Fractal Analysis in Machining PDF

SpringerBriefs in Applied Sciences and Technology
Computational Mechanics
For further volumes:

http://www.springer.com/series/8886
Prasanta Sahoo Tapan Barman
•
João Paulo Davim
Fractal Analysis
in Machining
123
Prasanta Sahoo João Paulo Davim
Department of Mechanical Engineering Department of Mechanical Engineering
Jadavpur University University of Aveiro
Kolkata 700032 Campus Universitário de Santiago
India 3810-193 Aveiro
e-mail: psjume@gmail.com Portugal
e-mail: pdavim@ua.pt
Tapan Barman
Department of Mechanical Engineering
Jadavpur University
Kolkata 700032
India
e-mail: tkbarman@gmail.com
ISSN 2191-5342 e-ISSN 2191-5350

ISBN 978-3-642-17921-1 e-ISBN 978-3-642-17922-8
DOI 10.1007/978-3-642-17922-8
Springer Heidelberg Dordrecht London New York
Ó Prasanta Sahoo 2011

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcast-
ing, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this
publication or parts thereof is permitted only under the provisions of the German Copyright Law of
September 9, 1965, in its current version, and permission for use must always be obtained from
Springer. Violations are liable to prosecution under the German Copyright Law.
The use of general descriptive names, registered names, trademarks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
Cover design: eStudio Calamar, Berlin/Figueres
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

Preface
The present book deals with fractal analysis of surface roughness in different
machining processes. Surface roughness is an important attribute of any machine
component. Conventionally several statistical roughness parameters are used for
describing surface roughness. But surface topography is a non-stationary random
process for which the variance of the height distribution of roughness features is
related to the length of the sample. Consequently, instruments with different res-
olutions and scan lengths yield different values of these statistical parameters for
the same surface. A logical solution to this problem is to use scale-invariant
parameters to characterize rough surfaces. In this context, to describe surface
roughness, the concept of fractals is considered. Fractals retain all the structural
information and are characterized by single descriptor, the fractal dimension,
D. Fractal dimension is intrinsic property of the surface and independent of the
filter processing of measuring instrument as well as the sampling length scale.
Four machining processes viz. CNC end milling, CNC turning, electrical dis-
charge machining and cylindrical grinding are considered for three different
materials. The generated machined surfaces are measured to find out fractal
dimension (D) of the surfaces. The experimental results are further analyzed with
response surface methodology (RSM) to consider the effects of process parameters
on fractal dimension. Also the effect of work-piece material variation on fractal
dimension of machined surfaces is considered. It is believed that the present book
will prove to add significant contribution to the existing literature from the point of
view of both industrial importance and academic interest.
v
Contents
1 Fundamental Consideration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Surface Metrology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Fractal Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.1 Fractal Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.2 Fractal Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.3 Self-Similarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.4 Self-Affinity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3.5 Fractal Description of Roughness . . . . . . . . . . . . . . . . . 9
1.3.6 Fractal Dimension Calculation . . . . . . . . . . . . . . . . . . . 11
1.3.7 Fractal Dimension Measurement in the Present Study . . . 12
1.4 Review of Roughness Study in Machining . . . . . . . . . . . . . . . . 12
1.5 Design of Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.5.1 Full Factorial Design . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.5.2 Central Composite Design . . . . . . . . . . . . . . . . . . . . . . 19
1.6 Response Surface Methodology. . . . . . . . . . . . . . . . . . . . . . . . 21
1.7 Closure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2 Fractal Analysis in CNC End Milling . . . . . . . . . . . . . . . . . . . . . . 29

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.2 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.1 Design of Experiments . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.2 Machine Used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.3 Cutting Tool Used. . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.2.4 Work-Piece Materials . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3 Results and Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.3.1 RSM for Mild Steel . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.3.2 RSM for Brass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3.3 RSM for Aluminium . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.4 Closure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
vii
viii Contents
3 Fractal Analysis in CNC Turning . . . . . . . . . . . . . . . . . . . . . . . . . 45

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.2.2 Machine Used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2.3 Cutting Tool Used. . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3.2 RSM for Brass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.4 Closure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4 Fractal Analysis in Cylindrical Grinding . . . . . . . . . . . . . . . . . . . 57

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.2.2 Machine Used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.3.2 RSM for Brass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.4 Closure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5 Fractal Analysis in EDM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.2.2 Machine Used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.2.4 Tool Electrode Used . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.3.2 RSM for Brass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.3.3 RSM for Tungsten Carbide . . . . . . . . . . . . . . . . . . . . . 77
5.4 Closure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Chapter 1
Fundamental Consideration
Abstract The importance and usefulness of fractal dimension in describing sur-

face roughness over the conventional roughness parameters are presented in this
chapter. The fundamental of fractal dimension and the methodology for evaluation
of fractal dimension are also discussed. Literature survey is carried out for four
different types of machining processes and shows that there is scarcity of litera-
tures which deal with fractal description of surface roughness. Fundamentals of
design of experiments and response surface methodology are also discussed.
1.1 Introduction
Surfaces are irregular though they may look like very smooth. When the surfaces
are magnified, the irregularities become prominent. This is true for the machining
surfaces as well. In a material removal process such as machining, unwanted
material is removed and altered surface topography is obtained. The surface
generated consists of inherent irregularities left by the cutting tool, which are
commonly defined as surface roughness. Such a surface is composed of a large
number of length scales of superimposed roughness that are generally character-
ized by the standard deviation of surface peaks. Three statistical characteristics are
generally used to describe the structure of machined surface topography: texture,
waviness and roughness. The texture determines the anisotropic property of the
surface. The waviness reflects the reference profile (or surface). The surface
roughness is formed by the micro deformation during the machining process.
Surface roughness plays an important role. It has large impact on the
mechanical properties like fatigue behavior, corrosion resistance, creep life, etc.
It also affects other functional attributes of machine components like friction,
wear, light reflection, heat transmission, lubrication, electrical conductivity, etc.
Surface roughness may depend on various factors like machining parameters,
work-piece materials, cutting tool properties, cutting phenomenon, etc. In a review
P. Sahoo et al., Fractal Analysis in Machining, 1

SpringerBriefs in Computational Mechanics,
DOI: 10.1007/978-3-642-17922-8_1, Prasanta Sahoo 2011
2 1 Fundamental Consideration
article, Benardos and Vosniakos (2003) have presented a fishbone diagram with
parameters that affect surface roughness. As a case study, they have considered
two machining operations—turning and milling. They broadly classified the fac-
tors as machining parameters, cutting tool properties, work-piece properties and
cutting phenomena. Machining parameters may include process kinematics, depth
of cut, cutting speed, feed rate, etc. Cutting tool properties may include tool
material, nose radius, tool shape, etc. Work-piece properties may include work-
piece hardness, work-piece size etc. and cutting phenomena includes vibration,
cutting force variations, chip formation, etc. It is obvious that for other machining
operations also, there are several factors that affect surface roughness. Many
researchers have attempted to model surface roughness but the developed models
are far from complete as it is not possible to consider all the controlling factors in a
particular study. So, researchers always pay attention to model surface roughness
in a better way so that surface roughness modeling can be done more accurately.
Surface roughness is generally expressed by three types of conventional
roughness parameters viz. amplitude parameters, spacing parameters and hybrid
parameters. Amplitude parameters are the measure of vertical characteristics of
surface deviation. Center line average roughness (Ra), root mean square roughness
(Rq), etc. are the examples of these types of parameters. Spacing parameters are
measures of the horizontal characteristics of surface deviations. Examples of such
parameters are mean line peak spacing (Rsm), high spot count, etc. On the other
hand, hybrid parameters are the combination of both vertical and horizontal
characteristics of the surface deviations e.g. root mean square slope of the profile,
root mean square wavelength, peak area, valley area, etc. Most commonly used
roughness parameters are centre line average value (Ra), root mean square value
(Rq), mean line peak spacing (Rsm), etc.
Conventionally, the deviation of a surface from its mean plane is assumed to be
a random process for which statistical parameters such as the variances of the
height, the slope and curvature are used for characterization (Nayak 1971).
However, it has been found that the variances of slope and curvature depend
strongly on the resolution of the roughness-measuring instrument or any other
form of filter and are hence not unique (Thomas 1982; Bhushan et al. 1988;
Majumdar and Tien 1990). It is also well known that surface topography is a non-
stationary random process for which the variance of the height distribution is
related to the length of the sample (Sayles and Thomas 1978). Consequently,
instruments with different resolutions and scan lengths yield different values of
these statistical parameters for the same surface. The conventional methods of
characterization are therefore fraught with inconsistencies which give rise to the
term ‘parameter rash’ (Whitehouse 1982) commonly used in contemporary liter-
ature. The underlying problem with the conventional methods is that although
rough surfaces contain roughness at a large number of length scales, the charac-
terization parameters depend only on a few particular length scales, such as the
instrument resolution or the sample length. A logical solution to this problem is to
use scale-invariant parameters to characterize rough surfaces. In this context, to
describe surface roughness, the concept of fractals is applied. The concept is based
1.1 Introduction 3
on the self-affinity and self-similarity of surfaces at different scales. Fractals retain

all the structural information and are characterized by single descriptor, the fractal
dimension, D. Fractal dimension is intrinsic property of the surface and inde-
pendent of the filter processing. Roughness measurements on a variety of surfaces
show that the power spectra of the surface profiles follow power laws. This sug-
gests that when a surface is magnified appropriately, the magnified image looks
very similar to the original surface. This property can be mathematically described
by the concepts of self-similarity and self-affinity. The fractal dimension, which
forms the essence of fractal geometry, is both scale-invariant and is closely linked
to the concepts of self-similarity and self-affinity (Mandelbrot 1982). It is therefore
essential to use fractal dimension to characterize rough surfaces and provide the
geometric structure at all length scales (Bigerelle et al. 2005). The possible
application of fractal geometry to tribology was explored (Ling 1990). The
influence of processing techniques on the fractal or non-fractal structure was also
examined (Majumdar and Bhushan 1990).
In a material removal process, mechanical intervention happens over length
scales which extend from atomic dimensions to centimeters. The machine vibra-
tion, clearances and tolerances affect the outcome of the process at the largest of
length scales (above 10-3 m). The tool form, feed rate, tool radius in the case of
single point cutting (Venkatesh et al. 1998) and grit size in multiple point cutting
(Venkatesh et al. 1999), affect the process outcome at the intermediate length
scales (10-6–10-3 m). The roughness of the tool or details of the grit surfaces
influence the final topography of the generated surface at the lowest length scales
(10-9–10-6 m). It has been shown that surfaces formed by electric discharge
machining, milling, cutting or grinding and worn surfaces (Brown and Savary
1991; Tricot et al. 1994; Hasegawa et al. 1996; He and Zhu 1997; Ge and Chen
1999; Zhang et al. 2001; Jiang et al. 2001; Zhu et al. 2003; Jahn and Truckenbrodt
2004; Kang et al. 2005; Han et al. 2005) have fractal structures, and fractal
parameters can reflect the intrinsic properties of surfaces to overcome the disad-
vantages of conventional roughness parameters. Thus, to characterize the rough-
ness of machined surfaces in different machining processes fractal dimension is
used as the roughness parameter.
1.2 Surface Metrology
Surface texture is a complex condition resulting from a combination of roughness

(nano and micro-roughness), waviness (macro-roughness), lay and flaw. Figure 1.1
shows a display of surface texture with unidirectional lay. Roughness is produced
by fluctuations of short wavelengths characterized by asperities (local maxima)
and valleys (local minima) of varying amplitudes and spacing. This occurs due to
the mechanism of the material removal process. Waviness is the surface irregu-
larities of longer wavelengths and may result from such factors as machine or work
piece deflections, vibration, chatter, heat treatment or warping strains. Lay is the
Fig. 1.1 Display of surface

texture
principal direction of the predominant surface pattern, usually determined by the

production process. Flaws are unexpected and unintentional interruptions in the
texture. Apart from these, the surface may contain large deviations from nominal
shape of very large wavelength, which is known as error of form. These are not
considered as part of surface texture.
Any engineering surface composes of a vast number of peaks and valleys and it
is not possible to measure the height and location of each of the peaks. So mea-
surement of a surface is carried out on a sampling length where it is assumed that
the surface outside and inside the sampling length is statistically similar. In order
to determine the numerical assessment of a sample’s surface texture, three char-
acteristic lengths are associated with the profile (ISO 4287, 1997) viz. sampling
length, evaluation or assessment or cut off length and traverse length. The sam-
pling length is the length over which the parameter to be measured will have
statistical significance. Cut off length is the length of the surface over which the
1.2 Surface Metrology 5
Motor and gearbox

Transducer
Skid
Amplifier
Stylus
A-D Chart
converter recorder
Specimen
Data logger
Fig. 1.2 Component parts of a typical stylus surface-measuring instrument
measurement is made. This length may include several sampling lengths—

typically five times. The measurement is the integration of the individual sampling
lengths. The total length of the surface traversed by the stylus in making a mea-
surement is called the traverse length. It will normally be greater than the eval-
uation length, due to the necessity of allowing run-up and over-travel at each end
of the evaluation length to ensure that any mechanical and electrical transients are
excluded from the measurement.
There are several methods to study the surface topography which are developed
over the years. The most common method of studying surface texture is the surface
profilometer (Fig. 1.2). In this method, a fine, very lightly loaded, stylus is traveled
smoothly at a constant speed across the surface under examination. The transducer
produces an electrical signal, proportional to displacement of the stylus, which is
amplified and fed to a chart recorder that provides a magnified view of the original
profile. But this graphical representation differs from the actual surface profile
because of difference in magnifications employed in vertical and horizontal
directions. Surface slopes appear very steep on profilometric record though they
are rarely steeper than 10 in actual cases. The shape of the stylus also plays a vital
role in incorporating error in measurement. The finite tip radius (typically 1–2.5
microns for a diamond stylus) and the included angle (of about 60 for pyramidal
or conical shape) results in preventing the stylus from penetrating fully into deep
and narrow valleys of the surface and thus some smoothing of the profile are done.
Some error is also introduced by the stylus in terms of distortion or damage of a
very delicate surface because of the load applied on it. In such cases non-
contacting optical profilometer having optical heads replacing stylus may be used.
Reflection of infrared radiation from the surface is recorded by arrays of photo-
diodes and analysis of the same in a microprocessor result in the determination of
the surface topography. Vertical resolution of the order of 0.1 nm is achievable
though maximum height of measurement is limited to few microns. This method is

clearly advantageous in case of very fine surface features.
1.3 Fractal Characterization
1.3.1 Fractal Geometry
Euclidean geometry describes ordered objects such as points, curves, surfaces and
cubes using integer dimensions of 0, 1, 2, and 3, respectively. A measure of the
object such as the length of a line, the area of a surface and the volume of a cube
are associated with each dimension. These measures are invariant with respect to
the unit of measurement. It means that the length of a line remains independent of
whether a centimeter or a micrometer scale is used. However, a multitude of
objects found in nature appear disordered and irregular for which the measures of
length, area and volume are scale-dependent. This suggests that the dimensions of
such objects cannot be integers. A generalized concept of a dimension and the
origins of fractal geometry are now discussed.
Mandelbrot (1967) founded fractal geometry when he showed that for
decreasing the unit of measurement, the length of a natural coastline does not
converge but, instead, increase monotonically. On plotting the length L as a
function of the unit of measurement [ on a log–log plot, he found a simple relation
of the form L * [1-D. Mandelbrot finally made an interesting conclusion that the
real number D associated with every coastline is the dimension of the coastline.
This study marked the origins of fractal geometry, which has now found numerous
applications in characterizing and describing disordered phenomena in science and
engineering.
1.3.2 Fractal Dimension
To measure the length of a line, let us break the line into small units of length
[ and then add the number of units in the form
L ¼ R 21 ð1:1Þ
Similarly to measure the area of surface, let us break up the surface into small
squares of size [ 9 [ and then add the number of units as
A ¼ R 22 ð1:2Þ
Here in Eqs. 1.1 and 1.2 the exponents 1 and 2 correspond to the dimensions of
the objects. These measures of length and area have a unique property that they are
independent of the unit of measurement [ and in the limit [ ? 0 these measures
1.3 Fractal Characterization 7
remain finite and non-zero. This concept of Euclidean dimension thus can be
generalized in the form
M ¼ R 2D ð1:3Þ
Here M is the measure and D is a real number. If the exponent D makes the
measure M independent of the unit of measurement [ in the limit of [ ? 0, then D
is the dimension of an object.
Contrary to common understanding of dimension, this generalization allows the
dimension of an object to take non-integer values. If, in this argument, it is
assumed that an object is broken into N equal parts then Eq. 1.3 can be written as
M = N[D. Since the measure is invariant with the unit of measurement, one can
write N * [-D. Now if the length of an object is evaluated, then the length would
vary as L = N[1 * [1-D, as was observed for the lengths of the coastlines. It can
be easily seen that the length will be independent of [ only when D = 1.
1.3.3 Self-Similarity
The generalized concepts of measure and dimension are fundamental to the issue
of self-similarity. Let us consider a one-dimensional line of unit length and break it
up into N equal segments. Each segment of the line, of size 1/m, is similar to the
whole line and needs a magnification of m to be an exact replica of the whole line.
Since the length of the line remains independent of 1/m, it follows that the number
of units is N * m. Now let us consider a square, which has a side of unit length.
Each small square of side 1/m is similar to the whole square and needs a mag-
nification of m to be an exact replica of the whole square. However, the number of
small squares in the whole is N * m2. In general, for an object of dimension D, it
follows that
N mD ð1:4Þ
Thus the dimension of the object can be written as
log N
D¼ ð1:5Þ
log m
This definition of dimension, which is based on the self-similarity of an object,
is called the similarity dimension. To perceive what an object of a non-integer
dimension looks like, one can follow the recursive construction in Fig. 1.3, which
yields the Koch curve of dimension 1.26.
In this construction the first step is to break a straight line into three parts and
replace the middle portion by two segments of equal lengths. In the subsequent
stages each straight segment is broken into three parts and the middle portion of
each segment is replaced by two parts. If this recursion is continued infinite times
then the Koch curve is obtained. This curve has some unique mathematical
properties. Firstly, the curve is continuous but it is not differentiable anywhere.
Fig. 1.3 Formation of Koch

curve
The non-differentiability arises because of the fact that if the curve is repeatedly
magnified, more and more details of the curve keep appearing. This means that
tangent cannot be drawn at any point and therefore the curve cannot be differ-
entiated. Secondly, the curve is exactly self-similar. This is because if a small
portion of the curve is appropriately magnified, it will be an exact replica of the
whole Koch curve. Thirdly, the dimension of the curve remains constant at all
scales, although the curve contains roughness at a large number of scales. This
scale-invariance of the dimension is an important property, which is utilized to
characterize rough surfaces. The coastline of an island is an example of a self-
similar object found in nature. Although these objects are not exactly self-similar,
they are statistically self-similar. Statistical self-similarity means that the proba-
bility distribution of a small part of an object will be congruent with the probability
distribution of the whole object if the small part is magnified appropriately.
However, not all fractal objects are self-similar. This leads to the more general
concept of self-affinity.
1.3.4 Self–Affinity
The definition of self-similarity is based on the property of equal magnification in

all directions. However, there are many objects in nature, which have unequal
Fig. 1.4 Qualitative description of statistical self-affinity for a surface profile
scaling in different directions. Thus these are not self-similar but self-affine. The
dimension of self-affine fractals cannot be obtained from Eq. 1.5, which is based
on the self-similarity of an object. Mandelbrot showed that the lengths of self-
affine fractal curves do not follow the relation L * [1-D for all values of [ and
therefore the dimension of self-affine curves cannot be obtained by measuring their
lengths. Instead, the dimension of self-affine functions can be obtained from their
power spectra.
1.3.5 Fractal Description of Roughness
The deviation of a surface from its mean plane is assumed to be a random process,
which is characterized by the statistical parameters such as the variance of the
height, the slope and the curvature. But, it has been observed that surface
topography is a non-stationary random process. It means the variance of the height
distribution is related to the sampling length and hence is not unique for a
particular surface. Rough surfaces are also known to exhibit the feature of geo-
metric self-similarity and self-affinity, by which similar appearances of the surface
are seen under the various degrees of magnification as quantitatively shown in
Fig. 1.4. Since increasing amounts of detail in the roughness are observed at
decreasing length scale, the concepts of slope and curvature, which inherently
assume the smoothness of the surface, cannot be defined. So the variances of slope
and curvature depend strongly on the resolution of the roughness-measuring
instrument or some other form of filter and are therefore not unique. In contem-
porary literature such a large number of characterizations parameter occurs that the
term ‘parameters rash’ is aptly used. The use of instrument-dependent parameters
shows different values for the same surface. Thus, it is necessary to characterize
rough surfaces by intrinsic parameters, which are independent of all scales of
roughness. This suggests the use of fractal geometry in characterizing the surface
roughness. The fractal dimension is an intrinsic property and should be used for
surface characterization. It is invariant with length scales and is closely linked to
the concept of geometric self-similarity.
The self-similarity or self-affinity of rough surfaces implies that as the unit of
measurement is continuously decreased, the surface area of the rough surface (a
two-dimensional measure) tends to infinity and the volume (a three-dimensional
measure) tends to zero. Here, self-similarity implies the property of equal mag-
nification in all directions while self-affinity refers to unequal scaling in different
directions. Thus, the Hausdorff or fractal dimension, D ? 1, of rough surfaces is a
fraction between 2 and 3. The profile of a rough surface z(x), typically obtained
from stylus measurements, is assumed to be continuous even at the smallest scales.
This assumption breaks down at atomic scale. But for engineering surfaces the
continuum is assumed to exist down to the limit of a zero-length scale. Since
repeated magnifications reveal the finer levels of detail, the tangent at any point
cannot be defined. Thus the surface profile is continuous everywhere but non-
differentiable at all points. This mathematical property of continuity, non-differ-
entiability and self-affinity (Berry and Lewis 1980) is satisfied by the modified
Weierstrass–Mandelbrot (W–M) fractal function, which is thus used to charac-
terize and simulate such profiles. The W–M function has a fractal dimension D,
between 1 and 2, and is given by
Xa
cos 2pcn x
zðxÞ ¼ GðD1Þ ð2DÞn
1\D\2; c[1 ð1:6Þ
n¼n1 c
where, G is a scaling constant. The parameter n1 corresponds to the low cut-off

frequency of the profile. Since surfaces are non-stationary random process the
lowest cut-off frequency depends on the length L of the sample and is given by
cn1 = 1/L.
The W–M function has the interesting mathematical property that the series for
z(x) converges but that for dz/dx diverges. It implies that it is non-differentiable at
all points. The power spectrum of this W–M function can be expressed by a
continuous function as
G2ðD1Þ 1
SðxÞ ¼ ð1:7Þ
2 ln c x52D
When this equation is compared with the power spectrum of a surface, the
dimension D is related to the slope of the spectrum on a log–log plot against x.
The constant G is the roughness parameter of a surface, which is invariant with
respect to all frequencies of roughness and determines the position of spectrum
along the power axis. In this characterization method both G and D are independent
of the roughness scales of the surface and hence intrinsic properties. The constants
of the W–M function, G, D, and n1 form a complete and fundamental set of scale-
independent parameters to characterize a rough surface. The physical significance
of D is the extent of space occupied by the rough surface, i.e., larger D values
correspond to denser profile or smoother topography (Yan and Komvopoulos 1998;
Sahoo and Ghosh 2007).
1.3.6 Fractal Dimension Calculation
Fractal calculation mainly includes the calculation of profile fractal dimension

(1 \ D \ 2) and the calculation of surface fractal dimension (2 \ D \ 3). Fractal
calculation is generally involved with computer assisted image analysis of
topography images of a surface obtained in analog or digital signals using profi-
lometer or microscopy, etc. An effective method to convert these signals into the
required data for calculating fractal dimensions must therefore be sought. Profile
instruments can be used to obtain data, which are then directly used to calculate
fractal dimension. The methods for calculating profile fractal dimension mainly
include the yardstick, the box counting, the variation, the structure function and the
power spectrum method (Sahoo 2005).
The yardstick method employs the technique of ‘walking’ around a profile
curve in a step length, r. A point on the profile curve is chosen as a starting point of
divider, whilst another point at a distance r from the starting point is taken as its
end point. Repetitively, find the point-pair of dividers in the same way until the
profile curve is entirely measured. Then, the summing up of the step lengths
enables the curve length to be determined. The repetition of this calculation
process at various step lengths allows all the curve length to be evaluated. Further,
plotting of the curve lengths verses the step lengths on a log–log scale gives the
slope m of a fitting line to be related to the fractal dimension D as D = 1 - m.
It is possible that this method has abandoned some pivotal points, resulting in
calculation error.
The principle of box counting method mainly involves an iteration operation to
an initial square, whose area is supposed to be 1 and which covers the entire graph.
The initial square is divided into four sub-squares and so on. After the n times
operations, the number of sub-squares, which contain the discrete points of the
profile graph are counted and the length L of the profile is approximately obtained.
Then the fractal dimension is calculated as D = 1 ? log L/(n.log2).
The variation method has the advantage of being proven theoretically for all
profiles (self-affine or not), and of giving quickly an estimation of the dimension of
mathematical profiles. A well-known technique used to analyze surfaces consists
in performing ‘slices’ through the surfaces, which allows one to transform a three-
dimensional problem to two-dimensional problem. In other words, a surface is
replaced by profiles, taken at different places, and the fractal dimension estimated
over profiles is then related to the three-dimensional fractal dimension by the
classical result: dimension of surface = 1 ? dimension of profiles. Such a tech-
nique obviously decreases the problem size. Accurate results are hard to obtain for
the surface dimension and the variation method gives the best approximations.
The variation method algorithm is based on the local oscillation of the profile
function Z.
The power spectrum method involves the evaluation of the power of the profile
function. The modified Weierstrass–Mandelbrot (W–M) function for a rough
surface is described by Eq. 1.6. The multi-scale nature of z(x) can be characterized
by its power spectrum, which gives the amplitude of the roughness at all length
scales. The parameters G and D can be found from the power spectrum of the W–
M function given by Eq. 1.7. Usually, the power law behavior would result in a
straight line if S(x) is plotted as a function of x on a log–log graph. Using fast
Fourier transform (FFT), the power spectrum of profile can be calculated and then
be plotted verses the frequency on a log–log scale. Thereafter, the fractal
dimension, D, can be related to the slope m of a fitting line on a log–log plot as:
D = (5 ? m).
The structure function method considers all points on the surface profile curve
as a time sequence z(x) with fractal character. The structure function s(s) of
sampling data on the profile curve can be described as s(s) = [z(x ? s) -
z(x)]2 = cs 4 - 2D where [z(x ? s) - z(x)]2 expresses the arithmetic average value
of difference square, and s is the random choice value of data interval. Different s
and the corresponding s(s) can be plotted verses the s on a log–log scale. Then, the
fractal dimension D can be related to the slope m of a fitting line on log–log plot
as: D = (4 - m).
1.3.7 Fractal Dimension Measurement in the Present Study
In the present study, roughness profile measurement is done using a stylus-type

profilometer, Talysurf (Taylor Hobson, UK). The profilometer is set to a cut-off
length of 0.8 mm, Gaussian filter, traverse speed 1 mm/sec and 4 mm traverse
length. Roughness measurements, in the transverse direction, on the work pieces
are repeated four times and average of four measurements of surface roughness
parameter values is recorded. The measured profile is digitized and processed
through the dedicated advanced surface finish analysis software Talyprofile. Then
fractal dimension is evaluated following the structure function method.
1.4 Review of Roughness Study in Machining
As surface roughness is an important parameter in the industry, many researchers

have tried to study surface roughness in machining. Though the present study
focuses on fractal dimension in describing surface roughness, both conventional
roughness parameters and fractal dimension are reviewed here. Four machining
processes viz. turning, grinding, milling and electrical discharge machining are
focused for this purpose and presented one by one.
In turning, many researchers have modeled surface roughness. Grzesik (1996)
has studied the effect of tribological interactions at the interface between the chip
and tool on surface roughness in finish turning of C45 carbon steel. Yang and
Tarng (1998) have showed that feed rate is the most significant factor affecting
surface roughness in S45C steel turning. Also, with increasing feed rate, surface
1.4 Review of Roughness Study in Machining 13
roughness decreases. Abouelatta and Madl (2001) have found a correlation

between surface roughness and cutting parameters and tool vibrations in turning
considering three conventional roughness parameters viz. center line average
roughness value, maximum height of the profile and skewness. Davim (2001) has
presented a study of the influence of cutting parameters on the surface roughness
obtained in turning of free machining steel using Taguchi design and shown that
the cutting velocity has a greater influence on the roughness followed by the feed
rate. Lin et al. (2001) have shown that in turning feed rate is the critical parameter
to affect the surface roughness, where increasing the feed rate will increase the
surface roughness. Suresh et al. (2002) have shown that surface roughness
decreases with an increase in cutting speed, and increases as feed increases in
turning of mild steel. Arbizu and Perez (2003) have developed models to deter-
mine surface quality of parts obtained through turning processes and shown that
surface roughness increases with increase in depth of cut and feed rate. Feng and
Wang (2003) have presented a nonlinear multiple regression analysis to predict
surface roughness in finish turning of Steel 8620 and Al 6061T materials. Dabnun
et al. (2005) have concluded that feed rate is the main influencing factor on the
roughness in turning of machinable glass–ceramic (Macor). Sahin and Motorcu
(2005) have developed a surface roughness model for turning of mild steel with
coated carbide tools and shown that feed rate is the main affecting factor on
surface roughness. Surface roughness increases with increase in feed rate but
decreases with increase in cutting speed and depth of cut. Kirby et al. (2006) have
shown that the feed rate and tool nose radius have the highest effects on surface
roughness in a turning operation of 6061-T6 aluminium alloy. Palanikumar et al.
(2006) have focused on the parametric influence of machining parameters on the
surface roughness in turning of glass fiber reinforced polymer (GFRP) and shown
that roughness increases with increase in feed rate but roughness decreases with
increase in cutting speed. Singh and Rao (2007) have developed a model to
determine the effects of cutting conditions and tool geometry on surface roughness
in the finish hard turning of the bearing steel (AISI 52100) and concluded that feed
rate is the dominant factor determining surface finish followed by nose radius and
cutting velocity. Ramesh et al. (2008) have found in their study that feed rate is the
main influencing factor on surface roughness in turning of titanium alloy.
Palanikumar (2008) has found that the most significant machining parameter for
surface roughness is feed followed by cutting speed in machining glass fiber
reinforced (GFRP). For modeling surface roughness in turning different method-
ologies are used viz. RSM (Suresh et al. 2002; Dabnun et al. 2005; Sahin and
Motorcu 2005; Palanikumar et al. 2006; Singh and Rao 2007; Ramesh et al. 2008;
Palanikumar 2008; Gupta 2010), Taguchi analysis (Yang and Tarng 1998; Davim
2001; Kirby et al. 2006; Nalbant et al. 2007; Palanikumar 2008;), artificial neural
network (Pal and Chakraborty 2005; Kohli and Dixit 2005; Bagci and Isik 2006;
Abburi and Dixit 2006; Feng et al. 2006; Zhong et al. 2006; Zhong et al. 2008;
Muthukrishnan and Davim 2009; Karayel 2009; Gupta 2010; Chavoshi and Tajdari
2010). Also, the literature survey shows that mainly three cutting parameters viz.
cutting speed, feed rate and depth of cut are the common parameters considered for
most of the studies (Yang and Tarng 1998; Davim 2001; Lin et al. 2001; Suresh
et al. 2002; Arbizu and Perez 2003; Jiao et al. 2004; Dabnun et al. 2005; Sahin and
Motorcu 2005; Bagci and Isik 2006; Ramesh et al. 2008; Palanikumar 2008;
Karayel 2009).
Grinding is the most commonly used manufacturing process in the industry and
this is a complex machining process with many interactive parameters and surface
quality produced is influenced by various parameters. Several researchers have
tried to model surface roughness in grinding and few of the recent literatures are
reviewed here. Zhang et al. (2001) have developed the relationships between the
fractal dimension and conventional roughness parameters (Ra or Rq or Rsm of
surface roughness) of different ground surfaces and justified the usefulness of
fractal theory. They concluded that fractal dimension D is relative to vertical
parameters and transverse parameters of surface topography. Zhou and Xi (2002)
have developed a model for predicting surface roughness in grinding taking into
consideration the random distribution of the grain protrusion heights. Maksoud
et al. (2003) have used artificial neural network to achieve desired surface
roughness under grinding wheel surface topography variations. Hassui and Diniz
(2003) have developed a relation between the process vibration signals and
roughness in a plunge cylindrical grinding operation of AISI 52100 quenched and
tempered steel. Hecker and Liang (2003) have presented the prediction of the
arithmetic mean surface roughness based on a probabilistic undeformed chip
thickness model. Bigerelle et al. (2005) have shown that grinding could be char-
acterized with an elementary function and the worn profile can be modeled by a
fractal curve defined by only two parameters (amplitude and fractal dimension)
with an infinite summation of these elementary functions. Krajnik et al. (2005)
have used response surface methodology to develop a model to minimize the
surface roughness in plunge center less grinding operation of 9SMn28, free-cutting
unalloyed steel. The analysis of variance shows that the grinding wheel dressing
condition most significantly affects the ground surface roughness. The surface
roughness is additionally affected by the geometrical grinding gap set-up factor
and the control wheel speed. Kwak (2005) has investigated the various grinding
parameters affected the geometric error in surface grinding process using com-
bined Taguchi method and response surface method. Four grinding parameters
such as grain size, wheel speed, depth of cut and table speed are selected for
experimentation. A second-order response model for the geometric error is
developed and the utilization of the response surface model is evaluated with
constraints of the surface roughness and the material removal rate. Fredj and
Amamou (2006) have tried to establish a model combining the application of
design of experiments (DOE) and neural network method for ground surface
roughness prediction. Kwak et al. (2006) have developed a model for grinding
power spent during the process and the surface roughness in the external cylin-
drical grinding of hardened SCM440 steel using the response surface method.
They have shown from the study that the grinding power seems to increase linearly
with increasing work-piece speed and the traverse speed and surface roughness is
dominantly affected by the change of the work-piece speed. Choi et al. (2008) have
established the generalized model for power, surface roughness, grinding ratio and
surface burning for grinding of various steel alloys using alumina grinding wheels
based on the systematic analysis and experiments. It is seen that steady-state
surface roughness is primarily dependent only on the effective chip thickness.
Mohanasundararaju et al. (2008) have developed a neural network and fuzzy-based
methodology for predicting surface roughness in a grinding process for work rolls
used in cold rolling. This methodology predicts the most likely estimates of sur-
face roughness along with lower and upper estimates using fuzzy numbers.
Siddiquee et al. (2010) have investigated the optimization of an in-feed centreless
cylindrical grinding process performed on EN52 austenitic valve steel (DIN:
X45CrSi93) considering dressing feed, grinding feed, dwell time and cycle time as
process parameters. They have optimized the multiple responses viz. surface
roughness, out of cylindricity of the valve stem and diametral tolerance using grey
relational Taguchi analysis.
Milling also is a popular machining process in modern industry. There are
several researchers who have tried to model the roughness in milling process. In
this section, few available literatures on surface roughness modeling in milling are
reviewed. Fuh and Wu (1995) have developed a model for prediction of surface
quality in end milling of 2014 aluminium alloy and shown that surface roughness
is mainly affected by the feed rate and tool nose radius. Alauddin et al. (1996) have
pointed out that feed rate is the most significant factor and with increase in feed,
surface roughness increases while with increase in cutting speed, surface rough-
ness decreases in end milling Inconel 718 using uncoated carbide inserts. Lou et al.
(1998) have used multiple regression models to develop a surface roughness model
to predict Ra in CNC end milling of 6061 aluminum and concluded that the feed
rate is the most significant factor. Yang and Chen (2001) found out the optimum
cutting parameters for milling of Al 6061 material using Taguchi design consid-
ering cutting speed, feed rate, depth of cut and tool diameter as the cutting
parameters. Lee et al. (2001) presented a method for the simulation of surface
roughness of the machined surface in high-speed end milling. Lin (2002) has
optimized cutting speed, feed rate and depth of cut with consideration of multiple
performance characteristics including removed volume, surface roughness and
burr height in face milling of stainless steel and shown that the most influence of
the cutting parameters is the feed rate. Mansour and Abdalla (2002) have con-
cluded that with increase in feed rate or in axial depth of cut, surface roughness
increases whilst with increase in cutting speed, surface roughness decreases in end
milling operations of EN32 materials. Ghani et al. (2004) have studied surface
roughness in end milling of hardened steel AISI H13 with TiN coated P10 carbide
insert tool and concluded that use of high cutting speed, low feed rate and low
depth of cut leads to better surface finish. Wang and Chang (2004) have analyzed
the influence of cutting condition and tool geometry on surface roughness in slot
end milling of AL2014-T6. Oktem et al. (2005) have developed an effective
methodology to determine the optimum cutting conditions leading to minimum
roughness in milling of Aluminum (7075-T6) molded surfaces considering feed,
cutting speed, axial depth of cut, radial depth of cut and machining tolerance as
cutting parameters. Reddy and Rao (2005) have developed a model to see the
effects of tool geometry, cutting speed and feed rate on surface roughness in end
milling of medium carbon steel. The investigations of this study indicate that the
parameters cutting speed, feed, radial rake angle and nose radius are the primary
factors influencing the surface roughness of medium carbon steel during end
milling. Reddy and Rao (2006a) have investigated the role of solid lubricant
assisted machining with graphite and molybdenum disulphide lubricants on sur-
face quality, cutting forces and specific energy while milling AISI 1045 steel using
cutting tools of different tool geometry (radial rake angle and nose radius). Reddy
and Rao (2006b) have studied the effect of various parameters such as cutting
speed, feed rate, radial rake angle and nose radius on surface roughness in milling
of AISI 1045 materials. They have shown that surface roughness decreases with
increasing cutting speed. Jesuthanam et al. (2007) have developed a hybrid neural
network trained with GA and Particle Swarm Optimization (PSO) for the pre-
diction of surface roughness in CNC end milling operation of mild steel materials.
For the development of network, spindle speed, feed, depth of cut and vibration
data are considered. Chang and Lu (2007) have applied a grey relational analysis
to determine the cutting parameters for optimizing the side milling process with
multiple performance characteristics and concluded that feeding-
direction roughness, axial-direction roughness and waviness are improved
simultaneously through the optimal combination of the cutting parameters
obtained from the proposed two-stage parameter design. El-Sonbaty et al. (2008)
have developed artificial neural network (ANN) models for the analysis and
prediction of the relationship between the cutting conditions and the corresponding
fractal parameters of machined surfaces in face milling operation using rotational
speed, feed, depth of cut, pre-tool flank wear and vibration level as input
parameters. Routara et al. (2009) have studied the influence of machining
parameters on conventional roughness parameters in CNC end milling of alu-
minium, steel and brass materials using response surface method. Berglund and
Rose’n (2009) have evaluated the connection between surface finish appearance
and measured surface roughness using scale sensitive fractal analysis in milling.
Öktem (2009) has developed an integrated study of surface roughness to model
and optimize the cutting parameters in end milling of AISI 1040 steel material
with TiAlN solid carbide tools under wet condition using ANN and GA. He has
shown that the axial depth of cut is the most important cutting parameters affecting
surface roughness (Ra). Zain et al. (2010a) have carried out a study using GA to
observe the optimal effect of the radial rake angle of the tool, combined with speed
and feed rate in influencing the surface roughness result. With the highest speed,
lowest feed rate and highest radial rake angle of the cutting conditions scale, the
GA technique recommends the best minimum surface roughness value. For end
milling also, to modeling surface roughness different tools are used like RSM
(Alauddin et al. 1996; Mansour and Abdalla 2002; Wang and Chang 2004; Oktem
et al. 2005; Reddy and Rao 2005; Reddy and Rao 2006b; Routara et al. 2009),
Taguchi analysis (Yang and Chen 2001; Lin 2002; Ghani et al. 2004; Bagci and
Aykut 2006), ANN (Tsai et al. 1999; Balic and Korosec 2002; Benardos and
Vosniakos 2002; El-Sonbaty et al. 2008; Öktem 2009; Zain et al. 2010b). From the
literature survey, it is seen that most of literatures deal with conventional rough-
ness parameters to describe surface roughness and also in the study, three
machining parameters viz. spindle speed, feed rate and depth of cut are the most
common machining parameters (Fuh and Wu 1995; Lou et al. 1998; Tsai et al.
1999; Yang and Chen 2001; Lin 2002; Ghani et al. 2004; Wang and Chang 2004;
Bagci and Aykut 2006; Zhang and Chen 2007; Routara et al. 2009).
Electrical discharge machining (EDM) is a non-conventional machining
process that can be used all types of conductive materials. It can also be used for
machining of difficult-to-machine shapes and materials. In this section, few of the
available literatures on surface roughness modeling in EDM are reviewed. Zhang
et al. (1997) have investigated the effects on material removal rate, surface
roughness and diameter of discharge points in electro-discharge machining (EDM)
on ceramics and shown that the material removal rate, surface roughness and the
diameter of discharge point all increase with increasing pulse-on time and
discharge current. Lee and Li (2001) have shown that the negative tool polarity
gives better surface finish in EDM of tungsten carbide. Also, surface roughness
increases with increasing peak current and pulse duration. Ramasawmy and Blunt
(2002) have illustrated the influencing process factors in modifying the surface
textures using Taguchi method in EDM on M300 tool steel and shown that the
direct current is the most dominant factor in modifying the surface texture. Lin and
Lin (2002) have studied an approach for the optimization of the electrical dis-
charge machining process (work-piece polarity, pulse on time, duty factor, open
discharge voltage, discharge current, and dielectric fluid) with multiple perfor-
mance characteristics viz. MRR, surface roughness and electrode wear ratio using
grey relational analysis. Lin and Lin (2005) have tried to optimize the electrical
discharge machining process using grey-fuzzy logic considering pulse on time,
duty factor and discharge current as process parameters. Puertas and Luis (2003)
have modeled centre line average value (Ra) and root mean square roughness value
(Rq) in terms of current, pulse on time and off time in EDM on soft steel (F-1110).
It has been seen that the current intensity has the most influence on surface
roughness and there is a strong interaction between the current intensity and the
pulse on time factors being advisable to work with high current intensity values
and low pulse on time values. They have justified the fact of having to employ high
current intensity values to obtain a better surface roughness because a better arc
stability causes a more uniform production of sparks and a narrow variation
interval of the Ra and Rq roughness parameters. Yih-fong and Fu-chen (2003) have
presented an approach for optimizing high-speed EDM using Taguchi methods.
They have concluded that the most important factors affecting the EDM process
robustness have been identified as pulse-on time, duty cycle, and pulse peak
current. Ramasawmy and Blunt (2004) have quantified the effect of process
parameters on the surface texture using Taguchi method in EDM of steel and
concluded that the pulse current is the most dominant factor in affecting the
surface texture. Puertas et al. (2004) have carried out a study on the influence of
the factors of intensity, pulse time and duty cycle over surface roughness, material
removal rate, etc. in EDM of a cemented carbide and observed that in the case of
Ra parameter the most influential factors are intensity, followed by the pulse time
factor. Petropoulos et al. (2004) have emphasized the interrelationship between
surface texture parameters and process parameters in EDM of Ck60 steel plates.
They have considered amplitude, spacing, hybrid, as well as random process and
fractal parameters. Puertas et al. (2005) have carried out a study on the influence of
EDM parameters over two spacing parameters in machining of siliconised or
reaction-bonded silicon carbide (SiSiC) and shown that intensity, pulse time and
duty cycle are most influential factors affecting mean spacing between peaks and
the number of peaks per cm whereas the dielectric flushing pressure is not an
influential factor. Amorima and Weingaertner (2005) have shown that the increase
of average surface roughness of the work-piece is directly related to the increase in
discharge current and discharge duration on the EDM of the AISI P20 tool steel
under finish machining. Ramakrishnan and Karunamoorthy (2006) have proposed
a multi objective optimization method in WEDM process using parametric design
of Taguchi method and identified that the pulse on time and ignition current
intensity are the influential parameters. Keskin et al. (2006) have shown that
surface roughness has an increasing trend with an increase in the discharge
duration in EDM on steel work-pieces. Sahoo et al. (2009) have investigated the
influence of machining parameters, viz., pulse current, pulse on time and pulse off
time on the quality of surface produced in EDM of mild steel, brass and tungsten
carbide materials using response surface methodology. It is seen that the pulse
current has the maximum influence on the roughness parameters while pulse on
time has some effect and pulse off time has no significant effect on roughness
parameters. Shah et al. (2010) have shown that the material thickness has little
effect on the material removal rate and kerf but is a significant factor in terms of
surface roughness in wire electrical discharge machining (WEDM) of tungsten
carbide samples. Now-a-days, artificial neural network is used as a tool in mod-
eling of EDM process (Spedding and Wang 1997; Tsai and Wang 2001; Sarkar
et al. 2006; Mandal et al. 2007; Assarzadeh and Ghoreishi 2008).
From the literature survey, it is revealed that there are many researches on
surface roughness modeling in different machining processes. However, most of
the literatures deal with conventional roughness parameters and there is scarcity of
literatures which deal with fractal dimension modelling in machining.
1.5 Design of Experiments
The design of experiments technique (DOE) is a very powerful tool, which permits
to carry out the modeling and analysis of the influence of process variables on the
response variables. The response variable is an unknown function of the process
variables, which are known as design factors. The purpose of running experiments
is to characterize unknown relations and dependencies that exist in the observed
design or process, i.e., to find out the influential design variables and the response
1.5 Design of Experiments 19
to variations in the design variable values. A scientific approach to planning the

experiment must be employed if an experiment is to be performed most efficiently.
The statistical design of experiments refers to the process of planning the exper-
iment so that appropriate data that can be analyzed by statistical methods will be
collected, resulting in valid and objective conclusions in a meaningful way. When
the problem involves data that are subject to experimental errors, statistical
methodology is the only objective approach to analysis. Sometimes, experiments
are repeated with a particular set of levels for all the factors to check the statistical
validation and repeatability by the replicate data. This is called replication. To get
rid of any biasness, allocation of experimental material and the order of experi-
mental runs are randomly selected. This is called randomization. To arrange the
experimental material into groups, or blocks, that should be more homogeneous
than the entire set of material is called blocking. So, when experiments are carried
out these things should be remembered. There are several methodologies for
design of experiments. Some of DOE methods are discussed below.
1.5.1 Full Factorial Design
Full factorial design creates experimental points using all the possible combina-
tions of the levels of the factors in each complete trial or replication of the
experiments. The experimental design points in a full factorial design are the
vertices of a hyper cube in the n-dimensional design space defined by the mini-
mum and the maximum values of each of the factors. These experimental points
are also called factorial points. For three factors having four levels of each factors,
considering full factorial design, total 43 (64) numbers of experiments have to be
carried out. If there are n replicates of complete experiments, then there will be n
times of the single replication experiments to be conducted. In the experimenta-
tion, it must have at least two replicates to determine a sum of squares due to error
if all possible interactions are included in the model.
1.5.2 Central Composite Design
A Box–Wilson Central Composite Design, commonly called ‘‘Central Composite

Design (CCD)’’ is frequently used for building a second order polynomial for the
response variables in response surface methodology without using a complete full
factorial design of experiments. To establish the coefficients of a polynomial with
quadratic terms, the experimental design must have at least three levels of each
factor. In CCD, there are three different point viz. factorial points, central points
and axial points. Factorial points are vertices of the n-dimensional cube which are
coming from the full or fractional factorial design where the factor levels are
coded to -1, +1. Central point is the point at the center of the design space. Axial
Fig. 1.5 Face centered

central composite design with
three factors
points are located on the axes of the coordinate system symmetrically with respect
to the central point at a distance a from the design center.
There are two main varieties of CCD namely Face centered CCD and Rotatable
CCD. In face centered CCD, a k factor 3-level experimental design requires
2k ? 2k ? C experiments, where k is the number of factors, 2k points are in the
corners of the cube representing the experimental domain, 2k axial points are in the
center of each face of the cube ½ða; 0; . . .0Þ; ð0; a; . . .0Þ; ð0; 0; . . . aÞ and C
points are the replicates in the center of the cube that are necessary to estimate the
variability of the experimental measurements, it is to say the repeatability of the
phenomenon which carry out the lack-of-fit or curvature test for the model. The
centre points may vary from three to six. The example of 3-level three factor FCC
design is shown in Fig. 1.5. In this figure, the deep black circles represent the
fractional points at the corner of cube while the white circles represent axial points
in the center of each face of the cube and the star mark represents the centre points.
For the three factor experiment, eight (23) factorial points, six axial points (2 9 3)
and six centre runs, a total of 20 experimental runs can be considered. The value of
a is chosen here as 1. The upper and lower limits of a factor are coded as +1 and
-1 respectively using the following relations Eq. 1.8. Generally, the experimental
runs are conducted in random order.
½2x ðxmax þ xmin Þ
xi ¼ ð1:8Þ
ðxmax xmin Þ
The rotatable central composite design is the most widely used experimental
design for modeling a second-order response surface. A design is called rotatable
when the variance of the predicted response at any point depends only on the distance
of the point from the center point of design. The rotatable design provides the
uniformity of prediction error and it is achieved by proper choice of a: In rotatable
designs, all points at the same radial distance (r) from the centre point have the same
magnitude of prediction error. For a given number of variables, the a required to
achieve rotatability is computed as a ¼ ðnf Þ1=4 ; where nf is the number of points in
the 2k factorial design. A rotatable CCD consists of 2k fractional factorial points,
augmented by 2 k axial points ½ða; 0; . . .0Þ; ð0; a; . . .0Þ; ð0; 0; . . . aÞ and nc
1.5 Design of Experiments 21
centre points (0, 0, 0, 0…,0). Here also, the centre points vary from three to six. With
proper choice of nc the CCD can be made orthogonal or it can be made uniform
precision design. It means that the variance of response at origin is equal to the
variance of response at a unit distance from the origin. Considering uniform preci-
sion, for three factor experimentation, eight (23) factorial points, six axial points
(2 9 3) and six centre runs, a total of 20 experimental runs may be considered and the
value of a is ð8Þ1=4 ¼ 1:682.
1.6 Response Surface Methodology
Response Surface Method (RSM) adopts both mathematical and statistical tech-
niques which are useful for the modeling and analysis of problems in which a
response of interest is influenced by several variables and the objective is to
optimize the response (Montgomery 2001). RSM helps in analyzing the influence
of the independent variables on a specific dependent variable (response) by
quantifying the relationships amongst one or more measured responses and the
vital input factors. The mathematical models thus developed relating the
machining responses and their factors facilitate the optimization of the machining
process. In most of the RSM problems, the form of the relationship between the
response and the independent variables is unknown. Thus the first step in RSM is
to find a suitable approximation for the true functional relationship between
response of interest ‘y’ and a set of controllable variables {x1, x2, …, xn}. Usually
when the response function is not known or non-linear, a second order model is
utilized (Montgomery 2001) in the form:
X
n X
n XX
y ¼ b0 þ bi x i þ bii x2i þ bij xi xj þ e ð1:9Þ
i¼1 i¼1 i\j
where, e represents the noise or error observed in the response y such that the
expected response is (y -eÞ and b’s are the regression coefficients to be estimated.
The least square technique is being used to fit a model equation containing the
input variables by minimizing the residual error measured by the sum of square
deviations between the actual and estimated responses. The calculated coefficients
or the model equations however need to be tested for statistical significance and
thus the following tests are performed.
To check the adequacy of the model for the responses in the experimentation,
Analysis of Variance (ANOVA) is used. ANOVA calculates the F-ratio, which is
the ratio between the regression mean square and the mean square error. The
F-ratio, also called the variance ratio, is the ratio of variance due to the effect of a
factor (the model) and variance due to the error term. This ratio is used to measure
the significance of the model under investigation with respect to the variance of all
the terms included in the error term at the desired significance level, a: If the
calculated value of F-ratio is higher than the tabulated value of F-ratio for
roughness, then the model is adequate at desired a level to represent the rela-
tionship between machining response and the machining parameters.
In the ANOVA Table, there is a P-value or probability of significance for each
independent variable in the model the value of which shows whether the variable
is significant or not. If the P-value is less or equal to the selected a-level, then the
effect of the variable is significant. If the P-value is greater than the selected
a-value, then it is considered that the variable is not significant. Sometimes the
individual variables may not be significant. If the effect of interaction terms is
significant, then the effect of each factor is different at different levels of the other
factors. ANOVA for different response variables are carried out in the present
study using commercial software Minitab (Minitab user manual 2001) with
confidence level set at 95%, i.e., the a-level is set at 0.05.
1.7 Closure
In this chapter, different basic considerations are discussed. The chapter starts with
the essence of fractal dimension to describe surface roughness. The basics of
surface metrology including the different roughness parameters along with the
surface roughness measurement technique are presented. Basics of fractal
dimension and its calculation are also discussed. Then the essence of design of
experiments and different design of experiment techniques are presented.
Response surface methodology (RSM) is discussed which is used to analyze the
experimental data in the subsequent chapters.
References
Abburi NR, Dixit US (2006) A knowledge-based system for the prediction of surface roughness
in turning process. Robotics Comput-Integr Manuf 22:363–372
Abouelatta OB, Madl J (2001) Surface roughness prediction based on cutting parameters and tool
vibrations in turning operations. J Mater Process Technol 118:269–277
Alauddin M, El Baradie MA, Hashmi MSJ (1996) Optimization of surface finish in end milling
Inconel 718. J Mater Process Technol 56:54–65
Amorima FL, Weingaertner WL (2005) The influence of generator actuation mode and process
parameters on the performance of finish EDM of a tool steel. J Mater Process Technol
166:411–416
Arbizu IP, Pérez CJL (2003) Surface roughness prediction by factorial design of experiments in
turning processes. J Mater Process Technol 143–144:390–396
Assarzadeh S, Ghoreishi M (2008) Neural-network-based modeling and optimization of the
electro-discharge machining process. Int J Adv Manuf Technol 39:488–500
Bagci E, Aykut S (2006) A study of Taguchi optimization method for identifying optimum
surface roughness in CNC face milling of cobalt-based alloy (stellite 6). Int J Adv Manuf
Technol 29:940–947
Bagci E, Isik B (2006) Investigation of surface roughness in turning unidirectional GFRP
composites by using RS methodology and ANN. Int J Adv Manuf Technol 31:10–17
References 23
Balic J, Korosec M (2002) Intelligent tool path generation for milling of free surfaces using
neural networks. Int J Mach Tools Manuf 42:1171–1179
Benardos PG, Vosniakos GC (2002) Prediction of surface roughness in CNC face milling using
neural networks and Taguchi’s design of experiments. Robotics Comput-Integr Manuf
18:343–354
Benardos PG, Vosniakos GC (2003) Predicting surface roughness in machining: a review. Int J
Mach Tools Manuf 43(8):833–844
Berglund J, Rose’n BG (2009) A method development for correlation of surface finish
appearance of die surfaces and roughness measurement data. Tribol Lett 36(2):157–164
Berry MV, Lewis ZV (1980) On the Weierstrass-Mandelbrot fractal function. Proc R Soc A
370:459–484
Bhushan B, Wyant JC, Meiling J (1988) A new three-dimensional non-contact digital optical
profiler. Wear 122:301–312
Bigerelle M, Najjar D, Iost A (2005) Multiscale functional analysis of wear a fractal model of the
grinding process. Wear 258:232–239
Brown CA, Savary G (1991) Describing ground surface texture using contact profilometry and
fractal analysis. Wear 141:211–226
Chang CK, Lu HS (2007) Design optimization of cutting parameters for side milling operations
with multiple performance characteristics. Int J Adv Manuf Technol 32:18–26
Chavoshi SZ, Tajdari M (2010) Surface roughness modelling in hard turning operation of AISI
4140 using CBN cutting tool. Int J Mater Form. doi:10.1007/s12289-009-0679-2
Choi TJ, Subrahmanya N, Li H, Shin YC (2008) Generalized practical models of cylindrical
plunge grinding processes. Int J Mach Tools Manuf 48:61–72
Dabnun MA, Hashmi MSJ, El-Baradie MA (2005) Surface roughness prediction model by design
of experiments for turning machinable glass-ceramic (Macor). J Mater Process Technol
164–165:1289–1293
Davim JP (2001) A note on the determination of optimal cutting conditions for surface finish
obtained in turning using design of experiments. J Mater Process Technol 116:305–308
El-Sonbaty IA, Khashaba UA, Selmy AI, Ali AI (2008) Prediction of surface roughness profiles
for milledsurfaces using an artificial neural network and fractal geometry approach. J Mater
Process Technol 200:271–278
Feng CX, Wang XF (2003) Surface roughness predictive modeling: neural networks versus
regression. IIE Trans 35:11–27
Feng CXJ, Yu ZG, Kusiak A (2006) Selection and validation of predictive regression and neural
network models based on designed experiments. IIE Trans 38:13–23
Fredj NB, Amamou R (2006) Ground surface roughness prediction based upon experimental
design and neural network models. Int J Adv Manuf Technol 31:24–36
Fuh KH, Wu CF (1995) A proposed statistical model for surface quality prediction in end-milling
of Al alloy. Int J Mach Tools Manuf 35(S):1187–1200
Ge S, Chen G (1999) Fractal prediction models of sliding wear during the running–in process.
Wear 231:249–255
Ghani JA, Choudhury IA, Hassan HH (2004) Application of Taguchi method in the optimization
of end milling parameters. J Mater Process Technol 145:84–92
Grzesik W (1996) A revised model for predicting surface roughness in turning. Wear
194:143–148
Gupta AK (2010) Predictive modelling of turning operations using response surface
methodology, artificial neural networks and support vector regression. Int J Prod Res
48(3):763–778
Han JH, Ping S, Shengsun H (2005) Fractal characterization and simulation of surface profiles of
copper electrodes and aluminum sheets. Mater Sci Eng A 403:174–181
Hasegawa M, Liu J, Okuda K, Nunobiki M (1996) Calculation of the fractal dimensions of
machined surface profiles. Wear 192:40–45
Hassui A, Diniz AE (2003) Correlating surface roughness and vibration on plunge cylindrical
grinding of steel. Int J Mach Tools Manuf 43:855–862
He L, Zhu J (1997) The fractal character of processed metal surfaces. Wear 208:17–24
Hecker RL, Liang SY (2003) Predictive model of surface roughness in grinding. Int J Mach Tools
Manuf 43:755–761
ISO 4287:1997 (1997) Geometrical product specification (GPS)—surface texture: profile
method—terms, definitions and surface texture parameters. International Organization of
Standardization, Geneva
Jahn R, Truckenbrodt H (2004) A simple fractal analysis method of the surface roughness.
J Mater Process Technol 145:40–45
Jesuthanam CP, Kumanan S, Asokan P (2007) Surface roughness prediction using hybrid neural
networks. Mach Sci Technol 11:271–286
Jiang Z, Wang H, Fei B (2001) Research into the application of fractal geometry in characterizing
machined surfaces. Int J Mach Tools Manuf 41:2179–2185
Jiao Y, Lei S, Pei ZJ, Lee ES (2004) Fuzzy adaptive networks in machining process modeling:
surface roughness prediction for turning operations. Int J Mach Tools Manuf 44:1643–1651
Kang MC, Kim JS, Kim KH (2005) Fractal dimension analysis of machined surface depending on
coated tool wear. Surf Coat Technol 193(1–3):259–265
Karayel D (2009) Prediction and control of surface roughness in CNC lathe using artificial neural
network. J Mater Process Technol 209:3125–3137
Keskin YH, Halkacı HS, Kizil SM (2006) An experimental study for determination of the effects
of machining parameters on surface roughness in electrical discharge machining (EDM). Int J
Adv Manuf Technol 28:1118–1121
Kirby ED, Zhang Z, Chen JC, Chen J (2006) Optimizing surface finish in a turning operation
using the Taguchi parameter design method. Int J Adv Manuf Technol 30:1021–1029
Kohli A, Dixit US (2005) A neural-network-based methodology for the prediction of surface
roughness in turning process. Int J Adv Manuf Technol 25:118–129
Krajnik P, Kopac J, Sluga A (2005) Design of grinding factors based on response surface
methodology. J Mater Process Technol 162–163:629–636
Kwak JS (2005) Application of Taguchi and response surface methodologies for geometric error
in surface grinding process. Int J Mach Tools Manuf 45:327–334
Kwak JS, Sim SB, Jeong YD (2006) An analysis of grinding power and surface roughness in
external cylindrical grinding of hardened SCM440 steel using response surface method. Int J
Mach Tools Manuf 46:304–312
Lee SH, Li XP (2001) Study of the effect of machining parameters on the machining
characteristics in electrical discharge machining of tungsten carbide. J Mater Process Technol
115:344–358
Lee KY, Kang MC, Jeong YH, Lee DW, Kim JS (2001) Simulation of surface roughness and
profile in high-speed end milling. J Mater Process Technol 113:410–4125
Lin TR (2002) Optimisation technique for face milling stainless steel with multiple performance
characteristics. Int J Adv Manuf Technol 19:330–335
Lin JL, Lin CL (2002) The use of orthogonal array with grey relational analysis to optimize the
electrical discharge machining process with multiple performance characteristics. Int J Mach
Tools Manuf 42:237–244
Lin JL, Lin CL (2005) The use of grey-fuzzy logic for the optimization of the manufacturing
process. J Mater Process Technol 160:9–14
Lin WS, Lee BY, Wu CL (2001) Modeling the surface roughness and cutting force for turning.
Ling FF (1990) Fractals, engineering surfaces and tribology. Wear 136:141–156
Lou MS, Chen JC, Li CM (1998) Surface roughness prediction technique for CNC end-milling.
J Ind Technol 15 (1), November 1998 to January 1999
Majumdar A, Bhushan B (1990) Role of fractal geometry in roughness characterization and
contact mechanics of surfaces. Trans ASME J Tribol 112:205–216
Majumdar A, Tien CL (1990) Fractal characterization and simulation of rough surfaces. Wear
136:313–327
References 25
Maksoud TMA, Atia MR, Koura MM (2003) Applications of artificial intelligence to grinding
operations via neural networks. Mach Sci Technol 7(3):361–387
Mandal D, Pal SK, Saha P (2007) Modeling of electrical discharge machining process using back
propagation neural network and multi-objective optimization using non-dominating sorting
algorithm-II. J Mater Process Technol 186:154–162
Mandelbrot BB (1967) How long is the coast of Britain? Statistical self-similarity and fractional
dimension. Science 156:636–638
Mandelbrot BB (1982) The fractal geometry of nature. W H freeman, New York
Mansour A, Abdalla H (2002) Surface roughness model for end milling: a semi-free cutting
carbon casehardening steel (EN32) in dry condition. J Mater Process Technol 124:183–191
Minitab User Manual Release 13.2 (2001) Making data analysis easier. MINITAB Inc. State
College, PA
Mohanasundararaju N, Sivasubramanian R, Gnanaguru R, Alagumurthy N (2008) A neural
network and fuzzy-based methodology for the prediction of work roll surface roughness in a
grinding process. Int J Comput Methods Eng Sci Mech 9:103–110
Montgomery DC (2001) Design and analysis of experiments. Wiley, New York
Muthukrishnan N, Davim JP (2009) Optimization of machining parameters of Al/SiC-MMC with
ANOVA and ANN analysis. J Mater Process Technol 209:225–232
Nalbant M, Gokkaya H, Sur G (2007) Application of Taguchi method in the optimization of
cutting parameters for surface roughness in turning. Mater Des 28:1379–1385
Nayak PR (1971) Random process model of rough surfaces. Trans ASME J Lubr Technol
93:398–407
Öktem H (2009) An integrated study of surface roughness for modeling and optimization of
cutting parameters during end milling operation. Int J Adv Manuf Technol 43:852–861
Oktem H, Erzurumlu T, Kurtaran H (2005) Application of response surface methodology in the
optimization of cutting conditions for surface roughness. J Mater Process Technol 170:11–16
Pal SK, Chakraborty D (2005) Surface roughness prediction in turning using artificial neural
network. Neural Comput Appl 14:319–324
Palanikumar K (2008) Application of Taguchi and response surface methodologies for surface
roughness in machining glass fiber reinforced plastics by PCD tooling. Int J Adv Manuf
Technol 36:19–27
Palanikumar K, Karunamoorthy L, Karthikeyan R (2006) Parametric optimization to minimise
the surface roughness on the machining of GFRP composites. J Mater Sci Technol
22(1):66–72
Petropoulos G, Vaxevanidis NM, Pandazaras C (2004) Modeling of surface finish in electro-
discharge machining based upon statistical multi-parameter analysis. J Mater Process Technol
155–156:1247–1251
Puertas I, Luis CJ (2003) A study on the machining parameters optimisation of electrical
discharge machining. J Mater Process Technol 143–144:521–526
Puertas I, Luis CJ, Álvarez L (2004) Analysis of the influence of EDM parameters on surface
quality, MRR and EW of WC-Co. J Mater Process Technol 153–154:1026–1032
Puertas I, Luis CJ, Villa G (2005) Spacing roughness parameters study on the EDM of silicon
carbide. J Mater Process Technol 164–165:1590–1596
Ramakrishnan R, Karunamoorthy L (2006) Multi response optimization of wire EDM operations
using robust design of experiments. Int J Adv Manuf Technol 29:105–112
Ramasawmy H, Blunt L (2002) 3D surface characterisation of elctropolished EDMed surface and
quantitative assessment of process variables using Taguchi Methodology. Int J Mach Tools
Manuf 42:1129–1133
Ramasawmy H, Blunt L (2004) Effect of EDM process parameters on 3D surface topography.
Ramesh S, Karunamoorthy L, Palanikumar K (2008) Surface roughness analysis in machining of
titanium alloy. Mater Manuf Process 23:174–181
Reddy NSK, Rao PV (2005) Selection of optimum tool geometry and cutting conditions using a
surface roughness prediction model for end milling. Int J Adv Manuf Technol 26:1202–1210
Reddy NSK, Rao PV (2006a) Experimental investigation to study the effect of solid lubricants on
cutting forces and surface quality in end milling. Int J Mach Tools Manuf 46:189–198
Reddy NSK, Rao PV (2006b) Selection of an optimal parametric combination for achieving a
better surface finish in dry milling using genetic algorithms. Int J Adv Manuf Technol
28:463–473
Routara BC, Bandyopadhyay A, Sahoo P (2009) Roughness modeling and optimization in CNC
end milling using response surface method: effect of workpiece material variation. Int J Adv
Manuf Technol 40:1166–1180
Sahin Y, Motorcu AR (2005) Surface roughness model for machining mild steel with coated
carbide tool. Mater Des 26:321–326
Sahoo P (2005) Engineering tribology. Prentice Hall of India, New Delhi
Sahoo P, Ghosh N (2007) Finite element contact analysis of fractal surfaces. J Phys D Appl Phys
40:4245–4252
Sahoo P, Routara BC, Bandyopadhyay A (2009) Roughness modeling and optimization in EDM
using response surface method for different workpiece materials. Int J Mach Mach Mater
5(2–3):321–346
Sarkar S, Mitra S, Bhattacharyya B (2006) Parametric optimisation of wire electrical discharge
machining of c titanium aluminide alloy through an artificial neural network model. Int J Adv
Sayles RS, Thomas TR (1978) Surface topography as a non-stationary random process. Nature
271:431–434
Shah A, Mufti NA, Rakwal D, Bamberg E (2010) Material removal rate, kerf, and surface
roughness of tungsten carbide machined with wire electrical discharge machining. J Mater
Eng Perform. doi:10.1007/s11665-010-9644-y
Siddiquee AN, Khan ZA, Mallick Z (2010) Grey relational analysis coupled with principal
component analysis for optimisation design of the process parameters in in-feed centreless
cylindrical grinding. Int J Adv Manuf Technol 46:983–992
Singh D, Rao PV (2007) A surface roughness prediction model for hard turning process. Int J Adv
Spedding TA, Wang ZQ (1997) Parametric optimization and surface characterization of wire
electrical discharge machining process. Precis Eng 20:5–15
Suresh PVS, Rao PV, Deshmukh SG (2002) A genetic algorithm approach for optimization of
surface roughness prediction model. Int J Mach Tools Manuf 42:675–680
Thomas TR (1982) Defining the microtopography of surfaces in thermal contact. Wear 79:73–82
Tricot C, Ferlans P, Baran G (1994) Fractal analysis of worn surfaces. Wear 172:127–133
Tsai KM, Wang PJ (2001) Predictions on surface finish in electrical discharge machining based
upon neural network models. Int J Mach Tools Manuf 41:1385–1403
Tsai YH, Chen JC, Lou SJ (1999) An in-process surface recognition system based on neural
networks in end milling cutting operations. Int J Mach Tools Manuf 39:583–605
Venkatesh K, Bobji MS, Biswas SK (1998) Some features of surface topographical power spectra
generated by conventional machining of a ductile metal. Mater Sci Eng A A252:153–155
Venkatesh K, Bobji MS, Gargi R, Biswas SK (1999) Genesis of workpiece roughness generated
in surface grinding and polishing of metals. Wear 225–229:215–226
Wang MY, Chang HY (2004) Experimental study of surface roughness in slot end milling
AL2014–T6. Int J Mach Tools Manuf 44:51–57
Whitehouse DJ (1982) The parameter rash, is there a cure? Wear 83:75–78
Yan W, Komvopoulos K (1998) Contact analysis of elastic-plastic fractal surfaces. J Appl Phys
84(7):3617–3624
Yang JL, Chen JC (2001) A systematic approach for identifying optimum surface roughness
performance in end-milling operations. J Ind Technol 17, 2 February 2001 to April 2001
Yang WH, Tarng YS (1998) Design optimization of cutting parameters for turning operations
based on the Taguchi method. J Mater Process Technol 84:122–129
Yih-fong T, Fu-chen C (2003) A simple approach for robust design of high-speed electrical
discharge machining technology. Int J Mach Tools Manuf 43:217–227
References 27
Zain AM, Haron H, Sharif S (2010a) Application of GA to optimize cutting conditions for
minimizing surface roughness in end milling machining process. Expert Syst Appl
37:4650–4659
Zain AM, Haron H, Sharif S (2010b) Prediction of surface roughness in the end milling
machining using artificial neural network. Expert Syst Appl 37:1755–1768
Zhang JZ, Chen JC (2007) The development of an in-process surface roughness adaptive control
system in end milling operations. Int J Adv Manuf Technol 31:877–887
Zhang JH, Lee TC, Lau WS (1997) Study on the electro-discharge machining of a hot pressed
aluminum oxide based ceramic. J Mater Process Technol 63:908–912
Zhang Y, Luo Y, Wang JF, Li Z (2001) Research on the fractal of surface topography of grinding.
Int J Mach Tools Manuf 41:2045–2049
Zhong ZW, Khoo LP, Han ST (2006) Prediction of surface roughness of turned surfaces using
neural networks. Int J Adv Manuf Technol 28:688–693
Zhong ZW, Khoo LP, Han ST (2008) Neural-network predicting of surface finish or cutting
parameters for carbide and diamond turning processes. Mater Manuf Process 23:92–97
Zhou X, Xi F (2002) Modeling and predicting surface roughness of the grinding process. Int J
Mach Tools Manuf 42:969–977
Zhu H, Ge S, Huang X, Zhang D, Liu J (2003) Experimental study on the characterization of
worn surface topography with characteristic roughness parameter. Wear 255:309–314
Chapter 2
Fractal Analysis in CNC End Milling
Abstract This chapter deals with the fractal dimension modeling in CNC end
milling operation. Milling operations are carried out for three different materials
viz. mild steel, brass and aluminium work-pieces for different combinations of
spindle speed, feed rate and depth of cut. The generated surfaces are measured
with Talysurf instrument and analyzed to get fractal dimension. The experimental
results are further processed to model fractal dimension using response surface
methodology (RSM). It is seen that spindle speed and depth of cut are the sig-
nificant factors affecting fractal dimension for mild steel. For brass material, the
significant factors are spindle speed and feed rate but for aluminium the significant
factor is depth of cut. In general, for mild steel and brass, with increase in spindle
speed, D increases. Comparing the developed response surface models, it is
concluded that the models are material specific and the tool-work-piece material
combination plays a vital role in fractal dimension of the generated surface profile.
Keywords Fractal dimension (D) CNC End Milling RSM Mild steel Brass
Aluminium
2.1 Introduction
CNC milling is a popular machining process in the modern industry because of its
ability to remove materials with a multi-point cutting tool at a faster rate with a
reasonably good surface quality. In order to get specified surface roughness,
selection of controlling parameters is necessary. There has been a great many
research developments in modeling surface roughness and optimization of the
controlling parameters to obtain a surface finish of desired level since only proper
selection of cutting parameters can produce a better surface finish. But such studies

DOI: 10.1007/978-3-642-17922-8_2, Ó Prasanta Sahoo 2011
30 2 Fractal Analysis in CNC End Milling
Table 2.1 Variable levels used in the experimentation

Levels Aluminium Brass Mild steel
d N f d N f d N f
-1 0.10 4,500 900 0.10 1,500 550 0.150 2,500 300
-0.5 0.15 4,750 950 0.15 1,800 600 0.175 2,750 350
0 0.20 5,000 1,000 0.20 2,100 650 0.200 3,000 400
0.5 0.25 5,250 1,050 0.25 2,400 700 0.225 3,250 450
1 0.30 5,500 1,100 0.30 2,700 750 0.250 3,500 500
are far from complete since it is very difficult to consider all the parameters that
control the surface roughness for a particular manufacturing process. In CNC
milling there are several parameters which control the surface quality. The analysis
of surface roughness on CNC end milling process is a big challenge for research
development. Several factors involved in machining process have to be optimized
to obtain a desired surface quality. In this study, three machining parameters are
considered viz. spindle speed, feed rate and depth of cut. Also the study is con-
ducted on three different materials, viz. mild steel, brass and aluminium to con-
sider the effect of work-piece material variation on fractal dimension of machined
surfaces. The experimental results are analyzed using RSM.
2.2 Experimental Details
2.2.1 Design of Experiments
A full factorial design is used with five levels of each of the three design factors
viz. depth of cut (d, mm), spindle speed (N, rpm) and feed rate (f, mm/min). Thus
the design chosen was five level-three factor (53) full factorial design consisting of
125 sets of coded combinations for each work-piece material. Three cutting
parameters are selected as design factors while other parameters have been
assumed to be constant over the experimental domain. The upper and lower limits
of a factor were coded as +1 and -1 respectively using Eq. 1.8. The process
variables/design factors with their values on different levels are listed in Table 2.1
for three different work-piece materials.
2.2.2 Machine Used
The machine used for the milling tests is a ‘DYNA V4.5’ CNC end milling
machine having the control system SINUMERIK 802 D with a vertical milling
head. The specification of CNC end milling machine has been shown in Table 2.2.
For generating the milled surfaces, CNC part programs for tool paths were created
with specific commands. The compressed coolant servo-cut was used as cutting
environment.
2.2 Experimental Details 31
Table 2.2 Specification of CNC end milling machine

Table size 450 9 250 mm
Table load capacity 200 Kgs
X Travel 250 mm
Y Travel 175 mm
Z Travel 175 mm
Spindle nose to table 300 mm
Spindle centre to column 280 mm
Taper of spindle nose BT 30
Spindle speed 9,000 rpm
Rapid on X and Y axis 15 m/min
Rapid on Z axis 10 m/min
Spindle motor 3.7 kW
X axis motor 3 Nm
Y axis motor 3 Nm
Z axis motor 6 Nm
Contro system 802 D SINUMERIK
Power requirement 7.5 kW/10 H.P.
Lubricating oil Tellus 33 or EN KLO 68
2.2.3 Cutting Tool Used
Coated carbide tools are known to perform better than uncoated carbide tools.
Thus commercially available CVD coated carbide tools were used in this inves-
tigation. The tools used were flat end mill cutters produced by WIDIA
(EM-TiAlN). The tools were coated with TiAlN coating. For each material a new
cutter of same specification was used. The details of the end milling cutters are
given below:
Cutter diameter = 8 mm
Overall length = 108 mm
Fluted length = 38 mm
Helix angle = 30°
Hardness = 1,570 HV
Density = 14.5 g/cc
Transverse rupture strength = 3,800 N/mm2
2.2.4 Work-Piece Materials
The present study was carried out with three different materials, viz., 6061-T4
Aluminium, AISI 1040 steel and Medium leaded Brass UNS C34000. The
chemical composition and mechanical properties of the work-piece materials are
shown in Table 2.3. All the specimens were in the form of 100 9 75 9 25 mm
blocks.
Table 2.3 Composition and mechanical properties of work-piece materials

Work material Chemical composition (W%t) Mechanical
property
Aluminium 0.2%Cr, 0.3%Cu, 0.85%Mg, 0.04%Mn, 0.5%Si, 0.04%Ti, Hardness—65
(6061-T4) 0.25%Zn, 0.5%Fe and balance Al BHN,
Density—2.7
g/cc,
Tensile
Strength—
241 MPa
Brass 0.095%Fe, 0.9%Pb, 34%Zn and balance Cu Hardness—68
(UNS HRF,
C34000) Density—
8.47 g/cc,
Tensile
strength—
340 MPa
Mild Steel 0.42%C, 0.48%Mn, 0.17%Si, 0.02%P, 0.018%S, 0.1%Cu, Hardness—201
(AISI 0.09%Ni, 0.07%Cr and balance Fe BHN,
1040) Density—
7.85 g/cc,
Tensile
strength—
620 MPa
2.3 Results and Discussion
CNC milling operations are carried out on mild steel, brass and aluminium work-
pieces to get machined surfaces for different combinations of spindle speed, feed rate
and depth of cut. The generated surfaces are measured using Talysurf instrument and
further processed to get fractal dimension (D). Full factorial design of experiments is
considered in the study and the experimental results are presented in Table 2.4.
The influences of the cutting parameters (d, N and f) on the profile fractal
dimension D have been assessed for three different materials. The second order
model was postulated in obtaining the relationship between the fractal dimension
and the machining variables using response surface methodology (RSM). The
analysis of variance (ANOVA) was used to check the adequacy of the second
order model. The results for the three different materials are presented one by one.
2.3.1 RSM for Mild Steel
The second order response surface equation for the fractal dimension in mild steel
milling is obtained in terms of coded values of design factors as:
D ¼1:3836 þ 0:0136d þ 0:0115N þ 0:0069f 0:0063dN þ 0:0003df
0:0106Nf 0:0283d2 þ 0:0169N 2 þ 0:0032f 2 ð2:1Þ
2.3 Results and Discussion 33
Table 2.4 Experimental results for CNC milling considering full factorial design
Sl Depth of Spindle Feed D for mild D for D for
No cut(d) speed(N) rate(f) steel brass aluminium
1 -1 -1 -1 1.31 1.28 1.34
2 -1 -1 -0.5 1.33 1.31 1.34
3 -1 -1 0 1.29 1.22 1.37
4 -1 -1 0.5 1.30 1.28 1.29
5 -1 -1 1 1.32 1.27 1.36
6 -1 -0.5 -1 1.29 1.30 1.38
7 -1 -0.5 -0.5 1.33 1.29 1.32
8 -1 -0.5 0 1.37 1.30 1.35
9 -1 -0.5 0.5 1.37 1.32 1.35
10 -1 -0.5 1 1.34 1.27 1.36
11 -1 0 -1 1.32 1.38 1.35
12 -1 0 -0.5 1.35 1.33 1.34
13 -1 0 0 1.38 1.31 1.33
14 -1 0 0.5 1.34 1.30 1.34
15 -1 0 1 1.39 1.31 1.34
16 -1 0.5 -1 1.38 1.36 1.35
17 -1 0.5 -0.5 1.36 1.33 1.36
18 -1 0.5 0 1.36 1.30 1.35
19 -1 0.5 0.5 1.40 1.31 1.34
20 -1 0.5 1 1.34 1.32 1.34
21 -1 1 -1 1.40 1.37 1.34
22 -1 1 -0.5 1.38 1.35 1.35
23 -1 1 0 1.41 1.34 1.35
24 -1 1 0.5 1.36 1.35 1.38
25 -1 1 1 1.37 1.30 1.38
26 -0.5 -1 -1 1.41 1.30 1.37
27 -0.5 -1 -0.5 1.39 1.27 1.36
28 -0.5 -1 0 1.35 1.26 1.35
29 -0.5 -1 0.5 1.39 1.25 1.31
30 -0.5 -1 1 1.38 1.28 1.37
31 -0.5 -0.5 -1 1.31 1.31 1.36
32 -0.5 -0.5 -0.5 1.37 1.29 1.39
33 -0.5 -0.5 0 1.40 1.31 1.31
34 -0.5 -0.5 0.5 1.41 1.29 1.35
35 -0.5 -0.5 1 1.40 1.29 1.32
36 -0.5 0 -1 1.38 1.38 1.35
37 -0.5 0 -0.5 1.32 1.34 1.34
38 -0.5 0 0 1.37 1.31 1.34
39 -0.5 0 0.5 1.39 1.28 1.34
40 -0.5 0 1 1.39 1.29 1.38
41 -0.5 0.5 -1 1.41 1.35 1.33
42 -0.5 0.5 -0.5 1.40 1.33 1.31
43 -0.5 0.5 0 1.36 1.32 1.37
44 -0.5 0.5 0.5 1.41 1.32 1.36
(continued)
Table 2.4 (continued)

45 -0.5 0.5 1 1.36 1.29 1.35
46 -0.5 1 -1 1.38 1.37 1.36
47 -0.5 1 -0.5 1.38 1.37 1.35
48 -0.5 1 0 1.32 1.36 1.34
49 -0.5 1 0.5 1.39 1.32 1.34
50 -0.5 1 1 1.41 1.31 1.38
51 0 -1 -1 1.38 1.26 1.41
52 0 -1 -0.5 1.42 1.25 1.35
53 0 -1 0 1.43 1.26 1.37
54 0 -1 0.5 1.43 1.27 1.34
55 0 -1 1 1.41 1.29 1.35
56 0 -0.5 -1 1.38 1.36 1.36
57 0 -0.5 -0.5 1.41 1.27 1.36
58 0 -0.5 0 1.38 1.32 1.35
59 0 -0.5 0.5 1.38 1.27 1.39
60 0 -0.5 1 1.40 1.29 1.31
61 0 0 -1 1.38 1.35 1.34
62 0 0 -0.5 1.37 1.35 1.34
63 0 0 0 1.34 1.32 1.37
64 0 0 0.5 1.41 1.31 1.35
65 0 0 1 1.38 1.31 1.31
66 0 0.5 -1 1.40 1.36 1.28
67 0 0.5 -0.5 1.39 1.34 1.34
68 0 0.5 0 1.36 1.32 1.34
69 0 0.5 0.5 1.40 1.32 1.37
70 0 0.5 1 1.38 1.36 1.35
71 0 1 -1 1.43 1.37 1.36
72 0 1 -0.5 1.41 1.33 1.35
73 0 1 0 1.40 1.35 1.37
74 0 1 0.5 1.39 1.34 1.36
75 0 1 1 1.43 1.36 1.36
76 0.5 -1 -1 1.40 1.24 1.38
77 0.5 -1 -0.5 1.39 1.27 1.32
78 0.5 -1 0 1.38 1.23 1.29
79 0.5 -1 0.5 1.43 1.26 1.33
80 0.5 -1 1 1.38 1.28 1.32
81 0.5 -0.5 -1 1.39 1.27 1.38
82 0.5 -0.5 -0.5 1.35 1.27 1.38
83 0.5 -0.5 0 1.37 1.33 1.33
84 0.5 -0.5 0.5 1.40 1.25 1.33
85 0.5 -0.5 1 1.41 1.28 1.34
86 0.5 0 -1 1.35 1.38 1.33
87 0.5 0 -0.5 1.32 1.33 1.36
88 0.5 0 0 1.37 1.32 1.36
(continued)

89 0.5 0 0.5 1.39 1.29 1.31
90 0.5 0 1 1.41 1.31 1.28
91 0.5 0.5 -1 1.36 1.39 1.33
92 0.5 0.5 -0.5 1.38 1.33 1.36
93 0.5 0.5 0 1.37 1.32 1.33
94 0.5 0.5 0.5 1.39 1.31 1.37
95 0.5 0.5 1 1.38 1.36 1.34
96 0.5 1 -1 1.44 1.36 1.34
97 0.5 1 -0.5 1.43 1.37 1.34
98 0.5 1 0 1.44 1.37 1.3
99 0.5 1 0.5 1.43 1.34 1.3
100 0.5 1 1 1.42 1.34 1.36
101 1 -1 -1 1.30 1.29 1.34
102 1 -1 -0.5 1.42 1.28 1.32
103 1 -1 0 1.38 1.26 1.32
104 1 -1 0.5 1.38 1.24 1.29
105 1 -1 1 1.39 1.26 1.36
106 1 -0.5 -1 1.35 1.31 1.37
107 1 -0.5 -0.5 1.38 1.28 1.24
108 1 -0.5 0 1.33 1.31 1.33
109 1 -0.5 0.5 1.36 1.27 1.33
110 1 -0.5 1 1.40 1.30 1.22
111 1 0 -1 1.39 1.37 1.36
112 1 0 -0.5 1.37 1.33 1.34
113 1 0 0 1.35 1.34 1.34
114 1 0 0.5 1.39 1.27 1.32
115 1 0 1 1.41 1.33 1.31
116 1 0.5 -1 1.40 1.39 1.32
117 1 0.5 -0.5 1.38 1.37 1.34
118 1 0.5 0 1.38 1.31 1.32
119 1 0.5 0.5 1.36 1.29 1.35
120 1 0.5 1 1.39 1.35 1.33
121 1 1 -1 1.41 1.37 1.35
122 1 1 -0.5 1.41 1.37 1.33
123 1 1 0 1.40 1.36 1.31
124 1 1 0.5 1.40 1.33 1.3
125 1 1 1 1.36 1.31 1.32
The developed model is checked for adequacy by ANOVA and F-test.

Table 2.5 presents the ANOVA table for the second order model proposed for
fractal dimension, D given in Eq. 2.1. It can be seen that the P-value is less than
0.05 which means that the model is significant at 95% confidence level. Also the
Table 2.5 ANOVA for second order model for D in CNC milling of mild steel
Source Degrees of freedom Sum of squares Mean squares Fcalculated F0.05 P
Regression 9 0.051657 0.005740 8.25 1.96 0
Residual error 115 0.080004 0.000696
Total 124 0.131661
Table 2.6 ANOVA for model coefficients for D in CNC milling of mild steel
d 4 0.0293648 0.0073412 13.15 2.52 0.000
N 4 0.0146848 0.0036712 6.58 2.52 0.000
f 4 0.0052688 0.0013172 2.36 2.52 0.063
d*N 16 0.0232112 0.0014507 2.60 1.82 0.004
d*f 16 0.0075072 0.0004692 0.84 1.82 0.636
N*f 16 0.0159072 0.0009942 1.78 1.82 0.054
Error 64 0.0357168 0.0005581
Total 124 0.1316608
Fig. 2.1 Main effect plot for

mild steel
calculated value of the F-ratio is more than the standard value of the F-ratio for
D. It means the model is adequate at 95% confidence level to represent the rela-
tionship between the machining response and the considered machining parame-
ters of the CNC end milling process on mild steel. Table 2.6 represents the
ANOVA table for individual model coefficients where it can be seen that there are
three effects with a P-value less than 0.05 which means that they are significant at
95% confidence level. These significant effects are: depth of cut, spindle speed and
the interaction between spindle speed and depth of cut. Figure 2.1 depicts the main
effects plot for the fractal dimension and the design factors considered in the
present study. From this figure also, it is seen that spindle speed and depth of cut
have the significant effect on fractal dimension. To see the effects of process
parameters on fractal dimension in the experimental regime, three dimensional
surface as well as contour plots are presented at high level and low level of the
parameters (Figs. 2.2, 2.3, 2.4).
Fig. 2.2 Surface and contour plot of fractal dimension for mild steel: a at high level of spindle
speed, b at low level of spindle speed
Fig. 2.3 Surface and contour plot of fractal dimension for mild steel: a at high level of depth of
cut, b at low level of depth of cut
Fig. 2.4 Surface and contour plot of fractal dimension for mild steel: a at high level of feed rate,
b at low level of feed rate
Table 2.7 ANOVA for second order model for D in CNC milling of brass
Regression 9 0.138293 0.015366 36.35 1.96 0
Residual Error 115 0.048614 0.000423
Total 124 0.186907
Table 2.8 ANOVA for model coefficients for D in CNC milling of brass
d 4 0.0006512 0.0001628 0.61 2.52 0.654
N 4 0.1095792 0.0273948 103.26 2.52 0.000
f 4 0.0264432 0.0066108 24.92 2.52 0.000
d*N 16 0.0043968 0.0002748 1.04 1.82 0.433
d*f 16 0.0092528 0.0005783 2.18 1.82 0.015
N*f 16 0.0196048 0.0012253 4.62 1.82 0.000
Error 64 0.0169792 0.0002653
Total 124 0.1869072
Fig. 2.5 Main effect plot for

brass
2.3.2 RSM for Brass
The second order response surface equation for fractal dimension in brass milling
is obtained in terms of coded values of design factors as:
D ¼1:3130 þ 0:0015d þ 0:0408N 0:0175f þ 0:0071dN þ 0:0014df

0:0098Nf 0:0008d2 0:0142N 2 þ 0:0163f 2 ð2:2Þ

D given in Eq. 2.2. It can be seen that the P-value is less than 0.05 which means
that the model is significant at 95% confidence level. Also the calculated value of
the F-ratio is more than the standard value of the F-ratio for D. It means the model
is adequate at 95% confidence level to represent the relationship between the
Fig. 2.6 Surface and contour plot of fractal dimension for brass: a at high level of spindle speed,
b at low level of spindle speed
Fig. 2.7 Surface and contour plot of fractal dimension for brass: a at high level of depth of cut,
b at low level of depth of cut
machining response and the considered machining parameters of the CNC end
milling process on brass. Table 2.8 represents the ANOVA table for individual
model coefficients where it can be seen that spindle speed, feed rate, the interaction
between spindle speed and feed rate and the interaction of depth of cut and feed
rate are significant factors at 95% confidence level. Figure 2.5 depicts the main
effects plot for the fractal dimension and the design factors considered in the
present study. From this figure also, it is seen that spindle speed and feed rate have
the significant effect on fractal dimension. Figures 2.6, 2.7, 2.8 show the estimated
three-dimensional surface as well as contour plots for fractal dimension as func-
tions of the independent machining parameters. All these figures clearly depict the
variation of fractal dimension with controlling variables within the experimental
regime.
Fig. 2.8 Surface and contour plot of fractal dimension for brass: a at high level of feed rate, b at
low level of feed rate
Table 2.9 ANOVA for second order model for D in CNC milling of aluminium
Regression 9 0.025241 0.002805 4.5 1.96 0
Residual error 115 0.0717 0.000624
Total 124 0.096941
2.3.3 RSM for Aluminium
The second order response surface equation has been fitted using Minitab software
for the response variable D. The equation can be given in terms of the coded values
of the independent variables as:
D ¼1:3433 0:0128d þ 0:0013N 0:0062 f 0:0011dN 0:0095 df

þ 0:0122Nf 0:0135d 2 þ 0:0041 N 2 þ 0:0056 f 2 ð2:3Þ
D given in Eq. 2.3. It can be appreciated that the P-value is less than 0.05 which
means that the model is significant at 95% confidence level. Also the calculated
value of the F-ratio is more than the standard value of the F-ratio for D. It means
the model is adequate at 95% confidence level to represent the relationship
between the machining response and the considered machining parameters of the
CNC end milling process. Table 2.10 represents the ANOVA table for individual
model coefficients where it can be seen that depth of cut and the interaction
between spindle speed and feed rate are significant at 95% confidence level.
Figure 2.9 depicts the main effects plot for the fractal dimension and the design
factors considered in the present study. From this figure also, it is seen that depth
of cut has the significant effect on fractal dimension. Figures 2.10, 2.11, 2.12 show
the estimated three-dimensional surface as well as contour plots for fractal
Table 2.10 ANOVA for model coefficients for D in CNC milling of aluminium
d 4 0.0146608 0.0036652 6.76 2.52 0.000
N 4 0.0004928 0.0001232 0.23 2.52 0.922
f 4 0.0032048 0.0008012 1.48 2.52 0.219
d*N 16 0.0110272 0.0006892 1.27 1.82 0.243
d*f 16 0.0102352 0.0006397 1.18 1.82 0.307
N*f 16 0.0226432 0.0014152 2.61 1.82 0.003
Error 64 0.0346768 0.0005418
Total 124 0.0969408
Fig. 2.9 Main effect plot for aluminium
Fig. 2.10 Surface and contour plot of fractal dimension for aluminium: a at high level of spindle
dimension as functions of the independent machining parameters. All these figures

clearly depict the variation of fractal dimension with controlling variables within
the experimental regime.
Fig. 2.11 Surface and contour plot of fractal dimension for aluminium: a at high level of depth
of cut, b at low level of depth of cut
Fig. 2.12 Surface and contour plot of fractal dimension for aluminium: a at high level of feed
rate, b at low level of feed rate
2.4 Closure
For three different work-piece materials, fractal dimension models are developed
in CNC end milling using response surface method. The second order response
models have been validated with analysis of variance. A comparison of the
response surface models for fractal dimension in different materials reveals the fact
that these models are material specific or in other words, the tool-work-piece
material combination plays a vital role in fractal dimension of the generated
surface profile. Also the effect of the cutting parameters on fractal dimension is
different for different materials as evidenced from Tables 2.6, 2.8 and 2.10.
Accordingly, optimum machining parameter combinations for fractal dimension
depend greatly on the work-piece material within the experimental domain.
2.4 Closure 43
However, it can be concluded that it is possible to select a combination of spindle

speed, depth of cut and feed rate for achieving the surface topography with desired
fractal dimension within the constraints of the available machine. Thus with the
known boundaries of desired fractal dimension and machining parameters,
machining can be performed with a relatively high rate of success.
Chapter 3
Fractal Analysis in CNC Turning
Abstract Modeling of fractal dimension in CNC turning of mild steel, brass and
aluminium work-pieces are presented in this chapter. Spindle speed, feed rate and
depth of cut are considered as the process parameters. The generated surface in
CNC tuning operations are measured and processed to calculate fractal dimension.
The experimental results are then analyzed with RSM. From the analysis, it is seen
that the work-piece speed is the most significant factor affecting the fractal
dimension for mild steel turning whereas feed rate is the significant factor for both
brass and aluminium materials. It can be concluded from the analysis that for all
the materials, with increase in feed rate, fractal dimension, D decreases. So, to get
smoother surface, feed rate should be at low level. With increase in spindle speed,
fractal dimension increases giving smoother surface for mild steel turning.
Keywords Fractal dimension (D) CNC turning RSM Mild steel Brass
Aluminium
3.1 Introduction
Turning operation is an old and very common machining process in the industry. In
recent times, uses of computer numerically controlled (CNC) machines have
become popular to minimize the operator input and to get higher surface finish.
Turning operations are carried out on a lathe. In turning, there are several machining
parameters which control the surface quality of the machined work-piece which
include cutting conditions, tool variables and work-piece variables. Cutting condi-
tions include speed, feed and depth of cut where as tool variables include tool
material, nose radius, rake angle, cutting edge geometry, tool vibration, tool over-
hang, tool point angle, etc. and work-piece variables include material hardness and
other mechanical properties. It is very difficult to consider all the parameters that

46 3 Fractal Analysis in CNC Turning
Table 3.1 Process parameters levels used in the experimentation for all the three materials
Process variables Unit Levels
-1.682 -1 0 1 1.682
A Depth of cut(d) mm 0.032 0.1 0.2 0.3 0.368
B Spindle speed(N) rpm 528 800 1,200 1,600 1,872
C Feed rate(f) mm/rev 0.0224 0.07 0.14 0.21 0.2576
Table 3.2 Design matrix of the rotatable CCD design with coded and actual value
Std. order Run order Coded values Actual values
d N f d N f
1 20 -1 -1 -1 0.1 800 0.07
2 1 1 -1 -1 0.3 800 0.07
3 9 -1 1 -1 0.1 1,600 0.07
4 11 1 1 -1 0.3 1,600 0.07
5 7 -1 -1 1 0.1 800 0.21
6 8 1 -1 1 0.3 800 0.21
7 13 -1 1 1 0.1 1,600 0.21
8 3 1 1 1 0.3 1,600 0.21
9 10 -1.682 0 0 0.032 1,200 0.14
10 6 1.682 0 0 0.368 1,200 0.14
11 5 0 -1.682 0 0.2 528 0.14
12 14 0 1.682 0 0.2 1,872 0.14
13 12 0 0 -1.682 0.2 1,200 0.0224
14 19 0 0 1.682 0.2 1,200 0.2576
15 2 0 0 0 0.2 1,200 0.14
16 4 0 0 0 0.2 1,200 0.14
17 17 0 0 0 0.2 1,200 0.14
18 16 0 0 0 0.2 1,200 0.14
19 18 0 0 0 0.2 1,200 0.14
20 15 0 0 0 0.2 1,200 0.14
control the surface quality. In a turning operation, it is the vital task to select the
cutting parameters to achieve the high quality performance. For this, modeling of the
surface roughness is necessary to predict and control the desired level of surface
roughness. In this study, CNC turning operations are carried out varying the
machining parameters, viz., depth of cut (mm), spindle speed (rpm) and feed rate
(mm/rev). Machining surfaces are further analyzed to find out the profile fractal
dimension. These experimental results are further analyzed using response surface
methodology.
In a turning operation, there are many factors that can affect the surface roughness.
But, the review of literature shows that the depth of cut (d, mm), spindle speed (N,
rpm) and feed rate (f, mm/rev) are the most widespread machining parameters
taken by the researchers. In the present study these are selected as design factors
while other parameters have been assumed to be constant over the experimental
domain. The process variables with their values are listed in Table 3.1. For the
experimentation, a rotatable central composite design (Sect. 1.5.2) is selected and
the experimental plan consists of experiment run order, standard order, coded
values and actual values of process parameters as shown in Table 3.2.
3.2.2 Machine Used
The machine used for the turning is a JOBBERXL CNC lathe having the control
system FANUC Series Oi Mate-Tc and equipped with maximum spindle speed of
3,500 rpm, feed rate 15–20 m/rev and KVA rating-16 KVA. For generating the
turned surfaces, CNC part programs for tool paths were created with specific
commands.
3.2.3 Cutting Tool Used
Coated carbide tools are known to perform better than uncoated carbide tools.
Thus commercially available CVD coated carbide tools were used in this inves-
tigation. The tool holder is used as the PTGNR-25-25 M16 050, WIDIA and insert
used as the TNMG 160404–FL, WIDIA. The tool is coated with titanium nitride
coating having hardness, density and transverse rupture strength as 1,570 HV,
14.5 g/cc and 3,800 N/mm2. The compressed coolant WS 50–50 with a ratio of
1:20 with water was used as cutting environment.
The present study was carried out with three different workpiece materials, viz.,
6061-T4 aluminium, mild steel (AISI 1040) and medium leaded brass UNS
C34000. All the specimens were in the form of bar with diameter 20 mm and
Table 3.3 Experimental results for CCD

Std. order D for mild steel D for brass D for aluminium
1 1.370 1.435 1.417
2 1.300 1.437 1.315
3 1.362 1.395 1.392
4 1.410 1.420 1.440
5 1.267 1.300 1.300
6 1.282 1.292 1.302
7 1.390 1.297 1.292
8 1.417 1.297 1.262
9 1.320 1.355 1.377
10 1.370 1.367 1.360
11 1.300 1.375 1.397
12 1.420 1.355 1.347
13 1.360 1.380 1.485
14 1.290 1.257 1.252
15 1.397 1.350 1.362
16 1.400 1.375 1.370
17 1.415 1.375 1.385
18 1.415 1.362 1.367
19 1.402 1.377 1.300
20 1.412 1.377 1.297
length 60 mm. The chemical and mechanical properties of the materials are
already given in Table 2.3 (Chap. 2).
To get machined surfaces, CNC turning operations are carried out for different
combinations of spindle speed, feed rate and depth of cut. Three different work-
piece materials are considered viz. mild steel, brass and aluminium. The generated
surfaces are measured using Talysurf instrument (Sect. 1.3.7) and further pro-
cessed to get fractal dimension (D). The experimental results are used for further
analyses using response surface methodology (RSM) to model fractal dimension.
For RSM, a rotatable central composite design of experiment is considered and the
experimental results are presented in Table 3.3.
The influences of the machining parameters on fractal dimension have been
assessed for three different materials using RSM. The whole analyses are done
using Minitab software. The results of RSM analyses are presented below.
Table 3.4 ANOVA for second order model for mild steel
Source DF SS MS F F0.05 P
Regression 9 0.049 0.005,409 16.96 3.02 0
Residual Error 10 0.0032 0.000319
Total 19 0.052
Table 3.5 Full ANOVA table for mild steel model

Source Sum of squares DF Mean square F value P value
Model 0.049 9 0.005409 16.96 0.0001
A–d 0.0007933 1 0.0007933 2.49 0.1458
B–N 0.023 1 0.023 72.48 0.0001
C–f 0.003009 1 0.003009 9.44 0.0118
AB 0.00211 1 0.00211 6.62 0.0277
AC 0.0005281 1 0.0005281 1.66 0.2271
BC 0.003003 1 0.003003 9.42 0.1190
A2 0.005608 1 0.005608 17.59 0.0018
B2 0.002998 1 0.002998 9.40 0.0119
C2 0.010 1 0.010 32.46 0.0002
Residual 0.00318 10 0.0003189
Lack-of-fit 0.002871 5 0.0005742 9.04 0.0152
Pure error 0.0003177 5 0.00006354
Cor total 0.05186 19
The second order response model is developed using Minitab in terms of coded
values of the independent machining parameters, viz., work-piece speed, feed rate
and depth of cut. The response model for mild steel material is given in the
following equation.
D ¼ 1:40674 þ 0:00453 d þ 0:02446 N 0:00883 f þ 0:005745 dN þ 0:002873 df

þ 0:006850Nf 0:00697 d 2 0:00510 N 2 0:00947 f 2 ð3:1Þ
The developed model is also checked for adequacy. Table 3.4 represents the
ANOVA table for the second order response model developed for D. It is clear that
the developed model is significant at 95% confidence level. The calculated value
of F ratio is greater than the tabulated value of F ratio and it can be concluded that
the model is adequate at 95% confidence level. ANOVA table for mild steel
(Table 3.5) shows that work speed, feed rate, interaction of depth of cut with
work-piece speed are significant factors at 95% confidence level. The main effects
plots for fractal dimension are shown in Fig. 3.1. From the main effect plots, it is
seen that work-piece speed and feed rate are significant. It can also be concluded
that with increase in work speed, D increases but with increase in feed rate,
D decreases in mild steel turning. Response surface plots are also generated using
Fig. 3.1 Main effect plots for mild steel
Minitab. Figs. 3.2, 3.3 and 3.4 show the estimated three dimensional surface as
well as contour plots for fractal dimension as functions of two independent
machining parameters while the third machining parameter is held constant. All
these figures clearly depict the variation of fractal dimension with controlling
variables within the experimental regime.
3.3.2 RSM for Brass
The second order response model for brass material is presented in terms of coded
values of work-piece speed, feed rate and depth of cut in Eq. 3.2.
D ¼ 1:36919 þ 0:00179 d 0:00386 N 0:03074 f þ 0:00133 dN 0:00155 df

þ 0:00265 Nf 1:21695 1004 d 2 þ 0:00035 N 2 0:00543 f 2 ð3:2Þ
The developed model is checked for adequacy and ANOVA result for the
model is presented in Table 3.6. From the ANOVA table, it is seen that the model
is significant and adequate at 95% confidence level. From the full ANOVA table
(Table 3.7), it is seen that feed rate is the main significant factor affecting fractal
dimension in brass turning. The calculated F-value of the lack-of-fit for D is much
lower than the tabulated value of the F-distribution (tabulated value 5.05) found
from the standard table at 95% confidence level. It implies that the lack-of-fit is not
significant relative to pure error. From the main effect plot (Fig. 3.5), it is seen that
only feed rate is significant and the other parameters are insignificant. It is also
Fig. 3.2 Surface and contour plot of fractal dimension, D for mild steel: a at high level of spindle
Fig. 3.3 Surface and contour plot of fractal dimension, D for mild steel: a at high level of depth
of cut, b at low level of depth of cut
Fig. 3.4 Surface and contour plot of fractal dimension, D for mild steel: a at high level of feed
Table 3.6 ANOVA for second order model for brass

Regression 9 0.041 0.004603 13.56 3.02 0
Residual Error 10 0.0034 0.000319
Total 19 0.045
Table 3.7 Full ANOVA table for brass model

Source Sum of squares DF Mean square F value P value
Model 0.041 9 4.603E-3 13.56 0.0002
A–d 1.232E-4 1 1.232E-4 0.36 0.5602
B–N 5.753E-4 1 5.753E-4 1.70 0.2221
C–f 0.036 1 0.036 107.57 0.0001
AB 1.125E-4 1 1.125E-4 0.33 0.5775
AC 1.531E-4 1 1.531E-4 0.45 0.5169
BC 4.500E-4 1 4.500E-4 1.33 0.2763
A2 1.707E-6 1 1.707E-6 5.032E-3 0.9448
B2 1.389E-5 1 1.389E-5 0.041 0.8437
C2 3.405E-3 1 3.405E-3 10.03 0.0100
Residual 3.393E-3 10 3.189E-4
Lack-of-fit 2.775E-3 5 5.551E-4 4.49 0.0624
Pure Error 6.177E-4 5 1.235E-4
Cor Total 0.045 19
Fig. 3.5 Main effect plots for brass
seen that with increase in feed rate, D decreases. The estimated three dimensional
surface as well as contour plots for fractal dimension are presented in Figs. 3.6, 3.7
and 3.8. To draw these surface plots, fractal dimension is plotted as functions of
two independent machining parameters while the third machining parameter is
Fig. 3.6 Surface and contour plot of fractal dimension, D for brass: a at high level of spindle
Fig. 3.7 Surface and contour plot of fractal dimension, D for brass: a at high level of depth of
cut, b at low level of depth of cut
Fig. 3.8 Surface and contour plot of fractal dimension, D for brass: a at high level of feed rate,
b at low level of feed rate
Table 3.8 ANOVA for second order model for aluminium

Regression 9 0.041 0.005409 16.96 3.02 0
Residual Error 10 0.0032 0.000319
Total 19 0.052
Table 3.9 Full ANOVA table for aluminium model

Source Sum of squares df Mean square F value P value
Model 0.052 9 5.841E-3 3.37 0.0359
A–d 9.174E-4 1 9.174E-4 0.53 0.4825
B–N 7.307E-5 1 7.307E-5 0.042 0.8410
C–f 0.047 1 0.047 27.08 0.0004
AB 1.726E-3 1 1.726E-3 1.000 0.3407
AC 9.453E-5 1 9.453E-5 0.055 0.8196
BC 2.720E-3 1 2.720E-3 1.580 0.2377
A2 1.751E-5 1 1.751E-5 0.010 0.9217
B2 8.496E-5 1 8.496E-5 0.049 0.8288
C2 1.751E-5 1 1.751E-5 0.010 0.9217
Residual 0.017 10 1.724E-3
Lack-of-fit 9.952E-3 5 1.990E-3 1.36 0.3707
Pure error 7.293E-3 5 1.459E-3
Cor total 0.070 19
held constant at high and low levels. All these figures clearly depict the variation
of fractal dimension with controlling variables within the experimental regime.
The second order response model for aluminium material is presented in terms of
coded values of the independent machining parameters, viz., work-piece speed,
feed rate and depth of cut in Eq. 3.3.
D ¼ 1:34809 0:00487 d 0:00138 N 0:03477 f þ 0:00519 dN þ 0:00122 df

0:00652 Nf þ 0:000390 d2 þ 0:00086 N 2 þ 0:00039 f 2 ð3:3Þ
D of aluminium material. It is observed that the model is significant and adequate
at 95% confidence level. From the full ANOVA table (Table 3.9), it is seen that
feed rate is the main significant factor affecting fractal dimension in aluminium
turning. The calculated F-value of the lack-of-fit for D is much lower than the
tabulated value of the F-distribution (tabulated value 5.05) found from the stan-
dard table at 95% confidence level. From the main effects plot (Fig. 3.9), it is seen
Fig. 3.9 Main effect plots for aluminium material
Fig. 3.10 Surface and contour plot of fractal dimension, D for aluminium: a at high level of
spindle speed, b at low level of spindle speed
that only feed rate is significant. It is also seen that with increase in feed rate,
fractal dimension, D decreases. Response surface plots are also generated using
Minitab. Figs. 3.10, 3.11 and 3.12 show the estimated three dimensional surface as
well as contour plots for fractal dimension as functions of two independent
machining parameters. The third machining parameter is held constant at high and
low levels. From these figures, variations of fractal dimension with machining
parameters can be observed within the experimental regime.
Fig. 3.11 Surface and contour plot of fractal dimension, D for aluminium: a at high level of
depth of cut, b at low level of depth of cut
Fig. 3.12 Surface and contour plot of fractal dimension, D for aluminium: a at high level of feed
3.4 Closure
Response surface models for three materials viz. mild steel, brass and aluminium
are developed in CNC turning. All the developed second order models are ade-
quate at 95% confidence level. From the analysis, it is seen that the work-piece
speed is the most significant factor affecting the fractal dimension for mild steel
turning whereas feed rate is the significant factor for both brass and aluminium
materials. It can be concluded from the analysis that for all the materials, with
increase in feed rate, fractal dimension, D decreases. So, to get smoother surface,
feed rate should be at low level. With increase in spindle speed, fractal dimension
increases giving smoother surface for mild steel turning.
Chapter 4
Fractal Analysis in Cylindrical Grinding
Abstract This chapter presents the fractal dimension modeling in cylindrical

grinding of mild steel, brass and aluminium work-pieces. The experimentations are
carried out for different combinations of work-piece speed, longitudinal feed and
radial infeed. The generated surfaces are measured and processed to calculate
fractal dimension. The experimental results are then analyzed with RSM. The
longitudinal feed rate is the most significant factor affecting the fractal dimension
for mild steel, whereas for brass, work-piece speed and longitudinal feed rate are
the most significant factors. For aluminium materials, all the three process
parameters are the significant factors affecting fractal dimension.
Keywords Fractal dimension (D) Cylindrical grinding RSM Mild steel

Brass Aluminium
4.1 Introduction
Grinding is one of the common machining processes. In today’s production, fin-

ishing of components is done by grinding due to the fact that it has the great
potential to replace other machining processes and to achieve significant reduction
in production time and cost. The acceptance of grinding as a finishing process is
connected with a high form and size accuracy, high surface finish and surface
integrity of the work-piece. In grinding there are several parameters which control
the surface quality. It is very difficult to consider all the parameters that control the
surface roughness for a particular manufacturing process. In this study, only three
machining parameters are considered viz. work-piece speed, longitudinal feed and
radial infeed. Also the study is conducted on three different materials, AISI 1040
mild steel, UNS C34000 brass and 6061-T4 aluminium to consider the effect
of workpiece material variation. The experimental results are analyzed using

58 4 Fractal Analysis in Cylindrical Grinding
Table 4.1 Process variables and their levels

Parameters Unit Notation 1 2 3 4
Work-piece speed rpm N 56 80 112 160
Long feed mm/rev f 11.33 17.00 22.66 28.33
Radial infeed mm d 0.02 0.04 0.06 0.08
response surface modeling (RSM). The experimental details and the results are
discussed below.
The process parameters chosen here are work-piece speed (N) in rpm, longitudinal
feed (f) in mm/rev and radial infeed (d) in mm. The process variables/design
factors with their values on different levels are listed in Table 4.1 for three dif-
ferent work-piece materials. The selection of the values of the variables is limited
by the capacity of the machine used in the experimentation as well as the rec-
ommended specifications for different work-piece-tool material combination. Four
levels, having nearly equal spacing, within the operating range of the parameters
are selected for each of the factors. By selecting four levels, the curvature or non-
linearity effects can be studied. In the present investigation, full factorial design of
experiment is considered for the experimentation and for four level three factors
total 64 experimental trials are carried out for each of the work materials.
4.2.2 Machine Used
The machine used for grinding is a HMT made, K130U grinding machine
equipped with maximum wheel speed of 1910 rpm. The wheel signature of the
machine is A70K5V10 and wheel diameter of 270 mm. The maximum grinding
length is about 340 mm. The compressed coolant WS 50–50 with a ratio of 1:20
was used as cutting environment. The details of the machine used in this study are
shown in the Table 4.2.
The present study is carried out with three different materials, viz., AISI 1040
steel, medium leaded brass UNS C34000 and 6061-T4 aluminium. The chemical
composition and mechanical properties of the work-piece materials are already
discussed in the Chap. 2 (Table 2.3). All the specimens are in the form of round
bars of diameter 48 mm and length 50 mm.
Table 4.2 Specification of the cylindrical grinding machine used in the experiment
Make HMT Model K130U Machine 57169
No
Maximum grinding length 340 mm
Maximum distance between 340 mm
centers
Maximum travel of the table 310 mm
Maximum swivel of the table 200 mm
Grinding wheel
Wheel speed 1910 and 2120 rpm
Wheel Signature A70K5V10
Wheel Diameter 270 mm
Face width 40 mm
Bore diameter 50 mm
Work head
Number of speed 8 (56-80-112-160-224-315-
450-630)
Swivel 90° towards wheel and 30°
away from wheel
Morse taper 3
Cylindrical grinding operations are carried out on mild steel, brass and aluminium
work-pieces to get machined surfaces for different combinations of work-piece
speed, longitudinal feed and radial infeed. The generated surfaces are measured
using Talysurf instrument and further processed to get fractal dimension (D). The
experimental results are used for further analyses using response surface meth-
odology (RSM) to model fractal dimension. For RSM, full factorial design of
experiments is considered and the design matrix and the experimental results of
cylindrical grinding of mild steel, brass and aluminium work-pieces are presented
in Table 4.3. The influences of the machining parameters viz. work-piece speed,
longitudinal feed, radial infeed on the profile fractal dimension for mild steel (AISI
1040), brass and aluminium grinding are presented below.
of the independent variables as the following:
D ¼ 1:53125 0:01081N 0:02637f 0:03162d 0:00095Nf

þ 0:00255Nd þ 0:00055fd þ 0:00219N 2 þ 0:00375f 2 þ 0:00375d 2 ð4:1Þ
Table 4.3 Design matrix of process variables and the experimental results
Std Run N Workpiece f Longitudinal d Radial D for D for D for
order order speed (rpm) Feed (mm/rev) infeed (mm) mild brass aluminium
steel
1 22 1 1 1 1.46 1.390 1.39
2 45 1 1 2 1.48 1.408 1.35
3 7 1 1 3 1.40 1.415 1.38
4 37 1 1 4 1.46 1.415 1.37
5 54 1 2 1 1.47 1.413 1.37
6 38 1 2 2 1.44 1.435 1.34
7 26 1 2 3 1.45 1.420 1.34
8 42 1 2 4 1.43 1.433 1.37
9 13 1 3 1 1.47 1.453 1.35
10 43 1 3 2 1.43 1.445 1.34
11 63 1 3 3 1.42 1.420 1.36
12 59 1 3 4 1.42 1.445 1.35
13 32 1 4 1 1.41 1.455 1.35
14 5 1 4 2 1.45 1.468 1.37
15 40 1 4 3 1.44 1.450 1.36
16 34 1 4 4 1.45 1.450 1.35
17 51 2 1 1 1.49 1.413 1.40
18 10 2 1 2 1.47 1.415 1.41
19 14 2 1 3 1.45 1.428 1.36
20 21 2 1 4 1.46 1.428 1.38
21 1 2 2 1 1.43 1.440 1.37
22 64 2 2 2 1.42 1.445 1.37
23 48 2 2 3 1.44 1.430 1.35
24 61 2 2 4 1.41 1.425 1.36
25 23 2 3 1 1.47 1.455 1.35
26 31 2 3 2 1.45 1.455 1.34
27 53 2 3 3 1.35 1.448 1.34
28 29 2 3 4 1.39 1.455 1.35
29 12 2 4 1 1.45 1.460 1.35
30 56 2 4 2 1.45 1.430 1.36
31 2 2 4 3 1.43 1.443 1.34
32 46 2 4 4 1.44 1.448 1.35
33 25 3 1 1 1.47 1.415 1.35
34 3 3 1 2 1.45 1.430 1.39
35 19 3 1 3 1.46 1.420 1.35
36 33 3 1 4 1.48 1.425 1.37
37 8 3 2 1 1.48 1.393 1.34
38 49 3 2 2 1.47 1.430 1.37
39 17 3 2 3 1.44 1.428 1.33
40 6 3 2 4 1.47 1.425 1.37
41 18 3 3 1 1.42 1.455 1.35
(continued)

Std Run N Workpiece f Longitudinal d Radial D for D for D for
order order speed (rpm) Feed (mm/rev) infeed (mm) mild brass aluminium
steel
42 44 3 3 2 1.47 1.450 1.34
43 15 3 3 3 1.43 1.453 1.32
44 57 3 3 4 1.45 1.468 1.35
45 36 3 4 1 1.43 1.463 1.36
46 35 3 4 2 1.39 1.468 1.36
47 28 3 4 3 1.45 1.428 1.37
48 41 3 4 4 1.42 1.455 1.39
49 11 4 1 1 1.45 1.440 1.41
50 47 4 1 2 1.48 1.408 1.38
51 30 4 1 3 1.44 1.435 1.38
52 27 4 1 4 1.47 1.430 1.38
53 39 4 2 1 1.47 1.453 1.39
54 16 4 2 2 1.45 1.445 1.35
55 9 4 2 3 1.46 1.460 1.35
56 52 4 2 4 1.47 1.455 1.35
57 24 4 3 1 1.47 1.470 1.35
58 50 4 3 2 1.41 1.453 1.34
59 4 4 3 3 1.46 1.450 1.34
60 20 4 3 4 1.42 1.465 1.35
61 62 4 4 1 1.45 1.472 1.40
62 60 4 4 2 1.45 1.465 1.39
63 58 4 4 3 1.45 1.470 1.40
64 55 4 4 4 1.44 1.445 1.38
Table 4.4 ANOVA for the response model of D for mild steel
Source DF Seq SS Adj MS F F0.05 P
Regression 9 0.012291 0.001366 2.47 2.04 0.020
Residual error 54 0.029903 0.000554
Total 63 0.042194
The analysis of variance (ANOVA) technique has been used to check the
adequacy of the developed model at 95% confidence level. As per this technique,
if the calculated value of the F-ratio of the regression model is more than the
standard tabulated value of table (F-table) for 95% confidence level, then the
model is considered adequate within the confidence limit. From Table 4.4, it is
observed that the developed model is adequate at 95% confidence level. From the
ANOVA table of individual parameters (Table 4.5), it can be concluded that the
longitudinal feed rate is the most significant factor affecting the fractal dimension
at 95% confidence level. The main effect plots of fractal dimension D is presented
in Fig. 4.1. From this figure, it is seen that longitudinal feed rate and radial infeed
have influences on fractal dimension. The estimated three dimensional surface as
Table 4.5 ANOVA for individual parameter of D for mild steel

Source DF SS MS Fcalculated F0.05 P
N 3 0.0021187 0.0007062 1.28 2.76 0.300
f 3 0.0074563 0.0024854 4.52 2.76 0.011
d 3 0.0034062 0.0011354 2.07 2.76 0.128
N*f 9 0.0059187 0.0006576 1.20 2.04 0.337
N*d 9 0.0034187 0.0003799 0.69 2.04 0.711
f*d 9 0.0050312 0.0005590 1.02 2.04 0.452
Error 27 0.0148437 0.0005498
Total 63 0.0421937
Fig. 4.1 Main effect plots for D in cylindrical grinding of mild steel
well as contour plots for D as function of the independent machining parameters

are presented in Figs. 4.2, 4.3, 4.4.
4.3.2 RSM for Brass
D ¼ 1:36367 þ 0:00134N þ 0:03714f þ 0:00450d 0:00114Nf 0:00060Nd

0:00300fd þ 0:00172N 2 0:00289f 2 þ 0:00094d 2
ð4:2Þ
Fig. 4.2 Surface and contour plots of D for mild steel: a at high level of radial infeed, b at low
level of radial infeed
Fig. 4.3 Surface and contour plots of D for mild steel: a at high level of work-piece speed, b at
low level of work-piece speed
Fig. 4.4 Surface and contour plots of D for mild steel: a at high level of longitudinal feed, b at
low level of longitudinal feed
Table 4.6 ANOVA for the response model of D for brass

Regression 9 0.016541 0.001838 12.46 2.04 0.000
Residual error 54 0.007968 0.000148
Total 63 0.024509
Table 4.7 ANOVA for individual parameter of D for brass

N 3 0.00305352 0.00101784 10.70 2.76 0.000
f 3 0.01316367 0.00438789 46.13 2.76 0.000
d 3 0.00016602 0.00005534 0.58 2.76 0.632
N*f 9 0.00210742 0.00023416 2.46 2.04 0.034
N*d 9 0.00153633 0.00017070 1.79 2.04 0.116
f*d 9 0.00191367 0.00021263 2.24 2.04 0.051
Error 27 0.00256836 0.00009512
Total 63 0.02450898
Fig. 4.5 Main effect plots for D in cylindrical grinding of brass
From ANOVA analysis of the second order model at 95% confidence level, it is
seen that the model is adequate (Table 4.6). From ANOVA table of individual
parameters (Table 4.7), it can be concluded that the work-piece speed, longitudinal
feed rate and interaction between work-piece speed and longitudinal feed are the
most significant factors affecting the fractal dimension. The main effect plots of
fractal dimension D is presented in Fig. 4.5. From this figure also, it is seen that
work-piece speed and longitudinal feed are significant while the radial infeed is
insignificant on fractal dimension in the studied range. The estimated three
dimensional surface as well as contour plots for D as function of the independent
machining parameters are presented in Figs. 4.6, 4.7, 4.8. It is seen that with
Fig. 4.6 Surface and contour plots of D for brass: a at high level of radial infeed, b at low level
of radial infeed
Fig. 4.7 Surface and contour plots of D for brass: a at high level of work-piece speed, b at low
level of work-piece speed
Fig. 4.8 Surface and contour plots of D for brass: a at high level of longitudinal feed, b at low
level of longitudinal feed
Table 4.8 ANOVA for the response model of D for aluminium

Regression 9 0.012886 0.001432 5.97 2.04 0.000
Residual error 54 0.012950 0.000240
Total 63 0.025836
Table 4.9 ANOVA for individual parameter of D for aluminium

N 3 0.0019672 0.0006557 4.75 2.76 0.009
f 3 0.0095922 0.0031974 23.15 2.76 0.000
d 3 0.0014672 0.0004891 3.54 2.76 0.028
N*f 9 0.0041516 0.0004613 3.34 2.04 0.007
N*d 9 0.0036266 0.0004030 2.92 2.04 0.015
f*d 9 0.0013016 0.0001446 1.05 2.04 0.431
Error 27 0.0037297 0.0001381
Total 63 0.0258359
increase in work-piece speed and longitudinal feed, the fractal dimension increases
i.e. the surface gets smoother while the radial infeed is kept constant at middle
level.
D ¼ 1:48062 0:01616N 0:07047f 0:02253d þ 0:00282Nf 0:00097Nd

þ 0:00202fd þ 0:00297N 2 þ 0:01078f 2 þ 0:00359d 2 ð4:3Þ
From the ANOVA analysis of the second order model at 95% confidence level,
it is seen that the model is adequate (Table 4.8). From the ANOVA table of
individual parameters (Table 4.9), it can be concluded that the work-piece speed,
longitudinal feed rate and radial infeed are the significant factors affecting the
fractal dimension at 95% confidence level. Also the interaction between work-
piece speed and longitudinal feed and between work-piece speed and radial infeed
are significant at 95% confidence interval. The main effect plots of fractal
dimension D is presented in Fig. 4.9. From this figure also, it is seen that work-
piece speed, longitudinal feed and radial infeed are significant in the studied range.
The variations of fractal dimension with two machining parameters are presented
in Figs. 4.10, 4.11, 4.12 while the third machining parameter is kept constant.
Fig. 4.9 Main effect plots for D in cylindrical grinding of aluminium
Fig. 4.10 Surface and contour plots of D for aluminium: a at high level of radial infeed, b at low
level of radial infeed
Fig. 4.11 Surface and contour plots of D for aluminium: a at high level of work-piece speed,
b at low level of work-piece speed
Fig. 4.12 Surface and contour plots of D for aluminium: a at high level of longitudinal feed, b at
low level of longitudinal feed
4.4 Closure
Response surface models for three materials viz. mild steel, brass and aluminium
are developed in cylindrical grinding. All the developed second order models are
adequate at 95% confidence level. For mild steel, the longitudinal feed rate is the
most significant factor affecting the fractal dimension whereas for brass materials,
the work-piece speed, longitudinal feed rate and interaction between work-piece
speed and longitudinal feed are the most significant factors. For brass materials,
with increase in work-piece speed and longitudinal feed, the fractal dimension
increases i.e. the surface gets smoother while the radial infeed is kept constant at
middle level. For aluminium materials, it is seen that the work-piece speed, lon-
gitudinal feed rate and radial infeed are the significant factors affecting the fractal
dimension.
Chapter 5
Fractal Analysis in EDM
Abstract In this chapter fractal dimension modeling in electrical discharge

machining is discussed. Machining operations are carried out for different com-
binations of pulse current, pulse-on time and pulse-off time on mild steel, brass and
tungsten carbide materials. The generated machined surfaces are measured to
calculate fractal dimension. The experimental results are then analyzed to model
fractal dimension using response surface methodology. From the response surface
models, it is seen that the effect of the cutting parameters on fractal dimension is
different for different materials. For tungsten carbide and brass, both pulse current
and pulse on time play a significant role in determining the fractal dimension while
for mild steel it is only the pulse current that plays the significant role. A com-
parison of the response surface models for fractal dimension in different materials
reveals the fact that these models are material specific.
Keywords Fractal dimension (D) EDM RSM Mild steel Brass Tungsten
carbide
5.1 Introduction
Electrical discharge machining (EDM) is a widespread machining technique used

for all types of conductive materials including metals, metallic alloys, graphite,
composites and ceramic materials. It is a non-conventional machining process
used for machining of difficult-to-machine materials and shapes with high degree
of accuracy (El-Hofy 2005). It is based on removing material from a part by means
of a series of repeated electrical discharges created by electric pulse generated at
short intervals between two electrodes; a tool electrode and a work-piece elec-
trode. The electrodes are separated by a dielectric fluid that makes it possible to
flush eroded particles from the gap between the electrodes. The electric spark

70 5 Fractal Analysis in EDM
Table 5.1 Variable levels used in the experimentation

Levels Current (I, amp) Pulse on time (ti, ls) Pulse off time (to, ls)
-1 3.125 50 50
0 6.250 100 75
1 9.375 150 100
Table 5.2 Design matrix of the FCC design (coded values and actual value of the factors)
Std. order Run order Coded value
Current (I) Pulse on time (ti) Pulse off time (t0)
1 5 -1 -1 -1
2 2 1 -1 -1
3 8 -1 1 -1
4 12 1 1 -1
5 18 -1 -1 1
6 16 1 -1 1
7 14 -1 1 1
8 1 1 1 1
9 9 -1 0 0
10 11 1 0 0
11 6 0 -1 0
12 13 0 1 0
13 19 0 0 -1
14 3 0 0 1
15 7 0 0 0
16 20 0 0 0
17 10 0 0 0
18 15 0 0 0
19 17 0 0 0
20 4 0 0 0
raises the surface temperature of both the tool and work-piece to a point that is in
excess of the melting or even boiling points of the substances. Thus material is
mainly removed in the liquid and vapor phases, and the surface generated consists
of debris either been melted or vaporized during machining. Since the tool does
not physically contact the work-piece, no mechanical stress is exerted on the work-
piece and the characteristics of the EDM process are thus not governed by the
mechanical properties of the work-piece material. Instead, the thermal and elec-
trical properties play a significant role in the process performance. The EDM
performance is characterized by three parameters, viz., material removal rate
(MRR), electrode wear rate (EWR) and surface roughness. In this study, surface
roughness is modeled based on fractal dimension for three different materials viz.
mild steel, brass and tungsten carbide materials in EDM using response surface
methodology (RSM). The experimental details and the results for different mate-
rials are presented below.
Table 5.3 Specification of the equipment used in the experimentation

Particulars Specification
Trade name TOOL CRAFT A 25
Type of construction ‘C’ type
Worktable 300 mm 9 200 mm
Fixed work tank 465 mm 9 270 mm 9 200 mm
Table longitudinal movement 100 mm
Table cross movement 175 mm
Maximum dielectric level over table 140 mm
Maximum work piece height 90 mm
Maximum work piece weight 45 kg
Servo head
Servo system Stepped drive
Quill travel 150 mm
Electrode platen size 100 mm sq
Accuracy of quill movement 0.01 mm over 200 mm
Dielectric system
Filtration flushing better than 10 l
Flushing side, 1.23 l/min (max)
Flushing pressure 15 kPa
Generator
Models A 25
Working current 25 A maximum through current selector
Pulse on time setting 2–2,000 ls
Pulse off time setting 2–2,000 ls
Power source connection 400/440 V, 50 Hz, 3-ph supply
There are a large number of factors that can be considered for machining of a
particular material in EDM. It is very difficult to conduct the experiment with
considering all the process variables. However, the review of literature shows that
the following three machining parameters are the most widespread among the
researchers and machinists to control the EDM process: pulse current (I, amp),
pulse-on time (ti, ls) and pulse-off time (to, ls). In the present study these are
selected as design factors while other parameters have been assumed to be constant
over the experimental domain. A face-centered central composite (FCC) design is
used with three levels of each of the three design factors. Considering three factors
and six replicates at the center point, the present design contains 20 experiments,
which were performed in a random order. The upper and lower limits of a factor
are coded as +1 and -1 respectively, the coded value being calculated using
Eq. 1.8. The process variables with their values on different levels are listed in
Table 5.1. The selection of the values of the variables is limited by the capacity of
Table 5.4 Composition and electrical/thermal properties of work-piece materials

Work Material Composition (%Wt) Electrical and thermal property
Tungsten carbide 94%WC–6%Co Electrical resistivity: 6 9 10-5
ohm-cm
Thermal conductivity: 84 W/m-K
Melting point: 2850°C
Mild Steel (AISI 0.42%C, 0.48%Mn, 0.17%Si, 0.02%P, Electrical resistivity:1.7 9 10-5
1040) 0.018%S, 0.1%Cu, 0.09%Ni, ohm-cm
0.07%Cr Thermal conductivity: 52 W/m-K
and balance Fe Melting point:1515°C
Brass (UNS 0.095%Fe, 0.9%Pb, 34%Zn and Electrical resistivity:6.6 9 10-6
C34000) balance Cu ohm-cm
Thermal conductivity:115°W/m-K
Melting point: 900°C
the machine used in the experimentation as well as the recommended specifica-

tions for different workpiece–tool material combinations. Table 5.2 shows the
experimental matrix of the FCC design employed in the present study.
5.2.2 Machine Used
The machine used for carrying out the machining operations is a ‘Toolcraft A25’
EDM machine having the stepped drive servo system and filtration flushing
capability. It is capable of generating maximum pulse current of 25 A, pulse on
time of 2,000 ls and pulse off time of 2,000 ls. The specification of the machine
is presented in Table 5.3.
The present study was carried out with three different work-piece materials, viz.,
tungsten carbide, AISI 1040 mild steel and medium leaded brass UNS C34000.
The chemical composition and electrical/thermal properties of the work-piece
materials are shown in Table 5.4. All the specimens were in the form of
20 mm 9 20 mm 9 4 mm blocks.
5.2.4 Tool Electrode Used
Electrolytic copper having 99.9% copper in composition and density 8,904 kg/m3
was used as tool electrode since it worked better in combination with the work-
piece materials considered in the present study. The tool electrode was in the form
of cylinder of diameter 15.9 mm and 50 mm in length mounted axially in line with
work-piece. The tool electrode was given negative polarity where as work-piece is
Table 5.5 Electrode material properties

Particulars Specifications
Material Electrolytic copper
Composition 99.09% copper
Density 8 904 kg/mm3
Melting point 1083 C°
Conductivity 101.41% IACS
Tensile strength 23.47 kg/mm2
Table 5.6 Experimental results

Std. order Run order D for WC D for MS D for Brass
1 5 1.383 1.413 1.440
2 2 1.350 1.310 1.406
3 8 1.356 1.330 1.430
4 12 1.250 1.276 1.400
5 18 1.410 1.426 1.453
6 16 1.313 1.306 1.423
7 14 1.356 1.426 1.420
8 1 1.216 1.283 1.386
9 9 1.390 1.333 1.413
10 11 1.270 1.286 1.410
11 6 1.323 1.346 1.440
12 13 1.250 1.343 1.420
13 19 1.320 1.346 1.430
14 3 1.386 1.316 1.406
15 7 1.263 1.356 1.420
16 20 1.313 1.306 1.423
17 10 1.310 1.363 1.410
18 15 1.343 1.306 1.400
19 17 1.310 1.330 1.416
20 4 1.293 1.316 1.416
positive polarity (Puertas et al. 2005). The properties of the tool electrode have
been given in Table 5.5. Kerosene was used as dielectric because of its high flash
point, good dielectric strength, transparent characteristics and low viscosity and
specific gravity.
As mentioned earlier, according to FCC design of experiments, machining oper-

ations are carried out to generate machined surfaces (EDMed). The generated
machined surfaces are measured with Talysurf and further processed to calculate
Table 5.7 ANOVA for second order model for D in EDM of mild steel
Source DF Seq SS Adj SS Adj MS F F0.05 P
Regression 9 0.029628 0.029628 0.003292 4.66 3.02 0.012
Residual Error 10 0.007057 0.007057 0.000706
Total 19 0.036685
Table 5.8 ANOVA for machining parameters for D in EDM of mild steel
I 2 0.021962 0.022226 0.011113 14.96 3.81 0
ti 2 0.004156 0.003419 0.001709 2.3 3.81 0.139
t0 2 0.000912 0.000912 0.000456 0.61 3.81 0.556
Error 13 0.009656 0.009656 0.000743
Total 19 0.036685
fractal dimension. Experimental results of fractal dimension for tungsten carbide,

mild steel and brass materials are presented in Table 5.6. The influences of the
machining parameters (I, ti and t0) on the profile fractal dimension D have been
assessed for three different materials using RSM. The second order model was
postulated in obtaining the relationship between the fractal dimension and the
machining variables. The analysis of variance (ANOVA) was used to check the
adequacy of the second order model. The results for the three different materials
are presented one by one.
The second order response surface equation for fractal in EDM of mild steel is
obtained in terms of coded values of design factors as:
D ¼ 1:33 0:0467 I 0:0143 ti þ 0:0083 to þ 0:0033 Iti 0:0133 Ito

þ 0:0117 ti to 0:0127 I 2 þ 0:0223 ti2 þ 0:0089 to2 ð5:1Þ
D given in Eq. 5.1. The developed model is significant at 95% confidence level as
the P-value is less than 0.05. Also the model is adequate at 95% confidence level
to represent the relationship between the machining response and the considered
machining parameters as the calculated value of the F-ratio is more than the
standard value of the F-ratio for D. Table 5.8 represents the ANOVA table for
individual machining parameters where it can be seen that only pulse current is the
significant parameter at 95% confidence level. Figure 5.1 shows the main effects
plot for the fractal dimension. From this figure also, it is seen that pulse current has
Fig. 5.1 Main effect plot of fractal dimension for mild steel
Fig. 5.2 Surface and contour plot of fractal dimension for mild steel: (a) at high level of pulse on
time, (b) at low level of pulse on time
Fig. 5.3 Surface and contour plot of fractal dimension for mild steel: (a) at high level of current,
(b) at low level of current
Fig. 5.4 Surface and contour plot of fractal dimension for mild steel: (a) at high level of pulse off
time, (b) at low level of pulse off time
Table 5.9 ANOVA for second order model for D in EDM of brass
Regression 9 0.003635 0.003635 0.000404 4.11 3.02 0.019
Residual Error 10 0.000982 0.000982 0.000098
Total 19 0.004616
significant effect on fractal dimension while pulse on time and pulse off time have
no effect on fractal dimension of the surface topography generated in EDM of mild
steel. The estimated three-dimensional surface as well as contour plots for fractal
dimension are presented in Figs. 5.2, 5.3, and 5.4. To draw these surface plots,
fractal dimension is plotted as functions of two independent machining parameters
while the third machining parameter is held constant. All these figures clearly
depict the variation of fractal dimension with controlling variables within the
experimental regime.
5.3.2 RSM for Brass
The second order response surface equation for the fractal dimension of brass sur-
faces machined in EDM is also obtained in terms of coded values of design factors as:
D ¼ 1:42 0:013 I 0:0107 ti 0:0017 to 0:0067 ti to 0:0068 I 2

þ 0:0115 ti2 0:0002 to2 ð5:2Þ

D given in Eq. 5.2. It is seen that the developed model is significant at 95%
Table 5.10 ANOVA for machining parameters for D in EDM of brass

I 2 0.001693 0.00182 0.00091 8.84 3.81 0.004
ti 2 0.001557 0.001502 0.000751 7.3 3.81 0.008
t0 2 2.83E - 05 2.83E - 05 1.42E - 05 0.14 3.81 0.873
Error 13 0.001338 0.001338 0.000103
Total 19 0.004616
Fig. 5.5 Main effect plot of

fractal dimension for brass
confidence level. Also the calculated value of the F-ratio is more than the standard
value of the F-ratio for D which implies the model is adequate at 95% confidence
level to represent the relationship between the machining response and the con-
sidered machining parameters of the EDM process on brass. Table 5.10 represents
the ANOVA table for individual machining parameters where it can be seen that
pulse current and pulse on time are the significant factors affecting fractal
dimension. Figure 5.5 depicts the main effects plot for the fractal dimension and
the design factors considered. From this figure also, it is seen that pulse current and
pulse on time have the significant effect on fractal dimension. Figures 5.6, 5.7, and
5.8 shows the estimated three-dimensional surface as well as contour plots for
fractal dimension. To draw these surface plots, fractal dimension is plotted as
functions of two independent machining parameters while the third machining
parameter is held constant. All these figures clearly show the variation of fractal
dimension with controlling variables within the experimental regime.
5.3.3 RSM for Tungsten Carbide
of the independent variables as:
Fig. 5.6 Surface and contour plot of fractal dimension for brass: (a) at high level of pulse on
time, (b) at low level of pulse on time
Fig. 5.7 Surface and contour plot of fractal dimension for brass: (a) at high level of current,
(b) at low level of current
Fig. 5.8 Surface and contour plot of fractal dimension for brass: (a) at high level of pulse off
time, (b) at low level of pulse off time
Table 5.11 ANOVA for second order model for D in EDM of tungsten carbide
Regression 9 0.046159 0.046159 0.005129 7.75 3.02 0.002
Residual error 10 0.006621 0.006621 0.000662
Total 19 0.05278
Table 5.12 ANOVA for machining parameters for D in EDM of tungsten carbide
I 2 0.026341 0.025185 0.012592 17.12 3.81 0
ti 2 0.01304 0.014658 0.007329 9.97 3.81 0.002
t0 2 0.003839 0.003839 0.00192 2.61 3.81 0.111
Error 13 0.00956 0.00956 0.000735
Total 19 0.05278
Fig. 5.9 Main effect plot of

fractal dimension for tungsten
carbide
D ¼ 1:31 0:0497 I 0:035ti þ 0:0023 to 0:0146Iti 0:0121 Ito

0:0029 ti to þ 0:0138 I 2 0:0296 ti2 þ 0:0371 to2 ð5:3Þ
The analysis of variance (ANOVA) and the F-ratio test have been performed to
check the adequacy of the developed model. Table 5.11 presents the ANOVA
table for the second order model proposed for D given in Eq. 5.3. It is seen that the
developed model is significant at 95% confidence level. Also the calculated value
of the F-ratio is more than the standard value of the F-ratio for D. It means the
model is adequate at 95% confidence level to represent the relationship between
the machining response and the considered machining parameters of the EDM
process. Table 5.12 represents the ANOVA table for individual machining
parameters where it can be seen that pulse current and pulse on time are significant
at 95% confidence level. Figure 5.9 shows the main effects plot for the fractal
dimension and the design factors considered in the present study. From this figure
also, it is seen that both pulse current and pulse on time have the significant effect
on fractal dimension while the effect of pulse off time is insignificant.
Fig. 5.10 Surface and contour plot of fractal dimension for tungsten carbide: (a) at high level of
pulse on time, (b) at low level of pulse on time
current, (b) at low level of current
pulse off time, (b) at low level of pulse off time
Figures 5.10, 5.11, and 5.12 shows the estimated three-dimensional surface as well
as contour plots for fractal dimension. To draw these surface plots, fractal
dimension is plotted as functions of two independent machining parameters while
the third machining parameter is held constant. These figures clearly depict the
variation of fractal dimension with controlling variables within the experimental
regime.
5.4 Closure
Response surface models are developed for fractal dimension in EDM of three
different materials. A comparison of the response surface models reveals the fact
that these models are material specific or in other words, the tool–workpiece
material combination plays a vital role in fractal dimension modeling. Also the
effect of the cutting parameters on fractal dimension is different for different
materials as evidenced from Table 5.8, Table 5.10 and Table 5.12. For tungsten
carbide and brass, both pulse current and pulse on time play a significant role in
determining the fractal dimension while for mild steel it is only the pulse current
that plays the significant role. Accordingly, optimum machining parameter com-
binations for fractal dimension depend greatly on the workpiece material within
the experimental domain. However, it can be concluded that it is possible to select
a combination of pulse current, pulse on time and pulse off time for achieving the
surface topography with desired fractal dimension within the constraints of the
available machine.
References
El-Hofy HAG (2005) Advanced machining processes. McGraw-Hill, New York

Puertas I, Luis CJ, Villa G (2005) Spacing roughness parameters study on the EDM of silicon
carbide. J Mater Process Technol 164–165:1590–1596

Fractal Analysis in Machining PDF

Diunggah oleh

Informasi Dokumen

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Fractal Analysis in Machining PDF

Diunggah oleh

Hak Cipta:

Format Tersedia

SpringerBriefs in Applied Sciences and Technology

For further volumes:

João Paulo Davim

ISSN 2191-5342 e-ISSN 2191-5350

Ó Prasanta Sahoo 2011

Cover design: eStudio Calamar, Berlin/Figueres

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

2 Fractal Analysis in CNC End Milling . . . . . . . . . . . . . . . . . . . . . . 29

3 Fractal Analysis in CNC Turning . . . . . . . . . . . . . . . . . . . . . . . . . 45

4 Fractal Analysis in Cylindrical Grinding . . . . . . . . . . . . . . . . . . . 57

5 Fractal Analysis in EDM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

Abstract The importance and usefulness of fractal dimension in describing sur-

P. Sahoo et al., Fractal Analysis in Machining, 1

on the self-affinity and self-similarity of surfaces at different scales. Fractals retain

1.2 Surface Metrology

Surface texture is a complex condition resulting from a combination of roughness

Fig. 1.1 Display of surface

principal direction of the predominant surface pattern, usually determined by the

Motor and gearbox

Fig. 1.2 Component parts of a typical stylus surface-measuring instrument

measurement is made. This length may include several sampling lengths—

though maximum height of measurement is limited to few microns. This method is

1.3 Fractal Characterization

1.3.1 Fractal Geometry

1.3.2 Fractal Dimension

Fig. 1.3 Formation of Koch

The definition of self-similarity is based on the property of equal magnification in

Fig. 1.4 Qualitative description of statistical self-affinity for a surface profile

1.3.5 Fractal Description of Roughness

where, G is a scaling constant. The parameter n1 corresponds to the low cut-off

1.3.6 Fractal Dimension Calculation

Fractal calculation mainly includes the calculation of profile fractal dimension

1.3.7 Fractal Dimension Measurement in the Present Study

In the present study, roughness profile measurement is done using a stylus-type

1.4 Review of Roughness Study in Machining

As surface roughness is an important parameter in the industry, many researchers

roughness decreases. Abouelatta and Madl (2001) have found a correlation

1.5 Design of Experiments

to variations in the design variable values. A scientific approach to planning the

1.5.1 Full Factorial Design

1.5.2 Central Composite Design

A Box–Wilson Central Composite Design, commonly called ‘‘Central Composite

Fig. 1.5 Face centered

1.6 Response Surface Methodology

P. Sahoo et al., Fractal Analysis in Machining, 29

Table 2.1 Variable levels used in the experimentation

2.2 Experimental Details

2.2.1 Design of Experiments

2.2.2 Machine Used

Table 2.2 Specification of CNC end milling machine

2.2.3 Cutting Tool Used

2.2.4 Work-Piece Materials

Table 2.3 Composition and mechanical properties of work-piece materials

2.3 Results and Discussion

2.3.1 RSM for Mild Steel

Table 2.4 (continued)

Table 2.4 (continued)

The developed model is checked for adequacy by ANOVA and F-test.

Fig. 2.1 Main effect plot for

Fig. 2.5 Main effect plot for

2.3.2 RSM for Brass

D ¼1:3130 þ 0:0015d þ 0:0408N 0:0175f þ 0:0071dN þ 0:0014df