Advanced DOE
Advanced DOE
Learning Objectives
Plackett-Burman Designs
Box-Behnken Designs
Central Composite (Box-Wilson)
Design
Seagate Confidential 2 Supplier Six Sigma Modular Training
Advanced DOE
Plackett-Burman Designs
Plackett-Burman Designs are two-level fractional factorial
designs, of Resolution III, used for studying K=N–1
variables in N runs, where N is a multiple of 4.
It is use to screen out a large number of factors with
small number of runs.
The design can be specified by giving just the first
column.
8Runs +++–+––
12 Runs ++–+++–––+–
16 Runs ++++–+–++––+–––
20 Runs ++--++++-+-+----++-
24 Runs +++++-+-++--++--+-+----
Seagate Confidential 3 Supplier Six Sigma Modular Training
Advanced DOE
Plackett-Burman Designs
Example: 8 Run Plackett-Burman Design: + + + – + – –
Run A B C D E F G
1 + – – + – + +
2 + + – – + – +
3 + + + – – + –
4 – + + + – – +
5 + – + + + – –
6 – + – + + + –
7 – – + – + + +
8 – – – – – – –
Each Plackett-Burman Design can have up to k = (n – 1) factors.
If k < (n – 1), only the first k columns are used.
Seagate Confidential 4 Supplier Six Sigma Modular Training
Advanced DOE
Example 1
An engineer plans to investigate the main effects
of 5 factors. He can only perform 10
experimental runs at best. Select an appropriate
experimental design.
Seagate Confidential 5 Supplier Six Sigma Modular Training
Advanced DOE
Example 1
Stat DOE Factorial Create Factorial Design
Seagate Confidential 6 Supplier Six Sigma Modular Training
Advanced DOE
Plackett-Burman Designs
StdOrder RunOrder CenterPt Blocks A B C D E
1 2 1 1 1 –1 –1 1 –1
2 5 1 1 1 1 –1 –1 1
3 3 1 1 1 1 1 –1 –1
4 1 1 1 –1 1 1 1 –1
5 7 1 1 1 –1 1 1 1
6 4 1 1 –1 1 –1 1 1
7 6 1 1 –1 –1 1 –1 1
8 8 1 1 –1 –1 –1 –1 –1
Note that only the first 5 columns are defined.
Seagate Confidential 7 Supplier Six Sigma Modular Training
Advanced DOE
Plackett-Burman Designs
Each of the k factors are run at 2 levels.
Pro:
a) Allows limited number of runs to evaluate large numbers of
factors.
b) Interactions are not evaluated, the effects for all factors are
placed in a Pareto chart and hence important factors will be
surfaced out for indepth testing.
Con:
a) Assume that interactions are not strong enough to mask
main effects and quadratic effects (if exist) are closely
related to the factor linear effects.
Seagate Confidential 8 Supplier Six Sigma Modular Training
Advanced DOE
Box-Behnken Designs
Run A B C It is used for 3-level design for
1 -1 -1 0
2 1 -1 0
modeling quantitative factors (time,
3 -1 1 0 temperature, pressure, etc).
4 1 1 0
5 -1 0 -1
6 1 0 -1
It uses 2k design while holding certain
7 -1 0 1 factors at their centerpoint.
8 1 0 1
9 0 -1 -1
10 0 1 -1
11 0 -1 1
12 0 1 1
13 0 0 0
14 0 0 0
15 0 0 0
Seagate Confidential 9 Supplier Six Sigma Modular Training
Advanced DOE
Box-Behnken Designs
Stat DOE Response Surface
Create Response Surface Design
(1)
(2)
(3)
Seagate Confidential 10 Supplier Six Sigma Modular Training
Advanced DOE
Box-Behnken Designs
Each of the k factors are run at 3 levels.
Pro:
a) Able to estimate all linear and quadratic effects, and
2-factor interactions.
b) Less runs required, compared vs 3k Factorial Designs.
c) Does not include any corner points.
Con:
a) Number of runs is large enough to estimate all
quadratic and 2-factor interactions, regardless of need.
b) Cannot be built-up from a 2k-p Factorial Design.
Seagate Confidential 11 Supplier Six Sigma Modular Training
Advanced DOE
Central Composite (Box-Wilson) Design
Run A B C
1 -1 -1 -1
2 1 -1 -1
3 -1 1 -1
4 1 1 -1
5 -1 -1 1
6 1 -1 1
7 -1 1 1
8 1 1 1
9 -1.682 0 0
10 1.682 0 0
11 0 -1.682 0
12 0 1.682 0
13 0 0 -1.682
14 0 0 1.682
15 0 0 0
16 0 0 0
17 0 0 0
18 0 0 0
19 0 0 0
20 0 0 0
= nf¼ where nf is number of runs
in factorial portion of CCD
Seagate Confidential 12 Supplier Six Sigma Modular Training
Advanced DOE
Central Composite (Box-Wilson) Design
+ =
Factorial Center Central
Points Points & Composite
Axial (Box-Wilson)
(8 runs) Points Design
(6 + 6 runs) (20 runs)
nc 4 nf 1 2k
na = 2k
Seagate Confidential 13 Supplier Six Sigma Modular Training
Advanced DOE
Central Composite (Box-Wilson) Design
Stat DOE Response Surface
Create Response Surface Design
(1)
(2)
(3)
Seagate Confidential 14 Supplier Six Sigma Modular Training
Advanced DOE
Central Composite (Box-Wilson) Design
Each of the k factors can be run at 5 levels.
Pro:
a) Able to estimate all linear effects, and selected
quadratic effects and 2-factor interactions.
b) Can be built-up from a 2k-q screening design, by
adding axial points.
Con:
a) Best suited for quantitative factors.
b) Some axial points may be in non-desirable
conditions.
Seagate Confidential 15 Supplier Six Sigma Modular Training
Advanced DOE
Comparison of 3-Level Designs
Scenario 3k Factorial Box-Behnken CCD
k=3 9+(0)=9 15+(3)=18 13+(3)=16
with no interaction
k=4 27+(0)=27 27+(3)=30 20+(4)=24
with 3 2-factor interactions
k=5 27+(0)=27 46+(6)=52 20+(2)=22
with 1 2-factor interaction
k=5 81+(0)=81 46+(6)=52 33+(7)=40
with all 2-factor interactions
k=7 81+(0)=81 62+(6)=68 33+(3)=36
with 7 2-factor interactions
Numbers in parenthesis = the number of replicated
center points.
For CCD, nc 4 nf 1 2k
Seagate Confidential 16 Supplier Six Sigma Modular Training
Advanced DOE
End of Topic
What question do you have?
Statistical significance does not always
mean practical significance.
Seagate Confidential 17 Supplier Six Sigma Modular Training
Advanced DOE
Reference Book
Author:
George EP Box
William G Hunter
J Stuart Hunter
Publisher:
John Wiley & Sons, Inc.
ISBN: 0-471-09315-7
Seagate Confidential 18 Supplier Six Sigma Modular Training
Advanced DOE
Reference Book
Author:
Stephen R Schmidt
Robert G Launsby
Publisher:
Air Academy Press
ISBN: 1-880156-05-9
Seagate Confidential 19 Supplier Six Sigma Modular Training
