RSM Simplified : Optimizing Processes Using Response Surface Methods for Design of Experiments

by Whitcomb; Patrick J.

• ISBN13: 9781563272974
• ISBN10: 1563272970
• Edition: CD
• Format: Paperback
• Copyright: 2004-11-17
• Publisher: Productivity Pr

Summary

RSM Simplified keeps formulas to a minimum and makes liberal use of figures, charts, graphs and checklists. It offers many relevant examples, with amusing sidebars and do-it-yourself exercises that will lead readers to the peak potential for their product quality and process efficiency. The authors, Mark J. Anderson and Patrick J. Whitcomb, are principals of Stat-Ease, a provider of DOE training, consulting, and software. They both are professional chemical engineers. Anderson has more of a business background while Whitcomb specializes in statistics. Incorporated into this book is the more advanced Design-Expert, version 7 software for Windows, with a 180-day trial, so the reader can do the complex statistical computations, generate the necessary graphics (2D and 3D maps) and perform the numerical optimization.

Author Biography

Mark J. Anderson and Patrick J. Whitcomb, are principals of Stat-Ease, a provider of DOE training, consulting, and software

Table of Contents

Preface	ix
Introduction to the Beauty of Response Surface Methods	1	(30)
Strategy of Experimentation: The Role for RSM	1	(4)
One Map Better Than 1000 Statistics (with Apologies to Confucius?)	5	(3)
Using Math Models to Draw a Map That's Wrong (But Useful!)	8	(16)
Finding the Flats to Achieve Six Sigma Objectives for Robust Design	24	(2)
Practice Problem 1-1	26	(1)
Appendix: Don't Let R2 Fool You	27	(4)
Lessons to Learn from Happenstance Regression	31	(26)
The Dangers of Happenstance Regression	31	(11)
The Storks-Bring-Babies Fallacy, in which Correlation Implies Causation	39	(3)
Diagnosing Input and Output Data for Outlying and/or Influential Points	42	(7)
Extrapolation Can Be Hazardous to Your Health!	49	(4)
Practice Problem 2-1	53	(1)
Appendix: A Brief Overview of Matrix Terminology	54	(3)
Factorials to Set the Stage for More Glamorous RSM Designs	57	(20)
Putting OFAT in the Fire	57	(4)
Factorial Designs as an Option to OFAT	61	(9)
Two-Level Factorial Design with Center Points: A Foundation for RSM	62	(5)
A Test for Curvature Tailored to 2k Designs with Center Points	67	(2)
Further Discussion on Testing for Lack of Fit	69	(1)
The Fork in the Road Has Been Reached!	70	(2)
Practice Problem 3-1	72	(1)
Practice Problem 3-2	73	(4)
Central Composite Desing: Stars Added---RSM Show Begins	77	(26)
Augmenting a Two-Level Factorial with Star Points to Fit Curvature	77	(2)
CCD Example: Part Two of a Chemical Reaction Example	79	(18)
How to Select a Good Predictive Model	79	(5)
Inspecting the Selected Model	84	(6)
Model Graphs	90	(7)
Practice Problem 4-1	97	(2)
Practice Problem 4-2	99	(1)
Appendix 4A: Details for the Sequential Model Sum of Squares Table	100	(1)
Appendix 4B: Details for Lack-of-Fit Tests	101	(2)
Three-Level Designs	103	(20)
The Box-Behnken Design	105	(3)
Case Study on Trebuchet	108	(8)
Practice Problem 5-1	116	(4)
Appendix: Box-Cox Plot for Transformations	120	(3)
Finding Your Sweet Spot for Multiple Responses	123	(26)
``Less filling'' versus ``Tastes great!''---Always a Trade-off	123	(10)
Making Use of Desirability Function to Evaluate Multiple Responses	124	(8)
Numerical Search for Optimum	132	(1)
Summary	132	(1)
Case Study with Multiple Responses: Making Microwave Popcorn	133	(8)
Practice Problem 6-1	141	(1)
Practice Problem 6-2	142	(2)
Appendix: Simplex Optimization	144	(5)
Computer-Generated Optimal Designs	149	(20)
Quantifying Operating Boundaries with Multilinear Constraints	152	(2)
Computer-Generated Design for Constrained Region of Interest	154	(4)
A Real-Life Application for Computer-Generated Optimal Design	158	(1)
Other Applications for Optimal Designs	159	(1)
Practice Problem 7-1	160	(1)
Practice Problem 7-2	161	(1)
Appendix 7A: How to Quantify Multilinear Constraints	162	(2)
Appendix 7B: How to Generate a Good Candidate Set	164	(1)
Appendix 7C: An Algorithm for D-Optimal Point Selection	165	(4)
Everything You Should Know About CCDs (but Dare Not Ask!)	169	(28)
Rotatable Central Composite Designs	172	(2)
Going Cuboidal when Your Region of Operability and Interest Coincide	174	(4)
Practical Placement of Stars---Outside the Box but Still in the Galaxy	178	(2)
Center Points in Central Composite Designs	180	(2)
Final Comments---How We Recommend You Construct CCDs	182	(2)
Practice Problem 8-1	184	(4)
Practice Problem 8-2	188	(2)
Appendix 8A: Some Details on Blocking CCDs and Issues of Orthogonality	190	(3)
Appendix 8B: Small Composite Designs (SCDs) to Minimize Runs	193	(4)
Good Alternative to SCDs: Minimum-Run Resolution V CCDs	195	(2)
RSM for Six Sigma	197	(18)
Developing POE at the Simplest Level of RSM: A One-Factor Experiment	198	(3)
An Application of POE for Three Factors Studied via a Box-Behnken Design	201	(5)
Practice Problem 9-1	206	(1)
Practice Problem 9-2	207	(1)
Appendix: Details on How to Calculate POE	207	(8)
POE for Transformed Responses	210	(5)
Other Applications for RSM	215	(20)
Adding Categoric Factors to RSM Designs	215	(7)
Response Surface Methods for Experimenting on Computer Simulations	222	(7)
Practice Problem 10-1	229	(2)
Appendix: Alternative RSM Designs for Experiments on Simulations	231	(4)
Applying RSM to Mixtures	235	(20)
Accounting for Proportional Effects by Creating Ratios of Ingredients	236	(10)
A Different Approach for Optimizing Formulations: Mixture Design	246	(3)
Practice Problem 11-1	249	(3)
Practice Problem 11-2	252	(3)
Glossary of Symbols and Terms	255	(18)
Glossary of Symbols	255	(2)
Glossary of Terms	257	(16)
References	273	(6)
About the Authors	279	(2)
Index	281	(8)
About the Software	289