MT4001 Experimental
Design
Monday, 12-3 pm
Dr. Ir. Budi Hartono
Setiamarga
Reference
Grading
Homework and Quiz (10%)
Midterm 1 (20%)
Midterm 2 (20%)
Class Project (20%)
Oral Final Test (30%)
Date (2016)
Notes
W1: Aug 22
Introduction: Chapter 1 Box
W2: Aug 29
Ch 2 Basic Statistics
W3: Sept 5
Ch 2 Basic Statistics + Class
Project Ch 5
W4: Sept 12
National Holiday: Idul Adha
W5: Sept 19
Class Project Proposal Submission
+Ch 3
W6: Sept 26
Ch 3
W7: Oct 3
Midterm 1: Covering Chapter 1-3
W8: Oct 10
MT review; Class Project Progress
Report
W9: Oct 17
Ch 4 Comparing more than 2
entities
W10: Oct 24
No class
W11: Oct 31
Ch 4 Comparing more than 2
entities
W12: Nov 7
Midterm 2 Covering Chapter 3
and 4
Introduction
Knowledge is power.
the key to innovation and
profit. But the getting of
new knowledge can be
complex, time consuming, and
costly.
To be successful in such
enterprise, you must learn
about learning.
Learning key to idea
generation, to process
improvement, to the
development of new and
robust products and
processes.
Aim: simplify and
accelerate the generation,
testing, and development
of new ideas
statistical experimental
design:
catalyze scientific method
and greatly increase its
efficiency
Scientific research is a process of guided
learning. The object of statistical methods is
to make that process as efficient as possible.
Convergence
Process
The Iterative Learning
Process
Example of Rita Stoveing, a Chemist
(p3)
The Iterative Problem
Solving
Important Considerations
Subject Matter Knowledge
To perceive and explore tentative
models
The route to Problem
Solving is not UNIQUE
-knowledge of strategy
illustration: The game of 20 Questions
There were many ways to get
to Rome
The Experimenters
Problems
1. Complexity
E.g. Experiment for process improvement
Input
Output
influen
Variabl
Variable
ce
es
s
Temperature,
Hardness, wear
(factors
(respons
heating duration,
rate,
yield
quenching media,
strength, etc.
)
es)
etc.
What does what to
Tempering of Tool Steel
-input variables: k factors (e.g. Ttemperature and t-time)
-output variables: p responses (e.g. Hardness
Rc and Wear [gram]
k x p entities to consider more complex
-Will the input variables behave
independently or dependently on on
another?
-Change one factor at a time is not
sufficient use Factorial Design (Class
Project)
Example of Complexity of
Problems
Secondary Hardening in Tempering of
Tool Steel
Factors: % Mo and Tempering Temperature;
Example of Complexity of
Problems:
X Ray White Radiation
Factors: Voltage;
Response: The
presence of
Characteristic X Ray
and
The Experimenters Problem
2. Experimental Error
Variability that is not explained by
known influences
E.g. due to Variations in raw
materials in sampling process and in
the settings of the experimental
factors
Good experimental design helps to
protect real effects from being
obscured by experimental error and
conversely having the investigator
mistakenly believe in effects that
do not exist
Parsnip plant
a long tapering cream-colored root with a sweet
flavor.
Hardness of Parsnip samples A
and B
The experimenters Problem
3. Confusion of
CORRELATION with
CAUSATION
Civilian Population if
Oldenburg (Germany) against
the number of Storks
What is Stork?
How to Use Statistical
Techniques
1. Find out as much as you can
about the problem
What is the object of this
investigation?
Who is responsible?
I am going to describe your problem:
Am I correct?
How to Use Statistical
Techniques
2. Dont forget non-statistical
Knowledge
statistical techniques are
useless unless combined with
appropriate subject matter
knowledge and experience
Do you have any past data?
How were the data collected?
In What order? On what days?
By Whom? How? May I see
them?
How were the responses
measured? Have the
necessary devices been
recently checked?
Do you have other data
like these?
How does the equipment
work? What does it look like?
May I see? May I see it work?
How much physical theory is
known about the phenomenon?
If the process concerns a
manufacturing process, what
are the sampling, measurement
and adjustment protocols?
How to Use Statistical
Techniques
3. Define Objectives
A. Define clearly the objectives of the
study
B. Be sure that all interested parties
concur in these objectives
C. Make sure that the necessary
equipment, facilities, scientific
personnel, time, money, and
adequate management support are
available to perform the proposed
investigation
D. Agree on the criteria that will
determine when the objectives have
been met
E. Arrange that if the objectives have
to be changed all interested parties
are made aware of the new
objectives and criteria.
Not giving these matter sufficient
attention can produce serious
dificulties and sometimes disaster
How to Use Statistical
Techniques
4. Learn from each other:
The Interplay Between
Theory and Practice
Statisticians VS Engineers
