International Journal of Engineering Trends and Technology (IJETT) – Volume 48 Number 1 June 2017

Optimization of Process Parameters for

Milling Operation using Taguchi Method
K.Prasadraju #1, M. Satish raja*2, V.Praveen#3, I.Ajith kumar,#4
#
Assistant professor & mechanical department & S.R.K.R Engineering. college
Bhimavaram, Ap, India

Abstract Quality and productivity play a major role TABLE: The table rest on ways on the saddle and
in today's manufacturing market. Due to this surface travels longitudinally. The top of the table is
finish & dimensional accuracy becomes very accurately finished and T-slots are provided for
important. Milling like any metal cutting operation clamping the work and other fixtures on it. A lead
is used with an objective of optimizing surface screw under the table engages a nut on the saddle to
roughness at micro level and economic performance move the table horizontally by hand or power.
at macro level. OVER HANGING ARM: Over hanging arm is
The experiments have been planned using Taguchi’s mounted on the top of column extends beyond the
experimental design technique. A L9 orthogonal column face and serve as a bearing support may be
array, taguchi method and analysis of variance provided nearest to the cutter. More than one bearing
(ANOVA) are used to formulate the experimental support may be provided for the arbor.
layout, to analyses the effect of each parameter on FRONT BRACE: The front brace is an extra support
the machining characteristics and to predict the that is fitted between the knee and over arm to
optimal choice for each milling parameter such as ensure further rigidity to the arbor and the knee. The
spindle speed, feed rate and depth of cut. front brace is slotted to allow for adjustment of the
In cutting process, optimization of cutting height of the knee relative to over arm.
parameters is considered to be a vital tool for SPINDLE: The spindle of the machine is locates in
improvement in output quality of a product as well the upper part of the column and receive power from
as reducing the overall production time. the motor through belts,.
ARBOR: An arbor is considered as an extension of
Keywords—Anova, MRR, surfaceroughness, the machine spindle on which cutters are securely
Taguchimethod . mounted and rotated..
.
I.INTRODUCTION PARTS OF MILLING MACHINE:
Milling is the machining process of using rotary A. Base
cutters to remove material from a work piece B. Column
advancing (or feeding) in a direction at an angle with C. Knee
the axis of the tool. It covers a wide variety of D. Saddle
different operations and machines, on scales from E. Table
small individual parts to large, heavy-duty gang F. Over Hanging
milling operations. It is one of the most commonly Arm
used processes in industry and machine shops today G. Front Brace
for machining parts to precise sizes and shapes H. Spindle
I. Arbor
BASE: The base of the machine is Grey iron casting
accurately machined on its top and bottom surface II. INVESTIGATION OF PROCESS
and serves as a foundation member for all the other.
COLUMN: The column is the main supporting There are number of statistical techniques
frame mounted vertically on the base. The column is available for engineering and scientific studies.
box shaped. Heavily ribbed inside and houses all the Taguchi prescribed a standard way to utilize DOE
driving mechanisms for the spindle and table feed.. (Design of Experiments) technique to enhance the
KNEE: The knee is the rigid gray iron casting that quality of products and process. DOE using Taguchi
slides up and down on the vertical way of the approach is a handy statistical tool to improve
column face. The adjustment of height is effected by consistency of performance, to build insensitivity
elevating screw on the base that also supports the towards uncontrollable factors in optimizing
knee. The knee houses the feed mechanism of the manufacturing process design, solving
table, and in different controls to operate it manufacturing and production problems and in
SADDLE: The saddle is placed on the top of the determining the best assembly method etc. It is
knee, which slides on guide ways set exactly at 90 to possible to reduce the time required for experimental
column face. investigation and improve process quality by
applying Taguchi technique.

ISSN: 2231-5381 http://www.ijettjournal.org Page 1

 International Journal of Engineering Trends and Technology (IJETT) – Volume 48 Number 1 June 2017

The L9 Orthogonal array has been selected as one randomization and blocking to ensure the validity of
of the processes of the investigation. In this process, the results
there are nine numbers of trials and in each one of
the trials, there will be two degrees of freedom. III EXPERIMENTALWORK AND ANALYSIS
There are three parameters and they are in three
levels (33), each of the parameters is changed in In this investigation an attempt was made to find out
three levels. The response characteristics of the job the optimum process parameters of MILLING on
are measured after completing the experiment. So, AISI 304 plates. Process parameters considered are
the process has to be conducted with L9 interactions. cutting speed, depth of cut and feed rate.
The response characteristic is surface roughness. A rectangular mild steel plate of size 100
The optimum process parameters identified are those mm×83mm×20mm in shaping machine for
which give smaller the better. performing vertical milling. In milling, the speed
and motion of the cutting tool is specified through
B.TAGUCHI METHOD several parameters. These parameters are selected
Taguchi outlined three step approach for for each operation based upon the work piece
assigning nominal values and tolerances to product material, tool material, tool size, and more.
and process design characteristics:
System design A.CUTTING SPEED
Parameter design and The speed of the work piece surface relative to the
Tolerance design edge of the cutting tool during a cut, measured in
System design is the basic prototype design to surface feet per minute (SFM).
achieve desired function and parameter design is to The cutting speed (Vc) for milling is defined as
specify levels of control factors that are relatively the peripheral speed of the cutter speed (Vc) is given
insensitive to noise factors. If parameter design fails Vc = лDN m/min
to produce adequately low functional variation of 1000
product, then tolerance design is helpful.
Parameter design experiments can be either B.FEED RATE
physical experiments or computer based simulation The feed (f) in a milling machine is defined as the
trails. Experimenter has to identify list of control movement of the work piece relative to cutter axis. It
parameters and levels of interactions array is is the rate at which the work piece is fed into the
selected based on degrees of freedom of all factors cutter.
and interactions put together Feed rate tooth (mm per tooth of the cutter)
Feed per revolution (mm per revolution of the cutter)
C.ORTHOGONAL ARRAYS Feed per minute (mm per minute)
Taguchi has developed a system of tabulated Feed per minute (table feed) Φ = feed per rev*
designs (arrays) that allow for the maximum number cutter speed (r.p.m), Ft=feed rate per tooth,
of main effects to be estimated in an unbiased F (mm/min) = ft x Z x N
(orthogonal) manner, with a minimum number of Z = Number of teeth on the cutter periphery.
runs in the experiment.
The Taguchi technique involves reducing the C.DEPTH OF CUT
variation in a process through robust design of In the milling process, the „depth of cut (d) is
experiments. The overall objective of the method is defined as the thickness of the layer of material of
to produce high quality product at low cost to the material removed in one pass of the work piece
manufacturer. Taguchi‟s parameter design is an under the cutter. A depth of cut from 3 mm to 8 mm
important tool for robust design. It offers a single is common for roughing cuts and is less than 1.5 mm
and systematic technique to optimize the design for finishing cuts.
performance, quality and cost
D.MATERIAL REMOVAL RATE
D.ANOVA TEST Material removal rate (MRR) is the volume of
Tests of hypothesis (null hypothesis) deal with material removed in unit time. For milling MRR is g
testing equality of at most two means. If one is MRR = B* d* f, mm3/min
interested in testing equality of several means at a Where, B= width of cut, D= depth of cut
time technique of analysis of variance is used. Based F= rate of feed
on F-ratios computed in ANOVA the significance of
each factor and interaction can be decided. E.MACHINING TIME
The terminology of ANOVA is largely from the Machining time tm is defined as the time required
statistical design of experiments. The experimenter for one pass of width of cut
adjusts factors and measures responses in an attempt tm =length of cut = L/ f = L/ ft*Z*N min
to determine an effect. Factors are assigned to feed rate
experimental units by a combination of

ISSN: 2231-5381 http://www.ijettjournal.org Page 2

 International Journal of Engineering Trends and Technology (IJETT) – Volume 48 Number 1 June 2017

Table 3.0: Milling Parameters and their Levels Mild steel plates are prepared as per the
S.No. Control Factors Factor Levels experimental plan given in above Table . For each of
Units L1 L2 L3
Factors Notation the plates made the responses are the surface
Spindle roughness and material removal rate which are
1 rpm N 1000 1250 1500
Speed experimentally determined. After getting the
Feed
experimental results, the results are analysed to
2 mm/min F 100 150 200 arrive at optimum values of process parameters.
Rate
Depth Table 3.3:Experimental Results for the Milling
3 mm dc 0.25 0.50 0.75
of Cut
Material
Spindle Depth Surface
In the present investigation three factors are Trail Feed rate Removal
considered at three levels which yield to total speed of cut roughness
no „mm/min‟ Rate
degrees of freedom of „6‟. So a L9 array with total „rpm‟ „mm‟ „Ra‟
degrees of freedom of „8‟ is chosen for „g/min‟
experimentation [16, 18, and 14]. This L9 array can
accommodate the entire 3-level factor considered is
1 1000 100 0.25 4.02 0.98
given in Table 3.1 which shows L9 array. Actual
level values of parameter are given in Table3. 2
2 1000 150 0.50 3.86 0.93
Table3.1: L9 (33) Orthogonal Array
Trial Milling parameter levels 3 1000 200 0.75 3.80 0.86
no A B C D
1 1 1 1 1 4 1250 100 0.75 3.85 0.89
2 1 2 2 2
3 1 3 3 3 5 1250 150 0.25 3.82 0.90
4 2 1 3 2
5 2 2 1 3 6 1250 200 0.50 3.21 0.83
6 2 3 2 1
7 3 1 2 3 7 1500 100 0.50 3.00 0.85
8 3 2 3 1
9 3 3 1 2 8 1500 150 0.75 3.24 0.97

Table3.2: L9 Orthogonal Array with Actual Values of

Factor Levels 9 1500 200 0.25 3.25 0.90

Milling parameter levels

Trial No IV RESULTS AND DISCUSSIONS
A B C I. OPTIMIZATION PROCESS PARAMETERS FOR
SURFACE ROUGHNESS
1 1000 100 0.25 Table 4.1: Parameter Levels and Response of Surface
Roughness
2 1000 150 0.50 Trial Milling parameter levels and response
3 1000 200 0.75 No A B C Surface roughness
4 1250 100 0.75 1 1 1 1 3.52
2 1 2 2 3.80
5 1250 150 0.25
3 1 3 3 3.50
6 1250 200 0.50 4 2 1 3 3.85
7 1500 100 0.50 5 2 2 1 3.82

8 1500 150 0.75 6 2 3 2 3.21

7 3 1 2 3.00
9 1500 200 0.25
8 3 2 3 3.24
9 3 3 1 3.25

ISSN: 2231-5381 http://www.ijettjournal.org Page 3

 International Journal of Engineering Trends and Technology (IJETT) – Volume 48 Number 1 June 2017

Table 4.2: Influence of each Process Parameter on C.OPTIMUM CONDITION

Surface Roughness After performing ANOVA it is observed that the
optimum condition for smaller surface roughness is
S.NO. Parameters L1 L2 L3 spindle speed at level-2, depth of cut at level-1 and
SPINDLE the values of each factor is given in Table 4.4.
1 3.60 3.62 3.16
SPEED
FEED Table 4.4: Optimum Condition for Surface Roughness
2 3.45 3.64 3.31
RATE
DEPTH OF
3 3.54 3.33 3.53 Factor % Level
CUT S.No. Notations
Name Contribution Description
Spindle
1 N 29.53 (1500)3
speed
Feed
2 F 9.92 (200)3
rate
Depth
3 Dc 53.64 (0.50)2
of cut

D.STEPS INVOLVED TO FIND THE OPTIMUM

Fig 4.1: Influence of each Process Parameter on CONDITION USING ANOVA
Surface Roughness 1. Total of all the results =
2. Correction Factor (C.F.)
A.RESPONSE GRAPH
C.F. = Where, T = Total number of
. Response graph for surface roughness is shown in
Fig 4.1. Quality characteristics for surface roughness results, n = Number of experiments
is smaller the better. At third level of spindle speed
i.e. rpm, at third level of feed rate i.e.mm and at 3. Total sum of squares, ST
second level of depth of cut i.e.mm ST = - C.F.
However, the significant and insignificant 4. Factor sum of squares, SA
parameter will be discriminated based on percentage 5. Degree of freedom, ft
contribution of each factor toward surface roughness. 6. Factor degree of freedom, fA
B.ANOVA 7. Variance, V
The surface roughness values obtained is different, 8. % Contribution, PA
the experimental trial combination given in Table
3.7, Analysis of variance (ANOVA) is performed SAMPLE CALCULATIONS
and results are given in Table 4.3. Percentage yA1 = = 3.60 yA2 = =3.62
contribution of each factor is depicted in the form of
bar graph in Fig 4.2. It can be observed from Fig 4.2 yA3 = = 3.16 yB1 = = 3.45
that Depth of cut has get major contribution towards yB2 = = 3.62 yB3 = = 3.32
variation in surface roughness, next best significant
parameters is spindle speed and next best significant yC1 = = 3.53 yC2 = = 3.33
parameters is feed rate. Hence, spindle speed and yC3 = = 3.53
Depth of cut are significant parameter which must be
maintained at the levels specified i.e. Depth of cut at = = 3.49
level-2 and Spindle speed at level-3 other parameter Degree of Freedom, ft Factor Degree of
can be maintained at any one of the level values Freedom, fA
specified based on cost consideration ft = n – 1 , ft = 9 – 1, ft = 8 fA = No.of
Levels – 1 , fA = 3 – 1, fA = 2
Table 4.3: Analysis of Variance (ANOVA) for Surface
Roughness E.SURFACE ROUGHNESS VALUES AT OPTIMUM
% CONDITION
Source DOF Seqss Adjms Grand Average of standard value of surface
contribution
Spindle roughness =3.49
2 0.4116 0.2058 29.53798459
speed Expected surface roughness at optimum condition
Feed Y Optimum = + + + =
2 0.1383 0.06915 9.924935054
rate 3.49+ (3.6-3.49) + (3.31-3.49) + (3.33-3.49) = 2.82
Depth (STANDARD VALUE based on the significance of
2 0.748 0.37378 53.64775451
of cut factors and their percentage contribution towards
Error 2 0.096 0.048 6.889325851 result)
Total 8 1.39 100

ISSN: 2231-5381 http://www.ijettjournal.org Page 4

 International Journal of Engineering Trends and Technology (IJETT) – Volume 48 Number 1 June 2017

Instead of conducting „27‟ experimental trial i.e. rpm, at second level of feed rate i.e.mm and at
combination by varying one factor at a time i.e. full first level of depth of cut i.e.mm
factorial experiments, with the help of Taguchi‟s L9 However, the significant and insignificant parameter
Orthogonal array optimized condition can be will be discriminated based on percentage
obtained by analysing the results of nine contribution of each factor toward material removal
experimental trials. Interestingly the obtained rate.
optimum condition does not match with any of the B.ANOVA
experimental trial combination existing in L9 array. The material removal rate values obtained is
It can be observed that the surface roughness value different, the experimental trial combination given in
obtained at optimum condition is less than any of the Table 3.7, Analysis of variance (ANOVA) is
surface roughness values obtained experimentally as performed and results are given in Table 4.3.
per L9 Orthogonal array. Percentage contribution of each factor is depicted in
the form of bar graph in Fig 4.2. It can be observed
II OPTIMIZATION PROCESSPARAMETERS FOR from Fig 4.2 that feed rate has get major contribution
MATERIAL REMOVAL RATE(MRR) towards variation in material removal rate, next best
significant parameters is Depth of cut and next best
Table 4.5: Parameter Levels and Response of significant parameters is spindle speed. Hence, feed
Material Removal Rate (MRR) rate and Depth of cut are significant parameters
which are higher percentage of contribution must be
Milling parameter levels and response maintained at the levels specified i.e., feed rate at
TRAIL MRR level-2, Depth of cut at level-1 and Spindle speed at
NO A B C
level-1.
1 1 1 1 0.98 Table 4.7: Analysis of Variance (ANOVA)
2 1 2 2 0.93 for Material Removal Rate
3 1 3 3 0.86
4 2 1 3 0.89 %
5 2 2 1 0.90 Source DOF Seqss Adjms
contribution
6 2 3 2 0.83 Spindle 3.89E-
7 3 1 2 0.85 2 0.001944 19.47
speed 03
8 3 2 3 0.97 Feed 9.62E-
2 0.004811 48.19
9 3 3 1 0.90 rate 03
Depth 4.96E-
2 0.002478 24.81
Table 4.6: Influence of each Process Parameter of cut 03
on Material Removal Rate Error 2 0.0015 0.00075 7.51
2.00E-
Parameters L1 L2 L3 Total 8 100
02
SPINDLE
2.77 2.62 2.72
SPEED C.OPTIMUM CONDITION
FEED
2.72 2.83 2.62
RATE After performing ANOVA it is observed that the
DEPTH OF optimum condition for higher material removal rate
2.78 2.61 2.72
CUT spindle speed at level-2, depth of cut at level-1 and
the values of each factor is given in Table 4.4.

Table 4.8 Optimum Conditions for Material

Removal Rate
Factor % Level
S .No Notations
Name Contribution Description
Spindle
1 N 19.47 (1000)1
Fig 4.3 Influence of each Process Parameter on
speed
Material Removal Rate Feed
2 F 48.19 (150)2
rate
A.RESPONSE GRAPH Depth
3 Dc 24.81 (0.25)1
This graph indicates the influence of each process of cut
parameter on material removal rate. Response graph
for material removal rate is shown in Fig 4.1.
Quality characteristics for material removal rate are
the higher the better. At first level of spindle speed

ISSN: 2231-5381 http://www.ijettjournal.org Page 5

 International Journal of Engineering Trends and Technology (IJETT) – Volume 48 Number 1 June 2017

D.MATERIAL REMOVAL RATE VALUES AT roughness, combinations of parameters and their

OPTIMUM CONDITION levels are also predicted by Taguchi method.
Grand Average of standard value of material By the experiment results it was found that the
removal rate =0.93 surface roughness quality characteristic is smaller
Expected material removal rate at optimum the better but the experimental value is 3.00mm i.e.,
condition at parameters S3, F1, D2 and for material removal
Y Optimum= + + + rate quality characteristic is bigger the better but
experimental value is i.e., 0.98 at S1, F1, D1.
= 0.903+ (2.77-0.903) + (2.83-0.903) + (2.78-0.903)
= 5.97
After applying Taguchi techniques the predicted
(STANDARD VALUE based on the significance of
values are 2.82mm and material removal rate is
factors and their percentage
5.97mm. The values obtaining after applying
contribution towards result)
Taguchi technique is more effective than the
Instead of conducting „27‟ experimental trial
experimental values.
combination by varying one factor at a time i.e. full
factorial experiments, with the help of Taguchi‟s L9
By ANOVA techniques, influence of each milling
Orthogonal array optimized condition can be
parameter is studied and the prediction of the surface
obtained by analysing the results of nine
roughness and material removal rate is done.
experimental trials. Interestingly the obtained
Analysis of surface roughness and material removal
optimum condition does not match with any of the
rate parameters such as spindle speed, feed rate and
experimental trial combination existing in L9 array.
It can be observed that the material removal rate depth of cut against variations in milling.
value obtained at optimum condition is less than any REFERENCES
of the material removal rate values obtained
experimentally as per L9 Orthogonal array. [1] Milon D. Selvam, Dr.A.K.Shaik Dawood,Dr. G.
Karuppusami, “Optimization Of Machining Parameters For
Face Milling Operation In A Vertical CNC Milling
V. CONCLUSIONS Machine Using Genetic Algorithm”, An International
After conducting the experiments and analysing the Journal (ESTIJ), ISSN: 2250-3498, Vol-2, August (2012).
[2] Anil Choubey, Vedansh Chaturvedi, Jyoti Vimal, “The
experimental results the following conclusions are Implementation Of Taguchi Methodology For
made Optimization Of End Milling Process Parameter Of Mild
Steel”, International Journal of Engineering Science and
Taguchi method has been successfully employed for Technology, ISSN : 0975-5462, Vol-2, (2007)
[3] Nafis Ahmad, Tomohisa Tanaka and Yoshio Saito,
optimizing the process parameters of Milling of mild “Optimization Of Cutting Parameters For End Milling
steel plates. It has been shown that the Taguchi Operation By Soap Based Genetic Algorithm”, ICME05-
method provides a systematic and efficient AM-08, August (2006).
methodology for searching the milling process [4] John D. Kechagias, Christos K. Ziogas, Menelaos K.
Pappas, Ioannis E. Ntziatzias, “Parameter Optimization
parameters with optimal milling parameters. during Finish End Milling of Al Alloy 5083 using Robust
Design”, ISBN: 978-988-18210-6-5 (2011).
As per L9 orthogonal array, we have 3 3=27 [5] R. JaliliSaffar, M.R. Razfar, A.H. Salimi and M.M. Khani,
combinations. Instead of 27 experiments, nine “Optimization of Machining Parameters to Minimize Tool
Deflection in the End Milling Operation Using Genetic
numbers of trials were conducted. The optimum Algorithm”, World Appl. Sci. J., 6 (1): 64-69, 2009.
value for surface roughness and material removal [6] S.S.K. Deepak, “Applications of Different Optimization
rate is not available in the nine numbers of Methods for Metal Cutting Operation – A Review”,
experiments. The optimum values of surface Research Journal of Engineering Sciences, Vol. 1(3), 52-58,
September (2012).

ISSN: 2231-5381 http://www.ijettjournal.org Page 6