Cover image for Practical experiment designs for engineers and scientists
Title:
Practical experiment designs for engineers and scientists
Author:
Diamond, William J., 1919-
Personal Author:
Edition:
Third edition.
Publication Information:
New York : Wiley, [2001]

©2001
Physical Description:
xvi, 423 pages : illustrations ; 25 cm
Language:
English
Subject Term:
Electronic Access:
Table of Contents http://www.loc.gov/catdir/toc/onix06/00043607.html
ISBN:
9780471390541
Format :
Book

Available:*

Library
Call Number
Material Type
Home Location
Status
Central Library QA279 .D5 2001 Adult Non-Fiction Non-Fiction Area
Searching...

On Order

Summary

Summary

Most books cover the subject from a statistical or theoretical point of view. Ideal for working engineers, this book uses real-world examples and boils statistical theory and analysis down to its simplest form.
∗ Features new and updated material, including cases and a larger focus on multivariate analysis.
∗ Uses simple analysis tools for practical implementation on the job.
∗ Targets experiment planning as the groundwork for quality experiments.


Author Notes

William J. Diamond is a consulting engineer based in Riverview, Florida


Table of Contents

Prefacep. xiii
Acknowledgmentsp. xvi
Part 1 Fundamentals of Experiment Designp. 1
1 Introduction to Experiment Design: Fundamental Conceptsp. 3
Origin of Bad Experimentsp. 3
Philosophy for Good Experimentsp. 5
Project Strategyp. 5
Experiment Strategyp. 6
Perceptions of Realityp. 7
Theology for Experimentersp. 7
Populationsp. 10
Distributionsp. 11
Means and Variancesp. 14
2 Introduction to Experiment Design: Elements of Decision Makingp. 18
Stating the Alternative Decisionsp. 18
Defining the Risksp. 20
Establishing an Objective Criterionp. 23
One Population Sampledp. 24
Two Populations Sampledp. 27
Computing the Sample Sizep. 28
Importance of Np. 31
Choosing [alpha], [beta], and [delta]p. 34
Operating Characteristic Curvesp. 39
Power Curvesp. 41
3 Simple Comparative Experiments: Decisions About Population Meansp. 44
Cases 1-3 Variance or Variances Are Knownp. 46
Case 1. H[subscript 0]:[mu subscript 1] = [mu subscript 0]; Variance is Knownp. 46
Case 2. H[subscript 0]:[mu subscript 1] = [mu subscript 2]; Variances Are Equal and Knownp. 48
Case 3. H[subscript 0]:[mu subscript 1] = [mu subscript 2]; Variances Are Not Equal and Both Are Knownp. 50
Cases 4-6 Variance or Variances Are Unknownp. 51
Case 4. H[subscript 0]:[mu subscript 1] = [mu subscript 0]; Variance is Unknownp. 52
Case 5. H[subscript 0]:[mu subscript 1] = [mu subscript 2]; Variances Are Equal and Unknownp. 56
Case 6. H[subscript 0]:[mu subscript 1] = [mu subscript 2]; Variances Are Not Equal and Are Unknownp. 58
Cases 7-9 Sample Estimate of Variance Is Knownp. 61
Case 7. H[subscript 0]:[mu subscript A] = [mu subscript 0]; S[superscript 2] Is Knownp. 62
Case 8. H[subscript 0]:[mu subscript A] = [mu subscript B]; S[superscript 2] Is Known and [sigma superscript 2 subscript A] = [sigma superscript 2 subscript B]p. 63
Case 9. H[subscript 0]:[mu subscript A] = [mu subscript B]; S[superscript 2 subscript A] and S[superscript 2 subscript B] Are Known and [sigma superscript 2 subscript A] = [sigma superscript 2 subscript B]p. 64
Special Casesp. 65
Special Case 1 (Paired Comparisons)p. 65
Special Case 2 (Some Data Lost)p. 67
4 Simple Comparative Experiments: Decisions About Population Variancesp. 71
Sample Sizep. 71
Criterion Valuesp. 73
Case 1. H[subscript 0]:[sigma superscript 2 subscript 1] = [sigma superscript 2 subscript 0]; H[subscript a]:[sigma superscript 2 subscript 1] [ [sigma superscript 2 subscript 0]p. 74
Case 2. H[subscript 0]:[sigma superscript 2 subscript 1] = [sigma superscript 2 subscript 0]; H[subscript a]:[sigma superscript 2 subscript 1] ] [sigma superscript 2 subscript 0]p. 75
Case 3. H[subscript 0]:[sigma superscript 2 subscript 1] = [sigma superscript 2 subscript 2]; H[subscript a]:[sigma superscript 2 subscript 1] [not equal] [sigma superscript 2 subscript 2]p. 76
Case 4. Combined Problem: H[subscript a1]:[mu subscript A] [not equal] [mu subscript B]; H[subscript a2]:[sigma superscript 2 subscript A] [not equal] [sigma superscript 2 subscript B]p. 77
Case 5. H[subscript 0]:[mu subscript A] = [mu subscript B]; Assumed [sigma superscript 2 subscript A] = [sigma superscript 2 subscript B]p. 78
5 Sequential Experimentsp. 82
Case 1. [sigma superscript 2] Is Known; H[subscript a]:[mu subscript 1] ] [mu subscript 0]p. 83
Case 2. [sigma superscript 2] Is Known; H[subscript a]:[mu subscript 1] [not equal] [mu subscript 0]p. 85
Case 3. [sigma superscript 2] Is Unknown; H[subscript a]:[mu subscript 1] ] [mu subscript 0]p. 87
Part 2 Two-Level Multivariable Experimentsp. 91
6 Two-Level Multivariable Experiments: Eight-Trial Hadamard Matrix Designsp. 93
Introduction to Matrix Experimentsp. 93
Other Advantages of Matrix Experiment Designp. 97
Construction of Hadamard Matricesp. 98
Use of the 8 x 8 Hadamard Matrixp. 100
One Variable, Each Treatment Combination Replicated Four Timesp. 101
Two Variables, Each Treatment Combination Replicated Twicep. 103
Three Variablesp. 109
Four Variables (Resolution IV Design)p. 115
Five, Six, or Seven Variables (Resolution III Designs)p. 121
7 Two-Level Multivariable Experiments: Hadamard Matrices Greater Than Order 8p. 130
16-Trial Designsp. 132
32-Trial Designsp. 134
64-Trial Designsp. 135
128-Trial Designsp. 136
Smoak Modified Designs for Resolution IV Designsp. 136
Other Hadamard Matrices Not of Order 2[superscript n]p. 139
Summaryp. 148
8 John's Three-Quarter Fractional Factorialsp. 150
Resolution V Designsp. 150
Four Variables in 12 Trialsp. 151
Seven and Eight Variables in 48 Trialsp. 158
Nine to 11 Variables in 96 Trialsp. 162
Resolution IV Designsp. 163
Three Variables in Six Trialsp. 163
Five or Six Variables in 12 Trialsp. 164
Nine to 12 Variables in 24 Trialsp. 166
Seventeen to 24 Variables in 48 Trialsp. 167
Thirty-Three to 48 Variables in 96 Trialsp. 167
9 Unbalanced Resolution V Designsp. 175
Class A Designsp. 176
Six-Variable, 28-Trial Designp. 176
Seven-Variable, 36-Trial Designp. 181
Eight-Variable, 44-Trial Designp. 183
Nine- to 11-Variable Designsp. 185
Class B Designsp. 185
Seven-Variable, 38-Trial Designp. 185
Eight-Variable, 46-Trial Designp. 188
Nine-Variable, 71-Trial Designp. 189
Twelve-Variable, 143-Trial Designp. 190
Class AB Designsp. 191
10 Resolution V Designs with Efficiency = 1p. 195
Three-Variable Designsp. 196
Four-Variable Designsp. 197
Five-Variable Designsp. 197
Six and Higher Numbers of Variablesp. 198
11 Hadamard Matrix Designs for Binomial and Poisson Responsesp. 200
Binomialp. 200
Poissonp. 204
12 Summary of Two-Level Matrix Designsp. 208
13 A Computer Program for Generating Hadamard Matrix Designs and Analyzing the Data from Such Designsp. 215
Designing the Experimentp. 216
Example of Saturated Resolution III Designs: 15 Variables in 16 Trials, Not Using the START Program, and Coding the Levels of the Variables as 0 or 1p. 217
Example of Saturated Resolution IV Designs: John's Three-Quarter Fractional Factorial with Six Variables in 12 Trials, Not Using the START Program, and Coding the Levels of the Variables as 0 or 1p. 218
Example of Resolution IV Designs: John's Three-Quarter Fractional Factorial with Nine Variables in 24 Trials, Using the START Program, and Using the Actual Levels of the Variablesp. 219
Example of Saturated Resolution III Designs: 15 Variables in 16 Trials, Not Using the START Program, and Using the Actual Levels of the Variablesp. 220
Analyzing the Datap. 221
Example of Resolution V Designs: Four Variables in 16 Trialsp. 221
Example of Saturated Resolution IV Designs: Eight Variables in 16 Trialsp. 222
Example of Resolution III Designs: 13 Variables in 16 Trialsp. 223
Example of John's Three-Quarter Fractional Factorial: Nine Variables in 24 Trialsp. 224
Appendix 1 APL Computer Program Codingp. 226
Appendix 2 Confounding of Two-Factor Interactions in Resolution IV Designp. 234
14 Analysis of Goodnessp. 243
15 Alternative Methods of Analysisp. 251
Part 3 Multilevel Multivariable Experimentsp. 261
16 Multilevel Experiments with Qualitative Variablesp. 263
Use of ANOVAp. 264
Latin and Greco-Latin Squaresp. 272
17 Multilevel Experiments with Quantitative Variablesp. 281
Central Composite Rotatable Designsp. 282
Designs for Experiments Where the Levels of Some of the Variables Are Differentp. 288
16-Trial Experiment Designs with Variables at Two Levels and Three Levelsp. 295
Two Variables at Two Levels and One Variable at Three Levelsp. 295
Three Variables at Two Levels and One Variable at Three Levelsp. 295
Main Effectsp. 296
Interactionsp. 297
Two Variables at Two Levels and Two Variables at Three Levelsp. 297
Four Variables at Two Levels and One Variable at Three Levelsp. 298
32-Trial Experiment Designs with Variables at Two Levels and Three Levelsp. 298
64-Trial Experiment Designs with Variables at Two Levels and Three Levelsp. 303
Other Multilevel Experiment Designsp. 306
18 Experiment Designs for Chemical Composition Experimentsp. 309
Extreme-Vertices Designsp. 309
Simplex Designsp. 315
19 Random-Strategy Experimentsp. 322
Part 4 Related Topicsp. 327
20 Blocking an Experimentp. 329
21 Validation of Test Methodsp. 335
22 Concepts for a Complete Project Strategyp. 348
Project Definitionp. 353
Planning the Projectp. 353
Design of the First Experimentp. 355
Design of the Second Experimentp. 357
Design of the Third Experimentp. 362
Design of the Fourth Experimentp. 367
23 Project Engineer's Gamep. 369
Project Assignmentp. 370
Use of the Computer Programp. 370
Playing the Gamep. 371
Other Project Engineer Gamesp. 372
Appendix Computer Program and Response Equation for Project Engineer's Gamep. 373
Instructions for Game Directorp. 376
Min and Max Values of Variablesp. 376
Response Equationsp. 376
24 Estimation of Variancep. 378
25 Testing Distributionsp. 384
Case 1. One Populationp. 384
Case 2. Two Populationsp. 386
Part 5 General References, Symbols, Tables, and Answers to Exercisesp. 389
General Referencesp. 391
Symbolsp. 397
Tablesp. 399
Answers to Exercisesp. 414
Indexp. 419

Google Preview