100 0.04 1.05 0.53 0.42 5.51
5 100 0.08 1.17 0.57 1.36 4.88
6 100 0.12 1.25 0.64 1.94 3.88
7 120 0.04 1.16 0.58 1.29 4.73
8 120 0.08 1.24 0.60 1.87 4.44
9 120 0.12 1.30 0.76 2.28 2.38
Table 6
Experimental results for cutting force and S/Nratio of 5A and 4A grade DSS.
S.no. Cuttingspeed
(m/min)
Feed rate(mm/rev) Cutting forceFc (N) S/Nratio (dB)
5A
Grade
4A
Grade
5A
Grade
4A
Grade
1 80 0.04 44.5 33.2 32.97 30.42
2 80 0.08 53.6 39.3 34.58 31.89
3 80 0.12 58.9 45.8 35.40 33.22
4 100 0.04 36.9 28.2 31.34 29.00
5 100 0.08 44.1 34.4 32.89 30.73
6 100 0.12 50.9 38.1 34.13 31.62
7 120 0.04 35.2 25.7 30.93 28.20
8 120 0.08 38.7 28.8 31.75 29.19
9 120 0.12 46.5 35.6 33.35 31.03
Table 7
Experimental results for tool wear and S/Nratio of 5A and 4A grade DSS.
S.
no.
Cutting
speed
(m/min)
Feed rate
(mm/rev)
Tool wear Vb
(mm)
S/Nratio (dB)
5A
Grade
4A
Grade
5A
Grade
4A
Grade
1 80 0.04 0.113 0.102 18.94 19.83
2 80 0.08 0.122 0.115 18.27 18.79
3 80 0.12 0.134 0.130 17.46 17.72
4 100 0.04 0.146 0.141 16.71 17.02
5 100 0.08 0.184 0.176 14.70 15.09
6 100 0.12 0.204 0.201 13.81 13.94
7 120 0.04 0.268 0.252 11.44 11.97
8 120 0.08 0.329 0.319 9.66 9.92
9 120 0.12 0.353 0.348 9.04 9.17
208 D. P. Selvaraj et al./ Measurement 49 (2014) 205215
http://-/?-http://-/?-http://-/?-http://-/?- -
7/25/2019 1-s2.0-S0263224113005903-main
5/11
1. Therefore, the optimal cutting parameters for surface
roughness of 5A and 4A grade DSS are the cutting speed
at level 2 (100 m/min) and the feed rate at level 1
(0.04 mm/rev).
TheS/Nresponse graph for cutting force of 5A and 4A
grade DSS is shown in Fig. 3. FromFig. 3, the higher S/N
ratio for cutting force of 5A and 4A grade are obtained at
cutting speed level 3 and feed rate level 1. Therefore, the
optimal cutting parameters for cutting force of 5A and 4A
grade DSS are the cutting speed at level 3 (120 m/min)
and the feed rate at level 1 (0.04 mm/rev).
The S/Nresponse graph for tool wear of 5A and 4A grade
DSS is shown inFig. 4. FromFig. 4, the higher S/Nratio for
tool wear of 5A grade are obtained at cutting speed level 1
and feed rate level 1. Hence the optimal cutting parameters
for tool wear of 5A and 4A grade DSS are the cutting speed
at level 1 (80 m/min) and the feed rate at level 1 (0.04 mm/
rev).
Similarly, Table 10 shows that the ANOVA results of
surface roughness, cutting force and tool wear of 5A grade
DSS. It was observed that the feed rate is the more signifi-
cant cutting parameter affecting the surface roughness.
The cutting parameters influencing the surface roughness
are feed rate followed by cutting speed. ANOVA results
showed that feed rate and cutting speed are affecting the
surface roughness of 5A grade DSS by approximately 64%,
and 31%, respectively. It was observed that the cutting
speed and the feed rate are the significant cutting parame-
ters affecting the cutting force. The feed rate and cutting
speed affecting the cutting force of 5A grade DSS by
approximately 53% and 45%, respectively. It was observed
that the cutting speed is the more significant and the feed
rate is the less significant cutting parameter affecting the
tool wear. The cutting speed and the feed rate affect the
tool wear of 5A grade DSS by approximately 92%, and 7%,
respectively.
Similarly, Table 11 shows that the ANOVA results of
surface roughness, cutting force and tool wear of 4A grade
Table 8
S/Nresponse table for surface roughness, cutting force and tool wear for 5A
grade DSS.
Cutting parameter MeanS/Nratio (dB) Maxmin
Level 1 Level 2 Level 3
Surface roughness
Cutting speed 1.98 1.24 1.81 0.74
Feed rate 1.10 1.72 2.21 1.11
Cutting force
Cutting speed 34.32 32.79 32.01 2.31
Feed rate 31.75 33.07 34.29 2.54
Tool wear
Cutting speed 18.22 15.07 10.05 8.17
Feed rate 15.77 14.21 13.44 2.33
Table 9
S/Nresponse table for surface roughness, cutting force and tool wear for 4A
grade DSS.
Cutting parameter MeanS/Nratio (dB) Maxmin
Level 1 Level 2 Level 3
Surface roughness
Cutting speed 3.07 4.76 3.85 1.69
Feed rate 4.89 4.22 2.56 2.33
Cutting force
Cutting speed 31.84 30.45 29.47 2.37
Feed rate 29.21 30.60 31.96 2.75
Tool wear
Cutting speed 18.78 15.35 10.35 8.43
Feed rate 16.27 14.60 13.61 2.66
Fig. 2. S/Ngraph for surface roughness-5A and 4A grade DSS.
Fig. 3. S/Ngraph for cutting force-5A and 4A grade DSS.
Fig. 4. S/Ngraph for tool wear 5A and 4A grade DSS.
D. P. Selvaraj et al. / Measurement 49 (2014) 205215 209
http://-/?- -
7/25/2019 1-s2.0-S0263224113005903-main
6/11
DSS. The feed rate and cutting speed affecting the surface
roughness of 4A grade DSS by about 63% and 31%, respec-
tively. The feed rate and cutting speed affect the cutting
force of 4A grade DSS by about 56% and 42%, respectively.
The cutting speed and the feed rate affect the tool wear of
4A grade DSS by about 91% and 9%, respectively.
4.2. Comparison of predicted and experimental results at the
optimal cutting conditions
The results at optimum cutting condition are predicted
using estimated signal to noise ratio equation discussed by
Yang and Tarng[21]. The predicted and experimental sur-
face roughness, cutting force and tool wear of 4A and 5A
grade DSS using the optimal cutting parameters are com-
pared inTables 12 and 13, respectively. The optimal cut-
ting parameters for surface roughness of 5A and 4A grade
DSS are the cutting speed at level 2 (100 m/min) and the
feed rate at level 1 (0.04 mm/rev). It is denoted as V2F1 in
theTables 12 and 13. The optimal cutting parameters for
cutting force of 5A and 4A grade DSS are the cutting speed
at level 3 (120 m/min) and the feed rate at level 1
(0.04 mm/rev). It is denoted as V3F1. The optimal cutting
parameters for tool wear of 5A and 4A grade DSS are the
cutting speed at level 1 (80 m/min) and the feed rate at le-
vel 1 (0.04 mm/rev). It is denoted as V1F1. The experimen-
tal results are closer to the predicted values within 8%
deviations.
4.3. Effect of cutting speed and feed rate on surface roughness
The influence of cutting speed on surface roughness of
5A and 4A grade DSS is illustrated in Fig. 5, for three differ-
ent feed rates. When cutting speed increases the surface
roughness value decreases up to 100 m/min. However, fur-
ther increase in cutting speed increases the surface rough-
ness values. The increasing cutting speed from 80 to
100 m/min reduces the surface roughness value due to
the reduction in BUE formation tendency. However, further
increase in cutting speed from 100 to 120 m/min, increases
the surface roughness due to the increase in cutting tool
nose wear at higher cutting speeds. Similar trend was re-
ported in turning operation of ASS[8].
The influence of feed rate on surface roughness of 5A
and 4A grade DSS is shown in Fig. 6, for three different cut-
ting speeds. The increasing feed rate increases the surface
Table 10
Results of the ANOVA for surface roughness, cutting force and tool wear for 5A grade DSS.
Cutting parameter Degrees of freedom Sum of squares Mean square Fratio Contribution (%)
Surface roughness
Cutting speed 2 0.9015 0.4507 11.40 30.91
Feed rate 2 1.8567 0.9283 23.49 63.67
Error 4 0.1581 0.0395 5.42
Total 8 2.9163 100
Cutting force
Cutting speed 2 8.2854 4.1427 56.92 45.38
Feed rate 2 9.6825 4.8412 66.52 53.03
Error 4 0.2911 0.0727 1.59
Total 8 18.2590 100
Tool wear
Cutting speed 2 101.8719 50.93595 233.62 92.05
Feed rate 2 7.9206 3.96030 18.1644 7.16
Error 4 0.8721 0.218025 0.79
Total 8 110.6646 100
Table 11
Results of the ANOVA for surface roughness, cutting force and tool wear for 4A grade DSS.
Cutting parameter Degrees of freedom Sum of squares Mean square Fratio Contribution (%)
Surface roughness
Cutting speed 2 4.2927 2.14635 12.07 31.48
Feed rate 2 8.6334 4.3167 24.28 63.31
Error 4 0.7112 0.1778 5.21
Total 8 13.6373 100
Cutting force
Cutting speed 2 8.50954 4.25475 66.38 42.31
Feed rate 2 11.3442 5.67210 88.49 56.41
Error 4 0.2564 0.06410 1.28
Total 8 20.1101 100
Tool wear
Cutting speed 2 107.8299 53.91495 637.67 90.60
Feed rate 2 10.8447 5.42235 64.13 9.11
Error 4 0.3382 0.08455 0.29
Total 8 119.0128 100
210 D. P. Selvaraj et al./ Measurement 49 (2014) 205215
http://-/?- -
7/25/2019 1-s2.0-S0263224113005903-main
7/11
roughness due to the increase in friction between work
piece and tool interface and increases the temperature in
the cutting zone. Hence, the shear strength of the material
reduces and behaves in a ductile fashion [31]. The sticky in
nature of DSS is the cause for the increased surface rough-
ness. The surface roughness can be minimized by employ-
ing the combination of lower level feed rate (0.04 mm/rev)
with medium level cutting speed (100 m/min).
4.4. Effect of cutting speed and feed rate on cutting force
The influence of cutting speed on cutting force of 5A
and 4A grade DSS is shown in Fig. 7, for three different feedrates. The cutting force decreases with increasing cutting
speed. Higher cutting force is required at lower cutting
speed due to the higher coefficient of friction between
the tool and work piece, results. At higher cutting speeds,
the temperature generation rate is higher which makes
the material soft at the cutting zone, which helps in remov-
ing the material at lower cutting forces. As the cutting
speed increases, the chip gets thinner and cutting forces re-
duced. The decrease in cutting force is due to reduction in
contact area and partly due to the drop in shear strength in
the flow zone as the temperature increases with increase
in cutting speed[31].
The influence of feed rate on cutting force of 5A and 4Agrade DSS is shown in Fig. 8, for three different cutting
speeds. The cutting force is increased with increasing feed
rate at all the selected cutting speeds. As the feed rate is in-
creased, the amount of material in contact with the tool
also increases. This implies an increased tool-work contact
Table 12
Comparison between predicted and experimental results for surface
roughness, cutting force and tool wear of 5A grade DSS at optimum cutting
condition.
Optimal cutting parameters
Prediction Experiment
Surface roughness
Level V2F1 V2F1Surface roughness (lm) 1.08 1.05
S/Nratio (dB) 0.66 0.42
Cutting force
Level V3F1 V3F1Cutting force (N) 34.36 35.2
S/Nratio (dB) 30.72 30.93
Tool wear
Level V1F1 V1F1Tool wear (mm) 0.105 0.113
S/Nratio (dB) 19.54 18.94
Table 13
Comparison between predicted and experimental results for surface
roughness, cutting force and tool wear of 4A grade DSS at optimum cuttingcondition.
Optimal cutting parameters
Prediction Experiment
Surface roughness
Level V2F1 V2F1Surface roughness (lm) 0.52 0.53
S/Nratio (dB) 5.76 5.51
Cutting force
Level V3F1 V3F1Cutting force (N) 25.38 25.7
S/Nratio (dB) 28.09 28.20
Tool wear
Level V1F1 V1F1Tool wear (mm) 0.098 0.102
S/Nratio (dB) 20.22 19.83
Fig. 5. Cutting speed vs. surface roughness-5A and 4A grade DSS.
Fig. 6. Feed rate vs. surface roughness-5A and 4A grade DSS.
Fig. 7. Cutting speed vs. cutting force-5A and 4A grade DSS.
D. P. Selvaraj et al. / Measurement 49 (2014) 205215 211
-
7/25/2019 1-s2.0-S0263224113005903-main
8/11
length. Due to this, the value of cutting force also increases.
In addition to increased contact length, the force resisting
deflection is high, which is due to the higher amount of
material in contact with the tool. This also contributes to
an increase in cutting forces. The minimum cutting force
can be obtained by employing the combination of lower le-
vel feed rate (0.04 mm/rev) with higher level cutting speed
(120 m/min).
4.5. Effect of cutting speed and feed rate on tool wear
The influence of cutting speed on tool wear of 5A and
4A grade DSS is shown in Fig. 9, for three different feed
rates. The tool wear increases with the increase in cutting
speed. Increase in cutting speed will increase the cutting
temperature at the cutting edge of the tool. The higher cut-
ting temperature causes the tool to lose its strength and
plastic deformation occurs. Therefore, the extent of tool
wear and cutting edge deformation increases. The tool
wear was significantly increased with increase in cutting
speed.The influences of feed rate on tool wear of 5A and 4A
grade DSS is depicted inFig. 10, for three different cutting
speeds. The tool wear increases with increase in feed rate.
The larger the feed, the greater is the cutting force per unit
area of chip-tool contact on the rake face and work-tool
contact on the flank face. However, it has been observed
that the effect of changes in feed rate on tool wear is rela-
tively lesser than that of proportionate changes in cutting
speed. The minimum tool wear was obtained by employing
the combination of lower feed rate (0.04 mm/rev) with
lower cutting speed (80 m/min).
4.6. Tool wear analysis
The worn surfaces of the cutting tool inserts used in the
machining processes of the 5A and 4A grade DSS work
piece material are examined using SEM. Figs. 11 and 12
illustrate the SEM images of the worn tool inserts used
for machining 5A and 4A grade DSS, respectively. It is ob-
served that wear predominantly occurred in three regions
during the tests, at the depth of cut line, rake surface and
the cutting edge. The SEM images inFig. 11ac are exam-
ined at three different cutting conditions. The lowest wear
was observed on the edge of the cutting tool used at
80 m/min cutting speed and 0.04 mm/rev feed rate
(Fig. 11a). The tool wear at the cutting edge is 0.113 mm.
The tool wear is extended on the cutting tool used at
100 m/min cutting speed and 0.08 mm/rev feed rate
(Fig. 11b). The tool wear at the cutting edge is 0.184 mm.
More wear was observed on the rake surface of the tool.
The highest wear was observed on the cutting tool used
at 120 m/min and 0.12 mm/rev feed rate (Fig. 11c).
Similarly, the less tool wear was observed on the edge
of the cutting tool used at 80 m/min cutting speed and
0.04 mm/rev feed rate (Fig. 12a). The tool wear at the cut-
ting edge is 0.102 mm. The tool wear is extended on the
cutting tool used at 100 m/min cutting speed and
0.08 mm/rev feed rate (Fig. 12b). The tool wear at the cut-
ting edge was 0.176 mm. More wear was observed on the
rake surface of the tool. The highest wear was observed
on the cutting tool used at 120 m/min and 0.12 mm/rev
feed rate (Fig. 12c). More wear was observed on the flank
and rake surface of the tool. The tool wear at the cutting
edge was 0.348 mm.
Tool wear is generally influenced by abrasion, diffu-
sion, thermal softening and notching. It was observed
from Figs. 11a and12a, there is less wear on the cutting
edge flank face and rake surface at a cutting speed of
80 m/min, a feed rate of 0.04 mm /rev and a depth ofFig. 8. Feed rate vs. cutting force-5A and 4A grade DSS.
Fig. 9. Cutting speed vs. tool wear-5A and 4A grade DSS.
Fig. 10. Feed rate vs. tool wear-5A and 4A grade DSS.
212 D. P. Selvaraj et al./ Measurement 49 (2014) 205215
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?- -
7/25/2019 1-s2.0-S0263224113005903-main
9/11
cut of 0.5 mm. It was due to the abrasive action of micro
hard particle present in the material and plastic defor-
mation. At this cutting speed, the cutting tool tempera-
ture was lower compared to the cutting speeds of
100 m/min and 120 m/min. Hence, abrasion is the major
factor influencing the tool wear at lower cutting speed. It
was also observed from Figs. 11b and12b, the wear rate
increases and there were more wear on the cutting edge,
flank face and rake surface at a cutting speed of 100 m/
min, a feed rate of 0.08 mm/rev and a depth of cut of
0.5 mm. Micro chipping was observed on the flank face
and rake surface. It was observed from Figs. 11c and
12c, the wear rate further increased, at a cutting speed
of 120 m/min, a feed rate of 0.12 mm/rev and a depth
of cut of 0.5 mm. Micro chipping and notch wear
were observed on the rake surface and flank face. At this
cutting speed, the cutting tool temperature was
higher compared to cutting speeds of 80 m/min and
100 m/min. The major factors such as diffusion, thermal
softening and notching influencing the tool wear at
120 m/min cutting speed. Hence, tool wear is mainly
due to abrasion at lower cutting speed and due to diffu-
sion, thermal softening and notching at higher cutting
speed. The results confirmed that the flank wear, rake
wear and notch wear were quite severe at higher cutting
speed [9].
Fig. 11. SEM images of tool inserts used for machining 5A grade DSS
(d= 0.5 mm), (a) V= 80m/min and f= 0.04 mm/rev, (b) V= 100 m/min
andf= 0.08mm/rev and (c)V= 120 m/min andf= 0.12 mm/rev. Fig. 12. SEM images of tool inserts used for machining 4A grade DSS
(d= 0.5 mm), (a) V= 80m/min and f= 0.04 mm/rev, (b) V= 100 m/min
andf= 0.08 mm/rev and (c)V= 120 m/min andf= 0.12 mm/rev.
D. P. Selvaraj et al. / Measurement 49 (2014) 205215 213
-
7/25/2019 1-s2.0-S0263224113005903-main
10/11
4.7. Effect of work piece material on surface roughness,
cutting force and tool wear
The surface roughness, cutting force and tool wear val-
ues for 4A grade DSS are lower compared to 5A grade DSS
due to the difference in chemical compositions, which
leads to the difference in the formation of micro grains
[4]. The 4A grade DSS material has lower strength and low-
er hardness compared to 5A grade DSS. The lower percent-
age of molybdenum presents in 4A grade DSS decreases the
strength and hardness[8]. Lower cutting force is required
for shearing 4A grade DSS leads to good surface finish
and less tool wear compared to 5A grade DSS.
4.8. Comparison between dry and wet turning operation
The comparison of surface roughness and cutting force
values of machined work piece during dry and wet turning
of 4A and 5A grade DSSs are depicted inFigs. 13 and 14,
respectively. The result reveals that the surface roughness
and the cutting force values of wet turning operation arereduced by about 510% compared to the dry turning oper-
ation. The application of cutting fluid reduces chip friction
during wet turning. Lower friction at the tool-chip inter-
face can lead to lower cutting temperature which results
in less tool wear and better surface finish. Moreover, low
friction at the tool-chip interface will reduce the tool-chip
contact length which in turn decreases the cutting force
[33,34].
5. Conclusion
The Taguchi optimization method was successfully
used to identify the optimal cutting parameters of two dif-
ferent grades of nitrogen alloyed DSS during dry turning
operations. The following specific conclusions are made
from this work.
A cutting speed of 100 m/min and a feed rate of
0.04 mm/rev are found to give the lowest surface
roughness for both 5A and 4A grade DSS. A cutting
speed of 120 m/min and a feed rate of 0.04 mm/rev
are found to give the lowest cutting force for both 5A
and 4A grade DSS. A cutting speed of 80 m/min and afeed rate of 0.04 mm/rev are found to give the lowest
tool wear for both 5A and 4A grade DSS.
ANOVA analysis indicates that the feed rate and the cut-
ting speed were affecting the surface roughness of 5A
grade DSS by approximately 64% and 31%, respectively.
The feed rate and the cutting speed were affecting the
cutting force of 5A grade DSS by approximately 53%
and 45%, respectively. The cutting speed and the feed
rate were affecting the tool wear of 5A grade DSS by
approximately 92% and 7%, respectively.
ANOVA analysis indicates that the feed rate and the cut-
ting speed were affecting the surface roughness of 4A
grade DSS by about 63% and 31%, respectively. The feedrate and the cutting speed were affecting the cutting
force of 4A grade DSS by about 56% and 42%, respec-
tively. The cutting speed and the feed rate were affect-
ing the tool wear of 4A grade DSS by about 91% and 9%,
respectively.
The tool wear was due to abrasion at lower cutting
speeds and due to diffusion, thermal softening and
notching at higher cutting speeds.
References
[1] J.R. Davis, ASM Specialty Handbook Stainless Steels, ASMInternational, Ohio, 1996.
[2] Malin Snis, Jan Olsson, Reduce costs for storage and distribution of
desalted water use duplex stainless steel, Desalination 223 (2008)
476486.
[3]Malin Snis, Jan Olsson, Duplex a new generation of stainless steels
for desalination plants, Desalination 205 (2007) 104113.
[4] S. Agrawal, S. Chakrabarti, A.B. Chattopadhyay, A study of the
machining of cast austenitic stainless steel with carbide tools, J.
Mater. Process. Technol. 52 (1995) 610620.
[5] Jukka Paro, Hannu Hanninen, Veijo Kauppinen, Tool wear and
machinability of X5 CrMnN 18 8 stainless steels, J. Mater. Process.
Technol. 119 (2001) 1420.
[6] Jukka Paro, Hannu Hanninen, Veijo Kauppinen, Tool wear and
machinability of HIPed P/M and conventional cast duplex stainless
steels, Wear 249 (2001) 279284.
[7] Ihsan Korkut, Mustafa Kasap, Ibrahim Ciftci, Ulvi Sekar,
Determination of optimum cutting parameters during machiningof AISI 304 austenitic stainless steel, Mater. Des. 25 (2004) 303305.
Fig. 13. Comparision of surface roughness between dry and wet turning
of 4A and 5A grade DSS.
Fig. 14. Comparision of cutting force between dry and wet turning of 4Aand 5A grade DSS.
214 D. P. Selvaraj et al./ Measurement 49 (2014) 205215
http://refhub.elsevier.com/S0263-2241(13)00590-3/h0175http://refhub.elsevier.com/S0263-2241(13)00590-3/h0175http://refhub.elsevier.com/S0263-2241(13)00590-3/h0010http://refhub.elsevier.com/S0263-2241(13)00590-3/h0010http://refhub.elsevier.com/S0263-2241(13)00590-3/h0010http://refhub.elsevier.com/S0263-2241(13)00590-3/h0015http://refhub.elsevier.com/S0263-2241(13)00590-3/h0015http://refhub.elsevier.com/S0263-2241(13)00590-3/h0020http://refhub.elsevier.com/S0263-2241(13)00590-3/h0020http://refhub.elsevier.com/S0263-2241(13)00590-3/h0020http://refhub.elsevier.com/S0263-2241(13)00590-3/h0020http://refhub.elsevier.com/S0263-2241(13)00590-3/h0025http://refhub.elsevier.com/S0263-2241(13)00590-3/h0025http://refhub.elsevier.com/S0263-2241(13)00590-3/h0025http://refhub.elsevier.com/S0263-2241(13)00590-3/h0030http://refhub.elsevier.com/S0263-2241(13)00590-3/h0030http://refhub.elsevier.com/S0263-2241(13)00590-3/h0030http://refhub.elsevier.com/S0263-2241(13)00590-3/h0035http://refhub.elsevier.com/S0263-2241(13)00590-3/h0035http://refhub.elsevier.com/S0263-2241(13)00590-3/h0035http://refhub.elsevier.com/S0263-2241(13)00590-3/h0035http://refhub.elsevier.com/S0263-2241(13)00590-3/h0035http://refhub.elsevier.com/S0263-2241(13)00590-3/h0035http://refhub.elsevier.com/S0263-2241(13)00590-3/h0030http://refhub.elsevier.com/S0263-2241(13)00590-3/h0030http://refhub.elsevier.com/S0263-2241(13)00590-3/h0030http://refhub.elsevier.com/S0263-2241(13)00590-3/h0025http://refhub.elsevier.com/S0263-2241(13)00590-3/h0025http://refhub.elsevier.com/S0263-2241(13)00590-3/h0025http://refhub.elsevier.com/S0263-2241(13)00590-3/h0020http://refhub.elsevier.com/S0263-2241(13)00590-3/h0020http://refhub.elsevier.com/S0263-2241(13)00590-3/h0020http://refhub.elsevier.com/S0263-2241(13)00590-3/h0015http://refhub.elsevier.com/S0263-2241(13)00590-3/h0015http://refhub.elsevier.com/S0263-2241(13)00590-3/h0010http://refhub.elsevier.com/S0263-2241(13)00590-3/h0010http://refhub.elsevier.com/S0263-2241(13)00590-3/h0010http://refhub.elsevier.com/S0263-2241(13)00590-3/h0175http://refhub.elsevier.com/S0263-2241(13)00590-3/h0175http://refhub.elsevier.com/S0263-2241(13)00590-3/h0175http://-/?-http://-/?- -
7/25/2019 1-s2.0-S0263224113005903-main
11/11
[8]Ibrahim Ciftci, Machining of austenitic stainless steels using CVD
multi-layer coated cemented carbide tools, Tribol. Int. 39 (2006)
565569.
[9]A. Senthikumar, A. Rajadurai, T. Sornakumar, The effect of tool wear
on tool life of alumina-based ceramic cutting tools while machining
hardened martensitic stainless steel, J. Mater. Process. Technol. 173
(2006) 151156.
[10] M.Y. Noordin, V.C. Venkatesh, S. Sharif, Dry turning of tempered
martensitic stainless tool steel using coated cermet and coated
carbide tools, J. Mater. Process. Technol. 185 (2007) 8390.
[11] S. Thamizhmanii, S. Hasan, Measurement of surface roughness andflank wear on hard martensitic stainless steel by CBN and PCBN
cutting tools, Ach. Mater. Manuf. Eng. 31 (2) (2008) 415421.
[12] G. Krolczyk, S. Legutko, M. Gajek, Predicting the surface roughness in
the dry machining of duplex stainless steel (DSS), Metalurgija 52
(2013) 259262.
[13] W. Bouzid Sai, N. Ben Salah, J.L. Lebrun, Influence of machining by
finishing milling on surface characteristics, J. Mach. Tools Manuf.
411 (2001) 443450.
[14] M. Muthukrishnan, J. Paulo Davim, Optimization of machining
parameters of Al/SiC-MMC with ANOVA and ANN analysis, J.
Mater. Process. Technol. 209 (2009) 225232.
[15] K. Palanikumar, Experimental investigation and optimization in
drilling of GFRP composites, Measurement 44 (10) (2011) 2138
2148.
[16] Nilrudra Mandal, B. Doloi, B. Mondal, Reeta Das, Optimization of
flank wear using Zirconia Toughened Alumina (ZTA) cutting tool:
Taguchi method and Regression analysis, Measurement 44 (2011)21492155.
[17] Ilhan Asiltrk, Harun Akkus, Determining the effect of cutting
parameters on surface roughness in hard turning using the Taguchi
method, Measurement 44 (2011) 16971704.
[18] J. Ross, Taguchi Technique for Quality Engineering, McGraw-Hill,
New York, 1993.
[19] G. Taguchi, M. El Sayed, Hashing, Quality Engineering and Quality
Systems, McGraw Hill, New York, 1989.
[20] K.R. Roy, A Primer on Taguchi Method, Van Nostrand Reinhold, New
York, 1990.
[21] W.H. Yang, Y.S. Tarng, Design optimization of cutting parameters for
turning operations based on the Taguchi method, J. Mater. Process.
Technol. 84 (1998) 122129.
[22] S. Basavarajappaa, G. Chandramohan, J. Paulo Davim, Some studies
on drilling of hybrid metal matrix composites based on Taguchi
techniques, J. Mater. Process. Technol. 196 (2008) 332338.
[23] J. Ghani, I.A. Chodhury, H.H. Hassan, Application of Taguchi method
in the optimization of end milling parameters, J. Mater. Process.
Technol. 145 (2004) 8492.
[24] D.I. Lalwani, N.K. Mehta, P.K. Jain, Experimental investigations of
cutting parameters influence on cutting forces and surface
roughness in finish hard turning of MDN250 steel, J. Mater.
Process. Technol. 206 (2008) 167179.
[25] Alakesh Manna, Sandeep Salodkar, Optimization of machiningconditions for effective turning of E0300 alloy steel, J. Mater.
Process. Technol. 203 (2008) 147153.
[26] M. Nalbant, H. Gokkaya, G. Sur, Application of Taguchi method in the
optimization of cutting parameters for surface roughness in turning,
Mater. Des. 28 (2007) 13791385.
[27] D. Philip Selvaraj, P. Chandramohan, Optimization of surface
roughness of AISI 304 austenitic stainless steel in dry turning
operation using Taguchi design method, J. Eng. Sci.Technol. 5 (2010)
293301.
[28] Z.Z. Julie, C.C. Joseph, E. Daniel Kirby, Surface roughness
optimization in an end-milling operation using the Taguchi design
method, J. Mater. Process. Technol. 184 (2007) 233239.
[29] C.H. Che Haron, A. Ginting, H. Arshad, Performance of alloyed
uncoated and CVD-coated carbide tools in dry milling of titanium
alloy Ti-6242S, J. Mater. Process. Technol. 185 (2007) 7782.
[30] M. Nouari, A. Ginting, Wear characteristics and performance of
multi-layer CVD-coated alloyed carbide tool in dry end milling oftitanium alloy, Surf. Coat. Technol. 200 (2006) 56635676.
[31] D.G. Thakur, B. Ramamoorthy, L. Vijayaraghavan, Study on the
machinability characteristics of superalloy Inconel 718 during high
speed turning, Mater. Des. 30 (2009) 17181725.
[32] H. Shao, L. Liu, H.L. Qu, Machinability study on 3% CO12% Cr
stainless steel in milling, Wear 263 (2007) 736744.
[33] L. De Chiffre, Function of cutting fluids in machining, J. Soc. Tribol.
Lubr. Eng. 44 (1988) 514518.
[34] P.K. Philip, A.S. Varadarajan, B. Ramamoorthy, Influence of cutting
fluid composition and delivery variables on performance in hard
turning using minimal fluid in pulsed jet form, J. Inst. Eng. (I) 82
(2001) 1219.
D. P. Selvaraj et al. / Measurement 49 (2014) 205215 215
http://refhub.elsevier.com/S0263-2241(13)00590-3/h0040http://refhub.elsevier.com/S0263-2241(13)00590-3/h0040http://refhub.elsevier.com/S0263-2241(13)00590-3/h0040http://refhub.elsevier.com/S0263-2241(13)00590-3/h0045http://refhub.elsevier.com/S0263-2241(13)00590-3/h0045http://refhub.elsevier.com/S0263-2241(13)00590-3/h0045http://refhub.elsevier.com/S0263-2241(13)00590-3/h0045http://refhub.elsevier.com/S0263-2241(13)00590-3/h0050http://refhub.elsevier.com/S0263-2241(13)00590-3/h0050http://refhub.elsevier.com/S0263-2241(13)00590-3/h0050http://refhub.elsevier.com/S0263-2241(13)00590-3/h0055http://refhub.elsevier.com/S0263-2241(13)00590-3/h0055http://refhub.elsevier.com/S0263-2241(13)00590-3/h0055http://refhub.elsevier.com/S0263-2241(13)00590-3/h0060http://refhub.elsevier.com/S0263-2241(13)00590-3/h0060http://refhub.elsevier.com/S0263-2241(13)00590-3/h0060http://refhub.elsevier.com/S0263-2241(13)00590-3/h0180http://refhub.elsevier.com/S0263-2241(13)00590-3/h0180http://refhub.elsevier.com/S0263-2241(13)00590-3/h0180http://refhub.elsevier.com/S0263-2241(13)00590-3/h0070http://refhub.elsevier.com/S0263-2241(13)00590-3/h0070http://refhub.elsevier.com/S0263-2241(13)00590-3/h0070http://refhub.elsevier.com/S0263-2241(13)00590-3/h0070http://refhub.elsevier.com/S0263-2241(13)00590-3/h0075http://refhub.elsevier.com/S0263-2241(13)00590-3/h0075http://refhub.elsevier.com/S0263-2241(13)00590-3/h0075http://refhub.elsevier.com/S0263-2241(13)00590-3/h0080http://refhub.elsevier.com/S0263-2241(13)00590-3/h0080http://refhub.elsevier.com/S0263-2241(13)00590-3/h0080http://refhub.elsevier.com/S0263-2241(13)00590-3/h0080http://refhub.elsevier.com/S0263-2241(13)00590-3/h0085http://refhub.elsevier.com/S0263-2241(13)00590-3/h0085http://refhub.elsevier.com/S0263-2241(13)00590-3/h0085http://refhub.elsevier.com/S0263-2241(13)00590-3/h0090http://refhub.elsevier.com/S0263-2241(13)00590-3/h0090http://refhub.elsevier.com/S0263-2241(13)00590-3/h0185http://refhub.elsevier.com/S0263-2241(13)00590-3/h0185http://refhub.elsevier.com/S0263-2241(13)00590-3/h0100http://refhub.elsevier.com/S0263-2241(13)00590-3/h0100http://refhub.elsevier.com/S0263-2241(13)00590-3/h0100http://refhub.elsevier.com/S0263-2241(13)00590-3/h0105http://refhub.elsevier.com/S0263-2241(13)00590-3/h0105http://refhub.elsevier.com/S0263-2241(13)00590-3/h0105http://refhub.elsevier.com/S0263-2241(13)00590-3/h0110http://refhub.elsevier.com/S0263-2241(13)00590-3/h0110http://refhub.elsevier.com/S0263-2241(13)00590-3/h0110http://refhub.elsevier.com/S0263-2241(13)00590-3/h0115http://refhub.elsevier.com/S0263-2241(13)00590-3/h0115http://refhub.elsevier.com/S0263-2241(13)00590-3/h0115http://refhub.elsevier.com/S0263-2241(13)00590-3/h0190http://refhub.elsevier.com/S0263-2241(13)00590-3/h0190http://refhub.elsevier.com/S0263-2241(13)00590-3/h0190http://refhub.elsevier.com/S0263-2241(13)00590-3/h0190http://refhub.elsevier.com/S0263-2241(13)00590-3/h0125http://refhub.elsevier.com/S0263-2241(13)00590-3/h0125http://refhub.elsevier.com/S0263-2241(13)00590-3/h0125http://refhub.elsevier.com/S0263-2241(13)00590-3/h0130http://refhub.elsevier.com/S0263-2241(13)00590-3/h0130http://refhub.elsevier.com/S0263-2241(13)00590-3/h0130http://refhub.elsevier.com/S0263-2241(13)00590-3/h0135http://refhub.elsevier.com/S0263-2241(13)00590-3/h0135http://refhub.elsevier.com/S0263-2241(13)00590-3/h0135http://refhub.elsevier.com/S0263-2241(13)00590-3/h0135http://refhub.elsevier.com/S0263-2241(13)00590-3/h0135http://refhub.elsevier.com/S0263-2241(13)00590-3/h0140http://refhub.elsevier.com/S0263-2241(13)00590-3/h0140http://refhub.elsevier.com/S0263-2241(13)00590-3/h0140http://refhub.elsevier.com/S0263-2241(13)00590-3/h0140http://refhub.elsevier.com/S0263-2241(13)00590-3/h0145http://refhub.elsevier.com/S0263-2241(13)00590-3/h0145http://refhub.elsevier.com/S0263-2241(13)00590-3/h0145http://refhub.elsevier.com/S0263-2241(13)00590-3/h0150http://refhub.elsevier.com/S0263-2241(13)00590-3/h0150http://refhub.elsevier.com/S0263-2241(13)00590-3/h0150http://refhub.elsevier.com/S0263-2241(13)00590-3/h0155http://refhub.elsevier.com/S0263-2241(13)00590-3/h0155http://refhub.elsevier.com/S0263-2241(13)00590-3/h0155http://refhub.elsevier.com/S0263-2241(13)00590-3/h0160http://refhub.elsevier.com/S0263-2241(13)00590-3/h0160http://refhub.elsevier.com/S0263-2241(13)00590-3/h0165http://refhub.elsevier.com/S0263-2241(13)00590-3/h0165http://refhub.elsevier.com/S0263-2241(13)00590-3/h0170http://refhub.elsevier.com/S0263-2241(13)00590-3/h0170http://refhub.elsevier.com/S0263-2241(13)00590-3/h0170http://refhub.elsevier.com/S0263-2241(13)00590-3/h0170http://refhub.elsevier.com/S0263-2241(13)00590-3/h0170http://refhub.elsevier.com/S0263-2241(13)00590-3/h0170http://refhub.elsevier.com/S0263-2241(13)00590-3/h0170http://refhub.elsevier.com/S0263-2241(13)00590-3/h0170http://refhub.elsevier.com/S0263-2241(13)00590-3/h0165http://refhub.elsevier.com/S0263-2241(13)00590-3/h0165http://refhub.elsevier.com/S0263-2241(13)00590-3/h0160http://refhub.elsevier.com/S0263-2241(13)00590-3/h0160http://refhub.elsevier.com/S0263-2241(13)00590-3/h0155http://refhub.elsevier.com/S0263-2241(13)00590-3/h0155http://refhub.elsevier.com/S0263-2241(13)00590-3/h0155http://refhub.elsevier.com/S0263-2241(13)00590-3/h0150http://refhub.elsevier.com/S0263-2241(13)00590-3/h0150http://refhub.elsevier.com/S0263-2241(13)00590-3/h0150http://refhub.elsevier.com/S0263-2241(13)00590-3/h0145http://refhub.elsevier.com/S0263-2241(13)00590-3/h0145http://refhub.elsevier.com/S0263-2241(13)00590-3/h0145http://refhub.elsevier.com/S0263-2241(13)00590-3/h0140http://refhub.elsevier.com/S0263-2241(13)00590-3/h0140http://refhub.elsevier.com/S0263-2241(13)00590-3/h0140http://refhub.elsevier.com/S0263-2241(13)00590-3/h0135http://refhub.elsevier.com/S0263-2241(13)00590-3/h0135http://refhub.elsevier.com/S0263-2241(13)00590-3/h0135http://refhub.elsevier.com/S0263-2241(13)00590-3/h0135http://refhub.elsevier.com/S0263-2241(13)00590-3/h0130http://refhub.elsevier.com/S0263-2241(13)00590-3/h0130http://refhub.elsevier.com/S0263-2241(13)00590-3/h0130http://refhub.elsevier.com/S0263-2241(13)00590-3/h0125http://refhub.elsevier.com/S0263-2241(13)00590-3/h0125http://refhub.elsevier.com/S0263-2241(13)00590-3/h0125http://refhub.elsevier.com/S0263-2241(13)00590-3/h0190http://refhub.elsevier.com/S0263-2241(13)00590-3/h0190http://refhub.elsevier.com/S0263-2241(13)00590-3/h0190http://refhub.elsevier.com/S0263-2241(13)00590-3/h0190http://refhub.elsevier.com/S0263-2241(13)00590-3/h0115http://refhub.elsevier.com/S0263-2241(13)00590-3/h0115http://refhub.elsevier.com/S0263-2241(13)00590-3/h0115http://refhub.elsevier.com/S0263-2241(13)00590-3/h0110http://refhub.elsevier.com/S0263-2241(13)00590-3/h0110http://refhub.elsevier.com/S0263-2241(13)00590-3/h0110http://refhub.elsevier.com/S0263-2241(13)00590-3/h0105http://refhub.elsevier.com/S0263-2241(13)00590-3/h0105http://refhub.elsevier.com/S0263-2241(13)00590-3/h0105http://refhub.elsevier.com/S0263-2241(13)00590-3/h0100http://refhub.elsevier.com/S0263-2241(13)00590-3/h0100http://refhub.elsevier.com/S0263-2241(13)00590-3/h0100http://refhub.elsevier.com/S0263-2241(13)00590-3/h0185http://refhub.elsevier.com/S0263-2241(13)00590-3/h0185http://refhub.elsevier.com/S0263-2241(13)00590-3/h0185http://refhub.elsevier.com/S0263-2241(13)00590-3/h0090http://refhub.elsevier.com/S0263-2241(13)00590-3/h0090http://refhub.elsevier.com/S0263-2241(13)00590-3/h0090http://refhub.elsevier.com/S0263-2241(13)00590-3/h0085http://refhub.elsevier.com/S0263-2241(13)00590-3/h0085http://refhub.elsevier.com/S0263-2241(13)00590-3/h0085http://refhub.elsevier.com/S0263-2241(13)00590-3/h0080http://refhub.elsevier.com/S0263-2241(13)00590-3/h0080http://refhub.elsevier.com/S0263-2241(13)00590-3/h0080http://refhub.elsevier.com/S0263-2241(13)00590-3/h0080http://refhub.elsevier.com/S0263-2241(13)00590-3/h0075http://refhub.elsevier.com/S0263-2241(13)00590-3/h0075http://refhub.elsevier.com/S0263-2241(13)00590-3/h0075http://refhub.elsevier.com/S0263-2241(13)00590-3/h0070http://refhub.elsevier.com/S0263-2241(13)00590-3/h0070http://refhub.elsevier.com/S0263-2241(13)00590-3/h0070http://refhub.elsevier.com/S0263-2241(13)00590-3/h0180http://refhub.elsevier.com/S0263-2241(13)00590-3/h0180http://refhub.elsevier.com/S0263-2241(13)00590-3/h0180http://refhub.elsevier.com/S0263-2241(13)00590-3/h0060http://refhub.elsevier.com/S0263-2241(13)00590-3/h0060http://refhub.elsevier.com/S0263-2241(13)00590-3/h0060http://refhub.elsevier.com/S0263-2241(13)00590-3/h0055http://refhub.elsevier.com/S0263-2241(13)00590-3/h0055http://refhub.elsevier.com/S0263-2241(13)00590-3/h0055http://refhub.elsevier.com/S0263-2241(13)00590-3/h0050http://refhub.elsevier.com/S0263-2241(13)00590-3/h0050http://refhub.elsevier.com/S0263-2241(13)00590-3/h0050http://refhub.elsevier.com/S0263-2241(13)00590-3/h0045http://refhub.elsevier.com/S0263-2241(13)00590-3/h0045http://refhub.elsevier.com/S0263-2241(13)00590-3/h0045http://refhub.elsevier.com/S0263-2241(13)00590-3/h0045http://refhub.elsevier.com/S0263-2241(13)00590-3/h0040http://refhub.elsevier.com/S0263-2241(13)00590-3/h0040http://refhub.elsevier.com/S0263-2241(13)00590-3/h0040