Atnaujinkite slapukų nuostatas

Prognostics and Remaining Useful Life (RUL) Estimation: Predicting with Confidence [Kietas viršelis]

4.00/5 (2 ratings by Goodreads)
(University of Maryland, Dept. of Mechanical Engineering, College Park, USA), (PARC, Palo Alto Research Center, USA), , (Luleå University of Technology, Sweden)
  • Formatas: Hardback, 461 pages, aukštis x plotis: 254x178 mm, weight: 1016 g, 32 Tables, black and white; 185 Line drawings, black and white; 18 Halftones, black and white; 203 Illustrations, black and white
  • Išleidimo metai: 28-Dec-2021
  • Leidėjas: CRC Press
  • ISBN-10: 0367563061
  • ISBN-13: 9780367563066
Kitos knygos pagal šią temą:
Prognostics and Remaining Useful Life (RUL) Estimation: Predicting with ConfidenceDidesnis paveikslėlis
  • Formatas: Hardback, 461 pages, aukštis x plotis: 254x178 mm, weight: 1016 g, 32 Tables, black and white; 185 Line drawings, black and white; 18 Halftones, black and white; 203 Illustrations, black and white
  • Išleidimo metai: 28-Dec-2021
  • Leidėjas: CRC Press
  • ISBN-10: 0367563061
  • ISBN-13: 9780367563066
Kitos knygos pagal šią temą:
Pastovi nuoroda: https://www.kriso.lt/db/9780367563066.html
Raktažodžiai:
Maintenance combines various methods, tools, and techniques in a bid to reduce maintenance costs while increasing the reliability, availability, and security of equipment. Condition-based maintenance (CBM) is one such method, and prognostics forms a key element of a CBM program based on mathematical models for predicting remaining useful life (RUL). Prognostics and Remaining Useful Life (RUL) Estimation: Predicting with Confidence compares the techniques and models used to estimate the RUL of different assets, including a review of the relevant literature on prognostic techniques and their use in the industrial field. This book describes different approaches and prognosis methods for different assets backed up by appropriate case studies.

FEATURES











Presents a compendium of RUL estimation methods and technologies used in predictive maintenance





Describes different approaches and prognosis methods for different assets





Includes a comprehensive compilation of methods from model-based and data-driven to hybrid





Discusses the benchmarking of RUL estimation methods according to accuracy and uncertainty, depending on the target application, the type of asset, and the forecast performance expected





Contains a toolset of methods and a way of deployment aimed at a versatile audience

This book is aimed at professionals, senior undergraduates, and graduate students in all interdisciplinary engineering streams that focus on prognosis and maintenance.
Preface xxi
Authors xxvii
Chapter 1 Information in Maintenance 1(44)
1.1 Traditional Maintenance: Corrective and Preventive
1(6)
1.1.1 Traditional Corrective Maintenance
1(3)
1.1.1.1 Corrective Maintenance
1(3)
1.1.2 Traditional Preventive Maintenance
4(3)
1.1.2.1 Preventive Maintenance
4(3)
1.2 CMMS and IT Systems Supporting Maintenance Function
7(5)
1.2.1 Computer Maintenance Management Systems (CMMS)
7(4)
1.2.1.1 CMMS Needs Assessment
7(1)
1.2.1.2 CMMS Capabilities
7(1)
1.2.1.3 CMMS Benefits
8(1)
1.2.1.4 Finding a CMMS
8(1)
1.2.1.5 Role of CMMS
8(1)
1.2.1.6 CMMS Implementation
9(2)
1.2.2 What Is IT System Support and Maintenance?
11(1)
1.2.2.1 Importance of Quality IT Assets and Maintenance
11(1)
1.3 Sensors for Health Monitoring and SCADA Systems: Operational Technologies (OTs)
12(9)
1.3.1 Health Monitoring
12(2)
1.3.1.1 Wireless Standards for Health Monitoring
12(1)
1.3.1.2 Sensors Facilitate Health Monitoring
12(1)
1.3.1.3 Different Types of Sensors
13(1)
1.3.1.4 Key Sensors and Applications
13(1)
1.3.2 Wearable Health Monitoring Systems (WHMS)
14(2)
1.3.3 SCADA Systems
16(5)
1.3.3.1 Basics of SCADA
17(1)
1.3.3.2 Architecture of SCADA
18(1)
1.3.3.3 Types of SCADA Systems
18(1)
1.3.3.4 Applications of SCADA
18(2)
1.3.3.5 Understanding SCADA
20(1)
1.4 Sensor Fusion, Data Fusion, and Information Fusion for Maintenance
21(14)
1.4.1 Sensor Fusion
21(8)
1.4.1.1 Sensor Fusion Architecture
24(1)
1.4.1.2 How Sensor Fusion Works
25(1)
1.4.1.3 Sensor Fusion Levels
26(1)
1.4.1.4 Leveraging Sensor Fusion for the Internet of Things (IoT)
26(2)
1.4.1.5 Sensor Fusion Advantages
28(1)
1.4.1.6 Challenges to Sensor Fusion
28(1)
1.4.2 Data Fusion
29(5)
1.4.2.1 Structures in Data Fusion
29(1)
1.4.2.2 Classification of Data Fusion Techniques
30(4)
1.4.3 Information Fusion
34(1)
1.4.3.1 An Introduction to Information Fusion
34(1)
1.4.3.2 Information Fusion Model
34(1)
1.5 Condition Monitoring and the End of Traditional Preventive Maintenance
35(5)
1.5.1 Condition Monitoring
35(4)
1.5.1.1 Condition Monitoring as Tool of Preventive Maintenance
37(2)
1.5.2 Optimizing Preventive Maintenance
39(1)
1.6 Predictive Maintenance as the Evolution of CBM Programs
40(5)
1.6.1 Condition-Based Maintenance (CBM)
40(1)
1.6.1.1 CBM Elements and Techniques
40(1)
1.6.1.2 Requirements for CBM Implementation
40(1)
1.6.2 Predictive Maintenance vs. Condition-Based Maintenance (CBM)
40(1)
1.6.3 Differences between Predictive Maintenance and Condition-Based Maintenance (CBM)
41(4)
1.6.3.1 Key Conclusions
42(3)
Chapter 2 Predictive Maintenance Programs and Servitization Maintenance as a Service (MaaS) Creating Value through Prognosis Capabilities 45(44)
2.1 Industry 4.0 and Servitization
45(13)
2.1.1 Industry 4.0
45(6)
2.1.1.1 Industry 4.0 Definition
46(1)
2.1.1.2 What Is Industry 4.0?
46(2)
2.1.1.3 Industry 4.0 Conception
48(1)
2.1.1.4 Industry 4.0 Components
49(1)
2.1.1.5 Industry 4.0: Design Principles
50(1)
2.1.2 Reference Architecture Model Industry 4.0 (RAMI 4.0)
51(1)
2.1.2.1 RAMI 4.0
51(1)
2.1.3 Industry 4.0 Component Model
52(2)
2.1.3.1 Specifications of Industry 4.0 Component Model
52(2)
2.1.4 Servitization
54(4)
2.1.4.1 Concept of Servitization
55(1)
2.1.4.2 Defining "Servitization"
56(1)
2.1.4.3 Features of Servitization
57(1)
2.1.5 Industry 4.0 Services
58(1)
2.1.5.1 Industry 4.0 Servitization Framework
58(1)
2.2 Performance-Based Contracting (PBC)
58(5)
2.2.1 Performance-Based Contracting Metrics
59(2)
2.2.1.1 Performance-Based Contracting Implementation Process
60(1)
2.2.2 Performance-Based Contracting in the Defense Industry
61(1)
2.2.3 Challenges and Opportunities for Performance-Based Contracting
62(1)
2.3 Virtual Engineering and Prognostics for Added Value Services
63(8)
2.3.1 Virtual Engineering: A Paradigm for the 21st Century
63(1)
2.3.2 Virtual Engineering Environments
64(1)
2.3.2.1 General Aspects
64(1)
2.3.2.2 Heterogeneous Data Formats
64(1)
2.3.2.3 Virtual Reality Devices
64(1)
2.3.3 Simple Definition of Value Added
65(2)
2.3.3.1 Four Types of Value-Added Work
66(1)
2.3.4 Virtual Engineering and R&D
67(1)
2.3.5 Ubiquitous Computing Technologies for Next Generation Virtual Engineering
68(3)
2.4 Product Lifecycle Management and Predictive Maintenance: Changing Role of Suppliers and End Users
71(7)
2.4.1 Product Lifecycle Management Approach
72(3)
2.4.1.1 "Outside-In": New Approach to PLM
73(2)
2.4.2 Evolution of PLM
75(1)
2.4.3 Product Lifecycle Management and Predictive Maintenance
75(1)
2.4.3.1 Repair before Standstill
76(1)
2.4.4 Digital Transformation
76(1)
2.4.5 PLM Business Value
76(2)
2.4.5.1 Increase Profitable Growth
77(1)
2.4.5.2 Reduce Build Costs
77(1)
2.5 RUL Estimation as Enabling Technology for Circular Economics
78(11)
2.5.1 Remaining Useful Life (RUL)
78(2)
2.5.1.1 Classification of Techniques for RUL Prediction
78(1)
2.5.1.2 Types of Prediction Techniques
79(1)
2.5.2 What Is the Circular Economy?
80(1)
2.5.3 Digital Technology: Enabling Transition
81(1)
2.5.4 Circular Economy Business Models
82(1)
2.5.5 Techniques for RUL Estimation and Maintenance Investment Outcomes
82(7)
2.5.5.1 Engineering Analysis
82(1)
2.5.5.2 Cost and Budget Models
83(1)
2.5.5.3 Operations Research Models
83(1)
2.5.5.4 Simulation Models
83(1)
2.5.5.5 Proprietary Models
84(5)
Chapter 3 RUL Estimation Powered by Data-Driven Techniques 89(46)
3.1 Approaches to Maintenance: Physical Model-Based vs. Data-Driven
89(1)
3.1.1 Where Run-to-Failure Data Come From
89(1)
3.1.1.1 Advantages
89(1)
3.1.1.2 Disadvantages
90(1)
3.2 Neural Networks (NNs), Fuzzy Logic, Decision Trees, Support Vector Machines (SVMs), Anomaly Detection Algorithms, Reinforcement Learning, Classification, Clustering and Bayesian Methods, and Data Mining Algorithms
90(25)
3.2.1 Neural Networks
90(3)
3.2.1.1 What Is a Neural Network?
90(1)
3.2.1.2 What Does a Neural Network Consist of9
90(1)
3.2.1.3 How Does a Neural Network Learn Things?
91(1)
3.2.1.4 Neural Network Architecture
92(1)
3.2.1.5 Architectural Components
92(1)
3.2.1.6 Neural Network Algorithms
92(1)
3.2.2 Fuzzy Logic
93(7)
3.2.2.1 Fuzzy Logic Operators
94(3)
3.2.2.2 Fuzzy Rules
97(1)
3.2.2.3 Fuzzy Sets and Crisp Sets
98(1)
3.2.2.4 Fuzzy Logic Applications
99(1)
3.2.3 Decision Trees
100(1)
3.2.3.1 Common Terms Used with Decision Trees
100(1)
3.2.3.2 How Decision Trees Work
100(1)
3.2.3.3 Types of Decision Trees
100(1)
3.2.3.4 Decision Tree Applications
100(1)
3.2.4 Support Vector Machines
101(4)
3.2.4.1 How an SVM Works
101(1)
3.2.4.2 Advantages of Support Vector Machines
101(1)
3.2.4.3 Disadvantages of Support Vector Machines
101(1)
3.2.4.4 SVM Applications
101(1)
3.2.4.5 Selecting the Right Hyper-Plane
102(3)
3.2.5 Anomaly Detection Algorithms
105(2)
3.2.5.1 What Are Anomalies?
106(1)
3.2.5.2 Anomaly Detection Algorithms
107(1)
3.2.6 Reinforcement Learning
107(3)
3.2.6.1 Terminologies Used in the Field of Reinforcement Learning
108(1)
3.2.6.2 Applications
109(1)
3.2.7 Classification, Clustering, and Bayesian Methods
110(3)
3.2.7.1 Classification
110(2)
3.2.7.2 Clustering
112(1)
3.2.7.3 Bayesian Methods
112(1)
3.2.8 Data Mining Algorithms
113(2)
3.2.8.1 Types of Data Mining Algorithms
114(1)
3.2.8.2 Top Data Mining Algorithms
114(1)
3.3 Conventional Numerical Techniques: Wavelets, Kalman Filters, Particle Filters, Regression, Demodulation, and Statistical Methods
115(8)
3.3.1 Wavelets
115(3)
3.3.1.1 Solving Partial Differential Equations (PDEs) Using Wavelets
117(1)
3.3.2 Kalman Filters
118(1)
3.3.3 Particle Filters
119(1)
3.3.4 Regression
120(1)
3.3.4.1 Linear Regression
121(1)
3.3.4.2 Multilinear Regression
121(1)
3.3.4.3 Nonlinear Regression
121(1)
3.3.5 Demodulation
121(2)
3.4 Statistical Approaches
123(12)
3.4.1 Gamma Process
123(1)
3.4.1.1 Gamma Process Model
123(1)
3.4.2 Hidden Markov Model
123(1)
3.4.3 Regression-Based Model
124(2)
3.4.4 Relevance Vector Machine (RVM)
126(1)
3.4.5 Autoregressive (AR) Model
127(1)
3.4.5.1 AR(p) Models
127(1)
3.4.6 Threshold Autoregressive (TAR) Model
128(1)
3.4.7 Bilinear Model
128(1)
3.4.8 Projection Pursuit
129(1)
3.4.9 Multivariate Adaptive Regression Splines (MARS)
129(1)
3.4.9.1 Explanation of MARS Method
130(1)
3.4.10 Volterra Series Expansion
130(5)
3.4.10.1 Volterra Series: Background and Definitions
131(4)
Chapter 4 Context Awareness and Situation Awareness in Prognostics 135(38)
4.1 IT vs. OT
135(4)
4.1.1 IT and OT - What's the Difference?
135(1)
4.1.2 When Worlds Collide - Industrial Internet
135(1)
4.1.3 New Concerns for Both Sides
135(2)
4.1.3.1 New Concerns for IT
136(1)
4.1.3.2 New Concerns for OT
137(1)
4.1.4 Finding Common Ground
137(1)
4.1.5 Differences between IT and OT
137(1)
4.1.5.1 Technological Needs
138(1)
4.1.5.2 Conditions of Conservation
138(1)
4.1.5.3 Security
138(1)
4.1.6 Regulations and Protocols
138(1)
4.1.7 Data vs. Processes
138(1)
4.1.8 Update Frequency
138(1)
4.1.9 IIoT Devices in Industry 4.0
138(1)
4.1.10 What Industrial Processes Will Improve IT and OT Integration?
139(1)
4.1.10.1 Energy
139(1)
4.1.10.2 Environment
139(1)
4.1.10.3 Production
139(1)
4.1.10.4 Quality Control
139(1)
4.1.10.5 Maintenance
139(1)
4.2 Context Definitions and Context Categorization
139(7)
4.2.1 Definition of Context
139(1)
4.2.2 Operating Context
140(1)
4.2.3 Context Awareness for Asset Maintenance Decisions
140(1)
4.2.4 Context-Driven Maintenance
141(1)
4.2.5 Classification of Context Types
141(1)
4.2.6 Categorization by Context
142(2)
4.2.6.1 Introduction
142(1)
4.2.6.2 Categories of Context
143(1)
4.2.7 Categorization of Characteristics of Context-Aware Applications
144(1)
4.2.8 Context Categorization, Acquisition, and Modeling
145(1)
4.2.9 Categorization of Context in Mobile Map Services
146(1)
4.3 Continuous Change, Temporality, and Spatiality
146(4)
4.3.1 What Happens When Transformation Becomes the Rule
146(1)
4.3.1.1 Real Change Takes Time
146(1)
4.3.1.2 Find the Right Perspective
147(1)
4.3.1.3 Continuous Engagement
147(1)
4.3.2 Cycle of Continuous Change
147(2)
4.3.2.1 Phase 1: Using Influence to Sell Ideas
148(1)
4.3.2.2 Phase 2: Using Authority to Change Practices
148(1)
4.3.2.3 Phase 3: Embedding Change in Technology
149(1)
4.3.2.4 Phase 4: Managing Culture to Fuel the Cycle of Change
149(1)
4.3.3 Temporal Data and Discovery
149(1)
4.4 Modeling Context and Representation Methods
150(3)
4.4.1 Context Model
150(1)
4.4.2 Evolution of Context Modeling and Reasoning
150(2)
4.4.2.1 Requirements
150(2)
4.4.2.2 Early Approaches: Key-Value and Markup Models
152(1)
4.4.2.3 Domain-Focused Modeling
152(1)
4.4.2.4 Toward More Expressive Modeling Tools
152(1)
4.4.3 Modeling Approaches
152(1)
4.4.3.1 Key Value Models
152(1)
4.4.3.2 Markup Scheme Models
153(1)
4.4.3.3 Graphical Models
153(1)
4.4.3.4 Object-Oriented Models
153(1)
4.4.3.5 Logic-Based Models
153(1)
4.4.3.6 Ontology-Based Models
153(1)
4.4.3.7 Spatial Context Model
153(1)
4.5 Ontologies and Context for Remaining Useful Life Estimation
153(7)
4.5.1 Definition of Ontology
153(1)
4.5.2 Ontology Use Cases
154(1)
4.5.3 Benefits of Using Ontologies
154(1)
4.5.4 Limitations of Ontologies
154(1)
4.5.5 Context Ontology
154(2)
4.5.5.1 Existing Upper Ontologies
155(1)
4.5.5.2 Scope of Logical Content
155(1)
4.5.5.3 Scope of Representational Framework
155(1)
4.5.6 Ontology Classifications
156(1)
4.5.6.1 Classification Based on Language Expressivity and Formality
156(1)
4.5.6.2 Classification Based on the Scope of the Ontology or the Domain Granularity
156(1)
4.5.7 Context Driven Remaining Useful Life (RUL) Estimation
157(1)
4.5.7.1 Introduction
157(1)
4.5.8 Methods for Prognostics and Remaining Useful Life Estimation,
158(1)
4.5.9 Data-Driven Methods for Remaining Life Estimation
158(2)
4.6 Context Uncertainty Management
160(3)
4.6.1 Uncertainty Management Theory
160(1)
4.6.2 Aspects of Uncertainty
160(1)
4.6.3 Uncertainty Management in Context-Aware Applications
160(1)
4.6.3.1 Uncertainty in Context-Aware Computing
160(1)
4.6.4 Locating and Modeling Uncertainty
161(2)
4.6.4.1 Context Uncertainty
161(1)
4.6.4.2 Model Uncertainty
162(1)
4.6.4.3 Input Uncertainty
163(1)
4.6.4.4 Parameter Uncertainty
163(1)
4.6.4.5 Model Outcome Uncertainty
163(1)
4.7 Prognosis in Prescriptive Analytics Powered by Context
163(10)
4.7.1 What Is Prescriptive Analytics?
163(1)
4.7.2 Prescriptive Maintenance: Building Alternative Plans for Smart Operations
164(2)
4.7.2.1 Prescriptive Maintenance Framework
164(2)
4.7.3 Methods and Techniques for Prescriptive Analytics
166(1)
4.7.4 Categories of Methods for Predictive and Prescriptive Analytics
167(6)
Chapter 5 Black Swans and Physics of Failure 173(40)
5.1 Prognosis Performance of Data-Driven Estimators
173(4)
5.1.1 Introduction
173(1)
5.1.2 Motivation
174(1)
5.1.3 Prognostic Performance Metrics
175(1)
5.1.3.1 Offline vs. Online Performance Metrics
175(1)
5.1.3.2 Offline Performance Evaluation
175(1)
5.1.3.3 Prognostic Horizon
175(1)
5.1.4 Data-Driven Techniques for Prognostics
176(1)
5.2 Black Swans in Risk Estimation
177(4)
5.2.1 Black Swan
177(1)
5.2.2 What Is a Black Swan Event?
177(1)
5.2.3 Basic Approaches to Managing Risk and Black Swans
178(1)
5.2.4 What Is a Black Swan in a Risk Context?
179(2)
5.2.4.1 Examples of Black Swans
179(1)
5.2.4.2 Three Types of Black Swans
180(1)
5.3 Failure Modes and Causes Missing in Data: Black Swans
181(4)
5.3.1 Failure Mode and Effects Analysis
181(1)
5.3.2 Functional Failure Mode and Effects Analysis
181(1)
5.3.3 Black Swans, Cognition, and the Power of Learning from Failure
182(2)
5.3.3.1 Definitions of Failure
182(2)
5.3.4 Black Swans and Fatigue Failures
184(1)
5.3.4.1 The Question
184(1)
5.3.4.2 Fatigue Failure and Fat Tails
185(1)
5.4 Probabilistic Physics of Failure Approach to Reliability
185(8)
5.4.1 Physics of Failure: An Introduction
185(1)
5.4.2 Physics of Failure Models
186(1)
5.4.2.1 Introduction to Physics of Failure Models
186(1)
5.4.3 Deterministic vs. Empirical Models
187(1)
5.4.3.1 Deterministic
187(1)
5.4.3.2 Empirical
187(1)
5.4.4 Reasons to Use PoF-Based Modeling in Reliability
187(1)
5.4.5 PoF Model Development Steps
187(1)
5.4.5.1 Strengths of PoF Modeling
187(1)
5.4.5.2 Weaknesses of PoF Modeling
187(1)
5.4.5.3 PoF Model Development Steps
188(1)
5.4.6 Physics of Failure Procedure
188(1)
5.4.7 Probabilistic Physics of Failure
189(1)
5.4.8 Prognostics and Health Management Using Physics of Failure
189(1)
5.4.8.1 PoF-Based PHM Implementation Approach
190(1)
5.4.9 Application of PoF for PHM
190(2)
5.4.10 Probabilistic Physics of Failure Degradation Models
192(1)
5.4.11 Why Physics of Failure is Preferred to Mean Time between Failures (MTBF) for Reliability Testing
192(1)
5.4.11.1 Mean Time between Failures
192(1)
5.4.11.2 Physics of Failure
193(1)
5.5 Mechanisms of Failure and Associated PoF Models
193(6)
5.5.1 Degradation Mechanisms
193(2)
5.5.1.1 Types of Degradation Mechanisms
193(2)
5.5.2 Review of Prognostics and Health Management
195(1)
5.5.2.1 Physics of Failure Approach
195(1)
5.5.3 Failure Modes, Causes, Mechanisms, and Models
196(1)
5.5.4 Failure Modes, Mechanisms, and Effects Analysis
197(2)
5.6 Time-Dependence of Materials and Device Degradation
199(3)
5.6.1 Condition-Based Prediction of Time-Dependent Reliability in Composites
199(2)
5.6.1.1 Introduction
199(1)
5.6.1.2 Fatigue Damage Modeling
200(1)
5.6.2 Mapping Degradation Mechanism and Techniques
201(1)
5.6.3 Through-Life Engineering Services, Degradation Mechanisms, and Techniques to Predict RUL
202(1)
5.7 Uncertainties and Model Validation
202(11)
5.7.1 Model Validation and Prediction
202(1)
5.7.2 Model Validation Statement
203(1)
5.7.3 Uncertainties in Physical Measurements
203(1)
5.7.4 Types of Uncertainty
204(1)
5.7.4.1 Aleatory Uncertainty
204(1)
5.7.4.2 Epistemic Uncertainty
205(1)
5.7.4.3 Epistemic vs. Aleatory Uncertainty
205(1)
5.7.5 Quantitative Validation of Model Prediction
205(2)
5.7.6 Uncertainty in Prognostics
207(6)
Chapter 6 Hybrid Prognostics Combining Physics-Based and Data-Driven Approaches 213(40)
6.1 Information Requirements for Hybrid Models in Prognosis
213(4)
6.1.1 Prognostics Technology
213(1)
6.1.1.1 Experience-Based Prognostic Models
213(1)
6.1.1.2 Data-Driven Models
213(1)
6.1.1.3 Physics-Based Prognostic Models
214(1)
6.1.2 Hybrid Prognostics Approach
214(3)
6.1.2.1 Prognostics Application
216(1)
6.2 Synthetic Data Generation vs. Model Tuning
217(6)
6.2.1 Synthetic Data Generation
217(3)
6.2.1.1 Types of Synthetic Data
218(1)
6.2.1.2 Approaches to and Methods for Synthetic Data Generation
218(2)
6.2.2 Model Tuning
220(3)
6.2.2.1 Definition of Model Tuning
220(3)
6.2.2.2 Why Model Tuning Is Important
223(1)
6.3 Hybrid Approach Incorporating Experience-Based Models and Data-Driven Models
223(5)
6.3.1 Hybrid Approach
225(1)
6.3.2 Hybrid Approaches
226(2)
6.4 Hybrid Approach Incorporating Experience-Based Models and Physics-Based Models
228(6)
6.4.1 Machine Learning vs. Physics-Based Modeling
228(1)
6.4.1.1 Hybrid Analytics: Combining Machine Learning and Physics-Based Modeling
229(1)
6.4.1.2 Why Use Machine Learning When Physics-Based Models Are Available?
229(1)
6.4.2 Physics of Failure-Based PHM Implementation Approach
229(4)
6.4.2.1 Failure Modes, Mechanisms, and Effects Analysis
229(1)
6.4.2.2 Lifecycle Load Monitoring
230(2)
6.4.2.3 Data Reduction and Load Feature Extraction
232(1)
6.4.2.4 Damage Assessment and Remaining Life Calculation
232(1)
6.4.2.5 Uncertainty Implementation and Assessment
233(1)
6.4.3 Physics-Based Modeling Approaches to Engine Health Management
233(1)
6.5 Hybrid Approach Incorporating Multiple Data-Driven Models
234(5)
6.5.1 Hybrid LSSVR/HMM-Based Prognostics Approach
236(3)
6.5.1.1 LSSVR/HMM-Based Prognostics
236(3)
6.6 Hybrid Approach Incorporating Data-Driven Models and Physics-Based Models
239(5)
6.6.1 Physics-Based Prognostics vs. Data-Driven Prognostics
239(1)
6.6.2 Fusion Prognostics Framework of Data-Driven and Physics-Based Methods
240(3)
6.6.2.1 Data-Driven Methods
241(1)
6.6.2.2 Physics-Based Method
241(2)
6.6.3 A New Hybrid Prognostic Methodology
243(1)
6.6.3.1 Methodology
244(1)
6.7 Hybrid Approach Incorporating Experience-Based Models, Data-Driven Models, and Physics-Based Models
244(9)
6.7.2 Proposed Prognosis Framework
246(1)
6.7.2.1 Dynamic Bayesian Networks
246(1)
6.7.2.2 Physics-Based Models
246(1)
6.7.2.3 Diagnosis
247(1)
6.7.2.4 Fault Diagnosis and Diagnosis Uncertainty Quantification
247(1)
6.7.2.5 Prognosis Validation
247(1)
6.7.3 Prognostics and Health Monitoring in the Presence of Heterogeneous Information
247(6)
6.7.3.1 Bayesian Networks and Dynamic Bayesian Networks
248(1)
6.7.3.2 Heterogeneous Information
248(1)
6.7.3.3 Learning Bayesian Networks and Dynamic Bayesian Networks
249(1)
6.7.3.4 Diagnosis
249(4)
Chapter 7 Prognosis in Prescriptive Analytics 253(26)
7.1 Evolution from Description to Prediction and Prescription
253(10)
7.1.1 Evolving Analytics: Descriptive to Prescriptive to Predictive
253(3)
7.1.1.1 Descriptive Analytics: Understanding the Past
253(1)
7.1.1.2 Predictive Analytics: Understanding What to Do
254(1)
7.1.1.3 Prescriptive Analytics: Seeing the Possible Future
255(1)
7.1.2 Categories of Methods for Predictive Analytics and Prescriptive Analytics
256(1)
7.1.3 Toward Prescriptive Analytics
256(3)
7.1.3.1 How Prescriptive Analytics Works
259(1)
7.1.4 Five Pillars of Prescriptive Analytics Success
259(2)
7.1.4.1 Hybrid Data
259(1)
7.1.4.2 Integrated Predictions and Prescriptions
260(1)
7.1.4.3 Prescriptions and Side Effects
260(1)
7.1.4.4 Adaptive Algorithms
261(1)
7.1.4.5 Feedback Mechanism
261(1)
7.1.5 Maintenance Analytics Concept
261(2)
7.1.5.1 Maintenance Descriptive Analytics
262(1)
7.1.5.2 Maintenance Diagnostic Analytics
262(1)
7.1.5.3 Maintenance Predictive Analytics
262(1)
7.1.5.4 Maintenance Prescriptive Analytics
263(1)
7.1.6 Maintenance Analytics and eMaintenance
263(1)
7.2 Role of Prognosis in a Dynamic Environment
263(3)
7.2.1 Procedure for Prognostics of Dynamic Systems
263(3)
7.2.1.1 Dynamic Bayesian Networks
263(2)
7.2.1.2 Prognostic Procedure for Dynamic Systems
265(1)
7.3 Probabilistic Models for Prescription
266(4)
7.3.1 Markov Decision Process
266(1)
7.3.2 Hidden Markov Model
266(1)
7.3.3 Markov Chain
267(3)
7.3.3.1 Types of Markov Chains
269(1)
7.3.3.2 Transitions
269(1)
7.4 Machine Learning and Data Mining in Prescriptive Analytics
270(4)
7.4.1 Machine Learning for Prescriptive Analytics
271(2)
7.4.1.1 Multi-Objective Reinforcement Learning for Prescriptive Analytics
271(2)
7.4.2 Data Mining for Prescriptive Analytics
273(1)
7.4.2.1 Predictive and Prescriptive Analytics
273(1)
7.4.2.2 Prescriptive Analytics in the Information Age (IA)
274(1)
7.5 Simulation and Logic-Based Methods for RUL Estimation in Prescriptive Analytics
274(5)
7.5.1 Simulation
274(1)
7.5.2 Logic-Based Methods
274(5)
7.5.2.1 Association Rules
275(1)
7.5.2.2 Decision Rules
276(3)
Chapter 8 Uncertainty Management and the Confidence of RUL Predictions 279(40)
8.1 Uncertainty Representation and Interpretation
279(3)
8.1.1 Uncertainty Interpretation
279(1)
8.1.2 Representing Uncertainty in Prognostic Tasks
280(1)
8.1.2.1 Probabilistic Representation of Epistemic Uncertainty
280(1)
8.1.2.2 Possibilistic Representation of the Epistemic Uncertainty
280(1)
8.1.2.3 Transformation from a Possibilistic Distribution to a Probabilistic Distribution
281(1)
8.1.3 Uncertainty within the Context of Risk Analysis
281(1)
8.2 Uncertainty Quantification, Propagation, and Management
282(13)
8.2.1 Uncertainty Quantification
282(7)
8.2.1.1 Uncertainty Quantification Classification
282(1)
8.2.1.2 Two Types of Uncertainty Quantification Problems
283(1)
8.2.1.3 Uncertainty Quantification in RUL Prediction
283(5)
8.2.1.4 Prognostics Challenges in Using Uncertainty Quantification
288(1)
8.2.2 Uncertainty Propagation
289(2)
8.2.2.1 Uncertainty Propagation Methods
289(2)
8.2.3 Uncertainty Management
291(4)
8.2.3.1 Uncertainty Management in Prognostics
292(1)
8.2.3.2 Uncertainty Management in Long-Term Predictions
293(2)
8.2.3.3 Implications for Uncertainty Management
295(1)
8.3 Sources of Uncertainty and Modeling Uncertainty
295(9)
8.3.1 Sources of Uncertainty
297(2)
8.3.1.1 Present Uncertainty
298(1)
8.3.1.2 Future Uncertainty
299(1)
8.3.1.3 Modeling Uncertainty
299(1)
8.3.1.4 Prediction Method Uncertainty
299(1)
8.3.2 Sources of Uncertainty in Prognostics
299(1)
8.3.2.1 Modeling Error
299(1)
8.3.2.2 Noise
300(1)
8.3.2.3 Sensors
300(1)
8.3.3 Uncertainty Source Analysis
300(1)
8.3.3.1 Model Parameter Uncertainty
301(1)
8.3.3.2 Measurement Noise
301(1)
8.3.3.3 Failure Threshold Uncertainty
301(1)
8.3.4 Other Sources of Uncertainty
301(1)
8.3.4.1 Parameter Uncertainty
302(1)
8.3.4.2 Parametric Variability
302(1)
8.3.4.3 Structural Uncertainty
302(1)
8.3.4.4 Algorithmic Uncertainty
302(1)
8.3.4.5 Experimental Uncertainty
302(1)
8.3.4.6 Interpolation Uncertainty
302(1)
8.3.5 Kinds of Uncertainty
302(1)
8.3.6 Uncertainty Model-Based Approaches
303(1)
8.3.7 Sources of Uncertainty in Prognostics and Health Management
304(1)
8.4 Uncertainty in Terms of Physical and Subjective Probabilities
304(5)
8.4.1 Physical Probabilities
304(4)
8.4.1.1 Frequentist View
305(2)
8.4.1.2 Confidence Intervals: Frequentist Approach
307(1)
8.4.2 Subjective Probabilities
308(1)
8.4.2.1 Subjective (Bayesian) View
308(1)
8.4.3 Choice of Interpretation
309(1)
8.5 Probability Distribution of Remaining Useful Life as an Uncertainty Propagation Problem
309(10)
8.5.1 Uncertainty Associated with RUL Estimation
309(2)
8.5.2 Uncertainty Propagation Methods for RUL Estimation
311(4)
8.5.2.1 Sampling-Based Methods
312(1)
8.5.2.2 Analytical Methods
313(1)
8.5.2.3 Hybrid Methods
313(1)
8.5.2.4 Discussion
314(1)
8.5.3 Probability Distribution: Bayesian Approach
315(4)
Chapter 9 RUL Estimation of Dynamic and Static Assets 319(42)
9.1 Physics of Failure in Dynamic and Non-Dynamic Assets
319(9)
9.1.1 Physics of Failure Prognostic Models
321(1)
9.1.2 Physics of Failure Procedure
321(3)
9.1.3 Physics of Failure and Its Role in Maintenance
324(3)
9.1.3.1 Predictive Maintenance and Prognostics
324(1)
9.1.3.2 Condition-Based Maintenance
325(2)
9.1.3.3 Root Cause Analysis
327(1)
9.1.4 Advantages of the PoF Approach in Reliability Engineering
327(1)
9.2 Reliability Estimation and Prediction
328(12)
9.2.1 Reliability Estimation
328(2)
9.2.1.1 Reliability Estimation Based on Event Data
329(1)
9.2.1.2 Reliability Estimation Based on Condition Monitoring Data
329(1)
9.2.2 Prognostics and Reliability
330(3)
9.2.2.1 From Maintenance to Prognostics
331(1)
9.2.2.2 From Prognostics to Predictions
331(1)
9.2.2.3 From Prediction to Reliability
332(1)
9.2.3 Reliability Prediction
333(7)
9.2.3.1 Role of Reliability Prediction
333(1)
9.2.3.2 Basic Concepts of Reliability Prediction
334(1)
9.2.3.3 Reliability Prediction Methods
335(1)
9.2.3.4 Reliability Prediction Definitions
336(2)
9.2.3.5 Types of Reliability Prediction
338(2)
9.2.3.6 Need for an Effective Approach in Reliability Prediction
340(1)
9.3 Sensing Technologies in Dynamic Assets and Failure Diagnosis
340(8)
9.3.1 Fault Sensing and Diagnosis
341(3)
9.3.1.1 Overview of Approaches to Fault Sensing and Diagnosis
343(1)
9.3.2 Fault Management Mechanism for Wireless Sensor Networks
344(2)
9.3.2.1 Fault Detection
344(2)
9.3.2.2 Fault Diagnosis
346(1)
9.3.3 Multi-Sensor Measurement and Data Fusion Technology
346(2)
9.4 Proportional Hazards Model and Physical Stressors
348(4)
9.4.1 Proportional Hazards Model with Time-Dependent Covariates
348(1)
9.4.2 Proportional Hazards Model
348(1)
9.4.3 Proportional Hazards Model Assumption
349(1)
9.4.4 Properties and Applications of the Proportional Hazards Model
350(1)
9.4.5 Cox Regression
350(2)
9.5 Hybrid Models for Dynamic and Non-Dynamic Assets
352(9)
9.5.1 Hybrid Models
352(1)
9.5.1.1 Series Approach
352(1)
9.5.1.2 Parallel Approach
352(1)
9.5.2 Hybrid Approach
352(2)
9.5.3 Combination Models
354(7)
Chapter 10 Principles of Digital Twin 361(46)
10.1 Principles of Digital Twin
361(14)
10.1.1 Introduction
361(1)
10.1.2 Why Digital Twins Matter
361(1)
10.1.3 How Digital Twins Work
362(1)
10.1.4 Value of Digital Twins
363(2)
10.1.5 Intrinsic Characteristics of Digital Twins
365(1)
10.1.6 Digital Twins: What, Why, and How?
365(2)
10.1.6.1 The "Why" Perspective
365(2)
10.1.6.2 The "How" Perspective
367(1)
10.1.7 A Digital Twin Example: SAP Digital Twin for Wind Power
367(1)
10.1.8 Digital Twin Origin: Physics and Simulation
368(1)
10.1.9 Use of Digital Twin in Operations
368(1)
10.1.9.1 Point Machine for Train Switches
368(1)
10.1.9.2 Planning the Digital Twin
369(1)
10.1.9.3 Digital Twin during Operation Phase
369(1)
10.1.9.4 Hybrid Analysis and Fleet Data
369(1)
10.1.10 Digital Twin Reference Model
369(4)
10.1.11 Physical Model-Based Digital Twins
373(2)
10.1.11.1 Model-Based Control
374(1)
10.2 Functional Mock-Up for Complex System Assembly
375(8)
10.2.1 Defining Complex Systems
375(1)
10.2.2 Complex Systems and Associated Problems
375(1)
10.2.3 Functional Mock-Up Interface
376(1)
10.2.3.1 Functional Mock-Up Interface for Model Exchange and Co-Simulation
376(1)
10.2.4 The Use of FMUs for the Digital Twin
377(1)
10.2.5 Objectives of FMI applied to Product Life Method
378(2)
10.2.6 Functions of Product Life Method
380(3)
10.2.6.1 Summary of PLM Functions
380(1)
10.2.6.2 Network Description
380(1)
10.2.6.3 Deployment Description
381(2)
10.3 Integration of Low-Level Digital Twins
383(15)
10.3.1 Digital Twin: Toward an Evaluation Framework
383(1)
10.3.2 Current Technologies Deployed in Digital Twin
384(14)
10.3.2.1 Industrial Internet of Things (IIoT) and Digital Twin
385(2)
10.3.2.2 Digital Twin, Cyber-Physical System, and Internet of Things
387(1)
10.3.2.3 Enabling Technologies
388(3)
10.3.2.4 Digital Twin and Simulation
391(1)
10.3.2.5 Machine Learning and Digital Twin
391(1)
10.3.2.6 Augmented and Virtual Reality and Digital Twin
392(1)
10.3.2.7 Cloud Technology and Digital Twin
392(1)
10.3.2.8 Extending the Relevance of Predictive Maintenance
393(1)
10.3.2.9 Operational Process Digital Twin: Diagnostic and Control Capability
393(1)
10.3.2.10 Operational Process Digital Twin: Predictive Capability
394(1)
10.3.2.11 Gamify Decision-Making with Prescriptive Analytics
394(1)
10.3.2.12 Enterprise Digital Twins
395(1)
10.3.2.13 Digital Twin Approach to Predictive Maintenance
396(2)
10.4 Failure Forecasting at the System Level
398(9)
10.4.1 Failure Forecasting
398(9)
10.4.1.1 Anomaly Detection and Analysis
398(1)
10.4.1.2 Failure Prediction
399(2)
10.4.1.3 Anomaly Detection Solutions for Predictive Maintenance of Industrial Equipment
401(6)
Chapter 11 Application of Prognosis in Industry, Energy, and Transportation 407(42)
11.1 Mechanical Systems: Maintenance Activities in Automotive and Railway Sectors, Aircraft Applications, Rotating Equipment (Bearings, Pumps, Gearboxes, Motors)
407(26)
11.1.1 Mechanical Systems: Maintenance Activities in Automotive Sector
407(5)
11.1.1.1 Current Composition of the Automotive Industry
407(1)
11.1.1.2 Technological Change, Skills, and Changing Job Roles
407(1)
11.1.1.3 Achievement of Reliability by Maintenance Activities and Tools in the Automotive Sector
408(3)
11.1.1.4 Total Productive Maintenance in Automotive Industry
411(1)
11.1.2 Mechanical Systems: Maintenance Activities in Railway Sector
412(2)
11.1.2.1 Scheduling Preventive Railway Maintenance Activities
412(1)
11.1.2.2 Current Maintenance Challenges in Railway Industry
413(1)
11.1.2.3 How to Implement Efficient Railway Maintenance through Digitalization
414(1)
11.1.3 Mechanical Systems: Maintenance Activities in Aircraft Applications
414(9)
11.1.3.1 Aircraft Maintenance and Repair
414(3)
11.1.3.2 Aircraft Maintenance Operations
417(2)
11.1.3.3 Aircraft Servicing, Maintenance, Repair, and Overhaul: Changed Scenarios through Outsourcing
419(4)
11.1.4 Mechanical Systems: Maintenance Activities in Rotating Equipment (Bearings, Pumps, Gearboxes, Motors)
423(10)
11.1.4.1 Maintenance Activities in Bearings
423(2)
11.1.4.2 Maintenance Activities in Pumps
425(3)
11.1.4.3 Maintenance Activities in Gearboxes
428(2)
11.1.4.4 Maintenance Activities in Motors
430(3)
11.2 Industrial Enterprises: Chemical, Continuous-Time Production Processes
433(6)
11.2.1 Continuous Production
433(2)
11.2.1.1 Characteristics of Continuous Production
433(1)
11.2.1.2 Types of Continuous Production
433(1)
11.2.1.3 When Is Continuous Production Suitable?
434(1)
11.2.2 Future Production Concepts in the Chemical Industry
435(1)
11.2.2.1 Traditional Batch Processing vs. Continuous Production Methods
435(1)
11.2.2.2 Continuous Manufacturing vs. Modularized Plant Systems
435(1)
11.2.3 Continuous Manufacturing in Pharmaceutical and Chemical Industries
436(3)
11.2.3.1 Factors behind the Rising Momentum toward Continuous Manufacturing
436(1)
11.2.3.2 What Is Continuous Manufacturing?
436(3)
11.3 Marine Systems: Shipboard Machinery and Logistics Maintenance
439(5)
11.3.1 Shipboard Machinery Maintenance
439(1)
11.3.2 Maintenance and Repair of Shipboard Machinery and Equipment
440(1)
11.3.3 Logistics Maintenance
440(4)
11.4 Medical: Hospital 4.0
444(5)
11.4.1 Hospital 4.0: Digital Transformation in Hospitals
445(4)
Index 449
Dr. Diego Galar is Full Professor of Condition Monitoring in the Division of Operation and Maintenance Engineering at LTU, Luleå University of Technology where he is coordinating several H2020 projects related to different aspects of cyber physical systems, Industry 4.0, IoT or Industrial AI and Big Data. He was also involved in the SKF UTC centre located in Lulea focused on SMART bearings and also actively involved in national projects with the Swedish industry or funded by Swedish national agencies like Vinnova. He is also principal researcher in Tecnalia (Spain), heading the Maintenance and Reliability research group within the Division of Industry and Transport. He has authored more than five hundred journal and conference papers, books and technical reports in the field of maintenance, working also as member of editorial boards, scientific committees and chairing international journals and conferences and actively participating in national and international committees for standardization and R&D in the topics of reliability and maintenance. In the international arena, he has been visiting Professor in the Polytechnic of Braganza (Portugal), University of Valencia and NIU (USA) and the Universidad Pontificia Católica de Chile. Currently, he is visiting professor in University of Sunderland (UK), University of Maryland (USA), and Chongqing University in China.

Dr. Kai Goebel is a Principal Scientist in the System Sciences Lab at Palo Alto Research Center (PARC). His interest is broadly in condition-based maintenance and systems health management for a broad spectrum of cyber-physical systems in the transportation, energy, aerospace, defense, and manufacturing sectors. Prior to joining PARC, Dr. Goebel worked at NASA Ames Research Center and General Electric Corporate Research & Development center. At NASA, he was a branch chief leading the Discovery and Systems Health tech area which included groups for machine learning, quantum computing, physics modeling, and diagnostics & prognostics. He founded and directed the Prognostics Center of Excellence which advanced our understanding of the fundamental aspects of prognostics. He holds 18 patents and has published more than 350 papers, including a book on Prognostics. Dr. Goebel was an adjunct professor at Rensselaer Polytechnic Institute and is now adjunct professor at Lulea Technical University. He is a co-founder of the Prognostics and Health Management Society, and associate editor of the International Journal of PHM.

Peter Sandborn is a Professor in the CALCE Electronic Products and Systems Center and the Director of the Maryland Technology Enterprise Institute at the University of Maryland. Dr. Sandborns group develops life-cycle cost models and business case support for long field life systems. This work includes: obsolescence forecasting algorithms, strategic design refresh planning, lifetime buy quantity optimization, return on investment models for maintenance planning and system health management, and outcome-based contract design and optimization. Dr. Sandborn is the developer of the MOCA refresh planning tool. Dr. Sandborn is an Associate Editor of the IEEE Transactions on Electronics Packaging Manufacturing and a member of the Board of Directors of the PHM Society. He is the author of over 200 technical publications and several books on electronic packaging and electronic systems cost analysis. He was the winner of the 2004 SOLE Proceedings, the 2006 Eugene L. Grant, the 2017 ASME Kos Ishii-Toshiba, and the 2018 Jacques S. Gansler awards. He has a B.S. degree in engineering physics from the University of Colorado, Boulder, in 1982, and the M.S. degree in electrical science and Ph.D. degree in electrical engineering, both from the University of Michigan, Ann Arbor, in 1983 and 1987, respectively. He is a Fellow of the IEEE, the ASME and the PHM Society.

Dr. Uday Kumar is the Chair Professor of Operation and Maintenance Engineering, Director of Research and Innovation (Sustainable Transport) at Luleå University of Technology and Director of Luleå Railway Research Center.

His teaching, research & consulting interests are equipment maintenance, reliability and maintainability analysis, product support, Life Cycle Costing(LCC) , Risk analysis, system analysis, eMaintenance, asset management etc. He is visiting faculty at the Center of Intelligent Maintenance System(IMS)- a centre sponsored by National Science Foundation , Cincinnati, USA since 2011, External examiner and Program Reviewer for Reliability and Asset Management Program of The University of Manchester, Distinguished visiting Professor at Tsinghua University Beijing, Honorary Professor at Beijing Jiaotong University, Beijing, etc. Earlier he has been visiting faculty at Imperial College London, Helsinki University of Technology, Helsinki, Univ of Stavanger , Norway, etc. He has more than 30 years of experiences in consulting and finding solutions to industrial problems directly or indirectly related to maintenance of engineering asserts. He has published more than 300 papers in International Journals and Conference Proceedings dealing with various aspects of maintenance of engineering systems, and has co-authored 4 books on Maintenance Engineering and contributed to World Encyclopaedia on Risk Management. He is an elected member of Royal Swedish Academy of Engineering Sciences.