Professional Documents
Culture Documents
Comparison of Intelligent Optimization Algorithms For Wire Electrical Discharge Machining Parameters
Comparison of Intelligent Optimization Algorithms For Wire Electrical Discharge Machining Parameters
Soheil Gohari
Amran Ayob
Email: soheil.gohari7@gmail.com
Email: amran@fkm.utm.my
Isfahan, Iran
Email: gabolfazl@gmail.com
I. INTRODUCTION
These days, with advent of new material exploitation
and era of technological advancement, particularly in
manufacturing industries, the use of machining and
methods of shaping the components has necessarily been
increasing. Meanwhile, methods of machining concerning
wire electrical discharge machining (WEDM) have high
importance and increasingly high applications in
industries. WEDM process is one of EDM machining in
which by creation of alternative spark between tool (wire)
and piece work, machining process of piece work is
implemented. In WEDM process, it is vital to choose best
machining parameters to economize choosing process
whereas WEDM is a nonconventional, applicable and
required machining process with high initial investment.
134
60 l
t
(1)
VMRR = F Dw H
(2)
TABLE 1.
MACHINING PARAMETERS AND THEIR LEVELS
Levels
Control
Parameters
Unit
Symbol
Current
[A]
Pulse-Off Time
[S]
Toff
14
10
22
Volt
Servo
[volt]
[volt]
V
S
140
28
130
32
110
30
120
26
1
11
2
10
3
9
4
12
TABLE 2.
EXPERIMENTAL RESULTS
S.No.
I
[A]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
11
11
11
11
10
10
10
10
9
9
9
9
12
12
12
12
Toff
[S]
14
10
8
22
14
10
8
22
14
10
8
22
14
10
8
22
VMRR
V
[volt]
S
[volt]
Ra
[S]
[mm3 / min]
140
130
110
120
130
140
120
110
110
120
140
130
120
110
130
140
28
32
30
26
30
26
28
32
26
30
32
28
32
28
26
30
3.169
3.410
3.229
2.707
3.018
3.035
3.046
2.754
2.421
2.845
3.270
2.575
3.462
3.293
3.528
3.408
12.555
13.785
12.645
7.995
11.108
12.255
10.913
6.698
7.725
8.205
9.593
6.000
11.033
12.435
14.288
8.250
135
TABLE 3.
FITTING EQUATIONS WITH THEIR CORRELATION PERCENTAGE
Respond Value
Model Type
Volumetric
Second-order
Material Removal
polynomial
Rate
equation
Surface
Roughness
Correlation
Fitting Equation
(%)
0.0319
Toff2
+ 0.19
V + 0.00188
Toff
S - 0.00220VS
I2-
S2-
0.1I
Toff
99.99
Toff
Toff
Exponential
V-0.00949 I S + 0.000606 Toff V + 0.00125 Toff S + 0.000078 V S )
TABLE 4.
RESULTS OF CONFORMATION TEST FOR Ra
I
[A]
Toff
10
12
Run
V
[volt]
S
[volt]
Results of Model
Results of Experiments
Error (%)
120
28
2.98
3.05
-2.3
14
120
32
3.33
3.46
-3.76
[S]
TABLE 5.
RESULTS OF CONFORMATION TEST FOR VMRR
I
[A]
Toff
10
12
Run
V
[volt]
S
[volt]
Results of Model
Results of Experiments
Error (%)
120
28
10.67
10.91
-2.2
14
120
32
12.21
11.03
9.66
[S]
136
(a)
(b)
Fig.1. Normal plot of residuals: (a). Normal plot of residuals for average surface roughness (b). Normal plot of residuals for volumetric material removal rate
V. MULTI-OBJECTIVE OPTIMIZATION
In this study, two objectives are put into consideration,
volumetric material removal rate (VMRR) and surface
roughness (Ra). It is noticed that if VMRR rises, Ra also
increases. But our objective is aimed at maximizing
VMRR and minimizing surface roughness. A single
optimal solution does not help to achieve our goals as our
purpose, since these objectives are opposing in nature.
Choice of VMRR and surface roughness is also dependent
on user and environment of the problem. Optimizing both
of the output parameters requires multi-objective
optimization [6].
TABLE 6.
RESULTS OF VARIANCE ANALYSIS ON VOLUMETRIC MATERIAL REMOVAL RATE
Factors
d.f.
Sum of Squares
Variance
Distribution Percentage
37.52164687
12.50721562
149.92
37.13
Toff
55.13897813
18.37965938
220.31
54.56
7.51227188
2.50409062
31.01
7.43
0.62510625
0.20836875
2.5
0.08
Error
0.2502844
0.0834281
0.8
Total
15
101.0482875
100
TABLE 7.
RESULTS OF VARIANCE ANALYSIS ON SURFACE ROUGHNESS
Factors
d.f.
sum of squares
Variance
Distribution percentage
0.89816618
0.29938873
23538.8
53.79
Toff
0.36541701
0.12180507
9576.7
21.88
0.20304333
0.06768111
5321.28
12.16
0.20300948
0.06766983
5320.39
12.15
Error
0.00003816
0.00001272
0.02
Total
15
1.66967416
100
137
(3)
138
SC
SC
F1 F1 ( x )
(5)
F1 ( x )
F2 ( x) F
F2 ( x )
(6)
Z = w1 F1SC + w2 F2SC
(7)
TABLE 8.
OPTIMUM MASHINING PARAMETERS ACQUIRED USING NSGA-II AND TABU SEARCH ALGORITHM
Level of Machining
Rough Machining
Semi-Polishing
Finish Polishing
Algorithm Type
Current (A)
Pulse-Off
Time (S)
Volt (v)
Servo (v)
Ra(m)
VMRR (mm3/min)
NSGA-II
11.94
110
26
3.60
14.80
Tabu Search
12
110
26
3.62
14.80
NSGA-II
10.97
13.22
140
26
3.22
14.21
Tabu Search
11
14
140
26
3.22
14.19
NSGA-II
22
110
26
2.05
3.39
Tabu Search
22
110
26
2.05
3.39
IX.CONCLUSION
As follows from forgoing analysis, the study based on
the influence of WEDM on surface roughness and
volumetric material removal rate was carried out. The
nonlinear polynomial models were developed for
volumetric material removal rate and average surface
roughness were used for optimization. In this study, two
multi-objective evolutionary algorithms concerning
efficient methodology including NSGA-II and Tabu
search algorithm were used to optimize machining
parameters in cold-work steel 2601.
The emphasis must be put on providing a preferred
solution for the process engineer in the short period of the
time. The choice of one solution over other ones is
dependent on the requirements of process engineer [13].
140