Vikas Gohil
Y. M. Puri
DOI: 10.21278/TOF.40401
ISSN 1333-1124
eISSN 1849-1391
EXPERIMENTAL INVESTIGATION ON SURFACE ROUGHNESS IN
ELECTRICAL DISCHARGE TURNING OF Ti-6Al-4V ALLOY
Summary
This study presents a novel EDM turning process specially designed and developed to
generate precise cylindrical forms on hard and difficult-to-machine materials. For this
purpose, a specially designed turning spindle is used. The spindle was mounted on a
conventional die-sinking EDM machine to rotate the workpiece. Axially symmetric parts can
be manufactured by feeding the shaped tool into the rotating workpiece. In this way an
axisymmetric workpiece can be made with small tools at both macro and micro levels. Effects
of machining parameters, such as the current, pulse-on time, rotational speed, flushing
pressure, and duty factor, on the surface roughness of Ti-6Al-4V alloy in electrical discharge
turning were investigated. Taguchi's design of experiment technique was used. Analysis of
variance and regression analysis were performed on experimental data. The signal-to-noise
ratio analysis was employed to find the optimal condition.
Key words: Electrical discharge machining, Electrical discharge turning, Surface
roughness, Axisymmetric workpieces
1. Introduction
Electrical discharge machining (EDM) is an electro-thermal process where material
removal is achieved by electrical discharges occurring between the tool and the workpiece
[1]. Among various EDM processes, the die-sinking EDM is used extensively to machine
complex cavities in full hardened and stabilized die steels, to avoid the problem of
dimensional variability [2]. Small cylindrical parts made of advanced conductive materials,
which are difficult to machine by a conventional cutting methods such as milling and turning,
can be produced by an electrical discharge machine. The EDM process can be improved by
setting up an external rotary axis on a conventional electrical discharge machine and by
feeding the tool electrode against the rotating workpiece in order to produce axisymmetric
geometries. This improved process named electrical discharge turning (EDT) can be utilized
for producing cylindrical forms and helical profiles on high-strength, temperature resistant
(HSTR) alloys widely used in the aeronautical and automotive industries. The concept of
EDM turning is illustrated in Fig.1.
TRANSACTIONS OF FAMENA XL-4 (2016)
1
Vikas Gohil, Y. M. Puri Experimental Investigation on Surface Roughness in
Electrical Discharge Turning of Ti-6Al-4V Alloy
Fig. 1 Concept of EDT process Fig. 2 Experimental set up of the EDT process
The EDT process has been developed from wire electrical discharge turning (WEDT) in
order to achieve a high degree of accuracy, surface finish, and cylindricity in the straight
turning and in the form turning of a hard-to-machine material with high slenderness ratio. Fig.
2 shows a precise rotary spindle setup which was added to the conventional EDM machine to
give rotary motion to the workpiece. A die-sinking EDM machine is widely used for turning
aerospace honey-comb seals where internal and external seals are machined to a burr-free
finish with internal or external diameters from 300mm-1400mm [3-6]. An example of a
machined part produced by the EDT method is shown in Fig. 3. The surface produced by
EDM is matte finish and complicated. The surface integrity explains the mechanical,
metallurgical, chemical, and topological conditions of the surface region, which in turn
depends on the machining parameters such as the applied current, voltage, and pulse duration.
As surface roughness (Ra) is one of the most important parameters in manufacturing,
various investigations have been carried out considering various input parameters in order to
improve the surface roughness of the part. Soni and Chakraverti [7] performed an
experimental study on the drilling operation performed on a rotary EDM. Titanium and
copper-tungsten were chosen as the workpiece and the tool material, respectively. The authors
found that rotation of the tool electrode improves the material removal rate. However, this
results in high surface roughness. The idea of employing EDM to machine cylindrical parts
for manufacturing small diameter pins was put forward by doctor Masuzawa at the University
of Tokyo [8]. Small pins can be utilized as the tool in the micro EDM application [9]. Guu et
al. [10] study the influence of workpiece rotation during electrical discharge machining of the
AISI D2 tool steel using a copper electrode. Experimental findings indicate that surface finish
improves with an increase in the workpiece rotational speed. Rohney et al. [11] presented the
application of cylindrical wire electrical discharge turning (CWEDT) for the profile turning of
metal bond diamond wheels. Using a mathematical model, Jun et al. [12] investigated the
surface integrity of a wire EDM turned part. Using statistical analysis, Mohammadi et al. [13]
investigated the effect of machining parameters on the surface finish and roundness of
cemented steel in wire electrical discharge turning. Investigation results demonstrate that
voltage, power, time-off, and spindle speed have the strongest effect on the surface roundness
Ra. As the values of power and voltage increase and those of spindle speed and time-off
decrease, Ra increases. Haddad et al. [14] used the response surface methodology approach to
investigate experimentally the surface roughness of the AISI D3 tool steel in the CWEDT
process. Janardhan et al. [15] used the pulse train data analysis to study the effect of process
parameters on the material removal, surface roughness, and roundness. Weingartner et al. [16]
investigated the WEDM process on high speed rotating workpieces and found that the volume
of the eroded craters increases with an increase in the relative speed.
2
TRANSACTIONS OF FAMENA XL-4 (2016)
Experimental Investigation on Surface Roughness in Vikas Gohil, Y. M. Puri
Electrical Discharge Turning of Ti-6Al-4V Alloy
Although extensive research has been done on the electrical discharge machining of
various materials and on optimizing EDM parameters, no research, to the best of our
knowledge, has been reported on electrical discharge machining of titanium in the context of
turning. In view of that fact, the surface roughness of a part turned by means of the
conventional EDM process is investigated in this study and possible ways of adjusting the
process parameters to achieve better surface finish are explored using statistical methods. Five
machining parameters were considered to study the effect of the EDM process on surface
roughness. In the present research, Taguchi's design of experiment technique was employed to
perform more accurate and less costly experiments. Three different analyses were performed
on the data obtained from the experiments. This was done by characterizing significant factors
by means of the analysis of variance (ANOVA) technique. Then, regression analysis was
conducted to establish the relationship between the factors and the response. Subsequently,
signal-to-noise (S/N) ratio analysis was conducted to find the optimal settings and factor
levels. Finally, optimal process parameters were verified by three experiments.
Fig. 3 A cylindrical EDT part Fig. 4 Tool electrode after machining
2. Experimental setup and equipment
The material selected for testing was titanium alloy because this material is difficult to
machine. It is extensively used in aerospace, automobile, chemical, and biomedical
applications due to its excellent physical, metallurgical, and mechanical properties such as
high strength-to-weight ratio, high temperature stability, and good corrosion resistance. A
cylindrical bar (10 mm in diameter) made of titanium alloy was tested. The machining time
for each experiment was fixed to 30 minutes. During that time, workpieces were machined
from 10 mm to 7.96mm. The tool electrode made of electrolytic copper with dimensions of 40
mm × 25 mm × 8 mm was manufactured on a CNC vertical machining centre (Agni
BMV45TC 24, BFW make). A photo of the tool electrode after machining is shown in Fig. 4.
Industrial grade EDM oil was used as a dielectric fluid for side flushing. All the experiments
were performed on an Electronica 500×300 ZNC EDM machine at reverse polarity. Surface
roughness was measured using a surface roughness tester (SV 514, Mitutoyo make) with 0.8
mm cut-off length (according to DIN EN ISO 3274:1998). The surface roughness was
measured at three different sections and the average of three values was considered. The
recast layer was revealed by a standard metallurgical procedure using cutting, grinding, and
polishing. Cylindrical workpieces were sliced in radial direction. The surface of sliced cross
sections were polished and etched to observe the sub-surface damage. Kroll's reagent (1-3 ml
HF, 2-6 ml HNO
3
and 100ml pure water) was used to etch the polished surface of titanium
workpieces.
2.1 Design of experiments
The accuracy and effectiveness of an experimental program depends on careful
planning and execution of the experimental procedure. In order to achieve the aforesaid
objective, the Taguchi orthogonal array (OA) L
18
(2
1
×3
7
) was adopted for the design of
experiments. Thus, 18 experiments were conducted at parameter levels shown in Table 1.
TRANSACTIONS OF FAMENA XL-4 (2016)
3
Vikas Gohil, Y. M. Puri Experimental Investigation on Surface Roughness in
Electrical Discharge Turning of Ti-6Al-4V Alloy
Each run is replicated three times so that the total number of runs is 54. Note that the
experiments were run in random order. Current, pulse on-time, spindle rotational speed,
flushing pressure, and duty factor, each at three levels, are adopted as factors (variable
parameters) which vary during the experiments according to the design of experiment. The
results of the surface roughness values measured in each experimental run are presented in
Table 1. In this research, only the main effects of factors are of interest and their interactions
are excluded from the data analysis. Therefore, the degree of freedom (DOF) for this study is
calculated as in Equation 1 [19].
DOF = number of factors × number of factor level - 1 (1)
= 5 × 3 - 1 = 10
Table 1 Experimental layout of an L
18
orthogonal array
Run C Ton RPM FP DF Surface roughness Ra (µm)
Replicate 1 Replicate 2 Replicate 3
1 5 5 50 0 1 2.40 2.15 2.42
2 5 10 100 0.25 5 2.40 2.57 2.28
3 5 15 150 0.50 9 2.80 3.54 3.20
4 20 5 50 0.25 5 3.17 2.42 2.86
5 20 10 100 0.50 9 3.58 3.76 3.37
6 20 15 150 0 1 4.18 4.19 4.61
7 35 5 100 0 9 3.17 3.25 3.55
8 35 10 150 0.25 1 4.93 4.40 4.67
9 35 15 50 0.50 5 5.16 5.28 5.43
10 5 5 150 0.50 5 2.00 2.12 2.20
11 5 10 50 0 9 2.40 3.60 3.42
12 5 15 100 0.25 1 4.30 3.27 3.98
13 20 5 100 0.50 1 3.00 3.35 3.16
14 20 10 150 0 5 3.38 3.52 3.29
15 20 15 50 0.25 9 4.20 4.41 4.26
16 35 5 150 0.25 9 3.26 3.50 3.15
17 35 10 50 0.50 1 4.30 3.95 4.44
18 35 15 100 0 5 4.55 5.10 5.14
C- Current, Ton- pulse-on time, RPM- spindle rotational speed, FP- flushing pressure, DF- duty factor
3. DATA ANALYSIS
Fig. 5 shows five plots of the main effects of five different input parameters on the
surface roughness of the workpiece machined surface. Note that the data mean is used to
determine the effect of each factor. Fig. 5 shows that the applied current and pulse-on time
have the most significant effects on the surface roughness. Both parameters and the surface
roughness are directly proportional, i.e. if the values of the current and pulse-on time increase,
the surface roughness increases significantly. This happens because when the value of current
increases, more violent electrical discharges strike the workpiece surface and produce deeper
craters. As a result, the surface processed by EDM is deteriorated. Similarly, when the value
of pulse-on time is increased, more heat energy is transferred to the workpiece surface and
more material is melted. The effects of spindle rotational speed, flushing pressure, and duty
factor on surface roughness are also presented in Fig. 5. As indicated, none of the parameters
4
TRANSACTIONS OF FAMENA XL-4 (2016)
Experimental Investigation on Surface Roughness in Vikas Gohil, Y. M. Puri
Electrical Discharge Turning of Ti-6Al-4V Alloy
have a significant effect on surface roughness except the duty factor which has the opposite
effect on it; the value of surface roughness decreases as that of the duty factor increases. Fig.
6 shows the plot of main effects of five selected parameters on the S/N ratio for surface
roughness. As it is clearly seen, by increasing the values of current and pulse-on time, the
sensitivity of the system to the noise factors increases for the surface roughness, whereas by
increasing the values of spindle rotational speed and duty factor, the sensitivity of the system
to the noise factors decreases for this response. The system is less sensitive to the noise
factors at the mean level of duty factor, and flushing pressure has a negligible effect on the
sensitivity of the system. ANOVA has often been employed by researchers since it overcomes
the shortcomings of graphical assessment. ANOVA is used here for discussing and
interpreting data. Before any inference can be made based on the ANOVA table data, the
assumptions based on the ANOVA process have to be checked and verified as follows.
Fig. 5 Effects of five factors on surface roughness Fig. 6 Plot of the main effects on the S/N ratio for Ra
Fig. 7 Normal probability plot of residuals Fig. 8 Plot of residuals versus fitted values
Fig. 7 shows the normal plot of residuals. This plot is used to test the normal
distribution of errors. If the underlying error distribution is normal, this plot will resemble a
straight line. The distribution of points in Fig. 7 shows that the error normality assumption is
valid [17]. The other two assumptions are shown valid by means of the plot of residuals
versus fitted values. Fig. 8 shows the plot of residuals versus fitted values for surface
roughness. The structureless distribution of data points above and below the abscissa (fitted
values) shows that the errors are independently distributed and the variance is constant [18].
In the above discussion, ANOVA assumptions (error normality, error independency, and
variance constancy) are proved not to be violated through this experimentation, so ANOVA
can be performed and the inference made on the basis of its table will be valid. Table 2
presents the ANOVA table for surface roughness. It can be seen from the table that the
current, pulse-on time, and duty factor have the most significant impacts on surface roughness
at a confidence level of 95%. Therefore, the P-values which are less than 0.05 indicate that the
null hypothesis should be rejected and thus the effect of the respective parameter is
significant.
35205
4.5
4.0
3.5
3.0
15105 15010050
0.500.250.00
4.5
4.0
3.5
3.0
951
IP
Mean
Ton RPM
FP DF
Main Effects Plot for Surface roughness
Data Means
35205
-9
-10
-11
-12
-13
15105 15010050
0.500.250.00
-9
-10
-11
-12
-13
951
IP
Mean of SN ratios
Ton RPM
FP DF
Main Effects Plot for SN ratios
Data Means
Signal-to-noise: Smaller is better
1.00.50.0-0.5
99
95
90
80
70
60
50
40
30
20
10
5
1
Residual
Percent
Normal Probability Plot
(response is Surface roughness)
5.55.04.54.03.53.02.52.0
1.00
0.75
0.50
0.25
0.00
-0.25
-0.50
Fitted Value
Residual
Versus Fits
(response is Surface roughness)
TRANSACTIONS OF FAMENA XL-4 (2016)
5