Performance Evaluation of WEDM on Inconel 718 (Publish)
description
Transcript of Performance Evaluation of WEDM on Inconel 718 (Publish)
PERFORMANCE EVALUATION OF WIRE ELECTRODE
DISCHARGE MACHINING (WEDM) ON INCONEL 718
MOHD NIZAM BIN ALI
MAY 2010
i
To my beloved mother and father Ali bin Mohd Jos
Badriah bt. Mat Noh
My beloved wife and son Roslina bt. Mamat
Muhammad Ali Imran bin Mohd Nizam
ii
ABSTRACT
Superalloys are known as unique materials ever produced in manufacturing
industries. It’s capable to withstand in high temperature and the excellent resistance in
mechanical and chemical degradations. Inconel 718 is one of the superalloy material
whichs is which is widely used in aeronautical and aerospace industries. This nickel-
based superalloy is a high strength, thermal resistance with extreme toughness and work
hardening characteristics materials. It is also noted for its excellent corrosion resistance
in many conditions of engineering applications. Due to it extremely tough nature, the
machinability studies of this material had been carried out by many researchers for the
past few years. This master project presents the machining of Inconel 718 using wire
electro-discharge machining with zinc coated brass electrode wire diameter of 0.25mm.
The objective of this master project is mainly to investigate the performance of wire
electro-discharge machining on Inconel 718. This is done by observing the influence of
the various WEDM machining characteristics namely, surface roughness (Ra), sparking
gap (Gap), material removal rate (MRR) and cutting speed (CS). A full factorial design
of experiment (DOE) approach with two-level was employed to conduct this
experiment. Design expert software was used to perform the ANOVA analysis and
confirmation test was also conducted to verify and compare the results from the
theoretical prediction using software. Overall result showed that pulse duration (ON)
was the most significant factor that appeared to influence on all machining
characteristics that had been investigated. The experimental results also acceptable due
to the results obtain fall in acceptable values with less than 15% of margin error.
iii
ABSTRAK
Superaloi telah diketahui umum sebagai bahan yang unik yang pernah dihasilkan
di dalam industri pembuatan. Ianya mampu untuk bertahan pada suhu yang tinggi dan
mempunyai ketahanan yang lasak di dalam pelbagai applikasi kejuruteraan. Inconel 718
adalah salah satu bahan superaloi yang digunakan secara meluas terutamanya di dalam
industri aeronatikal dan angkasa lepas. Superaloi berasaskan bahan nickel ini
mempunyai kekuatan, rintangan haba dan ketahanan karat yang tinggi serta dicirikan
juga dengan pengerasan kerja yang baik. Berdasarkan kepada sifat Inconel 718 yang
tahan lasak, kajian kebolehmesinan bahan ini telah menjadi minat pengkaji sejak
beberapa tahun kebelakangan ini. Projek sarjana ini bertujuan untuk menyiasat prestasi
kebolehmesinan Inconel 718 menggunakan proses pemotongan nyahcas-elekrik
menggunakan wayar elektrod tembaga bersalut zink berdiameter 0.25mm. Ianya
melibatkan ujikaji serta pemerhatian terhadap ciri-ciri pemesinan Inconel 718 seperti
kekasaran permukaan (Ra), jarak percikan api (Gap), kadar pemotongan bahan (MRR)
dan kelajuan pemotongan (CS). Rekabentuk ujikaji dengan pendekatan full factorial dua
tahap telah digunakan di dalam ujikaji ini. Perisian Design Expert juga telah digunakan
untuk tujuan analisa varian (ANOVA) bagi setiap keputusan ujikaji. Bagi tujuan
penentuan ralat, ujikaji pengesahan dilaksanakan untuk menguji kesahihan dan
perbandingan diantara keputusan yang dihasilkan oleh ujikaji dan juga secara teori.
Secara keseluruhannya, keputusan ujikaji menunjukkan tempoh masa denyutan (ON)
adalah faktor yang paling signifikan mempengaruhi semua ciri pemesinan yang dikaji.
Data ujikaji juga menunjukkan perbezaan margin di bawah nilai 15% dan ianya adalah
didalam julat yang boleh diterima pakai di dalam analisa ini.
iv
CONTENTS
CHAPTER TITLE PAGE
DEDICATION i
ABSTRACT ii
ABSTRAK iii
CONTENTS iv
LIST OF TABLES viii
LIST OF FIGURES x
LIST OF ABBREVIATIONS AND SYMBOLS xii
LIST OF APPENDICES xiv
1 INTRODUCTION
1.1 Project Background & Rationale 1
1.2 Research Statement 3
1.3 Research Objectives 4
1.4 Scope of Study 4
1.5 Expected Results 5
v
2 LITERATURE REVIEW
2.1 Introduction 6
2.2 Wire Electrical Discharge Machining (WEDM) 7
2.3 WEDM Machining Parameter 12
2.3.1 Pulse duration (On-time) 13
2.3.2 Pulse interval (Off-Time) 14
2.3.3 Servo voltage 14
2.3.4 Peak current 14
2.4 Machining Characteristic 15
2.4.1 Effect on surface finish, (Ra) 16
2.4.2 Effect on material removal rate, MRR 17
2.4.3 Effect on spark gap, Gap 19
2.5 Wire Electrode 20
2.5.1 Copper wire 21
2.5.2 Brass wire 21
2.5.3 Coated wire 22
2.6 Nickel Based Superalloy and Their Machinibility. 24
2.6.1 Inconel 718 physical properties and mechanical
properties 26
2.7 Design of Experiment (DOE) 27
2.7.1 Two-level full factorial design 28
3 METHODOLOGY
3.1 Introduction 30
3.2 Research Design and Data Analysis 31
3.3 Research Design Variable 33
3.3.1 Machining parameters 33
vi
3.3.2 Machining characteristic 36
3.3.3 Surface roughness, Ra 36
3.3.4 Sparking gap, Gap 37
3.3.5 Cutting speed, Cs 37
3.3.6 Material removal rate, MRR 37
3.4 Research Procedure 38
3.5 Experimental Set-up 41
3.6 Experimental Equipment 43
4 EXPERIMENTAL RESULTS AND DATA ANALYSIS
4.1 Introduction 46
4.2 Experimental Results 47
4.3 Analysis of Results 54
4.3.1 Analysis results for surface roughness, Ra 54
4.3.2 Analysis results for sparking gap, Gap 58
4.3.3 Analysis results for material removal rate, MRR 63
4.3.4 Analysis results for cutting speed, CS 68
4.4 Confirmation Test 74
4.4.1 Confirmation test and results 75
4.5 Comparisons of the Test Results 77
4.6 Verification of Mathematical Models 79
5 DISCUSSIONS
5.1 Introduction 83
5.2 Surface Roughness, Ra 84
5.3 Sparking Gap, Ra 85
5.4 Material Removal Rate, MRR 86
vii
5.5 Cutting Speed, CS 88
6 CONCLUSIONS AND RECOMMENDATIONS
6.1 Conclusions 90
6.2 Recommendations 91
REFERENCE 92
APPENDIX A 99
APPENDIX B 102
APPENDIX C 105
APPENDIX D 108
viii
LIST OF TABLES
NO. TITLE PAGE
2.1 The designation of various types of coated wire and their applications 23
2.2 Chemical properties of Inconel 718 (%) 26
2.3 Physical properties of Inconel 718 27
2.4 Mechanical properties of Inconel 718 27
3.1 Two-level full factorial experiment with four factors 32
3.2 The design of machining parameters 34
3.3 Actual value of experiment design 35
3.4 Machining parameters set-up (constant parameters) 41
4.1 Experimental results of surface roughness (Ra) 48
4.2 Experimental results of sparking gap (Gap) 49
4.3 Experimental results for cutting speed (CS) 50
4.5 Experimental results for material removal rate (MRR) 52
4.6 Overall experimental results corresponded to each run 53
4.7 ANOVA for surface roughness (Ra) 55
4.8 ANOVA for sparking gap (Gap) 59
4.9 ANOVA for material removal rate (MMR) 63
4.10 ANOVA for cutting speed (CS) 68
4.11 Summary of the significant factors in WEDM Inconel 718 74
4.12 Quality characteristic of the machining performance 75
4.13 True value of confirmation test experiment 75
ix
NO. TITLE PAGE
4.14 Confirmation test results for surface roughness (Ra) 76
4.15 Confirmation test results for sparking gap 76
4.16 Confirmation test results for cutting speed (CS) 76
4.17 Confirmation test results for material removal rates (MRR) 76
4.18 Comparison test results for surface roughness (Ra) 78
4.19 Comparison test results for sparking gap (Gap) 78
4.20 Comparison test results for material removal rate (MRR) 78
4.21 Comparison test results for cutting speed (CS) 78
4.22 Margin of error for actual results and predicted values (%) 82
x
LIST OF FIGURES
NO. TITLE PAGE
2.1 Typical product by WEDM (AGIE Charmilles groups,
Charmilles the solution when to EDM, Geneva 2004) 7
2.2 Classification of EDM processes 8
2.3 Wire electrical discharge machining (WEDM) processes 9
2.4 Schematic of the thermal removal processes of WEDM
(Spark between electrode and workpiece perform the material removal) 10
2.5 Classification of major EDM research areas 11
2.6 Definition of kerf and overcut in WEDM 19
3.1 Research methodology 31
3.2 Flowchart of experiment steping 40
3.3 WEDM linear motor 5 axes – Sodick series AQ537L 44
3.4 Mitutoyo surface roughness measuring machine 44
3.5 Zeiss Axiotech high power optical microscope 45
4.1 Normal probability plot of residuals for surface roughness (Ra) 56
4.2 Residual vs. predict response for surface roughness (Ra) 57
4.3 Main effect plot for surface roughness (Ra) 57
4.4 Normal probability plot of residuals for sparking gap (Gap) 60
4.5 Residual vs. predict response for sparking gap (Gap) 61
4.6 Main effect plot for sparking gap (Gap) 61
4.7 Normal probability plot of residuals for material removal rate (MRR) 65
xi
NO. TITLE PAGE
4.8 Residual vs. predict response for material removal rate (MRR) 65
4.9 Main effect plot for material removal rate (MRR) 66
4.10 Interaction between SV*OFF for material removal rate (MRR) 67
4.11 Normal probability plot of residuals for cutting speed (CS) 70
4.12 Residual vs. predict responses for cutting speed (CS) 70
4.13 Main effect plot for cutting speed (CS) 71
4.14 Interaction plot of cutting speed (CS) 72
5.1 3D interaction graph for surface roughness (Ra) 84
5.2 3D interaction graph for sparking gap (Gap) 85
5.3 3D interaction graph for material removal rate 87
5.4 3D interaction graph of IP*SV for cutting speed (CS) 88
5.5 3D interaction graph of OFF*SV for cutting speed (CS) 89
xii
LIST OF ABBREVIATIONS AND SYMBOLS
AA - Arithmetic arrange
ANOVA - Analysis of variance
CI - Confidence interval
CLA - Centre line average
CNC - Computer numerical control
CS - Cutting speed
d - Machining distance
DC - Direct current
DOE - Design of experiment
EDM - Electrical discharge machining
FCC - Face centre cubic
FP - Flushing pressure
Gap - Sparking gap
HSTR - High strength thermal resistance
IACS - International Annealed Copper Standard
IP - Peak current
MRR - Material removal rate
OFF - Pulse interval
ON - Pulse duration
Ra - Surface roughness
Rav - Overall surface roughness
RSM - Response surface method
xiii
Rx - Surface roughness at x-direction
Ry - Surface roughness at y-direction
SF - Server voltage
t - Machining time in second
V - Voltage
WEDM - Wire electrical discharge machining
WT - Wire tension
xiv
LIST OF APPENDICES
NO. TITLE PAGE
A1 Schedule for Master project part I (Semester 1 – 2009/2010) 100
A2 Schedule for Master project part I (Semester 1 – 2009/2010) 101
B1 Summary of finding related to EDM performance 103
C1 Experimental results of sparking gap (top surface) 106
C2 Experimental results of sparking gap (bottom surface) 107
D1 Confirmation experimental results of sparking gap (top surface) 109
D2 Confirmation experimental results of sparking gap (bottom surface) 109
1
CHAPTER 1
INTRODUCTION
1.1 Project Background and Rationale
Electrical discharge machining (EDM) is a non-traditional concept of
machining which has been widely used to produce dies and molds. This technique
has been developed in the late 1940s and has been one of the fast growing methods
in manufacturing during 1980s and 1990s [1].
This non-traditional machining method is commonly used for very hard metals
that would be impossible to machine with traditional techniques. It has been
extensively used, especially for cutting intricate contours or delicate cavities that also
would be difficult to produce with a conventional machining methods or tools.
However, one critical limitation is that EDM only works with electrically conductive
materials. Metal that can be machined by using EDM include nickel-based alloy
(such as inconel), hardened tool steels and carbides.
Wire electrical discharge machining (WEDM) is introduced in the late 60’s.
The process was fairly simple, not complicated and wire choices were limited to
copper and brass only. WEDM is a thermo-electrical process in which material is
2
eroded from the workpiece by a series of discrete sparks between the workpiece and
the wire electrode (tool) separated by a thin film of dielectric fluid (deionized water)
that is continuously fed to the machining zone to flush away the eroded particles. The
movement of wire is controlled numerically to achieve the desired three-dimensional
shape and accuracy of the workpiece [2]. The degree of accuracy of workpiece
dimensions obtainable and the fine surface finishes make WEDM particularly
valuable for applications involving manufacture of stamping dies, extrusion dies and
prototype parts. Without WEDM, the fabrication of precision workpieces requires
many hours of manual grinding and polishing [3].
In recent years, the technology of wire electrical discharge machining
(WEDM) has been improved significantly to meet the requirements in various
manufacturing needs, especially in the precision mold and die industry. WEDM is
being used to machine a wide variety of miniature and micro-parts from metals,
alloys, sintered materials, cemented carbides, ceramics and silicon. This tremendous
achievement in WEDM technology has been achieved by many researchers from
some of the world leading institution and research centre, but still cannot coped with
the new materials introduced to the market.
The selection of cutting parameters for obtaining higher cutting efficiency or
accuracy in WEDM is still not fully solved, even with the most up-to-date CNC
WEDM machine. This is mainly due to the nature of the complicated stochastic
process mechanisms in wire-EDM. As a result, the relationships between the cutting
parameters and the process performance are hard to achieve accurately [4]. There is
still lack of research on WEDming of material such as nickel based super alloy
which include Inconel 718. It is widely used; mostly in aerospace and marine
applications which are classified as difficult to machine material by conventional
method due to high cutting temperature and rapid tool wear [5].
3
1.2 Research Statement
Studies on WEDM using coated wire somehow is limited and manufactures
claimed that the outstanding performance is achieved through their lab test. But the
results are not disclosed to the public and researcher for further study and
understanding. As such the machine materials information and the WEDM
parameters setting for the subjected wire are somehow limited. The only information
given/set by manufactures is commonly applicable to the common steel grades [6].
Inconel 718 is a high strength and thermal resistance (HSTR) [7] known to
play increasingly important in the aviation, space navigation and shipping industries
because of its outstanding multi-properties [8]. Broad bases of Inconel 718
knowledge are now exist due to its great acceptance in industries. However, the
parameter setting on WEDM of Inconel 718 is still lacking. The available
technological data which is based on manufacturers for in house experimentation is
helpful but insufficient.
Inconel 718 is assigned to be machined with WEDM in this project with the
attention to study the parameters setting for an optimum machining. A comparative
study will be carried out between previous study using brass wire and the proposed
study using coated wire electrode.
4
1.3 Research Objectives
The objectives of the research are:
a. To determine the significant parameters that influences the machining
responses during Wire Electro-Discharge Machining (WEDM) of Inconel 718.
b. To evaluate the performance of Electro-Discharge Machining (WEDM) on
Inconel 718 with respect to various responses such as spark gap, material
removal rate, cutting speed and surface finish.
c. To establish mathematical model for spark gap, surface finish, cutting speed
and material removal rate during WEDM of Inconel 718 alloy.
1.4 Scope of Study
The scope of the research consists of:
a. Wire Electro-Discharge Machining, (WEDM) linear motor 5-axis – Sodick
series AQ537L will be employed.
b. Nickel based superalloy, Inconel 718 will be used as the workpiece material.
c. Zinc coated brass wire of diameter 0.25mm will be used as electrode.
d. Parameters to be studied include voltage, peak current, pulse duration and
interval time.
e. Response variable to be study are surface finish, spark gap, material removal
rate and cutting speed.
f. The DOE and analysis of variance (ANOVA) will be processed using Design
Expert software version 7.0.0.
5
1.5 Expected Results
The expected outcomes of this study are as follows:
a. To obtain the optimum condition for WEDM on Inconel 718 in various
parameters setting using zinc coated brass wire.
b. Establishment of mathematical models for various responses during WEDM on
Inconel 718
c. The outcome of the study can be used to assist the industrial practitioners that
involved in machining of superalloy materials such as nickel alloys and to
select the most suitable cutting parameters for machining nickel alloys
application.
d. This will help in improving the quality of Inconel products as well as
minimizing the machining cost to realize the economical potential to the
fullest.
6
CHAPTER 2
LITERATURE REVIEW
2.1 Introduction
Wire electrical discharge machining (WEDM) is a metalworking process with
the help of which a material is separated from a conductive work piece, by means of
rapid, repetitive spark discharges from a pulsating direct-current power supply with
dielectric flow between the workpiece and the tool [8].
Research in areas WEDM has
become a considerable interest due to the various advantageous offered by this
process. Among the various non-conventional machining processes, WEDM is the
most widely and successfully used method for machining difficult to machined
materials such as super alloys [9].
Considering the increasing number of high-strength, non-corrosion and wear
resistant materials such as Inconel 718, WEDM has brought many improvements in
recent years. Researchers are struggling to reveal a new method to improve WEDM
efficiency; the objectives are the same: to enhance the capability of machining
performance, to get better output product, to develop technique to machine new
materials and to have better working conditions [1]. This is due to WEDM
technologies offers no readily available standard for setting the cutting parameters
7
such as current, polarity, duty cycle, etc, [9] to achieve the desire machining
characteristics of the nickel alloys in particular Inconel 718.
The selection of cutting parameters for obtaining higher cutting efficiency or
accuracy in WEDM is still not fully solved, even with the most up-to-date WEDM
machine. This is mainly due to the nature of the complex stochastic process
mechanisms in WEDM [10].
2.2 Wire Electrical Discharge Machining (WEDM)
Wire electrical discharge machining (WEDM) has been found to be an
extremely potential electro-thermal process in the field of conductive material
machining. Owing to high process capability it is widely used in manufacturing of
cam wheels, special gears, bearing cage, various press tools, dies, and similar
intricate parts etc (Figure 2.1) [11].
Figure 2.1: Typical product by WEDM (AGIE Charmilles groups, Charmilles the
solutions when to EDM, Geneva, 2004.)
8
According to previous researcher, Sommer [12] EDM can be categorized into
two: die sink EDM and wire EDM. Pandey and Shah [13] classified EDM processes
into three main categories as shown in Figure 2.2. EDM techniques have developed
in many areas. Trends on activities carried out by researchers depend on the interest
of the researchers and the availability of the technology. In 1994, Rajurkar [14] has
indicated some future trends activities in EDM: machining advanced materials,
mirror surface finish using powder additives, ultrasonic-assisted EDM control and
automation.
Figure 2.2: Classification of EDM processes [13]
The concept of WEDM is shown in Figure 2.3. In this process, a slowly
moving wire travels along a prescribed path and removes material from the
workpiece. WEDM uses electro-thermal mechanisms to cut electrically conductive
materials. The material is removed by a series of discrete discharges between the
wire electrode and the workpiece in the presence of die-electirc fluid, which creates a
path for each discharge as the fluid becomes ionized in the gap. The area where
discharge takes place is heated to extremely high temperature, so that the surface is
melted and removed. The removed particles are flushed away by the flowing
dielectric fluids as shown in Figure 2.4. The taper can ranging from 15º for a 100mm
thick to 30º for a 400mm thick workpiece can be obtained on the cut surface material
[15].
9
The wires for WEDM are made of brass, copper, tungsten, molybdenum (0.05
– 0.3mm in diameter) which capable to achieve very small corner radii. Zinc or brass
coated wires are also used extensively in this process. The wire used in WEDM
process should posse’s high tensile strength and good electrical conductivity.
Figure 2.3: Wire electrical discharge machining (WEDM) process [16].
10
Figure 2.4: Schematic of the thermal removal process of WEDM (spark between the
electrode and workpiece perform the material removal) [17].
WEDM process is usually used in conjunction with CNC and will only work
when a part is to be cut completely through. The melting temperature of the parts to
be machined is an important parameter for this process rather than strength or
hardness. The surface quality and MRR of the machined surface by wire EDM will
depend on different machining parameters such as applied peak current, and wire
materials. WEDM process is commonly conducted on submerged condition in a tank
fully filled with dielectric fluid; nevertheless it also can be conducted in dry
condition. This method is used due to temperature stabilization and efficient flushing
in cases where the workpiece has varying thickness [18].
Although both conditions (submerged or dry machining) can be performed,
most important is to produce a good quality of machined surface and dimensional
accuracy. The main goals of WEDM manufacturer and users are to achieve a better
stability and higher productivity of the WEDM process. As newer and more exotic
materials are developed, and more complex shapes are required, conventional
machining operation will continue to reach their limitations and the increased use of
WEDM in manufacturing will continue to grow at an accelerated rated [19].
However, due to a large number of variables in WEDM, it is difficult to achieve the
optimal performance of WEDM processes [20] and the effective way of solving this
11
problem is to establish the relationship between the performance measures of the
process and its controllable input parameters.
Ho and Newman [6] have classified research areas in EDM machining process
as shown in the Figure 2.5. Investigation into the influences of machining input
parameters on the performance of WEDM have been widely reported [1, 4, 8, 10].
Several attempts have been made to develop mathematical model of the process [4,
11]
Figure 2.5: Classification of major EDM research areas [6]
In this project, focusing is more on improving the performance measures
including MRR, spark gap and surface finish. These responses are mainly depend on
the discharge energy, electrical pulse parameters and discharge distribution in space
and time and flushing condition [21]. Scot et al. [22] found that current, pulse
duration and pulse frequency were the main significant control factors for both the
MRR and surface finish. Whereby, wire speed, wire tension, and dielectric flow were
relatively significant. In addition, Ahmet Hascalyk and Ulas Caydas [23] reported
that, surface roughness primarily depend on pulse duration, open circuit voltages and
dielectric fluid pressure and wire speed not seeming to have much influence.
However, WEDM also have some limitation associated during machining
processes including the use of electrically conductive material only, also, material
removal rates are very low as compared to other processes and the work surface layer
12
is damaged after processing with this technique [6]. In addition, the selection of the
appropriate parameters is also difficult and relies heavily on the operator’s
experienced and machining parameters tables provided by the WEDM machine
builder [24].
2.3 WEDM Machining Parameters
According to Wang and Yan [25], EDM parameters consist of two functional
group:
a. Electrical Parameters: polarity, peak current, pulse duration, power supply
voltage.
b. Non-electrical Parameters: rotational of speed electrode, injection flushing
pressure.
Van Tri [26] categorized the EDM parameters into five groups:
a. Dielectric fluid - type of dielectric, temperature, pressure, flushing system.
b. Machine characteristics - servo system and stability stiffness, thermal stability
and accuracy.
c. Tool - material, shape, accuracy
d. Workpiece.
e. Adjustable parameters - discharge current, gap voltage, pulse duration,
polarity, charge frequency, capacitance and tool materials.
Other studies have been carried out in order to determine the significant
WEDM machining parameters that affect the performance of the WEDM processes.
According to Mas Ayu [27], WEDM machining parameters had more significant
effect than the electrical parameters. Her finding concluded that the most significant
WEDM machining parameters are pulse duration, voltage, peak current and interval
13
time. Liao et al. [24] proposed the significant factors affecting the machining
performance are spark frequency, average gap voltage and ratio of normal sparks to
total sparks. Nihat Tosun et al. [28] described the highly effective parameters on both
kerf and MRR were found as open circuit voltage and pulse duration whereas wire
speed and dielectric flushing pressure were less effective factors. Previous
researchers findings indicate that the electrical parameters are more significant than
non-electrical parameters on the machining characteristic.
Several attempts also have been carried out by many researchers to investigate
the effect of non-electrical parameters on WEDM machining characteristic. Erden
[29] reported that dielectric flushing affected the EDM performance due to the
changing of erosion rate, mirror like finishing achieved by multi divided electrode
method. Kinoshita et al. [30] proved dielectric pressure greatly influence the WEDM
parameters during recuts.
Furthermore, trend on activities carried out by researchers depends on the
interest of the researchers and the availability of the technology. Rajurkar [14] has
indicated some future trends activities in EDM: machining advanced materials,
mirror surface finish using powder additive, ultrasonic-assisted EDM, control and
automation.
2.3.1 Pulse duration (On-Time)
During WEDM all the work is done during pulse duration (On time). The
erosion rates are affected mainly by pulse parameter. The spark gap is bridged,
current is generated and the work is accomplished. The longer the spark is sustained,
the higher is the material removal. Consequently the resulting craters will be broader
and deeper; therefore the surface finish will be rougher. Obviously with shorter
duration of sparks the surface finish will be better.
14
2.3.2 Pulse interval (Off-Time)
While most of the machining takes place during on time of the pulse, the off
time during which the pulse rests and the re-ionization of the die-electric takes place,
can affect the speed of the operation in a large way. Longer is the off time greater
will be the machining time. But this is an integral part of the EDM process and must
exist. The off time also governs the stability of the process. An insufficient off time
can lead to erratic cycling and retraction of the advancing servo, slowing down the
operation cycle. In addition, the interval time also provides the time to clear the
disintegrated particles from the gap between the electrode and workpiece for efficient
cut removal [1]. Too short pulse interval will increase the relative wear ratio and will
increase the surface roughness of the machine surface [31].
2.3.3 Servo voltage
The preset voltage determines the width of the spark gap between the leading
edge of the electrode and the work piece. High voltage settings increase the gap and
hence the flushing and machining. Some material may be necessary for high open-
open voltage due to high electrical resistance and high discharge voltage [1, 27].
2.3.4 Peak current
Peak current is also another important primary input of WEDM process. The
stronger the discharge current, MRR, overcut and surface roughness will increase. In
other hand, decrease the rate of electrode wear [31]. To minimize the electrode wear
15
and keep the current density between the tolerance limit it is necessary to select an
appropriate value of current [18, 28].
2.4 Machining Characteristic
WEDM performance is mainly measured by the material removal rate (MRR),
spark gap (kerf) and surface roughness of the workpiece. This three machining
characteristics have been identified by the previous researchers as the most
significant machining criteria that can influence the WEDM performance [22 - 24].
Determining the optimum machining parameters of WEDM to machine certain
material is very crucial. Thicker oxide layer formed due to thermal oxidation during
WEDM process is expected [31] and can reflect the surface finish of the machined
surface.
Any machined surface during machining processes will experiences a disturbed
layer which has different characteristics from those of the base metal. Surface
integrity entails the study and control of surface topography, as well as surface
metallurgy [9]. O.A. Abu Zeid [32] claimed, that thermal nature of the WEDM
process always produce a recast and underlying heat-effected zone on the surface
being machined and develops a residual stress that often causes micro cracks.
Thermal sensitivity or chemical complexity of the material can also affect the surface
integrity [23]. Several other studies also have been carried out to determine the
appropriate EDM machining parameters combination from the aspect of surface
integrity. The surface crack formation for AISI D2 and H13 has been studied by H.T.
Lee & T.Y. Tai [33]. It was reported that crack formation and white layer thickness
are related to the EDM parameters.
Ahmet et. al, [23] concluded surface integrity can be divided into two
important categories; surface texture, which concern principally on the surface
roughness and surface metallurgy, which concern to the nature of the surface layer
16
produced during machining. Hence, the selection of the machining parameters
including pulse-on time, pulse off-time, table feed rate, flushing pressure, wire
tension, wire velocity, etc should be chosen properly according to the workpiece
properties so that better performance can be obtained [23]. However based on the
previous findings, the significant selection of the appropriate machining responses
for cutting Inconel 718 using WEDM for this project is surface roughness (Ra), spark
gap (Gap) and material removal rate (MRR).
2.4.1 Effect on surface finish, Ra
Surface topography or surface finish, also known as surface texture are terms
used to describe the general quality of machined surface, which is concerned with the
geometric irregularities and the quality of a surface [9]. The quality of a machined
surface is becoming more and more important to satisfy the increasing demands of
sophisticated component performance, longevity, and reliability.
Fine surface finish is obtained by a combination of the proper electrode
material, good flushing conditions, and the proper power supply settings. High
frequency, low power and orbiting produce the best finish, as these conditions
produce smaller, less defined craters in the work metal [8, 11, 27, 28]. Pandey and
Shah [13] have found that surface finish to be inversely proportional to the frequency
of discharge. Assuming that each spark leads to a spherical crater formation on the
surface of workpiece, the volume of metal removed per crater will be proportional to
the cube of the crater depth. Zhang et al. [34] has investigated the effect of discharge
voltage, discharge current and pulse duration on AISI 1045 steel as workpiece. The
investigation revealed that surface finish increase with an increase of this factor.
In 2005, M. S. Hewidy et al. [4] has investigated the WEDM performance on
Inconel 601. This work has been established based on the response surface
methodology (RSM). They have confirmed surface roughness increase with the
17
increase of peak current and decrease with the increase of duty factor and wire
tension. Many researchers concluded that the ideal surface finishes are rare to happen
due various factors. Zhang et al. [34] developed a theoretical model to estimate the
surface roughness. Investigation have been carried out using AISI 1045 steel as work
piece material and copper as the electrode. Results showed that the roughness of
finished surface increases with an increase in the discharge voltage, discharge current
and pulse duration. Guo et al. [39] also concluded that with ultrasonic aid the cutting
efficiency of WEDM can improve the surface finish quality.
2.4.2 Effect on material removal rate, MRR
The removal of material in electrical discharge machining is based upon the
erosion effect of electric sparks occurring between two electrodes. Several theories
have been forwarded in attempts to explain the complex phenomenon of "erosive
spark". The following are the theories:
a. Electro-mechanical theory
b. Thermo-mechanical theory
c. Thermo-electric theory
Electro-mechanical theory suggests that abrasion of material particles takes
place as a result of the concentrated electric field. The theory proposes that the
electric field separates the material particles of the workpiece as it exceeds the forces
of cohesion in the lattice of the material. This theory neglects any thermal effects.
Experimental evidence lacks supports for this theory.
Whereby, thermo-mechanical theory suggests that material removal in EDM
operations is attributed to the melting of material caused by "flame jets". These so -
called flame jets are formed as a result of various electrical effects of the discharge.
18
However, this theory does not agree with experimental data and fails to give a
reasonable explanation of the effect of spark erosion.
Thermo-electric theory is best-supported by experimental evidence, suggests
that metal removal in EDM operations takes place as a result of the generation of
extremely high temperature generated by the high intensity of the discharge current.
Although well supported, this theory cannot be considered as definite and complete
because of difficulties in interpretation.
Material removal rate is proved by the previous researchers as one of the most
important output parameters, which decide the performance of WEDM machining
processes [1, 2, 4, 9, 11]. Rival [9] have discovered factors such as current, voltage,
pulse on time and interval time which to have significant effect on MRR and EWR.
Most researchers also concluded EDM electrical parameters such as polarity, peak
current, pulse duration and power supply voltage are highly influence the MRR for
performance of EDM processes [20, 21, 24, 27].
In 1991, Kunieda et al. [40] has revealed a new method to improve EDM
efficiency by supplying oxygen gas into this gap. This new method results shows,
material removal rate is increased due to the enlarged volume of discharged crater
and more frequent occurrence of discharge. Powder additive method also has been
carried out by previous researcher in order to improve the EDM efficiency. Jeswani
[41] revealed that the addition of about 4 g/l of fine graphite powder in kerosene
increases MRR by 60% and tool wear by 15%.
19
2.4.3 Effect on spark gap, Gap
The workpiece and wire electrode represent positive and negative terminal DC
electric circuit and separated by a controlled gap which constantly controlled by the
machine [31]. This gap is fulfill with dielectric fluid which act as insulator, cooling
as well as flushing agent in order to flushed away the eroded particles from the
cutting zone (Figure 2.6).
Figure 2.6: Definition of kerf and overcut in WEDM [42]
Spark gap is the most crucial parts of the EDM system. The size of the gap is
governed by the servo control system whose motion is controlled by gap width
sensors. They control the motion of the ram head or the quill which in turn governs
the gap size. Spark gaps in WEDM make the kerf larger than the wire diameter as
shown in Figure 2.6. This overcut is in the range of 0.020 – 0.050 mm [42]. The most
common sparking gap is 0.03 mm. Once cutting condition has been establish for a
given cut, the overcut remains fairly constant and predictable.
In this project, water is used as the dielectric. During WEDM processes,
sparking occurs between the side and machine surface of the workpiece. The
sparking area consists only the front of the electrode diameter (180º) as it progress
into the cut while the clearance is equal to the spark length of the wire electrode [31].
Sparks are formed through a sequence of rapid electrical pulse, generated by the
20
WEDM machine power supply thousands of times per second. Each spark forms an
ionization channel under extremely high heat and pressure resulting in vaporization
of localized sections. The vaporized metallic debris created by this process, from
both the workpiece and wire material, is subsequently quenched and flushed away by
the flow of dielectric fluid through the gap [43].
Literature studies showed there is fewer researchers investigated the correlation
between machining parameters and spark gap in the WEDM process. Nihat Tosun et
al. [28] investigated the correlation between the machining parameters and spark gap
as a factor in determining the WEDM performance. The results concluded that open
circuit voltage and pulse duration have the significant impact to both MRR and kerf
width. Whereby, wire speed and dielectric pressure were less effective factors.
Appendix B shows the summary of the researches that have been done in evaluating
the WEDM machine performance.
2.5 Wire Electrode
Previous researchers M.S. Hewidy et al. [4]; S. Sarkar et al. [36] and R.
Ramakrisnan et al. [38] used brass wire as the electrode to WEDM the workpiece.
Brass wire is widely used in WEDM processes due to its good machining properties
and can be die casted or extruded for specialized application. It possesses high tensile
strength, high electrical conductivity, and good wire drawability to close tolerances.
Other researchers such as S.S Mahapatra and Amar Parnaik [35]; Kuang Yuan
and Ko Ta Chiang [37] used coated wire electrode to investigate WEDM machining
performance. Coated brass wire can perform at higher cutting speed compared to
brass wire electrode. Coated brass wire also can produce exceptional surface finish
especially when WEDM tungsten carbide and often utilized for cutting PCD and
graphite.
21
The ideal wire electrode material for this process has three important criteria:
a. High electrical conductivity.
b. Sufficient mechanical strength.
c. Optimum sparks and flushes characteristics.
As discussed above, there is no “perfect” wire that excels in every criteria, and
some compromises become necessary, depending upon the desired results and
application. And all the three factors are very closely related and interdependent.
2.5.1 Copper wire
Copper was the original material first used in WEDM. It is an excellent
conductor with 100 IACS (International Annealed Copper Standard) value [31].
Although its conductivity rating is excellent, its low tensile strength, high melting point
and low vapor pressure rating severely limited its potential. Under the electro-thermal
condition, predominant during WEDM, copper wire wears rapidly and its tension ability
is rather poor, resulting, therefore, in machining instabilities, due to high degree of short
circuits, especially in the machining of small curvature [27].
2.5.2 Brass wire
Brass was the first logical alternative to copper when early EDM researchers
were looking for better performance. Brass EDM wire is a combination of copper
and zinc, typically alloyed in the range of 63–65% Cu and 35–37% Zn [43].
22
The addition of zinc element provides significantly higher tensile strength,
lower melting point and higher vapor pressure rating, which more than offsets the
relative losses in conductivity. Brass quickly became the most widely used electrode
material for general purpose wire EDM. It is now commercially available in a wide
range of tensile strengths and hardness.
2.5.3 Coated wire
Coated wire is commonly employed in WEDM process to increase
substantially the cutting speed and cutting precision. Since brass wires cannot be
efficiently fabricated with any higher concentration of zinc, the logical next step was
the development of coated wires, sometimes called plated or “stratified” wire. They
typically have a core of brass or copper, for conductivity and tensile strength, and are
electroplated with a coating of pure or diffused zinc for enhanced spark formation
and flush characteristics.
Originally called “speed wire” due to their ability to cut at significantly higher
metal removal rates [2], coated wires are now available in a wide variety of core
materials, coating materials, coating depths and tensile strengths, to suit various
applications and machine requirements. Although more expensive than brass, coated
wires currently represent the optimum choice for top all-around performance, and
their relative economics are covered in a later section. Table 2.1 indicates the
designation of these coated wire and their applications [31].
23
Table 2.1: The designation of various types of coated wire and their applications
Types, Ø used /
Charmilles Designation Basic Material Applications
1. Zinc coated brass: Half
hard zinc coated brass
(Ø 0.20 – 0.25mm),
SS20-SS25.
CuZn37
Characteristic and application
similar to those of half hard
brass wire
2. Diffused zinc coated
copper (Ø 0.25 –
0.30mm), XS25-XS30
Cu
Better usage for cylindrical cut,
1 roughness and 2 finishing
passes
3. Brass coated with
special alloy (Ø 0.30 -
0.07mm)
Special alloy
Allows high wire tensions
making it possible to produce
high precision punches and
dies with fines detail.
Based on the literature and considering all the important criteria, zinc coated
brass wire will be employed to machine Inconel 718 in this investigation. Zinc
coated brass wire was one of the first attempts to present more zinc to the wire’s
cutting surface. This wire consists of a thin (approximately 5 µm) zinc coating over a
core which is one of the standard EDM brass alloys [43]. This wire offers a
significant increase in cutting speed over plain brass wires, without any sacrifice in
any of the other critical properties. Zinc coated brass wires produce exceptional
surface finishes when cutting tungsten carbide and are often utilized for cutting PCD
and graphite. These wires are also utilized in those circumstances in which brass
wires produce unacceptable brass plating on the workpiece.
Due to the low electrical conductivity of Inconel 718, zinc coated brass wire is
suitable choice due to its high electrical conductivity. IACS (International Annealed
Copper Standard) number is one of the methods to identify the types of wire in
accordance to its conductivity percentage. Copper has known to be excellent
conductor with 100 IACS value. Brass alloy 63%Cu + 37% Zn = (brass) has 29
24
IACS value, while molybdenum wire has 34 IACS value. As for these projects zinc
coated with copper or brass wire was chosen as it has 84 IACS value.
Zinc coated high conductivity copper alloy offers a number of superiority; high
temperature toughness, high current efficiency and high discharge performance make
it the best electrode wire for high speed, high precision fine machining. The core is
made of copper-alloy, therefore good in workability and superior straightness which
is optimum for automatic wire connection.
2.6 Nickel Based Super Alloy and Their Machinibility
The number of superalloys that have been developed and used over the years is
large. In reality, the solid solution strengthened alloys are strengthened both by solid
solution hardening and by the presence of carbides, while the precipitation hardened
alloys are strengthened by the combination of precipitates, solid solution hardening
and the presence of carbides.
Nickel based superalloys are the most complex of the superalloys and are used
in the hottest parts of aircraft engines, constituting over 50% of the engine weight.
They are either solid solution hardened for lower temperature use or precipitation
hardened for higher temperature use. The nickel based alloys contain at least 50%
nickel and are characterized by the high phase stability of the FCC austenitic (γ)
matrix. Many nickel based alloys contain 10–20% chromium, up to about 8%
aluminum and titanium combined 5–15% cobalt, and small amounts of boron,
zirconium, hafnium and carbon. Other common alloying additions are molybdenum,
niobium, tantalum, tungsten and rhenium. Chromium and aluminum are important in
providing oxidation resistance by forming the oxides Cr2O3 and Al2O3 respectively
[46].
25
The most commercially superalloy is Inconel 718, listed as an iron–nickel
based alloy even though it contains more nickel than iron. This classification fits
with the traditional classification for this alloy, although many newer works list it as
a nickel based alloy. Also note that for the cobalt based alloys, they are none listed as
being precipitation hardened, because unfortunately, these alloys do not precipitation
harden like the nickel and iron–nickel alloys.
Nickel based superalloy namely; Inconel 718 is difficult to machine material
[7], perhaps second only to titanium in machining difficulty, although those who
machine Inconel 718 would probably maintain that these superalloys are the most
difficult to machine material. Many of the same characteristics that make Inconel 718
good high temperature materials also make them difficult to machine, namely
[46,47]:
a. Retention of high strength levels at elevated temperature
b. Rapid work hardening during machining
c. Presence of hard abrasive carbide particles
d. Generally low thermal conductivities and
e. Tendency of chips to weld to cutting edges and form built-up edges.
Due to their high temperature strength, Inconel 718 remains hard and stiff at
the cutting temperature, resulting in high cutting forces that promote chipping or
deformation of the tool cutting edge. In addition, since superalloys retain a large
percentage of their strength at elevated temperatures, more heat is generated in the
shear zone resulting in greater tool wear than with most metals. Since the forces
required to cut superalloys are about twice those required for alloy steels, tool
geometry, tool strength, and rigidity are all important variables.
Their low thermal conductivities cause high temperatures during machining.
The combination of high strength, toughness, and ductility impairs chip
segmentation, while the presence of abrasive carbide particles accelerates tool wear.
Inconel 718 also have a tendency to rapidly work harden which can create a
26
hardened surface layer that degrades the surface integrity and can lead to lower
fatigue life. General guidelines for machining superalloys are very similar to those
for titanium alloys [48].
2.6.1 Inconel 718 physical and mechanical properties
The workpiece chosen for this study was Inconel 718. It is a precipitation-
hardenable nickel-chromium alloy containing significant amounts of iron, niobium,
and molybdenum along with lesser amount of aluminum and titanium (Table 2.2). Its
combines corrosion resistance and high strength with outstanding weldability,
including resistance to postweld cracking [46].
The alloy has creep-rupture strength at high temperatures up to 700ºC. Used in
gas turbines, rocket motors, spacecraft, nuclear reactors, pumps and tooling [47]. The
properties of Inconel 718 are listed in the Tables 2.3 and 2.4.
Table 2.2: Chemical properties of Inconel 718 (%)
Element Percentage (%) Ni (+Co) 50 - 55
Fe Bal Mo 2.3 – 3.3 Ti 0.65 – 1.15 C 0.08 Si 0.35 Cu 0.3 Cr 17 – 21 Co 1
Nb (+Ta) 4.75 – 5.5 Al 0.2 – 0.8 Mn 0.35 B 0.006
27
Table 2.3: Physical properties of Inconel 718
Table 2.4: Mechanical properties of Inconel 718
2.7 Design of Experiment (DOE)
According to Lochner, R.H., and Matar, J.E. [49], design of experiment (DOE)
is series of tests in which purposeful changes are made to the input variables of a
process or system so that the reasons for change in the output responses can be
observed and identified.
There are several reasons for designing complete factorial experiments rather
than, for example, using a series of experiments investigating one factor at a time.
The first is that factorial experiments are much more efficient for estimating main
effects, which are the averaged effects of a single factor over all units. The second
and very important reason is that interaction among factors can be assessed in a
factorial experiment but not from series of one-at-a-time experiment. Interaction
Physical Properties Value Density 8.19 g/cm3 Melting Point Range 1260 - 1336ºC Specific Heat 435 J/kg.K Average Coefficient of Thermal Expansion 13.0 µm/m.K
Thermal Conductivity 11.4 W/m.K Electrical Resistivity 1250 n.m Curie Temperature -112ºC
Mechanical Properties (Room Temperature) Value
Ultimate Tensile Strength 1240 MPa Yield Strength 1036 MPa Eleongation (in 50mm) 12% Elastic Modulus 211 Gpa
28
effects are important in determining how the conclusions of the experiment might
apply more generally. Complete factorial systems are often large, especially if an
appreciable number of factors are to be tested. Often an initial experiment will set
each factor at just two levels, so that important main effects and interactions can be
quickly identified and explored further.
The choice of factors and the choice of levels for each factor are crucial aspects
of the design of any factorial experiment, and will be dictated by the subject matter
knowledge and constraints of time or cost on the experiment. The levels of factors
can be qualitative or quantitative. The range of values for quantitative factor must be
decide on how they are going to be measured and the level at which they will be
control during the trials. Meanwhile, the quantitative factor is parameters that will be
determine discretely [50].
2.7.1 Two-level full factorial design
Experiments with large numbers of factors are often used as a screening
device to assess quickly important main effects and interactions. For this, it is
common to set each factor at just two levels, aiming to keep the size of the
experiment manageable. The levels of each factor are conventionally called low and
high, or absent and present [50]. One of the advantages when implemented full
factorial design is that it offers the capability to estimate the correlation between two
or more factors in one time, where it is possible with other quality tool. Furthermore,
this tool also capable to identify the importance factor in the experiment under a
wide range of condition without sacrifices any factors.
Two-factor experiment is the simplest type of factorial design in DOE, in
which effects of two factors on one or more response variables are tested
simultaneously. It is common to use two levels for each factor studied, where k, is
the number of factors and 2 indicates the level of experiment, then the total number
29
of combination will be 2k
. In this experiment, four-factor experiments design was
employed with two level of full factorial design experiment. Experiment shall
include all the possible combination factors at two levels (low and high value of the
machining parameters). The arrangement of the entire factor will be based on Design
Expert version 7.0.0 software. This program will randomly and automatically
analyze all the combination of the possible combination for the experiment. This
software also can automatically analyze all the experimental results in order to
investigate the influence of the machining parameters to the machining outputs or
responses.
30
CHAPTER 3
METHODOLOGY
3.1 Introduction
The purpose of this project is to evaluate the performance of WEDM on Inconel
718. To achieve this objective proper experimental plan is necessary to achieve good
results. This experiment consist of four main elements namely, research design and data
analysis, variables, research procedure and instrumentation. Figure 3.1 shows the four
elements involved in research methodology in achieving the objective of this
experiment.
Design of experiment with full factorial using Design Expert version 7.0.0
software was applied as a tool for design of experiment and data analysis. The
confirmation test was also implemented in order to give the reliability of the WEDM
results for Inconel 718.
31
Figure 3.1: Research methodology
3.2 Research Design and Data Analysis
Full factorial DOE will be employed for the whole design and analysis. DOE
includes determining controllable factors and the levels to be investigated. In this
experiment, four-factor experiment design will be use with two levels of full factorial
design experiment. Based on this, the total number of experiments (combinations)
required was 16 (24
) experiment.
This experiment design includes all the possible combination factors at two levels
(low and high value) for each parameter. Table 3.1 showed the notation used to denoted
these levels; plus (+) for high value and minus (-) for low value. The arrangement of the
factors for this project was based on Design Expert version 7.0.0 software. This program
randomly choose the combination of factors to run the experiment and automatically
analyse all the experimental results to investigate the influence of WEDM machining
parameter on the surface integrity of Inconel 718.
32
Table 3.1: Two-level full factorial experiment with four factors.
Exp. No.
Factor Servo Voltage
(SV)
(V)
Peak Current (IP)
(A)
Pulse Duration (ON) (µs)
(µs)
Pulse Interval (OFF)
(µs)
1. - - - - 2. + - - - 3. - + - - 4. + + - - 5. - - + - 6. + - + - 7. - + + - 8. + + + - 9. - - - + 10. + - - + 11. - + - + 12. + + - + 13. - - + + 14. + - + + 15. - + + + 16. + + + + 17. cp Cp cp cp 18. cp Cp cp cp 19. cp Cp cp cp 20. cp Cp cp cp
Notes: + (high value), - (low value), cp (centre point).
33
3.3 Research Design Variable
The design variables are described into two main groups, which are response
parameters and machining parameters. Response parameters (machining characteristic
@ dependent variables) include:
a. Surface roughness, Ra
b. Sparking gap, Gap
c. Material removal rate, MRR
d. Cutting speed, CS
Machining parameters or also known as independent variables involves in this
experiment:
a. Pulse duration (ON, µs)
b. Pulse interval (OFF, µs)
c. Peak current (IP, ampere)
d. Servo voltage (SV, V)
Note: these parameters were donated as ON, OFF, IP, SV respectively.
3.3.1 Machining parameters
Based on previous studies, several numbers of machining factors have been used
in WEDM operation. As mentioned earlier, electrical parameters are the factor that
significantly influences the machining characteristic whereby, non-electrical parameters
have less significant to the machining characteristic [25 - 27].
34
Most researchers identified four WEDM cutting parameters that greatly affect
the machining output; ON, OFF, IP, SV which are known as electrical factors while non-
electrical factor (mechanical factor) include wire tension, wire speed, dielectric pressure
etc are held constant. Table 3.2 shows the setting of the parameters studies. All the
suggested value in Table 3.2 was based on previous studies. The actual values of these
setting are shown in Table 3.3
Table 3.2: The design of machining parameters
Machining Parameters
Level
1 (Low) 2 (High)
Code Value
True Value
Code Value
True Value
Pulse Duration ON (µs) 001 0.65 003 0.75
Pulse Interval OFF (µs) 007 4.0 015 8.0
Peak Current IP (Ampere)
2210 8.0 2215 12.0
Servo Voltage (Volt) 030 30.0 060 60.0
35
Table 3.3: Actual value of experiment design
Machining Voltage : 80V
Wire Speed : 10 m/min
Wire Tension : 800 g
Injection Pressure : 12 bar
SV IP ON OFF
Exp. No. Servo Voltage
(V)
Peak Current
(A)
Pulse Duration
(µs)
Pulse Interval
(µs)
1. 30 8 0.65 4
2. 60 8 0.65 4
3. 30 12 0.65 4
4. 60 12 0.65 4
5. 30 8 0.75 4
6. 60 8 0.75 4
7. 30 12 0.75 4
8. 60 12 0.75 4
9. 30 8 0.65 8
10. 60 8 0.65 8
11. 30 12 0.65 8
12. 60 12 0.65 8
13. 30 8 0.75 8
14. 60 8 0.75 8
15. 30 12 0.75 8
16. 60 12 0.75 8
17. 45 10 0.70 6
18. 45 10 0.70 6
19. 45 10 0.70 6
20. 45 10 0.70 6
36
3.3.2 Machining characteristic
This study investigates the machining characteristics such as surface roughness
(Ra), spark gap (Gap), material removal rate (MRR) and cutting speed (CS). These are
the most common key indicators used by many manufacturers to determine the quality
machine surface through surface roughness, while spark gap is a reflection of degree of
accuracy the WEDM machining can achieved and material removal rate with the input
of cutting speed is the key indicator of the efficiency of the WEDMing process [31].
3.3.3 Surface roughness, Ra
Surface topography or surface roughness, also known as surface texture are terms
used to express the general quality of a machined surface, which is concerned with the
geometric irregularities and the quality of a surface [9]. According to Armarego and
Brown [44], ideal surface roughness may be specified in a variety of ways, but two
common methods are the peak to valley height (h) and the arithmetic average, Ra (μm).
The Ra value, also known as centre line average (CLA) and arithmetic average
(AA) is obtained by averaging the height of the surface above and below the centre line.
The Ra will be measured using a surface roughness tester from Mitutoyo, Model:
Formtracer CS-5000. Before conducting the measurement, all the samples were cleaned
with acetone. The Ra values of the WEDMed surface were obtained by averaging the
surface roughness values of 5mm measurement length.
37
3.3.4 Sparking gap, Gap
Sparking gap (SG) or also known as overcut is one of the responses investigate in
this study. Sparking gap is measure using optical microscope in order to study the
correlation between machining parameters and the spark gap. The unit used is mm.
Sparking gap (Gap) can be calculated by the following formula:
3.3.5 Cutting speed, CS
Cutting speed (CS) is measured after WEDMing 10mm distance and recorded
using the WEDM machine controller.
3.3.6 Material Removal Rate, MRR
The material removal rate (MRR) of the workpiece is the volume of the material
removal per minute. As for these the following equation is used to determine the
material removal rate (MRR) value:
Spark Gap =(kerf width − wire diameter)
2
38
Volume = Spark Gap (mm) x Machine Distance (10mm) x Workpiece
Thickness (25mm)
By knowing the density of Inconel 718 was 0.00819 g/mm3
.The mass of material that
removed by the WEDM process:
Mass = Density (g/mm3 ) x Volume (mm3
)
Therefore, MRR is then measured by:
3.4 Research Procedure
Based on previous studies, the following four machining parameters i.e. pulse on-
time, pulse off-time, peak current and servo voltage were chosen as the input
parameters. WEDM performance on Inconel 718 is measured by four important
response parameters such as surface finish (Ra), sparking gap (Gap), material removal
rate (MRR) and cutting speed (CS).
Full factorial design was employed with two level of full factorial design
experiment. The total number of experiment (combinations) required is 24 = 16
experiments with additional of 4 centre point. The experiment design will include all the
combination factors at two levels. The arrangement of the factor for this project will be
MRR =Mass (g)
Machining Time (min)
39
based on design expert version 7.0.0 software. Table 3.4 shows the flowchart of the step
involves in the overall experiment.
40
To verify the experimental results
START
Design Plan (Full Factorial Design)
Run experiment according to design plan
Experiment resulted obtained
Performed analysis based on full factorial design
Determine the optimum machining condition for the work material
Confirmation test
Final results recorded
END
20 experiment including 4 centre points
Four responses: surface finish, spark gap, MRR & cutting speed
ANOVA
Figure 3.2: Flowchart of experiment stepping
41
3.5 Experimental Set-up
The experiments were performed on a WEDM linear motor 5 axes – Sodick series
AQ537L machine. Zinc coated brass wire with diameter of 0.25mm was chosen as
electrode in machining Inconel 718. Table 3.3 indicates the list of machining parameters
that involve in this experiment. These parameters were kept constant throughout the
experiment trials.
Table 3.4: Machining parameters set-up (constant parameters)
Parameter Setting Value
Main Power Supply Voltage, V (Volt) 80
Servo Speed, SF (mm/mmin) At no load) – normal servo control
Wire Tension, WT (g) 800
Wire Speed, WS (m/min) 10
Flushing Pressure, FP (bar) 12
Wire Electrode Zinc coated brass wire, Ø 0.25mm
Polarity Workpiece : Negative
Wire Electrode : Positive
Dielectric Fluid Submerge deionizer water.
The size of Inconel 718 was a rectangular bar with dimension of 48mm x 44mm x
25mm. The workpice material was cut to size for 10mm length with a 3mm gap between
two cutting experiments. Permanent marker will be used to mark the cut out workpiece
to indicate the orientation accordingly to the parameter. Usage of scriber to mark the
workpiece need to be avoided to preventing coated being snapped when passes the
scriber mark. The wire is so sensitive that any irregularities on the machine surface can
cause it to break.
42
To measure the surface roughness, 5.0mm will be cut from the workpiece and the
remaining is left to the remaining workpiece as a backup sample and a tool to measure
the spark gap.
The spark gap was measure using Zeiss Axiotech High Power Optical Microscope
with 100x magnification. All the measurement of spark gap was measured before the
original workpiece was cut into smaller size of 3mm x 5mm x 25mm. This is required in
order to facilitate post measurements on other equipment such as SEM and surface
roughness tester.
After the machining trials were completed, the machined workpiece were cut out
into smaller specimen’s perpendicular to the cutting surface with the same machine. This
is to reveal the section of machined surface layer for measuring the surface roughness.
The machine surface roughness was the measured using Mitutoyo surface roughness
machine. In accordance to the research made on measuring the surface roughness,
Mustafa et al. [45] reported that the measurement must consider x & y direction
perpendicular to the lay direction. Horizontal direction (x-direction) is expected to be
more crucial than the averages surface roughness along the vertical direction (y-
direction) due to the position of lay direction, the average of surface roughness along
horizontal and vertical direction is used as an indication of the total surface roughness of
each test section as shown below.
Roughness on X-Axis (Rx) = (RX1 + RX2 + RX3
)/3
Roughness on Y-Axis (Ry) = (RY1 + RY2 + RY3
)/3
Overall Roughness (RAV) = (Rx + RY
)/2
43
The method was followed since other published papers available do not provide
any specific information on the selection of machining parameters for various machining
conditions and materials.
3.6 Experimental Equipments
The equipment involved in this study are as follows:
a. WEDM machine – WEDM linear motor 5 axes – Sodick series AQ537L (Figure
3.3).
b. Measuring equipment – surface roughness was measured with the Mitutoyo
surface roughness measuring machine (Figure 3.4). Whereby, Zeiss Axiotech High
Power Optical Microscope will used to measure the spark gap (Figure 3.5).
44
Figure 3.3: WEDM linear motor 5 axes – Sodick series AQ537L
Figure 3.4: Mitutoyo surface roughness measuring machine
45
Figure 3.5: Zeiss Axiotech high power optical microscope
46
CHAPTER 4
EXPERIMENTAL RESULTS AND DATA ANALYSIS
4.1 Introduction
This chapter discusses on the experimental results on WEDM of Inconel 718 using
zinc coated brass wire of diameter 0.25mm. The main purpose of this research is to
investigate the performance of the WEDM on Inconel 718 based on predetermined
WEDM machining parameters such as servo voltage (SV), peak current (IP), Pulse
Duration (ON) and Pulse Interval (OFF).
Design Expert software version 7.0.0 was employed to analyse all the data of the
20 experiment trial runs by using ANOVA approach. The quadratic mathematical
models were proposed for the response variables such as surface roughness (Ra),
sparking gap (GAP), material removal rate (MRR) and cutting speed (CS). These
relationships between machining factors and responses were evaluated by the F-test of
ANOVA and the fit summary reveals that the fitted quadratic model is statistically
significant to be considered.
47
4.2 Experimental Results
Full factorial design of four factors with two levels each was conducted which
consist of 20 runs (including of four center points). The machine responses were surface
roughness (Ra), sparking gap (Gap), material removal rates (MRR) and cutting speed
(CS) respectively. Tables 4.1 to 4.4 show the summary of the machining responses
corresponding to the various setting of WEDM machine parameters.
Table 4.1 shows the summary of surface roughness measurement of the
experimental trials. Taylor Mitutoyo surface roughness measuring machine was used to
conduct all the surface roughness value. The measurement length of each specimen was
5mm and it was divided into 3 sections with sampling of 0.25mm each. Every section
was measured 3 times before average results of each section were obtained. Table 4.1
shows the summary of the measurement obtain for the surface roughness measurement.
48
Table 4.1: Experimental results of surface roughness
No. of trial
Rx
(µm)
Ry
(µm)
Total Average
(µm)
1. 2.13 2.09 2.11 2. 1.79 1.68 1.74 3. 2.11 1.68 1.90 4. 2.29 2.06 2.18 5. 2.64 2.54 2.59 6. 2.66 2.23 2.45 7. 2.59 2.07 2.33 8. 2.74 2.54 2.64 9. 1.96 1.91 1.94 10. 1.94 2.01 1.98 11. 2.06 1.93 2.00 12. 2.11 1.89 2.00 13. 2.62 2.57 2.60 14. 2.51 2.10 2.31 15. 3.01 2.42 2.72 16. 2.36 1.91 2.14 17. 2.19 1.73 1.96 18. 2.16 1.98 2.07 19. 2.24 2.41 2.33 20. 2.56 2.06 2.31
Table 4.2 shows the summary of kerf width and sparking gap of the experimental
runs. All measurement was taken using Zeiss Axiotech High Optical Microscope under
100 times of magnification. In order to reduce the uncertainty errors, the kerf width
measurement was taken three times at three different points along to the cutting line.
This measurement was also taken on the top and the bottom of the cutting line before the
average measurement value of kerf width was calculated. The sparking gap was
calculated using the following equation:
49
Where,
Kerf width was obtained from the measurement shown on Appendix D.
Wire diameter = 0.25mm
Table 4.2: Experimental results of sparking gap (Gap)
No. of trial
Sparking gap on top
surface (mm)
Sparking gap on bottom
surface (mm)
Total Average
(µm)
1. 0.031 0.024 0.028 2. 0.031 0.026 0.029 3. 0.032 0.026 0.029 4. 0.039 0.031 0.035 5. 0.043 0.038 0.040 6. 0.038 0.042 0.040 7. 0.039 0.036 0.037 8. 0.039 0.039 0.039 9. 0.036 0.029 0.032 10. 0.034 0.026 0.030 11. 0.035 0.028 0.031 12. 0.034 0.025 0.029 13. 0.036 0.037 0.036 14. 0.037 0.041 0.039 15. 0.042 0.039 0.041 16. 0.047 0.029 0.038 17. 0.037 0.016 0.027 18. 0.036 0.025 0.030 19. 0.032 0.026 0.029 20. 0.032 0.027 0.030
50
Other machine response that has been considered was cutting speed (CS). This
response was recorded by the WEDM machine time indicator. The distance of 10mm for
cutting distance was fixed for all experiment. Cutting speed measurement value showed
in Table 4.3 is obtained with following equation:
Where,
Machining distance, d = 10mm
Machining time, t was obtained from the WEDM machine time indicator
Table 4.3: Experimental results for cutting speed (CS)
No. of trial
Machining distance
(mm)
Machining time
(min)
Cutting speed, CS
(mm/min)
1. 10 13.65 0.733 2. 10 9.57 1.045 3. 10 12.95 0.772 4. 10 8.78 1.139 5. 10 8.85 1.130 6. 10 6.57 1.523 7. 10 9.00 1.111 8. 10 6.23 1.604 9. 10 11.93 0.838 10. 10 10.13 0.987 11. 10 11.77 0.850 12. 10 9.47 1.056 13. 10 8.28 1.207 14. 10 7.15 1.399 15. 10 8.53 1.172 16. 10 6.85 1.460 17. 10 7.18 1.392 18. 10 7.25 1.379 19. 10 7.20 1.389 20. 10 7.30 1.370
51
Meanwhile, Table 4.3 indicated the experimental results for material removal rates
(MRR) for each 20 run of experiments. MRR shown on Table 4.5 were determined by
using the following equation:
Where, Mass is defined by: Density of the Inconel 718 = 0.00819 g/mm
3
Volume is definied by:
Volume = sparking gap, SG (mm) x machining distance (10mm)
x thickness, t (25mm)
Machining time, t was recorded by the WEDM machine time indicator.
52
Table 4.5: Experimental results for material removal rate (MRR)
No. of trial
Sparking gap
(mm)
Volume
(mm3
Mass
)
(g)
Machining time
(min)
MRR
(g/mm3) 1. 0.028 7.000 0.057 13.650 0.0042 2. 0.029 7.250 0.059 9.567 0.0062 3. 0.029 7.250 0.059 12.950 0.0050 4. 0.035 8.750 0.072 8.783 0.0082 5. 0.040 10.000 0.082 8.850 0.0093 6. 0.040 10.000 0.082 6.567 0.0125 7. 0.037 9.250 0.076 9.000 0.0084 8. 0.039 9.750 0.080 6.233 0.0128 9. 0.032 8.000 0.066 11.933 0.0055 10. 0.030 7.500 0.061 10.133 0.0060 11. 0.031 7.750 0.063 11.767 0.0054 12. 0.029 7.250 0.059 9.467 0.0062 13. 0.036 9.000 0.074 8.283 0.0089 14. 0.039 9.750 0.080 7.150 0.0112 15. 0.041 10.250 0.084 8.533 0.0098 16. 0.038 9.500 0.078 6.850 0.0114 17. 0.027 6.750 0.055 7.183 0.0077 18. 0.030 7.500 0.061 7.250 0.0084 19. 0.029 7.250 0.059 7.200 0.0082 20. 0.030 7.500 0.061 7.300 0.0084
Table 4.6 present the overall results for wire electrical discharge machining on
Inconel 718 in terms of surface roughness (Ra), sparking gap (Gap), material removal
rates (MRR) and cutting speed (CS). All these machining responses were used as input
to the Design Expert version 7.0.0 software for further analysis. By using ANOVA,
information such as main effects, percentage contribution for each factor and estimation
of the optimum results can be produced and analysed.
53
Table 4.6: Overall experimental results corresponded to each run
Exp. No.
Factors Responses SV (V)
IP (A)
ON (µs)
OFF (µs)
Ra (µm)
Gap (mm)
MRR (g/min)
CS (mm/min)
1. 30 8 0.65 4 2.11 0.028 0.0042 0.733
2. 60 8 0.65 4 1.74 0.029 0.0062 1.045
3. 30 12 0.65 4 1.90 0.029 0.0050 0.772
4. 60 12 0.65 4 2.18 0.035 0.0082 1.139
5. 30 8 0.75 4 2.59 0.040 0.0093 1.130
6. 60 8 0.75 4 2.45 0.040 0.0125 1.523
7. 30 12 0.75 4 2.33 0.037 0.0084 1.111
8. 60 12 0.75 4 2.64 0.039 0.0128 1.604
9. 30 8 0.65 8 1.94 0.032 0.0055 0.838
10. 60 8 0.65 8 1.98 0.030 0.0060 0.987
11. 30 12 0.65 8 2.00 0.031 0.0054 0.850
12. 60 12 0.65 8 2.00 0.029 0.0062 1.056
13. 30 8 0.75 8 2.60 0.036 0.0089 1.207
14. 60 8 0.75 8 2.31 0.039 0.0112 1.399
15. 30 12 0.75 8 2.72 0.041 0.0098 1.172
16. 60 12 0.75 8 2.14 0.038 0.0114 1.460
17. 45 10 0.7 6 1.96 0.027 0.0077 1.392
18. 45 10 0.7 6 2.07 0.030 0.0084 1.379
19. 45 10 0.7 6 2.33 0.029 0.0082 1.389
20. 45 10 0.7 6 2.31 0.030 0.0084 1.370
54
4.3 Analysis of Results
This section discusses the experimental finding of parametric influences on the
performance characteristics of WEDM machined on Inconel 718. The analysis of
variance (ANOVA) was used to generate statistically the significant machining
parameters and the percentage contribution of each parameter. As mention earlier,
Design Expert version 7.0.0 software was used to analyze the ANOVA analysis.
ANOVA table is commonly used to summarize the experimental results. This table
concludes all information of analysis of variance and case statistics for further
interpretation [9].
In the next section, all the analyses were presented in normal probability plot,
main effect plot and interaction plot for the dependent parameters that significant to the
responses. The interpretation was done unilaterally, meaning that ANOVA analysis for
all 4 responses was done separately at one time.
4.3.1 Analysis results for surface roughness (Ra)
According to the analysis done by the Design Expert software, if the values of
probability (Prob>F) are less than 0.05, it indicated that the factors is significant to the
response parameters. As observed in ANOVA result (Table 4.7), there is only one factor
that influences surface roughness (Ra). In this case, the pulse on (ON) were significant
to the Ra. Other factors, namely servo voltage (V), peak current (IP), pulse off (OFF) are
not significant since the probability values are greater than 0.1, therefore, these factor
did not appear on Table 4.7. In this investigation, 95% of confidence interval (CI) was
used.
55
The lack of fit was not significant which satisfy the model to be fitted. The value
of R2 was quiet high and closed to 1 (≈ 0.6589) which is desirable. The adjusted R2 and
predicted R2
were in agreement as the difference between the values was below 0.2 (≈
0.1126). The adequate prediction is above value of 4 (≈ 7.398), thus indicated that the
model discrimination was adequate.
Table 4.7: ANOVA for surface roughness (Ra)
The information was better illustrated in normal probability plot as shown in
Figure 4.1. Normal probability plot is needed in order to check for the normality of
residuals of the factors studied.
Figure 4.1 reveals that residuals are spread on a straight line implying that errors
are distributed normally. Meanwhile, the plot in Figure 4.2 shows no obvious pattern
56
and unusual structure and all the results fall in the acceptance range. Therefore, it can be
concluded that the model proposed was adequate. For clearer observation, the main plot
in Figure 4.3 indicates how the significant variables affected the Ra.
Figure 4.1: Normal probability plot of residuals for surface roughness (Ra)
57
Figure 4.2: Residual vs. predicted response for surface roughness (Ra)
Figure 4.3: Main effect plot for surface roughness (Ra)
2.4725
1.9813
58
As shown in Figure 4.3, the factor that influence the Ra is pulse on (ON) and its
clearly indicated that whenever ON is increased from 0.65µs up to 0.75µs, the value of
Ra is increased dramatically. The percentage of the increment of Ra was approximately
24.8%. From the correlation obtained in this investigation, the judgement in terms of
selecting the most suitable setting for ON for future optimization can be made. In order
to obtain better Ra during WEDM of Inconel 718, ON should be set at the lowest value,
0.65µs.
The mathematical model for Ra was also developed by the ANOVA analysis in
order to identify the relationship between independent variables, namely, ON and Ra.
The following equation is the final empirical models in terms of coded and factors and
actual factors for Ra respectively. This equation was generated by the Design Expert
software after the transformation had been carried out.
The final equation in terms of coded factors:
Ra = +2.33 + 0.25C
The final equation in terms of actual factors: Ra = -1.21187 + 4.91250*ON
4.3.2 Analysis results for sparking gap (Gap)
Similar approach is also employed to the next response, sparking gap (Gap). The
final ANOVA table for sparking gap (Gap) is shown in Table 4.8
59
Table 4.8: ANOVA for sparking gap (Gap)
In terms of sparking gap (Gap), it was observed that only one factor without any
interaction have the major influence to the sparking gap, it’s also the factor of pulse on
(ON). The value of probability for ON effect was below 0.05 with confidence interval
(CI) is 95%. The lack of fit was not significant which satisfy the model to be fitted. The
value of R2 was high and closed to 1 (≈ 0.8025) which is desirable. The adjusted R2 and
predicted R2 were in agreement as the difference between the values was below 0.2 (≈
0.0019). The adequate prediction value was above 4 (≈ 13.249), thus indicated that the
model discrimination was adequate.
60
Figure 4.4: Normal probability plot of residuals for sparking gap (Gap)
61
Figure 4.5: Residual vs. predicted response for sparking gap (Gap)
Figure 4.6: Main effect plot for sparking gap (Gap)
0.0304
0.0386
62
Figure 4.4 reveals that residuals are spread on a straight line implying that errors
are distributed normally. In other hand, the plot in Figure 4.5 shows no obvious pattern
and unusual structure and all the results fall in the acceptance range. Accept for the run
number 13 which lies far from other runs number, however it still fall in the acceptable
range. Therefore, it can be concluded that the model proposed was adequate. For clearer
observation, the main plot in Figure 4.6 indicates how the significant variables affected
the Gap.
As shown in Figure 4.6, increasing in pulse on (ON) led to an increase of sparking
gap (Gap). Once again, ON revealed an interesting plot with an increment of 27% as it
increased from 0.65µs up to 0.75µs. Previous studies reported that, it is expected that
sparking gap continues to increase if the range of ON is widen. Based on the graph, the
judgement in term of selecting the most suitable setting for ON for future optimization
can be made. In order to obtain better Gap during WEDM of Inconel 718, ON should be
set at the lowest value of 0.65µs. The following equations are the final empirical models
in terms of coded and factors and actual factors for Gap respectively.
The final equation in terms of coded factors: Gap = +0.035 + 4.18 x 10-3
*C
The final equation in terms of actual factors: Gap = -0.024062 + 0.083750*ON
63
4.3.3 Analysis results for material removal rate (MRR)
Material removal rate (MRR) in WEDM processes is an important factor because
of its vital effect on the production cost [4]. Table 4.9 indicates the final analysis of
ANOVA for material removal rate (MRR).
Table 4.9: ANOVA for material removal rate (MRR)
By considering the results from the ANOVA for MRR, there were three main
significant main effects that influence the MRR. ANOVA of MRR also indicated one
interaction of main effect for the MRR. The significant main effects and interaction were
identified by the probability value (Prob>F) justification. With the confidence interval
64
(CI) of 95%, whatever main effect or interaction with their “Prob>F” value of 0.05 or
lower, the main effect are considered as the significant factors that affecting MRR.
Results from ANOVA (Table 4.9) showed the main significant factors were servo
voltage (SV), pulse on (ON) and pulse off (OFF). Meanwhile, interaction between servo
voltage and pulse off (SV*OFF) were observed to be the significant (Prob>F ≈ 0.0076)
interaction in this study.
The lack of fit was not significant which satisfy the model to be fitted. The value
of R2 was high and closed to 1 (≈ 0.9558) which is desirable. The adjusted R2 and
predicted R2
were in agreement as the difference between the values was below 0.2 (≈
0.0358). The adequate prediction value was above 4 (≈ 23.661), thus indicated that the
model discrimination was adequate.
Figure 4.7 reveals that residuals are spread on a straight line implying that errors
are distributed normally. In other hand, the plot in Figure 4.8 show no obvious pattern
and unusual structure and all the results fall in the acceptance range. Accept for the run
number 4 which lies far from other runs number, however it still fall in the acceptable
range. Therefore, it can be concluded that the model proposed was adequate. For clearer
observation, the main plots in Figures 4.9 to 4.10 indicate how the significant variables
affected the MRR.
65
Figure 4.7: Normal probability plot of residuals for material removal rate (MRR)
Figure 4.8: Residual vs. predicted response for material removal rate (MRR)
66
Figure 4.9: Main effect plot for material removal rate (MRR)
From the main plot graph shown in Figure 4.9, it was obvious that an increase in
both SV and ON factors leads to the increase of the MRR. As observed in Figure 4.9,
MRR experienced an increment percentage of approximately 31% when SV is increase
from 30V up to 60V. Similar pattern was revealed when ON was increased from 0.65µs
to 0.75µs, the MRR increases with huge percentage of approximately 81%. Apparently,
opposite result is revealed from the pulse off (OFF) factor. From the graph shown in
Figure 4.9, when OFF increases from 4µs up to 8µs, the MRR slightly decreases about
2.4% respectively. Based on these relationships, maximum MRR can be obtained when
the parameters are set at SV = 60V, ONN = 0.75µs and OFF = 4µs.
0.0071
0.0093
0.0081
0.0083
0.0105
0.0058
67
Figure 4.10: Interaction between SV*OFF for material removal rate (MRR)
Figure 4.10 shows an interaction of the factors graphically for MRR. It was
observed when SV and OFF were set at 60V and 8µs respectively the MRR increase
about 17.6%. Meanwhile, MRR increases more rapidly to 47.8% with an increment of
SV as OFF were set at 4µs. In order to gain maximum MRR the parameters of SV and
OFF should be set up to 60V and 4µs respectively. The following equations are the final
empirical models in terms of coded factors and actual factors for MRR respectively.
The final equation in terms of coded factors: MRR = +8.187 x 10-3 + 1.125 x 10-3*A + 2.350 x 10-3*C – 1.375 x 10-4
– 4.750 x 10
*D -4
*A*D
The final equation in terms of actual factors: MRR = -0.031950 + 1.7 x 10-4*SV + 0.047*ON + 6.43750 x 10-4
- 1.58333 x 10
*OFF -5*SV*OFF
SV = 30V OFF = 8µs MRR = 0.0074
SV = 30V OFF = 4µs MRR = 0.0067
SV = 60V OFF = 4µs MRR = 0.0099
SV = 60V OFF = 8µs MRR = 0.0087
68
4.3.4 Analysis results for cutting speed (CS)
Lastly, the final analysis in this chapter is to determine the factors and interaction
that affect the cutting speed (CS). Similar procedures were applied for the other
responses; the significant factors and the possible interaction of CS are referred to the
ANOVA result shows in Table 4.10. Whatever factors that have the “Prob>F” less than
0.05 are considered as significant factor for CS with the confident interval (CI) used is
95%.
Table 4.10: ANOVA for cutting speed
69
Based on ANOVA in Table 4.10, results show that there are five effects with
“Prob>F” value below 0.05, which are significant for a 95% of confident interval (CI).
These are servo voltage (SV), peak current (IP), pulse on (ON), the interaction between
servo voltage (SV) and peak current (IP) and finally, the interaction between servo
voltage (SV) and pulse off (OFF). Meanwhile from the ANOVA table, factor OFF with
probability value of 0.7386 (> 0.05) is shown because it is required in order to support
the hierarchy in the software and this factor was not discussed any further. Apparently,
the analysis for CS was quite complicated since the number of effects and interactions
were higher than the previous responses.
The lack of fit was not significant which satisfy the model to be fitted. The value
of R2 was high and closed to 1 (≈ 0.9949) which is desirable. The adjusted R2 and
predicted R2
were in agreement as the difference between the values was below 0.2 (≈
0.0054). The adequate prediction value was above 4 (≈ 62.340) as obtained in this case,
thus indicated that the model discrimination was adequate.
Figure 4.11 showed that residuals are spread on a straight line implying that errors
are distributed normally. In other hand, the plot in Figure 4.12 shows no obvious pattern
and unusual structure and all the results fall in the acceptance range. Accept for runs
number 4 and 17 which lies far from other runs number, however they still fall in the
acceptable range. Therefore, it can be concluded that the model proposed was adequate.
For clearer observation, the main plots in Figures 4.13 to 4.14 indicate how the
significant variables affected the CS.
70
Figure 4.11: Normal probability plot of residuals for cutting speed (CS)
Figure 4.12: Residual vs. predicted responses for cutting speed (CS)
71
Figure 4.13: Main effect plot for cutting speed (CS)
The final ANOVA results for CS obtained from the modified model are shown in
Table 4.10. Results show that there are three main effects with “Prob>F” value below
0.05, which are significant for a 95 confident interval (CI) level. As displayed in the
Figure 4.13, CS was affected by SV, IP and ON. For SV factor, increasing in SV from
30 to 60V increases the CS up to 30.8%. Meanwhile, CS was observed to slightly
increase about 3.4% by the increased of the IP from 8A up to 12A. Lastly, as shown
graphically in Figure 4.13, ON shows the significant factor for CS until up to 40.1% of
total increasing in CS value. Based on these relationships, maximum CS can be obtained
when the parameters are set at SV = 60V, IP = 12A and ON = 0.75µs.
0.9679
1.3196
0.9222
1.1307
1.0936
1.2663
72
(a) Interaction between SV*IP
(b) Interaction between SV*OFF
Figure 4.14: Interaction plot of cutting speed (CS)
SV = 30V IP = 8A CS = 0.9669
SV = 60V IP = 12A CS = 1.3051
SV = 60V IP = 8A CS = 1.2280
SV = 30V OFF = 8µs CS = 1.0093
SV = 30V OFF = 4µs CS = 0.9273
SV = 60V OFF = 8µs CS = 1.2167
SV = 60V OFF = 4µs CS = 1.3268
73
Figures 4.14 (a) and (b) show the significant interaction factors of the parameters
for the CS. Based on the ANOVA results (Table 4.10), the “Prob>F” value of SV*IP,
SV*OFF interaction were 0.0057 and < 0.0001 respectively. As the value of “Prob>F”
mentioned is closer to zero, it indicates that the model of interaction factor was
significant. Therefore, the first interaction to be considered was the interaction between
servo voltage and peak current (SV*IP) as shown graphically on Figure 4.14 (a). When
IP was set at 8A and SV was vary from 30V to 60V, CS increased about 27%. On the
other hand, CS percentage increased about 35% when IP was set at low level, 4µs with
SV vary remain unchanged. This 35% increment for CS is obtain when SV = 60V and IP
= 12A.
Next interaction was the interaction between servo voltage and pulse off
(SV*OFF).as shown in Figure 4.14 (b). When SV is set between 30V and 60V with a
constant OFF of 8µs, the CS experienced a 20.6% increment. Similar pattern was
recorded as OFF is set at low value of 4µs with the same setting of SV, it contributed
about 43.1% of increment in CS value. These results implied that in order to achieved
maximum value of CS the SV and OFF should be set at 60V and 4µs respectively. The
following equations are the final empirical models in terms of coded factors and actual
factors for CS respectively.
The final equation in terms of coded factors: Sqrt(CS) = +1.05 + 0.071*A + 8.814 x 10-3*B + 0.094*C - 6.955 x 10-4
+ 8.315 x 10
*D -3
*A*B - 0.022*A*D
The final equation in terms of actual factors: Sqrt(CS) = -0.58760 + 6.25238 x 10-3*SV - 8.06548 x 10-3
+ 0.031963*OFF + 2.77165 x 10
*IP + 1.88401*ON -4*SV*IP - 7.18012 x 10-4*SV*OFF
74
Summary of the significant factors that were achieved from WEDM experimental
run is shown in Table 4.11. These include all the responses investigated in this
experiment.
Table 4.11: Summary of the significant factors in WEDM Inconel 718
Responses Significant Factor Surface roughness (Ra) C Sparking gap (Gap) C Material removal rate (MRR) A-C-D-AD Cutting speed (CS) A-B-C-D-AB-AD
Note: A = Servo voltage, B = Peak current, C = Pulse on, D = Pulse off
4.4 Confirmation Test
The confirmation test is the final step undertaken during this experiment. The
purpose of the confirmation runs is to validate the conclusion drawn during the analysis
phases [35]. In addition, the confirmation tests need to be carried out in order to ensure
that the theoretical predicted model for optimum results using the software was accepted
or in other word to verify the adequacy of the models that were developed. All
parameters used in the confirmation test were suggested by Design Expert software.
Three (3) confirmation tests were carried out in order to compare the experimental
results from the prediction made by the ANOVA. Table 4.12 shown in this section
indicates the optimization of quality characteristic needed for each response.
75
Table 4.12: Quality characteristic of the machining performance
Machining Characteristic Quality Characteristic Surface roughness (Ra) Minimum Sparking gap (Gap) Minimum Material removal rate (MRR) Maximum Cutting speed (CS) Maximum
4.4.1 Confirmation tests and results
Table 4.13 shows the three series of parameters settings for the confirmation test.
The parameters values were selected between the high and low range of the machining
factor that have been studied from previous experiment.
Table 4.13: True value of confirmation test experiment
Machining Voltage : 80V Wire Speed : 10 m/min Wire Tension : 800 g Injection Pressure : 12 bar SV IP ON OFF Exp. No. Servo Voltage
(V) Peak Current
(A) Pulse Duration
(µs) Pulse Interval
(µs) 1. 30 8 0.65 4 2. 30 11.2 0.65 5.5 3. 30 11.6 0.65 7
Tables 4.14 to 4.17 show the results of the machining responses for surface finish
(Ra), sparking gap (Gap), material removal rate (MRR) and cutting speed (CS)
respectively.
76
Table 4.14: Confirmation test results for surface roughness (Ra)
No. of trial
Rx
(µm)
Ry
(µm)
Total Average
(µm) 1. 1.92 1.75 1.84 2. 2.19 2.13 2.16 3. 2.13 1.84 1.99
Table 4.15: Confirmation test results for sparking gap (Gap)
No. of trial
Sparking gap on top
surface (mm)
Sparking gap on bottom
surface (mm)
Total Average
(µm)
1. 0.032 0.022 0.027 2. 0.029 0.026 0.028 3. 0.031 0.031 0.031
Table 4.16: Confirmation test results for cutting speed (CS)
No. of trial
Machining distance
(mm)
Machining time
(min)
Cutting speed, CS
(mm/min)
1. 10 14.20 0.704 2. 10 13.70 0.730 3. 10 13.65 0.733
Table 4.17: Confirmation test results for material removal rates
No. of trial
Sparking gap
(mm)
Volume
(mm3
Mass
)
(g)
Machining time
(min)
MRR
(g/mm3) 1. 0.027 6.750 0.055 14.20 0.0039 2. 0.028 7.000 0.057 13.70 0.0042 3. 0.031 7.750 0.063 13.65 0.0047
77
Full confirmation test measurement results for sparking gap (Gap) can be referred
to Appendix D.
4.5 Comparison of the Test Results
Based on the flow charts of experiment steps discussed in chapter three, the
comparison of the test results between the theoretically prediction and confirmation test
results was the final consideration that will evaluate whether the optimum parameters
predicted were in the allowable range. The margin of error from the prediction and
experimental results was set below than 15%. Margin error was calculated using the
equation below:
Tables 4.18 to 4.21 show the comparison of test results between theoretical
prediction and confirmation test for surface roughness (Ra), sparking gap (Gap),
material removal rate (MRR) and cutting speed (CS) respectively.
78
Table 4.18: Comparison test results for surface roughness (Ra)
No. of confirmation
run
Experimental (Confirmation test)
Prediction (Design Expert)
Error Margin (%)
1. 1.84 1.98 7.68 2. 2.16 1.98 8.33 3. 1.95 1.98 1.54
Table 4.19: Comparison test results for sparking gap (Gap)
No. of confirmation
run
Experimental (Confirmation test)
Prediction (Design Expert)
Error Margin (%)
1. 0.027 0.030 11.11 2. 0.028 0.030 7.14 3. 0.031 0.030 3.23
Table 4.20: Comparison test results for material removal rate (MRR)
No. of confirmation
run
Experimental (Confirmation test)
Prediction (Design Expert)
Error Margin (%)
1. 0.0039 0.0044 12.82 2. 0.0042 0.0046 9.52 3. 0.0047 0.0049 4.26
Table 4.21: Comparison test results for cutting speed (CS)
No. of confirmation
run
Experimental (Confirmation test)
Prediction (Design Expert)
Error Margin (%)
1. 0.704 0.754 7.10 2. 0.730 0.783 7.26 3. 0.733 0.810 10.51
79
4.6 Verification of the Mathematical Models
From ANOVA analysis, the mathematical models for each response were
generated by the Design Expert software. These mathematical models identified the
relationship between the independent variables (SV, IP, ON and OFF) and the dependent
variables (Ra, Gap, MRR and CS). Even though the mathematical models are able to
predict the results automatically when setting parameters are inserted into the system,
these mathematical models still require verification.
This section verifies the mathematical models developed by the Design Expert
software. Some examples of experimental data were selected and manually calculated
using the equation. This is to make sure that the predicted results given by the software
is correct. In this case, all calculation was based on the setting parameters used in the
experimental data trial # 1 (SV = 30V, IP = 8A, ON = 0.65µs, OFF = 4.0 µs). The
mathematical models for all responses were presents as below:
a. Surface roughness (Ra)
From Table 4.7, the surface roughness (Ra) can be obtain from:
Ra = +2.33 + 0.25(C)
By using experimental data trial #1, the predicted Ra was calculated as follows;
Ra = +2.33 + 0.25(C), where;
C = -ve (low)
Ra = 2.33 + 0.25(-1)
= 2.08 µm
#
From the experimental data trial #1 results, the value of Ra is 2.11µm. So the margin
error is 1.42%.
80
b. Sparking Gap (Gap)
Similarly from Table 4.8, the sparking gap (Gap) can be obtained from:
Gap = +0.035 + 4.18 x 10-3
By using experimental data trial #1, the predicted Gap was calculated as follows;
(C)
Ra = +0.035 + 4.18 x 10-3
C = -ve (low)
(C) where;
Ra = +0.035 + 4.18 x 10-3
=
(-1)
0.0308mm
#
From the using experimental data trial #1 results, the value of Gap is 0.0304µm. So the
margin error is 1.32%.
c. Material removal rate (MRR)
Similarly from Table 4.9, material removal rate (MRR) can be obtained from:
MRR = +8.187 x 10-3 + 1.125 x 10-3*A + 2.350 x 10-3*C – 1.375 x 10-4
– 4.750 x 10
*D -4
By using experimental data trial #1, the predicted MRR was calculated as follows;
*A*D
MRR = +8.187 x 10-3 + 1.125 x 10-3(A) + 2.350 x 10-3(C) – 1.375 x 10-4
– 4.750 x 10
(D) -4
A = -ve (low)
(A)(D), where;
B = -ve (low)
C = -ve (low)
D = -ve (low)
MRR = +8.187 x 10-3 + 1.125 x 10-3(-1) + 2.350 x 10-3(-1) – 1.375 x 10-4
– 4.750 x 10
(-1) -4
=
(-1)(-1)
0.0044 g/min
#
81
From the experimental data trial #1 results, the value of MRR is 0.0042µm. So the
margin error is 4.76%.
d. Cutting speed (CS)
Similarly from Table 4.8, the cutting speed (CS) can be obtained from:
Sqrt(CS) = +1.05 + 0.071(A) + 8.814 x 10-4(B) + 0.094(C) - 6.955 x 10-4
+ 8.315 x 10
(D) -3
By using experimental data trial #1, the predicted CS was calculated as follows;
(A)(B) - 0.022(A)(D)
Sqrt(CS) = +1.05 + 0.071(A) + 8.814 x 10-4(B) + 0.094(C) - 6.955 x 10-4
+ 8.315 x 10
(D) -3
A = -ve (low)
(A)(B) - 0.022(A)(D)
B = -ve (low)
C = -ve (low)
D = -ve (low)
Sqrt(CS) = 1.05 + 0.071(-1) + 8.814 x 10-3
-6.955 x 10
(-1) + 0.094(-1) -4(-1) + 8.315 x 10-3
=
(-1)(-1) - 0.022(-1)(-1)
0.7451 mm/min
#
From the experimental data trial #1 results, the value of CS is 0.733 mm/min. So the
margin error is 2.25%.
Full verification results of the mathematical models for all responses are shown in Table
4.22.
82
Table 4.22: Margin of error for actual results and predicted values (%)
Exp.
No.
Responses (Actual) Responses (Predicted) Ra
(µm) SG
(mm) MRR
(g/min) CS
(mm/min) Ra
(µm) %
(error) SG
(mm) %
(error) MRR
(g/min) %
(error) CS
(mm/min) %
(error) 1. 2.11 0.028 0.0042 0.733 1.98 -6.5 0.0304 8.4 0.0044 4.0 0.754 2.8 2. 1.74 0.029 0.0062 1.045 1.98 12.4 0.0304 6.2 0.0076 18.2 1.074 2.6 3. 1.90 0.029 0.0050 0.772 1.98 4.4 0.0304 4.5 0.0044 -14.3 0.756 -2.2 4. 2.18 0.035 0.0082 1.139 1.98 -9.8 0.0304 -14.7 0.0076 -8.3 1.146 0.6 5. 2.59 0.040 0.0093 1.130 2.47 -4.8 0.0388 -3.9 0.0091 -2.5 1.117 -1.2 6. 2.45 0.040 0.0125 1.523 2.47 1.1 0.0388 -2.4 0.0123 -1.8 1.500 -1.5 7. 2.33 0.037 0.0084 1.111 2.47 5.8 0.0388 4.1 0.0091 7.4 1.119 0.7 8. 2.64 0.039 0.0128 1.604 2.47 -6.8 0.0388 -1.1 0.0123 -4.3 1.585 -1.2 9. 1.94 0.032 0.0055 0.838 1.98 2.3 0.0304 -6.7 0.0051 -8.9 0.828 -1.2 10. 1.98 0.030 0.0060 0.987 1.98 0.3 0.0304 2.1 0.0064 5.5 0.984 -0.3 11. 2.00 0.031 0.0054 0.850 1.98 -0.7 0.0304 -3.2 0.0051 -6.9 0.830 -2.4 12. 2.00 0.029 0.0062 1.056 1.98 -0.9 0.0304 3.7 0.0064 2.4 1.053 -0.4 13. 2.60 0.036 0.0089 1.207 2.47 -5.0 0.0388 6.9 0.0098 8.7 1.206 -0.1 14. 2.31 0.039 0.0112 1.399 2.47 6.8 0.0388 -1.3 0.0111 -1.4 1.393 -0.4 15. 2.72 0.041 0.0098 1.172 2.47 -9.8 0.0388 -4.5 0.0098 -0.5 1.209 3.0 16. 2.14 0.038 0.0114 1.460 2.47 13.7 0.0388 2.2 0.0111 -3.2 1.475 1.0 17. 1.96 0.027 0.0077 1.392 2.23 12.0 0.0346 22.6 0.0082 6.0 1.112 -25.2 18. 2.07 0.030 0.0084 1.379 2.23 7.0 0.0346 12.0 0.0082 -2.6 1.112 -24.0 19. 2.33 0.029 0.0082 1.389 2.23 -4.4 0.0346 16.8 0.0082 -0.2 1.112 -24.9 20. 2.31 0.030 0.0084 1.370 2.23 -3.7 0.0346 14.6 0.0082 -2.6 1.112 -23.2
83
CHAPTER 5
DISCUSSION
5.1 Introduction
The main purpose of this research was to study the effect of WEDM performance on
Inconel 718 by using various setting of selected parameter. Performance of the WEDM on
Inconel 718 is determined in terms of machining outputs such as surface roughness (Ra),
sparking gap (Gap), material removal rate (MRR) and cutting speed (CS). This chapter
elaborates more clearly about the relationship between the performance measures to the
main parameters and its influence.
84
5.2 Surface Roughness (Ra)
Based on observation from the ANOVA result for surface roughness (Ra), pulse
on (ON) is the only significant factor that influence the Ra. It was revealed when ON
was set from 0.65µs up to 0.75µs the Ra was increases about 24.8%. The 3D surface
graph for Ra is given in Figure 5.1. Results show that the surfaces profile was in
accordance to the model fitted. It was understood that generally Ra increases only by a
single factor, ON. Factor of OFF clearly shows that there is no effect on Ra even though
the value of OFF was varied from 4µs to 8µs.
Figure 5.1: 3D interaction graph for surface roughness (Ra)
Similar trend of Ra behavior was reported by Mas Ayu [27], the Ra increases
when ON increases due to the longer time of machining, leading to the higher possibility
of re-sparking and localized sparking to occur. In other words, re-sparking can cause
poor surface finish since only the initial phase spark contribute to the material removal
rate, while the following spark were poorly distributed along the kerf surface, debris and
removed particles.
85
Meanwhile, Ahmet Hascalyk and Ulas Caydas [23] concluded that, the increasing
pattern of Ra occurred when intense heat was generated during each electrical discharge.
Since the greater the discharge energy conducted into the machining zone, the greater
the melted depth of the workpiece that is created. Furthermore, greater discharge energy
will produce a larger crater, causing a larger surface roughness value on the workpiece.
5.3 Sparking Gap (Gap)
In term of sparking gap (Gap), the significant factor that influences this response is
also a pulse on (ON). It was recorded by ANOVA when the pulse on (ON) is increase
it’s led to increasing of sparking gap (Gap). As for this study, the increment of sparking
gap was 27% from low level to the higher level setting of ON. From Figure 5.2, it was
understood graphically that the Gap values builds up simultaneously with the pulse on
increment of pulse on from 0.65µs to 0.75µs. In this case, pulse off (OFF) also clearly
showed there is no effect to the Gap even though the value of OFF varies from 4µs to
8µs.
Figure 5.2: 3D interaction graph for sparking gap (Gap)
86
According to Mas Ayu [27] and Mohd Faisal [31], the wider the ON time, the
longer the machining to takes place resulting in a wider spark gap. They reported that
servo voltage (SV) also contributed to the effect to the Gap on workpiece materials of
tungsten carbide [27] and titanium alloy [31]. As for this study, WEDM Inconel 718
indicated no significant interaction of servo voltage (SV) was recorded on ANOVA.
S.S. Mahapatra and Amar Patnaik [35] also suggested that factor like pulse on
(ON) have been found to be significant in effect on sparking gap (Gap). This study
revealed similar pattern of Gap behavior when ON is vary from low to the high value.
This is due to the increment of power density for the wire to discharge sparks and to
elevate the temperature in the gap, hence the higher the power the larger the sparking
gap (Gap.
5.4 Material Removal Rate (MRR)
Results obtained from the ANOVA in Table 4.9 clearly show that the most
significant factors in affecting material removal rate (MRR) were servo voltage (SV),
pulse on (ON) and pulse off (OFF). At the same time, interaction between servo voltage
and pulse off (SV*OFF) were also observed to be the significant (Prob>F ≈ 0.0076)
interaction in this study. Apparently, results obtain from Figure 5.3 indicated that SV
and ON contribute to the increment of MRR about 31% and 81% respectively.
Meanwhile, the opposite effect of MRR was obtained from the OFF value, whereby
MRR was slightly decreased about 2.4% when OFF is varied from 4µs to 8µs. In
addition, SV and ON are also interacted to each other which contribute to the increase
on MRR. When SV and OFF were set at 60V and 8µs respectively the MRR increase
about 17.6%. Meanwhile, MRR increased more rapidly up to 47.8% with an increment
of SV when OFF was set at 4µs.
87
Figure 5.3: 3D interaction graph of SV*OFF for material removal rate (MRR)
This result corresponds to the previous researchers finding of Kuang-Yuan Kung
and Ko-Ta Chiang [37] whereby they reported that, the electrical spark-erosion process
occurs successively and then the removal of melt results in the form of crater on the
machined surface. The amount of melt removal determines the level of material removal
rate (MRR). The addition of zinc coated in electrode wire provide significantly increase
the tensile strength, lowers the melting point and increase the vapor pressure rating
resulting in higher MRR.
M.S Hewidy et al.[4] proposed, that the increment in the rate of the heat energy
hence in the rate of melting and evaporation. Increase in peak current above a certain
limit, leads to arcing which decreases the discharge number and the machining
efficiency, and subsequently decreases the MRR. Meanwhile, increase in ON time
means applying the same heating temperature for a longer time. This will cause an
increase in the evaporation rate and number of gas bubbles, which explodes with high
ejecting force when the discharge ceases causing removal of bigger volume of the
molten metal. Increasing of MRR is continued with the increase of the ejecting force
until reaching a situation in which the ejecting force will have no more increase in MRR
since the molten metal decreases.
88
5.5 Cutting Speed (CS)
Figure 5.4 presents the 3D effect of cutting speed at various setting of servo
voltage (SV) and peak current. The ANOVA results shown in Table 4.10, has shown
that the significant parameters for CS were SV, IP and ON. Pulse off (OFF) was only
significant during the interaction with SV. However, before ANOVA analysis can be
proceed, all the data for CS were obtained from the calculation by dividing the
machining distance, d with the machining time, t.
It can be seen that increasing in IP seemly not affected much on CS. From the
calculation, CS only increase about 3.45% with increase in IP from 8A up to 12A.
Improvement in CS increases dramatically as shown in Figure 5.4 during interaction
between IP and SV. Increase in SV from 30V to 60V while IP is maintain at 8A,
resulting in increment in CS about 27%. Otherwise, 35% of increment in CS were
recorded when SV is varied from 30V up to 60V during maintaining the IP value at 12A.
From this finding, the best setting for maximum CS are set at SV = 60V, IP = 12A and
ON = 0.75µs.
Figure 5.4: 3D interaction graph of IP*SV for cutting speed (CS)
89
Second interaction that had a significant effect on CS with value of “Prob>F”
below 0.05 was interaction between OFF*SV. When SV is varied between 30V and 60V
with a constant OFF of 8µs, the CS experienced a 20.6% increment.
Figure 5.5 indicates graphically a similar pattern was recorded as OFF is set at low
value of 4µs with the same setting of SV. It contributes about 43.1% of increment in the
CS value. In this case, to achieve maximum value of CS, the SV and OFF should be set
at 60V and 4µs respectively.
Figure 5.5: 3D interaction graph of OFF*SV for cutting speed (CS)
Analogously, higher OFF leads to a lower machining time and reduces the CS.
This may be due to the fact that during OFF time, the operating impulse was switched
off and no current flow at this stage. Too long off time will increase the machining time
and reduced CS simultaneously [27]. Based on Figure 5.5, it was obvious that highest
CS can dramatically be achieved by setting OFF at low level while SV is set at high
level. This setting condition is able to maximize the time for machining and increase the
CS. Sufficient setting of OFF time is very important because it can lead to erratic
cycling and retraction of the advancing servo, thus slowing down the operation cycle
[51].
90
CHAPTER 6
CONCLUSIONS AND RECOMMENDATIONS
6.1 Conclusions
Inconel 718 is a high strength thermal resistant material alloy. It is also a highly
strain rate sensitive material which work hardens readily, and contains hard particles
making it a very difficult-to-cut material. As for this research, Inconel 718 was
machined by using Sodick WEDM linear motor series AQ537L using zinc coated brass
wire diameter 0.25mm as the electrode. This research presents an investigation on the
effect of machining parameters on WEDM in terms of surface roughness (Ra), sparking
gap (Gap), material removal rate (MRR) and cutting speed (CS). The level of
importance of the machining parameters on the machining responses was determined by
using ANOVA. A total of 20 runs of experiment including centre point were performed
in this study which done using Design Expert software version 7.0.0. The following
conclusions were drawn based on the performance of machining responses namely;
surface roughness (Ra), sparking gap (Gap), material removal rate (MRR) and cutting
speed (CS).
91
a. The results of ANOVA and comparisons of experimental data proved that
the mathematical models of the value of Ra, Gap, MRR and CS were fairly
well fitted with the experimental values with a 95% confidence level.
b. The confirmation test show that the errors associated with Ra, Gap, MRR
and CS are within the range of 1.54% ~ 12.82%.
c. Pulse on (ON) was found to be the most significant factor influencing all
responses investigated. Increasing in ON will lead to the low quality of
machining responses such as Ra and Gap. Meanwhile, the opposite were
observed for MRR and CS whereby increasing of ON will result in better
rate of MRR and CS.
d. Higher value of thermal conduction and specific heat capacity of Inconel 718
causes the decrease of efficiency of WEDM using zinc coated brass wire as
electrode.
6.2 Recommendations
Based on the observation and finding in this study, the future works might attempt
to consider the other performance criteria proposed as follows:
a. The used of different type of wire materials as electrode need to be
considered for better understanding for WEDM of Inconel 718.
b. Surface integrity study can be evaluated in order to understand the effect of
the machining parameters on the surface quality and microstructure of the
machined surface.
c. Consideration of others performance criteria, such as surface waviness, form
accuracy and surface flatness as additional output parameter fors WEDM of
Inconel 718.
92
REFERENCE
1. Norliana Mohd Abbas, Darius G. Solomon, Md. Fuad Bahari. (2007). “A Review
on Current Research Trends in Electrical Discharge Machining (EDM).”
International Journal of Machine Tools & Manufacture. 47: pg. 1214–
1228.
2. J. Proshka, A.G Mamalis, N.M. Vaxevanidis. (1997). “The Effect of Electrode
Material on Machinability in Wire Electro Discharge Machining.”
Journal of Material Processing Technology. 69: pg. 233-237.
3. Lin, Y.C., Yan, B.H., and Chang, Y.S. (2000). “Machining Characteristics of
Titanium Alloy (Ti-6Al-4V) Using a Combination process of EDM and
USM.” Journal of Material Processing Technology. 104: pg. 171-177.
4. M.S. Hewedy, T.A. El-Taweel, M.F. El-Safty. (2005) “Modelling the Machining
Parameters of Wire Electrical Discharge Machining of Inconel 601 Using
RSM.” Journal of Material Processing Technology. 169: pg. 328-336.
5. S. Sharif. (1999). “Face Milling of Titanium Alloys Using Coated and Uncoated
Carbide Tools.” University of Conventry, UK: PhD Thesis.
93
6. H. Ho, S.T Newman, S. Rahimifard, R.D. Allen. (2004). “State of the Art in Wire
Electrical Discharge Machining (WEDM).” International Journal of
Machine Tool Manufacture 44: pg. 1247-1259.
7. M. Rahman, W.K.H. Seah, T.T. Teo. (1997). “The Machinability of Inconel 718.”
Journal of Material Processing Technology 63 (1997) 199-204.
8. J.C. Rebelo, A. Dias Morao, D. Kremer, J.L. Lebrun. (1998). “Influence of EDM
Pulse Energy on the Surface Integrity of Martensitic Steel.” Journal of
Material Processing Technology 84: pg. 90–96.
9. Rival. (2005). “Electrical Discharge Machining of Titanium Alloy Using Copper
Tungsten Electrode With SiC Powder Suspension Dielectric Fluid.”
Universiti Teknologi Malaysia: Master Thesis.
10. I. Puertas, C.J. Luis. (2003). “Modeling the Manufacturing Parameters in
Electrical Discharge Machining of Siliconized Silicon Carbide.” Proc.
Inst. Mech., J. Eng. Manuf. 217 (Part B): pg. 791–803.
11. S. Sarkar, S. Mitra, B. Bhattacharyya. (2005). “Parametric Analysis and
Optimization of Wire Electrical Discharge Machining of Titanium
Aluminide Alloy.” Journal of Materials Processing Technology 159: pg.
286–294.
12. Sommer, C. (2000). “Non-Traditional Machining Handbook.” First edition.
Advanced Publishing Houston.
13. Pandey, P.C., and Shah, H.S. (1980). “Modern Machining Processes.” New Delhi.
Tata Mcgraw-Hill.
14. K.P. Rajurkar. (1994). “Handbook of Design, Manufacturing and Automation,
Chapter 13: Nontraditional Manufacturing Processes.” Wiley, USA.
94
15. George F. Schrader, Ahmad K. Elshennawy, Lawrence E. (2000). “Manufacturing
Processes & Materials.” Forth Edition. SME, ISBN: 0872635171.
16. Helmi A. Youssef, Hassan El-Hofy. (2008). “Machining Technology – Machine
Tools and Operation, Chapter 11: Nontraditional Machine Tools and
operation.” CRR Press. ISBN: 978-1-4200-4339-6.
17. Jack M. Walker. (1996). “Hanbook of Manufacturing Engineering, Chapter 16:
Nontraditional Machining Methods.” Marcel Decker, Inc., ISBN: 0-8247-
8962-8.
18. Mchado A.R. and J. Wallbank. (1990). “Machining of Titanium and its Alloys – A
Review.” Proc. Instn. Mech. Engrs. Vol. 204: pg. 53-60.
19. Guitrau, E.B. (1991). “Wire EDM – An Overview of Technology and Trends.”
SME Technical Paper, MR 91-519.
20. William, R.E., Rajurkar, K.P. (1991). “Study of Wire Electrical Discharge
Machining Surface Characteristic.” Journal of Materials Processing
Technology, Vol. 28: pg. 486-493.
21. Jerzy Kozak, Kamlakar P. Rajukar, Niraj Chandarana. (2004). “Machining Low
Electrical Conductive Materials by Wire electrical Discharge Machining
(WEDM).” Journal of Materials Processing Technology 149: pg. 266-
271.
22. D. Scot, S. Boyina, K.P. Rajurkar. (1991). “Anallysis and Optimization of
Parameters Combination in Wire Electrical Discharge Machining.” Int. J.
Prod. Res. 29 (11): pg. 2189-2207.
23. Ahmet Hascalyk, Ulas Caydas. (2004). “Experimental Study of Wire Electrical
Discharge Machining of AISI D5 Tool Steel.” Journal of Material
Processing Technology 148: pg. 362-367.
95
24. Y.S. Liao, J.T. Huang, H.C. Su. (1997). “A Study on the Machining Parameters
Optimization of Wire Electrical Discharge Machining.” Journal of
Material Processing Technology 71: pg. 487-493.
25. Wang, C.C., and. Yan, B.H. (2000). “Blind Hole Drilling of Al2O3
/6061 Al
Composite Using Rotary Electro Discharge Machining.” Journal of
Material Processing Technology 102: pg. 90-102.
26. Van Tri, N. (2002). “Electrical Discharge Machining of Aluminum Alloy Using
Classical Design of Experiment Approach.” Universiti Teknologi
Malaysia: Master Thesis.
27. Mas Ayu. (2006). “Investigation of WEDM Machining Tungsten Carbide.”
Universiti Teknologi Malaysia: Master Thesis.
28. Nihat Tosun, Can Cogun, Gul Tosun. (2004). “A Study on Kerf and Material
Removal Rate in Wire Electrical Discharge Machining Based on Taguchi
Method.” Journal of Material Processing Technology 152: pg. 316-322.
29. Erden, A. (1982). “Role of Dielectric Flushing on Electric Discharge Machining
Performance.” Machine Tool Design and Research Conference (1982)
September 14-15. Manchester. 23: pg. 283-289.
30. Kinoshita, N., Fukui, M., Shichida, H., Gamo, G., sata, T. (1976). “Study on EDM
with Wire Electrode; Gap Phenomena.” Ann. CIRP 25 (1976) (1): pg.
141-145.
31. Faisal Othman. (2008). “EDM Wirecut of Titanium Alloy (Ti-6A14V) Using
Brass Coated Wire Electrode.” Universiti Teknologi Malaysia: Master
Thesis.
32. O. A. Abu Zeid. (1997). “On the Effect of Electro Discharge Machining Parameter
on the Fatigue Life of AISI D6 Tool Steel.” Journal of Material
Processing Technology 68: pg. 27-32.
96
33. H.T. Lee, T.Y. Tai. (2003). “Relationship between EDM Parameters and Surface
Crack Formation.” Journal of Material Processing Technology 142: pg.
676-683.
34. Q. H. Zhang, J.H. Zhang, S.F. Ren, Z.W. Niu, X. Ai. (2005). “A Theoritical Model
of Surface Roughness in Ultrasonic Vibration Assisted Electrical
Discharge Machining in Gas.” International Journal of Manufacturing
Technology and Management 7: 381-390.
35. S.S Mahapatra, Amar Panaik. (2007). “Optimization of Wire Electrical Discharge
Machining (WEDM) Using Taguchi Method” International Journal of
Manufacturing Technology 34: pg. 911-925.
36. S. Sarkar, M. Sekh, S. Mitra, B. Bhattacharyya. (2008). “Modelling and
Optimization of Wire Electrical Discharge Machining of γ-TiAl in Trim
Cutting Operation.” Journal of Material Processing Technology 205: pg.
376-387.
37. Kuang-Yuan Kung, Ko-Ta Chiang. (2008). “Modeling and Analysis of
Machinability Evaluation in the Wire Electrical Discharge Machining
(WEDM) Process of Aluminum Oxide-Based Ceramic.” Material and
Manufacturing Processes 23: pg. 241-250.
38. R. Ramakrishnan, L. Karunamoorthy. (2008). “Modelling and Multi-Responses
Optimization of Inconel 718 on Machining of CNC WEDM Processes.”
Journal of Material Processing Technology 207: pg. 342-349.
39. Z.N. Guo, T.C. Lee, T.M. Yue, W.S. Lau. (1997). “A Study of Ultrasonicaided
Wire Electrical Discharge Machining.” Journal of Materials Processing
Technology 63: pg. 823–828.
97
40. M. Kunieda, S. Furuoya, N. Taniguchi. (1991). “Improvement of EDM Efficiency
by Supplying Oxygen Gas into gap.” CIRP Annals - Manufacturing
Technology 40: pg. 215–218.
41. M.L. Jeswani. (1981). “Effect of the Addition of Graphite Powder to Kerosene
Used as the Dielectric Fluid in Electrical Discharge Machining.” Wear
70: pg. 133–139.
42. Mikell P. Groover. (2007). “Fundamentals of Modern Manufacturing – Materials,
Process and Systems, Chapter 26.” John Wiley & Sons, Inc. ISBN: 13-
978-0-74485-6.
43. Intech EDM. (1995). “EDM Wire: A Reference to Understanding, Selecting and
Using Wire on Wire-Cut EDM Machine.” 2nd
Edition USA.
44. Armarego, E.J.A., and Brown, R.H. (1969). “The Machining of Metals.” Prentice
Hall New Jersey.
45. Mustafa Ihan Gokler, Alp Mihat Ozanozgu. (1997). “Experimental Investigation
of Effects of Cutting Parameters on Surface Roughness in the WEDM
Process.”
46. F.C. Campbell. (2006). “Manufacturing Technology for Aerospace Structural
Materials, Charter 6: Super Alloys.” Elsevier Ltd. ISBN-13: 978-1-85-
617495-4.
47. I.A. Choudhury, M.A. El-Baradie. (1999). “Machinability Assessment of Inconel
718 by Factorial Design of Experiment Coupled with Response Surface
Methodology.” Journal of Material Processing Technology 95: pg. 30-39.
48. Graham, D. (2002). “Turning Difficult-to-Machined Alloy.” Modern Machine
Shop.
98
49. Lochner, R.H., and Matar, J.E. (1990). “Designing for Quality; an Introduction to
the Best of Taguchi and Western Methods of Statistical Experiment
Design.” Chapman and Hall, New York.
50. Montogory. (2001). “Practical Experiment Design for Engineers and Scientist.”
3rd
Edition, John Waley & Sons, Inc. New York.
51. Lee, S.H. and Li, X.P. (2001). “Study of the Effect of Machining Parameters on
the Machining Characteristics in Electrical Discharge Machining of
Tungsten Carbide.” Journal of Materials Processing Technology. 115: pg.
344-358.
99
APPENDIX A
MASTER PROJECT PLANNING
100
Appendix A1: Schedule for Master project part I (Semester 1 – 2009/2010)
100
Appendix A2: Schedule for Master project part II (Semester 2 – 2009/2010)
101
102
APPENDIX B
SUMMARY OF FINDING RELATED TO EDM
PERFORMANCE
103
Appendix B1: Summary of finding related to EDM performance
Researcher
Issue and Methodology Used Result/Conclusion
M.S. Hewidy et al. [4].
Development of mathematical model for correlating the inter-relationship of various WEDM machining parameter (peak current, duty factor, wire tension, and water pressure) at MRR, spark gap & surface finish. Material: Inconel 601 (6mm Thickness) Electrode: Brass Wire – CuZn377, Ø.25mm Method: RSM
Result show surface finish greatly influence by peak current, duty factor and wire tension. Surface finish increase by increase in peak current whereby decrease with the increase in duty factor and wire tension.
S.S Mahapatra & Amar Panaik [35]
Optimization of WEDM process parameter (WEDM parameters: discharge current, pulse duration, pulse frequency, wire speed, wire tension and dielectric flow) Material: AISI D2 tool steel (10mm thickness) Electrode: Zink-coated cooper wire (Stratified, copper Ø.25mm) Method: Taguchi Method
Determine the discharge current, pulse duration and dielectric flow rate as significant role in rough cutting to minimize surface finish and increase the MRR.
S. Sarkar et al. [36]
Modeling and optimization of WEDM in trim cutting operation (WEDM parameters:cutting speed). Material: γ-TiAl (15mm thickness) Electrode: Brass wire Ø 0.25mm Method: RSM
Determine the surface finish decrease as the cutting speed increase.
104
Kuang-Yuan Kung & Ko-Ta Chiang
Modeling and analysis of machinibility evaluation in the WEDM process of aluminum oxide-based ceramic (WEDM parameters: peak current, pulse-on time, duty factor and wire speed). Material: Aluminium oxide Al2O3
Electrode: Cylindrical electrolytic copper, Ø 0.20mm.
(10mm thickness)
Method: RSM
Conclude the values of MRR and surface finish increase with the increase in pulse on-time and duty factor up to certain limits then decrease with the further increase in the pulse on-time and duty factor.
R. Ramakrishnan & L. Karunamoorthy
[38]
Modeling and multi-responses optimization of Inconel 718 on machining of CNC WEDM process (WEDM parameters: pulse on-time, pulse off-time, wire feed speed & ignition current) Material: Inconel 718 (14mm) Electrode: Brass wire, Ø 0.25mm Method: Taguchi Method.
Results shows by an increase of pulse time and ignition current effect on MRR was improved. But at higher rates of pulse on-time and ignition current, the surface finish of the Inconel 718 was affected.
Nihat Tosun et al. [28]
Study on kerf and MRR in WEDM machining proceses (WEDM parameters: pulse duration, open circuit voltage, wire speed & dielectric pressure) Material: AISI 4140 steel (10 mm thickness) Electrode: Brass wire – CuZn37, Ø 0.25mm Method: Taguchi Methods
Revealed the highly effective parameters on both kerf and MRR were found as open circuit voltage and pulse duration. Wire speed & dielectric pressure were less effective factors.
105
APPENDIX C
EXPERIMENTAL RESULTS OF SPARKING GAP
(TOP AND BOTTOM SURFACE)
106
Appendix C1: Experimental results of sparking gap (top surface)
Exp. No.
Kerf Width (mm) Kerf
Width Average
(mm)
Sparking Gap (mm) Sparking
Gap Average
(mm) 1 2 3 1 2 3 1. 0.314 0.309 0.315 0.313 0.032 0.030 0.033 0.031 2. 0.307 0.303 0.326 0.312 0.029 0.027 0.038 0.031 3. 0.319 0.315 0.307 0.314 0.035 0.033 0.029 0.032 4. 0.328 0.328 0.327 0.328 0.039 0.039 0.039 0.039 5. 0.334 0.338 0.334 0.335 0.042 0.044 0.042 0.043 6. 0.320 0.331 0.325 0.325 0.035 0.041 0.038 0.038 7. 0.322 0.340 0.320 0.327 0.036 0.045 0.035 0.039 8. 0.328 0.323 0.334 0.328 0.039 0.037 0.042 0.039 9. 0.321 0.317 0.329 0.322 0.036 0.034 0.040 0.036 10. 0.318 0.314 0.319 0.317 0.034 0.032 0.035 0.034 11. 0.327 0.316 0.317 0.320 0.039 0.033 0.034 0.035 12. 0.329 0.326 0.297 0.317 0.040 0.038 0.024 0.034 13. 0.326 0.315 0.323 0.321 0.038 0.033 0.037 0.036 14. 0.336 0.315 0.323 0.325 0.043 0.033 0.037 0.037 15. 0.343 0.332 0.329 0.335 0.047 0.041 0.040 0.042 16. 0.371 0.342 0.321 0.345 0.061 0.046 0.036 0.047 17. 0.319 0.316 0.338 0.324 0.035 0.033 0.044 0.037 18. 0.320 0.317 0.330 0.322 0.035 0.034 0.040 0.036 19. 0.308 0.308 0.325 0.314 0.029 0.029 0.038 0.032 20. 0.318 0.303 0.322 0.314 0.034 0.027 0.036 0.032
107
Appendix C2: Experimental results of sparking gap (bottom surface)
Exp. No.
Kerf Width (mm) Kerf
Width Average
(mm)
Sparking Gap (mm) Sparking
Gap Average
(mm) 1 2 3 1 2 3 1. 0.297 0.295 0.304 0.299 0.024 0.023 0.027 0.024 2. 0.315 0.290 0.301 0.302 0.033 0.020 0.026 0.026 3. 0.291 0.310 0.306 0.302 0.021 0.030 0.028 0.026 4. 0.304 0.308 0.323 0.312 0.027 0.029 0.037 0.031 5. 0.319 0.330 0.328 0.326 0.035 0.040 0.039 0.038 6. 0.325 0.338 0.337 0.333 0.038 0.044 0.044 0.042 7. 0.316 0.321 0.327 0.321 0.033 0.036 0.039 0.036 8. 0.318 0.338 0.329 0.328 0.034 0.044 0.040 0.039 9. 0.305 0.319 0.298 0.307 0.028 0.035 0.024 0.029 10. 0.297 0.301 0.308 0.302 0.024 0.026 0.029 0.026 11. 0.297 0.305 0.314 0.305 0.024 0.028 0.032 0.028 12. 0.295 0.296 0.308 0.300 0.023 0.023 0.029 0.025 13. 0.319 0.329 0.321 0.323 0.035 0.040 0.036 0.037 14. 0.327 0.334 0.336 0.332 0.039 0.042 0.043 0.041 15. 0.331 0.323 0.328 0.327 0.041 0.037 0.039 0.039 16. 0.286 0.313 0.322 0.307 0.018 0.032 0.036 0.029 17. 0.271 0.291 0.286 0.283 0.011 0.021 0.018 0.016 18. 0.300 0.295 0.303 0.299 0.025 0.023 0.027 0.025 19. 0.296 0.308 0.300 0.301 0.023 0.029 0.025 0.026 20. 0.287 0.316 0.308 0.304 0.019 0.033 0.029 0.027
108
APPENDIX D
EXPERIMENTAL RESULTS OF SPARKING GAP FOR
CONFIRMATION TEST (TOP AND BOTTOM SURFACE)
109
Appendix D1: Confirmation experimental results of sparking gap (top surface)
Appendix D2: Confirmation experimental results of sparking gap (bottom surface)
Exp. No.
Kerf Width (mm) Kerf
Width Average
(mm)
Sparking Gap (mm) Sparking
Gap Average
(mm) 1 2 3 1 2 3 1. 0.313 0.318 0.309 0.313 0.032 0.034 0.030 0.032 2. 0.313 0.312 0.300 0.308 0.032 0.031 0.025 0.029 3. 0.310 0.313 0.315 0.313 0.030 0.032 0.033 0.031
Exp. No.
Kerf Width (mm) Kerf
Width Average
(mm)
Sparking Gap (mm) Sparking
Gap Average
(mm) 1 2 3 1 2 3 1. 0.313 0.318 0.309 0.313 0.032 0.034 0.030 0.032 2. 0.313 0.312 0.300 0.308 0.032 0.031 0.025 0.029 3. 0.310 0.313 0.315 0.313 0.030 0.032 0.033 0.031