18MEO113T - DOE - Unit 3 - AY2023 - 24 ODD

Subject Code
18MEO113T - Design of Experiments
Handled by
S. Murali, Ph.D.,
Associate Professor,
Department of Mechanical Engineering,
SRM IST, Kattankulathur.
Disclaimer
The content prepared in the presentation are from various sources, only used for education pur-
pose. Thanks to all the sources.
18MEO113T – Design of Experiments
• To learn the fundamentals of design of experiment techniques.
• To familiarize in how to setup experiments and accomplish all analyze tasks using software
packages like Minitab, etc.
• After this course, You will be ready to apply the technique confidently to all of your projects .
Unit 3
1: Introduction to Robust design, Loss functions.

2: Eight steps in Taguchi methodology
3: Orthogonal array, Selecting the interaction, Linear graphs
4: S/N ratio: Larger-the-better, Smaller-the-better, Nominal-the-best
5: Analyze the data, factor effect diagram
6: Levels of parameters
7: Confirmation test
8: Augmented design with simple case
9: Solving Case studies on robust design with statistics software
Introduction
• Robust Design method, also called the Taguchi Method,

pioneered by Dr. Genichi Taguchi, greatly improves
engineering productivity.
• Genichi Taguchi (1 January 1924 – 2 June 2012) was a
statistician and engineer. He obtained his Ph.D. in Statistics
and Mathematics Science at Kyushu University University,
Japan in 1962.
Introduction
• Genichi Taguchi Quotes:
“Cost is more important than quality but quality is the best way to reduce cost.”
“A scientific or technical study always consists of the following three steps:

1. One decides the objective.
2. One considers the method.
3. One evaluates the method in relation to the objective.”
Taguchi Method is quite different:
For examples, if seven variables can each be set at two levels, then the total
number of experiments possible is given by
Level of the power of factors, that is (27) = 128.
• It allows us to vary all the factors at a time and still let us evaluate the effect
of each individual factor. Also, it allows us to do the limited number of
experiments and then provides a method for predicting the other untried
combinations.
• For examples, in the cited example, Taguchi method suggests doing only 8
experiments to understand the 128 combinations.
6
A well-known example of Taguchi designs is from the Ina Tile Company of Japan in the 1950s.
The company was manufacturing too many tiles outside specified dimensions.
A quality team discovered that the temperature in the kiln used to bake the tiles varied, causing
nonuniform tile dimension. They could not eliminate the temperature variation because building a
new kiln was too costly. Thus, temperature was a noise factor. Using Taguchi designed
experiments, the team found that by increasing the clay's lime content, a control factor, the tiles
became more resistant, or robust, to the temperature variation in the kiln, letting them
manufacture more uniform tiles.
7
• In Taguchi designs, a measure of robustness used to identify control
factors that reduce variability in a product or process by minimizing the
effects of uncontrollable factors (noise factors).
• In a Taguchi designed experiment, you manipulate noise factors to force
variability to occur and from the results, identify optimal control factor
settings that make the process or product robust, or resistant to variation
from the noise factors.
8
Background of the Taguchi Method
• Robust Design method (or) Taguchi Method, pioneered by Dr. Genichi Taguchi, in 1980.
– Comparable to importance to Statistical Process Control (SPC), the Deming approach
and the Japanese concept of TQC.
TQM expresses about continuous

improvement in the processes while QC
is about maintaining the quality
standards throughout the process.
QC is used to verify the quality of the

output; TQM is the process of
managing for quality.
Taguchi Terminology
• A schematic representation of a product / process is shown in a figure. The
response is denoted by y and is called quality characteristics.
10
Taguchi Terminology
• The parameters that influence the quality characteristic are classified as
follows:
• Signal factor (M): These are the parameters set by the user or operator of
the product to express the intended value for the response of the product
based on the knowledge of the product being developed. Sometimes 2 or
more signal factors may also be used in combination to express the desired
response.
11
Taguchi Terminology
• Noise factor (x): Certain parameters are not controllable by the designer
and are known as noise factors. Parameters whose settings (also called
levels) are difficult to control in the field or whose levels are expensive to
control are also considered noise factors. The levels of the noise factors
change from one unit to another, from one environment to another, and from
time to time. The noise factors cause the response y to deviate from the
target specified by the signal factor M and lead to quality loss.
• Control factor (z): These are parameters that can be specified by the
designer. Control factor can take multiple values, called levels.
12
Control or Noise Factor (Example)
• Most automotive sub-
assemblies like the
alternator, the ignition coil,
and the electronic control
module must undergo testing
to determine if they are
resistant to salt water that
may be splashed on them
from the road.
• An automotive supplier is
testing an ignition coil to
determine if it will withstand
salt water. The following
factors are tested :
13
• Unique aspects of the Taguchi Method
– The Taguchi definition of quality
– The concept of Robust Design
– The Taguchi Quality Loss Function (QLF)
1. The Taguchi definition of quality

– Ideal quality refers to a target value for determining the quality level
– Ideal quality is delivered if a product or service tangible performs its intended function
throughout its projected life under reasonable operating conditions without harmful side
effects
– Ideal quality is a function of customer perception and satisfaction
– Service quality is measured in terms of loss to society
What is robust design?
2. The concept of Robust Design

• Robust design is an “engineering methodology for improving productivity
during research and development so that high-quality products can be
produced.
• Robust Design is a technique that reduces variation in a product by reducing
the sensitivity of the design of the product to sources of variation rather
than by controlling their sources. The end result is a robust design, a design
that has minimum sensitivity to variations in uncontrollable factors
• What is Robust?
– A system (product or process) is robust if it performs properly in a wide range of conditions.
Robust Products Products that are not robust
A pen that writes until the ink is empty Pen that stops writing after a few months
A car that starts at -20 deg A car that does not start
A vacuum cleaner that maintains suction A vacuum cleaner that loses suction
levels
17
• What is Robust?
– In the design of a new product, any design activity can be called robust, if it leads the
product;
1. To have longer life (higher reliability)
2. To be more consistent with use
3. To be more consistent from product to product
4. To perform consistently as temperature and other conditions change
18
• Products and services should be designed to be inherently defect free and high
quality
– Meet customer’s expectations also under non-ideal conditions
• Disturbances are events that cause the design performance to deviate from its
target value
• Taguchi divide disturbances into three categories
– External disturbances: variation in the environment where the product is used
– Internal disturbances: wear and tear inside a specific unit
– Disturbances in the production process: deviation from target values

• A three step method for achieving robust design (Taguchi)

1. Concept design
2. Parameter design
3. Tolerance design
1. Concept Design / System Design

– The process of examining competing technologies for producing a product – Includes choices of
technology and process design.
– A prototype design that can be produced and meets customers needs under ideal conditions
without disturbances
2. Parameter Design
– The selection of control factors (parameters) and their optimal levels
• The objective is to make the design Robust!
– Control factors are those process variables in a process that management

can influence
• Ex. The procedures used and the type and amount of training
• Often a complex (non-linear) relationship between the control factors and product / design
performance
– The optimal parameter levels can be determined through experimentation.

3. Tolerance Design
– Tolerance design deals with developing specification limits.
• Necessary because there will always be some variation in the production process
• Taguchi fiercely advocates aiming for the target value not just settle for inside the specification limits.
– Tolerance design occurs after parameter design has been used to reduce variation and the
resulting improvement has been insufficient.
– Often results in increased production costs
• More expensive input material might have to be used to meet specifications.
– Interchangeability.
3. Taguchi Quality/Quadratic Loss Function (QLF)

– Taguchi defines Quality as “the loss imparted to the society by the product from the time the
product is designed to the time it is shipped to the customer”
– LOSS = Cost to operate, Failure to function, maintenance and repair cost, customer
satisfaction, poor design.
<Traditional approach> <Taguchi approach>

The Taguchi Quality/ Quadratic Loss Function (QLF)
<Taguchi approach> <Traditional approach> <Taguchi approach>
25
The Taguchi Quality/Quadratic Loss Function (QLF)
Where y is the critical performance
parameter value,
L is the loss associated with a particular
parameter y,
m is the nominal value of the parameter
specification (ie. Target value of y),
k is a constant that depends on the cost at
the specification limits. (ie. Quality loss
coefficient)
26
The Taguchi Quality Loss Function (QLF)
27
• The loss is proportional to square of the deviation of y from the target value.
• k is proportionality constant.
28
29
Quality Evaluations
• According to the traditional concept of loss function that as long as the
product’s quality characteristics is within the specification limits or
tolerances, there are no losses incurred. The lose takes the form of a
step function with a constant value A outside of the specification limits or
tolerances.
• Taguchi defines quality as “the total loss imparted to society from the
time the product is designed to the time product is shipped to the
customer.” 30
Quality Evaluations
• Loss function L(y) makes no distinction between products

– Whose quality characteristic is exactly on target at m, or
– Whose quality characteristic is just above lower specification limit (LSL), or
– Whose quality characteristics is just below the upper specification limit (USL)
31
Quality Evaluations
• The conventional method of computing the cost of quality is based on the number of parts rejected and
reworked in a production environment. This method of quality evaluation is incapable of distinguishing
between two samples, both within the specification limits but with different distributions of targeted
properties.
Taguchi’s definition clearly puts more emphasis on

customer satisfaction, whereas previously all definitions
were concerned with the producer. Optimum customer
satisfaction can be achieved by developing the products
that meet the target value on a consistent basis.
https://brharnetc.edu.in/br/wp-content/uploads/2018/11/11.pdf
32
Quality Evaluations
• It is also doubtful that the loss function L(y) remains constant at a value of A
beyond the specification limits USL and LSL.
• Since the Taguchi loss function increases quadratically with increasing
deviation from the target value, the Taguchi loss function can be applied
to three different situations or cases, namely

– Target is best (or) Nominal is best
– Smaller is better
– Larger is better
33
Quality Evaluations
Target is best (or) Nominal is best
Larger is better
Smaller is better
Ranjit K . Roy, A Primer on the Taguchi Method, Second
34
Quality Evaluations
Target is best or The Nominal-the-Best Case
35
Quality Evaluations
Target is best or The Nominal-the-Best Case
36
Quality Evaluations
• Example Sony TV
37
Quality Evaluations (S/N Ratio)
• Taguchi experiments often use a 2-step optimization process.
• In step 1, use the signal-to-noise ratio to identify those control factors that
reduce variability.
• In step 2, identify control factors that move the mean to target and have a
small or no effect on the signal-to-noise ratio.
38
• Higher values of the signal-to-noise ratio (S/N) identify control factor
settings that minimize the effects of the noise factors.
• The signal-to-noise ratio measures how the response varies relative to
the nominal or target value under different noise conditions. You can
choose from different signal-to-noise ratios, depending on the goal of your
experiment.
• The important contribution of Taguchi is proposing the signal-to-noise (S/N)
ratio. It was developed as a proactive equivalent to the reactive loss
function.
39
Mean
Standard Deviation
or Variation
https://www.slideshare.net/rbalisnomo/Introduction-To-Taguchi-Method-
05Sep08
40
Signal-to-Noise Ratio
41
42
43
44
45
S/N Ratio
46
Taguchi Analysis (Example)
L(y) = k (y – m)2 L(y) = ky2 L(y) = k (1/y2)

47
Quality Evaluations
Smaller – the – better loss function
• No loss if the quality characteristics is
below the upper limit
• Loss is A0 if the quality characteristics
is above upper limit
• L(y) = ky2
Larger – the – better loss function

• No loss if the quality characteristics is
above the lower limit
• Loss is A0 if the quality characteristics
is below lower limit
• L(y) = k (1/y2)
48
Tolerance Design and Tolerancing
49
50
51
Example (Quality Loss Function)
52
Example (Quality Loss Function)
53
8-Steps in Taguchi Methodology
1. Choose control factors and their levels
2. Identify uncontrollable (noise) factors and decide on how they will be simulated.
3. Select the response variable(s) and determine the performance measures (mean, standard
deviation, SNR, etc.)
4. Setup the experimental layout (choose appropriate design array (s))
5. Conduct the experiments and collect data
6. Analyze the data (effects, ANOVA, regression)
7. Choose optimal control factor levels and predict the performance measure at these levels
8. Confirm the optimal levels by experimentation.
54
Improving the Grinding Ratio.

• Imagine we are setting up an experiment in a grinding machine in our
machine shop. We want to find out what settings in the grinding machine can
give us maximum material removal efficiency, for a grinding wheel. We want
to use Grinding ratio (GR) as the factor to be optimized.
• GR is given by the expression material removed divided by wheel wear in
grams.
55

• Step 1: Problem definition: We want to maximize the GR by selecting
suitable machine parameters for the given wheel formulation.
• Step 2: Identify the Noise factor-i.e. Factors not in our control but are liable
to affect the measured output (GR). Let us assume that all wheels of the
same formulation may not work at the same efficiency level due to slight
variations in processing. Let us also assume that different persons will
experiment differently. This will also affect the results, but not in our control.
56

• Step 3: Plan the evaluation method to take care of the noise factor:
 To overcome the first noise factor, namely wheel inconsistency, we can arrange to test two
wheels of the same category.
 To overcome the second noise factor, namely operator difference, we can plan to have the
wheels evaluated by two different operators.
 The correct identification of noise factors and correct planning to capture these effects is the
key to the success of this method. Thus in our case, for each selected machine settings, we
have to get GR for 4 wheels of the same formulation (2 wheels by 1 operator and another 2
by the second operator).
57

• Step 4: Identify the objective function:
 Taguchi suggests different formula for different needs like Maximizing, minimizing,
normalizing etc. Our aim is to maximize GR.
 So the formula for objective function ȵ is given as
• ȵ = -10log10 (1/)
• Where y is the measured output function to be optimized. So, in our case,
y=Grinding ratio.
58

• Step 5: Identify control factors and their levels:
 A control factor is one which has an effect on the output (Our case G.R) and which is
controllable by us. A level indicates a setting.
Item Control Factor Level 1 Level 2

A Spindle speed 33 mps 45 mps
B Feed rate 2.5mm/rev 5.0 mm / rev
C Depth of cut 0.5mm 1.0mm
59

• Step 6: Design of experiment
 Taguchi method suggests the number of experiments to be used based on the number of
factors. There are separate tables for two level and three level factors. The table giving the
number of experiments is called Orthogonal array table.
For 2 Level factors For 3 Level factors
No of control fac- Array to be Selected No of control Array to be Selected

tors factors
2-3 L4 2-4 L9
4-7 L8 5-7 L18
8-11 L12
From the above table, in our case, since there are three
12-15 L16 control factors of 2 level each, we choose orthogonal
array L4, from the first table
60

• L4 looks as below. (Taguchi method lists such table for all Orthogonal
arrays).
Expt. Spindle speed mps Feed rate mm/rev Depth of cut mm
L4 orthogonal Array No
Control Factor levels

Expt.No A B C 1 33 2.5 0.5
1 1 1 1 2 33 5.0 1.0
2 1 2 2
3 45 2.5 1.0
3 2 1 2
4 2 2 1 4 45 5.0 0.5
The number of experiments at each setting is decided by us from steps 2 &3 to capture the
noise factors. Please note that this is different from the number of machine settings as given in
orthogonal array.
61

• In our case, since we have assumed 2 operators evaluating two wheels each
, we will be doing experiments in 4 different settings with 4 repetitions for
each setting.
62

• Step 7: Conduct the experiments and tabulate the ‘y’ values
 As decided in step 6, we conduct the experiments and measured the 'y' , which in our case
is grinding ratio defined as.
 G.R =/Weight of wheel loss
 Values in brackets denote the level whether 1 or 2.
Expt. No Speed Feed Depth of Measured values of 'y' G.R
rate cut
1 2 3 4
1 33(1) 2.5(1) 0.5(1) 21.5 20.2 20.5 19.0
2 33(1) 5.0 (2) 1.0 (2) 18.5 17.5 18.0 18.0
3 45(2) 2.5 (1) 1.0 (2) 19.0 19.5 18.8 18.8
16.6
4 45(2) 5.0 (2) 0.5 (1) 16.0 16.2 16.6
63

• Step 8: Calculate the objective function
• From step 4, for our type of problem of larger the better category
η = - 10 log10 {1/4 (1/+ 1/+ 1/+ 1/) } = 26.13
64

• Step 8: Calculate the objective function
Expt. No Speed Feed Depth of Measured values of 'y' G.R
rate cut
η
1 2 3 4
1 33(1) 2.5(1) 0.5(1) 21.5 20.2 20.5 19.0 26.13
2 33(1) 5.0 (2) 1.0 (2) 18.5 17.5 18.0 18.0 25.10
3 45(2) 2.5 (1) 1.0 (2) 19.0 19.5 18.8 18.8 25.58
16.6 24.27
4 45(2) 5.0 (2) 0.5 (1) 16.0 16.2 16.6
Mean 25.27
65

• Step 9: Determine the factor effects
This method helps in segregating the individual effect of each control factor at each level and is
denoted as mA1 mB1 etc.
mA1 means the effect of factor A at level 1
mC2 means the effect of factor C at level 2
Let us see how this can be calculated From the table under step 7.
mA1 should be the effect of all those η values where factor A (speed) is used in level 1 (33mps).
This occurs in expt.1 and expt. 2. So, using the values of these experiments,
mA1 =1/2 (η1 + η2) = 0.5 (26.13 +25.10) = 25.62
66
L4 orthogonal Array
Improving the Grinding Ratio. Expt.No A

B C
1 1 1 1
• Step 9: Determine the factor effects 2 1 2 2
3 2 1 2
Similarly, we can calculate the effect of other factors also as below: 4 2 2 1
mA1 =1/2 (η1 + η2) = 0.5 (26.13 +25.10) = 25.62

mA2 =1/2 (η3 + η4) = 0.5 (25.58 +24.27) = 24.92 Factor Level
1 2
mB1 =1/2 (η1 + η3) = 0.5 (26.13 +25.58) = 25.86
A (speed) 25.62 24.92
mB2 =1/2 (η2 + η4) = 0.5 (25.10 +24.27) = 24.68 B (Feed rate) 25.86 24.68
C (depth of cut) 25.2 25.34

mC1 =1/2 (η1 + η4) = 0.5 (26.13 +24.27) = 25.2
mC2 =1/2 (η2 + η3) = 0.5 (25.10 +25.58) = 25.34
These are tabulated and plotted with control factors at three individual points along x axis.
67
25.8
• Improving the Grinding Ratio. 25.6
Grinding Ratio
25.4
25.2
From the graph, level 1 of A, level 1 25
24.8
of B and level 2 of C give the
24.6
maximum effect in improving G.R. 24.4
Level 1 Level 2
Levels
A (speed)
26 25.4
25.6 25.35
Grinding Ratio
Grinding Ratio
25.3
25.2
25.25
24.8
25.2
24.4 25.15
24 25.1
Level 1 Level 2 Level 1 Level 2
Levels Levels
B (feed rate) C (dept of cut)

68
26
25.8
25.6
25.4
Grinding Ratio
25.2
25
24.8
24.6
24.4
24.2
24
Level 1 Level 2
Levels
A (speed) B (feed rate) C (dept of cut)
69
• Step 9: Determine the factor effects
• Naturally, Al, B1, C2 is the best combination. i.e. a speed of 33mps , a feed rate of 2.5m/min and a
depth of cut of 1mm will give the maximum G.R.
• It is of interest to see that we have not tried this combination. The combinations tried by us are given
below:
Expt 1 was Al, B1, C1
Expt 2 was A1, B2, C2
Other inferences from the graph Factor B has the largest effect on GR; Factor C has the least effect
on GR. The lowest GR is from A2, B2, C1.
70
• Step 10: Predicting GR for Untried combinations
This is the most exciting part of Taguchi method. Any combination of control factor settings,
even for experiments not conducted, can be estimated, using the formula.
η = µ + (mA - µ) + (mB - µ) + (mC - µ) Factor Level
1 2
where µ = mean A (speed) 25.62 24.92
for example, for the combination of A2 B2 C1. B (Feed rate) 25.86 24.68
η = µ + (m- µ) + (m- µ) + (m- µ) C (depth of cut) 25.2 25.34
= 25.27 + (24.92 - 25.27) + (24.68 – 25.27) + (25.20 - 25.27)

= 24.26
Which is very close to 24.27 in the table under step 7 for expt 4.
71
Already did on the following combination
Step 10: Predicting GR for Untried combinations
So, for the expected combination of A1 B1 C2 ,
η expectable = µ + (mA1 - µ) + (mB1 - µ) + (mC2 - µ)
= 25.27 + (25.62 - 25.27) + (25.86 – 25.27) + (25.34 - 25.27)
= 26.28
From our definition, η = -10 log10 ( 1 / ) = 26.28
log10 ( 1 / ) = 26.28/(-10) Factor Level
log10 ( 1 / ) = -2.628 1 2
(1 / ) = antilog ( -2.628 ) A (speed) 25.62 24.92
(1 / ) = 10^(-2.628) B (Feed rate) 25.86 24.68
C (depth of cut) 25.2 25.34

(1 / ) = 0.002355
() = 1/(0.002355) = 424.6196
So, y = SQRT (424.6196) = 20.61 GR max.
72
• Instead of 8 combinations, we tried only four combinations of setting
and from these four we were able to predict the GR achievable for those
combinations of controls which we have not tried.
• For 3 — two level factors, the combinations possible are only = 8. Here
we tried 4 experiments.
• But suppose there are seven factors like speed, infeed, dressing
traverse etc. each with three setting levels, then the possibility is =
2187 combinations.
• The effect of all these 2187 experiments can be predicted with just 18
experiments. (for 3 level factors of 7 numbers, we use L18 array).
73
Taguchi Analysis (Orthogonal Array)
For 2 Level factors For 3 Level factors
No of control fac- Array to be Selected No of control Array to be Selected

tors factors
2-3 L4 2-4 L9
4-7 L8 5-7 L18
8-11 L12
12-15 L16
L4 orthogonal Array
Expt.No A B C
1 1 1 1
2 1 2 2
3 2 1 2
4 2 2 1
74
L8 (2^7) ORTHOGONAL ARRAY
Expt. Columns
No
1 2 3 4 5 6 7
1 1 1 1 1 1 1 1
2 1 1 1 2 2 2 2
3 1 2 2 1 1 2 2
4 1 2 2 2 2 1 1
5 2 1 2 1 2 1 2 L9 (3^4) ORTHOGONAL ARRAY
6 2 1 2 2 1 2 1
7 2 2 1 1 2 2 1
8 2 2 1 2 1 1 2 Expt.No Columns
1 2 3 4
1 1 1 1 1
2 1 2 2 2
3 1 3 3 3
4 2 1 2 3
5 2 2 3 1
6 2 3 1 2
7 3 1 3 2
8 3 2 1 3
9 3 3 2 1
75
76
Taguchi Analysis (Example Problem)
• The objective of the experiment is to obtain minimum surface roughness of
the parts machined by wire cut EDM machine using Taguchi’s DOE
technique of process parameters.
• Surface roughness(Ra) values were obtained using Taylor-Hobson Surtronic

instruments and the sequence are 2.5275, 2.3520, 2.2540, 2.4650, 2.7000,
2.8125, 2.350, 2.2875.
• Find out the Main effect and draw the factor effect diagram and identify the
optimum WEDM process parameters.
77
• Since, the number of factors given are 7 with 2 levels. Hence, we can
choose L8 orthogonal array.
Expt. No Columns
1 2 3 4 5 6 7
1 1 1 1 1 1 1 1
2 1 1 1 2 2 2 2
3 1 2 2 1 1 2 2
4 1 2 2 2 2 1 1
5 2 1 2 1 2 1 2
6 2 1 2 2 1 2 1
7 2 2 1 1 2 2 1
8 2 2 1 2 1 1 2
78
• L8 orthogonal array.
• Smaller-the-Better: η = -10 log10( )

• Hence solving for η
Expt. No Columns
Surface
Pull on Timet Pulse off time Peak Wire Tension Fluid Pres- Roughness η
Wire Feed (E) Fluid Rate (F)
(A) (B) Current(C) (D) sure (G) (y)
1 125 (1) 35 (1) 11 (1) 1000 (1) 7 (1) 8 (1) 13 (1) 2.5275 η1
2 125 (1) 35 (1) 11 (1) 1200 (2) 8 (2) 9 (2) 15 (2) 2.3520 η2
3 125 (1) 40 (2) 12 (2) 1000 (1) 7 (1) 9 (2) 15 (2) 2.2540 η3
4 125 (1) 40 (2) 12 (2) 1200 (2) 8 (2) 8 (1) 13 (1) 2.4650 η4
5 130 (2) 35 (1) 12 (2) 1000 (1) 8 (2) 8 (1) 15 (2) 2.7000 η5
6 130 (2) 35 (1) 12 (2) 1200 (2) 7 (1) 9 (2) 13 (1) 2.8125 η6
7 130 (2) 40 (2) 11 (1) 1000 (1) 8 (2) 9 (2) 13 (1) 2.350 η7
8 130 (2) 40 (2) 11 (1) 1200 (2) 7 (1) 8 (1) 15 (2) 2.2875 η8
µ
79
• L8 orthogonal array.
• Smaller-the-Better: η = -10 log10( )

• Hence solving for η
Expt. No Columns
Surface
Peak current Pulse on time Peak Wire Tension Fluid Pres- Roughness η
Wire Feed (E) Fluid Rate (F)
(A) (B) Current(C) (D) sure (G) (y)
-
1 125 (1) 35 (1) 11 (1) 1000 (1) 7 (1) 8 (1) 13 (1) 2.5275 8.053823285 η1
-
2 125 (1) 35 (1) 11 (1) 1200 (2) 8 (2) 9 (2) 15 (2) 2.3520 7.428746348 η2
-
3 125 (1) 40 (2) 12 (2) 1000 (1) 7 (1) 9 (2) 15 (2) 2.2540 7.059078234 η3
-
4 125 (1) 40 (2) 12 (2) 1200 (2) 8 (2) 8 (1) 13 (1) 2.4650 7.836338472 η4
-
5 130 (2) 35 (1) 12 (2) 1000 (1) 8 (2) 8 (1) 15 (2) 2.7000 8.627275283 η5
-
6 130 (2) 35 (1) 12 (2) 1200 (2) 7 (1) 9 (2) 13 (1) 2.8125 8.981850622 η6
-
7 130 (2) 40 (2) 11 (1) 1000 (1) 8 (2) 9 (2) 13 (1) 2.350 7.421357245 η7
-
8 130 (2) 40 (2) 11 (1) 1200 (2) 7 (1) 8 (1) 15 (2) 2.2875 7.187222055 η8
µ -7.82446144
80
• Determine the factor effects

mA1 =1/4 (η1 + η2 + η3 + η4) = 0.25 * (-8.053-7.428-7.059-7.836) = -7.5945
…
…
…
…
…
…
…
…
…
….
81

mA1 =1/4 (η1 + η2 + η3 + η4) = 0.25 * (-8.053-7.428-7.059-7.836) = -7.5945
mA2 =1/4 (η5 + η6 + η7 + η8) = 0.25 * (-8.627-8.981-7.42-7.187) = -8.053
mB1 =1/4 (η1 + η2 + η5 + η6) =
mB2 =1/4 (η3 + η4 + η7 + η8) =
mC1 =
mC2 =
mD1 =
mD2 =
mE1 =
mE2 =
mF1 =
mF2 =
mG1 =
mG2 =
82

mA1 =1/4 (η1 + η2 + η3 + η4) = 0.25 * (-8.053-7.428-7.059-7.836) = -7.5945
mA2 =1/4 (η5 + η6 + η7 + η8) = 0.25 * (-8.627-8.981-7.42-7.187) = -8.053
mB1 =1/4 (η1 + η2 + η5 + η6) = -8.272
mB2 =1/4 (η3 + η4 + η7 + η8) = -7.37
mC1 =1/4 (η1 + η2 + η7 + η8) = -7.522
mC2 =1/4 (η3 + η4 + η5 + η6) = -8.125
mD1 =1/4 (η1 + η3 + η5 + η7) = -7.789
mD2 =1/4 (η2 + η4 + η6 + η8) = -7.858
mE1 =1/4 (η1 + η3 + η6 + η8) = -7.82
mE2 =1/4 (η2 + η4 + η5 + η7) = -7.827
mF1 =1/4 (η1 + η4 + η5 + η8) = -7.925
mF2 =1/4 (η2 + η3 + η6 + η7) = -7.722
mG1 =1/4 (η1 + η4 + η6 + η7) = -8.072
mG2 =1/4 (η2 + η3 + η5 + η8) = -7.575
83
• Determine the factor effects (graph)
84
85
86
87
• A1, B2, C1, D1, E2, F2, G2 giving the minimum effect in reducing the
surface roughness.
η = µ + (mA1 - µ) + (mB2 - µ) + (mC1 - µ) + (mD1 - µ) + (mE2 - µ) + (mF2 - µ) + (mG2 - µ)
= -7.823 + (-7.594+7.823) + (-7.37+7.823) + (-7.597+7.823) + (-7.789+7.823) + (-7.827+7.823) + (-

7.722+7.823) + (-7.573+7.823)
= -6.533
η = -6.533
i.e., η = -10 log10( ) = -6.533
log10( ) = 0.6533
= antilog (0.6533)
= 4.5009
= 2.12
Hence, Surface Roughness = 2.12 with the above combination.
88
Problem: The objective of the research is an experimental investigation to find
optimization of machining parameters of EDM machine for machining steel
material EN-8 using Taguchi’s DOE.
• Ra values obtained according to experiment sequence are 2.625,

1.905,2.5275,2.99,4.4325,4.10,3.38,3.015, 3.312.
• Find out the Main effect and draw the factor effect diagram and identify the
optimum process parameters.
89
Solution:
How many factors ?
How many levels ?
Which Orthogonal Array?
90
Solution: Hence, 4 factors and 3 levels, we have to choose L9 orthogonal array.
L9 (3^4) ORTHOGONAL ARRAY
Expt.No Columns Expt.No Columns

1 2 3 4 1 2 3 4
1 1 1 1 1 1 3 25 45 0.5
2 1 2 2 2 2 3 52 50 .75
3 1 3 3 3 3 3 100 55 1
4 2 1 2 3 4 5 25 50 1
5 2 2 3 1 5 5 52 55 0.5
6 2 3 1 2 6 5 100 45 0.75
7 3 1 3 2 7 7 25 55 0.75
8 3 2 1 3 8 7 52 45 1
9 3 3 2 1 9 7 100 50 0.5
91
Solution: solving for signal-to-noise ratio (Efficiency)
Expt.No Columns
1 2 3 4 Ra S.N Ratio Smaller-the-Better: η = -10 log10( )
Hence solving for η
1 3(1) 25(1) 45(1) 0.5(1) 2.625 -8.3826
2 3(1) 52(2) 50(2) 0.75(2) 1.905 -5.5979
3 3(1) 100(3) 55(3) 1(3) 2.5275 -8.0538
4 5(2) 25(1) 50(2) 1(3) 2.99 -9.5134
5 5(2) 52(2) 55(3) 0.5(1) 4.4325 -12.9330
6 5(2) 100(3) 45(1) 0.75(2) 4.10 -12.2557
7 7(3) 25(1) 55(3) 0.75(2) 3.38 -10.5783
8 7(3) 52(2) 45(1) 1(3) 3.015 -9.5857
9 7(3) 100(3) 50(2) 0.5(1) 3.312 -10.4018
92
Solution: Determination of factor effects
Expt.No Columns
1 2 3 4 Ra S.N Ratio mA1 =1/3 (η1 + η2 + η3) = -7.345
1 3(1) 25(1) 45(1) 0.5(1) 2.625 -8.3826 mA2 =1/3 (η4 + η5 + η6) = -11.567
mA3 =1/3 (η7 + η8 + η9) = -10.189
2 3(1) 52(2) 50(2) 0.75(2) 1.905 -5.5979
mB1 =1/3 (η1 + η4 + η7) = -9.491
3 3(1) 100(3) 55(3) 1(3) 2.5275 -8.0538 mB2 =1/3 (η2 + η5 + η8) = -9.372
4 5(2) 25(1) 50(2) 1(3) 2.99 -9.5134 mB3 =1/3 (η3 + η6 + η9) = -10.237
5 5(2) 52(2) 55(3) 0.5(1) 4.4325 -12.9330 mC1 =1/3 (η1 + η6 + η8) = -10.075
6 5(2) 100(3) 45(1) 0.75(2) 4.10 -12.2557
mC2 =1/3 (η2 + η4 + η9) = -8.504
mC3 =1/3 (η3 + η5 + η7) = -10.522
7 7(3) 25(1) 55(3) 0.75(2) 3.38 -10.5783 mD1 =1/3 (η1 + η5 + η9) = -10.572
8 7(3) 52(2) 45(1) 1(3) 3.015 -9.5857 mD2 =1/3 (η2 + η6 + η7) = -9.477
9 7(3) 100(3) 50(2) 0.5(1) 3.312 -10.4018 mD3 =1/3 (η3 + η4 + η8) = -9.051
93
Solution: Determination of factor
effects A B C D
1 -7.345 -9.491 -10.075 -10.572
2 -11.567 -9.372 -8.504 -9.477
3 -10.489 -10.237 -10.522 -9.054
Factor effect graphs.

From the graph, A1 B2 C2 D3 giving the
minimum surface roughness. So, we can
finanlize
94
Problem: Heat treatment process used to harden steel components, using
Taguchi analysis. Determine which process parameters have the greatest
impact on the hardness of the steel components. Hardness values are 57, 59,
65, 45, 67, 73, 51, 48. Hardness value (HV)
Level 2 Level 1 Parameters Parameter number

900 760 Temperature(OC) 1
140 35 Quenching rate(OC/s) 2
300 1 Cooling time(sec) 3
6 1 Carbon contents (Wt% c) 4
20 5 Co 2concentration (%) 5
95
Problem: Heat treatment process
96
Problem: Determine which process parameters have the greatest impact on
the hardness of the steel components. Hardness values are 57, 59, 65, 45, 67,
73, 51, 48. Hardness value (HV)
Level 2 Level 1 Parameters Parame-
ter num-
ber
140 35 Quenching 2
rate(OC/s)
6 1 Carbon contents 4
(Wt% c)
20 5 Co 2concentration 5
(%)
97
Solution: How many factors? How any levels?
Factors: 5
Level: 2
Orthogonal array: L8
Level 2 Level 1 Parameters Parame-
ter num-
ber
Expt.
Columns 900 760 Temperature(OC) 1
No
140 35 Quenching 2
1 2 3 4 5 6 7
rate(OC/s)
1 1 1 1 1 1 1 1
2 1 1 1 2 2 2 2
3 1 2 2 1 1 2 2 6 1 Carbon contents 4
4 1 2 2 2 2 1 1 (Wt% c)
5 2 1 2 1 2 1 2
6 2 1 2 2 1 2 1
7 2 2 1 1 2 2 1 (%)
8 2 2 1 2 1 1 2
98
Solution: L8 orthogonal array Level 2 Level 1 Parameters Parame-
ter num-
ber
Expt. Columns
No 140 35 Quenching 2
1 2 3 4 5
rate(OC/s)
1 760 35 1 1 5
2 760 35 1 6 20
3 760 140 300 1 5 6 1 Carbon contents 4
4 760 140 300 6 20 (Wt% c)
5 900 35 300 1 20
6 900 35 300 6 5
7 900 140 1 1 20 (%)
8 900 140 1 6 5
99
Solution: L8 orthogonal array.
Hardness value (HV) of steel: larger is better
Hence, S/N ratio: -10 * Log10(sum(1/y^2)/n)
S/N ratio = -10 * Log10(sum(1/y^2)/n) = 35.1175 = 35.12

.. Expt. Columns
.. No
1 2 3 4 5 HV S/N ratio
.. 1 760 35 1 1 5 57 35.12
2 760 35 1 6 20 59 35.42
.. 3 760 140 300 1 5 65 36.26
4 760 140 300 6 20 45 33.06
S/N ratio = -10 * Log10(sum(1/y^2)/n) = 33.62 5 900 35 300 1 20 67 36.52
6 900 35 300 6 5 73 37.27
7 900 140 1 1 20 51 34.15
8 900 140 1 6 5 48 33.62
100
Solution: main effects of factors
mA1 =1/4 (η1 + η2 + η3 + η4) = 0.25 * (35.12+35.42+36.26+33.06) = 34.96
mA2 =1/4 (η5 + η6 + η7 + η8) = 35.39
Expt. Columns
mB1 =1/4 (η1 + η2 + η5 + η6) = 36.08
No
mB2 =1/4 (η3 + η4 + η7 + η8) = 34.27 1 2 3 4 5 HV S/N ratio η
mC1 =1/4 (η1 + η2 + η7 + η8) = 34.58 1

760
35(1) 1 (1) 1 (1) 5 (1) 57 35.12 1
(1)
mC2 =1/4 (η3 + η4 + η5 + η6) = 35.78
760
mD1 =1/4 (η1 + η3 + η5 + η7) = 35.51 2 35(1) 1 (1) 6 (2) 20 (2) 59 35.42 2
(1)
mD2 =1/4 (η2 + η4 + η6 + η8) = 34.84 3
760
140(2) 300 (2) 1 (1) 5 (1) 65 36.26 3
(1)
mE1 =1/4 (η1 + η3 + η6 + η8) = 35.57 760
4 140(2) 300 (2) 6 (2) 20 (2) 45 33.06 4
mE2 =1/4 (η2 + η4 + η5 + η7) = 34.79 (1)
900
5 35(1) 300 (2) 1 (1) 20 (2) 67 36.52 5
(2)
900
6 35(1) 300 (2) 6 (2) 5 (1) 73 37.27 6
(2)
900
7 140(2) 1 (1) 1 (1) 20 (2) 51 34.15 7
(2)
900
8 140(2) 1 (1) 6 (2) 5 (1) 48 33.62 8
(2)
101
Solution: main effects of factors
mA1 =1/4 (η1 + η2 + η3 + η4) = 0.25 * (35.12+35.42+36.26+33.06) = 34.96
mA2 =1/4 (η5 + η6 + η7 + η8) = 35.39
mB1 =1/4 (η1 + η2 + η5 + η6) = 36.08 Main Effects Plot for SN ratios
mB2 =1/4 (η3 + η4 + η7 + η8) = 34.27 Data Means
A B C D E
mC1 =1/4 (η1 + η2 + η7 + η8) = 34.58
36.0
mC2 =1/4 (η3 + η4 + η5 + η6) = 35.78
mD1 =1/4 (η1 + η3 + η5 + η7) = 35.51
Mean of SN ratios
mD2 =1/4 (η2 + η4 + η6 + η8) = 34.84 35.5
mE1 =1/4 (η1 + η3 + η6 + η8) = 35.57

mE2 =1/4 (η2 + η4 + η5 + η7) = 34.79 35.0
34.5
760 900 35 140 1 300 1 6 5 20
Signal-to-noise: Larger is better
102
Solution: A2, B1, C2, D1, E1 has largest value in S/N ratio. Hence, we can
consider this as best combination of factors and its level to get higher hardness
value (HV).
Main Effects Plot for SN ratios
Data Means
A B C D E
36.0
Mean of SN ratios
35.5
35.0
34.5
760 900 35 140 1 300 1 6 5 20
Signal-to-noise: Larger is better
103
Problem: Find out the optimum value for MS plate & Value of these shown in
table.
104
Problem:
Surface roughness:
Kerf:
105
Solution:
How many factors? How many level? Which Orthogonal Array? Which S/N
Ratio?
106
Solution:
How many factors? 3 factors
How many level? 3 levels
Which Orthogonal Array? L9 orthogonal array
Which S/N Ratio? Both surface roughness and Kerf needs to be smaller.
Hence, Smaller is better to be used to calculate S/N ratio.
107
Solution:
How many factors? 3 factors
How many level? 3 levels
Which Orthogonal Array? L9 orthogonal array
Which S/N Ratio? Surface roughness needs to be smaller.
Hence, Smaller is better to be used to calculate S/N ratio.
Smaller-the-Better: η = -10 log10( )
Hence solving for η
108
Smaller is better
Solution: Level A B C
1 -45.96 -45.97 -45.98
Smaller-the-Better: η = -10 log10( ) 2 -45.98 -45.99 -45.94
Hence solving for η 3 -45.96 -45.95 -45.98
Delta 0.02 0.04 0.04
Rank 3 1 2
A B C Surface
Main Effects Plot for SN ratios
A B C Roughness S/N Ratio
1 1 1 Data Means
500 1 1.5 198.01 -45.9337
1 2 2 -45.94
A B C
500 2 1.8 198.9 -45.9727

1 3 3
500 3 2 199.05 -45.9792
2 1 2
-45.95
Mean of SN ratios
600 1 1.8 199.01 -45.9775
2 2 3 -45.96
600 2 2 199.14 -45.9832
2 3 1
600 3 1.5 199.237 -45.9874 -45.97
3 1 3
700 1 2 199.31 -45.9906
3 2 1 -45.98
700 2 1.5 199.842 -46.0137
3 3 2
700 3 1.8 196.79 -45.8801 -45.99
500 600 700 1 2 3 1.5 1.8 2.0
Signal-to-noise: Smaller is better
η = -10 log10( )
109
Application of Optimization Techniques
• Design of aircraft structures at minimum weight.
• Finding the optimum trajectory of space vehicle.
• Design of civil engineering structures like frames, foundation, bridges, towers, chimney at
minimum weight.
• Selection of machining condition in metal cutting process for minimum production cost.
• Optimum design of electrical networks.
• Allocation of resources / services to several activities to maximize the benefits.
• Controlling the idle time, waiting time and queuing in production line to reduce the cost.
• Design of pumps, turbines and heat transfer equipments for maximize the efficiency.
• Shortest route to be taken by sales person to visit different cities.
• Design of material handling equipments for minimize the cost.
110
Thank you
https://support.minitab.com/en-us/minitab/18/
help-and-how-to/modeling-statistics/doe/sup-
porting-topics/taguchi-designs/taguchi-
designs/
111

18MEO113T - DOE - Unit 3 - AY2023 - 24 ODD

Uploaded by

Copyright:

Available Formats

18MEO113T - DOE - Unit 3 - AY2023 - 24 ODD

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

18MEO113T - DOE - Unit 3 - AY2023 - 24 ODD

Uploaded by

Copyright:

Available Formats

Subject Code

18MEO113T - Design of Experiments

• To learn the fundamentals of design of experiment techniques.

packages like Minitab, etc.

1: Introduction to Robust design, Loss functions.

• Robust Design method, also called the Taguchi Method,

• Genichi Taguchi Quotes:

“A scientific or technical study always consists of the following three steps:

• In Taguchi designs, a measure of robustness used to identify control

factors that reduce variability in a product or process by minimizing the

effects of uncontrollable factors (noise factors).

• In a Taguchi designed experiment, you manipulate noise factors to force

settings that make the process or product robust, or resistant to variation

from the noise factors.

TQM expresses about continuous

QC is used to verify the quality of the

1. The Taguchi definition of quality

2. The concept of Robust Design

Robust Products Products that are not robust

1. To have longer life (higher reliability)

2. To be more consistent with use

3. To be more consistent from product to product

4. To perform consistently as temperature and other conditions change

– Internal disturbances: wear and tear inside a specific unit

– Disturbances in the production process: deviation from target values

• A three step method for achieving robust design (Taguchi)

1. Concept Design / System Design

– Control factors are those process variables in a process that management

– The optimal parameter levels can be determined through experimentation.

resulting improvement has been insufficient.

– Often results in increased production costs

• More expensive input material might have to be used to meet specifications.

3. Taguchi Quality/Quadratic Loss Function (QLF)

<Traditional approach> <Taguchi approach>

<Taguchi approach> <Traditional approach> <Taguchi approach>

• According to the traditional concept of loss function that as long as the

product’s quality characteristics is within the specification limits or

step function with a constant value A outside of the specification limits or

time the product is designed to the time product is shipped to the

• Loss function L(y) makes no distinction between products

– Whose quality characteristic is just above lower specification limit (LSL), or

Taguchi’s definition clearly puts more emphasis on

beyond the specification limits USL and LSL.

• Since the Taguchi loss function increases quadratically with increasing

to three different situations or cases, namely

Ranjit K . Roy, A Primer on the Taguchi Method, Second

Target is best or The Nominal-the-Best Case

Target is best or The Nominal-the-Best Case

• Taguchi experiments often use a 2-step optimization process.

small or no effect on the signal-to-noise ratio.

L(y) = k (y – m)2 L(y) = ky2 L(y) = k (1/y2)

Larger – the – better loss function

deviation, SNR, etc.)

4. Setup the experimental layout (choose appropriate design array (s))

5. Conduct the experiments and collect data

6. Analyze the data (effects, ANOVA, regression)

8. Confirm the optimal levels by experimentation.

Improving the Grinding Ratio.