Neil W. Polhemus, CTO, StatPoint Technologies, Inc.

Design of Experiments Using

Statgraphics Centurion

Copyright 2011 by StatPoint Technologies, Inc.

Web site: www.statgraphics.com

Design of Experiments Wizard

Phase 1: Creating an experiment

Phase 2: Analyzing the results

Phase 3: Further experimentation

2

Phase 1: Creating an experiment

Phase 1: Creating an experiment

Step 1: Define responses

Step 2: Define experimental factors

Step 3: Select design

Step 4: Specify model

Step 5: Select runs

Step 6: Evaluate design

Step 7: Save design

Phase 2: Analyzing the results

Phase 3: Further experimentation

3

Phase 2: Analyzing the results

Phase 1: Creating an experiment

Phase 2: Analyzing the results

Step 8: Analyze data

Step 9: Optimize responses

Step 10: Save results

Phase 3: Further experimentation

4

Phase 3: Further experimentation

Phase 1: Creating an experiment

Phase 2: Analyzing the results

Phase 3: Further experimentation

Step 11: Augment design

Step 12: Extrapolate

5

Example – from Response Surface

Methodology by Myers and

Montgomery

6

Find the settings of time, temperature and catalyst that maximize

the conversion percentage of a chemical process while keeping

the thermal activity as close as possible to 57.5.

DOE Wizard – Main Window

7

Step 1: Define responses

8

Step 2: Define experimental factors

9

Step 3: Select design

10

Step 3: Select design (cont.)

11

Step 3: Select design (cont.)

12

Step 3: Select design (cont.)

13

Datasheet

14

Step 4: Select model

15

Step 5: Select runs (D-optimal

design only)

16

Step 6: Evaluate design

17

Plot of design space

18

RSM Example

8.3 10.3 12.3 14.3 16.3 18.3

time

160170

180190

200210

temperature

1.6

1.9

2.2

2.5

2.8

3.1

3.4

ca

taly

st

Prediction variance plot

19

Prediction Variance Plotcatalyst=2.5

8.110.1

12.114.1

16.118.1

time

150

170

190

210

230

temperature

0

0.2

0.4

0.6

0.8

1

Stn

d.

err

or

Power curve

20

Power Curve for A:time

alpha = 5.0%, sigma estimated from total error with 10 d.f.

-4 -2 0 2 4

True effect/sigma

0

0.2

0.4

0.6

0.8

1

Pro

bab

ilit

y o

f d

ete

cti

on

Do the experiment!

Then enter the results:

21

Step 8: Analyze data

22

Step 8: Analyze data (cont.)

23

ANOVA table

24

Pareto chart

25

Standardized Pareto Chart for conversion

0 2 4 6 8

Standardized effect

A:time

AB

AA

BC

BB

B:temperature

CC

C:catalyst

AC +

-

Exclude effects

26

Final models

27

Standardized Pareto Chart for conversion

0 2 4 6 8

Standardized effect

A:time

BC

BB

B:temperature

CC

C:catalyst

AC +

-

Standardized Pareto Chart for thermal activity

0 2 4 6 8 10

Standardized effect

AA

C:catalyst

A:time

+

-

28

Step 9: Optimize responses

29

Step 9: Optimize responses (cont.)

Step 9: Optimize responses (cont.)

30

3-D contour plot

31

Desirability Plotcatalyst=2.31889

8.1 10.1 12.1 14.1 16.1 18.1

time

150170

190210

230

temperature

1.6

1.9

2.2

2.5

2.8

3.1

3.4

cata

lyst

Desirability

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Desirability

Overlaid contour plots

32

Overlay Plot

catalyst=2.34412

conversion

thermal activity

8.1 10.1 12.1 14.1 16.1 18.1

time

150

170

190

210

230

tem

pera

ture

75.0

83.0

91.0

99.0

107.0115.0123.057.0 59.0 61.0 63.0 65.0 67.0 69.0 71.0

Step 10: Save results

33

Step 11: Augment design

34

Step 12: Extrapolate

35

Step 12: Extrapolate (cont.)

36

More… (next webinar)

Creating and using RPDs (Robust Parameter Designs)

Analyzing designs with both process and mixture

components

Using D-optimal designs to fix a botched experiment

37

More Information

Go to www.statgraphics.com

Or send e-mail to info@statgraphics.com

38