Engineering Science and Technology, an International Journal 18 (2015) 664e668

H O S T E D BY Contents lists available at ScienceDirect

Engineering Science and Technology,

an International Journal
journal homepage: http://www.elsevier.com/locate/jestch

Short communication

Modelling and analysis of material removal rate and surface roughness

in wire-cut EDM of armour materials
Ravindranadh Bobbili*, V. Madhu, A.K. Gogia
Defence Metallurgical Research Laboratory, Hyderabad 500058, India

a r t i c l e i n f o a b s t r a c t

Article history: The current work presents a comparative study of wire electrical discharge machining (WEDM) of ar-
Received 3 March 2015 mour materials such as aluminium alloy 7017 and rolled homogeneous armour (RHA) steel using
Received in revised form buckingham pi theorem to model the input variables and thermo-physical characteristics of WEDM on
22 March 2015
material removal rate (MRR) and surface roughness (Ra) of Al 7017 and RHA steel. The parameters of the
Accepted 29 March 2015
Available online 13 May 2015
model such as pulse-on time, flushing pressure, input power, thermal diffusivity and latent heat of
vaporization have been determined through design of experiment methodology. Wear rate of brass wire
increases with rise in input energy in machining Al 7017. The dependence of thermo-physical properties
Keywords:
Armour materials
and machining variables on mechanism of MRR and Ra has been described by performing scanning
Buckingham pi theorem electron microscope (SEM) study. The rise in pulse-on time from 0.85ms to 1.25ms causes improvement in
Wire electrical discharge machining MRR and deterioration of surface finish. The machined surface has revealed that craters are found on the
machined surface. The propensity of formation of craters increases during WEDM with a higher current
and larger pulse-on time.
© 2015 Karabuk University. Production and hosting by Elsevier B.V. This is an open access article under
the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction Tosun [3] evaluated the significance machine variables on re-

sponses i.e kerf width and Ra. An optimum combination of process
Wire electrical discharge machining (WEDM) of armour mate- parameters was derived for large MRR and small Ra by applying
rials such as aluminium alloy 7017 and RHA steel has been analysis of variance (ANOVA). Tzeng et al. [4] studied the effect of
considered in this work. Al 7017 is an AleMgeZn-based alloy, machine variables on Ra by employing taguchi technique. Kumar
which is having superior impact strength, corrosion resistance and [5] employed grey relational methodology to optimize input pa-
low density [1]. This alloy is potential for armour applications due rameters of EDM to maximize MRR. The optimum machine vari-
to excellent properties. Armour steels [2] are the widely used ables were validated by performing confirmation experiments.
metallic armour materials nowadays due to excellent strength and Wang [6] explored the possibility of removing recast layer using
hardness with superior toughness making them suitable for armour etching by means of EDM. An L9 orthogonal array was selected to
applications. WEDM plays significant role in cutting conductive design experiments for attaining optimum process parameters.
materials to produce intricate profiles and complex shapes. The Somasekhar [7] presented the modelling and optimization of
material removal takes place due to melting and evaporation of micro-EDM using back propagation and genetic algorithms. The
workpiece because of the heat produced by discharges. The wire neural net work model has been established and simulated using
traverse is regulated by numerical controlled systems to accomplish MATLAB. Lin et al. [8] attempted to improve the multiple response
desired accuracy and precision of components. characteristics using taguchi technique to optimize machine vari-
The most significant response variables in WEDM are material ables of EDM. Tsai [9] developed a semi-empirical model of surface
removal rate (MRR) and surface roughness (Ra) of workpiece. Spark finish for various materials by adopting dimensional analysis in the
gap voltage, discharge current and pulse duration are the electrical discharge machining process. Fan et al. [10] employed
machining parameters which influence the performance measures. Buckingham pi theorem to develop a model for the erosion rate in
micro abrasive air jet machining of glasses.
* Corresponding author. Tel.: þ91 040 24586355; fax: þ91 040 24342252. Limited research has been done to develop mathematical
E-mail address: ravindranadhb@gmail.com (R. Bobbili). models for MRR and Ra in wire EDM based on thermo-mechanical
Peer review under responsibility of Karabuk University.

http://dx.doi.org/10.1016/j.jestch.2015.03.014
2215-0986/© 2015 Karabuk University. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).
 R. Bobbili et al. / Engineering Science and Technology, an International Journal 18 (2015) 664e668 665

approaches. Authors have made an attempt to establish the relation

between these parameters by employing buckingham pi theorem
and understanding the mechanism of material removal by per-
forming microscopic study. RHA steel and aluminium 7017 are the
most widely used armour materials. Using these materials, a variety
of armour configurations have to be fabricated to provide ballistic
protection against different threat levels. WEDM plays an impor-
tant role in processing these materials to the desired sizes to
accomplish the above objective. The aim of this investigation is to
analyse the effect of input variables and thermo-physical charac-
teristics of Al 7017 and RHA steel on MRR and Ra. Based on the
literature survey, several preliminary experiments were performed
to select the influencing factors on performance characteristics. The
chosen machining variables are: pulse on-time (TON), flushing
pressure (FP), peak current (IP), input power (E), thermal diffusivity
(a) and latent heat of vaporization (Hv). The design of experiment
technique is a dominant experimental planning tool, uses efficient
and orderly approach for obtaining optimum process variables.
Buckingham pi theorem is employed to develop mathematical Fig. 1. Photograph of the experimental system.
model between process variables and performance characteristics.
The complex nature of WEDM involves interaction of thermal,
mechanical and physical phenomena, which makes process model recognise the machining variables by screening experiments. Two
difficult. The current work presents a comparative study of WEDM aspects employed in taguchi method are (i) S/N ratio to estimate
of Al 7017 and RHA steel using buckingham pi theorem to model the quality [8] and (ii) orthogonal arrays [11,12] to accommodate
the input variables and thermo-physical properties. Experiments many factors affecting simultaneously to evaluate the machining
were carried out on workpieces under various conditions. performances. The machining parameters and corresponding
values selected are furnished in Table 2. Using Taguchi technique
2. Experimental details [9], an L18 (21  33) orthogonal arrays table was chosen. In the
present study all the designs, plots and analysis have been carried
2.1. Material and methods out using Minitab statistical software. The effect of various WEDM
machining parameters has been assessed by adopting ANOVA
The experiments were performed using CNC WEDM make technique. Based on ANOVA tables (Table 3a,b) for both MRR and
ELECTRONICA MACHINE TOOLS LTD. The Wire cut Electric Ra, it was observed that TON is the influencing parameter followed
Discharge Machine usually consists of a machine tool, a power by peak current. Preliminary trials were carried out by varying TON
supply unit and flushing unit. Wire travels through the work piece and IP, but at higher levels of IP frequent wire breakage was wit-
from upper and lower wire guides. In wire-cut EDM process the nessed. Peak current, pulse-off time, flushing pressure and wire
spark is generated between continuously travelling brass wire tension remain same during the experiments.
(0.25 mm diameter) and work piece. Aluminium alloy 7017 (0.3% Si,
0.5% Cu, 3.0% Mg, 5.00% Zn and Al remainder) and RHA steel 2.3. Buckingham pi theorem
(0.3e0.35% C, 1.5% Cr, 1.5% Ni, 0.14% V, 0.25% Si, 0.4% Mn, 0.03% Al,
remaining Fe) were used as cutting materials (Table 1). A picture of A correlation between n parameters (properties such as density,
the machine unit is given in Fig. 1. The cutting performance was thermal conductivity and specific heat) can be stated as a relation
evaluated by MRR and Ra. between (nem) dimensionless groups of variables, where m is the
The MRR was determined by equation, number of basic dimensions [8]. The machine parameters: pulse-on
time (TON), input power (E), flushing pressure (FP) were selected as
MRR ðmm3 =minÞ ¼ cutting rate  width  depth of work: parameters. The thermal properties [13] of workpiece materials
(1) such as heat of vaporization (Hv) and thermal diffusivity (a) were
chosen to establish model [14]. Variables used in the model for
The surface roughness, usually expressed as a Ra value in mi- workpiece materials are given in Table 1. The fundamental di-
crons was measured by Taylor Hobson Surtronic 25 Roughness mensions of chosen parameters are given in Table 4. The group of
Checker. significant variables and material properties for determining MRR
and Ra can be expressed as:
2.2. Design of experiment
MRR ¼ f ðTON; E; FP; a; HV Þ (2)
In this study, two stages of experimental work were necessary to
develop mathematical models for MRR and Ra. The first stage is to Ra ¼ f ðTON; E; FP; a; HV Þ (3)

Table 1
Properties of materials. Table 2
Input process parameters and their levels.
Properties Al 7017 Armour steel Units

Parameters Symbol Level 1 Level 2 Level 3 Units
Melting point 950 1800 K
Density 2800 7800 kg/m3 Pulse On time Ton 0.85 1.35 e ms
Thermal conductivity 210 80 w/m-k Pulse off time Toff 18 36 56 ms
Specific heat 950 410 J/kg-k Peak current IP 10 13 16 A
Heat of vaporization 10542 6230 kJ/kg Spark voltage SV 10 15 20 Volt
666 R. Bobbili et al. / Engineering Science and Technology, an International Journal 18 (2015) 664e668

Table 3 Table 6
(a) Analysis of Variance for MRR. (b) Analysis of Variance for Surface roughness (Ra). Coefficients of the model of Ra.

Source DF Seq SS Adj SS Adj MS F P Al 7017 Armour steel

(a) R2 0.99 0.96

Ton 1 1.478 1.478 0.478 8.28 0.016 RMSE 0.3955 1.102
Toff 2 0.00541 0.0054 0.00271 0.02 0.985 A2 2.443 1.824
IP 2 2.939 2.939 1.469 8.23 0.008 B2 0.1962 0.578
SV 2 1.0936 1.0934 0.5468 3.06 0.092 C2 1.74 0.968
Error 10 1.7864 1.7864 0.17865
Total 17 7.303
(b)
Ton 1 18.453 18.453 18.452 8.15 0.017 B1
Toff 2 1.26 1.260 0.6299 0.28 0.763 MRR ¼ A1  ðTON1=2  a3=2 Þ  ða3=2  TON1=2  FP  E1
IP 2 67.339 67.339 33.669 14.86 0.001
SV 2 2.674 2.674 1.337 0.59 0.572  C1
Error 10 22.653 22.653 2.265  TON  HV  a1
Total 17 112.37
(5)

B2
Table 4 Ra ¼ A2  ðTON1=2  a1=2 Þ  ða3=2  TON1=2  FP  E1
Dimensions of parameters in WEDM.
 C2
Parameters Dimensions  TON  HV  a1
Melting point q
(6)
Density ML3
Thermal conductivity MLT3 q1
Specific heat L2 T2 q1 where A1, B1 and C1 are power indexes of MRR model.
Heat of vaporization L2 T2 A2, B2 and C2 are power indexes of Ra model.
Pulse-on time T The power indices of model for MRR and Ra are presented in
Input power ML2 T3
Tables 5 and 6.
Surface roughness L
Material removal rate L3 T1
Thermal diffusivity L2 T1
3. Results and discussion
Flushing pressure ML1 T2

A mathematical model for MRR and Ra has been developed with

process variables of pulse-on time, input power, flushing pressure
Based on theory of WEDM, the material removal takes place
and other thermal properties. The coefficients and power indexes
significantly by evaporation. Thermal properties [15] like specific
of models were evaluated for different workpiece materials by
heat (CP), thermal conductivity (K) and density (r) can be written in
employing various optimization methods. By noticing the coeffi-
one parameter i.e. called thermal diffusivity (a). Input power (E) is
cient of the models, the B1 and C1 are smaller than A1. This study
expressed as the product of applied voltage and current. As the
corroborates the results from design of experiment that the ther-
non-dimensional homogeneous equations of performance mea-
mal properties are also prominent on MRR and Ra. It is observed
sures have six variables and only three basic dimensions, the so-
that R2 values of RHA steel are less than aluminium alloy due to
lution can be expressed as a product of three terms (p1, p2 and p3).
variation in thermal properties. Hence these metals can be distin-
The p variables can be expressed for estimating MRR as follows.
guished by different set of power indexes. Figs. 2e3 depict com-
parison between experimental and modelling results. It is noticed
from the above figures that the average prediction error is less than
p1 ¼ Ea1 TONb1 ac1 MRR (4a)
10%. This model is different from the earlier, since it mainly con-
centrates on thermal end electrical properties of the selected
materials.
p2 ¼ Ea2 TONb2 ac2 FP (4b) Total input energy transported between wire and workpiece is
distributed into three main components namely wire, work piece

p3 ¼ Ea3 TONb3 ac3 HV (4c)

p2 and p3 can be expressed as a function of p1 as p1 ¼ f (p2, p3)

By comparing the powers of basic units on both sides, the
following expressions are derived. Equations (1) and (2) are
deduced as (4) and (5) respectively.

Table 5
Coefficients of the model of MRR.

Al 7017 Armour steel

R2 0.99 0.97
RMSE 0.3201 1.195
A1 1.89 2.818
B1 0.6846 0.8067
C1 1.008 0.7836
Fig. 2. Comparison between experimental results and model predictions of MRR.
 R. Bobbili et al. / Engineering Science and Technology, an International Journal 18 (2015) 664e668 667

At a large current, a stronger discharge generates more heat energy.

By virtue of the size of workpiece, some amount of heat is absorbed
by it. The remaining portion of energy is accumulated at the wire
resulting into higher wear rate, this leads to frequent wire
breakages.
Higher thermal conductivity of Al 7017 helps easy energy
dissipation and its low melting point facilitates larger MRR. At low
input power, a small amount of thermal energy is produced and a
significant portion is absorbed by the surroundings, this makes
available energy will be less. But the rise in input power generates
intense discharge, which impacts the surface of the workpiece and
causes more molten material to be driven out of the crater. The SEM
images of Al 7017 machined surface are shown in Fig. 4. The
machined surface revealed that discharge craters were large at
higher peak current and pulse duration. The Increase in TON from
Fig. 3. Comparison between experimental results and model predictions of Ra. 0.85 to 1.25 msec resulting into creation of larger craters on the

Fig. 4. Micrographs of machined surface of Al 7017.

and gap between them. The energy loss comprises heat carried surface. This is reason for the increase in Ra with input power and
away by the debris by conduction and heat loss due to convection pulse-on time. The higher the input power, the smaller is the
and radiation. machining time, as the machining rate is proportional to input
The maximum MRR was obtained for cutting of Al 7017 than power. It directly depends on the number of sparks generated per
RHA steel. The reason may be due to lower melting point. Heat of second. Flushing pressure (FP) has a significant influence on MRR.
vaporization and thermal conductivity were found to be important Higher MRR can be achieved by supplying dielectric fluid at low
characteristics of the material for ascertaining MRR. Reduction in velocity into the spark gap, thereby short-circuit effect is negligible.
MRR is attributed to larger heat capacity and thermal conductivity. This enhances improvement in efficiency and thus increases MRR.

Fig. 5. Micrographs of machined surface of RHA steel.

668 R. Bobbili et al. / Engineering Science and Technology, an International Journal 18 (2015) 664e668

Higher FP hampers creation of ionized bridges across the gap and Paman of armour design and development group for participating
reduces spark energy and diminishes MRR. Fig. 5(a) shows the SEM in technical discussions.
micrograph of RHA steel machined surface observed at minimal
surface roughness condition. Fig. 5(b) shows cross sectional view of References
specimen, depicts the outer layer called recast layer which cooled
too quickly to and were cast to the material. [1] P.K. Jena, N. Jagtap, K. Siva kumar, T. Balakrishna Bhat, Some experimental
studies on angle effect in penetration, Int. J. Impact Eng. 37 (2010) 489e501.
[2] P.K. Jena, B. Mishra, M. Rameshbabu, A. Babu, A.K. Singh, K. Sivakumar,
4. Conclusions T.B. Bhat, Effect of heat treatment on mechanical and ballistic properties of a
high strength armour steel, Int. J. Impact Eng. 37 (2010) 242e249.
An application of buckingham pi theorem to model the input [3] N. Tosun, C. Cogun, Gul Tosun, A study on kerf and material removal rate in
WEDM based on Taguchi method, J. Mater. Process. Technol. 152 (2004)
variables of WEDM on MRR and Ra of Al 7017 and RHA steel alloy 316e322.
has been studied. The dependence of thermo-physical parameters [4] C. Tzeng, Y. Lin, Y. Yang, M.C. Jeng, Optimization of turning operations with
on mechanism of MRR and Ra has been interpreted. The conclu- multiple performance characteristics using the taguchi method and grey
relational analysis, J. Mater. Process. Technol. 209 (2009) 2753e2759.
sions are as follows:
[5] A. Kumar, S. Maheswari, S. Sharma, B. Naveen, A study of multiobjective
parametric optimization of silicon abrasive mixed electrical discharge mach-
 Mathematical models were developed using buckingham pi ing of tool steel, Mater. Manuf. Process. 25 (2010) 1041e1047.
theorem for MRR and Ra to establish the relationship between [6] C.C. Wang, H.M. Chow, L.D. Yang, C.T. Lu, Recast layer removal after electrical
discharge machining via taguchi analysis: a feasibility study, J. Mater. Process.
machine variables and performance measures. Results predicted Technol. 209 (2009) 4134e4140.
by the model are well matching with experiment results. [7] K.P. Somashekhar, N. Ramachandran, J. Mathew, Optimization of material
 Owing to thermal properties of WEDM, MRR of machining is removal rate in micro-EDM using artificial neural network and genetic algo-
rithms, Mater. Manuf. Process. 25 (2010) 467e475.
higher for materials with a low melting temperature and spe- [8] J.L. Lin, C.L. Lin, The use of the orthogonal array with grey relational analysis to
cific heat. optimize the electrical discharge machining process with multiple perfor-
 The rise in pulse-on time from 0.85ms to 1.25ms causes mance characteristics, Int. J. Mach. Tools Manufac. 42 (2002) 237e244.
[9] K. Tsai, P. Wang, Semi-empirical model of surface finish on electrical discharge
improvement in MRR and deterioration of surface finish. The machining, Int. J. Mach. Tools Manufac. 41 (2001) 1455e1477.
machined surface revealed that craters are found on the [10] J.M. Fan, C.Y. Wang, J. Wang, Modeling the erosion rate in micro abrasive air
machined surface. The propensity of formation of craters in- jet machining of glasses, J. Wear 266 (2009) 968e974.
[11] M.S. Phadke, Quality Engineering Using Robust Design, Prentice- Hall, New
creases during WEDM with a higher current and larger pulse-on Jersey, 1989.
time. [12] Electronica Machine Tools Ltd., Technology Manual for Wire Cut EDM -
 Wear rate of brass wire increases with increase in input energy Elektra Sprintcut 734, p. 1e2.
[13] M.C. Chua, M. Rahman, Y.S. Wong, H.T. Loh, Determination of optimal cutting
leads to wire breakage.
conditions using design of experiments and optimization techniques, Int. J.
Mach. Tools Manufac. 33 (1993) 297e305.
Acknowledgements [14] H.S. Payal, R. Choudhary, S. Singh, Analysis of electro discharge machined
surfaces of EN-31 tool steel, J. Sci. Ind Res. 67 (2008) 1072e1077.
[15] W. Yu-kui1, X. Bao-cheng, W. Zhen-long, P. Zi-long, Micro EDM deposition in
The authors are thankful to DRDO, India for rendering monetary air by single discharge thermo simulation, Trans. Non Ferr. Met. Soc. China 21
help to this research. The authors would also thank Mr.Ashish (2011) s450es455.