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