18MEO113T

Design of Experiments

SRM Institute of Science and Technology ,

Kattankulathur

1
 Purpose /aim of this
course
• 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, Qualitek , etc.

• After this course , You will be ready to apply the

technique confidently to all of your projects .

2
Course Plan

3
Reference books

4
ACADEMIC YEAR : 2023 - 24 EVEN
18MEO113T DESIGN OF EXPERIMENTS

UNIT-I: Introduction to DoE

Introduction in Design of Experiments (DOE)

– The Fundamental and Potential Practical
Problems in Experimentation
Lecture: Day-1

Instructor: Dr. Sumit Pramanik

Research Associate Professor
Department of Mechanical Engineering
FACULTY OF ENGINEERING AND TECHNOLOGY
SRM Institute of Science and Technology, Kattankulathur

Venue: TP 601 (Block: Tech Park) BTech Student: Batch- 1/G

18/01/2024
 Design of Experiments(DOE)
DOE : Sir R.A Fisher in England in the early 1920s-
Powerful technique used for exploring new processes;
gaining increased knowledge of the existing processes ;
optimizing these processes for achieving world class
performance.
―The great New York Yankees catcher Yogi Berra said that ― . . . you
can observe a lot just by watching.‖ However, to understand what
happens to a process when you change certain input factors, you have
to do more than just watch—you actually have to change the factors‖

DOE : DOX or Experimental design

6
 Brief history of DOE
• The agricultural origins, 1918-1940s
– R. A. Fisher & his co-workers
– Profound impact on agricultural science
– Factorial designs, ANOVA
• The First industrial era, 1951- late 1970s
– Box & Wilson, response surfaces
– Applications in the chemical & process industries
• The Second industrial era, late 1970s – 1990s
– Quality improvement initiatives in many
companies
– Taguchi and robust parameter design, process
robustness
• The modern era, beginning circa 1990
7
 What is experiment?
• Experiment is an important activity performed by
all in day to day activity.

• An experiment is a test or a series of tests

conducted to analyse, understand or improve
the system.

• The only way to learn anything about a system is

to disturb it and then observe it.
 Example of an experiment

Aluminium Pot Ceramic Pot

Constant
Heat, Same
amount of
water

Outcome : Time taken to Boil

Other Variables : source of heat, without lid, Material, shape of

pot etc…
 Example…

Spike shoe

Same person,
day,
Sports shoe playground

Outcome : Time taken to reach Target

Other Factors : Weather, Cloth, location
 Basic Terminologies
Outcomes = Response
Variables / Attributes = Factors

Types of factors:
Numeric: Quantitative
Categorical : Qualitative

Objective:
Combine outcome by adjusting choosing the variable
Some eg.
Maximize the profit , yield
Minimise the expense, loss, energy
 Experiment …..
• Efficiency is doing the least amount of work
and getting the most amount of information.

• Objective is to improve outcome and there

are two or more factors we can change to
make that improvement or influence the
outcome.

• So there's an outcome/response and there

are factors that we want to improve.
 Identify
Aluminium Pot Ceramic Pot
Constant
Heat, Same
amount of
water

Control Variable : Aluminium Pot

Experiment Variable : Ceramic Pot
Independent Variable : Type of material used for pot
Dependant Variable : Time taken to Boil
Outcome
Objective : Minimise the time taken to Boil

Other Factors : source of heat, without lid, Material,

shape of pot etc…
 Process or System Model
⮚ A process can be any manufacturing process or service Process
which is used for experiment
⮚ An experiment involves a test or series of test in which purposeful
changes are made to the input variables of a process or system so
that changes in the output responses can be observed.
 Objective of experiment

1. Determining which Control variables (Input)

X, are most influential on the response
(output), y
2. Determining where to set the influential ‘X’ so
that ‘y’ is near the nominal requirement.
3. Determining where to set the influential ‘x’ so
that variability in ‘y’ is small.
4. Determining where to set the influential ‘x’ so
that the effect of uncontrollable variables ‘z’
are minimized.
ACADEMIC YEAR : 2023 - 24 EVEN
18MEO113T DESIGN OF EXPERIMENTS

UNIT-I: Introduction to DoE

Statistical Thinking and its Role within DOE

Lecture: Day-1

Instructor: Dr. Sumit Pramanik

Research Associate Professor
Department of Mechanical Engineering
FACULTY OF ENGINEERING AND TECHNOLOGY
SRM Institute of Science and Technology, Kattankulathur

Venue: TP 601 (Block: Tech Park) BTech Student: Batch- 1/G

18/01/2024
17
18
19
 Terms used in Design of Experiments
RESPONSE:
A measurable outcome of interest. e.g., time, etc.
FACTORS:
Controllable variables that are deliberately manipulated to
determine the individual and joint effects on the response(s),
OR Factors are those quantities that affect the outcome of an
experiment ,e.g.: heat source, lid, shape of pot etc.
LEVELS:
Levels refer to the values of factors i.e., the values for which
the experiments will be conducted ‖, e.g.:
Level–1 heat input=low
Level–2 heat input=high

20
 Terms used in Design of Experiments
TREATEMENT:
A set of specified factor levels for an experiment or run ,e.g.:
Treatment–1 : heat input=low and with lid
Treatment–2 : heat input = high and without lid
NOISE:
Variables that affect product/ process performance, whose
values cannot be controlled or are not controlled for economic
reasons. Eg. Climate, environment etc….
REPLICATION:
Replication is a systematic duplication of series of experimental
runs. It provides the means of measuring precision by
calculating the experimental error.

21
ACADEMIC YEAR : 2023 - 24 EVEN
18MEO113T DESIGN OF EXPERIMENTS

UNIT-I: Introduction to DoE

Basic principles of DOE and Degrees of Freedom -

Selection of Quality Characteristics for Experiments

Lecture: Day-2

Instructor: Dr. Sumit Pramanik

Research Associate Professor
Department of Mechanical Engineering
FACULTY OF ENGINEERING AND TECHNOLOGY
SRM Institute of Science and Technology, Kattankulathur

Venue: TP 601 (Block: Tech Park) BTech Student: Batch- 1/G

24/01/2024
 DOE Definition
• Design of experiments is a systematic method to
determine the relationship between factors affecting
a process and the output of that process

• In other words, it is used to find cause and effect

relationships.

23
Cause and Effect

24
 What is DOE all about?
• Helps us to study many factors simultaneously and most
economically
• By studying the effects of each factor on the results , the best
factor combination can be determined.
• When applied to product or process design, the technique
helps to seek out the best design among the many
alternatives
• Used to solve scientifically problems whose solutions lies in
the proper combination of ingredients rather than innovations
or a single identifiable cause.
25
 DOE : Necessity
• Experiments are performed today in many manufacturing
organizations

• For continuous improvement in product/process quality, it

is fundamental to understand the process behavior, the
amount of variability and its impact on processes

• Relationship between input (process variables or factors)

and output performance characteristics (quality
characteristics or response)

26
 Experiments ?
In an engineering environment, experiments are often
conducted to explore, estimate or confirm

• Exploration refers to understanding the data from the

process

• Estimation refers to determining the effects of process

variables or factors on the output performance
characteristic

• Confirmation implies verifying the predicted results

obtained from the experiment

27
 Experiments?
• Ex Metal cutting operation ; Input : cutting speed, feed rate, type of

coolant, depth of cut, etc. Output : surface finish of the finished part

• One of the common approaches employed by many engineers

today in manufacturing companies is One Variable At a

Time(OVAT), where we vary one variable at a time keeping all

other variables in the experiment fixed

• Depends upon guesswork, luck, experience and intuition for its

success. Unreliable, inefficient, time consuming and may yield false

optimum condition for the process

28
 Few Salient Points
• Statistical thinking and statistical methods play an

important role in planning, conducting, analyzing

and interpreting data from engineering experiments

• When several variables influence a certain

characteristic of a product, the best strategy is then
to design an experiment so that valid, reliable and
sound conclusions can be drawn effectively,
efficiently and economically
29
 Few Salient Points
• In DOE , engineers can make deliberate changes in input
variables and determines how output functional performance
varies
• Importantly , not all variables affect the performance in same
manner
• Some may have strong influence, some may have medium,
some have no influence at all
• Therefore, the objective of a carefully planned designed
experiment is to understand which set of variables in a
process affects the performance most and then determine the
best levels for these variables to obtain satisfactory output
functional performance in products 30
 Industrial Experiments
Industrial experiments involves a sequence of activities
1. Hypothesis an assumption that motivates the
experiment
2. Experiment a series of tests conducted to
investigate the hypothesis
3. Analysis involves understanding the nature of data
and performing statistical analysis of the data
collected from the experiment
31
 Industrial Experiments
4. Interpretation is about understanding the results of
the experimental analysis

5. Conclusion involves whether or not the originally

set hypothesis is true or false Very often more
experiments are to be performed to test the
hypothesis and sometimes we establish new
hypothesis which requires more experiments

32
 DOE : Benefits
• All kinds of industries(products and processes) , service
industry
• Produce maximum returns when applied to research,
concept design, and product development.
• Engineers and scientific personals should learn and
apply DOE
• DOE should be an essential skill for all manufacturing
and process specialists
• Engineers at a manufacturing plant should routinely use
DOE techniques to improve production processes.
33
 Applications of DOE
• Improved process yield , stability , profits , return on
investment , and process capability.

• Reduced process variability and hence better product

performance consistency; Reduced manufacturing costs;
Reduced process design and development time

• Increased understanding of the relationship between key

process inputs and output(s); Increased business profitability
by reducing scrap rate, defect rate, rework, retest, etc.

34
35
36
37
38
39
40
41
 DOE Terminology
• Factor :
– A variable or attribute which influences or is suspected of
influencing the characteristic being investigated.

– All input variables which affect the output of a system are

factors

– Factors are varied in the experiment

– They can be controlled at fixed levels.

– They can be varied or set at levels of our interest.

42
 DOE Terminology
• Factor :
– Two types of variables or factors : quantitative and
qualitative

– Quantitative – how the range of settings to be measured

and controlled e.g. temperature , pressure in the process
yield experiment

– Qualitative – discrete in nature ; requires more levels

e.g. types of raw materials, type of catalysts, type of supplier,
etc.

– These are also called independent variables.

43
 DOE Terminology
• Level :
– Specified value or setting of the factor being examined in

the experiment.

– The values of a factor/independent variable being

examined in an experiment.

– If the factor is an attribute, each of its state is a level. For

example, setting of a switch on or off are the two levels of

the factor switch setting.

44
 DOE Terminology
• Level :
– If the factor is a variable, the range is divided into required

number of levels. For example, the factor temperature

ranges from 1000 to 1200°C and it is to be studied at three

values say 1000°C, 1100°C and 1200°C, these three

values are the three levels of the factor temperature.

– The levels can be fixed or random.

45
 DOE Terminology
• Treatment:
– One set of levels of all factors employed in a given
experimental trial. For example, an experiment conducted
using temperature T1 and pressure P1 would constitute
one treatment.
– In the case of single factor experiment, each level of the
factor is a treatment.

• Run or trial :
– a trial or run is a certain combination of factor levels whose
effect on the output (or performance characteristic) is of
interest.
46
 DOE Terminology
• Experimental unit :
– Facility with which an experimental trial is conducted such as samples
of material, person, animal, plant, etc.

• Response :
– The result/output obtained from a trial of an experiment. This is also
called dependent variable. Examples are yield, tensile strength,
surface finish, number of defectives, etc.

• Effect :
– Effect of a factor is the change in response due to change in the level
of the factor.
47
48
 DOE Terminology
• Experimental error:
– It is the variation in response when the same experiment is
repeated, caused by conditions not controlled in the
experiment.

– It is estimated as the residual variation after the effects

have been removed.

49
50
 Example
Example 1: A flow solder machine is used in the manufacturing process for
printed circuit boards. The process engineer would like to conduct a designed
experiments to determine which controllable variables is significantly affecting
the output quality. And also, to find out the way to reduce the variability of
output due to change in uncontrollable variable.

Control variables
Output
Input (Reducing
(solder) solder defects)
Process

Un-Controlled
variables

51
 Example

Controllable factors Un-Controllable factors

1. Solder temperature 1. Thickness of the printed
2. Preheat temperature circuit board
3. Conveyor speed 2. Types of components
4. Flux Type used on the board
5. Flux specific gravity 3. Layout of the components
6. Solder wave depth on the board
7. Conveyor angle 4. Operator
5. Production rate

52
 Example
Example 2: In this example, we are interested to improve the yield of a
chemical process through using the two most important process variables:
reaction time (x1) and operating temperature (x2). As the given response
surface is unknown to the process personnel, experimental set up will be
required to investigate it.

x2
Temperature & Reaction Time
(80,180) (90,180)
Output (Improving
the yield of a
Chemical x1
chemical process)
Process
(80,170) (90,170)

53
 Example
Example 3: The response variables of interest in designing an aircraft engine
are fuel consumption and engine thrust. The engineer in-charge uses a
computer model of the system where he can vary the following designed
parameters (inlet flow, fan pressure ratio, overall pressure, stator outlet
temperature) to optimize the performance of the engine.

1 2 3 4

Fuel consumption
Computer
model
Process
Engine thrust

54
Experimental Strategies

55
 Best-Guess Approach
• Best-Guess Approach
– Eg: Play Golf / Machining a part / Preparing a concrete
/ Preparing a sample for new application / etc.

56
 Best-Guess Approach
• Best-Guess Approach

– Disadvantages.
– First, suppose the initial best-guess does not produce
the desired results. Now the experimenter has to take
another guess at the correct combination of factor levels.
This could continue for a long time, without any
guarantee of success.
– Second, suppose the initial best-guess produces an
acceptable result. Now the experimenter is tempted to
stop testing, although there is no guarantee that the best
solution has been found.

57
 Conventional Strategies
One Factor at a Time (OFAT)
– The OFAT method consists of selecting a starting point, or baseline
set of levels, for each factor, and then successively varying each
factor over its range with the other factors held constant at the
baseline level. After all tests are performed, a series of graphs are
usually constructed showing how the response variable is affected by
varying each factor with all other factors held constant.

58
 Conventional Strategies
One Factor at a Time (OFAT)
– Disadvantage. OFAT
method fails to consider
any possible interaction
between the factors.
– Interactions between
factors are very common,
and if they occur, the
OFAT strategy will usually
produce poor results

59
 Conventional Strategies
One Factor at a time
Trial Factor Test Result Test Average
1 A1 - - Y1
2 A2 - - Y2

Several Factor One at a time

Factor Test Result Test
Trial A B C D Average
1 1 1 1 1 - - Y1
2 2 1 1 1 - - Y2
3 1 2 1 1 - - Y3
4 1 1 2 1 - - Y4
5 1 1 1 2 - - Y5

60
 Conventional Strategies

Several Factors at same time

Factor Test Result Test
Trial A B C D Average
1 1 1 1 1 - - Y1
2 2 2 2 2 - - Y2

Drawbacks:
• Not possible to attribute the change in result to any of the
factor/s.
• Effect of interaction between factors cannot be studied
• These are poor experimental strategies and there is no
scientific basis. The results cannot be validated.
61
 Strategic Design
Full Factorial Design
Factor
Trial A B Response
1 1 1 - -
2 1 2 - -
3 2 1 - -
4 2 2 - -
❖ It is balanced and also orthogonal
❖ Factor A does not influence the estimate of the effect of factor B and
vice versa.
❖ Both factor and interaction effects can be estimated.

Limitations:
✔ Only few factors can be investigated.
✔ When several factors are to be investigated, the number of experiments
to be run under full factorial design is very large.
62
 Need for Strategic Design
❑ To obtain unambiguous results at minimum cost.
❑ Permits consideration of all possible variables/factors
simultaneously
❑ A valid evaluation of the main factors effects and the
interaction effects can be obtained.
❑ With limited number of experiments, the crucial factors can
be obtained.
❑ The statistical concepts used in the design form the basis for
validating the results

63
 Statistical Thinking and its
Role within DOE
• One of the success factors for the effective deployment of
DOE in any organization is the uncompromising commitment
of the senior management team and visionary leadership.
• However, it is not essential that the senior managers have a
good technical knowledge of the working mechanisms of
DOE.
• Statistical thinking is a philosophy of learning and action
based on the following three fundamental principles (Snee, 1990)
(1) All work occurs in a system of interconnected processes.
(2) Variation exists in all processes.
(3) Understanding and reducing variation are the key to success.

64
 Statistical Thinking and its
Role within DOE

• The importance of statistical thinking derives from the

fundamental principle of quality put forth by Deming: ‗Reduce
variation and you improve quality‘.
• Customers of today and tomorrow value products and
services that consistently perform, which can be achieved by
systematically eliminating variation in business processes
(ASQ, 1996). 65
 Statistical Thinking and its
Role within DOE

However, our managers lack statistical thinking and some of

the possible reasons for this are as follows:

1. A shift in the organization's priorities - Global

competition has forced managers to rethink how
organisations are run and to search for better ways to
manage. Problem solving in manufacturing and R&D, while
important, is not seen as particularly relevant to the needs of
management.

66
 Statistical Thinking and its
Role within DOE
2. Managers view statistics as a tool for ‘fire fighting’
actions - One of the most difficult challenges for every
manager is to figure out how to use statistical thinking
effectively to help them make effective decisions.
When a problem arises in the business, managers
want to fix it as soon as possible so that they can deal with
their day-to-day activities. However, what they do not realise
is that the majority of problems are in systems or processes
that can only be tackled with the support of senior
management team.
The result is that management spends too much time
‗fire fighting‘, solving the same problem again and again
because the system was not changed.
67
 Statistical Thinking and its
Role within DOE
3. A change in the mindset of people in the enterprise - ‗If
you cannot change your mind, you cannot change anything‘
by Philosopher George Bernard Shaw.
It is clear that managers, quality professionals and
statisticians all have new roles that require new skills.
Change implies discontinuity and the destruction of familiar
structures and relationships. Change can be resisted
because it involves confrontation of the unknown and loss of
the familiar (Huczynski and Buchanan, 2001)

68
 Statistical Thinking and its
Role within DOE
4. Fear of statistics by managers - ‗Even if managers were
taught statistics at university, it was usually focused on
complex maths and formulas rather than the application of
statistical tools for problem solving and an effective decision
making process. Usually managers have their first
experience with statistical thinking in a workshop inside the
company, applying some tools with the guidance of an
expert. Although this is the best learning method for
understanding and experiencing statistical thinking,
managers may still struggle to apply the principles to a
different problem. This fundamental problem can be tackled
by teaching usable and practical statistical techniques
through real case studies at the university level.
69
 Steps in Experimentation
Little and Hills 1978 suggested that:
• Define the problem
• Determine the objectives
• Select the treatments
• Select the experimental material
• Select the experimental design
• Select the experimental unit and number of replications
• Ensure proper randomization and layout
• Ensure proper means of data collection
• Outline the statistical analysis before doing the experiment
• Conduct the experiment
• Analyze the data and interpret the results
• Prepare complete and readable reports.
 Steps in Experimentation

1. Problem statement
2. Selection of factors, levels and ranges
3. Selection of response variable
4. Choice of experimental design
5. Conducting the experiment
6. Analysis of data
7. Conclusions and recommendations
 1. Problem statement
• A clear definition of the problem contributes to better
understanding of the process being studied and the final
solution of the problem.
• An insight into the problem leads to a better design of the
experiment.
Example:
1. If there is a rejection in bore diameter, specify whether the
rejection is due to oversize or undersize or both.
2. A variation in the compression force of a shock absorber.,
specify clearly if there is less force or more force or both.
3. A defect like crack, blister, etc., specify whether its
observed as random phenomenon on the product or
concentrated to one specific area.
 2. Selection of factors,
levels and ranges
Factors
• Several factors may influence the performance of a
process or product. The following methods may be
used for factor selection keeping customers in mind
1. Brainstorming
2. Flowcharting (especially for processes)
3. Cause-effect diagrams

73
 2. Selection of factors,
levels and ranges
Cause-Effect diagram / Fishbone / Ishikawa diagram

74
 2. Selection of factors,
levels and ranges
Cause-Effect diagram / Fishbone / Ishikawa diagram

75
 2. Selection of factors,
levels and ranges
Cause-Effect diagram / Fishbone / Ishikawa diagram

76
 2. Selection of factors,
levels and ranges
Levels
– To estimate factor effect, a minimum of two levels should
be used in preliminary experimentation.
– When large number of factors are included in the study, the
number of levels is limited to two.
– Detailed experimentation with a few factors, more number
of levels can be considered

77
 2. Selection of factors,
levels and ranges
Range
– Selection of levels for qualitative variables is easy as the
level would be Sometimes ‗on‘ and ‗off‘ or ‗yes‘ and ‗no‘ or
‗high‘ and ‗low‘.
– Selection of the extreme (within the operating range) is
always safe because the system would function safely.
– If levels are chosen too close together, one may not see
any difference and may miss something important outside
the range of experimentation.
– Better to include experts and operators who are
knowledgeable.

78
 3. Selection of response
variable

– The response variable selected must be

appropriate to the problem under study and
provide useful information.
– A clear statement is required of the response to be
studied, i.e., the dependent variables to be
evaluated.
– Example: tensile strength, surface finish,
%elongation, leakages, %defectives, etc.

79
 4. Choice of experimental
design

– Several experimental designs like factorial designs - single

factor and multifactor, fractional factorial designs,
confounding designs etc.
– Selecting the right type of design for the problem under
study is very important, else the results will be misleading.
– Choice of design involves the consideration of the number
of factors and their levels, resources required such as
sample size (number of replications), availability of
samples, time to complete one replication, randomization
restrictions, applicability of blocking etc…

80
 5. Conducting the
experiment
– Experiments should be monitored & carried out as per the plan.
– The allocation of sample as well as the order of experimentation should
be random to even out the effect of unknown & controlled factors.
– Complete Randomization
• any trial has an equal chance of being selected for the first test.
– Simple Repetition –
• any trial has an equal chance of being selected for the first test, but
once that trial is selected all the repetitions are tested for that trial.
– Complete Randomization within Blocks
• where one factor may be very difficult or expensive to change the
test set up for, but others are very easy. If factor A were difficult to
change, the experiment could be completed in two blocks.
• All A1 trials could be selected randomly and then all A2 trials could
be randomly selected.
81
 6. Analysis of data
• Statistical methods should be used to analyse the data.
• Analysis of variance (ANOVA) is widely used to test the statistical
significance of the effects through F-Test.
• Confidence interval estimation is also part of the data analysis. Empirical
models are developed relating the dependent (response) and independent
variables (factors).
• Residual analysis and model adequacy checking are also part of the
data analysis procedure.

• Statistical analysis of data is a must for academic and scientific purpose.

• Graphical analysis and normal probability plot of the effects may be
preferred by Industry.

82
6. Analysis of data

83
 7. Conclusions and
recommendations
• The experimenter must interpret the results and
recommend the same for possible implementation.
• The recommendations include the settings of levels
for all the factors (input variables studied) that
optimizes the output (response).
• It is always better to conduct conformation tests
using the recommend levels for the factors to
validate the conclusions.

84
 Selection of Quality Characteristics
for Industrial Experiments
• The selection of an appropriate quality characteristic
is vital for the success of an industrial experiment.
• To identify a good quality characteristic, it is
suggested to start with the engineering or economic
goal.
• Having determined this goal, identify the fundamental
mechanisms and the physical laws affecting this
goal.

85
Selection of Quality Characteristics
for Industrial Experiments

86
 Selection of Quality Characteristics
for Industrial Experiments
1. Try to use quality characteristics that are easy to measure.
2. Quality characteristics should, as far as possible, be
continuous variables.
3. Use quality characteristics which can be measured precisely,
accurately and with stability.
4. For complex processes, it is best to select quality
characteristics at the subsystem
level and perform experiments at this level before attempting
overall process optimisation.
5. Quality characteristics should cover all dimensions of the ideal
function or the input–output relationship.

87
Selection of Quality Characteristics
for Industrial Experiments

88
Selection of Quality Characteristics
for Industrial Experiments

89
 Choice of sample size
Variable Data:
• Selection of sample size is important.
• A minimum of one test for each trial is required.
• More than one test per trial increases the sensitivity of
experiment to detect small changes in averages of
populations.
• •he sample size (the number of replications)depends on the
resources available (the experimental units) and the time
required for conduct of the experiment.
• The sample size should be as large as possible so that it
meets the statistical principles and also facilitate a better
estimate of the effect.

90
 Choice of sample size
Attribute Data:
• Provides less discrimination than variable data.
• When an item is classified as defective, the measure of how
bad is not indicated. Because of this reduced discrimination,
many pieces of attribute data are required.

Example: In a study on defective parts, if 20 defectives are required and

Suppose the past defective percent in a process is 5%. Then, the total
sample size required (all trials/ runs) shall be 400. 91
ACADEMIC YEAR : 2023 - 24 EVEN
18MEO113T DESIGN OF EXPERIMENTS

UNIT-I: Introduction to DoE

Basic Principles of DOE and Degrees of

Freedom - Selection of Quality
Characteristics for Experiments
Lecture: Day-3

Instructor: Dr. Sumit Pramanik

Research Associate Professor
Department of Mechanical Engineering
FACULTY OF ENGINEERING AND TECHNOLOGY
SRM Institute of Science and Technology, Kattankulathur

Venue: TP 601 (Block: Tech Park) BTech Student: Batch- 1/G

29/01/2024
 Introduction to Experimental Design

Controllable variables (factors)

X1 X2 Xn

Input (s) Output (s)

Process/system Y

Z1 Z2 Zn

Uncontrollable variables (factors)

•Output(s) – performance characteristics

•Controllable factors – easily varied - vital for process characterization
•Uncontrollable factors – difficult to control – attributed to variability/inconsistency in product
performance
•Fundamental strategy of robust design is to minimize the effect of Z‘s by optimally control X‘s

93
 Basic Principles of DOE
• The success of any industrially designed experiment
depends on sound planning, appropriate choice of
design, statistical analysis of data and teamwork skills.

• Principles of experimental design :

– Randomization; Replication; Blocking.

• Applied to reduce or even remove experimental bias (in
some cases, could mask the effect of the really
significant factors).

94
 Basic principles of
experimentation
• The basic principles of experimental design are
Randomization, Replication and Blocking
(Montgomery 2014).
• These principles make a valid test of significance
possible.

95
 Randomization
• A statistical tool used to minimize potential
uncontrollable biases in the experiment by randomly
assigning material, people, order in the experimental
trails to be conducted.
x2
(80,180) (90,180)

• Purpose: to remove bias and variation

x1
• Example: chemical processing
• X1: Reaction Time (80~90 min)
(80,170) (90,170)
• X2: Temperature (170 ~180 degree)

96
 Randomization
• The use of statistical methods requires randomization in any
experiment.
• Dorian Shainin - ‗failure to randomize the trial conditions
mitigates the statistical validity of an experiment‘.
• The allocation of experimental units (samples) for conducting
the experiment as well as the order of experimentation should
be random.
• When complete randomization is not possible, appropriate
statistical design methods shall be used to tackle restriction
on randomization.
97
 Randomization
• The following questions are useful if you decide to apply
randomization strategy to your experiment.

● What is the cost associated with change of factor levels?

● Have we incorporated any noise factors in the experimental
layout?

● What is the set-up time between trials?

● How many factors in the experiment are expensive or difficult
to control?
● Where do we assign factors whose levels are difficult to change
from one to another level?

62
 Replication
The replication is a repetition of the basic
experiment without changing any factor settings.

Each run needs to be R1,R2,….R5 x2 R1,R2,….R5

conduct more than one
(80,180) (90,180)
time in random order.

x1

(80,170) (90,170)

R1,R2,….R5 R1,R2,….R5

99
 Replication
• Replication involves the repetition of the experiment and
obtaining the response from the same experimental set up
once again on different experimental unit (samples).
• An experimental unit may be a material, animal, person,
machine, etc.

• Purpose:
• The purpose of replication is to obtain the magnitude of
experimental error.
• This error estimate (error variance) is used for testing statistically
the observed difference in the experimental data.
• To increase the precision of estimate of error, which is a measure of
the variability of the experimental error
• Obtain a more precise estimate of the mean effect of a treatment.
10
 Replication
• Replications also permit the experimenter to obtain a precise
estimate of the effect of a factor studied in the experiment.
Finally, it is to be noted that replication is not a repeated
measurement.

• Suppose in an experiment five hardness measurements are

obtained on five samples of a particular material using the
same tip for making the indent. These five measurements are
five replicates.

10
 Blocking
• Blocking is a design technique used to improve the precision of
the experiment. Blocking is used to reduce or eliminate the effect
of nuisance factors or noise factors.
• A block is a portion of the experimental material that should be
more homogeneous than the entire set of material or a block is a
set of more or less homogeneous experimental conditions.
• It is also a restriction on complete randomization. More about
blocking will be discussed in factorial designs.
• Purpose: the main purpose of the principle of blocking is to
increase the
• efficiency of an experimental design by decreasing the
experimental error.

102
Blocking

103
 Blocking
Examples :

104
105
 Degrees of freedom (DOF)
• In the context of statistics, DOF refers to the number of
independent and fair comparisons that can be made in a set
of data.

• In the context of DOE,

Degrees of freedom for a main effect = Number of levels - 1

• The number of degrees of freedom for the entire experiment
is equal to one less than the total number of data points or
observations.

106
107
 Degrees of freedom (DOF)
Dosage ‘a’ Dosage ‘b’

Degrees of freedom for Interaction = (a-1) x (b-1)

108
 Degrees of Freedom
(DOF)
• In an 8 trial experiment , where each trial was replicated twice
• So the total observations are 2*8 = 16 , therefore the DOF for
each experiment is equal to 15 (i.e. 16 -1)

• The degrees of freedom for an interaction is equal to the

product of the degrees of freedom associated with each factor
involved in that particular interaction effect.

• An experimenter wishes to study the effect of four process or

design parameters at 3-levels.

109
 Robust Design
• Reduce time to design/develop new products &
processes
• Improve performance of existing processes
• Improve reliability and performance of products
• Achieve product & process robustness
• Evaluation of materials, design alternatives, setting
component & system tolerances, etc.

110
 Selection of Quality Characteristics for
Industrial Experiments
• The selection of an appropriate quality characteristic is
vital for the success of an industrial experiment.

• To identify a good quality characteristic, it is suggested to

start with the engineering or economic goal.

• Having determined this goal, identify the fundamental

mechanisms and the physical laws affecting this goal.

• The quality characteristics is to increase the

understanding of these mechanisms and physical laws.
111
 Selection of Quality Characteristics for
Industrial Experiments
1. Try to use quality characteristics that are easy to measure.
2. Quality characteristics should, as far as possible, be
continuous variables.
3. Use quality characteristics which can be measured precisely,
accurately and with stability.
4. For complex processes, it is best to select quality
characteristics at the sub-system level and perform
experiments at this level prior to attempting overall process
optimisation.
5. Quality characteristics should cover all dimensions of the
ideal function or the input–output relationship.

112
Skills required for an industrial
designed experiment

113
 Characteristics of a well-
planned experiment
• As per, Cox 1958
1. Degree of precision  the probability should be high that
the experiment will be able to measure differences with the
degree of precision the experimenter desires. This implies an
appropriate design and sufficient replication. (i.e., the
standard deviation should be minimal from the experiment)

2. Simplicity  the design should be as simple as possible,

consistent with the objective of the experiment.

114
 Characteristics of a well-
planned experiment
• As per, Cox 1958

3. Absence of systematic error  Experimental units

receiving one treatment should not differ in any systematic
way from those receiving another treatment so that an
unbiased estimate of each treatment effect can be obtained.

4. Range of validity of conclusions  Conclusions should

have as wide a range of validity as possible.

5. Calculation of degree of uncertainty  the experiment

should be designed so that it is possible to calculate the
possibility of obtaining the observed result by chance alone.

115
Characteristics of a well-
planned experiment

Unbiased

116
 Example 1
• Consider a certain painting process which results in various
problems such as
– Orange peel
– Poor appearance
– Voids, etc.
• Often, experimenters measure these characteristics as data and try
to optimize the quality characteristic.
• It is not the function of the coating process to produce an orange
peel.
• The problem could be due to excess variability of the coating
process due to noise factors such as variability in viscosity,
ambient temperature, etc.
117
 Example 1
• We should make every effort to gather data that relate to the engineering
function itself and not to the symptom of variability.
• One fairly good characteristic to measure for the coating process is the
coating thickness.
• It is important to understand that excess variability of coating thickness
from its target value. This could lead to problems such as orange peel or
voids.
• The sound engineering strategy is to design and analyse an experiment so
that best process parameter settings can be determined in order to yield a
minimum variability of coating thickness around the specified target
thickness.

118
 Example 2
• In the service organisations, the selection of quality
characteristics is not very straight forward due to the human
behavioural characteristics present in the delivery of the
service.

In the banking sector

- The number of processing errors

- The processing time for certain transactions

- transactions, the waiting time to open a bank account, etc.

• It is important to measure those quality characteristics which

have an impact on customer satisfaction.
119
 Example 3
• Health care services
– The proportion or fraction of medication errors

– The proportion of cases with inaccurate diagnosis

– The waiting time to get a treatment

– The waiting time to be admitted to an A&E department

– The number of malpractice claims in a hospital every week
or month, etc.

120
ACADEMIC YEAR : 2023 – 24 EVEN
18MEO113T DESIGN OF EXPERIMENTS

UNIT-I: Introduction to DoE

Understanding Key Interaction in Processes

- An Alternative Method for Calculating Two-
Order Interaction Effect
Lecture: Day-3

Instructor: Dr. Sumit Pramanik

Research Associate Professor
Department of Mechanical Engineering
FACULTY OF ENGINEERING AND TECHNOLOGY
SRM Institute of Science and Technology, Kattankulathur

Venue: TP 601 (Block: Tech Park) BTech Student: Batch- 1/G

29/01/2024
Understanding key
interactions in a
process

12
 Introduction to Understanding
Key Interactions in a Process
• For modern industrial processes, the interactions between the
factors or process parameters are a major concern to many
engineers and managers, and therefore should be studied,
analyzed and understood properly for problem solving and
process optimisation problems.
• For many process optimisation problems in industries, the root
cause of the problem is sometimes due to the interaction
between the factors rather than the individual effect of each
factor on the output performance characteristic (or response).
• Here performance characteristic is the characteristic of a
product/ service which is most critical to customers
(Logothetis, 1994)

12
 Significance of interactions
Example : Wave-soldering process of a PCB assembly line
in a certain electronic industry.
Aim : To reduce the number of defective solder joints obtained
from the soldering process.

• The average defect rate based on the existing conditions is

410 ppm (parts per million)
• To perform a simple experiment to understand the
influence of wave-soldering process parameters on the
number of defective solder joints.

124
 What is OFAT ?

One Factor at a Time (OFAT)

– The OFAT method consists of selecting a starting point, or
baseline set of levels, for each factor, and then successively
varying each factor over its range with the other factors held
constant at the baseline level. After all tests are performed, a
series of graphs are usually constructed showing how the
response variable is affected by varying each factor with all
other factors held constant.

125
 Significance of
Interactions
Method of Experimentation :
First Attempt through OVAT approach – 2 levels , – low level
(represented by −1) and high level (represented by +1) for
each parameter

Table 1 List of Process Parameters and Their Levels

126
Understanding Key Interaction in
Processes

Table 2 OVAT Approach to Wave-Soldering Process

First Trial with all

process parameters
kept at low level

127
 Understanding Key
Interaction in Processes

• The difference in the responses between the trials 1 and 2 provides

an estimate of the effect of process parameter ‗A‘. From Table 3.2,
the effect of ‗A‘ (370 − 420 = −50) was estimated when the levels of
‗B‘ and ‗C‘ were at low levels. There is no guarantee whatsoever that
‗A‘ will have the same effect for different conditions of ‗B‘ and ‗C‘.
• Similarly, the effects of ‗B‘ and ‗C‘ can be estimated.
• In the above experiment, the response values corresponding to the
combinations A (−1) B (+1), A (−1) C (+1) and B (−1) C (+1) are
missing.
• Therefore OVAT to experimentation can lead to unsatisfactory
conclusions and in many cases it would even lead to false optimum
conditions. In this case, the team failed to study the effect of each
factor at different conditions of other factors. In other words, the team
failed to study the interaction between the process parameters. 128
 Understanding Key
Interaction in Processes
• Interactions occur when the effect of one process parameter depends on
the level of the other process parameter. In other words, the effect of one
process parameter on the response is different at different levels of the
other process parameter.
• In order to study interaction effects among the process parameters, we
need to vary all the factors simultaneously (Anderson and Whitcomb, 2000).
• For the above wave-soldering process, the engineering team employed a
Full Factorial Experiment (FFE) and each trial or run condition was
replicated twice to observe variation in results within the experimental trials.
• The results of the FFE are given in previous Table.
• Each trial condition was randomised to minimise the effect of undesirable
disturbances or external factors which were uncontrollable or expensive to
control during the experiment.

129
 Understanding Key
Interaction in Processes
• In order to determine whether two process parameters are interacting or
not, one can use a simple but powerful graphical tool called interaction
graphs.
• If the lines in the interaction plot are parallel, there is no interaction
between the process parameters (Barton, 1990).
• This implies that the change in the mean response from low to high level
of a factor does not depend on the level of the other factor.
• On the other hand, if the lines are non-parallel, an interaction exists
between the factors. The greater the degree of departure from being
parallel, the stronger the interaction effect.

130
 Full Factorial Experiment
It is possible to study all the interactions among the factors A,
B and C.

Table 3 Results from a 23 FFE Design with interaction

131
 Understanding Key
Interaction in Processes
For the above example, three two-order interactions and a
third-order interaction can be studied. Third-order and higher
order interactions are not often important for process
optimisation problems and therefore not necessary to be
studied.
• In order to study the interaction between A (flux density) and
B (conveyor speed), it is important to form a table for average
ppm values at the four possible combinations of A and B (i.e.
A(−1) B(−1), A(−1) B(+1), A(+1) B(−1) and A(+1) B(+1)).

132
 Full Factorial Experiment
It is possible to study all the interactions among the factors A,
B and C.

Table 3 Results from a 23 FFE

133
 Interactions
Table 4 Average ppm Values

• From the previous Table, the effect of ‗A‘ (i.e., going from low
level (–1) to high level

134
 Interaction graph
Figure 1 Interaction plot between flux density and conveyor speed

The interaction graph between flux

density and conveyor speed shows
that the effect of conveyor speed on
ppm at two different levels of flux
density is not the same. This
implies that there is an interaction
between these two process
parameters. The defect rate (in
ppm) is minimum when the
conveyor speed is at high level and
flux density at low level.

Effect of conveyor speed on ppm at two different levels of flux density is not the same.
This implies that there is an interaction between these two process parameters.
The defect rate (in ppm) is minimum when the conveyor speed is at high level and flux
density at low level.
10
 Alternative Method for Calculating the
Two-Order Interaction Effect

To calculate the average ppm at high level of (A × B) and low

level of (A × B). The difference between these will provide an
estimate of the interaction effect.

Table 5 Alternative Method to Compute the Interaction Effect

136
 Alternative Method for Calculating the
Two-Order Interaction Effect
As it is an FFE, it is possible to study all the interactions
among the factors A, B and C. The interaction between two
process parameters (say, A and B) can be computed using
the following equation:

137
 Alternative Method for Calculating the
Two-Order Interaction Effect

To calculate the average ppm at high level of (A × B) and low

level of (A × B). The difference between these will provide an
estimate of the interaction effect.

138
 Alternative Method for Calculating the
Two-Order Interaction Effect

To calculate the average ppm at high level of (A × B) and low

level of (A × B). The difference between these will provide an
estimate of the interaction effect.

139
ACADEMIC YEAR : 2023 - 24 EVEN
18MEO113T DESIGN OF EXPERIMENTS

UNIT-I: Introduction to DoE

An Alternative Method For Calculating Two-order
Interaction Effect – Synergistic Interaction,
Antagonistic Interaction – Synergistic Interaction
Versus Antagonistic Interaction
Lecture: Day-4,5,6

Instructor: Dr. Sumit Pramanik

Research Associate Professor
Department of Mechanical Engineering
FACULTY OF ENGINEERING AND TECHNOLOGY
SRM Institute of Science and Technology, Kattankulathur

Venue: TP 601 (Block: Tech Park) BTech Student: Batch- 1/G 05/02/2024
03/02/2024
02/02/2024
2-Level Factorial Designs

141
Interactions

142
Drawbacks of Full Factorial
Designs

143
 Number of Runs vs.
Number of Factors

Number of runs ~ 2.5 x Number of factors ^ 1.7

144
Full Factorial

145
Fractional Factorial for 2-
Level Designs

24 – 1 = 24 / 2 = 8 FFD matrix

146
147
148
149
150
151
152
153
154
155
156
 Alternative Method for Calculating
the Two-Order Interaction Effect
Now consider the interaction
between flux density (A) and
solder temperature.
• The graph shows that the
effect of solder temperature at
different levels of flux density is
almost the same. Moreover, the
lines are almost parallel, which
indicates that there is little
interaction between these two
factors. between the factors.
• The interaction plot suggests that the mean solder defect rate is
minimum when solder temperature is at high level and flux density at low
level.
• Note: Non-parallel lines are an indicator of the existence of interactions
between two factors and parallel lines indicate no interactions between
the factors.
157
 Type of Interactions
1. Additive interaction
2. Suppressive interaction
3. Antagonistic interaction
4. Synergistic interaction

Main Main Effects of B on A

Effect Effect
of A of B
Effect

A B 1-AB 2-AB 3-AB 4-AB

158
 Synergistic Interaction Vs.
Antagonistic Interaction
• The effects of process parameters can be either fixed or
random.
• Fixed process parameter effects occur when the process
parameter levels included in the experiment are controllable
and specifically chosen because they are the only ones for
which inferences are desired.
• For example, if you want to determine the effect of
temperature at 2-levels (180°C and 210°C) on the viscosity of
a fluid, then both 180°C and 210°C are considered to be
fixed parameter levels.
159
 Synergistic Interaction Vs.
Antagonistic Interaction
• Random process parameter effects are associated with those
parameters whose levels are randomly chosen from a large
population of possible levels.

• Inferences are not usually desired on the specific parameter

levels included in an experiment, but rather on the population
of levels represented by those in the experiment.

• Factor levels represented by batches of raw materials drawn

from a large population are examples of random process
parameter levels.
160
 Synergistic interaction
The lines on the plot do not cross each other

Figure 3 Synergistic interaction between two factors A and B.

161
 Antagonistic interaction
The lines on the plot cross each other
Figure 4 Antagonistic interaction between two factors A and B.

162
 Synergistic Interaction Vs
Antagonistic Interaction
The effects of process parameters can be either
fixed or random.
• Fixed Process Parameters
o Fixed process parameter effects occur when
the process parameter levels included in the
experiment are controllable and specifically
chosen because they are the only ones for
which inferences are desired.
o For example, if you want to determine the
effect of temperature at 2-levels (180°C and
210°C) on the viscosity of a fluid, then
both180°C and 210°C are considered to be
fixed parameter levels. 163
 Synergistic Interaction Vs
Antagonistic Interaction
The effects of process parameters can be either fixed or
random.
• Random Process Parameters
o Random process parameter effects are associated
with those parameters whose levels are randomly
chosen from a large population of possible levels.
o Inferences are not usually desired on the specific
parameter levels included in an experiment, but
rather on the population of levels represented by
those in the experiment.
o For example, Factor levels represented by batches of
raw materials drawn from a large population

164
 Synergistic Interaction Vs
Antagonistic Interaction
Synergistic interaction means that the effect of two
chemicals taken together is greater than the sum of
their separate effect at the same doses. An
example is pesticide and fertilizer. (i.e., Combined
effect of two factors is greater than the individual
factors on the response. In other words, when
higher level of one independent variable enhances
the effect of another independent variable)

165
 Synergistic Interaction Vs
Antagonistic Interaction
Antagonistic interaction means that the effect of two
chemicals is actually less than the sum of the effect
of the two drugs taken independently of each other.
This is because the second chemical increases the
excretion of the first, or even directly blocks its toxic
actions. (i.e., Combined effect of two factors is
lesser than the individual factors on the response.
In other words, When main effect is non-significant
and interaction is significant. In this situation the
two independent variables tend to reverse each
others effect).

166
 Synergistic Interaction

The lines on the plot do not cross each other

167
 Antagonistic Interaction

The lines on the plot cross each other

168
 References
• Antony, J. (2014). Design of experiments for
engineers and scientists. Elsevier.

• Krishnaiah, K., & Shahabudeen, P. (2012). Applied

design of experiments and Taguchi methods. PHI
Learning Pvt. Ltd.

169
Thank you

170