Optimization of Surface Roughness Parameters of Al-6351 Alloy in EDC Process: A Taguchi Coupled Fuzzy Logic Approach

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Optimization of Surface Roughness

Parameters of Al-6351 Alloy in EDC


Process: A Taguchi Coupled Fuzzy Logic
Approach

Siddhartha Kar, Sujoy Chakraborty,


Vidyut Dey & Subrata Kumar Ghosh

Journal of The Institution of


Engineers (India): Series C
Mechanical, Production, Aerospace and
Marine Engineering

ISSN 2250-0545

J. Inst. Eng. India Ser. C


DOI 10.1007/s40032-016-0297-y

1 23
Your article is protected by copyright and all
rights are held exclusively by The Institution
of Engineers (India). This e-offprint is for
personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

1 23
Author's personal copy
J. Inst. Eng. India Ser. C
DOI 10.1007/s40032-016-0297-y

ORIGINAL CONTRIBUTION

Optimization of Surface Roughness Parameters of Al-6351 Alloy


in EDC Process: A Taguchi Coupled Fuzzy Logic Approach
Siddhartha Kar1 • Sujoy Chakraborty1 • Vidyut Dey1 • Subrata Kumar Ghosh1

Received: 21 December 2015 / Accepted: 18 May 2016


 The Institution of Engineers (India) 2016

Abstract This paper investigates the application of Abbreviations


Taguchi method with fuzzy logic for multi objective EDC Electro discharge coating
optimization of roughness parameters in electro discharge P/M Powder metallurgy
coating process of Al-6351 alloy with powder metallurgi- SiC Silicon carbide
cal compacted SiC/Cu tool. A Taguchi L16 orthogonal Cu Copper
array was employed to investigate the roughness parame- Ra Centre line average (lm)
ters by varying tool parameters like composition and Rz Maximum height of profile (lm)
compaction load and electro discharge machining param- Sa Mean spacing of local peaks (lm)
eters like pulse-on time and peak current. Crucial rough- S/N Signal to noise (dB)
ness parameters like Centre line average roughness, COM Comprehensive output measure
Average maximum height of the profile and Mean spacing Al Aluminium
of local peaks of the profile were measured on the coated EDM Electro discharge machining
specimen. The signal to noise ratios were fuzzified to Amp Ampere
optimize the roughness parameters through a single com- ls Micro second
prehensive output measure (COM). Best COM obtained P Compaction load (Ton)
with lower values of compaction load, pulse-on time and Adj Adjusted
current and 30:70 (SiC:Cu) composition of tool. Analysis
List of Symbols
of variance is carried out and a significant COM model is
C Composition
observed with peak current yielding highest contribution
lm Micro metre
followed by pulse-on time, compaction load and compo-
Ton Pulse on time (ls)
sition. The deposited layer is characterised by X-Ray
Toff Pulse off time (ls)
Diffraction analysis which confirmed the presence of tool
Ip Peak current (Amp)
materials on the work piece surface.
DOE Design of experiment
MSD Mean square deviation
Keywords Electro discharge coating  Surface roughness 
VS Very small
Taguchi  Fuzzy logic  Comprehensive output measure
S Small
M Medium
L Large
VL Very large
Con Condition
& Sujoy Chakraborty ANOVA Analysis of variance
ersujoymech05@gmail.com XRD X-Ray Diffraction
1
National Institute of Technology Agartala, Agartala 799046,
Tripura, India

123
Author's personal copy
J. Inst. Eng. India Ser. C

Introduction parameters of surface roughness were centre line average


(Ra), maximum height of profile (Rz) and mean spacing
EDM is a well-established non-traditional machining pro- of local peaks (Sa). The fuzzy logic unit was used to
cess for machining very hard materials (like Inconel, transform the three correlated response into a single
Tungsten etc.) which are difficult to machine by conven- index named COM (comprehensive output measure) to
tional machining processes [1]. Recently, EDM has emerged determine the optimum EDC parameters to achieve good
as a tool for surface modification by intentional deposition of surface finish.
tool materials over work piece surface. The intrinsic feature
of EDM to have tool wear and formation of hard layers over
surface of substrate by the reaction of tool materials and Proposed Taguchi Coupled Fuzzy Logic
decomposed hydrocarbon from dielectric fluid has led to the Methodology
technique of surface modification by EDM, which is also
known as electro discharge coating (EDC) [2]. For proper Taguchi is an effective tool of design of experiment which
material transfer, the tool material must be loosely bonded evaluates the individual effect of each process parameters
together to be able to get eroded sufficiently by thermo- on the responses by conducting minimum number of
electric energy of the discharge. This type of special tool can experiments [8]. Signal to noise ratio finds the best level of
be made by P/M compact, either in green or semi sintered each parameter for a particular response based on quality
state [3]. The intentional transfer of tool material is done to characteristic such as lower the better, higher the better,
modify the mechanical properties such as strength, tough- and nominal the better [9]. Irrespective of the chosen
ness, hardness along with surface properties such as wear quality characteristics for a particular response, the levels
resistance and abrasion resistance with good surface finish. with higher S/N ratios of the corresponding factors is
For saleability of a product, its surface finish is the outmost deemed to be optimum for that particular response. Tagu-
demand at any field. Most researchers define the surface chi method alone is sufficient enough for handling single
roughness based on a single parameter which is Ra (Centre objective problems but not adequate to handle multi
line average). But surface roughness depends on many objective problems. These types of problems are normally
parameters such as average, peak, surface, hybrid etc., and solved by assigning individual weights to the responses
Ra is just one of the average parameters. Optimization of all according to priority. Utility theory, grey relational anal-
the parameters together would properly justify the quality of ysis theory, desirability function approach are few of the
surface more accurately. examples. In all these theories weights are assigned based
In this study, electrodes were prepared by special on assumptions which may lead to uncertainty in result
process of green compaction P/M method, which is [10]. Fuzzy logic approach is used in this study to over-
composed of SiC and Cu of mesh 325 each. The com- come all these assumptions and to get over the regime of
plete process is intended to modify the surface of Al- uncertainty.
6351 alloy in terms of hardness, wear resistance and Fuzzy logic is a mathematical theory of inexact reasoning
abrasion resistance with good surface quality to with- that allows modelling of the reasoning process of human in
stand rigorous working condition. Such change in linguistic terms [11]. It is based on intuition and judgment
properties along with high surface finish would further and it does not need a mathematical model to construct [12].
enhance the applicability and saleability of Aluminium Fuzzy logic system (Mamdani system) comprises of a
alloys in various engineering services. SiC-coated Alu- fuzzifier, membership functions, a fuzzy rule base, an
minium has also found many applications in the aero- inference engine, and defuzzifier. Fuzzifier is used to convert
space [4] and automotive industry. SiC coated Al have the crisp values into values with membership function [13].
high modulus, strength values, wear resistance, high The membership function is a graphical representation of the
thermal stability, less weight and a more effective load magnitude of participation of each input. The rules use the
carrying capacity compared to many other materials input membership values as weighting factors to determine
[5, 6]. Moreover, any turbine engine or rotary com- their influence on the fuzzy output sets of the final output
pressor that suffers from erosion due to airborne particle conclusion. Once the functions are inferred, scaled, and
impact will also benefit from this type of coating [7]. combined, they are defuzzified into a crisp output which
The key process parameters selected in making Taguchi drives the system. Next, the inference engine (Mamdani
L-16 array were compaction load, composition, pulse fuzzy inference system) performs fuzzy reasoning on fuzzy
duration and peak current. The investigated process rules to generate a fuzzy value [14]. Finally, the defuzzifier

123
Author's personal copy
J. Inst. Eng. India Ser. C

Table 1 Composition of Al-6351 alloy


Si Fe Cu Mn Mg Zn Ti Al

0.8 % 0.5 % 0.1 % 0.4 % 0.4 % 0.2 % 0.2 % Rest

converts fuzzy predicted value into a single index named


COM.

Deposition Process by EDC

Selection of Tool Materials

Silicon carbide powder was selected as tool material due to


its high hardness, wear resistance and abrasion resistance.
Copper (Cu) powder was added to the SiC powder to
impart good binding strength and increase conductivity of
the tool that is prerequisite for performing EDC.

Selection of Substrate
Fig. 1 Pellets at different load and composition
Aluminium 6351 alloy was selected as the work-piece
material due to its wider application in the field of piping,
aerospace industries etc. The constituents of Al-6351 alloy
are shown in Table 1.

Preparation of P/M Compact Tool Electrode Using


SiC and Cu

Electrodes to be used for EDC were made with mixtures of


copper and silicon carbide at different compositions
(C) arbitrarily (50:50 and 70:30 wt%) by powder metal-
lurgy method. Here, Cu act as a binder and improve the
conductivity of tool. Again, Cu helps to adhere the compact
tool or pellet with the extension material. Electrodes con-
sist of two parts; a tool extension part for proper holding of
the tool electrode in the EDM system which may be cop-
per, brass, graphite, mild steel or any other conductive
material and a P/M compacted pellet which actually act as
Fig. 2 Pellets glued to copper rod
electrodes and from which material would be eroded and
deposited on the work piece surface during the EDC pro-
cess. For making pellets, copper and silicon carbide pow- Preparation of Substrate
ders were mixed together using mortar and pestle made of
ceramic. Powders were compacted at different compaction Al-6351 alloy plate of necessary dimension (25 mm 9
loads, P (5–20 Ton) in a pellet press consisting of a die and 25 mm 9 5 mm) complementing the tool surface dimen-
punch assembly of 13 mm diameter of maximum load sion was used as work piece.
25 Ton. The dimensions of the pellets were 5 mm in height
and 13 mm in diameter. The compacted pellets (Fig. 1) and Experimental Set Up
tool extension parts were glued together (Fig. 2) by an
electrically conductive paste. Microscopic images of pel- The experiments were carried out in die sinking EDM
lets with different composition of SiC and Cu is shown in (Model: Sparkonix 25F). Based on the outcomes of pilot
Fig. 3, where red colour represents Cu and grey colour experiments, for better deposition, tool and work piece are
represents SiC. connected to anode and cathode respectively. Kerosene oil

123
Author's personal copy
J. Inst. Eng. India Ser. C

Fig. 3 Microscopic images of


the pellets of different
composition of SiC and Cu.
a Pellet composed of 30:70 of
SiC and Cu. b Pellet composed
of 50:50 of SiC and Cu

Fig. 4 Experimental setups. a Spark EDM. b Pellet press

is used as dielectric fluid in this study. The experimental set melting point, thermal conductivity, specific heat etc.
ups of EDM and pellet press are shown in Fig. 4. The Sometimes two or more powders are mixed to produce
technical specifications of EDM are as follows: some composite coating which enhances the mechanical
properties of the coating.
Current: 1–25 A
Pulse on time: 3–2000 ls
Compaction Load
Pulse off time: 2–400 ls
This is one of the most valuable parameter, when preparing
Process Parameters a powder compact. Depending on the size of the die the
loads in tons are calculated and compact is prepared.
Composition of Tool Electrode Higher the compaction load, higher will be the bonds
between the particle and compact will be more dense and
Composition of powder or powder mixture is a very with higher strength. However higher strength of compact
important factor, because it decides the properties of the sometimes reduces the tool wear during EDC process and
developed coating i.e. hardness, wear resistance, density, deposition rate decreases.

123
Author's personal copy
J. Inst. Eng. India Ser. C

Table 2 Factors and their levels selected for Taguchi L16 orthogonal array
Factors Symbol Level 1 Level 2 Level 3 Level 4

Compaction load (Tons) P 5 10 15 20


Peak current (Amp) Ip 2 4 6 8
Pulse duration (ls) Ton 11 21 50 100
Composition (SiC:Cu) C 30:70 50:50 – –

resulting in higher cutting rate. With higher values of Ton,


Table 3 Modified Taguchi L16 array
however, surface roughness tends to be higher.
Exp. no. Input parameters
P Ip Ton C (SiC–Cu) Peak Current

1 1 1 1 1
The peak current (Ip) is the maximum value of the current
2 1 2 2 1
passing through the electrodes for the given pulse. Increase
3 1 3 3 2 in the Ip value will increase the pulse discharge energy
4 1 4 4 2 which in turns improves the cutting rate further. For higher
5 2 1 2 2 value of Ip, gap conditions may become unstable with
6 2 2 1 2 improper combination of Ton, Toff or other parameter set-
7 2 3 4 1 tings. As and when the discharge conditions become
8 2 4 3 1 unstable one must reduce the Ip value.
9 3 1 3 1
10 3 2 4 1 DOE by Taguchi Method
11 3 3 1 2
12 3 4 2 2 The main intention is to deposit silicon and its carbide
13 4 1 4 2 layer over Al-6351 alloy. Increasing the peak current above
14 4 2 3 2 eight Ampere and pulse duration above 100 ls produces
15 4 3 2 1 arcing rather than sparking thus hindering uniform depo-
16 4 4 1 1 sition and resulting in irregular surface and formation of
carbide layer. Based on pilot experiments, the lower level
and higher level of factors were selected [15] as given in
Pulse Duration Table 2. To accommodate this type of unequally dis-
tributed level of factors, mixed level design of Taguchi was
The pulse on time is referred as Ton and it represents the approached and a modified L-16 orthogonal array was
duration of time in micro seconds (ls), for which the selected (Table 3). By preliminary experiments it was
current is flowing in each cycle. During this time, the gap observed that reverse polarity was more suitable for coat-
voltage is applied across the electrodes. The single pulse ing. Duty cycle was kept between 0.5 and 0.57 throughout
discharge energy increases with increasing Ton period, the experimentations. The gap sensing knob was adjusted

Fig. 5 Photograph of
specimens after experiment in
order of experiment number
shown in tabular set. a Coated
specimens. b Tabular set Exp. 1 Exp. 2 Exp. 3 Exp. 4

Exp. 5 Exp. 6 Exp. 7 Exp. 8

Exp. 9 Exp. 10 Exp. 11 Exp. 12

Exp. 13 Exp. 14 Exp. 15 Exp. 16

(a) (b)

123
Author's personal copy
J. Inst. Eng. India Ser. C

Fig. 6 Surface roughness plot


(Ra = 1.42 lm,
Rz = 8.76 lm)

Fig. 7 3-D view of a surface roughness plot (Sa = 2.46 lm) Fig. 8 Membership function of Ra

A sample surface roughness plot is shown in Fig. 6 and


to maintain the gap voltage in the range of 55–60 V a 3-D view is shown in Fig. 7 where different colours
throughout each experiment. The set of coated specimens denote different heights.
along with a tabular set indicating experiment number is
shown in Fig. 5.
Multi Response Optimization Using Taguchi
Output Parameters Coupled Fuzzy Logic
The geometrical features of a surface includes Micro-ir- The crisp values in this project work are the S/N ratios
regularities, Macro-deviations and Surface waviness. corresponding to Ra, Rz and Sa. These values are then
Basically the surface roughness is evaluated by the height, converted to membership function. The input values are
Rz and mean roughness index, Ra of the micro-irregulari- subjected to triangular membership function of three sub-
ties. In this present study, the following surface roughness sets named small (S), medium (M) and large (L) as shown
parameters have been taken: in Figs. 8, 9 and 10. Similarly, the output value (COM) was
Ra It is also known as centre line average (CLA) or subjected to triangular membership function of five subsets
arithmetic average (AA). Mathematically, Ra is the named very small (VS), small(S), medium (M), large
arithmetic average value of the profile departure from (L) and very large (VL) as shown in Fig. 11. The fuzzy rule
the mean line, within a sampling length base consists of a group of if–then control rules [9] 13 with
Rz It is defined as the maximum height of profile within the three inputs, y1, y2 and y3 and one output z. The rules
a sampling length, usually analysed as a mean over a are formulated as:
minimum of 5 sampling lengths • Rule 1: if y1 is A1 and y2 is B1 and y3 is C1 then z is D1
Sa It is known as mean spacing of local peaks which else
usually indicates fine surface deviations caused by • Rule 2: if y1 is A2 and y2 is B2 and y3 is C2 then z is D2
material breakdown else

123
Author's personal copy
J. Inst. Eng. India Ser. C

Table 4 Fuzzy Rules expression


Sr. no. If Then
Ra con Rz con Sa COM

1. S and S and S VS
2. S and S and M S
3. S and S and L S
4. S and M and S S
5. S and M and M M
6. S and M and L M
7. S and L and S S
8. S and L and M M
9. S and L and L L
10. M and S and S S
Fig. 9 Membership function of Rz
11. M and S and M M
12. M and S and L M
13. M and M and S M
14. M and M and M M
15. M and M and L M
16. M and L and S M
17. M and L and M M
18. M and L and L L
19. L and S and S S
20. L and S and M M
21. L and S and L L
22. L and M and S M
23. L and M and M M
24. L and M and L L
25. L and L and S L
Fig. 10 Membership function of Sa 26. L and L and M L
27. L and L and L VL

where, Ai, Bi, Ci are fuzzy subsets of input variables and D


is fuzzy subset of output value defined by the correspond-
ing membership functions. Table 4 shows the 27 proposed
fuzzy rules. Based on the values obtained from experiment,
various degrees of membership of the fuzzy sets are
calculated. The inference system performs a fuzzy reason-
ing on fuzzy rules to generate a fuzzy value [16]. By taking
the max–min compositional operation, the fuzzy reasoning
of these rules yields a fuzzy output. Assuming that y1, y2
and y3 are the three input values of the fuzzy logic unit, the
membership function of the output of fuzzy reasoning can
be expressed as [14, 16, 17]:
 
lC0 ðzÞ ¼ lA1 ðy1 Þ ^ lB1 ðy2 Þ ^ lC1 ðy3 Þ ^ lD1 ðzÞ
 
Fig. 11 Membership function of COM _ lA2 ðy1 Þ ^ lB2 ðy2 Þ ^ lC2 ðy3 Þ ^ lD2 ðzÞ
 
_    _ lAn ðy1 Þ ^ lBn3d ðy2 Þ ^ lCn ðy3 Þ ^ lDn ðzÞ

• ………………… ð1Þ
• ………………… where, lAi ,lBi and lCi are the membership functions cor-
• Rule n: if y1 is An and y2 is Bn and y3 is Cn then z is Dn responding to the input values and lDi is the membership

123
Author's personal copy
J. Inst. Eng. India Ser. C

Table 5 Responses and their corresponding S/N ratios and COM


Exp. no. Responses COM
Ra (lm) S/N Ratio Rz (lm) S/N Ratio Sa (lm) S/N Ratio

1 2.75 -8.79 12.711 -22.08 2.609 -8.33 0.917


2 2.89 -9.22 17.712 -24.97 3.527 -10.95 0.622
3 3.47 -10.81 24.411 -27.75 4.108 -12.27 0.500
4 4.62 -13.29 21.002 -26.44 3.681 -11.32 0.395
5 2.7 -8.63 15.313 -23.7 2.731 -8.72 0.769
6 2.77 -8.85 22.425 -27.01 3.974 -11.98 0.546
7 3.61 -11.15 24.226 -27.69 8.408 -18.49 0.250
8 2.87 -9.16 25.408 -28.1 8.566 -18.66 0.250
9 2.74 -8.76 16.468 -24.33 3.228 -10.18 0.682
10 2.86 -9.13 22.534 -27.06 4.69 -13.42 0.500
11 2.92 -9.31 16.652 -24.43 3.576 -11.07 0.613
12 3.24 -10.21 20.37 -26.18 5.196 -14.31 0.500
13 2.83 -9.04 16.25 -24.22 3.333 -10.46 0.662
14 2.85 -9.1 17.983 -25.1 4.246 -12.56 0.500
15 2.73 -8.72 18.479 -25.33 4.146 -12.35 0.510
16 2.63 -8.4 15.64 -23.88 3.748 -11.48 0.600

function of the output value. Here, _ denotes the maximum


operation and ^ denotes the minimum operation.
Centroid method of defuzzification was adopted to
transform the fuzzy inference output to a non-fuzzy value
called COM by the following equation [18]:
P
zlC ðzÞ
z0 ¼ P 0 ð2Þ
lC0 ðzÞ

Results and discussion

Roughness parameters (Ra, Rz and Sa) of the deposited


layer were measured by Taylor Hobson 3d Profilometer.
Ten readings were taken for each specimen and averages of
them were taken for the present study. S/N ratios of the Fig. 12 Fuzzy Rule COM graph
observed data’s were also evaluated based on the quality
characteristics of smaller the better type which may be
defined as [8, 9]: Table 5. The Fuzzy Rule COM graph is also shown in
Fig. 12. It represents fuzzy rules stated in Table 4 in
S=N ratio; g ¼ 10log10 ðMSDÞ ð3Þ
graphical form indicating the range of the triangular
where, MSD is the mean-square deviation for the output membership functions of the input variables Ra (-13.29
characteristic. The mean-square deviation (MSD) for and -8.4), Rz (-28.1 and -22.08), Sa (-18.66 and
smaller the better quality characteristic can be expressed -8.33) and output variable COM (0 and 1) respectively. It
as: also demonstrates a working model of the system where the
1X n output COM value is taken as 0.5, when the input values of
MSD ¼ O2 ð4Þ Ra is -10.8, Rz is -25.1 and Sa is -13.5. The yellow
n i¼1 i
coloured triangles of the input variables indicate the
where, n is the number of tests and O is the observed value membership functions to which the input value belongs and
at the ith test. The Ra, Rz and Sa response with their blue colour triangle of COM indicate the membership
respective S/N ratios and corresponding COM are shown in function of output value yielding from the input values of

123
Author's personal copy
J. Inst. Eng. India Ser. C

Fig. 13 Surface view of COM vs Ra and Rz Fig. 14 Surface view of COM vs Ra and Sa

the system based on the adapted fuzzy rules. Surface view


of COM against S/N ratio of Ra and Rz is plotted in
Fig. 13. Steep slopes of contrasting nature of COM surface
are observed when S/N ratio of Ra and Rz are higher
(higher COM) and when S/N ratio of Ra and Rz are lower
(lower COM). A plane COM surface corresponding to 0.5
COM value is observed when Ra and Rz are of contrasting
nature (one is in higher side and other in lower side).
Similar trend of surfaces are also observed for COM
against S/N ratio of Ra and Sa in Fig. 14 and COM against
S/N ratio of Rz and Sa in Fig. 15 respectively, which
complements the fact that all the roughness parameters are
correlated to each other and have similar effect on the
outcome.
From Table 5, it is observed that, experiment no. 1 with
COM value of 0.917 yielded the best result corresponding Fig. 15 Surface view of COM vs Rz and Sa
to lowest current (2 A), pulse duration (11 ls) and com-
paction load (5 Ton) and SiC30:Cu70 composition roughness (Ra, Rz and Sa) increases with increase in cur-
respectively for combination of all the roughness parame- rent and pulse duration. It may be due to increase in spark
ters as comprehensive output measure gives higher value intensity hitting the tool material heats up the outer layer so
within this proposed model. Analysis of variance much (about 10,000 C) that the material evaporates. Due
(ANOVA) is a method to find statistically significant pro- to this, the metal gases are formed and then condense in the
cess parameters [19, 20]. Thus, with the help of ANOVA, dielectric, usually in the form of hollow balls. These hol-
the optimal combination of process parameters for each low balls or gas bubbles are having a sharp edge which
response can be determined. The Taguchi analysis and forms craters on the workpiece. So surface roughness (Ra,
ANOVA is carried out by statistical software MINITAB 16 Rz and Sa) increases with increase in current and pulse
in this present study. From the ANOVA table for COM duration, may be due to the increase in gas bubble pro-
(Table 6), the correlation coefficient [R2 (adj)] value is duced and temperature [21–23]. It may also be due to
fairly high (93.26 %) indicating better significance of the arcing, instead of sparking at higher on-time values [24].
proposed COM model. The percentage contributions of Again surface roughness (Ra, Rz and Sa) decreases with
process parameters (Fig. 16) were evaluated from sum of increase in compaction load which may be because of the
square terms. It is seen that peak current has 57.05 % material becomes denser and less porous. From Table 5, it
contribution to the Comprehensive Output Measure (COM) is also seen that surface roughness (Ra, Rz and Sa)
model followed by pulse duration (28.03 %), compaction increases with increasing percentage of SiC in tool elec-
load (12.34 %) and tool composition (0.34 %) respec- trode material with higher current and pulse duration. It
tively. It is also observed from Table 5 that surface may be because of higher deposition of tool electrode

123
Author's personal copy
J. Inst. Eng. India Ser. C

Table 6 Result of ANOVA for COM


Source Degree of freedom Sum of square Mean square F ratio P Contribution (%)

Compaction load 3 0.054281 0.018094 9.15 0.018 12.34


Peak current 3 0.250954 0.083651 42.32 0.001 57.05
Pulse duration 3 0.123296 0.041099 20.79 0.003 28.03
Composition 1 0.001482 0.001482 0.75 0.426 0.34
Error 5 0.009882 0.001976 – – 2.24
Total 15 0.439896 – – – 100
R2 = 97.75 % R2(adj) = 93.26 %

Compaction Load
0.34% 2.24% 12.34%
Peak Current
28.03%
Pulse Duration

57.05% Composition

Error

Fig. 16 Percentage of Contribution of input parameters to the COM Fig. 17 Optical Microscopic images of Parent material and deposited
model (color figure online) layer

material. As the percentage of SiC increases in the SiC/Cu in Fig. 18. The peaks of Si, SiC, Cu, Al along with the
mixture, the compactness of the tool electrode reduces and peaks of inter-metallic phases Al2Cu and Al2CuMg are
the tool electrode material is more ready to being deposited clearly visible from the figure which confirms the transfer
on the surface of workpiece and formation of more craters of tool material on the workpiece surface in elemental state
may also take place on the work piece due to higher current as well as in the form of carbides (SiC) [2, 27] and thus
and pulse on time value [25]. Krishna et al. [24] conducted forming a hard layer of composite.
a study on the surface of C-40 grade steel and they have
found the similar reasons which support the results
obtained in this present study. So, lower peak current and Conclusions
lower pulse duration settings can be used for obtaining
good surface finish in EDC process and also higher com- In this study, the multiple surface roughness parameters
paction load can be used in preparing powder tool for EDC like Centre line average roughness (Ra), Average maxi-
process. Thus, electric parameters like current and pulse mum height of the profile (Rz) and Mean spacing of local
duration should be chosen with precision. peaks of the profile (Sa) are efficiently optimized using the
Fuzzy theory along with Taguchi method in EDC of Al-
Surface Characterization 6351 alloy with powder metallurgical (P/M) compacted
SiC/Cu tool. The signal to noise (S/N) ratios were fuzzified
Microscopic image of the top surface of work piece high- to optimize the roughness parameters through a single
lighting both parent material and deposited layer is shown comprehensive output measure (COM). The following are
in Fig. 17. From this image, it is clearly seen the presence the conclusions:
of foreign material on the top surface of work piece. To
confirm the presence of this foreign material on the 1. All the surface roughness parameters (Ra, Rz and Sa)
workpiece, the top surface of work piece is also charac- are correlated to each other and have similar effect on
terized by X-Ray Diffraction (XRD) technique [26]. The the outcome as per the surface view of COM.
XRD plot of a sample deposited with tool of 15 Ton 2. In simultaneous optimization of three roughness
compaction load and composition of SiC50–Cu50 is shown parameters (Ra, Rz and Sa), 1-1-1-1(experiment no.

123
Author's personal copy
J. Inst. Eng. India Ser. C

15 Ton SiC50:Cu50 References


1,4
70 1 - Al
2 - Si 1. S.H. Wankhade, S.B. Sharma, Study of the productivity and
60 3 - SiC surface quality of hybrid EDM. J. Inst. Eng. India Ser. C 97(1),
1 4 - Cu 71–76 (2016)
50 5 - Al2Cu 2. P.K. Patowari, U.K. Mishra, P. Saha, Surface modification of C40
3 steel using WC-Cu P/M green compact electrodes in EDM. Int.
6 - Al2CuMg J. Manuf. Technol. Manage 21(1/2), 83–98 (2010)
Intensity →

40
3. P.K. Patowari, P. Saha, P.K. Mishra, Taguchi analysis of surface
modification technique using W-Cu powder metallurgy sintered
30
tools in EDM and characterization of the deposited layer. Int.
1
J. Adv. Manuf. Technol. 54(5), 593–604 (2011)
20
4. K. Suryanarayanan, R. Praveen, S. Raghuraman, Silicon carbide
reinforced aluminium metal matrix composites for aerospace
10
6 2,3 3 5 5
applications: a literature review. Int. J. Innov. Res. Sci. Eng.
35
3 6 2 46 4 2,3 Technol. 2(11), 6336–6344 (2013)
0 5. R. Arsenault, The strengthening of aluminum alloy 6061 by fiber
30 40 50 60 70 80 90
and platelet silicon carbide. Mater. Sci. Eng. 64(2), 171–181
2θ angle → (1984)
6. M. Dave, K. Kothari, Composite material-aluminium silicon
Fig. 18 XRD plot of a deposited layer using 15 Ton, SiC50:Cu50 alloy: a review. Paripex Indian J. Res. 2(3), 148–150 (2013)
compact tool 7. International advanced research centre for powder metallurgy and
new materials (1997), (http://www.arci.res.in/index.php)
8. W.H. Yang, Y.S. Tarng, Design optimization of cutting param-
1 with COM value of 0.917) came out as the optimal eters for turning operations based on the Taguchi method.
level of parameters, i.e., compaction load at 5 Ton, J. Mater. Process Technol. 84(1–3), 122–129 (2007)
current at 2 Ampere, pulse duration at 11 ls and tool 9. M. Nalbant, H. Gokkaya, G. Sur, Application of Taguchi method
in the optimization of cutting parameters for surface roughness in
composition of 30:70 by % weight of SiC and Cu. turning. Mater. Des. 28(4), 1379–1385 (2007)
3. ANOVA result shows that peak current is the most 10. B. Das, S. Roy, R.N. Rai, S.C. Saha, Surface roughness of Al-
influencing parameter (57.05 %) that significantly 5Cu alloy using a Taguchi-fuzzy based approach. J. Eng. Sci.
affects the roughness parameters followed by pulse Technol. Rev. 7(2), 217–222 (2014)
11. C.C. Kao, J.S. Albert, S.F. Miller, Fuzzy logic control of Micro-
duration (28.03 %), compaction load (12.34 %) and hole electrical discharge machining. J. Manuf. Sci. Eng.
tool composition (0.34 %). 130(064502), 1–6 (2008)
4. To achieve a good surface finish, higher compaction 12. C. Venkatesha, N.M. Arunb, R. Venkatesan, Optimization of
load can be used for preparing powder tool in EDC micro drilling parameters of B4C DRMM Al 6063composite in
lECM using Taguchi coupled fuzzy logic. Proc. Eng. 97,
process. 975–985 (2014)
5. Surface roughness (Ra, Rz and Sa) increases with 13. R. Thanigaivelan, R.M. Arunachalam, J. Jerald, T. Niranjan,
increasing percentage of SiC in tool electrode material Applications of Taguchi technique with fuzzy logic to optimise
with higher current and pulse duration. an electrochemical micromachining process. Int. J. Exp. Des.
Process Optim. 2(4), 283–298 (2011)
6. Lower peak current and lower pulse duration settings 14. A.A.D. Sarhan, L.S. Fen, M.W. Yip, M. Sayuti, Fuzzy modeling
are also suggested for obtaining a good surface finish for micro EDM parameters Optimization in drilling of biomedical
in EDC process. implants Ti-6Al-4V alloy for higher machining performance. Int.
7. Microscopic image and XRD analysis of the deposited J. Mech. Aerosp. Ind. Mech. Eng. 9(1), 197–201 (2015)
15. S.S. Agrawal, V. Yadava, Development and experimental study
layer confirmed the formation of a composite layer of surface-electrical discharge diamond grinding of Al–10 wt%
composed of Si, SiC, Cu, Al, Al2Cu, and Al2CuMg on SiC composite. J. Inst. Eng. India Ser. C 97(1), 1–9 (2016)
the work surface due to the transfer of tool materials. 16. R. Mohan, P.C. Josephkunju, B. Paul, Multi output optimization
of CNC high speed hard turning of AISI 52100 bearing steel
From this study, it can be concluded that the proposed using Taguchi method and fuzzy logic unit. Int. J. Eng. Trends
methodology can be taken as an effective approach for Technol. 15(3), 118–123 (2014)
17. P. Sengottuvel, S. Satishkumar, D. Dinakaran, Optimization of
optimizing the multiple surface roughness parameters in multiple characteristics of EDM parameters based on desirability
industrial experiments and also helps researchers. approach and fuzzy modeling. Proc. Eng. 64, 1069–1078 (2013)
18. C.K. Biswas, S. Dewangan, Optimisation of EDM Process with
Acknowledgments The present paper is an extended version of the Fuzzy Logic Technique. in International conference on metal-
paper entitled ‘Optimization of Surface Roughness Parameters of Al- lurgical, manufacturing and mechanical engineering 26–27 Dec
6351 Alloy in EDC Process: A Taguchi Coupled Fuzzy Logic 2012, pp. 346–349
Approach’, presented during 30th National Convention of Production 19. S.Y. Martowibowo, A. Wahyudi, Taguchi method implementa-
Engineers, held at NIT Tripura during July 18–19, 2015. The authors tion in Taper motion wire EDM process optimization. J. Inst.
are thankful to the Organising Committee. Eng. India Ser. C 93(4), 357–364 (2012)

123
Author's personal copy
J. Inst. Eng. India Ser. C

20. R.A. Fisher, Statistical methods for research worker (Oliver & 25. A. Das, J.P. Misra, Experimental investigation on surface modi-
Boyd, London, 1925) fication of aluminum by electro discharge coating process using
21. Y. Chen, S.M. Mahdavian, Parametric study into erosion wear in TiC/Cu green compact tool electrode. Mach. Sci. Technol. Int. J.
a computer numerical controlled electro-discharge machining 16(4), 601–623 (2012)
process. Wear 236(1–2), 350–354 (1999) 26. P.K. Patowari, U.K. Mishra, P. Saha, P.K. Mishra, Surface
22. I. Puertas, C.J. Luis, L. Álvarez, Analysis of the influence of integrity of C-40 steel processed with WC-Cu powder metallurgy
EDM parameters on surface quality, MRR and EW of WC-CO. green compact tools in EDM. Mater. Manuf. Process. 26(5),
J. Mater. Process Technol. 153–154, 1026–1032 (2004) 668–676 (2011)
23. I. Puertas, C.J. Luis, G. Villa, Spacing roughness parameters 27. P.K. Patowari, P. Saha, P.K. Mishra, An experimental investi-
study on the EDM of silicon carbide. J. Mater. Process Technol. gation of surface modification of C-40 steel using W-Cu powder
164–165, 1590–1596 (2005) metallurgy sintered compact tools in EDM. Int. J. Adv. Manuf.
24. M.E. Krishna, P.K. Patowari, Parametric study of electric dis- Technol. 80(1), 343–360 (2015)
charge coating using powder metallurgical green compact elec-
trodes. Mater. Manuf. Process. 29(9), 1131–1138 (2014)

123

You might also like