Chemical Process Performance Evaluation Book

Nomenclature
Introduction     1
Motivation and Historical Perspective     2
Outline     4
Univariate Statistical Monitoring Techniques     7
Statistics Concepts     8
Univariate SPM Techniques     11
Shewhart Control Charts     11
Cumulative Sum (CUSUM) Charts     18
Moving Average Monitoring Charts for Individual Measurements     19
Exponentially Weighted Moving Average Chart     22
Monitoring Tools for Autocorreleated Data     22
Monitoring with Charts of Residuals     26
Monitoring with Detecting Changes in Model Parameters     27
Limitations of Univariate Monitoring Techniques     32
Summary     35
Multivariate Statistical Monitoring Techniques     37
Principal Components Analysis     37
Canonical Variates Analysis     43
Independent Component Analysis     43
Contribution Plots     46
Linear Methods for Diagnosis     48
Clustering     48
Discriminant Analysis     50
Fisher's Discriminant Analysis     53
Nonlinear Methods for Diagnosis     58
Neural Networks     58
Kernel-Based Techniques     64
Support Vector Machines     66
Summary     69
Empirical Model Development     73
Regression Models     75
PCA Models     78
PLS Regression Models     79
Input-Output Models of Dynamic Processes     83
State-Space Models     89
Summary     97
Monitoring of Multivariate Processes     99
SPM Methods Based on PCA     100
SPM Methods Based on PLS     105
SPM Using Dynamic Process Models     108
Other MSPM Techniques     112
Summary     114
Characterization of Process Signals     115
Wavelets     115
Fourier Transform     116
Continuous Wavelet Transform     119
Discrete Wavelet Transform     123
Filtering and Outlier Detection     127
Simple Filters     128
Wavelet Filters     131
Robust Filter     133
Signal Representation by Fuzzy Triangular Episodes     135
Development of Markovian Models     138
Markov Chains     139
Hidden Markov Models     141
Wavelet-Domain Hidden Markov Models     145
Summary     147
Process Fault Diagnosis     149
Fault Diagnosis Using Triangular Episodes and HMMs     149
CSTR Simulation     152
Vacuum Column     155
Fault Diagnosis Using Wavelet-Domain HMMs     157
pH Neutralization Simulation     161
CSTR Simulation     164
Fault Diagnosis Using HMMs     166
Case Study of HTST Pasteurization Process     167
Fault Diagnosis Using Contribution Plots     174
Fault Diagnosis with Statistical Methods     179
Fault Diagnosis Using SVM     191
Fault Diagnosis with Robust Techniques     192
Robust Monitoring Strategy     192
Pilot-Scale Distillation Column     198
Summary     202
Sensor Failure Detection and Diagnosis     203
Sensor FDD Using PLS and CVSS Models     204
Real-Time Sensor FDD Using PCA-Based Techniques     215
Methodology     218
Case Study     224
Summary     230
Controller Performance Monitoring     231
Single-Loop Controller Performance Monitoring     233
Multivariable Controller Performance Monitoring      237
CPM for MPC     238
Summary     248
Web and Sheet Processes     251
Traditional Data Analysis     252
MD/CD Decomposition     252
Time Dependent Structure of Profile Data     256
Orthogonal Decomposition of Profile Data     257
Gram Polynomials     259
Principal Components Analysis     262
Flatness of Scanner Data     264
Controller Performance     268
MD Control Performance     269
Model-Based CD Control Performance     271
Summary     274
Bibliography     277
Index     305