5th International & 26th All India Manufacturing Technology, Design and Research Conference (AIMTDR 2014) December

12th14th, 2014, IIT

Guwahati, Assam, India

Utility Fuzzy Multiobjective Optimization of Process Parameters for CNC

Turning of GFRP/Epoxy Composites
Hari Vasudevan1*,Naresh Deshpande2, Ramesh Rajguru3, Sandip Mane4
1*,2,4

Department of Production Engineering, D.J. Sanghvi College of Engineering,

Mumbai, India, 400056.
1*
E-mail:principaldjs@gmail.com;2E-mail: ncdeshpande72@yahoo.co.in;
3
Department of MechanicalEngineering, D.J. Sanghvi College of Engineering,
Mumbai, India, 400056.
3
E-mail: ramesh.rajguru9@gmail.com;4E-mail:sandip_dabade@yahoo.co.in
Abstract

Although Glass Fibre Reinforced Plastic (GFRP) composites are usually moulded near-net shape for obtaining close
fits and tolerances, certain amount of machining has to be carried out on them.A number of axisymmetric GFRP
composite parts are finish machined by turning. These include axles, spindles, columns, rolls, bearings, drag links
and steering columns. Quality and productivity are two important, but contradictory parameters while performing
machining operations. Hence, it becomes essential to evaluate the optimal cutting parameters setting in order to
satisfy contradictory requirements of quality and productivity.In this study,a hybrid multiobjective optimization
algorithm involving utility and fuzzy coupled with Taguchi methodology is used. Four process parameters, each at
three levels are selected for the study viz. cutting tool nose radius, cutting speed, feed rate and depth of cut. Surface
roughness parameter Ra, cutting force Fz and material removal rate MRR are the chosen output performance
measures. The experimental plan is laid according to Taguchis orthogonal array L27. Woven fabric based GFRP/
Epoxy tubes produced using hand layup process are finish turned using Poly Crystalline Diamond (PCD) cutting
tool. Utility values of the three performance measures are converted into a single Multi Performance Characteristics
Index (MPCI) using Mamdani type fuzzy inference system. This MPCI is then optimized using Taguchi analysis.
The parameter combination of A2B3C1D2, i.e. tool nose radius of 0.8 mm, cutting speed of 200 m/min, feed rate of
0.05 mm/rev and depth of cut of 1mm, is evaluated as the optimum combination. The confirmatory experiment at
these settings gives maximum value of MPCI validating the results.
Keywords: GFRP/Epoxy, Utility values, Fuzzy inference system, Multiobjective optimization.

1 Introduction
Machining of GFRP is different in comparison to
that of metals. Most homogeneous and ductile metals
can be machined by shearing and plastic deformation.
Continuous chips are usually formed during machining
of such metals. Machining of GFRPs, is usually
characterized by uncontrolled intermittent fracture.
Oscillating cutting forces are typical, because of the
intermittent fracture of the fibres (Jamal, 2009).
Quality and productivity are two important, but
contradictory parameters while performing machining
operations. Quality mainly concerns with dimensional
accuracy and surface roughness of the machined part,
whereas productivity is directly related to Material
Removal Rate (MRR) during machining. Minimizing
the cutting force is also important as it affects tool wear,
tool life and stability of machine tool. Surface finish
seems to be inversely related to MRR; hence it becomes
essential to evaluate the optimum cutting parameters

setting in order to satisfy contradictory requirements of

quality and productivity.
Isik and kentli (2009) proposed a multiple criteria
optimization approach using sensitivity. Minimizing
cutting forces and maximizing the material removal
were considered as objectives, while turning of
unidirectional glass fiber reinforced polyester rods.
Palanikumar et al. (2007) used grey relational grade &
Taguchi method for minimizing tool wear, surface
roughness and specific cutting pressure, while
maximizing material removal. They carried out turning
on GFRP/Epoxy composites using carbide (K10)
tool.Routara et al. (2010) applied utility concept
coupled with Taguchi method in order to evaluate the
best process environment,which could simultaneously
satisfy multiple requirements of surface quality. They
presented a case study on CNCend milling of UNS
C34000 medium leaded brass.Hari Singh and Pradeep
Kumar (2006) used the utility concept coupled with
Taguchi method for optimizing multiple quality
characteristics during turning of En24 steel.

221-1

Utility Fuzzy Multiobjective Optimization of Process Parameters for CNC Turning of GFRP/Epoxy Composit
Composites

It is observed that the machinability oof composite

materials is highly influenced by the ty
type of fibre, type
of resin, fibre orientation and method of manufacturing.
The extant literature survey also reveals that woven
glass fibre reinforced epoxy composites manufactured
by hand lay-up process have not been widely explored
for their machining characteristics, despi
despite their wide
applications. The present study is an attempt to bridge
this gap.

2 Methodology
The methodology used for this study is as shown
below in Fig. 1.

Ishikawa fish bone diagram was constructed for the

turning process of GFRP composites to sort out the
process parameters influencing the cutting force,
roughness and material removal rate. As the outcome of
this cause and effect analysis, process parameters
selected for the study are cutting tool nose radius,
cutting speed, feed rate and depth of cut. The range of
values for these parameters, areselected afterperforming
pilot experiments and referring
ng to the past contributions
of researchers in this context.
The experiments are planned using Taguchis design of
experiments (DOE). The total degrees of freedom
(DOF) for four parameters, each at three levels are
eight. Hence, a three level orthogonal array (OA) with
at least eight DOF is to be selected. The L27 OA (DOF
= 26) is thus selected for this study. The factors are
assigned to column no. 1, 2, 5 and 8 respectively. The
unassigned columns are treated as error. Also the trials
are carried out in random order.The
The process parameters
selected for the present work and their levels are as
given in Table 1.
Table 1 Control factors and their levels
Label
A
B
C
D

Process
parameters
Tool nose
radius
Cutting
speed
Feed rate
Depth of
cut

Units

Levels
L1

L2

mm

0.4

0.8

1.2

m/min

120

160

200

mm/rev 0.05 0.15

0.25

mm

0.6

L3

1.6

2.1 Experimentation
The work material selected for the study is glass
fibre reinforced epoxy composite. The E-glass
E
reinforcement is of woven fabric form having following
specifications. Type of weave: plain, weight: 1805
gm/m2 and 0.18mm thickness.
thickness
Epoxy resin
manufactured by Huntsman,
ntsman, product Araldite LY3297
and hardener Aradur 3298 is used as polymer matrix
material. The work specimens are tubular in shape,
shape 50
mm long, with inner diameter of 20 mm and outer
diameter of 55 mm. They are manufactured using hand
lay-up
up process and cured at room temperature. The
volume fraction of the reinforcement is 70%. The work
specimens before & after machining are as shown in
Fig. 2 & Fig. 3 respectively.

Figure 1 Methodology used for the study.

The cutting tool selected for turning is Poly

Crystalline Diamond (PCD) insert of the fine grade.
Three different types of inserts are used. They have ISO
coding as CNMA 120404, CNMA 120408 and CNMA

221-2

5th International & 26th All India Manufacturing Technology, Design and Research Conference (AIMTDR 2014) December 12th14th, 2014, IIT
Guwahati, Assam, India

dynamometer to a computer and using KistlerDynoware

type- 2825A software. The material removal rate is
calculated as per Eq. (1), by measuring the weight of
component before and after turning operation, with
precision digital weighing machine and recording the
machining time with a stop watch.

M .R.R. =

Figure 2 Work specimens before machining

Wi W f
tm

( gms / sec) (1)

Where, Wiis the initial weight of work specimen in gms;

Wf is the final weight of work specimen after machining
in gms. and tm is the machining time in sec. Table 2
shows L27 OA used for this experimentation.
Table 2 Taguchi L27 OA.

Figure 3 Work specimen after machining

120412. The tool holder is of WIDEX-ID1G with ISO
coding, PCLNL 25X25 M12.
The experiments are conducted on a Ace Jobber XL
CNC lathe machine with the following specifications:
swing over bed 500 mm, swing over carriage 260 mm,
max. turning dia. 270 mm, max. turning length 400 mm,
max. spindle speed 4000 rpm, spindle motor power 7.5
KW and Fanuc series Oi-TD Mate CNC controller. The
machining tests are carried out without any coolant.
The work piece is mounted on specially designed
mandrel, which is subsequently clamped by the lathe
chuck. One repeat run is conducted for each of the 27
trials. The response measures selected are, surface
roughness parameter Ra, tangential cutting force Fz and
material removal rate MRR. The surface roughness
parameter is measured using Taylor Hobson Talysurf-5
with Gaussian filter, cut-off length of 0.8mm, 5 cut-offs,
and total traverse length 4mm. Data acquisition is
accomplished by connecting this profiler to computer
and using SESURF software. The tangential cutting
force is measured with KistlerPiezo electric
dynamometer of type-5233A with built in, charge
amplifier up to 10 KN and a least count of 1mN. Data
acquisition is accomplished by connecting this

Exp.
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

A
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3

Factors levels (coded)

B
C
1
1
1
2
1
3
2
1
2
2
2
3
3
1
3
2
3
3
1
1
1
2
1
3
2
1
2
2
2
3
3
1
3
2
3
3
1
1
1
2
1
3
2
1
2
2
2
3
3
1
3
2
3
3

D
1
2
3
2
3
1
3
1
2
1
2
3
2
3
1
3
1
2
1
2
3
2
3
1
3
1
2

2.2 Optimization
2.2.1 Utility Theory
Utility refers to the satisfaction that each attribute
provides to the decision maker. Thus, utility theory
assumes that any decision is made on the basis of
the utility maximization principle, according to
which the best choice is the one that provides the
highest satisfaction to the decision maker (Kaladhar et

221-3

Utility Fuzzy Multiobjective Optimization of Process Parameters for CNC Turning of GFRP/Epoxy Composit
Composites

al., 2011).The preference number or utility value can be

expressed on a logarithmic scale as follows:

X
U i = A log i`
Xi

of Ui are given as inputs to the system. The given model

is a MISO (Multi Input and Single Output) model as
shown in Fig. 4.

(2)

Here, is the value of any quality characteristics or

attribute i, is the just acceptable value of quality
characteristic or attribute i and A is a constant. Value of
A can be found by the condition that if
(where
is the optimum or best value), then
therefore,

A=

9
X*i
log `
Xi

(3)
Figure 4 The FIS model

The individual utility values are usually aggregated to

calculate the overall utility using the following Eq. ((4).
Here, UO is the overall utility value, Ui is the individual
utility value of the ith quality characteristic and n is the
total number of responses. Wi is the weight for
ithattribute. Sum of all the attribute
ute weights should be
equal to unity.
n

Uo = WiU i

(4)

i =1

Subject to the condition:

n

W = 1
i

(5)

i =1

However, the problem in treating overall utilityUO, as

equivalent aggregated quality index is in assigning
priority weights of various responses. Extant literature
survey confirms that previous investigators have
determined optimal setting of process parameters by
maximizing UO within the experimental doma
domain. Results
obtained by such method can be inaccurate, as the exact
value of priority weight to be assigned to each and
individual responses is difficult to predict. Therefore,
slight change in priority weight may shift the optimal
setting, if these weightss are found sensitive to predict
the optima. To avoid
void this uncertainty, fuzzy inference
system is used in the present study to couple individual
utilityvalues into a single performance index i.e. MPCI.
2.2.2 Fuzzy Inference System
Fuzzy sets and systems were introduced by Prof.
Lotfi A. Zedah in 1965. A fuzzy rule based system
consists of four parts:
ts: Fuzzifier, knowledge base,
inference engine and defuzzifier. Detailed analysis on
fuzzy can be found in numerous literature (Zadeh 1976;
Mendel 1992).
This study has made an attemptto use fuzzy
inference system to estimate the MPCI, when values

The number of input variables (Ui) obtained in

utility value analysis & labelled as U1, U2, U3, etc. are
used as inputs. In three inputs (Ui) and one output
(MPCI) system, both the inputs and the output are taken
in the form of linguistic format.
rmat. A linguistic variable is
a variable, whose values are words or sentences in a
natural or man-made language. Forexample,
For
d1 = {low,
medium, high}, d2 = {low, medium, high}, and d3 =
{low, medium, high).The output (MPCI) is similarly
divided into MPCI = {very low, low, medium, high,
very high}. Fuzzy values are determined by the
membership functions, however so far there has been
no standard method for choosing the proper shape
of the membership functions for control variables. In
the proposed model, Gaussian type membership
functions are used for input as well as output variable,
In this proposed model, centroid
d of area (COA) method
of defuzzification is used for determining the
output.This crisp value is the MPCI. Table-3 shows the
observations and calculations of this method. Ra
indicates the average of roughness parameter, Fz
indicates the average tangential cutting force, and MRR
indicates the average of material removal rate as
calculated using Eq. (1), for each of the 27 trials. U_Ra,
U_Fz, and U_MRR are the individual
ndividual utility values of
the responses viz. roughness, tangential cutting force
and
material removal rate respectively.Multi
performance characteristics index MPCI values are as
given by the FIS. The Signal-to-Noise
Noise ratios (S/N) of
MPCI are calculated using higher the better criteria as
per Taguchi method.
2.2.3 Taguchi Optimization
To determine the optimal parameter settings, it is
required to find out the highest MPCI. Optimization of
MPCI has been carried out using Taguchi
method.Taguchi
Taguchi method converts response value
intocorresponding S/N ratio. The Signal-to-Noise
Signal
(S/N)
ratio is the ratio of mean to deviation of the response

221-4

5th International & 26th All India Manufacturing Technology, Design and Research Conference (AIMTDR 2014) December 12th14th, 2014, IIT
Guwahati, Assam, India

Table 33Utility Fuzzy calculation table

Ex.
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

Ra
(microns)
2.5802
2.4347
3.6033
3.0963
2.0938
3.2542
2.3977
2.6735
3.1297
2.3362
2.6505
2.6837
2.1363
1.9337
3.2880
2.0878
2.1220
2.2728
3.7797
1.9260
1.6973
2.1115
2.5267
2.8962
1.7735
2.2870
2.6573

Fz
(N)
9.2450
38.2350
80.6850
12.4650
54.0150
28.1650
16.7050
17.7300
48.9000
7.4300
30.2150
74.4050
12.4800
47.7300
24.4450
17.3950
17.4250
42.8650
10.5750
26.3700
68.4350
15.1800
49.7900
21.0750
18.5550
21.5750
43.2300

MRR
(gms/sec)
0.1398
0.6825
0.8889
0.2372
0.8721
0.5952
0.4737
0.4960
1.0000
0.1037
0.4879
2.8889
0.1630
0.6869
0.6389
0.6490
0.4286
1.7778
0.1198
0.5794
1.0556
0.4022
0.8426
1.0000
0.4240
0.4101
1.1111

U_Ra

U_Fz

U_MRR

MPCI

S/n MPCI

4.2919
4.9445
0.5371
2.2418
6.6399
1.6829
5.1166
3.8925
2.1214
5.4087
3.9896
3.8498
6.4140
7.5345
1.5666
6.6722
6.4897
5.7177
0.0000
7.5792
9.0000
6.5454
4.5275
2.9931
8.5065
5.6479
3.9607

8.1753
2.8181
0.0000
7.0476
1.5143
3.9716
5.9427
5.7180
1.8897
9.0000
3.7064
0.3058
7.0430
1.9811
4.5061
5.7900
5.7835
2.3868
7.6681
4.2200
0.6214
6.3040
1.8216
5.0658
5.5464
4.9774
2.3548

0.8082
5.0967
5.8116
2.2383
5.7599
4.7268
4.1091
4.2336
6.1302
0.0000
4.1888
9.0000
1.2225
5.1141
4.9183
4.9609
3.8382
7.6866
0.3891
4.6537
6.2765
3.6664
5.6669
6.1302
3.8092
3.7187
6.4152

4.5700
4.1500
1.5200
4.3100
3.6400
3.1200
5.0100
4.8900
3.4500
4.8800
4.7000
4.8100
3.8700
4.6700
3.0600
5.4700
5.1300
5.0500
2.6200
6.0800
5.0900
5.1000
3.3600
4.5500
6.6400
4.8200
4.1500

13.1983
12.3610
3.6369
12.6895
11.2220
9.8831
13.9968
13.7862
10.7564
13.7684
13.4420
13.6429
11.7542
13.3863
9.7144
14.7597
14.2023
14.0658
8.3660
15.6781
14.1344
14.1514
10.5268
13.1602
16.4434
13.6609
12.3610

Figure 5 Main effects plot for S/N ratios of MPCI

221-5

Utility Fuzzy Multiobjective Optimization of Process Parameters for CNC Turning of GFRP/Epoxy Composites

from targeted value. Optimal parametric combination

has been evaluated from the plot in Fig.5. The optimum
settings are A2B3C1D2.
The estimated mean of the response characteristic
S/N ratio () could be computed by using the following
Eq. (6), (Phadke M.S., 1989). Where = overall mean
of S/N ratio () for MPCI, A2= average value of S/N
ratio () for MPCI at second level of nose radius, B3 =
average value of S/N ratio () for MPCI at third level of
speed and C1 = average value of S/N ratio () for MPCI
at first level of feed. A, B & C are the most significant
factors, affecting the S/N ratio () for MPCI.
opt = + (A2 - ) + (B3 - ) + (C1 )
(6)
Predicted value (S/N Ratio) of MPCI becomes17.14
(highest among all entries of values in Table 3.),
whereas in confirmatory test it has been computed
as18.51,thus indicating that the quality has
beenimproved by this optimal setting (increment of S/N
ratio).

3 Conclusion
In this study, the fuzzy rule based model has been
developed using three input variables and one output
variable i.e. MPCI. By this way, a multi-response
optimization problem has been converted into an
equivalent single objective optimization problem, which
has been solved by Taguchi philosophy. The proposed
procedure is simple and effective in developing a robust
finish turning process for GFRP/Epoxy composites. The
proposed approach converts numerical response into a
linguistic term so that the issue of response correlation
could be avoided. Within the selected experimental
domain the optimal parameter settings obtained by
using this approach are A2B3C1D2, i.e. tool nose radius
of 0.8 mm, cutting speed of 200 m/min, feed rate of
0.05 mm/rev and depth of cut of 1mm.

References
Hari Singh and Pradeep Kumar (2006), Optimizing
multi-machining characteristics through Taguchi's
approach and utility concept, Journal of Manufacturing
Technology Management, Vol. 17 Issue 2, pp.255 - 274.
Ik B. &Kentli A. (2009), Multicriteria optimization of
cutting parameters in turning of UD-GFRP materials
considering sensitivity, Int J AdvManufTechnol, Vol.44,
pp.11441153.
Jamal Y. Sheikh-Ahmad (2009), Machining of Polymer
Composites, Springer, New York.
Kaladhar M., Subbaiah K.V., RaoC.S.andRao K. N.
(2011), Application of Taguchi approach and Utility
Concept in solving the Multi-objective Problem when
turning AISI 202 Austenitic Stainless Steel, Journal of
Engineering Science and Technology Review, Vol. 4,
Issue1, pp.55-61.
Mendel J. M. (1992), Fuzzy logic systems for
engineering: A tutorial, Proceedings of the IEEE,
Vol.83, pp.345377.
Palanikumar K., Karunamoorthy L. &Karthikeyan
R.(2007), Multiple Performance Optimization of
Machining Parameters on the Machining of GFRP
Composites Using Carbide (K10) Tool, Materials and
Manufacturing Processes, Vol.21, Issue 8, pp.846-852.
Phadke M. S. (1989), Quality Engineering Using Robust
Design, P T R Prentice-Hall, New Jersey.
Routara B. C., Mohanty S. D., Datta S.,
BandyopadhyayA.andMahapatra
S.
S.(2010),
Optimization in CNC end milling of UNS C34000
medium leaded brass with multiple surface roughnesses
characteristics, Sadhana, Vol.35, Issue 5, pp. 619-629.
Zadeh L. A. (1976), Fuzzy-algorithm approach to the
definition of complex or imprecise concept,

International Journal of Man-Machines Studies,Vol.8,

pp.249291.

221-6