Full Scale Plant Optimization in Chemical Engineering: A Practical Guide

Full Scale Plant Optimization in Chemical Engineering Highlights the basic principles and applications of the primary three methods in plant and process optimization for responsible operators and engineers. Chemical engineers are a vital part of the creation of any process development—lab-scale and pilot-scale—for any plant. In fact, they are the lynchpin of later efforts to scale-up and full-scale plant process improvement. As these engineers approach a new project, there are three generally recognized methodologies that are applicable in industry generally: Design of Experiments (DOE), Evolutionary Operations (EVOP), and Data Mining Using Neural Networks (DM). In Full Scale Plant Optimization in Chemical Engineering, experienced chemical engineer Živorad R. Laziċ offers an in-depth analysis and comparison of these three methods in full-scale plant optimization applications. The book is designed to provide the basic principles and necessary information for complete understanding of these three methods (DOE, EVOP, and DM). The application of each method is fully described. Full Scale Plant Optimization in Chemical Engineering readers will also find: A thorough discussion of the advantages, disadvantages and applications for the five different EVOP methods (BEVOP, ROVOP, REVOP, QSEVOP & SEVOP) with examples and simulations An overview of EVOP tools that responsible operators and engineers utilize in deciding which EVOP method is the most appropriate for the certain type of the process Particular attention is given to the simple but powerful technique Evolutionary Operation or EVOP, which provides the experimental tools for the full scale plant optimization Full Scale Plant Optimization in Chemical Engineering is a useful reference for all chemists in industry, chemical engineers, pharmaceutical chemists, and process engineers.

Table of Contents:
Preface ix Biography xii 1 The Basic Ideas 1 1.1 Introduction 1 2 Design of Experiments – DOE 3 2.1 The 2 2 Factorial Designs 4 2.2 Effects for 2 2 Factorial Designs 6 2.3 Interactions Between Factors 6 2.4 Standard Error for the Effects 7 2.5 2 3 Factorial Design 7 2.6 Effects for the 2 3 Factorial Designs 9 2.7 Standard Errors of Effects for Two- and Three-Level Factorial Designs 10 3 Neural Network Modeling – Data Mining 17 3.1 Data Preprocessing 18 3.2 Building, Training, and Verifying Model 19 3.3 Model Analyzing 21 3.4 What-Ifs Optimization 25 3.5 DOE Experiment Using Neural Networks Model 26 4 Evolutionary Operation – EVOP 29 4.1 Small-Scale and Plant-Scale Investigation 29 4.2 Scale-up 29 4.3 Static and Evolutionary Operation 30 4.4 Analysis of Information Board 34 4.5 Three-Factor Scheme 35 4.6 Current Best-Known Conditions 36 4.7 Change in Mean for a 2 2 Factorial Design with Center Point 38 4.8 Standard Errors for the Effects 38 4.9 The Effects and Their Standard Errors for a 2 2 Design with Center Point 40 4.10 Analysis of Information Board for Three Responses Using Factorial Effects 40 4.11 2 3 Factorial Design Effects, Interpretation, and Information Board 41 4.11.1 An Estimate of Standard Deviation 43 4.12 Dividing the 2 3 Factorial Design Into Two Blocks 45 4.13 2 3 Design with Two Center Points Run in Two Blocks 45 4.13.1 Two Standard Error Limits for the Overall Change in Mean 46 5 Different Techniques of EVOP 49 5.1 Box EVOP – BEVOP 49 5.2 Calculation Procedure for Two-Factor EVOP 50 5.3 Calculation Procedure for Three-Factor EVOP 54 5.4 BEVOP in Plant-Scale Experiments 93 5.5 BEVOP Applications 96 5.6 BEVOP Advantages and Disadvantages 99 5.7 BEVOP Simulation 100 5.7.1 2 2 BEVOP Simulation 100 5.7.1.1 Simulation No. 1 100 5.7.1.2 Simulation No. 2: 2 2 BEVOP 117 5.7.1.3 Simulation No. 3: 2 2 BEVOP 127 5.7.2 2 3 BEVOP Simulation 134 5.7.2.1 Simulation No. 4 134 5.8 Rotating Square Evolutionary Operation – ROVOP 155 5.8.1 2 2 Rovop 155 5.8.2 Method of Analysis 157 5.8.3 2 2 ROVOP Simulation 158 5.8.3.1 Simulation No. 5 158 5.8.4 2 3 ROVOP Simulation 185 5.8.4.1 Simulation No. 6: 2 3 ROVOP 186 5.9 Random Evolutionary Operation – REVOP 198 5.9.1 REVOP Simulation 200 5.9.1.1 Simulation No. 7 200 5.10 Quick-Start EVOP – QSEVOP 204 5.10.1 The way QSEVOP works 204 5.10.2 How to Recover From “Hang-ups” 207 5.11 QSEVOP Simulation 208 5.11.1 Simulation No. 8 208 5.12 Simplex Evolutionary Operation – SEVOP 214 5.12.1 The Basic Simplex Method 214 5.12.2 Simplex Evolutionary Operation – SEVOP 223 5.12.3 SEVOP Simulation 228 5.12.3.1 Simulation S-9 228 5.12.3.2 Simulation S-10 231 5.13 Some Practical Advice About Using EVOP 233 6 EVOP Software 235 Appendix A The Approximate Method of Estimating the Standard Deviation in EVOP 237 Appendix B 2 2 -and2 3 -Factor Box EVOP Calculations with Center Point 239 Appendix C Short Table of Random Normal Deviates 243 Appendix d How Many Cycles Are Necessary to Detect Effects of Reasonable Size 245 Appendix E Multiple Responses: The Desirability Approach 247 References 253 Index 257

About the Author :
Živorad R. Laziċ is the author of “Design of Experiments in Chemical Engineering: A Practical Guide”, published by J. Wiley in January 2004. He has produced a unique, “how to do it”, a practical guide for the statistical design of experiments. It is the ideal book for the industrial scientist or engineer who wants to take an advantage of DOE techniques without becoming a statistician. Basic statistical ideas are presented clearly and simply with numerous examples. This is one of the few books that are practically suited for self-study by a busy technologist, engineers and scientists. He is a Certified Six-Sigma Black Belt professional with interests in advanced statistical tools, Design of Experiments(DOE), Statistical Process Control(SPC), Evolutionary Operation(EVOP) and process modeling via application of neural networks.