Available via license: CC BY-NC 4.0
Content may be subject to copyright.
Journal of
Environmental
Research Article Nanotechnology
J. Environ. Nanotechnol.,Vol. 13(3), 197-206 (2024)
https://doi.org/10.13074/jent.2024.06.243848
Experimental Analysis of EDM Parameters on D2 Die
Steel Using Nano-aluminum Composite Electrodes
Y. Justin Raj1, A. Bovas Herbert Bejaxhin1*, R. Madhumitha2, A. S. Anitha3, Hitesh Gehani4,
Aslam Abdullah5 and V. Naveenprabhu6
1Department of Mechanical Engineering, Saveetha School of Engineering, Saveetha Institute of Medical and Technical Science (SIMATS),
Chennai, TN, India
2ECE Department, St. Joseph’s College of Engineering, Chennai, TN, India
3Department of Biotechnology, Karpaga Vinayaga College of Engineering and Technology, Padalam, TN, India
4School of Computer Science & Engineering, Shri Ramdeobaba College of Engineering and Management, Ramdeobaba University, Nagpur, MH,
India
5School of Chemical Engineering, Vellore Institute of Technology, Vellore, TN, India
6Department of Mechanical Engineering, Sri Eshwar College of Engineering, Coimbatore, TN, India
Received: 20.07.2024 Accepted: 25.09.2024 Published: 30.09.2024
*herbert.mech2007@gmail.com
ABSTRACT
This paper presents an experimental study on the electrical discharge machining (EDM) of AISI D2 die steel using an
Al-Ni composite electrode. The investigation focuses on the influence of input parameters like current, pulse duration, and
pulse interval, on key output parameters such as material removal rate (MRR), tool wear rate (TWR), and surface roughness
(SR). EDM oil was employed as the dielectric fluid. Grey relational analysis (GRA) was utilized for designing and
conducting the experiments using the Taguchi L9 method. The ANN model showed excellent predictive accuracy with a
perfect correlation coefficient (R) of 1.00, indicating strong capability in forecasting MRR based on machining parameters.
GRA further confirmed that higher current settings and longer pulse-off times effectively reduce tool wear, suggesting that
the ANN model accurately reflects the conditions that minimize TWR. The ANN model achieved strong predictive accuracy
for SR, with high correlation coefficients, although with slightly higher mean squared error (MSE) in testing.
Keywords: EDM machining; GRA; ANN; MRR; EWR; SR.
1. INTRODUCTION
The material removal rate (MRR) was greatly
influenced by the optimization of EDM process
parameters by (Jeykrishnan et al. 2018). Recent research
revealed that the applied current contributed 69.7% of the
variance in MRR, underscoring the importance of
accurate current management in improving EDM
performance (Anbuchezhiyan et al. 2022). Powder-
mixed EDM and Nanopowder-mixed EDM have been
developed as advanced methods for improving
machining performance in difficult-to-cut materials, such
as Al-Z-Mg composites reinforced with Si3N4,
particularly when using nickel-coated and uncoated brass
electrodes. The process of micro-hole machining, taking
into account variables like pulse on time, voltage, input
current, and capacitance, has demonstrated significant
effects on MRR and EWR, with SEM analysis providing
valuable insights into surface morphology alterations
(Jana et al. 2021). Taguchi’s L9 orthogonal array and
multi-criteria decision-making techniques, such as
TOPSIS and grey relational analysis, were used in this
study to improve die-sinking EDM settings for AISI D2
steel machining using a copper electrode. With canola oil
acting as the dielectric, the research found that the
optimal settings provided enhanced output responses.
These settings included a level 3 pulse on time, level 2
peak current, and level 2 servo voltage (Padhi et al. 2020)
to maximize wire EDM of EN-31 alloy steel. This
research used the grey-Taguchi technique, paying
particular attention to cutting rate, surface roughness, and
dimensional deviation. The study determined the ideal
process parameters using Taguchi's L27 orthogonal array
and grey relational analysis. A confirmatory test verified
the enhanced machining performance. Najm et al. (2023)
optimized the hybrid electrochemical discharge
machining process for tungsten carbide using grey
relation analysis and artificial neural networks, achieving
significant enhancements in MRR and machining depth.
Analysis revealed that electrolyte concentration and
current were the most influential factors, with a resulting
average surface roughness of 0.9275 µm, and
confirmation of surface composition was done with
Energy-dispersive X-ray spectroscopy. Paswan et al.
(2023) found that using a carbonated liquid as an
electrolytic solution in EDM significantly increased
discharge energy density and material removal rate
compared to EDM oil, primarily due to the liquid's higher
Y. Justin Raj
et al
. / J. Environ. Nanotechnol., Vol. 13(3), 197-206 (2024)
198
viscosity. Higher peak currents and optimized pulse-on-
time values enhanced MRR, although increased viscosity
led to higher micro-crack density.
Jeykrishnan et al. (2016) found that the
incorporation of tungsten powder into the dielectric
medium significantly enhanced the MRR in PMEDM of
D2 die steel. Optimization using taguchi techniques and
validation through confirmation experiments indicated
that adjusting the applied input factors further improved
MRR, with ANOVA highlighting the individual
contributions of these parameters. Naveen et al. (2023)
examined how solar energy may be integrated into
thermochemical processes for biomass conversion and
presented the techno-economic and life cycle
evaluations. Sharma et al. (2021) demonstrated that a
grey-fuzzy logic approach effectively optimized EDM
parameters for hexagonal hole formation in pearlite SG
iron, identifying peak current, pulse duration, and gap
settings as key factors. The optimal settings for improved
MRR and reduced overcut were established, and the
developed regression model for predicting responses
showed reliable performance with minimal prediction
error. Raza et al. (2018) found that current intensity
significantly affected output responses in EDM of
Al6061-SiC composites. Among the electrodes tested,
brass offered superior MRR and surface finish compared
to stainless steel, while stainless steel provided a lower
EWR than brass. Venkateswaran et al. (2023), based on
different coating techniques and nanoparticle
concentrations, examined how nanomaterial coatings
affect the performance of fin and tube condensers and
showed notable gains in heat transfer efficiency and
system effectiveness. Ansari et al. (2023) found that for
Wire-EDM of Al-10% SiC composites, pulse duration
and sparking voltage were crucial for optimizing MRR
and SR, while tool wear rate and spark gap were majorly
influenced by sparking voltage. The multi-variable
optimization identified the optimal parameters for high-
precision microchannel fabrication. With input factors
like Ip, Ton, and Toff, two neuro-fuzzy models and a
neural network model were created to predict MRR,
TWR, and radial overcut (G) in EDM for AISI D2 tool
steel (Pradhan et al. 2010). The models showed excellent
accuracy and useful prediction skills for these EDM
process responses when verified against experimental
data. Rizwee et al. (2019) reviewed the use of EDM for
machining reinforced metal matrix composites,
highlighting methodologies such as ANOVA, RSM, and
Taguchi for process optimization. The review identified
key process parameters affecting machining performance
and suggested future research directions in optimizing
EDM for MMCs. Nano-coating on electrodes
significantly enhances the EDM machinability of Inconel
825, with graphene and CNT coatings, improving MRR,
EWR, and surface roughness using silicon powder-mixed
deionized water as dielectric (George et al. 2021). Raj et
al. (2024) examined how different dielectric fluids and
nanocomposite electrodes were used in EDM, with an
emphasis on how they affected the machining of tougher
materials. Important technical features were analyzed,
highlighting improvements in EDM for challenging and
complicated applications. Using RSM and ANOVA for
parameter optimization, Y. Justin Raj et al. (2024)
examined the effects of Al-Ni Nano-electrodes on
surface roughness and machining time in Inconel 625
EDM. The findings provided valuable information for
improving EDM efficiency and surface quality, as they
showed that pulse-off time mostly impacted machining
time, while current primarily affected surface roughness.
Naveenprabhu et al. (2023) investigated how finned heat
exchangers with different cooling pads might improve
the performance of water chillers. Jute fiber, when
positioned between the fan and condenser, exhibits better
heat rejection and quicker chilling rates. Naveenprabhu
et al. (2023) based on experimental data collected in a
variety of climate settings, assessed the performance of
evaporative heat exchangers in a range of designs and
provided correlations for Sherwood and Nusselt
numbers. The dual strategy of hydrogen generation and
dye degradation using photocatalysis was reviewed,
emphasizing the need for improvements in hydrogen
storage and transportation as well as the efficacy of
different photo catalysts (Priyadharsini et al. 2023).
The goal of this study is to enhance the output
machining performance of Inconel 718, a hard super
alloy, by EDM. To optimize the process, experimental
trials were carried out using different discharge input
parameter settings. The Taguchi L9 with Grey Relational
Analysis technique was used. The best settings were
determined via the use of confirmation experiments. The
investigation showed that Inconel 718's EDM
performance was significantly enhanced, especially
when using a unique Cu-Ni-B4C composite electrode.
This resulted in lower wear, better surface polish, and
higher MRR. The intended function technique made it
easier to determine the ideal EDM factor values for the
least amount of electrode wear, and main effect graphs
helped identify the important elements driving electrode
wear. This work contributes significantly to the area of
advanced material processing by addressing important
issues in super alloy machining.
2. MATERIALS AND METHODS
2.1 Electrode Materials
The powder metallurgy process was used to
create the Al-Ni electrode, which is 15 mm in diameter
and 30 mm in length. This ensures a uniform combination
and the ideal qualities for EDM machining. The
procedure started with the selection of 10% by weight of
nickel powder, with an average particle size of 40
microns, and high-purity aluminum powder, 90% by
weight, with an average particle size of 50 microns. To
get a consistent dispersion, these powders were
meticulously weighed and combined in a ball mill
Y. Justin Raj
et al
. / J. Environ. Nanotechnol., Vol. 13(3), 197-206 (2024)
199
running at 200 rpm for 4 hours. The resultant mixture was
compressed to improve density and decrease porosity in
a die cavity at a pressure of 600 MPa using a uniaxial
hydraulic press. The compressed green body was then
carefully cooled to room temperature after being sintered
for two hours at 600 °C in an argon environment to avoid
oxidation. To achieve the required dimensions and
surface smoothness for efficient EDM application, the
sintered electrode was lastly machined. Fig. 2 shows the
electrode image.
Fig. 1: EDM setup
Fig. 2: EDM aluminum electrode tool
2.2 Work piece Material
D2 die steel, valued for its remarkable hardness
and wear resistance, was the material utilized to create
the work piece. A Sinker EDM EMS 350 machine, which
provides exact control over several process parameters,
was used to do the EDM machining. EDM oil's superior
cooling and flushing qualities made it suitable for use as
a dielectric fluid. The work piece measured 30 mm in
diameter and 10 mm in thickness as shown in Fig. 3. It
was aimed to use the powder metallurgy-fabricated
composite Al-Ni electrode to drill a 3 mm hole. The
electrode's particular size and material composition
allowed for regulated conditions to be met when
machining the D2 die steel efficiently and accurately.
Fig. 3: EDM D2 Die steel workpiece
2.3 Process Parameters
Three variables make up the EDM input factors:
pulse duration at 60, 90, and 120 µs, pulse interval at 10,
20, and 30 µs; and current (Ip) set at 4, 8, and 12 A. To
investigate the impact of these combinations on
machining performance, they were methodically
changed.
2.4 Output Responses
MRR and TWR were determined by measuring
the weight loss of the work piece and electrode with a
precision balance. Surface roughness was assessed using
a surface profilometer.
2.5 Grey Relational Analysis (GRA)
Grey relational analysis is a methodological tool
used in engineering and optimization research to evaluate
and improve the performance of complex systems. To get
the greatest quality features, several input factors are
optimized using GRA. When assessing or reviewing the
performance of a large project with little information,
grey relational analysis is often used. By assigning
weights to each answer, GRA may be used to determine
the ideal situation for multi-objective issues.
3. RESULTS AND DISCUSSION
Table 1 displays the findings of the EDM
experiment, emphasizing how roughness, wear rate, and
removal rate of material are affected by changes in input
factors.
Y. Justin Raj
et al
. / J. Environ. Nanotechnol., Vol. 13(3), 197-206 (2024)
200
Table 1. Output Response of EDM Experiment
Ip,
A
Pulse on
(Ton), µs
Pulse off
(Toff), µs
MRR
TWR
SR
4
60
10
0.1321
0.0444
3.342
4
90
20
0.3465
0.0611
4.434
4
120
30
0.7649
0.0523
5.523
8
60
20
0.3765
0.0869
3.501
8
90
30
1.1122
0.0742
6.609
8
120
10
0.2014
0.0378
5.736
12
60
30
0.7602
0.1639
5.341
12
90
10
0.5826
0.0539
5.525
12
120
20
0.5568
0.0339
6.251
min
0.1321
0.0339
3.342
max
1.1122
0.1639
6.609
The TWR varied from 0.0339 to 0.1639
mm3/min, the SR from 3.342 to 6.609 µm, and the MRR
from 0.1321 to 1.1122 mm3/min. These outcomes show
how the machining performance measures and the input
parameters are related.
The equation (1) is used for normalization in
Grey Relational Analysis (GRA). It transforms the
original data into a dimensionless value between 0 and 1,
facilitating comparison. This method is applied when
higher values are better, scaling the values relative to the
minimum and maximum within the dataset. The process
ensures that the lowest value becomes 0 and the highest
becomes 1, standardizing the data for further analysis.
Equation (2) normalizes data in Grey Relational
Analysis (GRA) where lower values are better. It scales
the original value
to a range between 0 and 1, with
the highest value becoming 0 and the lowest value
becoming 1. This standardization facilitates comparison
across different criteria by transforming the data into a
dimensionless form, ensuring uniformity in multi-criteria
decision-making processes.
Equation (3) calculates the absolute difference
between the normalized reference sequence and the
comparative sequence for the k-th criterion in GRA. It
quantifies the deviation of each alternative from the
reference, essential for computing the Grey relational
coefficient (GRC). Smaller differences indicate closer
similarity to the reference sequence, aiding in
performance comparison and ranking. The grey
relational coefficient correlation between the
comparative and reference sequences is shown in Table
2. The deviation sequences and normalized values are
used in the calculation.
Table 2. Normalized and Deviation GRA analysis
NORMALIZED VALUE
DEVIATION
MRR
EWR
SR
MRR
EWR
SR
0.0000
0.0808
0.0000
1.0000
0.9192
1.0000
0.2188
0.2092
0.3343
0.7812
0.7908
0.6657
0.6456
0.1415
0.6676
0.3544
0.8585
0.3324
0.2494
0.4077
0.0487
0.7506
0.5923
0.9513
1.0000
0.3100
1.0000
0.0000
0.6900
0.0000
0.0707
0.0300
0.7328
0.9293
0.9700
0.2672
0.6409
1.0000
0.6119
0.3591
0.0000
0.3881
0.4596
0.1538
0.6682
0.5404
0.8462
0.3318
0.4333
0.0000
0.8904
0.5667
1.0000
0.1096
0
0
0
0
0
0
1
1
1
1
1
1
Equations (4) and (5) define the maximum and
minimum Grey relational differences in grey relational
analysis. Equation (4), identifies the greatest absolute
difference across all criteria k and all alternatives j,
establishing the upper bound of the differences.
Conversely, Equation (5), determines the smallest
absolute difference, setting the lower bound. These
bounds are crucial for calculating the Grey relational
coefficient, as they normalize the differences, ensuring
comparability across different criteria and alternatives.
Equation (6), calculates the Grey relational
coefficient in grey relational analysis. Here, and
are the minimum and maximum Grey relational
differences, respectively while represents the
absolute difference between the reference and
comparative sequences for the k-th criterion. The
coefficient ζ is a distinguishing coefficient, typically set
between 0 and 1, used to adjust the sensitivity of the
analysis. This equation normalizes the differences,
providing a relative measure of how closely each
alternative matches the reference sequence, with higher
values indicating a closer relationship.
Equation (7) calculates Grey relational grade
(GRG) in GRA Here, ζi(k) is the Grey relational
coefficient for the i-th alternative and k-th criterion, and
Y. Justin Raj
et al
. / J. Environ. Nanotechnol., Vol. 13(3), 197-206 (2024)
201
n is the total number of criteria. The Grey relational grade
γi is the average of the GRCs across all criteria, providing
an overall measure of the performance of each alternative
relative to the reference sequence. Higher values of γi
indicate better overall performance and closer similarity
to the reference sequence. This metric is essential for
ranking and selecting the best alternative among multiple
options in multi-criteria decision-making processes.
Table 3 illustrates the grey relational coefficient
values for output responses for 9 experiments, alongside
their corresponding grey relational grades and rankings.
The GRC values for MRR range from 0.3333 to 1.0000,
for EWR from 0.3333 to 1.0000, and for SR from 0.3333
to 1.0000. The GRG, which is the mean of these
coefficients, ranges from 0.3397 to 0.8067. Experiment
5, with a GRG of 0.8067, ranks highest, indicating
superior overall performance, whereas experiment 1,
with a GRG of 0.3397, ranks lowest. This evaluation
highlights the most effective experimental setup by
integrating the performance metrics of MRR, EWR, and
SR.
Table 3. Gray relation coefficient GRA analysis
MRR
Gray Relation Coefficient (GRC)
EWR
SR
GRG
GR RANK
0.3333
0.3523
0.3333
0.3397
9
0.3902
0.3874
0.4289
0.4022
7
0.5852
0.3681
0.6007
0.5180
4
0.3998
0.4577
0.3445
0.4007
8
1.0000
0.4202
1.0000
0.8067
1
0.3498
0.3401
0.6517
0.4472
6
0.5820
1.0000
0.5630
0.7150
2
0.4806
0.3714
0.6011
0.4844
5
0.4687
0.3333
0.8202
0.5408
3
3.1 ANN analysis of MRR
The ANN model demonstrated high accuracy in
predicting the MRR of D2 die steel machined by EDM
with an aluminum composite electrode, as evidenced by
low mean squared error values and high correlation
coefficients (R). The model effectively captured the
impact of machining parameters on MRR, indicating its
potential for optimizing EDM processes. The analysis at
epoch 4 shows (Fig. 4) a gradient of
2.5344×10−122.5344 \times 10^{-12}2.5344×10−12,
and zero validation checks, indicating highly stable and
well-converged training at this early stage.
Fig. 4: ANN analysis of MRR
Fig. 5 displays a line graph that demonstrates
the performance of a machine learning model across four
epochs. The number of epochs is represented by the
horizontal axis, while the vertical axis represents the
mean squared error on a logarithmic scale. Three distinct
data sets' MSE values are shown versus the number of
epochs in the graph:
• Train (blue line): The MSE on the training data is
shown here, showing how well the model can match
the training set over time.
• Validation (green line): To keep an eye on the
model's performance on untested data and avoid
over fitting, this line displays the MSE on the
validation set.
• Test (red line): After training, the model's
performance is assessed by this line, which displays
the MSE on the test set of data.
• Top Validation Results: The best validation
performance is 0.0011946 at epoch 4. This indicates
that at this stage, the model may have reached its
lowest validation error. With the use of this graph,
one may examine the model's performance across
the epochs and determine the optimal location on
the validation set.
The histogram image (Fig. 6) shows the
distribution of prediction errors from an ANN model,
indicating the frequency of errors within specific ranges.
It provides insight into the model's accuracy and
performance by highlighting the concentration of errors
around the mean value.
Y. Justin Raj
et al
. / J. Environ. Nanotechnol., Vol. 13(3), 197-206 (2024)
202
Fig. 5: ANN best validation performance analysis of MRR
MRR, TWR, and SR were assessed for nine
experimental runs in the study's ANN analysis of MRR
using Minitab. The training (6 samples), validation (2
samples), and testing (2 samples) sets of the dataset were
separated. The correlation coefficient and mean squared
error were used to assess the ANN model's performance.
A perfect linear connection between the predicted and
real MRR values was shown by the training, validation,
and testing MSE values of 0.0000, 0.0012, and 0.3289,
respectively, in Fig. 7. These results showed a high level
of model accuracy with an R-value of 1.00. Significant
differences in MRR were found to occur when machining
settings were altered, with the lowest and greatest MRR
values being recorded at 0.1321 and 1.1122, respectively.
Fig. 6: ANN error histogram analysis of MRR
3.2 ANN analysis of EWR
This study explores the application of artificial
neural networks to predict the tool wear rate of D2 die
steel machined by electrical discharge machining using
an aluminum composite electrode. The objective is to
improve the accuracy and reliability of TWR predictions
in advanced EDM operations. The ANN TWR analysis
at epoch 4 (Fig. 8) shows a gradient value and three
validation checks, indicating moderate convergence and
stability with early signs of overfitting.
Fig. 7: ANN training and validation analysis of MRR
Fig. 8: ANN analysis of EWR
Fig. 9: ANN best validation performance analysis of EWR
Fig. 9 displays a training plot for a neural
network showing MSE against epochs. The blue line
represents training error, which decreases significantly
over 4 epochs. The green and red lines represent
validation and test errors, which remain relatively stable.
Y. Justin Raj
et al
. / J. Environ. Nanotechnol., Vol. 13(3), 197-206 (2024)
203
The best validation performance (0.0016387 MSE) is
achieved at epoch 1, highlighted by a green circle.
In an artificial neural network analysis of tool
wear rate with a histogram having 20 bins and zero error,
shown in Fig. 10, the model accurately predicts TWR,
showing no discrepancies between predicted and actual
values. The zero error indicates perfect model
performance and precise bin classification.
Fig. 10: ANN error histogram analysis of EWR
Fig. 11: ANN training and validation analysis of EWR
In the ANN analysis of wear rate using
MATLAB, the model uses input parameters from a
dataset of nine experiments. The performance metrics for
the model include training, validation, and testing MSE
values of 0.0001, 0.0016, and 0.0018, respectively. The
training and testing correlation coefficients are both 1.00,
indicating perfect linear relationships, while the
validation R is 0.8074, showing a strong but slightly less
perfect correlation. These results suggest that the model
accurately predicts TWR with high precision, capturing
the complex relationships between the input parameters
and TWR. The TWR values in the dataset range from a
minimum of 0.0339 to a maximum of 0.1639,
highlighting the variability in tool wear based on
different machining conditions (Fig. 11).
Fig. 12: ANN analysis of SR
Fig. 13: ANN best validation performance analysis of SR
3.3 ANN analysis of SR
This study investigates the use of ANN to
predict the surface roughness of D2 die steel machined
by EDM using an aluminum composite electrode. The
analysis aims to enhance the precision and efficiency of
surface quality predictions in advanced machining
processes.The ANN SR analysis at epoch 4 shows a
gradient value and three validation checks, indicating
reasonable convergence with potential overfitting
concerns at this training stage in Fig. 12. The graph,
shown in Fig. 13, illustrates the MSE over 4 epochs for
training, validation, and test datasets. The blue line,
representing training error, decreases significantly. The
green and red lines show relatively stable validation and
test errors. The best validation performance (0.12908
MSE) occurs at epoch 1, highlighted by a green circle.
Fig. 14 shows an ANN analysis of surface
roughness with a histogram having 20 bins and zero
Y. Justin Raj
et al
. / J. Environ. Nanotechnol., Vol. 13(3), 197-206 (2024)
204
error. The model accurately predicts SR, showing no
discrepancies between predicted and actual values. The
zero error indicates perfect model performance and
precise bin classification.
Fig. 14: ANN error histogram analysis of EWR
Fig. 15: ANN training and validation analysis of EWR
In Fig. 15 the model uses input parameters from
a dataset of different experimental runs in the ANN
analysis of roughness using MATLAB. The dataset has a
range of SR values, from 3.342 at the least to 6.609 at the
highest. Training, validation, and testing datasets are
used to assess the model's performance; the
corresponding MSE values are 0.1608, 0.1291, and
2.5913, respectively. For training, validation, and testing,
the corresponding correlation coefficients are 0.9775,
1.00, and 1.00, respectively. According to these findings,
the ANN model predicts SR with a high degree of
accuracy and a strong linear connection between the
predicted and actual values. The slightly higher MSE
throughout the testing phase indicates that, overall, the
model has excellent surface roughness prediction skills
based on the supplied machining parameters, however
there may be some fluctuation in the model's
performance.
4. CONCLUSION
Based on the combined ANN and GRA of EDM
machining of D2 die steel with an aluminum composite
electrode, the following conclusions were drawn: Based
on systematic testing and analysis, pulse on time was
shown to be the key influencing factor for surface
roughness, followed by pulse off time and current.
• Material removal rate (MRR): The ANN model
showed excellent predictive accuracy with a perfect
correlation coefficient (R) of 1.00, indicating strong
capability in forecasting MRR based on machining
parameters. Complementarily, GRA highlighted
that higher current and pulse duration optimize
MRR, underscoring the model's effectiveness in
capturing the impact of these parameters on
machining efficiency. The research focused on
current as the most important parameter in terms of
machining time reduction, highlighting its
relevance in reducing the amount of time needed to
mill Inconel 625.
• Tool wear rate (TWR): ANN results demonstrated
high precision in predicting TWR, with perfect
correlation coefficients (R) and minimal MSE
values, reflecting reliable model performance. GRA
further confirmed that higher current settings and
longer pulse-off times effectively reduce tool wear,
suggesting that the ANN model accurately reflects
the conditions that minimize TWR.
• Surface roughness (SR): The ANN model achieved
strong predictive accuracy for SR, with high
correlation coefficients, although with slightly
higher MSE in testing. GRA results corroborated
that optimal SR is achieved with balanced
parameters, validating the model's capability in
predicting surface quality and highlighting the
importance of parameter optimization for improved
surface finish.
FUNDING
This research received no specific grant from
any funding agency in the public, commercial, or not-for-
profit sectors.
CONFLICTS OF INTEREST
The authors declare that there is no conflict of
interest.
COPYRIGHT
This article is an open-access article distributed
under the terms and conditions of the Creative Commons
Y. Justin Raj
et al
. / J. Environ. Nanotechnol., Vol. 13(3), 197-206 (2024)
205
Attribution (CC BY) license
(http://creativecommons.org/licenses/by/4.0/).
REFERENCES
Anbuchezhiyan, G., Saravanan, R., Pugazhenthi, R.,
Palani, K. and Mamidi, V. K., Influence of Coated
Electrode in Nanopowder Mixed EDM of Al–Zn–
Mg–Si3N4 Composite, Adv. Mater. Sci. Eng., 2022,
1–11 (2022).
https://doi.org/10.1155/2022/9539790
Ansari, M. and Khan, I. A., Investigation on the
performance of wire electrical discharge machining
(WEDM) using aluminium matrix composites
(AMCs) micro-channel, Eng. Res. Express, 5(3),
035065 (2023).
https://doi.org/10.1088/2631-8695/acf5ca
George, J., Chandan, R., Manu, R. and Mathew, J.,
Experimental Investigation of Silicon Powder Mixed
EDM Using Graphene and CNT Nano Particle
Coated Electrodes, Silicon, 13(11), 3835–3851
(2021).
https://doi.org/10.1007/s12633-020-00658-0
Jana, A. K., Ranjith Kumar, R., Mohanty, S. C.,
Mangapathi Rao, K., Shanker, V. G. and Reddy, A.
Y., Parametric Optimization of Die Sinking EDM in
AISI D2 Steel considering Canola oil as Dielectric
using TOPSIS and GRA, IOP Conf. Ser. Mater. Sci.
Eng., 1057(1), 012061 (2021).
https://doi.org/10.1088/1757899X/1057/1/012061
Jeykrishnan, J., Vijaya Ramnath, B., Chenbaga Ram, N.,
Naveen Babu, R. and Naveen, B., Experimental
investigation on powder mixed electro - discharge
machining (EDM) of D2 die steel with De-ionized
water as dielectric medium, IOP Conf. Ser. Mater.
Sci. Eng., 402, 012084 (2018).
https://doi.org/10.1088/1757-899X/402/1/012084
Jeykrishnan, J., Vijaya Ramnath, B., Jude Felix, A.,
Rupan Pernesh, C. and Kalaiyarasan, S., Parameter
Optimization of Electro-Discharge Machining
(EDM) in AISI D2 Die Steel using Taguchi
Technique, Indian J Sci Technol.
https://doi.org/10.17485/ijst/2016/v9i43/101972
Najm, V., Abbas, T. and Aghdeab, S., Integrating Grey
Relation Analysis and Artificial Neural Networks for
Optimal Machining of Tungsten Carbide Composite
Using Hybrid Electrochemical Discharge, Eng.
Technol. J., 41(12), 1594–1610 (2023).
https://doi.org/10.30684/etj.2023.143347.1581
Naveen, S., Aravind, S., Yamini, B., Vasudhareni, R.,
Gopinath, K. P., Arun, J. and Pugazhendhi, A., A
review on solar energy intensified biomass
valorization and value-added products production:
Practicability, challenges, techno economic and
lifecycle assessment, J. Clean. Prod., 405, 137028
(2023).
https://doi.org/10.1016/j.jclepro.2023.137028
Naveenprabhu, V. and Suresh, M., Performance studies
on a water chiller equipped with natural fiber cooling
pad based evaporative condenser, Ind. Crops Prod.,
201, 116923 (2023).
https://doi.org/10.1016/j.indcrop.2023.116923
Padhi, P. C. and Tripathy, D. K., Multi-response
optimisation of machining parameters in wire EDM
process using grey relational analysis, Int. J. Manuf.
Technol. Manag., 34(4), 376 (2020).
https://doi.org/10.1504/IJMTM.2020.108019
Paswan, K., Pramanik, A., Chattopadhyaya, S., Sharma,
S., Singh, G., Khan, A. M. and Singh, S., An Analysis
of Machining Response Parameters, Crystalline
Structures, and Surface Topography During EDM of
Die-Steel Using EDM Oil and Liquid-Based Viscous
Dielectrics: A Comparative Analysis of Machining
Performance, Arab. J. Sci. Eng., 48(9), 11941–11957
(2023).
https://doi.org/10.1007/s13369-023-07626-x
Pradhan, M. K. and Biswas, C. K., Neuro-fuzzy and
neural network-based prediction of various responses
in electrical discharge machining of AISI D2 steel,
Int. J. Adv. Manuf. Technol., 50(5–8), 591–610
(2010).
https://doi.org/10.1007/s00170-010-2531-8
Priyadharsini, P., SundarRajan, P., Pavithra, K. G.,
Naveen, S., SanjayKumar, S., Gnanaprakash, D.,
Arun, J. and Pugazhendhi, A., Nanohybrid
photocatalysts in dye (Colorants) wastewater
treatment: Recent trends in simultaneous dye
degradation, hydrogen production, storage and
transport feasibility, J. Clean. Prod., 426, 139180
(2023).
https://doi.org/10.1016/j.jclepro.2023.139180
Raj, Y. J., Bejaxhin, A. B. H. and Rajkumar, S., Review
about removal rates and wear rate of EDM using nano
composite electrodes with variant electrolytic
solutions, In: (2024).
https://doi.org/10.1063/5.0197422. p 020249
Raza, M. H., Wasim, A., Ali, M. A., Hussain, S. and
Jahanzaib, M., Investigating the effects of different
electrodes on Al6061-SiC-7.5 wt% during electric
discharge machining, Int. J. Adv. Manuf. Technol.,
99(9–12), 3017–3034 (2018).
https://doi.org/10.1007/s00170-018-2694-2
Y. Justin Raj
et al
. / J. Environ. Nanotechnol., Vol. 13(3), 197-206 (2024)
206
Rizwee, M., Minz, S. S., Md. Orooj, Hassnain, M. Z. and
Khan, M. J., Electric Discharge Machining Method
for various Metal Matrix Composite Materials, Int. J.
Innov. Technol. Explor. Eng., 8(9), 1796–1800
(2019).
https://doi.org/10.35940/ijitee.I8112.078919
Sharma, A., Kumar, V., Babbar, A., Dhawan, V.,
Kotecha, K. and Prakash, C., Experimental
Investigation and Optimization of Electric Discharge
Machining Process Parameters Using Grey-Fuzzy-
Based Hybrid Techniques, Materials (Basel).,
14(19), 5820 (2021).
https://doi.org/10.3390/ma14195820
Venkateswaran, N., Subbaiyan, N., Punniyakotti
Varadharajan, G., Vellingiri, S., Asary, A. R., Giri, J.
M. and Petchimuthu, P., A comprehensive review of
the energy efficiency on nano coated fin and tube
condenser, Environ. Qual. Manag., 33(1), 293–301
(2023).
https://doi.org/10.1002/tqem.22055
Y. Justin Raj, A. Bovas Herbert Bejaxhin, S. Raj Kumar
and G. M. Balamurugan, Influence of Novel Al-Ni
Electrodes on Roughness and Machining Time in
Inconel 625 EDM Machining: A Comprehensive
RSM Performance Analysis, J. Environ.
Nanotechnol., 13(1), 243–253 (2024).
https://doi.org/10.13074/jent.2024.03.241547
