G Model

PRE-6331; No. of Pages 11 ARTICLE IN PRESS

Precision Engineering xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Precision Engineering
journal homepage: www.elsevier.com/locate/precision

Experimental investigation, prediction and optimization of

cylindricity and perpendicularity during drilling of WCB material
using grey relational analysis
Saurin Sheth a,∗ , P.M. George b
a
Mechatronics Engineering Department, G. H. Patel College of Engineering and Technology, Vallabh Vidyanagar 388120, Gujarat, India
b
Mechanical Engineering Department, B V M Engineering College, Vallabh Vidyanagar 388120, Gujarat, India

a r t i c l e i n f o a b s t r a c t

Article history: Manufacturing is always the heart of majority of industries. Drilling is an extremely important and an
Received 25 July 2015 essential machining process which requires a lot of attention as in most of the cases it is required for
Received in revised form assembly purposes. Majority of the holes produced during drilling are made with the help of Vertical
19 December 2015
Machining Centre (VMC) meant for pin- hole assembly. Though the tolerance is within limit, assembly
Accepted 4 January 2016
problems arise due to the improper geometry of these holes. Various geometrical tolerances like cylin-
Available online xxx
dricity, circularity, perpendicularity and position errors are responsible for such assembly problems. This
investigation is focussed on cylindricity and perpendicularity in the drilling of Wrought Cast Steel Grade
Keywords:
Vertical Machining Centre
B (WCB) material using SOMX 050204 DT insert. In this work, effect of machining variables like cutting
Drilling speed, feed rate and depth of cut (canned cycle) are investigated and optimized using grey relational
Coordinate Measuring Machine analysis (GRA). Reliable experiments are conducted based on a 33 full factorial, replicated twice. Second
Cylindricity order regression models are developed for predicting cylindricity and perpendicularity. The models’ ade-
Perpendicularity quacy has been checked by calculating correlation coefficient. It shows that the developed models are
Analysis of variance well fitted for the prediction of responses within the specific range of input variables.
Regression © 2016 Elsevier Inc. All rights reserved.
GRA

1. Introduction [37]. Schmitz et al. have presented a case study on comparison of

Achieving better quality to improve performance of the prod-
uct is the basic necessity of every industry. Certain challenges
like dimensional/geometrical tolerances, good surface finish, high
production rate, reduced tool wear are encountered by many pro-
duction industries. Increase in the metal removal rates lead to
high productivity in machining operations. It is very difficult to
balance the high metal removal rate and tight tolerance targets,
because of their nature. The degree of conformance of the fin-
ished part to dimensional and geometric specifications has four
major contributors: geometric errors of the machine construction,
thermally induced errors from heat sources associated with the
machine/cutting process, trajectory following errors caused by con-
troller and machine structural dynamics, and errors due to cutting
forces [36]. Bryan J. explained the significance of heat generated
during machining. Machine axes, spindle motor and shearing action
during cutting are responsible for heat generation. This heat dis-
tort the machine geometry and causes error in the machined parts
∗ Corresponding author. Tel.: +91 2692 231651/+91 09428076097.
E-mail address: saurinsheth@gcet.ac.in (S. Sheth).
error sources in high- speed milling. They have considered geomet-
rical, thermal, contouring and cutting force related errors. Their
analysis shows that the cutting force error is the most dominant
one compared to all other errors for certain choices of spindle
speeds [38]. Ramesh et al. reviewed various error compensation
techniques related to machine tools. They have considered three
major types of error such as geometric, thermal and cutting force
induced errors. The paper reports the work done in analysing the
various sources of geometric errors that are usually encountered
on machine tools and the methods of elimination or compensation
employed in these machines. Even cutting force induced errors as
well as thermal errors and its compensation techniques are very
well discussed [39–41]. Thus it can be said that deflection of tool,
workpiece, machine spindle and fixture can be regarded as the
sources of dimensional errors in machining [7,8]. As high preci-
sion assemblies cannot be analysed with the assumption that form
errors are negligible. So, serge et al. have proposed and built a geo-
metrical model based on the modal shapes of the ideal surface
[51].
Many components have more than one geometrical feature
required for assembly. In such cases it is required to find several
geometric tolerances on a single part, which is very crucial from the

http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
0141-6359/© 2016 Elsevier Inc. All rights reserved.

Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
G Model
PRE-6331; No. of Pages 11 ARTICLE IN PRESS
2 S. Sheth, P.M. George / Precision Engineering xxx (2016) xxx–xxx
inspection point of view. Moroni et al. proposed a methodology for
planning CMM sampling strategies for multiple tolerance based on
the minimization of inspection costs [44]. Estimating any geomet-
rical tolerance requires to sample a cloud of points on the feature
to check and then to fit an ideal substitute geometry with respect
to evaluation of errors. Generally points are sampled by means of
a CMM and then a suitable software algorithm will be applied.
Moroni et al. reviewed that some algorithms are slow and some
of them give approximated solution. So, a new algorithm has been
proposed by them [45]. Antonio et al. consider how deviation on
fixturing elements effects the location tolerance of a hole’s pattern.
They have designed locator positioning, which in turn minimizes
the positional tolerance [48].
Exploring the literature it was found that a lot of potential is
there in the research domain of geometrical tolerances, as it plays
vital role during the assembly attributed to better conformity of
components. Majority of the researches are done for various types
of errors associated in achieving these tolerances. The error com-
pensation strategies are also developed. Various innovative models
and techniques were developed to measure various geometrical
tolerances. But the work related to cutting parameters, its investi-
gation, its contribution and its selection to optimize the geometrical
errors need some amount of focus, which is presented here.
Geometric Dimensioning & Tolerancing (GD&T) is created to
insure smooth assembly of mating parts, to improve quality, and
to reduce the cost [1]. GD & T is a symbolic language used to
specify the size, form, orientation and location of part features.
It is based on the standard, Dimensioning and Tolerancing ASME
Y14.5M-1994 which is later on updated as ASME Y14.5-2009 [3].
ISO 1101: 2012 (E), which is an international standard of Geo-
metrical product specifications (GPS) – Geometrical tolerancing –
Tolerances of form, orientation, location and run-out, contains basic
information and gives requirements for the geometrical toleranc-
ing of workpieces [4,47]. Drawings with properly applied geometric
tolerancing provide the best opportunity for unambiguous inter-
pretation and cost-effective assembly. Thus GD & T is a design tool,
which helps in reducing the rework due to assembly problems,
eventually leading to most economized manufacturing [2]. There
are five types of tolerances, namely, Form, Profile, Orientation, Run
out and Location. While the four form characteristics are Flatness,
Straightness, Circularity and Cylindricity [3,4].
2. Literature review
Kumar et al. [9] investigate the effects of drilling parame-
ters such as cutting speed (5, 6.5, 8 m/min), feed (0.15, 0.20,
0.25 mm/rev) and drill tool diameter (10, 12, 15 mm) on surface
roughness, tool wear by weight, material removal rate and hole
diameter error in drilling of OHNS material using HSS spiral drill.
Their results show that the feed (62.24%) has greater influence on
surface roughness and drill tool diameter, cutting speed accounts
13.94%, 11.48% percent contribution respectively on the holes sur-
face roughness. Also feed (83.38%) has greater influence on hole
diameter error and cutting speed and drill toll diameter accounts
1.17%, 0.50% percent contribution respectively on the hole diameter
error. Shahrajabian et al. [10] investigated the effects of machining
parameter and tool geometry (spindle speed, feed rate and tool
point angle) during the drilling process of carbon fibre reinforced
polymer composites (CFRP) on surface roughness, delamination
and thrust force. The experiments on CFRP were conducted to
obtain surface roughness, delamination factor and thrust force val-
ues based on the full factorial design of experiments, and then
analysis of variance (ANOVA) is performed.
Madhavan et al. [11] investigate effect of thrust force during
drilling of 10 mm diameter holes in 20 mm thick carbon fibre rein-
forced plastic composite laminate using HSS, Solid Carbide (K20)
Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
and Poly Crystalline Diamond insert drills. Experiments are con-
ducted on a Vertical Machining Centre using Taguchi design of
experiments. A model is developed to correlate the drilling param-
eters with thrust force using Response surface methodology (RSM).
Kilickap et al. [12] investigated the cutting parameters (cutting
speed, feed rate, tool geometry) affecting delamination of drilling
operation. Taguchi L16 method is performed in drilling machine
on GFRP composite. Tyagi et al. [13] applied Taguchi methodology
for the drilling of mild steel with the help of CNC drilling machin-
ing operation with tool use high speed steel. A L9 orthogonal array
and analysis of variance (ANOVA) were applied to study the per-
formance characteristics of machining parameter (spindle speed,
feed, and depth) with consideration of good surface finish as well
as high material removal rate (MRR).
Bharti et al. [14] analysed effect of machining parameters spin-
dle speed, feed rate and tool point on the hole diameter produced
and the material removal rate in the Micro-drilling. Taguchi based
method along with ANOVA (analysis of variance) and DOE (design
of experiments) is implemented for optimized result. In percentage
contribution feed rate has about 73.61%, spindle speed has 23.31%
and tool point angle has 1.5% influence on the optimum diameter
and material removal rate combine. Palanikumar et al. [15] evalu-
ated the effect of spindle speed and feed on delamination in drilling
of GFRP composites. They have used twist drill and 4-flute cutter
to perform the experiments. The influence of speed is 3.04% and
feed is 93.15% on the delamination when drilling with twist drill,
while 0.15% and 97.67% when drilling with 4-flute cutter. Results
show that the 4-flute cutter has better results compared to twist
drill. Reddy et al. [16] analysed effect of machining parameters on
surface roughness and roundness error during drilling of AL 6061
alloy. They have used grey relational analysis for optimizing the
responses. Kuram et al. [17] developed model to predict thrust force
and surface roughness using fuzzy logic and regression. Thrust force
and surface roughness during the drilling of AISI 304 with the HSS-E
tool measured. Spindle speed, Feed and depth of cut are considered
as machining parameters. Sheth et al. [18–20] investigated effect
of pressure, RPM and no of grooves on MRR and Spread during the
flashing operation of precision steel ball manufacturing. They have
developed a fuzzy logic and regression based model to predict MRR
and Spread during flashing operation.
Literature survey shows that researchers are working with the
surface roughness of the drilled holes. Even though the surface
roughness and dimensions are within limit, majority of the assem-
bly problems are occurring due to form error. Various geometrical
tolerances like cylindricity, circularity, perpendicularity and posi-
tion are responsible for such problems in case of holes [21–23].
These geometric tolerances must be applied at the design stage
itself to meet the functional requirement along with the assembly
producibilityi. However, the production design should be made in
terms of machine tool capability so that DFMA can be achieved.
In the present study an attempt is made to develop a predictive
model of geometric tolerances for WCB material in the context of
various drilling input variables. This experimental study may give
a new insight to the practising engineers for utilizing the machine
tool at optimized level to achieve the best geometry and precision.
The same methodology may be adopted for modelling of other geo-
metrical errors also. It may be extended to various machine tools
as well for various materials too.
3. Experimentation
3.1. Work piece
The test specimen material is WCB, which is carbon steel with
hardness 255 HB and 240 mm × 160 mm × 55 mm size. It has
excellent strength properties at high temperature. It is generally
G Model
PRE-6331; No. of Pages 11 ARTICLE IN PRESS
S. Sheth, P.M. George / Precision Engineering xxx (2016) xxx–xxx 3

Fig. 1(b). Around 80% of the valves are manufactured using WCB
as base material due to its less cost along with its functionality
[25]. This implies the selection of material as WCB and geometrical
errors associated with it like cylindricity and perpendicularity for
further investigation. The chemical compositions and mechanical
properties are listed in Table 1.

3.2. Tool

The tool insert used is SOMX 050204 DT, which has been pro-
duced by ISCAR Ltd [26]. It is an insert for DR drills with DT chip
former for medium to high speed general use. It is recommended
and already being used by valve manufacturers to drill such kind of
holes. Fig. 2(a) shows the drilling tool with insert. Fig. 2(b) shows
the specifications of the insert.

3.3. VMC Machine

Machinability tests are carried out at the 3-axis VMC VF2SS

machine as shown in Fig. 3. It has a spindle motor power of 22.4 kW
with maximum spindle speed of 12,000 rev/min [27].

3.4. Experimental procedure

WCB Block is clamped using a hydraulic vice. The bottom surface

is machined to create a qualified datum feature, which acts as a pri-
mary datum. Then workpiece is located on the qualified datum for
further machining. The slot is created on the top surface to gener-
ate a new qualified surface, which will be useful for measurement.
Holding the plate in the same position the holes are drilled. As
here 33 full factorial with twice replication design is selected to
perform experiments, total 54 experimental runs are performed
for various combinations of the input factors. Three plates of size
240 mm × 160 mm × 55 mm are drilled with each plate having 18
holes of 20 mm diameter and 40 mm depth. The machined com-
ponents are shown in Fig. 4. It is assumed that the remaining
parameters are constant during the machining.

3.5. Design of experiments

Fig. 1. (a) Exploded view of Nozzle Check Valve. (b) Cut section of diffuser showing
the mating surfaces.
used in manufacturing of valves, flanges, fittings, or other
pressure-containing parts for high-temperature service pertaining
to chemical industries. Various materials like WCB, CF8, CF8M, CF3,
CF3M, Inconel, Aluminium Bronze are used to manufacture the
nozzle check valves [24,52]. Fig. 1(a) shows the exploded view of
nozzle check valve with different parts. Here part 3, i.e. disc stem,
has to slide in the part 2, i.e. diffuser. It seems that the geometry
of mating parts play a vital role in assembly, leakage and fluttering
at high pressure. So, these two parts have to be produced carefully
with proper tolerances for fulfilling the functional requirements.
Cylindricity and perpendicularity tolerances are playing vital role
in reducing the fluttering at high pressure which can be seen in
Different variables that affect the drilling process are spindle
speed, feed rate, depth of cut, coolant ratio, tool geometry (point
angle, drill diameter, chisel edge angle) etc. From the literature
it seems that speed, feed and depth of cut are significant in sur-
face roughness, delamination, Thrust force, MRR and hole diameter
error [9,10,13–16]. Sheth et al. investigated the effect of drilling
parameters during machining of mild steel in context of geomet-
rical tolerances such as cylindricity and perpendicularity [42,43].
They found that three most significant variables are speed, feed and
depth of cut. So, here speed, feed and depth of cut vary for experi-
mentation and investigation. Cylindricity and perpendicularity are
considered as responses in the drilling process as it is desired during
the manufacturing of nozzle check valve. Input variables in terms
of the coded factors along with its levels are defined in Table 2.

Table 1
Mechanical properties and chemical composition of WCB material.

Carbon steel Metal code WCB Standard ASTM A216

Tensile strength (N/mm2 ) 485–655 Yield strength (N/mm2 ) ≥250 Hardness HB 145

C Si Mn P S Cr Ni Mo Cu V

≤0.30 ≤0.60 ≤1.00 ≤0.040 ≤0.045 ≤0.50 ≤0.50 ≤0.20 ≤0.30 ≤0.03

Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
G Model
PRE-6331; No. of Pages 11 ARTICLE IN PRESS
4 S. Sheth, P.M. George / Precision Engineering xxx (2016) xxx–xxx

Fig. 2. (a) Drilling tool with insert. (b) Specification of Insert [21].

Table 2
Factors and levels.

Factors Coded factors Low (1) Medium (2) High (3)

Spindle speed (rpm) A 1200 1400 1600

Feed rate (mm/min) B 100 150 170
DOC/CAN cycle (mm) C 1 1.5 2

3.6. Measuring techniques

Fig. 3. Vertical milling machine at Cipriani Harrison Engineering.
3.6.1. CMM
The prompt advances in manufacturing techniques justify the
use of Coordinate Measuring Machines (CMMs), as basic instru-
ments in modern metrology. They are used broadly for carrying
out on/off line inspection with various uncertainty. Four main
uncertainties are machine uncertainty, sampling uncertainty,
thermally induced uncertainty and datum uncertainty. Hocken
et al. have suggested uncertainty estimation tool as a module tied
to the CMM off-line programming software, so that non experts
can also use CMM easily [28,46]. Hexagon made CNC Coordinate
Measuring Machine (CMM) has been used to measure cylindricity

Fig. 4. Machined work piece.

Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
G Model
PRE-6331; No. of Pages 11 ARTICLE IN PRESS
S. Sheth, P.M. George / Precision Engineering xxx (2016) xxx–xxx 5

Fig. 5. Hexagon co-ordinate measuring machine with probe at BVM College.

and perpendicularity. CMM as shown in Fig. 5, has servo controlled are measured with 15◦ angular pitch and 30/48 mm linear pitch.
axes with resolution of 0.1 ␮m. The figure also has bird eye view Even the centre axis of the cylinder has been obtained by measur-
for head and probe of CMM. It has Motorized index able probe ing cylindricity. The perpendicularity is measured in the context of
head with Touch trigger probing system and temperature com- top plane/surface. A schematic diagram for measurement is shown
pensation. Axes range of CMM is 500 mm, 500 mm and 400 mm in Fig. 7(a). As minimum three points on the top surface decide the
for X axis, Y axis and Z axis, respectively. datum plane [3–6]. Three extremum points on the machined sur-
As per ISO 1101 the cylindricity of a single toleranced feature is face decided the datum as shown in Fig. 7(a). Central axis deviation
deemed to be correct when the feature is confined between two of the cylinder with respect to the top datum plane gives the value
coaxial cylinders such that the difference in radii is equal to or of perpendicularity. Fig. 7(b) shows the schematic sketch showing
less than the value of the specified tolerance. The location of the cylindricity and perpendicularity error in 2D and 3 D View of the
axes of these cylinders and the value of their radii shall be chosen plate. Sample measurement of cylindricity and perpendicularity for
so that the difference in radii between the two coaxial cylinders
is the least possible value [4,47]. So, here the variation between
measured cylindrical surface and ideal cylindrical surface is consid-
ered as cylindricity error. The perpendicularity error is the variation
between ideal cylindrical axes with the measured one with respect
to specified datum. These features act as important parameter for
assembly of nozzle check valve. [24,25,28]. Fig. 6(a) and (b) give
the conceptual idea of cylindricity and perpendicularity. Fig. 6(a)
shows the radial tolerance zone, with 0.0X wide, where X is the tol-
erance value. Fig. 6(b) shows the possible axis orientation, which is
present in the tolerance zone of perpendicularity. Here, surface A is
considered as datum for measurement of perpendicularity. Stanis-
law et al. have discussed a method of quantitative comparison of
cylindricity profiles measured with different strategies. The four
cylindricity measuring strategies are: (a) the strategy for measure-
ment of roundness profiles (b) the strategy for measurement of
generatrix lines (c) the bird-cage strategy (which is a combination
of measurement of roundness profiles & generatrix lines) (d) point
strategy. A new measuring device has been prepared, which works
on reference measurement of cylindricity [49,50].
The measurement of cylindricity was carried out by moving a
probe into the spiral fashion for 720◦ rotation which covered 30 mm
depth. The intermediate points during the measurement of one hole Fig. 6. (a) Tolerance zone of cylindricity. (b) Tolerance zone of perpendicularity.

Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
G Model
PRE-6331; No. of Pages 11 ARTICLE IN PRESS
6 S. Sheth, P.M. George / Precision Engineering xxx (2016) xxx–xxx
Fig. 7. (a) Schematic plate showing the three extremum points used to define the
datum. (b) Schematic sketch showing cylindricity and perpendicularity error in 2D
and 3 D View.
8th run order is shown in Fig. 8. Table 3 shows the experimen-
tal results of cylindricity and perpendicularity for each treatment
combination.
4. Results and discussion
Statistical design of experiments (DOE) is widely used in the
analysis of machining experiments [30,31]. DOE helps to plan the
experiments, which in turn is useful for analysis and modelling of
the process.
4.1. Analysis of variance (ANOVA)
Significance of the input variables can be analyzed using ANOVA.
Tables 4 and 5 indicate the results of ANOVA for the cylindric-
ity and perpendicularity. The results indicate that spindle speed,
feed and depth of cut are the significant parameters, which affect
Fig. 8. Sample reading of cylindricity and perpendicularity at 8th run order using CMM.
Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
cylindricity and perpendicularity. Interactions between the spin-
dle speed and feed also have significant effect on cylindricity and
perpendicularity both. From Fig. 9(a), it can be observed that the
rise of spindle speed escalates cylindricity error and reduces per-
pendicularity error. The surges of feed rate rise cylindricity error
significantly, while increase in perpendicularity error is quite rea-
sonable. Increase of depth of cut increases both cylindricity error
and perpendicularity error. Interaction effect between the drilling
parameters is presented in Fig. 9(b) in terms of the coded variables.
It shows that there is no interaction among spindle speed, feed
rate and depth of cut for cylindricity and perpendicularity within
the limit. The coded variables are defined with levels and units in
Table 2. As the modelling is done in terms of the coded factors so
no units will be available at vertical scale. If we co relate Fig. 7(b)
with Table 2 then it shows that coded factor A is representing spin-
dle speed in rpm with three levels. Level 1(Low) = 1200 rpm, level
2 (medium) = 1400 rpm and level 3 (High) = 1600 rpm. The same is
applicable for feed and depth of cut.
4.2. Regression modelling
To predict the responses in context of input and output parame-
ters, regression models are widely used. Regression model consists
of constants and coefficient of predictor. The general regression
model is shown in Eq. (1). The value of (ˇ0 ) is zero.
y = ˇ0 + ˇ1 x1 + ˇ2 x2 + ˇ12 x1 x2 + · · · + ˇjk xj xk (1)
The regression model for cylindricity (Eq. (2)) and perpendicu-
larity (Eq. (3)) is developed with the help of MINITAB 16 software.
Here x1 , x2 , x3 are the predictor variables associated with the spin-
dle speed, feed rate and depth of cut/canned cycle in terms of the
coded factors.
Cylindricity = 0.12413 − 0.0386667x1 − 0.0335x2
− 0.0268333x3 + 0.00577778x12 +
0.0145417x1 x2 + 0.00544444x22 − 0.00125x2 x3 (2)
+ 0.000833333x1 x3 + 0.00752778x32 +
0.0003125x1 x2 x3
S = 0.00351979 R-Sq = 98.04% R-Sq (adj) = 97.58%
PRESS = 0.000805213 R-Sq (pred) = 97.03%
Perpendicularity = 0.0797037 − 0.0615278x1
+ 0.0514167x2 + 0.0245278x3 +
0.0104722x12 − 0.008625x1 x2 − 0.00386111x22 (3)
− 0.00295833x2 x3 − 4.16667e − 005x1 x3 −
0.00169444x32 + 0.000125x1 x2 x3
S = 0.00591858 R-Sq = 97.71% R-Sq (adj) = 97.18%
PRESS = 0.00246767 R-Sq (pred) = 96.25%
G Model
PRE-6331; No. of Pages 11 ARTICLE IN PRESS
S. Sheth, P.M. George / Precision Engineering xxx (2016) xxx–xxx 7

Table 3
Experimental results.

Run Order Treatment combinations Responses

Cylindricity (mm) Perpendicularity (mm)

A B C Replicate 1 Replicate 2 Replicate 1 Replicate 2

1 1 1 1 0.057 0.055 0.088 0.093

2 1 1 2 0.053 0.052 0.100 0.101
3 1 1 3 0.061 0.066 0.102 0.127
4 1 2 1 0.055 0.060 0.093 0.125
5 1 2 2 0.048 0.050 0.125 0.138
6 1 2 3 0.059 0.063 0.134 0.144
7 1 3 1 0.06 0.063 0.138 0.140
8 1 3 2 0.056 0.057 0.147 0.150
9 1 3 3 0.063 0.068 0.153 0.151
10 2 1 1 0.056 0.050 0.049 0.050
11 2 1 2 0.051 0.052 0.064 0.067
12 2 1 3 0.067 0.061 0.079 0.080
13 2 2 1 0.067 0.070 0.067 0.068
14 2 2 2 0.061 0.059 0.082 0.081
15 2 2 3 0.07 0.067 0.091 0.099
16 2 3 1 0.082 0.083 0.075 0.080
17 2 3 2 0.08 0.081 0.082 0.088
18 2 3 3 0.09 0.091 0.093 0.091
19 3 1 1 0.059 0.057 0.025 0.035
20 3 1 2 0.057 0.058 0.046 0.045
21 3 1 3 0.068 0.070 0.059 0.057
22 3 2 1 0.073 0.093 0.036 0.047
23 3 2 2 0.078 0.080 0.052 0.054
24 3 2 3 0.094 0.095 0.060 0.062
25 3 3 1 0.124 0.122 0.044 0.046
26 3 3 2 0.121 0.123 0.056 0.054
27 3 3 3 0.132 0.134 0.062 0.060

Table 4
ANOVA for cylindricity.

Source Degree of freedom Sum of squares Adj. sum of squares Adj. mean squares F P % Contribution

A 2 0.01013 0.01013 0.00506 431.65 0.000 37.35

B 2 0.00970 0.00970 0.00485 413.10 0.000 35.77
C 2 0.00117 0.00117 0.00058 49.88 0.000 4.31
A×B 4 0.00563 0.00563 0.00140 120.07 0.000 20.75
A×C 4 0.00005 0.00005 0.00001 1.25 0.315 0.18
B×C 4 0.00005 0.00005 0.00001 1.25 0.313 0.18
A×B×C 8 0.00004 0.00004 0.000005 0.45 0.881 0.14

Error 27 0.00031 0.00031 0.00001 1.14

Total 53 0.0271226

R2 = 98.83%. R2 (adj) = 97.71%

4.3. Adequacy check of the model
R-Sq is the co-efficient of determination, which helps to judge
the adequacy of the developed regression model. R-Sq (adj)
values for both the models are very much close to R-Sq value,
which indicates that non-significant terms are not present in both
the models. The R-Sq (pred) value is the variability in the data
accounted by the model in percentage. The R-Sq (pred) value
obtained for cylindricity and perpendicularity models is near to 1.
This indicates that high correlation exists between experimental

Table 5
ANOVA for perpendicularity.

Source Degree of freedom Sum of squares Adj. sum of squares adj. mean squares F P % Contribution

A 2 0.05186 0.05186 0.02593 598.20 0.000 78.89

B 2 0.00563 0.00563 0.002815 64.94 0.000 8.56
C 2 0.00459 0.00459 0.00295 52.95 0.000 6.98
A×B 4 0.00202 0.00202 0.00050 11.70 0.000 3.07
A×C 4 0.00002 0.00002 0.000005 0.13 0.969 0.03
B×C 4 0.00031 0.00031 0.000077 1.79 0.160 0.47
A×B×C 8 0.00011 0.00011 0.0000143 0.33 0.947 0.16

Error 27 0.00117 0.00117 0.000043 1.78

Total 53 0.06573

R2 = 98.22%. R2 (adj) = 96.5%.

Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
G Model
PRE-6331; No. of Pages 11 ARTICLE IN PRESS
8 S. Sheth, P.M. George / Precision Engineering xxx (2016) xxx–xxx

Fig. 9. (a) Main effects plot for cylindricity and perpendicularity. (b) Interaction plot for cylindricity and perpendicularity.

Table 6 was used to normalize the data. Eq. (4) is used for normalizing the
Errors during prediction.
data [32–35].
Model Minimum Maximum Average
error (%) error (%) error (%) max Xi 0 (k) − Xi 0 (k)
Xi ∗ (k) = (4)
Cylindricity 0.21 6.62 2.34 max Xi 0 (k) − min Xi 0 (k)
Perpendicularity 0.36 5.88 2.39
here i = 1 . . . m; k = 1 . . . n; m is the number of experimental data
item and n is the number of parameters. Xi 0 (k) denotes the original
sequence, Xi * (k) the sequence after the data pre-processing, max
and predicted values. The model is also adequate in prediction Xi 0 (k) is the largest value of Xi 0 (k), min Xi 0 (k) is the minimum value
owing to small value of PRESS, which is a measure of how well of Xi 0 (k). The normalized values are tabulated in Table 7.
the model is able to predict the data [30]. Values predicted by the Step 2: Calculating the grey relational co-efficient & grey rela-
regression model are very close to experimental values. Percent- tional grade
age errors are shown in Table 6. Error between the predicted and After data pre-processing, the grey relation coefficient i (k) was
experimental readings for each run is shown in Fig. 10(a) and (b) calculated for performance characteristics and is expressed by the
for cylindricity and perpendicularity respectively. following equation:

min + ς max
ς i (k) (5)
5. Grey relational analysis oi (k) + ς max

where oi is the deviation sequence.

From Fig. 9(a) it seems that to obtain the minimal cylindricity
 
oi (k) =  xo ∗ (k) − Xi ∗ (k)
lower spindle speed is required and to minimize perpendicular-
ity error maximum speed is required, which leads to trade off of
spindle speed and leads to multi objective optimization. Grey rela-
 
tional analysis (GRA) gives the optimum process parameters to get max (k) = max max  xo ∗ (k) − Xi ∗ (k)
the desire responses in multi objective optimization [14,29,30].
Step 1: Normalize the data  
The data to be used in grey analysis must be preprocessed into min (k) = min min  xo ∗ (k) − Xi ∗ (k)
quantitative indices for normalizing raw data for another analy-
sis. Preprocessing raw data is a process of converting an original here Xo * (k) denotes the reference sequence and Xi * (k) denotes the
sequence into a decimal sequence between a 0.00 and 1.00 for comparability sequence.  is distinguishing or identification coef-
comparison. Since the objective is to minimize cylindricity and per- ficient:  ∈ [0, 1] (the value may be adjusted based on the actual
pendicularity, so “lower is better” characteristics of grey relational system requirements).  = 0.5 is considered for these analysis. The

Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
G Model
PRE-6331; No. of Pages 11 ARTICLE IN PRESS
S. Sheth, P.M. George / Precision Engineering xxx (2016) xxx–xxx 9

7 Table 8
Grey relational coefficients and grades.
6
Run order Grey relational coefficients Grey relational grade
5 Cylindricity Perpendicularity

1 0.857142857 0.502057613 0.6796

% Error

4
2 0.923076923 0.463878327 0.693478
3 3 0.743362832 0.419243986 0.581303
4 0.831683168 0.435714286 0.633699
2 5 1 0.375384615 0.687692
6 0.777777778 0.358823529 0.568301
1 7 0.770642202 0.358823529 0.564733
8 0.848484848 0.339832869 0.594159
0 9 0.717948718 0.333333333 0.525641
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 10 0.913043478 0.757763975 0.835404
Run Order 11 0.943820225 0.632124352 0.787972
12 0.736842105 0.552036199 0.644439
(a)
13 0.682926829 0.61928934 0.651108
7 14 0.79245283 0.542222222 0.667338
15 0.682926829 0.484126984 0.583527
6 16 0.556291391 0.562211982 0.559252
17 0.571428571 0.525862069 0.548645
5 18 0.502994012 0.495934959 0.499464
19 0.823529412 1 0.911765
20 0.831683168 0.797385621 0.814534
% Eroro

4
21 0.677419355 0.685393258 0.681406
3 22 0.552631579 0.84137931 0.697005
23 0.583333333 0.726190476 0.654762
2 24 0.48000000 0.663043478 0.571522
25 0.362068966 0.802631579 0.58235
1 26 0.365217391 0.709302326 0.53726
27 0.333333333 0.663043478 0.498188
0
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
Run Order
(b)

Fig. 10. (a) Error percentage during predicting cylindricity for each run. (b) Error
percentage during predicting perpendicularity for each run.

Table 7
Normalized values.

Run order Cylindricity Perpendicularity

Smaller the better Smaller the better

1 0.9167 0.5041
2 0.9583 0.4221
3 0.8274 0.3074
Fig. 11. Run order vs grey relational grade.
4 0.8988 0.3525
5 1.0000 0.1680
6 0.8571 0.1066
7 0.8512 0.1066 calculated grey relational coefficients are tabulated in Table 8. Eq.
8 0.9107 0.0287
(6) gives values of grey relational grade.
9 0.8036 0.0000
10 0.9524 0.8402
1
n
11 0.9702 0.7090
12 0.8214 0.5943 i = ς i (k) (6)
n
13 0.7679 0.6926 k=1
14 0.8690 0.5779
15 0.7679 0.4672 The average values of the grey relational coefficients and grades
16 0.6012 0.6107
17 0.6250 0.5492
are shown in Table 8. From Fig. 11 it can be observed that the run
18 0.5060 0.4918 order 19 is having highest grey relational grade, means 19th run
19 0.8929 1.0000 order gives optimum set of cutting parameters for the cylindricity
20 0.8988 0.8730 and perpendicularity.
21 0.7619 0.7705
22 0.5952 0.9057
23 0.6429 0.8115
6. Conclusion
24 0.4583 0.7459
25 0.1190 0.8770
26 0.1310 0.7951 The response parameters like cylindricity and perpendicular-
27 0.0000 0.7459 ity in drilling of WCB have been measured under different cutting
conditions using 33 full factorial with twice replicates. The experi-
mental and analytical results lead to the following conclusions:

Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
G Model
PRE-6331; No. of Pages 11 ARTICLE IN PRESS
10 S. Sheth, P.M. George / Precision Engineering xxx (2016) xxx–xxx
• Spindle speed, feed, depth of cut and interaction between spindle
speed and feed are significant parameters (p < 0.05) for cylindric-
ity and perpendicularity.
• Minimal cylindricity on WCB material can be achieved by lower
spindle speed, lower feed and moderate depth of cut. Minimal
perpendicularity can be achieved by higher spindle speed, lower
feed and lower depth of cut.
• A second order regression model has been developed from the
measured experimental results. Fig. 8(a) and (b) show that the
predicted and experimental results are fairly close to each other.
The average prediction errors for cylindricity and perpendicular-
ity are 2.34% and 2.39% respectively. This justifies the use of model
for further prediction with more than 95% confidence level.
• Cylindricity is minimal at lower spindle speed while perpendic-
ularity is minimal at higher spindle speed, which leads to multi
objective optimization using GRA.
• The largest value of GRA is achieved at the spindle speed of
1600 rpm, feed rate of 100 mm/min and depth of cut of 1 mm.
Thus machining at this combination optimizes cylindricity and
perpendicularity.
Acknowledgement
The Authors would like to thank Cippriani Harrison Pvt. Ltd, V U
Nagar for providing resources during experimentation and to Flovel
Valves Pvt. Ltd., for helping and justifying the need of GD & T in the
context of machining parameter.
References
[1] Cogorno Gene R. Geometric dimensioning and tolerancing for mechanical
design. New York, NY: McGraw-Hill; 2011.
[2] Madsen David A. Geometric dimensioning and tolerancing: based on ASME
Y14.5-2009. Goodheart Willcox; 2010.
[3] ASME. ASME Y14.5-2009; 2015, http://mach.jlu.edu.cn/jxcx/standars/AGMA-
ANSI/ASME%20Y14.5M-2009.pdf (18/10/2015 at 2:00 pm).
[4] ISO. ISO 1101-2012; 2015, http://mach.jlu.edu.cn/jxcx/standars/ISO/ISO1101-
2012.pdf (18/10/2015 at 12.00 noon).
[5] Drake Jr Paul J. Dimensioning and tolerancing handbook. New York: McGraw-
Hill; 1999.
[6] Meadows JD. Geometric dimensioning and tolerancing. Marcel Dekker Inc.;
1995.
[7] Hinds BK, Ong TS. End milling of circular pockets to meet geometric tolerances.
J Mater Process Technol 2004;152:339–45.
[8] Ong TS, Hinds BK. The application of tool deflection knowledge in process
planning to meet geometric tolerances. Int J Mach Tools Manuf 2003;43:
731–7.
[9] Kumar J Pradeep, Packiaraj P. Effect of drilling parameters on surface roughness,
tool wear, material removal rate and hole diameter error in drilling of OHNS.
Int J Adv Eng Res Stud 2012;3(1):150–4.
[10] Shahrajabian H, Hadi M, Farahnakian M. Experimental investigation of machin-
ing parameters on machinability of carbon fibre/epoxy composites. Int J Eng
Innovative Technol 2012;2(3):30–6.
[11] Madhavan S, Balasivanadha Prabu S. Experimental investigation and analysis
of thrust force in drilling of carbon fibre reinforced plastic compos-
ites using response surface methodology. Int J Mod Eng Res 2012;2(4):
2719–23.
[12] Kilickap Erol. Determination of optimal parameter on delamination in
drilling of GFRP composite by Taguchi. Indian J Eng Mater Sci 2010;17:
265–74.
[13] Tyagi Yogendra, Chaturvedi Vedansh, Vimal Jyoti. Parametric optimization of
CNC drilling machine for mild steel using Taguchi design and single to noise
ratio analysis. Int J f Eng Sci Technol 2012;4(08):3758–66.
[14] Bharti Ashish, Moulick SK. Parametric optimization of multi response factors
in micro drilling operation. Int J Sci Eng Res 2013;4(7):1157–63.
[15] Palanikumar K, Prakash S, Shanmugam K. Evaluation of delamination in drilling
GFRP composites. Mater. Manuf. Processes 2008;23(8):858–64.
[16] Reddy S, Rao Srinivasa Ch. Application of grey relational analysis for surface
roundness error in drilling of Al 6061 alloy. Int J Lean Thinking 2012;3(2):
67–78.
[17] Kuram Emel, Ozcelik Babur. Fuzzy logic and regression modelling of cutting
parameters in drilling operation using vegetable based cutting fluid. Int J Eng
Mater Sci 2013;20:51–8.
[18] Sheth Saurin, Modi Bhavin S, George PM, Patel Pratik. A fuzzy logic based model
to predict MRR in flashing operation of precision steel ball manufacturing pro-
cess. Procedia Mater Sci 2014;5:1837–45.
Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002
[19] Patel Pratik J, Sheth Saurin, Chauhan Purvi. Effect of various parameters on
spread in flashing operation of precision steel ball manufacturing process. Pro-
cedia Mater Sci 2014;5:2224–32.
[20] Patel Pratik J, Sheth Saurin. Effect of various parameters on material removal
rate in flashing operation of precision steel ball manufacturing process. In:
First international and 16th national conference on machines and mechanisms
(iNaCoMM2013). 2013. p. 332–8.
[21] Tailor Arpit G, Patel Dipal, Sheth Saurin, Tailor Vikrant. Redesign the air con-
trol valve assembly by incorporating geometric dimensioning and tolerancing
(GD & T). In: Proceedings of the international conference on advances in pro-
duction and industrial engineering. 2015. p. 55–60, http://dx.doi.org/10.13140/
RG.2.1.4913.968.
[22] Misty Bhavesh, George PM, Sheth Saurin. Study and scope of DFMA and
GD&T in manufacturing process: a case study on dual plate check valve. In:
Proceedings of seventh international conference on advanced computing and
communication technologies. 4. 2013. p. 64–8, http://dx.doi.org/10.13140/
RG.2.1.4631.7600.
[23] Jayalakshmi George PM, Joshi SP. Modelling of GD&T requirements of
crankshaft flange using DOE. In: Proceedings of the fifth international & 26th
all India manufacturing technology, design and research conference (AIMTDR
2014), 310. 2014. p. 1–6.
[24] http://www.flovelvalves.com/ (15/7/2014, 22/9/2014).
[25] FM. Installation and operating manual for nozzle check vale. In: FM 20150923,
Rev:00. 2015.
[26] http://www.iscar.com/eCatalog/item.aspx?cat=5507922&fnum=2246&mapp
=DR&app=0&GFSTYP=M (18/4/2015 2:00 pm).
[27] http://int.haascnc.com/we spec1.asp?intLanguageCode=1033&id=VF2SS&
sizeID=30 40INCH VM pp. 67-78C (6/1/2015 2:00 pm).
[28] Wen Xiu-Lan, Huang Jia-Cai, Sheng Dang-Hong, Wang Feng-Lin. Conicity and
cylindricity error evaluation using particle swarm optimization. Precis Eng
2010;34:338–44.
[29] Xianqing Lei, Hongwei Song, Yujun Xue, Jishun Li, Jia Zhou, Mingde Duan.
Method for cylindricity error evaluation using geometry optimization search-
ing algorithm. Measurement 2011;44:1556–63.
[30] Douglas C Montgomery. Design and analysis of experiments. 7th ed. India:
Wiley; 2010. Wiley Student Edition.
[31] Antony Jiju. Design of experiments for engineers and scientists. Elsevier Science
and Technology Books; 2003.
[32] Senthilkumar N, Tamizharasan T, Anandakrishnan V. A hybrid Taguchi-grey
relational technique and cuckoo search algorithm for multi-criteria optimiza-
tion in hard turning of AISI D3 steel. J Adv Eng Res 2014;1(1):16–31.
[33] Senthilkumar N, Tamizharasan T, Anandakrishnan V. Experimental investi-
gation and performance analysis of cemented carbide inserts of different
geometries using Taguchi based grey relational analysis. Measurement
2014;58:520–36.
[34] Singh Hari, Kamboj Abhishek, Kumar Sudhir. Multi response optimization in
drilling Al6063/SiC/15% metal matrix composite. Int J Chem Nuclear Metall
Mater Eng 2014;8(4):276–81.
[35] Aydin Hakan, Bayram Ali, Esme Ugur, Kazancoglu Yigit, Guven Onur.
Application of grey relational analysis (GRA) and Taguchi method for the para-
metric optimization of friction stir welding process. Mater Technol 2010;44:
205–11.
[36] Hocken R, editor. Technology of machine tools, machine tool accuracy, UCRL-
52960-5, vol. 5. Livermore, CA: Lawrence Livermore National Laboratory; 1980.
[37] Bryan J. International status of thermal error research. CIRP Ann
1990;39(2):645–56.
[38] Schmitz TL, Ziegert JC, Canning S, Zapata R. Case study: a comparison of error
sources in high-speed milling. Precis Eng 2008;32:126–33.
[39] Ramesh R, Mannan MA, Poo AN. Error compensation in machine tools—a
review. Part I: Geometric, cutting-force induced and fixture-dependent errors.
Int J Mach Tools Manufa 2000;40:1235–56.
[40] Ramesh R, Mannan MA, Poo AN. Error compensation in machine tools—a
review. Part II: Thermal errors. Int J Mach Tools Manufa 2000;40:
1257–84.
[41] Ramesh R, Mannan MA, Poo AN. Thermal error measurement and modelling
in machine tools. Part I: Influence of varying operation conditions. Int J Mach
Tools Manufa 2003;43:391–404.
[42] Pancholi Akanksha. Investigation of cutting parameters during drilling of Mild
steel in context of cylindricity and perpendicularity. India: BVM Engineering
College; 2014. Thesis.
[43] Sheth Saurin, Chauhan Purvi, Modi Bhavin S, Pancholi Akanksha. Study
and investigate effect of cutting parameters in drilling on cylindricity
and perpendicularity. In: Proceedings of the fourth national confer-
ence on recent advances in manufacturing (RAM-2014). 2014. p. 43–8,
http://dx.doi.org/10.13140/RG.2.1.4631.7600.
[44] Moroni Giovanni, Petro Stefano. Inspection strategies and multiple geometric
tolerances. Procedia CIRP 2013;10:54–60.
[45] Moroni Giovanni, Petro Stefano. Geometric tolerance evaluation: a discussion
on minimum zone fitting algorithms. Precis Eng 2008;32(3):232–7.
[46] Hocken RJ, Pereira PH, editors. Coordinate measuring machines and systems.
2nd ed. CRC Press; 2012. ISBN-13: 978-1574446524.
[47] http://www.gdandtbasics.com/perpendicularity/ (29/10/2015 at 5:14 pm).
[48] Armillotta Antonil, Moroni Giovanni, Poloni Wilma. To analytically estimate
the 3D position deviation of a holes pattern due to fixturing. Procedia CIRP
2013;10:186–93.
G Model
PRE-6331; No. of Pages 11 ARTICLE IN PRESS
S. Sheth, P.M. George / Precision Engineering xxx (2016) xxx–xxx 11

[49] Adamczak Stanislaw, Janecki Dariusz, Makiela Wlodzimierz, Stepien
Krzysztof. Quantitative comparison of cylindricity profiles measured with
different methods using Legendre–Fourier coefficients. Metrol Meas Syst
2010;XVII(3):397–404.
[50] Adamczak Stanislaw, Janecki Dariusz, Stepien Krzysztof. Concept of reference
measurements of cylindricity profiles of machine parts. In: Proceedings, XVII
IMEKO world congress. 2003. p. 1764–8.
[51] Samper Serge, Adragna A Pierre, Favreliere Hugues, Pillet Maurice. Modeling
of 2D and 3D assemblies taking into account form errors of plane surfaces. J
Comput Inf Sci Eng 2009;9(4).
[52] Sheth Saurin, George PM. Experimental investigation and fuzzy modelling of
flatness and surface roughness for WCB material using face milling opera-
tion. In: CAD/CAM, Robotics and Factories of the Future. Springer India; 2016.
p. 769–77.

Please cite this article in press as: Sheth S, George PM. Experimental investigation, prediction and optimization of
cylindricity and perpendicularity during drilling of WCB material using grey relational analysis. Precis Eng (2016),
http://dx.doi.org/10.1016/j.precisioneng.2016.01.002