Details

Process Control System Fault Diagnosis


Process Control System Fault Diagnosis

A Bayesian Approach
Wiley Series in Dynamics and Control of Electromechanical Systems 1. Aufl.

von: Ruben Gonzalez, Fei Qi, Biao Huang

107,99 €

Verlag: Wiley
Format: PDF
Veröffentl.: 21.07.2016
ISBN/EAN: 9781118770580
Sprache: englisch
Anzahl Seiten: 360

DRM-geschütztes eBook, Sie benötigen z.B. Adobe Digital Editions und eine Adobe ID zum Lesen.

Beschreibungen

<p><b>Process Control System Fault Diagnosis: A Bayesian Approach</b></p> <p>Ruben T. Gonzalez, University of Alberta, Canada</p> <p>Fei Qi, Suncor Energy Inc., Canada</p> <p>Biao Huang, University of Alberta, Canada</p> <p> </p> <p><b><i>Data-driven Inferential Solutions for Control System Fault Diagnosis</i></b></p> <p><b><i> </i></b></p> <p>A typical modern process system consists of hundreds or even thousands of control loops, which are overwhelming for plant personnel to monitor. The main objectives of this book are to establish a new framework for control system fault diagnosis, to synthesize observations of different monitors with a prior knowledge, and to pinpoint possible abnormal sources on the basis of Bayesian theory.</p> <p><i>Process Control System Fault Diagnosis: A Bayesian Approach</i> consolidates results developed by the authors, along with the fundamentals, and presents them in a systematic way. The book provides a comprehensive coverage of various Bayesian methods for control system fault diagnosis, along with a detailed tutorial. The book is useful for graduate students and researchers as a monograph and as a reference for state-of-the-art techniques in control system performance monitoring and fault diagnosis. Since several self-contained practical examples are included in the book, it also provides a place for practicing engineers to look for solutions to their daily monitoring and diagnosis problems.</p> <p> </p> <p>Key features:</p> <p>•             A comprehensive coverage of Bayesian Inference for control system fault diagnosis.</p> <p>•             Theory and applications are self-contained.</p> <p>•             Provides detailed algorithms and sample Matlab codes.</p> <p>•             Theory is illustrated through benchmark simulation examples, pilot-scale experiments and industrial application.</p> <p> </p> <p><i>Process Control System Fault Diagnosis: A Bayesian Approach </i>is a comprehensive guide for graduate students, practicing engineers, and researchers who are interests in applying theory to practice.</p>
<p>Preface xiii</p> <p>Acknowledgements xvii</p> <p>List of Figures xix</p> <p>List of Tables xxiii</p> <p>Nomenclature xxv</p> <p><b>Part I FUNDAMENTALS</b></p> <p><b>1 Introduction 3</b></p> <p>1.1 Motivational Illustrations 3</p> <p>1.2 Previous Work 4</p> <p>1.2.1 Diagnosis Techniques 4</p> <p>1.2.2 Monitoring Techniques 7</p> <p>1.3 Book Outline 12</p> <p>1.3.1 Problem Overview and Illustrative Example 12</p> <p>1.3.2 Overview of Proposed Work 12</p> <p>References 16</p> <p><b>2 Prerequisite Fundamentals 19</b></p> <p>2.1 Introduction 19</p> <p>2.2 Bayesian Inference and Parameter Estimation 19</p> <p>2.2.1 Tutorial on Bayesian Inference 24</p> <p>2.2.2 Tutorial on Bayesian Inference with Time Dependency 27</p> <p>2.2.3 Bayesian Inference vs. Direct Inference 32</p> <p>2.2.4 Tutorial on Bayesian Parameter Estimation 33</p> <p>2.3 The EM Algorithm 38</p> <p>2.4 Techniques for Ambiguous Modes 44</p> <p>2.4.1 Tutorial on Θ Parameters in the Presence of Ambiguous Modes 46</p> <p>2.4.2 Tutorial on Probabilities Using Θ Parameters 47</p> <p>2.4.3 Dempster–Shafer Theory 48</p> <p>2.5 Kernel Density Estimation 51</p> <p>2.5.1 From Histograms to Kernel Density Estimates 52</p> <p>2.5.2 Bandwidth Selection 54</p> <p>2.5.3 Kernel Density Estimation Tutorial 55</p> <p>2.6 Bootstrapping 56</p> <p>2.6.1 Bootstrapping Tutorial 57</p> <p>2.6.2 Smoothed Bootstrapping Tutorial 57</p> <p>2.7 Notes and References 60</p> <p>References 61</p> <p><b>3 Bayesian Diagnosis 62</b></p> <p>3.1 Introduction 62</p> <p>3.2 Bayesian Approach for Control Loop Diagnosis 62</p> <p>3.2.1 Mode M 62</p> <p>3.2.2 Evidence E 63</p> <p>3.2.3 Historical Dataset D 64</p> <p>3.3 Likelihood Estimation 65</p> <p>3.4 Notes and References 67</p> <p>References 67</p> <p><b>4 Accounting for Autodependent Modes and Evidence 68</b></p> <p>4.1 Introduction 68</p> <p>4.2 Temporally Dependent Evidence 68</p> <p>4.2.1 Evidence Dependence 68</p> <p>4.2.2 Estimation of Evidence-transition Probability 70</p> <p>4.2.3 Issues in Estimating Dependence in Evidence 74</p> <p>4.3 Temporally Dependent Modes 75</p> <p>4.3.1 Mode Dependence 75</p> <p>4.3.2 Estimating Mode Transition Probabilities 77</p> <p>4.4 Dependent Modes and Evidence 81</p> <p>4.5 Notes and References 82</p> <p>References 82</p> <p><b>5 Accounting for Incomplete Discrete Evidence 83</b></p> <p>5.1 Introduction 83</p> <p>5.2 The Incomplete Evidence Problem 83</p> <p>5.3 Diagnosis with Incomplete Evidence 85</p> <p>5.3.1 Single Missing Pattern Problem 86</p> <p>5.3.2 Multiple Missing Pattern Problem 92</p> <p>5.3.3 Limitations of the Single and Multiple Missing Pattern Solutions 93</p> <p>5.4 Notes and References 94</p> <p>References 94</p> <p><b>6 Accounting for Ambiguous Modes: A Bayesian Approach 96</b></p> <p>6.1 Introduction 96</p> <p>6.2 Parametrization of Likelihood Given Ambiguous Modes 96</p> <p>6.2.1 Interpretation of Proportion Parameters 96</p> <p>6.2.2 Parametrizing Likelihoods 97</p> <p>6.2.3 Informed Estimates of Likelihoods 98</p> <p>6.3 Fagin–Halpern Combination 99</p> <p>6.4 Second-order Approximation 100</p> <p>6.4.1 Consistency of Θ Parameters 101</p> <p>6.4.2 Obtaining a Second-order Approximation 101</p> <p>6.4.3 The Second-order Bayesian Combination Rule 103</p> <p>6.5 Brief Comparison of Combination Methods 104</p> <p>6.6 Applying the Second-order Rule Dynamically 105</p> <p>6.6.1 Unambiguous Dynamic Solution 105</p> <p>6.6.2 The Second-order Dynamic Solution 106</p> <p>6.7 Making a Diagnosis 107</p> <p>6.7.1 Simple Diagnosis 107</p> <p>6.7.2 Ranged Diagnosis 107</p> <p>6.7.3 Expected Value Diagnosis 107</p> <p>6.8 Notes and References 111</p> <p>References 111</p> <p><b>7 Accounting for Ambiguous Modes: A Dempster–Shafer Approach 112</b></p> <p>7.1 Introduction 112</p> <p>7.2 Dempster–Shafer Theory 112</p> <p>7.2.1 Basic Belief Assignments 112</p> <p>7.2.2 Probability Boundaries 114</p> <p>7.2.3 Dempster’s Rule of Combination 114</p> <p>7.2.4 Short-cut Combination for Unambiguous Priors 115</p> <p>7.3 Generalizing Dempster–Shafer Theory 116</p> <p>7.3.1 Motivation: Difficulties with BBAs 117</p> <p>7.3.2 Generalizing the BBA 119</p> <p>7.3.3 Generalizing Dempster’s Rule 122</p> <p>7.3.4 Short-cut Combination for Unambiguous Priors 123</p> <p>7.4 Notes and References 124</p> <p>References 125</p> <p><b>8 Making Use of Continuous Evidence Through Kernel Density Estimation 126</b></p> <p>8.1 Introduction 126</p> <p>8.2 Performance: Continuous vs. Discrete Methods 127</p> <p>8.2.1 Average False Negative Diagnosis Criterion 127</p> <p>8.2.2 Performance of Discrete and Continuous Methods 129</p> <p>8.3 Kernel Density Estimation 132</p> <p>8.3.1 From Histograms to Kernel Density Estimates 132</p> <p>8.3.2 Defining a Kernel Density Estimate 134</p> <p>8.3.3 Bandwidth Selection Criterion 135</p> <p>8.3.4 Bandwidth Selection Techniques 136</p> <p>8.4 Dimension Reduction 137</p> <p>8.4.1 Independence Assumptions 138</p> <p>8.4.2 Principal and Independent Component Analysis 139</p> <p>8.5 Missing Values 139</p> <p>8.5.1 Kernel Density Regression 140</p> <p>8.5.2 Applying Kernel Density Regression for a Solution 141</p> <p>8.6 Dynamic Evidence 142</p> <p>8.7 Notes and References 143</p> <p>References 143</p> <p><b>9 Accounting for Sparse Data Within a Mode 144</b></p> <p>9.1 Introduction 144</p> <p>9.2 Analytical Estimation of the Monitor Output Distribution Function 145</p> <p>9.2.1 Control Performance Monitor 145</p> <p>9.2.2 Process Model Monitor 146</p> <p>9.2.3 Sensor Bias Monitor 148</p> <p>9.3 Bootstrap Approach to Estimating Monitor Output Distribution Function 150</p> <p>9.3.1 Valve Stiction Identification 150</p> <p>9.3.2 The Bootstrap Method 153</p> <p>9.3.3 Illustrative Example 156</p> <p>9.3.4 Applications 160</p> <p>9.4 Experimental Example 164</p> <p>9.4.1 Process Description 164</p> <p>9.4.2 Diagnostic Settings and Results 167</p> <p>9.5 Notes and References 170</p> <p>References 170</p> <p><b>10 Accounting for Sparse Modes Within the Data 172</b></p> <p>10.1 Introduction 172</p> <p>10.2 Approaches and Algorithms 172</p> <p>10.2.1 Approach for Component Diagnosis 173</p> <p>10.2.2 Approach for Bootstrapping New Modes 176</p> <p>10.3 Illustration 181</p> <p>10.3.1 Component-based Diagnosis 184</p> <p>10.3.2 Bootstrapping for Additional Modes 188</p> <p>10.4 Application 194</p> <p>10.4.1 Monitor Selection 195</p> <p>10.4.2 Component Diagnosis 195</p> <p>10.5 Notes and References 198</p> <p>References 199</p> <p><b>Part II APPLICATIONS</b></p> <p><b>11 Introduction to Testbed Systems 203</b></p> <p>11.1 Simulated System 203</p> <p>11.1.1 Monitor Design 203</p> <p>11.2 Bench-scale System 205</p> <p>11.3 Industrial Scale System 207</p> <p>References 207</p> <p><b>12 Bayesian Diagnosis with Discrete Data 209</b></p> <p>12.1 Introduction 209</p> <p>12.2 Algorithm 210</p> <p>12.3 Tutorial 213</p> <p>12.4 Simulated Case 216</p> <p>12.5 Bench-scale Case 217</p> <p>12.6 Industrial-scale Case 219</p> <p>12.7 Notes and References 220</p> <p>References 220</p> <p><b>13 Accounting for Autodependent Modes and Evidence 221</b></p> <p>13.1 Introduction 221</p> <p>13.2 Algorithms 222</p> <p>13.2.1 Evidence Transition Probability 222</p> <p>13.2.2 Mode Transition Probability 226</p> <p>13.3 Tutorial 228</p> <p>13.4 Notes and References 231</p> <p>References 231</p> <p><b>14 Accounting for Incomplete Discrete Evidence 232</b></p> <p>14.1 Introduction 232</p> <p>14.2 Algorithm 232</p> <p>14.2.1 Single Missing Pattern Problem 232</p> <p>14.2.2 Multiple Missing Pattern Problem 236</p> <p>14.3 Tutorial 238</p> <p>14.4 Simulated Case 241</p> <p>14.5 Bench-scale Case 242</p> <p>14.6 Industrial-scale Case 244</p> <p>14.7 Notes and References 246</p> <p>References 246</p> <p><b>15 Accounting for Ambiguous Modes in Historical Data: A Bayesian Approach 247</b></p> <p>15.1 Introduction 247</p> <p>15.2 Algorithm 248</p> <p>15.2.1 Formulating the Problem 248</p> <p>15.2.2 Second-order Taylor Series Approximation of p(E|M,Θ) 248</p> <p>15.2.3 Second-order Bayesian Combination 250</p> <p>15.2.4 Optional Step: Separating Monitors into Independent Groups 252</p> <p>15.2.5 Grouping Methodology 253</p> <p>15.3 Illustrative Example of Proposed Methodology 254</p> <p>15.3.1 Introduction 254</p> <p>15.3.2 Offline Step 1: Historical Data Collection 255</p> <p>15.3.3 Offline Step 2: Mutual Information Criterion (Optional) 255</p> <p>15.3.4 Offline Step 3: Calculate Reference Values 256</p> <p>15.3.5 Online Step 1: Calculate Support 257</p> <p>15.3.6 Online Step 2: Calculate Second-order Terms 258</p> <p>15.3.7 Online Step 3: Perform Combinations 260</p> <p>15.3.8 Online Step 4: Make a Diagnosis 262</p> <p>15.4 Simulated Case 265</p> <p>15.5 Bench-scale Case 268</p> <p>15.6 Industrial-scale Case 269</p> <p>15.7 Notes and References 270</p> <p>References 271</p> <p><b>16 Accounting for Ambiguous Modes in Historical Data: A Dempster–Shafer Approach 272</b></p> <p>16.1 Introduction 272</p> <p>16.2 Algorithm 272</p> <p>16.2.1 Parametrized Likelihoods 272</p> <p>16.2.2 Basic Belief Assignments 273</p> <p>16.2.3 The Generalized Dempster’s Rule of Combination 275</p> <p>16.3 Example of Proposed Methodology 276</p> <p>16.3.1 Introduction 276</p> <p>16.3.2 Offline Step 1: Historical Data Collection 277</p> <p>16.3.3 Offline Step 2: Mutual Information Criterion (Optional) 277</p> <p>16.3.4 Offline Step 3: Calculate Reference Value 278</p> <p>16.3.5 Online Step 1: Calculate Support 279</p> <p>16.3.6 Online Step 2: Calculate the GBBA 280</p> <p>16.3.7 Online Step 3: Combine BBAs and Diagnose 283</p> <p>16.4 Simulated Case 283</p> <p>16.5 Bench-scale Case 284</p> <p>16.6 Industrial System 286</p> <p>16.7 Notes and References 287</p> <p>References 287</p> <p><b>17 Making use of Continuous Evidence through Kernel Density Estimation 288</b></p> <p>17.1 Introduction 288</p> <p>17.2 Algorithm 289</p> <p>17.2.1 Kernel Density Estimation 289</p> <p>17.2.2 Bandwidth Selection 289</p> <p>17.2.3 Adaptive Bandwidths 290</p> <p>17.2.4 Optional Step: Dimension Reduction by Multiplying Independent Likelihoods 291</p> <p>17.2.5 Optional Step: Creating Independence via Independent Component Analysis 291</p> <p>17.2.6 Optional Step: Replacing Missing Values 292</p> <p>17.3 Example of Proposed Methodology 293</p> <p>17.3.1 Offline Step 1: Historical Data Collection 295</p> <p>17.3.2 Offline Step 3: Mutual Information Criterion (Optional) 296</p> <p>17.3.3 Offline Step 4: Independent Component Analysis (Optional) 298</p> <p>17.3.4 Offline Step 5: Obtain Bandwidths 298</p> <p>17.3.5 Online Step 1: Calculate Likelihood of New Data 301</p> <p>17.3.6 Online Step 2: Calculate Posterior Probability 302</p> <p>17.3.7 Online Step 3: Make a Diagnosis 302</p> <p>17.4 Simulated Case 302</p> <p>17.5 Bench-scale Case 304</p> <p>17.6 Industrial-scale Case 304</p> <p>17.7 Notes and References 307</p> <p>References 307</p> <p>Appendix 308</p> <p>17.A Code for Kernel Density Regression 308</p> <p>17.A.1 Kernel Density Regression 308</p> <p>17.A.2 Three-dimensional Matrix Toolbox 310</p> <p><b>18 Dynamic Application of Continuous Evidence and Ambiguous Mode Solutions 313</b></p> <p>18.1 Introduction 313</p> <p>18.2 Algorithm for Autodependent Modes 313</p> <p>18.2.1 Transition Probability Matrix 314</p> <p>18.2.2 Review of Second-order Method 314</p> <p>18.2.3 Second-order Probability Transition Rule 315</p> <p>18.3 Algorithm for Dynamic Continuous Evidence and Autodependent Modes 316</p> <p>18.3.1 Algorithm for Dynamic Continuous Evidence 316</p> <p>18.3.2 Combining both Solutions 318</p> <p>18.3.3 Comments on Usefulness 319</p> <p>18.4 Example of Proposed Methodology 320</p> <p>18.4.1 Introduction 320</p> <p>18.4.2 Offline Step 1: Historical Data Collection 320</p> <p>18.4.3 Offline Step 2: Create Temporal Data 320</p> <p>18.4.4 Offline Step 3: Mutual Information Criterion (Optional, but Recommended) 321</p> <p>18.4.5 Offline Step 5: Calculate Reference Values 322</p> <p>18.4.6 Online Step 1: Obtain Prior Second-order Terms 322</p> <p>18.4.7 Online Step 2: Calculate Support 323</p> <p>18.4.8 Online Step 3: Calculate Second-order Terms 323</p> <p>18.4.9 Online Step 4: Combining Prior and Likelihood Terms 324</p> <p>18.5 Simulated Case 325</p> <p>18.6 Bench-scale Case 326</p> <p>18.7 Industrial-scale Case 326</p> <p>18.8 Notes and References 327</p> <p>References 327</p> <p>Index 329</p>
<p>Ruben Gonzalez completed his Bachelor's degree in chemical engineering in 2008 at the University of New Brunswick. Under the supervision of Dr. Biao Huang, he completed his Master's degree in 2010 and his Doctorate in 2014, both in chemical engineering, at the University of Alberta. His research interests include Bayesian diagnosis, fault detection and diagnosis, data reconciliation, and applied kernel density estimation.<br /><br />Fei Qi obtained his Ph.D. degree in Process Control from the University of Alberta, Canada, in 2011. He had his M.Sc. degree (2006) and B.Sc. degree (2003) in Automation from the University of Science and Technology of China. Fei Qi joined Suncor Energy Inc. in 2010 as an Advance Process Control Engineer. He has extensive experiences in applying system identification, model predictive control, and control performance monitoring in real industrial processes. His Ph.D. research was on applying Bayesian statistics to control loop diagnosis. His current research interests include model predictive control, soft sensor, fault detection, and process optimization.<br /><br />Biao Huang obtained his PhD degree in Process Control from the University of Alberta, Canada, in 1997. He is currently a Professor in the Department of Chemical and Materials Engineering, University of Alberta, NSERC Industrial Research Chair in Control of Oil Sands Processes and AITF Industry Chair in Process Control. He is a Fellow of the Canadian Academy of Engineering, Fellow of the Chemical Institute of Canada, and recipient of numerous awards including Germany’s Alexander von Humboldt Research Fellowship, Bantrel Award in Design and Industrial Practice, APEGA Summit Award in Research Excellence, best paper award from Journal of Process Control etc. Biao Huang’s main research interests include: Bayesian inference, control performance assessment, fault detection and isolation. Biao Huang has applied his expertise extensively in industrial practice. He also serves as the Deputy Editor-in-Chief for Control Engineering Practice, the Associate Editor for Canadian Journal of Chemical Engineering and the Associate Editor for Journal of Process Control.</p>
<p><b>Process Control System Fault Diagnosis: A Bayesian Approach</b></p> <p>Ruben T. Gonzalez, University of Alberta, Canada</p> <p>Fei Qi, Suncor Energy Inc., Canada</p> <p>Biao Huang, University of Alberta, Canada</p> <p> </p> <p><b><i>Data-driven Inferential Solutions for Control System Fault Diagnosis</i></b></p> <p><b><i> </i></b></p> <p>A typical modern process system consists of hundreds or even thousands of control loops, which are overwhelming for plant personnel to monitor. The main objectives of this book are to establish a new framework for control system fault diagnosis, to synthesize observations of different monitors with a prior knowledge, and to pinpoint possible abnormal sources on the basis of Bayesian theory.</p> <p><i>Process Control System Fault Diagnosis: A Bayesian Approach</i> consolidates results developed by the authors, along with the fundamentals, and presents them in a systematic way. The book provides a comprehensive coverage of various Bayesian methods for control system fault diagnosis, along with a detailed tutorial. The book is useful for graduate students and researchers as a monograph and as a reference for state-of-the-art techniques in control system performance monitoring and fault diagnosis. Since several self-contained practical examples are included in the book, it also provides a place for practicing engineers to look for solutions to their daily monitoring and diagnosis problems.</p> <p> </p> <p>Key features:</p> <p>•             A comprehensive coverage of Bayesian Inference for control system fault diagnosis.</p> <p>•             Theory and applications are self-contained.</p> <p>•             Provides detailed algorithms and sample Matlab codes.</p> <p>•             Theory is illustrated through benchmark simulation examples, pilot-scale experiments and industrial application.</p> <p> </p> <p><i>Process Control System Fault Diagnosis: A Bayesian Approach </i>is a comprehensive guide for graduate students, practicing engineers, and researchers who are interests in applying theory to practice.</p>

Diese Produkte könnten Sie auch interessieren:

Turbulent Drag Reduction by Surfactant Additives
Turbulent Drag Reduction by Surfactant Additives
von: Feng-Chen Li, Bo Yu, Jin-Jia Wei, Yasuo Kawaguchi
PDF ebook
156,99 €
Turbulent Drag Reduction by Surfactant Additives
Turbulent Drag Reduction by Surfactant Additives
von: Feng-Chen Li, Bo Yu, Jin-Jia Wei, Yasuo Kawaguchi
EPUB ebook
156,99 €
Wear
Wear
von: Gwidon W. Stachowiak
PDF ebook
159,99 €