Measurement 125 (2018) 336–349

Contents lists available at ScienceDirect

Measurement
journal homepage: www.elsevier.com/locate/measurement

Multi-objective optimization of micro-electrical discharge machining of T

nickel-titanium-based shape memory alloy using MOGA-II
⁎
Mustufa H. Abidia, Abdulrahman M. Al-Ahmaria,b, Usama Umera, , Mohammed Sarvar Rasheedc
a
Princess Fatima Alnijris Research Chair for Advanced Manufacturing Technology (FARCAMT Chair), Advanced Manufacturing Institute, King Saud University, Riyadh,
Saudi Arabia
b
Industrial Engineering Department, College of Engineering, King Saud University, Riyadh, Saudi Arabia
c
Baynounah Institute of Science and Technology, Abu Dhabi Vocational Education Training Institute, Madinat Zayed, Abu Dhabi, United Arab Emirates

A R T I C LE I N FO A B S T R A C T

Keywords: Shape memory alloys (SMAs) have received significant attention especially in biomedical and aerospace in-
Micro-EDM (µEDM) dustries owing to their unique properties. However, they are difficult-to-machine materials. Electrical discharge
Micro-machining machining (EDM) can be used to machine difficult to cut materials with good accuracy. However, several
Shape memory alloy challenges and issues related with the process at micro-level continue to exist. One of the aforementioned issues
MOGA-II
is that the micro-EDM (µEDM) process is extremely slow when compared to other non-conventional processes,
such as laser machining, although it offers several other benefits. The study considers the analysis and opti-
mization of µEDM by using a multi-objective genetic algorithm (MOGA-II). Drilling of micro-holes is performed
by using a tabletop electrical discharge machine. Nickel-Titanium (Ni-Ti) based SMA (a difficult to cut advance
material) is used as a specimen. The objective involves determining optimal machining parameters to obtain
better material removal rate with good surface finish. The results of the study indicate that MOGA-II is an
efficient tool to optimize input parameters. Optimum results are obtained with tungsten electrode at low to
moderate capacitance values and low discharge voltage. Conversely, brass electrode yields high MRR at the
expense of tool wear and micro-holes quality.

1. Introduction
Various difficult-to-machine materials are available in the market
owing to advancements in the field of material science and engineering.
Non-conventional processes, such as electrical discharge machining
(EDM), laser machining, and ultrasonic vibrations, are widely used to
process these materials [1–3]. Specifically, EDM is widely used because
it provides good surface finish and accuracy without undesirable results
such as a heat affected zone [4,5]. Furthermore, there is a significant
demand for miniaturization given developments in the field of auto-
motive, aerospace, and biomedical industries. Therefore, a precise and
accurate machining process is required to fulfill the industrial demand.
Micro-EDM (µEDM) offers a solution by providing an accurate process
to machine difficult to cut materials even at the micro-level [6,7].
The conventional machining processes are incompetent in terms of
machining advanced materials within an acceptable tolerance limit [8].
Especially, in the case of micro-holes drilling, the drill bits are easily
broken when the conventional drills are used, and chip evacuation is
also an issue [9]. The µEDM process results in no mechanical stresses
and machining vibrations because there is no direct contact between
the tool and the workpiece [10]. A further advantage of the process is
that it can machine materials irrespective of their hardness. However,
the workpiece material should be electrically conductive. Workpiece
hardness plays a significant role in the case of conventional machining
processes. In the EDM process, there is no direct contact between tool
and workpiece, and electrical sparks are used for material removal, and
thus, the hardness of workpiece does not come into effect.
Murali and Yeo explained EDM as a process in which a cathode and
an anode (two electrodes (i.e., workpiece and tool) that are separated
by a dielectric medium come close to each other and make the dielectric
conductive by breaking it down [11]. Hence, a spark is generated be-
tween cathode and anode, and thermal energy released by the phe-
nomenon is used for material removal by melting and vaporization. If
the amount of energy released is reduced to a lower value, then micro-
machining is possible [11]. Furthermore, EDM is also known as spark
machining and removes material without any physical contact between
the workpiece and tool electrode. It removes material by using re-
petitive pulse discharges obtained from an electric pulse generator with
dielectric fluid flowing in between the workpiece and tool electrode.
Each generated spark melts and vaporizes material from both the

⁎
Corresponding author at: P.O. Box-800, Advanced Manufacturing Institute, College of Engineering, King Saud University, Riyadh 11421, Saudi Arabia.
E-mail address: uumer@ksu.edu.sa (U. Umer).

https://doi.org/10.1016/j.measurement.2018.04.096
Received 15 February 2017; Received in revised form 17 April 2018; Accepted 30 April 2018
Available online 02 May 2018
0263-2241/ © 2018 Elsevier Ltd. All rights reserved.
M.H. Abidi et al.
workpiece and tool electrode, and this leads to material removal. The
discharge current causes the heating of the dielectric, workpiece, and
tool electrode. The dielectric forms a plasma channel of partially io-
nized gas. The channel provides a heat source and heats up the work-
piece beyond its melting temperature, and the material is removed and
solidified with the dielectric cooling effect. The µEDM method is a
subset of EDM that operates at the miniature level [12].
The characteristic that differentiates µEDM from macro-EDM is the
size of tool electrode and discharge energy. The plasma size exceeds in
the macro-EDM and axes movement resolutions by several orders of
magnitude, which are at micron levels in case of µEDM [13]. µEDM can
be used to machine almost burr free micro-holes ranging from a few
hundred microns to 5 µm, and complex 3D cavities [14]. The µEDM
method is mainly selected owing to its high efficiency and accuracy for
the micro-machining of difficult-to-machine material irrespective of
their hardness, such as metals, metallic alloys, graphite, and a few se-
lected ceramics [15]. Hence, µEDM is considered as an efficacious
process to drill any types of holes including blind, small, or deep holes.
Extant studies successfully drilled holes ranging from a diameter of
5 µm to few hundred µm by using the process [6,9].
Shape Memory Alloys (SMAs) are special type of advanced en-
gineering materials that exhibit the ability to remember their shape or
the pseudo-plasticity. Shape memory phenomenon was observed in Ni-
Ti alloys in 1963 [16]. Specifically, SMAs exhibit several exceptional
properties including high mechanical strength, wear resistance, hard-
ness, high efficiency in converting thermal into mechanical energy, and
excellent biocompatibility [17,18]. Otsuka and Kakeshita indicated that
Ni-Ti-based SMA is widely used in industrial applications (such as in the
aerospace industry), in MEMS, and as sensors, actuators, and couplings
[19,20]. Additionally, SMAs are used in a broad range of applications in
various fields including dentistry [21]; thin film coatings [22]; ad-
vanced actuators [23]; safety of civil structures [24]; bone tissue en-
gineering [25]; and biomedical engineering [26]. Hsieh et al. examined
the machining characteristics of TiNiX SMA by using wire electro-dis-
charge machining (WEDM). They examined the thickness of the recast
layer that is an issue in the EDM process and concluded that the
thickness of the recast layer varies with the pulse duration[27].
Several studies are underway to increase the efficiency of the pro-
cess. Specifically, µEDM is the key process for drilling micro-holes (with
diameters less than 200 µm) in diesel and gasoline injection nozzles and
in turbine blades for cooling effect [28]. The µEDM method is a com-
petent technique to create simple and complex shapes, cavities, and 3D
micro-contours. Furthermore, µEDM is used to produce inkjet nozzles
[29]. Sato et al. used the µEDM process to drill holes on inkjet nozzles
for the first time [30]. Jeong et al. proposed an algorithm to predict the
shape of tool and holes drilled by µEDM [31]. Yeo et al. investigated the
µEDM input parameters (discharge energy and electrode material) to
machine zirconium-based bulk metallic glass. The output parameters
that were considered included surface roughness, tool wear, and burr
width. It was reported that the lower discharge energy mitigates the
surface roughness and tool wear. Additionally, it was concluded that
tube electrodes performed better than rod electrodes in terms of tool
wear [32]. The effect of the process parameters of µEDM on plastic and
mold steel workpiece was investigated in an extant study [33]. The
blind micro-holes were analyzed for their shape and dimensional ac-
curacy. The µEDM of titanium alloy Ti–6Al–4V was modeled by using
response surface methodology and an artificial neural network (ANN).
Process parameters that were considered included peak current, pulse
on time, and dielectric flushing pressure. The output responses were
material removal rate (MRR), tool wear rate (TWR), and overcut. The
set of optimal process parameters obtained were the pulse-on-time of
14.2093 ms, peak current of 0.8363 A, and flushing pressure of 0.10 kg/
cm2. The obtained values were successfully validated by experiments
[34]. Prihandana et al. suggested a novel method of vibrating the di-
electric and the mix MoS2 powder in it to increase the MRR and ob-
tained an improved surface finish in the µEDM process [35].
337
Measurement 125 (2018) 336–349
Somashekhar et al., optimized the gap voltage, capacitance, federate,
and speed in µEDM process for drilling holes in a 1-mm thick alumi-
nium plate. The tool material corresponded to tungsten carbide with a
diameter of 500 µm. Moreover, ANN and genetic algorithm (GA) was
used to model and optimize the process. The results revealed that a
combination of ANN and GA was effective in modeling and optimizing
the µEDM process [36].
Multi-objective optimization of µEDM process using GA was per-
formed. Responses that were considered included TWR and overcut.
Pulse-on time, peak current, and flushing pressure were the input
variables. A set of Pareto optimal non dominated points were presented
[37]. Output responses of micro electrical discharge milling were op-
timized by using GA. Responses that were considered included MRR
and TWR. Input parameters were voltage, capacitance, electrode rota-
tion speed, and feedrate. In an extant study, multi-objective optimiza-
tion was performed while machining Ti-6Al-4V specimen. Optimized
values of MRR, and TWR were obtained at a parameter setting of 150 V,
0.01 μF, 734.24 rpm, and 18 mm/min [38]. A study was conducted to
drill accurate micro-holes with least electrode depletion by using
µEDM. The Grey based Taguchi approach was applied to optimize the
input parameters (gap voltage, capacitance, pulse-on-time, aspect ratio,
and electrode rotation) for output responses (overcut at the top and
bottom and taper angle). Aluminum sheets were used as the workpiece
material, and a tungsten rod was used as an electrode. Validating tests
performed at optimal machining parameters to yield good results [39].
Grey relational analysis technique was implemented to optimize the
µEDM drilling process of Inconel 718 nickel-based super alloy. Ex-
periments were performed by using full factorial design. Pulse duration
and discharge current were input variables that were considered, and
taper ratio and hole dilation were measured as the output response. The
findings indicated that the discharge current affects the outputs more
than the pulse duration [6].
Manjaiah et al. examined the WEDM of SMA by using the Taguchi or-
thogonal array. The results revealed that pulse duration affects the MRR and
surface roughness substantially [40]. The µWEDM process of titanium alloy
(Ti-6Al-4V) was examined by using the response surface methodology
(RSM). The input variables included voltage, capacitance, and feed rate. The
examined responses included MRR, kerf width, and surface roughness.
Additionally, GA was used to implement multi-objective optimization [41].
The µWEDM process was modeled, and multi-objective optimization was
conducted for the micro-machining of Inconel-718. An L18 mixed ortho-
gonal array design was used in the experiments. Grey relational analysis was
implemented for multi-objective optimization. The examined process
parameters included voltage, capacitance, feed rate, wire tension and wire
feed velocity. The output responses included MRR and surface roughness
[42]. Furthermore, RSM and GA were used to optimize the input para-
meters (discharge current, pulse-on time and pulse-off time) while in-
vestigating their effect on output responses (MRR and TWR) during the
µEDM of SS 316L by using a brass electrode. The results indicated that the
resulting optimal input parameters improved the selected objectives sub-
stantially [43].
The µEDM input parameters (discharge current, pulse duration,
pulse off, and jump distance) were optimized by using grey relational
analysis for machining high-speed steel alloy (SKH59) using a tungsten-
carbide (WC) electrode. Furthermore, a comparative study was per-
formed between coated electrodes to achieve better MRR, surface
finish, and lower TWR. The results indicated that WC-coated Cu elec-
trodes achieved highest MRR, WC-coated Ag produced a better surface
finish while the WC electrodes exhibited the least wear rate [44]. The
µED milling was investigated to optimize its parameters (voltage, ca-
pacitance, electrode rotational speed, and feed rate) while machining
Ti-6Al-4V. The responses under study were MRR and TWR. Box-
Behnken design of RSM was used to conduct the experiments. It was
revealed that capacitance and feed rate were the most significant
parameters that were affect the responses [45]. Jahan et al. investigated
the effect of different functional parameters on the µEDM behavior of
M.H. Abidi et al.
Table 1
Output responses specified by extant studies for the µEDM process.
Output response Refs.
MRR [9,10,35,34,5,36,48,38,41,42,43,44,45,46,47]
TWR [9,10,5,49,37,48,38,43,44,45,46,47]
Surface Roughness [35,10,12,41,42,46]
Overcut [9,10,49,37,39,44]
Taper Angle [9,39,6]
Circularity [6,5]
Duralumin (Al alloy AA 2024). Micro-grooves were engraved on the
surface of AA 2024 by using a resistance-capacitance (RC) type pulse
generator. The output responses included MRR, TWR, surface rough-
ness, and depth of grooves. An increase in the capacitance and voltage
resulted in an increase in discharge energy, and this leads to higher
MRR although TWR also increases. Thus a set of recommended para-
meters were suggested to obtain better outputs [46]. Porous stainless
steel samples were machined by using µEDM. The machining char-
acteristics were compared with those of austenitic stainless steel (AISI
304), and it was observed that porous stainless steel exhibited a better
machining performance. Furthermore, it was revealed that porosity and
pore size significantly influenced the machining characteristics [47].
Table 1 shows the output responses considered by various studies while
examining the µEDM process.
Based on the literature and Table 1, it is observed that no credible
studies were performed on machining difficult to cut material, such as
Ni-Ti-based SMAs, by using micro-EDM process. Most extant studies
focused on optimizing MRR and TWR. Furthermore, various studies
focused on different materials. However, to the best of the authors'
knowledge, previous studies did not consider all the output responses
specified in Table 1 in a single study in conjunction with the application
of multi-objective optimization. It is necessary to apply the multi-ob-
jective optimization technique for µEDM as the output responses by
nature contradict each other. For example, an increase in the MRR
owing to high voltage and current increases the TWR and deteriorates
the surface finish. Hence, a proper balance should exist between the
input parameters to achieve better machining in all aspects. Therefore,
in the study, a multi-objective genetic algorithm (MOGA-II) is applied
to optimize the MRR, TWR, surface roughness (Ra), overcut, taper
angle, and circularity in the process of drilling micro-holes in the Ni-Ti-
based SMA by using the µEDM process.
2. Experimental procedure
2.1. Experimental setup
A special experimental setup is used to investigate the μEDM pro-
cess. The apparatus consists of a tabletop electro-discharge machine
Fig. 1. Illustrative diagram of μEDM.
338
Measurement 125 (2018) 336–349
integrated with microscope lenses. The advantage of the setup is that
preliminary investigation of the micro-holes can be performed without
removing the workpiece from the worktable. Hence, machining is
continued from the same position if a few further improvements are
required, and this improves the accuracy. Fig. 1 shows the line diagram
of the micro electro-discharge machine. Fig. 2 shows the re-
sistor–capacitor (RC) circuit of the machine’s power supply unit.
Fig. 3 shows the tabletop μEDM setup used in the study that consists
of resistor–capacitor generator. It generates pulses ranging from a few tens
of nanoseconds to a few micro-seconds. The voltage level varies from 45 V
to 120 V. The setup is attached with an optical microscope that provides the
option to assess the micro-hole without removing the workpiece from the
setup. The inverted polarity of the tool electrode offers significant benefits.
In case of conventional EDM, the tool (i.e., electrode) is charged as anode
owing to the polarity effect, and this results into high MRR and lower tool
wear. However, short pulse duration is used in μEDM, and thus the effect is
reversed and the tool is charged as cathode.
2.2. Specimen, tool, and dielectric material
As mentioned above in section 1, SMAs exhibit several advantages
and applications. However, they are considered as materials that are
difficult to cut by using conventional machining and especially when
micro-level manufacturing is required. Hence, Ni-Ti-based SMA is used
in the study. The specimen is cut into a rectangular plate corresponding
to 3 × 1.5 × 0.5 mm. The properties of the specimen material are listed
in Table 2.
Tungsten and brass cylindrical electrodes with a diameter of 100 µm
are used for drilling micro holes. The material properties of the tool are
given in Table 3.
Kerosene is used as a dielectric fluid although several other fluids
are available.
2.3. Experiment procedure
In μEDM, proper selection of polarity of an electrode is considered
as an important aspect. Therefore, suitable electrode polarity is selected
based on the responses. Extant studies indicated that an increase in the
MRR decreases TWR and improves the surface quality given the use of
negative electrode polarity [50,51]. Therefore, in μEDM, all the ex-
periments were performed by using the negative polarity of the tool
electrode. Micro-holes were drilled by using two types of tool material
on SMA plate with different sets of input parameters obtained by using
response surface methodology as described in next section.
2.4. Experimental design
Design of experiments is a statistical technique and powerful ana-
lyzing tool that is used to simultaneously analyze the effect of more
M.H. Abidi et al. Measurement 125 (2018) 336–349

Fig. 2. Resistor–capacitor (RC) circuit of the machine.

Table 3
Tool’s material properties.
Tool Material Tungsten (Super Brass (C2680)
FSK)

Composition W > 99.99% Cu:64–68%, Pb: < 0.05%,

Zn: 36–32%
Density (kg/m3) 19,300 8400
Melting Point (°C) 3370 930
Coefficient of Thermal 4.6 × 10−6 18.7 × 10−6
Expansion (/°C)
Thermal Conductivity 180 119
(W/m-C)

Table 4
Input variables and their levels.
Levels Capacitance (pF) Discharge Voltage (V) Electrode Material

1 475 100 Brass

2 315 90 Tungsten
3 155 80

2.4.2. Output responses

The performance of μEDM process is assessed by calculating the
Fig. 3. Tabletop μEDM (Developed by T. Masuzawa) [13]. MRR, TWR, and accuracy of micro-holes (i.e., overcut, taper angle, and
circularity of micro-holes produced) as shown in Fig. 4.
Table 2 Material removal rate (MRR): This is an extremely important re-
Material properties of the specimen. sponse because machining speed is always an issue in electrical dis-
Work piece Material Ni-Ti SMA
charge machining. In the μEDM process, both the workpiece material
and tool electrode material are eroded and the MRR is calculated as the
Composition Ni: 55.8%, Ti: 44.2%, C < 0.02% average volume of the material removed to the machining time and is
Density (kg/m3) 6500 generally expressed in cubic millimeters per minute. The general vo-
Melting Point (°C) 1310
lume formula considered for the MRR in a workpiece is the volume of a
Electrical Resistivity (µΩ-m) 820
Modulus of Elasticity (MPa) 41–75 × 103 conical frustum. In tool wear, the rate is considered as cylindrical vo-
Coefficient of Thermal Expansion (/°C) 11 × 10−6 lume over the machining time to the make through hole and is math-
Ultimate Tensile Strength (MPa) 1070 ematically expressed as given below.
Total Elongation (%) 10

MRR = {( ∏ /3) ∗ [R 2t + R t R b + R2b] ∗h}/t (mm3/min)

than one factor on the several responses. A set of input variables with
their levels are selected to examine their effects on the selected output
responses.
2.4.1. Input variables
The input variables that were considered for performing μEDM of
Ni-Ti SMA include capacitance, discharge voltage, and electrode ma-
terial. The parameters and their levels are given in Table 4.
where
Rt is the radius at the entrance of the micro-hole produced (mm).
Rb is the radius at the bottom of the micro-hole produced (mm).
h is the thickness of workpiece material (mm).
t is the machining time to make a micro-hole (min).
Tool wear rate (TWR): This is important because tool life is vital for
machining cost and accuracy. In μEDM, tool wear is responsible for the

339
M.H. Abidi et al.
dimensional inaccuracy of micro-holes. An increase in the TWR is ob-
served at higher discharge energy levels, and thus the use of shorter
pulses is recommended to reduce the TWR. The TWR is calculated as
the ratio of the volume of material removed from the tool to the ma-
chining time. In the study, frontal tool wear is considered to calculate
the TWR as the tool first erodes from the bottom and subsequently from
the sides.
Tool Wear Rate (TWR) = {(∏ D2q)/4t} (mm3/min)
where
D is the tool diameter (after erosion) (µm).
q is the frontal electrode wear or eroded length of the tool (mm).
t is the time to machine through hole.
Overcut: The overcut of micro-hole is one of the accuracy aspects of
micro hole that is evaluated by calculating the difference between the
average diameter of micro-hole after machining and the diameter of
tool electrode as shown in Fig. 4b. Mathematically, the formula is ex-
pressed as follows:
Overcut = {(Da −D)/2} (μm)
where Da is the average diameter of the micro-hole produced (µm).
Da = {(Dt + D b)/ 2} (μm)
where
Dt is the Top diameter of the micro-hole produced (µm).
Fig. 4. (a) Overcut (b) Measuring taperness (c) Circularity of Micro-holes.
340
Measurement 125 (2018) 336–349
Db is the Bottom diameter of the micro-hole produced (µm).
D is the tool diameter (µm).
Taper angle: The taperness is measured as the difference between
the entrance diameter and the exit diameter of micro-hole, and the
angle between them is termed as the taper angle (Fig. 4b), and is given
as follows:
Taperangle (α) = tan−1 {(Dt −D b)/ 2h} (°)
where α denotes the taper angle, and h denotes the thickness of
workpiece (mm).
Surface roughness (Ra): In any manufacturing processes, after ma-
chining surface roughness is the most important property that should be
considered. It is defined as the surface irregularities in the machined
area that exist owing to various reasons. Surface roughness is important
in all types of machining processes either it is conventional or non-
conventional. In micro-EDM, the surface roughness is an important
factor that should be considered. In the present study, at different en-
ergy levels, the average surface roughness (Ra) is measured by Talysurf
CCI600 and subsequently observed by using an optical microscope and
a scanning electron microscope (SEM).
Circularity error: The error of circularity is defined as the radial
distance between the minimum circumscribing circle and the maximum
inscribing circle (Fig. 4c). Circularity is measured as the difference
between the maximum and minimum diameters of the drilled hole. A
Buehler optical microscope was used to examine the circularity by
measuring the various diameters around the profile.
M.H. Abidi et al. Measurement 125 (2018) 336–349

Table 5
Design of the experiment matrix and measured output responses.
Run Input Variables Responses

Cap. Elect. Volt. Ra (µm) TWR (mm3/min) × 10−4 MRR (mm3/min) × 10−4 Overcut (µm) Taper Angle (°) Circularity (µm)

1 155 1 80 0.0982 1.7508 3.5015 7.6523 2.2161 6.1480

2 155 1 90 0.0985 2.0417 3.8390 8.2751 1.9657 5.4330
3 155 1 100 0.0991 2.0895 3.7313 8.8979 1.8989 5.8610
4 155 2 80 0.0829 0.4649 1.4900 6.0169 1.1229 0.9700
5 155 2 90 0.0864 0.6724 1.3780 6.1034 1.3376 2.5530
6 155 2 100 0.0912 0.7884 1.8772 6.1898 1.4554 3.2650
7 315 1 80 0.1103 2.1963 4.8689 8.0733 1.5820 3.3528
8 315 1 90 0.1102 3.4545 6.3784 8.5059 1.5488 3.4443
9 315 1 100 0.1167 4.6477 7.7828 8.9385 1.5025 3.5098
10 315 2 80 0.0895 1.1573 2.9537 6.7259 1.1651 2.6399
11 315 2 90 0.0969 1.5469 3.2940 6.6229 1.1614 2.6810
12 315 2 100 0.1047 2.0016 3.7399 6.5198 1.1708 2.7480
13 475 1 80 0.1081 2.2986 5.8637 8.4944 0.9969 0.5910
14 475 1 90 0.1214 4.8023 8.8122 8.7367 1.1188 1.4296
15 475 1 100 0.1334 7.8185 12.6110 8.9791 1.1118 1.2330
16 475 2 80 0.0972 1.9518 4.5392 7.4348 1.1637 4.2570
17 475 2 90 0.1078 2.8465 4.9555 7.1423 0.9984 2.8349
18 475 2 100 0.1181 2.8431 5.0770 6.8499 0.8751 2.1760

Cap. = Capacitance in pF; Elect. = Electrode Material; Volt. = Discharge voltage in Volts.

2.4.3. Full factorial design

Table 6 Experiments were conducted according to the design of experiments
Optimization problem with objective functions and constraints. based on full factorial design (FFD), and results of the selected re-
Objective functions (1) Minimize Ra sponses are illustrated in Table 5.
(2) Minimize TWR

Constraints (1) MRR ≥ 3.0 × 10−4 mm3/min

(2) Taper ≤ 1.85°
3. Multi-objective optimization
(3) Overcut ≤ 8.5 µm
(4) Circularity ≤ 4.75 µm In order to perform multi-objective optimization, response surfaces
were developed by using radial basis functions based on the full fac-
torial matrix as shown in Table 5. A multi-objective optimization pro-
blem was formulated to minimize the surface roughness of micro holes

Fig. 5. Optimization work flow.

341
M.H. Abidi et al. Measurement 125 (2018) 336–349

Circularity

Taper
Capacitance*Electrode
Overcut Voltage*Electrode
Capacitance*Voltage

MRR Electrode
Voltage
Capacitance
TWR

Ra

0 0.2 0.4 0.6 0.8 1

Fig. 6. Relative strength of input variables for output responses.

Fig. 7. Effect of capacitance and discharge voltage on Ra for the brass electrode.

and tool wear rate of the electrode. The remaining four responses were
considered as constraints in the optimization problem. The results with
extremely low MRR and high values of the overcut, taper angle, and
circularity error were considered as unfeasible. The optimization pro-
blem with objective functions and constraints is listed in Table 6, and
the optimization workflow with all variables is shown in Fig. 5. The
response surface generation and optimization study were performed by
using a general purpose optimization software modeFRONTIER®.
Specifically, a multi-objective genetic algorithm termed as MOGA-II
is used for the optimization search. It is an efficient algorithm that uses
a smart multi-search elitism. The new elitism operator preserves a few
excellent solutions without causing premature convergence to local-
optimal frontiers. For the purpose of simplicity, MOGA-II requires only
extremely few user-provided parameters, and several other parameters
are internally settled to provide robustness and efficiency to the opti-
mizer. The algorithm attempts a total number of evaluations that is
equal to the number of points in the DOE table (the initial population)
multiplied by the number of generations. For further details on the
MOGA-II, readers are directed to the modeFRONTIER® documentation
[52].
4. Results and discussion
The relative strengths of input variables and their interactions on
output responses by using analysis of variance (ANOVA) are shown in
Fig. 6. Capacitance is indicated as the most dominant factor for Ra,

342
M.H. Abidi et al. Measurement 125 (2018) 336–349

Fig. 8. Effect of capacitance and discharge voltage on Ra for the tungsten electrode.

Fig. 9. Effect of the capacitance and electrode material on TWR at discharge voltage = 90 V.

TWR, and MRR and is followed by electrode material and discharge
voltage. The overcut is mainly influenced by the electrode material
while others display negligible effects. The taper angle is mainly af-
fected by capacitance and electrode material followed by their inter-
action. The interaction of capacitance and electrode material sig-
nificantly affects the circularity error and is followed by the individual
effect of capacitance and electrode material.
The effects of input variables on output responses are shown in
Figs. 7–14 by using the results obtained from response surfaces. The
effects of capacitance and discharge voltage on Ra for brass and tung-
sten electrodes are shown in Figs. 7 and 8. As shown in the figures, Ra
increases with capacitance and discharge voltage. However, at low
capacitance (i.e., below 300 pF), Ra is almost constant irrespective of
the discharge voltage for the brass electrode. The effect is slightly

343
M.H. Abidi et al. Measurement 125 (2018) 336–349

Fig. 10. Effect of the capacitance and discharge voltage on the MRR for the brass electrode.

Fig. 11. Effect of the capacitance and discharge voltage on the MRR for the tungsten electrode.

reduced for the tungsten electrode, and Ra always increases with dis-
charge voltage for all capacitance values. For both the brass and
tungsten electrodes, the effect of capacitance on Ra increases with the
discharge voltage. The maximum surface roughness is low for the
tungsten electrode owing to its low electrical conductivity and tool
wear rate, which results in small and uniform craters depth along the
machined surface. This is in accordance with results in extant studies
[53,54]. As shown in Fig. 7, the anomalous change in surface roughness
for the brass electrode at low voltage may be attributed owing to poor
flushing conditions at the low spark gap. An increase in the voltage
increases the strength of the electric field and results in a stabilized
spark and predictable behavior.
Fig. 9 shows the effect of capacitance and electrode material on
TWR at a discharge voltage of 90 V. Along Y-axis, an electrode value of

344
M.H. Abidi et al. Measurement 125 (2018) 336–349

Fig. 12. Effect of the capacitance and electrode material on the taper angle at discharge voltage = 90 V.

Fig. 13. Effect of the capacitance and electrode material on the overcut at the discharge voltage = 90 V.

1.0 corresponds to the brass electrode and 2.0 for the tungsten elec-
trode. For both electrodes, the effect of an increase in capacitance is
extremely identical, namely tool wear rate increases with capacitance
except that it displays higher values for the brass electrode. The be-
havior does not change significantly with the increase in discharge
voltage, and the TWR simply displayed higher values. The low TWR for
the tungsten electrode is mainly owing to its excellent wear and heat
resistant properties when compared to those of the brass electrode.
The effects of capacitance and discharge voltage on the MRR for the
brass and tungsten electrodes are shown in Figs. 10 and 11, respec-
tively. Generally, the MRR increases with capacitance although there
are differences in the effect of discharge voltage for both electrodes. At
low capacitance (i.e., below 200 pF), an increase in the discharge vol-
tage does lead to any significant change in MRR for both electrodes.
With respect to the brass electrode, the effect of an increase in voltage is
prominent at higher capacitance values, and the MRR reaches its peak

345
M.H. Abidi et al. Measurement 125 (2018) 336–349

Fig. 14. Effect of the capacitance and electrode material on the circularity error at the discharge voltage = 90 V.

Fig. 15. 3D bubble chart denoting the optimum, feasible, and unfeasible design points.

at the top right corner of the diagram as shown in Fig. 10. In the case of
the tungsten electrode (Fig. 11), the discharge voltage does not sig-
nificantly affect the MRR and is mostly dependent on capacitance va-
lues. Overall, MRR is higher for the brass electrode for the same dis-
charge energy in the selected domain. An increase in capacitance and
voltage increases the discharge energy and results in larger craters that
contribute to high MRR and surface roughness. Brass exhibits less
thermal conductivity than tungsten, and thus additional heat is ab-
sorbed by the workpiece and results in high MRR than that of the
tungsten electrode. The minimal effect of the voltage for tungsten
electrode is owing to its low electrical conductivity that results in a less
variation in the spark gap when compared to that of the brass electrode.
Fig. 12 shows the effect of capacitance and electrode material on
taper angle at a discharge voltage of 90 V. The results indicate that the

346
M.H. Abidi et al.
Fig. 16. Parallel coordinate chart showing all the parameters used in the study.
Table 7
Details of the optimal solution along with the neighboring points.
Design Input Variables Responses
Cap. Elect. Volt. Ra (µm) TWR (mm3/min) × 10−4
A 355 2 80 0.0914 1.3564
B 315 2 85 0.0928 1.3577
C 395 2 80 0.0934 1.5598
taper angle of micro holes decreases with the increase in capacitance
for both electrodes. The lowest taper angle is achieved with tungsten
electrode at higher capacitance while maximum taper angle is obtained
with the brass electrode at low capacitance values. As shown in Table 5,
the effect of discharge voltage is minimal in general, and the taper angle
is observed to slightly decrease with the increase in the discharge vol-
tage for the brass electrode at low to moderate capacitances. At the low
capacitance, the discharge energy is low and results in low machining
depth per tool feed and requires increased flushing cycle. The efficiency
of spark decreases owing to the accumulation of debris, and conse-
quently machined area decreases resulting in a high taper.
Fig. 13 shows the effect of the capacitance and electrode material on
the overcut of micro holes at a discharge voltage of 90 V. The overcut
increases with the increase in capacitance for both electrodes although
the variation is less for brass electrodes and higher values of overcut are
obtained. The minimum overcut for tungsten is obtained at low capa-
citance and despite rapid variation, and the maximum overcut is less
than minimum obtained with the brass electrode. The high overcut with
the brass electrode is attributed to the low thermal conductivity of brass
electrode that results in increased heat absorption by the workpiece.
Fig. 14 depicts the variations in circularity error of micro holes with
capacitance and electrode material at a discharge voltage of 90 V. The
347
Measurement 125 (2018) 336–349
MRR (mm3/min) × 10−4 Overcut (µm) Taper Angle (°) Circularity (µm)
3.3482 6.9031 1.1751 3.0418
3.1036 6.6744 1.1625 2.6584
3.7464 7.0804 1.1794 3.4438
circularity error is almost constant for the tungsten electrode and is set
to lower values irrespective of the capacitance. Conversely, a significant
variation is observed for brass electrode, and the circularity error de-
creases with an increase in capacitance. The low circularity error for
tungsten electrode is owing to low tool wear rate and low MRR.
A total of 2000 generations are requested by using the full factorial
DOE matrix for the MOGA-II search engine. The characteristics of the
experimental runs (i.e., design points) obtained after optimization runs
are shown by using bubble charts in Fig. 15. Design points contained in
the DOE matrix are real while the values predicted from the response
surface are virtual. The unfeasible design points correspond to the va-
lues that violate the constraints in the optimization study as listed in
Table 6. The 3D bubble chart is plotted by using the design points re-
lative to the two objective functions (i.e., Ra and TWR) as shown in
Fig. 16. One of the output variables (i.e., MRR) is represented by the
diameter of the bubbles. Evidently, Ra increases with TWR and MRR in
an almost linear pattern. The bubbles distribution is traced by using two
line segments corresponding to two different electrodes as shown in
Fig. 15. The red bubbles correspond to the tungsten electrode, and the
blue bubbles correspond to the brass electrodes. The diameter of the
bubble denote the capacitance values. This clearly distinguishes the
performance of the two electrodes in terms of the selected output
M.H. Abidi et al.
responses. The design points associated with the brass electrode are
associated with high Ra, TWR, and MRR. The unfeasible design points
mostly occupy the lower left and top right corners of the bubble chart
and are characterized by low MRR values and poor micro-holes quality,
respectively.
The objective of the optimization problem involves minimizing both
Ra and TWR, and thus, the feasible design points corresponding to the
lower left corner of the bubble chart are candidates for the optimal
solution. The design point mark A is considered as a candidate for the
optimal solution along with the neighboring points B and C. However,
both points B and C are inferior to A with respect to Ra and TWR. The
design points correspond to tungsten electrode have low to moderate
capacitance and low discharge voltage values. The characteristics of
points A, B, and C are shown in Table 7.
An alternate way to analyze the performance of the two electrodes is
by using a parallel coordinate chart as shown in Fig. 16. A parallel
coordinate chart can show design points with all the parameters used in
the study. The chart is useful in visualizing all the input and output
variables simultaneously. It is used to verify the range of input variable
associated with a specific domain of a few output variables Evidently,
the tungsten electrode outperforms brass except in terms of the MRR,
and it is associated with low Ra and TWR in conjunction with improved
micro-holes quality resulting in low taper angles, overcuts, and circu-
larity errors.
5. Conclusions
In this study, the micro electric discharge machining of the Ni-Ti-
based SMA was optimized in terms of machining with brass and tung-
sten electrodes. The adopted methodology was based on radial basis
response surfaces and a new multi-objective genetic algorithm MOGA-
II. The results indicated that capacitance and electrode material are the
dominant factors that affect the micro EDM process. Good quality
micro-holes with low tool wear rates were observed with respect to the
tungsten electrode at low to moderate capacitance and low discharge
voltage values. In contrast, brass electrode exhibited a high material
removal rate at the expense of tool wear and micro-hole quality.
Acknowledgement
The authors are grateful to the Deanship of Scientific Research, King
Saud University for funding through Vice Deanship of Scientific
Research Chairs. Moreover, the authors thank the Deanship of Scientific
Research and RSSU at King Saud University for their technical support.
References
[1] E. Aliakbari, H. Baseri, Optimization of machining parameters in rotary EDM pro-
cess by using the Taguchi method, Int. J. Adv. Manuf. Technol. 62 (9–12) (2012)
1041–1053, http://dx.doi.org/10.1007/s00170-011-3862-9.
[2] A. Pandey, A. Dubey, Multiple quality optimization in laser cutting of difficult-to-
laser-cut material using grey–fuzzy methodology, Int. J. Adv. Manuf. Technol. 65
(1–4) (2013) 421–431, http://dx.doi.org/10.1007/s00170-012-4181-5.
[3] X. Xiao, K. Zheng, W. Liao, Theoretical model for cutting force in rotary ultrasonic
milling of dental zirconia ceramics, Int. J. Adv. Manuf. Technol. 75 (9–12) (2014)
1263–1277, http://dx.doi.org/10.1007/s00170-014-6216-6.
[4] A.M.A. Al-Ahmari, M.S. Rasheed, M.K. Mohammed, T. Saleh, A hybrid machining
process combining micro-EDM and laser beam machining of nickel–titanium-based
shape memory alloy, Mater. Manuf. Processes (2015) 1–9, http://dx.doi.org/10.
1080/10426914.2015.1019102.
[5] M. Jahan, Y. Wong, M. Rahman, A comparative experimental investigation of deep-
hole micro-EDM drilling capability for cemented carbide (WC-Co) against austenitic
stainless steel (SUS 304), Int. J. Adv. Manuf. Technol. 46 (9–12) (2010) 1145–1160,
http://dx.doi.org/10.1007/s00170-009-2167-8.
[6] M. Ay, U. Çaydaş, A. Hasçalık, Optimization of micro-EDM drilling of Inconel 718
superalloy, Int. J. Adv. Manuf. Technol. 66 (5) (2013) 1015–1023, http://dx.doi.
org/10.1007/s00170-012-4385-8.
[7] M.P. Jahan, M. Rahman, Y.S. Wong, A review on the conventional and micro-
electrodischarge machining of tungsten carbide, Int. J. Mach. Tools Manuf 51 (12)
(2011) 837–858, http://dx.doi.org/10.1016/j.ijmachtools.2011.08.016.
[8] A.K. Dubey, V. Yadava, Multi-objective optimization of Nd:YAG laser cutting of
348
Measurement 125 (2018) 336–349
nickel-based superalloy sheet using orthogonal array with principal component
analysis, Opt. Lasers Eng. 46 (2) (2008) 124–132, http://dx.doi.org/10.1016/j.
optlaseng.2007.08.011.
[9] M.P. Jahan, Y.S. Wong, M. Rahman, A study on the quality micro-hole machining of
tungsten carbide by micro-EDM process using transistor and RC-type pulse gen-
erator, J. Mater. Process. Technol. 209 (4) (2009) 1706–1716, http://dx.doi.org/
10.1016/j.jmatprotec.2008.04.029.
[10] M.P. Jahan, Y.S. Wong, M. Rahman, A study on the fine-finish die-sinking micro-
EDM of tungsten carbide using different electrode materials, J. Mater. Process.
Technol. 209 (8) (2009) 3956–3967, http://dx.doi.org/10.1016/j.jmatprotec.2008.
09.015.
[11] M. Murali, S.H. Yeo, Rapid biocompatible micro device fabrication by micro
electro-discharge machining, Biomed. Microdev. 6 (1) (2004) 41–45, http://dx.doi.
org/10.1023/B:Bmmd.0000013364.71148.51.
[12] M.S. Rasheed, M.H. Abidi, A.M. El-Tamimi, A.M. Al-Ahmari, Investigation of micro-
edm input parameters on various outputs in machining Ni-Ti shape memory alloy
Using full factorial design, Adv. Mater. Res. 816–817 (2013) 173–179.
[13] T. Masuzawa, State of the art of micromachining, CIRP Ann. - Manuf. Technol. 49
(2) (2000) 473–488, http://dx.doi.org/10.1016/S0007-8506(07)63451-9.
[14] Z.Y. Yu, T. Masuzawa, M. Fujino, Micro-EDM for three-dimensional cavities - de-
velopment of uniform wear method, CIRP Ann. - Manuf. Technol. 47 (1) (1998)
169–172, http://dx.doi.org/10.1016/S0007-8506(07)62810-8.
[15] I. Puertas, C.J. Luis, L. Álvarez, Analysis of the influence of EDM parameters on
surface quality, MRR and EW of WC–Co, J. Mater. Process. Technol. 153–154
(2004) 1026–1032, http://dx.doi.org/10.1016/j.jmatprotec.2004.04.346.
[16] T. Tadaki, K. Otsuka, K. Shimizu, Shape memory alloys, Annu. Rev. Mater. Sci. 18
(1) (1988) 25–45, http://dx.doi.org/10.1146/annurev.ms.18.080188.000325.
[17] S.A. Shabalovskaya, Surface, corrosion and biocompatibility aspects of Nitinol as an
implant material, Biomed. Mater. Eng. 12 (1) (2002) 69–109.
[18] N. Sharma, T. Raj, K. Jangra, Applications of Nickel-Titanium alloy, J. Eng. Technol.
5 (1) (2015) 1–7, http://dx.doi.org/10.4103/0976-8580.149472.
[19] K. Otsuka, T. Kakeshita, Science and technology of shape-memory alloys: New
developments, MRS Bull. 27 (02) (2002) 91–100, http://dx.doi.org/10.1557/
mrs2002.43.
[20] K. Otsuka, X. Ren, Physical metallurgy of Ti–Ni-based shape memory alloys, Prog.
Mater Sci. 50 (5) (2005) 511–678, http://dx.doi.org/10.1016/j.pmatsci.2004.10.
001.
[21] S.A. Thompson, An overview of nickel-titanium alloys used in dentistry, Int. Endod.
J. 33 (4) (2000) 297–310.
[22] A. Ishida, V. Martynov, Sputter-deposited shape-memory alloy thin films: properties
and applications, MRS Bull. 27 (2) (2002) 111–114, http://dx.doi.org/10.1557/
mrs2002.46.
[23] C. Mavroidis, Development of advanced actuators using shape memory alloys and
electrorheological fluids, Res. Nondestr. Eval. 14 (1) (2002) 1–32, http://dx.doi.
org/10.1007/s00164-001-0018-6.
[24] M.S. Alam, M.A. Youssef, M. Nehdi, Utilizing shape memory alloys to enhance the
performance and safety of civil infrastructure: a review, Can. J. Civ. Eng. 34 (9)
(2007) 1075–1086, http://dx.doi.org/10.1139/l07-038.
[25] J.Y. Xiong, Y.C. Li, X.J. Wang, P.D. Hodgson, C.E. Wen, Titanium–nickel shape
memory alloy foams for bone tissue engineering, J. Mech. Behav. Biomed. Mater. 1
(3) (2008) 269–273, http://dx.doi.org/10.1016/j.jmbbm.2007.09.003.
[26] A. Biesiekierski, J. Wang, M. Abdel-Hady Gepreel, C. Wen, A new look at biome-
dical Ti-based shape memory alloys, Acta Biomaterialia 8 (5) (2012) 1661–1669,
http://dx.doi.org/10.1016/j.actbio.2012.01.018.
[27] S.F. Hsieh, S.L. Chen, H.C. Lin, M.H. Lin, S.Y. Chiou, The machining characteristics
and shape recovery ability of Ti–Ni–X (X=Zr, Cr) ternary shape memory alloys
using the wire electro-discharge machining, Int. J. Mach. Tools Manuf. 49 (6)
(2009) 509–514, http://dx.doi.org/10.1016/j.ijmachtools.2008.12.013.
[28] K. Nakakita, T. Kondoh, S. Watanabe, A study on diesel combustion with high-
pressure fuel injection: Improvement of combustion and exhaust emissions using
small-hole-diameter nozzles, Trans. Jpn. Soc. Mech. Eng. B 60 (577) (1994)
3198–3206.
[29] D.M. Allen, A. Lecheheb, Micro electro-discharge machining of ink jet nozzles:
optimum selection of material and machining parameters, J. Mater. Process.
Technol. 58 (1) (1996) 53–66, http://dx.doi.org/10.1016/0924-0136(95)02107-8.
[30] T. Sato, T. Mizutani, K. Yonemochi, K. Kawata, The development of an electro-
discharge machine for micro-hole boring, Precis. Eng. 8 (3) (1986) 163–168, http://
dx.doi.org/10.1016/0141-6359(86)90035-8.
[31] Y.H. Jeong, B.-K. Min, Geometry prediction of EDM-drilled holes and tool electrode
shapes of micro-EDM process using simulation, Int. J. Mach. Tools Manuf 47
(12–13) (2007) 1817–1826, http://dx.doi.org/10.1016/j.ijmachtools.2007.04.011.
[32] S.H. Yeo, P.C. Tan, E. Aligiri, S.B. Tor, N.H. Loh, Processing of Zirconium-based
Bulk metallic glass (BMG) using micro electrical discharge machining (Micro-EDM),
Mater. Manuf. Processes 24 (12) (2009) 1242–1248, http://dx.doi.org/10.1080/
10426910903129661.
[33] T.T. Öpöz, B. Ekmekci, A. Erden, An experimental study on the geometry of mi-
croholes in microelectric discharge machining, Mater. Manuf. Processes 24 (12)
(2009) 1236–1241, http://dx.doi.org/10.1080/10426910903129422.
[34] B.B. Pradhan, B. Bhattacharyya, Modelling of micro-electrodischarge machining
during machining of titanium alloy Ti—6Al—4V using response surface metho-
dology and artificial neural network algorithm, Proc. Inst. Mech. Eng., Part B: J.
Eng. Manuf. 223 (6) (2009) 683–693, http://dx.doi.org/10.1243/
09544054JEM1343.
[35] G.S. Prihandana, M. Mahardika, M. Hamdi, Y.S. Wong, K. Mitsui, Effect of micro-
powder suspension and ultrasonic vibration of dielectric fluid in micro-EDM
processes—Taguchi approach, Int. J. Mach. Tools Manuf 49 (12–13) (2009)
M.H. Abidi et al.
1035–1041, http://dx.doi.org/10.1016/j.ijmachtools.2009.06.014.
[36] K.P. Somashekhar, N. Ramachandran, J. Mathew, Optimization of material removal
rate in micro-EDM using artificial neural network and genetic algorithms, Mater.
Manuf. Processes 25 (6) (2010) 467–475, http://dx.doi.org/10.1080/
10426910903365760.
[37] R. Phipon, B.B. Pradhan, Process parameters optimization of micro electric dis-
charge machining process using genetic algorithm, Int. J. Eng. Res. Appl. 2 (5)
(2012) 1986–1993.
[38] S.K.B. Jithin, J. Mathew, Multi-objective optimization of micro ED milling of Ti-6Al-
4V using genetic algorithm (ga), in: Paper presented at the International Conference
on Precision, Meso, Micro and Nano Engineering, NIT Calicut, Kerala, India, 13–15
December, 2013.
[39] H.K. Dave, V.J. Mathai, S. Kumar, K.P. Desai, H.K. Raval, Study on micro hole
accuracy and electrode depletion during micro EDM process through grey based
Taguchi approach, in: Paper Presented at the 3rd International Conference on
Production and Industrial Engineering (CPIE), Jalandhar, Punjab, India, 29–31
March, 2013.
[40] M. Manjaiah, S. Narendranath, S. Basavarajappa, V.N. Gaitonde, Wire electric dis-
charge machining characteristics of titanium nickel shape memory alloy, Trans.
Nonferrous Met. Soc. China 24 (10) (2014) 3201–3209, http://dx.doi.org/10.1016/
S1003-6326(14)63461-0.
[41] P. Sivaprakasam, P. Hariharan, S. Gowri, Modeling and analysis of micro-WEDM
process of titanium alloy (Ti–6Al–4V) using response surface approach, Eng. Sci.
Technol. Int. J. 17 (4) (2014) 227–235, http://dx.doi.org/10.1016/j.jestch.2014.
06.004.
[42] D.P. Babu, D.G. Dilip, K.P. Somashekhar, K. Allesu, J. Mathew, Numerical modeling
and multi-objective optimization of micro-wire EDM process, in: Paper Presented at
the 5th International & 26th All India Manufacturing Technology, Design and
Research Conference (AIMTDR), IIT Guwahati, Assam, India, 2014.
[43] P. Suresh, R. Venkatesan, T. Sekar, N. Elango, V. Sathiyamoorthy, Optimization of
intervening variables in microEDM of SS 316L using a genetic algorithm and re-
sponse-surface methodology, Strojniški vestnik – J. Mech. Eng. 60 (10) (2014)
656–664, http://dx.doi.org/10.5545/sv-jme.2014.1665.
[44] A.-H. Chiou, C.-C. Tsao, C.-Y. Hsu, A study of the machining characteristics of micro
EDM milling and its improvement by electrode coating, Int. J. Adv. Manuf. Technol.
349
Measurement 125 (2018) 336–349
78 (9) (2015) 1857–1864, http://dx.doi.org/10.1007/s00170-014-6778-3.
[45] B. Kuriachen, J. Mathew, Experimental investigations into the effects of micro-
electric-discharge milling process parameters on processing Ti–6Al–4V, Mater.
Manuf. Processes 30 (8) (2015) 983–990, http://dx.doi.org/10.1080/10426914.
2014.984206.
[46] M.P. Jahan, P. Kakavand, E.L.M. Kwang, M. Rahman, Y.S. Wong, An experimental
investigation into the micro-electro-discharge machining behaviour of aluminium
alloy (AA 2024), Int. J. Adv. Manuf. Technol. 78 (5) (2015) 1127–1139, http://dx.
doi.org/10.1007/s00170-014-6712-8.
[47] K. Wang, Q. Zhang, Q. Liu, G. Zhu, J. Zhang, Experimental study on micro electrical
discharge machining of porous stainless steel, Int. J. Adv. Manuf. Technol. (2016)
1–7, http://dx.doi.org/10.1007/s00170-016-9611-3.
[48] M.S. Rasheed, A.M. Al-Ahmari, A.M. El-Tamimi, M.H. Abidi, Analysis of influence
of micro-EDM parameters on MRR, TWR and Ra in machining Ni-Ti shape memory
alloy, Int. J. Recent Technol. Eng. 1 (4) (2012) 32–37.
[49] G. Kibria, B.R. Sarkar, B.B. Pradhan, B. Bhattacharyya, Comparative study of dif-
ferent dielectrics for micro-EDM performance during microhole machining of Ti-
6Al-4V alloy, Int. J. Adv. Manuf. Technol. 48 (5–8) (2010) 557–570, http://dx.doi.
org/10.1007/s00170-009-2298-y.
[50] S. Kumar, R. Singh, A. Batish, T.P. Singh, Investigations on effect of electrode po-
larity on machining performance of Ti-5Al-2.5Sn alloy using Electric Discharge
Machining Process, Adv. Mater. Res. 1137 (2016) 39–51, http://dx.doi.org/10.
4028/www.scientific.net/AMR.1137.39.
[51] T. Muthuramalingam, B. Mohan, A. Jothilingam, Effect of tool electrode re-
solidification on surface hardness in electrical discharge machining, Mater. Manuf.
Processes 29 (11–12) (2014) 1374–1380, http://dx.doi.org/10.1080/10426914.
2014.930956.
[52] Optimization MOGA-II, Mode Frontier Version 4.0, User Manual: Esteco, SPA, 2014.
[53] A. Giridharan, G.L. Samuel, Analysis on the effect of discharge energy on machining
characteristics of wire electrical discharge turning process, Proc. Inst. Mech. Eng.,
Part B: J. Eng. Manuf. 230 (11) (2015) 2064–2081, http://dx.doi.org/10.1177/
0954405415615732.
[54] S.S. Habib, Study of the parameters in electrical discharge machining through re-
sponse surface methodology approach, Appl. Math. Model. 33 (12) (2009)
4397–4407, http://dx.doi.org/10.1016/j.apm.2009.03.021.