Fault Detection and Diagnosis in Industrial Systems

Couverture
Springer Science & Business Media, 11 déc. 2000 - 279 pages

Early and accurate fault detection and diagnosis for modern manufacturing processes can minimise downtime, increase the safety of plant operations, and reduce costs. Such process monitoring techniques are regularly applied to real industrial systems. Fault Detection and Diagnosis in Industrial Systems presents the theoretical background and practical methods for process monitoring. The coverage of data-driven, analytical and knowledge-based techniques include:

• principal component analysis

• Fisher discriminant analysis

• partial least squares

• canonical variate analysis;

• parameter estimation;

• observer/state estimators

• parity relations;

• artificial neural networks;

• expert systems.

Application of the process monitoring techniques to a number of processes, including to a manufacturing plant, demonstrates the strenghts and weaknesses of each approach in detail. This aids the reader in selecting the right method for a particular application. A plant simulator and homework problems are included in which students apply the process monitoring techniques to a nontrivial simulated process, and can compare their performance with that obtained in the case studies in the text. A number of additional homework problems encourage the reader to implement and obtain a deeper understanding of the techniques.

 

Pages sélectionnées

Table des matières

Introduction
3
11 Process Monitoring Procedures
4
12 Process Monitoring Measures
5
13 Process Monitoring Methods
7
14 Book Organization
10
Background
13
Multivariate Statistics
15
22 Data Pretreatment
16
85 Simulation Program
107
86 Control Structure
109
9 Application Description
113
93 Sampling Interval
114
94 Sample Size
115
95 Lag and Order Selection
117
96 Fault Detection
118
97 Fault Identification
119

23 Univariate Statistical Monitoring
17
24 T Statistic
21
25 Thresholds for the T Statistic
22
26 Data Requirements
24
27 Homework Problems
25
Pattern Classification
27
32 Discriminant Analysis
28
33 Feature Extraction
30
34 Homework Problems
31
Datadriven Methods
33
Principal Component Analysis
35
42 Principal Component Analysis
36
43 Reduction Order
41
44 Fault Detection
42
45 Fault Identification
45
46 Fault Diagnosis
48
47 Dynamic PCA
52
48 Other PCAbased Methods
54
49 Homework Problems
55
5 Fisher Discriminant Analysis
57
53 Reduction Order
60
54 Fault Detection and Diagnosis
62
55 Comparison of PCA and FDA
63
56 Dynamic FDA
69
57 Homework Problems
70
Partial Least Squares
71
62 PLS Algorithms
72
63 Reduction Order and PLS Prediction
77
64 Fault Detection Identification and Diagnosis
78
65 Comparison of PCA and PLS
79
66 Other PLS Methods
81
67 Homework Problems
83
Canonical Variate Analysis
85
72 CVA Theorem
87
73 CVA Algorithm
89
74 State Space Model and System Identifiability
91
75 Lag Order Selection and Computation
92
76 State Order Selection and Akaikes Information Criterion
94
77 Subspace Algorithm Interpretations
95
78 Process Monitoring Statistics
97
79 Homework Problems
98
Application
101
Tennessee Eastman Process
103
82 Process Flowsheet
104
Results and Discussion
121
103 Case Study on Fault 4
124
104 Case Study on Fault 5
129
105 Case Study on Fault 11
131
106 Fault Detection
133
107 Fault Identification
142
108 Fault Diagnosis
146
109 Homework Problems
166
Analytical and Knowledgebased Methods
171
Analytical Methods
173
112 Fault Descriptions
175
113 Parameter Estimation
179
114 Observerbased Method
190
1141 Fullorder State Estimator
191
1142 Reducedorder Unknown Input Observer
195
115 Parity Relations
197
1152 Detection Properties of the Residual
201
1153 Specification of the Residuals
203
1154 Implementation of the Residuals
204
1155 Connection Between the Observer and Parity Relations
208
1156 Isolation Properties of the Residual
211
1157 Residual Evaluation
214
116 Homework Problems
218
12 Knowledgebased Methods
223
122 Causal Analysis
224
1222 Symptom Tree Model
227
123 Expert Systems
228
1231 ShallowKnowledge Expert System
229
1233 Combination of ShallowKnowledge and DeepKnowledge Expert Systems
230
1235 Knowledge Representation
231
1236 Inference Engine
232
1241 Artificial Neural Networks
233
1242 SelfOrganizing Map
239
125 Combinations of Various Techniques
242
1252 Fuzzy Logic
243
1253 Fuzzy Expert Systems
245
1254 Fuzzy Neural Networks
248
1255 Fuzzy Signed Directed Graph
249
1256 Fuzzy Logic and the Analytical Approach
250
1257 Neural Networks and the Analytical Approach
251
126 Homework Problems
252
References
255
Index
275
Droits d'auteur

Autres éditions - Tout afficher

Fault Detection and Diagnosis in Industrial Systems
L.H. Chiang,E.L. Russell,R.D. Braatz
Aperçu limité - 2012
Fault Detection and Diagnosis in Industrial Systems
L. H. Chiang,E L Russell,R. D. Braatz
Aucun aperçu disponible - 2001

Expressions et termes fréquents

analytical applied approach Chapter components computed covariance matrix CVA-based data-driven described DFDA/DPCA1 diagnosing faults dimensionality reduction discriminant analysis discriminant function discriminant PLS DPCA DPCAM eigenvectors equations error expert systems fault class fault detection fault diagnosis fault identification fault occurs FDA vectors FDA/PCA1 Figure Fisher's data flow rate fuzzy logic fuzzy sets Homework Problems information criterion knowledge-based leak in Stream linear loading vectors measured input measured output missed detection rates model parameters multivariate statistics neural networks neurons nodes noise normal operating conditions number of observations observation variables overall misclassification rates parameter estimation parity relations partial least squares pattern classification PCA1 PCAM PLS1 and PLS2 PLS2adj process monitoring methods process variables Q statistic Reactor Cooling Water reduction order sampling interval self-organizing maps Shewhart sigmoid function simulation singular values state-space model T2 statistic Table testing set threshold training set variance variations zero

Références à ce livre

Informations bibliographiques

TitreFault Detection and Diagnosis in Industrial Systems
Advanced Textbooks in Control and Signal Processing
AuteursL.H. Chiang, E.L. Russell, R.D. Braatz
Éditionillustrée
ÉditeurSpringer Science & Business Media, 2000
ISBN1852333278, 9781852333270
Longueur279 pages
  
Exporter la citationBiBTeX EndNote RefMan