Artificial Intelligence in Process Fault Diagnosis: Methods for Plant Surveillance

✍ Scribed by Fickelscherer, Richard J.

• Publisher: Wiley
• Year: 2024
• Tongue: English
• Leaves: 436
• Category: Library

✦ Synopsis

Artificial Intelligence in Process Fault Diagnosis A comprehensive guide to the future of process fault diagnosis Automation has revolutionized every aspect of industrial production, from the accumulation of raw materials to quality control inspections. Even process analysis itself has become subject to automated efficiencies, in the form of process fault analyzers, i.e., computer programs capable of analyzing process plant operations to identify faults, improve safety, and enhance productivity. Prohibitive cost and challenges of application have prevented widespread industry adoption of this technology, but recent advances in artificial intelligence promise to place these programs at the center of manufacturing process analysis. Artificial Intelligence in Process Fault Diagnosis brings together insights from data science and machine learning to deliver an effective introduction to these advances and their potential applications. Balancing theory and practice, it walks readers through the process of choosing an ideal diagnostic methodology and the creation of intelligent computer programs. The result promises to place readers at the forefront of this revolution in manufacturing. Artificial Intelligence in Process Fault Diagnosis readers will also find: Coverage of various AI-based diagnostic methodologies elaborated by leading experts Guidance for creating programs that can prevent catastrophic operating disasters, reduce downtime after emergency process shutdowns, and more Comprehensive overview of optimized best practices Artificial Intelligence in Process Fault Diagnosis is ideal for process control engineers, operating engineers working with processing industrial plants, and plant managers and operators throughout the various process industries.

✦ Table of Contents

COVER
TITLE PAGE
COPYRIGHT PAGE
DEDICATION PAGE
CONTENTS
LIST OF CONTRIBUTORS
FOREWORD
PREFACE
ACKNOWLEDGMENTS
CHAPTER 1 MOTIVATIONS FOR AUTOMATING PROCESS FAULT ANALYSIS
OVERVIEW
CHAPTER HIGHLIGHTS
1.1 INTRODUCTION
1.2 THE CHANGING ROLE OF THE PROCESS OPERATORS IN PLANT OPERATIONS
1.3 TRADITIONAL METHODS FOR PERFORMING PROCESS FAULT MANAGEMENT
1.4 LIMITATIONS OF HUMAN OPERATORS IN PERFORMING PROCESS FAULT MANAGEMENT
1.5 THE ROLE OF AUTOMATED PROCESS FAULT ANALYSIS
REFERENCES
CHAPTER 2 VARIOUS PROCESS FAULT DIAGNOSTIC METHODOLOGIES
OVERVIEW
CHAPTER HIGHLIGHTS
2.1 INTRODUCTION
2.2 VARIOUS ALTERNATIVE DIAGNOSTIC STRATEGIES OVERVIEW
2.2.1 Fault Tree Analysis
2.2.2 Alarm Analysis
2.2.3 Decision Tables
2.2.4 Sign-Directed Graphs
2.2.5 History-Based Statistical Methods
2.2.6 Diagnostic Strategies Based upon Qualitative Models
2.2.7 Diagnostic Strategies Based upon Quantitative Models
2.2.8 Artificial Neural Network Strategies
2.2.9 Artificial Immune System Strategies
2.2.10 Knowledge-Based System Strategies
2.2.11 Role of the Process Operators in Automated Fault Detection and Diagnosis
2.3 DIAGNOSTIC METHODOLOGY CHOICE CONCLUSIONS
REFERENCES
CHAPTER 2.A FAILURE MODES AND EFFECTS ANALYSIS
2.A.1 Introduction
2.A.2 FMEA Procedure
2.A.3 Conclusion
CHAPTER 3 ALARM MANAGEMENT AND FAULT DETECTION
CHAPTER HIGHLIGHTS
ABBREVIATIONS USED
3.1 INTRODUCTION
3.2 APPLICABLE DEFINITIONS AND GUIDELINES
3.3 THE ALARM MANAGEMENT LIFE CYCLE
3.3.1 Introduction
3.3.2 Life Cycle Model
3.3.3 Alarm Philosophy
3.3.4 Alarm Identification
3.3.5 Alarm Rationalization
3.3.6 Alarm Design
3.3.7 Implementation
3.3.8 Operation
3.3.9 Maintenance
3.3.10 Monitoring and Assessment
3.3.11 Management of Change
3.3.12 Audit
3.4 GENERATION OF DIAGNOSTIC INFORMATION
3.4.1 Introduction
3.4.2 As Part of the Basic Process Control System
3.4.3 As a Separate Application
3.5 PRESENTATION OF THE DIAGNOSTIC INFORMATION
3.5.1 Introduction
3.5.2 As Part of the Alarm Text
3.5.3 As a Link to the Diagnostic Application
3.5.4 As an Indication in the HMI
3.6 INFORMATION RATES
3.6.1 Introduction
3.6.2 Nuisance Alarms
REFERENCES
CHAPTER 4 OPERATOR PERFORMANCE: SIMULATION AND AUTOMATION
CHAPTER HIGHLIGHTS
4.1 BACKGROUND
4.2 AUTOMATION
4.2.1 Smart Alarming
4.2.2 Safe Park Applications
4.3 SIMULATION
4.4 RESEARCH
4.4.1 Method
4.4.2 Testing and Results
4.4.3 Operator Performance
4.4.4 Implications
4.5 AI INTEGRATION
4.5.1 Pattern Recognition
4.5.2 Training
4.6 CASE STUDY: TURBO EXPANDERS OVER-SPEED
4.7 HUMAN-CENTERED AI
4.7.1 Case Study: Boeing 737 MAX
4.7.2 AI Mental Models
REFERENCES
CHAPTER 5 AI AND ALARM ANALYTICS FOR FAILURE ANALYSIS AND PREVENTION
OVERVIEW
5.1 INTRODUCTION
5.2 POST-ALARM ASSESSMENT AND ANALYSIS
5.2.1 Alarm Configuration Database
5.3 REAL-TIME ALARM ACTIVITY DATABASE AND OPERATOR ACTION JOURNAL
5.4 PRE-ALARM ASSESSMENT AND ANALYSIS
5.5 UTILIZING ALARM ASSESSMENT INFORMATION
5.6 EXAMINING THE ALARM SYSTEM TO RESOLVE FAILURES ON A WIDER SCALE
5.6.1 Sequence of Events (SOE) Module
5.6.2 Use of First Principles to Determine Likely Root Causes
5.6.3 Use of Simple Data Analytics to Identify Redundant/Repetitive Alarms
5.6.4 Use of Data Analytics to Identify Problem Areas with Upsets Related to Transitions, out of Service, and out of Suppression States
5.6.5 Use of Data Analytics to Identify Problem Areas with Chronic Alarm Shelving
5.7 EMERGING METHODS OF ALARM ANALYSIS
5.7.1 Use of Advanced Modeling Methods to Determine Remediation
5.7.2 Use of Automated Machine Learning to Determine Causes and Assess Interventions
5.8 DEEP REINFORCEMENT LEARNING FOR ALARMING AND FAILURE ASSESSMENT
5.9 SOME TYPICAL AI AND MACHINE LEARNING EXAMPLES FOR FURTHER STUDY
5.9.1 Boolean Logic Tables
5.9.2 Statistical Regression and Variance
5.9.3 Artificial Neural Networks (ANNs)
5.9.4 Expert Systems
5.9.5 Sensitivity Analysis
5.9.6 Fuzzy Logic
5.9.7 Bayesian Networks
5.9.8 Genetic Algorithms
5.9.9 SmartSignal, PRiSM (AVEVA), and PPCL
5.9.10 Control System Effectiveness Study
5.10 WRAP-UP
CHAPTER 5.A PROCESS STATE TRANSITION LOGIC EMPLOYED BY THE ORIGINAL FMC FALCONEER KBS
5.A.1 INTRODUCTION
5.A.2 POSSIBLE PROCESS OPERATING STATES
5.A.3 SIGNIFICANCE OF PROCESS STATE IDENTIFICATION AND TRANSITION DETECTION
5.A.4 METHODOLOGY FOR DETERMINING PROCESS STATE IDENTIFICATION
5.A.4.1 Present Value States of all Key Sensor Data
5.A.4.2 Predicted Next Value States of all Key Sensor Data
5.A.5 PROCESS STATE IDENTIFICATION AND TRANSITION LOGIC PSEUDO-CODE
5.A.5.1 Attributes of the Current Data Vector
5.A.5.2 Method that is Applied to Each Updated Data Vector
5.A.6 SUMMARY
APPENDIX 5.A REFERENCES
CHAPTER 5.B PROCESS STATE TRANSITION LOGIC AND ITS ROUTINE USE IN FALCONEER™ IV
5.B.1 TEMPORAL REASONING PHILOSOPHY
5.B.2 INTRODUCTION
5.B.3 STATE IDENTIFICATION ANALYSIS CURRENTLY USED IN FALCONEER™ IV
5.B.4 STATE IDENTIFICATION ANALYSIS SUMMARY
APPENDIX 5.B REFERENCES
CHAPTER 6 PROCESS FAULT DETECTION BASED ON TIME-EXPLICIT KIVIAT DIAGRAM
OVERVIEW
CHAPTER HIGHLIGHTS
6.1 INTRODUCTION
6.2 TIME-EXPLICIT KIVIAT DIAGRAM
6.3 FAULT DETECTION BASED ON THE TIME-EXPLICIT KIVIAT DIAGRAM
6.4 CONTINUOUS PROCESSES
6.5 BATCH PROCESSES
6.6 PERIODIC PROCESSES
6.7 CASE STUDIES
6.8 CONTINUOUS PROCESSES
6.9 BATCH PROCESSES
6.10 PERIODIC PROCESSES
6.11 CONCLUSIONS
ACKNOWLEDGMENT
REFERENCES
ACKNOWLEDGMENTS
APPENDIX 6.A REFERENCES
CHAPTER 6.A VIRTUAL STATISTICAL PROCESS CONTROL ANALYSIS
6.A.1 OVERVIEW
6.A.2 INTRODUCTION
6.A.3 EWMA CALCULATIONS AND SPECIFIC VIRTUAL SPC ANALYSIS CONFIGURATIONS
6.A.3.1 Controlled Variables
6.A.3.2 Uncontrolled Variables and Performance Equation Variables
6.A.4 VIRTUAL SPC ALARM TRIGGER SUMMARY
6.A.5 VIRTUAL SPC ANALYSIS CONCLUSIONS
ACKNOWLEDGMENTS
APPENDIX 6.A REFERENCES
CHAPTER 7 SMART MANUFACTURING AND REAL-TIME CHEMICAL PROCESS HEALTH MONITORING AND DIAGNOSTIC LOCALIZATION
CHAPTER HIGHLIGHTS
7.1 INTRODUCTION TO PROCESS OPERATIONAL HEALTH MODELING
7.2 DIAGNOSTIC LOCALIZATION – KEY CONCEPTS
7.2.1 Qualitative Modeling and Symptomaticand Topographic Search
7.2.2 Functional Representation as a Qualitative Modeling Construct
7.2.3 Causal Link Assessment for Combined Topographical and Symptomatic Assessment
7.3 TIME
7.3.1 Discretization and Single Time-Step Analysis
7.3.2 Dynamics in an Individual Functional Representation
7.3.3 Time Window and Feature Extraction
7.4 THE WORKFLOW OF DIAGNOSTIC LOCALIZATION
7.5 DL-CLA USE CASE IMPLEMENTATION: NOVA CHEMICAL ETHYLENE SPLITTER
7.5.1 CPP Generation
7.5.2 CPP Interpretation
7.5.3 Diagnostic Localization
7.6 ANALYZING POTENTIAL MALFUNCTIONS OVER TIME
7.7 ANALYSIS OF VARIOUS OPERATIONAL SCENARIOS
7.7.1 Event Manifestation, Sensor Reliability/Sensor Malfunctions
7.7.2 Hypothetical and Function-Only Devices
7.7.3 Unaccounted Malfunctions, Graceful Degradation, Multiple Malfunctions, Etc.
7.7.4 Sensor Availability and Reliability
7.7.5 Process Complexity
7.8 DL-CLA INTEGRATION WITH SMART MANUFACTURING (SM)
7.9 AN FR MODEL LIBRARY
7.9.1 Reusing FR Device Models
7.9.2 Complex FR Models
7.9.3 Analysis and Results
7.10 CONCLUSIONS
REFERENCES
CHAPTER 8 OPTIMAL QUANTITATIVE MODEL-BASED PROCESS FAULT DIAGNOSIS
OVERVIEW
CHAPTER HIGHLIGHTS
8.1 INTRODUCTION
8.2 PROCESS FAULT ANALYSIS CONCEPT TERMINOLOGY
8.3 MOME QUANTITATIVE MODELS OVERVIEW
8.4 MOME QUANTITATIVE MODEL DIAGNOSTIC STRATEGY
8.5 MOME SV&PFA DIAGNOSTIC RULES’ LOGIC COMPILER MOTIVATIONS
8.6 MOME FUZZY LOGIC ALGORITHM OVERVIEW
8.6.1 MOME Fuzzy Logic Algorithm Details
8.6.2 Single Fault Fuzzy Logic SV&PFA Diagnostic Rule
8.6.3 Multiple Fault Fuzzy Logic SV&PFA Diagnostic Rule
8.7 SUMMARY OF THE MOME DIAGNOSTIC STRATEGY
8.8 ACTUAL PROCESS SYSTEM KBS APPLICATION PERFORMANCE RESULTS
8.9 CONCLUSIONS
ACKNOWLEDGMENTS
REFERENCES
CHAPTER 8.A FALCONEERTM IV FUZZY LOGIC ALGORITHM PSEUDO-CODE
8.A.1 Introduction
8.A.2 Single and Non-Interactive Multiple Faults
8.A.3 Pairs of Interactive Multiple Faults
8.A.4 Summary
CHAPTER 8.B MOME CONCLUSIONS
8.B.1 Overview
8.B.2 Summary of the Mome Diagnostic Strategy
8.B.3 Falconeer™ IV KBS Application Project Procedure
8.B.4 Optimal Automated Process Fault Analysis Conclusions
CHAPTER 9 FAULT DETECTION USING ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING
CHAPTER HIGHLIGHTS
ABBREVIATIONS USED
9.1 INTRODUCTION
9.2 ARTIFICIAL INTELLIGENCE
9.3 MACHINE LEARNING
9.4 ENGINEERED FEATURES
9.4.1 Fast Fourier Transformation and Signal Processing
9.4.2 Principal Component Analysis
9.5 MACHINE LEARNING ALGORITHMS
9.5.1 Decision Trees and Ensemble Trees
9.5.2 Artificial Neural Networks
9.5.3 Bayesian Networks
9.5.4 High-Density Clustering
9.5.5 Large Language Models and the Future AI-Driven Factories
REFERENCES
CHAPTER 10 KNOWLEDGE-BASED SYSTEMS
CHAPTER HIGHLIGHTS
ABBREVIATIONS USED
10.1 INTRODUCTION
10.2 KNOWLEDGE
10.2.1 Definition
10.2.2 Mathematical – Physical Contextualization
10.2.3 Procedural Knowledge
10.2.4 Heuristic Knowledge
10.2.5 Topological Knowledge
10.3 INFORMATION REQUIRED FOR DIAGNOSIS
10.3.1 Introduction
10.3.2 Association
10.3.3 State and State Transition
10.4 KNOWLEDGE REPRESENTATION
10.4.1 Introduction
10.4.2 Mathematical
10.4.3 Linguistic
10.4.4 Graphical
10.4.5 Conclusion
10.5 MAINTAINING, UPDATING, AND EXTENDING KNOWLEDGE
10.5.1 Introduction
10.5.2 Progressing Insight
10.5.3 Software as a Knowledge-Based System
10.5.4 A Database as a Knowledge-Based System
10.5.5 Search Engines
10.5.6 Online Encyclopedia
10.5.7 Learning
10.6 EXPERT SYSTEMS
10.6.1 Introduction
10.6.2 Inference Engine
10.6.3 Generic – Specific
10.6.4 Interactive – Real-Time
10.6.5 Polling
10.6.6 Conclusions
10.6.7 Documentation
10.7 DIGITIZATION, DIGITALIZATION, DIGITAL TRANSFORMATION, AND DIGITAL TWINS
10.7.1 Digitizing Information
10.7.2 Digitalization
10.7.3 Digital Transformation
10.7.4 Digital Twins
10.8 FAULT DIAGNOSIS WITH KNOWLEDGE-BASED SYSTEMS
10.8.1 Introduction
10.8.2 Elimination
10.8.3 Inductive (Forward) Reasoning
10.8.4 Deductive (Backward) Reasoning
10.8.5 Conclusions
10.9 GRAPHICAL REPRESENTATION OF FAULT DIAGNOSIS
10.9.1 Introduction
10.9.2 Fault Trees
10.9.3 FMEA
10.9.4 Bowtie Analysis
10.9.5 Complex Event Processing
10.10 CONCLUSIONS
REFERENCES
CHAPTER 10.A COMPRESSOR TRIP PREDICTION
CHAPTER 11 THE FALCON PROJECT
OVERVIEW
CHAPTER HIGHLIGHTS
11.1 INTRODUCTION
11.2 THE DIAGNOSTIC PHILOSOPHY UNDERLYING THE FALCON SYSTEM
11.3 TARGET PROCESS SYSTEM
11.4 THE FIELDED FALCON SYSTEM
11.4.1 The Inference Engine
11.4.2 The Human/Machine Interface
11.4.3 The Dynamic Simulation Model
11.4.4 The Diagnostic Knowledge Base
11.5 THE DERIVATION OF THE FALCON DIAGNOSTIC KNOWLEDGE BASE
11.5.1 First Rapid Prototype of the FALCON System KBS
11.5.2 The Fielded FALCON System Development
11.5.3 The Fielded FALCON System’s Performance Results
11.6 THE IDEAL FALCON SYSTEM
11.7 USE OF THE KNOWLEDGE-BASED SYSTEM PARADIGM IN PROBLEM SOLVING
ACKNOWLEDGMENTS
REFERENCES
CHAPTER 12 FAULT DIAGNOSTIC APPLICATION IMPLEMENTATION AND SUSTAINABILITY
OVERVIEW
CHAPTER HIGHLIGHTS
NOMENCLATURE
12.1 KEY PRINCIPLES OF SUCCESSFULLY IMPLEMENTING NEW TECHNOLOGY
12.2 EXPECTATION OF ADVANCED TECHNOLOGY
12.2.1 What Are the Expected Actions?
12.2.2 Who Is the Audience?
12.2.3 What Are the Failure Modes?
12.2.4 When Is an Alert Expected and Valid?
12.3 DEFINING SUCCESS
12.4 LEARNING FROM HISTORY
12.5 EXAMPLE: REGULATORY CONTROL LOOP MONITORING
12.5.1 Regulatory Control Failure Modes
12.5.2 Expectations of Loop Monitoring
12.6 WHAT SUCCESS LOOKS LIKE
12.7 EXAMPLE: SYSTEMATIC STEWARDSHIP
12.8 CONCLUSIONS
12.8.1 Motivational Requirement
12.8.2 Setup Requirements
12.8.3 Usage Requirements
12.8.4 Sustainment and Continuous Improvement
REFERENCES
CHAPTER 13 PROCESS OPERATORS, ADVANCED PROCESS CONTROL, AND ARTIFICIAL INTELLIGENCE-BASED APPLICATIONS IN THE CONTROL ROOM
CHAPTER HIGHLIGHTS
OVERVIEW
13.1 INTRODUCTION
13.2 HISTORY OF SUSTAINABLE APC
13.3 OPERATORS AS ULTIMATE APC APPLICATION END USERS
13.4 APC APPLICATION DESIGN CONSIDERATIONS
13.5 APC DEVELOPMENT – INTERNAL VERSUS EXTERNAL EXPERTS
13.6 APC TECHNOLOGY
13.7 APC SUPPORT
13.8 CONCLUSIONS
REFERENCES
INDEX
EULA