Prognostics and Health Management of Electronics. Fundamentals, Machine Learning, and Internet of Things. 2nd Edition. Wiley - IEEE

  • ID: 4471935
  • Book
  • 750 Pages
  • John Wiley and Sons Ltd
1 of 4

An indispensable guide for engineers and data scientists in design, testing, operation, manufacturing, and maintenance

A road map to the current challenges and available opportunities for the research and development of Prognostics and Health Management (PHM), this important work covers all areas of electronics and explains how to:

  • assess methods for damage estimation of components and systems due to field loading conditions
  • assess the cost and benefits of prognostic implementations
  • develop novel methods for in situ monitoring of products and systems in actual life–cycle conditions
  • enable condition–based (predictive) maintenance
  • increase system availability through an extension of maintenance cycles and/or timely repair actions;
  • obtain knowledge of load history for future design, qualification, and root cause analysis
  • reduce the occurrence of no fault found (NFF)
  • subtract life–cycle costs of equipment from reduction in inspection costs, downtime, and inventory

Prognostics and Health Management of Electronics also explains how to understand statistical techniques and machine learning methods used for diagnostics and prognostics. Using this valuable resource, electrical engineers, data scientists, and design engineers will be able to fully grasp the synergy between IoT, machine learning, and risk assessment. 

Note: Product cover images may vary from those shown
2 of 4




Chapter 1 Introduction to PHM

1.1 Reliability and Prognostics

1.2 PHM of Electronics

1.3 PHM Approaches

1.3.1 PoF–based Approach

1.3.2 Canaries

1.3.3 Data–Driven Approach

1.3.4 Fusion Approach

1.4 Implementation of PHM in a System of Systems

1.5 PHM in the Internet of Things (IoT) Era

1.5.1 IoT–Enabled PHM Applications: Manufacturing

1.5.2 IoT–Enabled PHM Applications: Energy Generation

1.5.3 IoT–Enabled PHM Applications: Transportation and Logistics

1.5.4 IoT–Enabled PHM Applications: Automobiles

1.5.5 IoT–Enabled PHM Applications: Medical Consumer Products

1.5.6 IoT–Enabled PHM Applications: Warranty Services

1.5.7 IoT–Enabled PHM Applications: Robotics



Chapter 2 Sensor Systems for PHM

2.1 Sensor and Sensing Principles

2.1.1 Thermal Sensors

2.1.2 Electrical Sensors

2.1.3 Mechanical Sensors

2.1.4 Chemical Sensors

2.1.5 Humidity Sensors

2.1.6 Biosensors

2.1.7 Optical Sensors

2.1.8 Magnetic Sensors

2.2 Sensor Systems for PHM

2.2.1 Parameters to Be Monitored

2.2.2 Sensor System Performance

2.2.3 Physical Attributes of Sensor Systems

2.2.4 Functional Attributes of Sensor Systems

2.2.5 Reliability

2.2.6 Availability

2.2.7 Cost

2.3 Sensor Selection

2.4 Examples of Sensor Systems for PHM Implementation

2.5 Emerging Trends in Sensor Technology for PHM

2.6 References

Chapter 3 Physics–of–Failure Approach to PHM

3.1 PoF–based PHM Methodology

3.2 Hardware Configuration

3.3 Loads

3.4 Failure Modes, Mechanisms, and Effects Analysis

3.4.1 Examples of FMMEA for Electronic Devices

3.5 Stress Analysis

3.6 Reliability Assessment and Remaining–Life Predictions

3.7 Outputs from PoF–based PHM

3.8 Caution and Concerns in the Use of PoF–based PHM

3.9 Combining PoF with Data–Driven Prognosis

3.10 References

Chapter 4 Machine Learning: Fundamentals

4.1 Types of Machine Learning

4.1.1 Supervised, Unsupervised, Semi–supervised, and Reinforcement Learning

4.1.2 Batch and Online Learning

4.1.3 Instance–based and Model–based Learning

4.2 Probability Theory in Machine Learning: Fundamentals

4.2.1 Probability Space and Random Variables

4.2.2 Distributions, Joint Distributions, and Marginal Distributions

4.2.3 Conditional Distributions

4.2.4 Independence

4.2.5 Chain Rule and Bayes Rule

4.3 Probability Mass Function and Probability Density Function

4.3.1 Probability Mass Function

4.3.2 Probability Density Function

4.4 Mean, Variance, and Covariance Estimation

4.4.1 Mean

4.4.2 Variance

4.4.3 Robust Covariance Estimation

4.5 Probability Distributions

4.5.1 Bernoulli Distribution

4.5.2 Normal Distribution

4.5.3 Uniform Distribution

4.6 Maximum Likelihood and Maximum a Posteriori Estimation

4.6.1 Maximum Likelihood Estimation

4.6.2 Maximum a Posteriori Estimation

4.7 Correlation and Causation

4.8 Kernel Trick

4.9 Performance Metrics

4.9.1 Diagnostic Metrics

4.9.2 Prognostic Metrics

4.10 References

Chapter 5 Machine Learning: Data Pre–processing

5.1 Data Cleaning

5.1.1 Missing Data Handling

5.2 Feature Scaling

5.3 Feature Engineering

5.3.1 Feature Extraction

5.3.2 Feature Selection

5.4 Imbalanced Data Handling

5.4.1 Sampling Methods for Imbalanced Learning

5.4.2 Effect of Sampling Methods for Diagnosis

5.5 References

Chapter 6 Machine Learning: Anomaly Detection

6.1 Introduction

6.2 Types of Anomalies

6.2.1 Point Anomalies

6.2.2 Contextual Anomalies

6.2.3 Collective Anomalies

6.3 Distance–based Methods

6.3.1 MD Calculation Using an Inverse Matrix Method

6.3.2 MD Calculation Using a Gram–Schmidt Orthogonalization Method

6.3.3 Decision Rules

6.4 Clustering–based Methods

6.4.1 K–Means Clustering

6.4.2 Fuzzy C–Means Clustering

6.4.3 Self–Organizing Maps

6.5 Classification–based Methods

6.5.1 One–class Classification

6.5.2 Multi–class Classification

6.6 Statistical Methods

6.6.1 Sequential Probability Ratio Test

6.6.2 Correlation Analysis

6.7 Anomaly Detection with No System s Health Profile

6.8 Challenges in Anomaly Detection

6.9 References

Chapter 7 Machine Learning: Diagnostics and Prognostics

7.1 Overview of Diagnosis and Prognosis

7.2 Techniques for Diagnostics

7.2.1 Supervised Machine Learning Algorithms

7.2.2 Ensemble Learning

7.2.3 Deep Learning

7.3 Techniques for Prognostics

7.3.1 Regression Analysis

7.3.2 Particle Filtering

7.4 References

Chapter 8 Uncertainty Representation, Quantification, and Management in Prognostics

8.1 Introduction

8.2 Sources of Uncertainty in PHM

8.3 Formal Treatment of Uncertainty in PHM

8.3.1 Problem 1: Uncertainty Representation and Interpretation

8.3.2 Problem 2: Uncertainty Quantification

8.3.3 Problem 3: Uncertainty Propagation

8.3.4 Problem 4: Uncertainty Management

8.4 Uncertainty Representation and Interpretation

8.4.1 Physical Probabilities and Testing–based Prediction

8.4.2 Subjective Probabilities and Condition–based Prognostics

8.4.3 Why is RUL Prediction Uncertain?

8.5 Uncertainty Quantification and Propagation for RUL Prediction

8.5.1 Computational Framework for Uncertainty Quantification

8.5.2 RUL Prediction: An Uncertainty Propagation Problem

8.5.3 Uncertainty Propagation Methods

8.6 Uncertainty Management

8.7 Case Study: Uncertainty Quantification in the Power System of an Unmanned Aerial Vehicle

8.7.1 Description of the Model

8.7.2 Sources of Uncertainty

8.7.3 Results: Constant Amplitude Loading Conditions

8.7.4 Results: Variable Amplitude Loading Conditions

8.7.5 Discussion

8.8 Existing Challenges

8.8.1 Timely Predictions

8.8.2 Uncertainty Characterization

8.8.3 Uncertainty Propagation

8.8.4 Capturing Distribution Properties

8.8.5 Accuracy

8.8.6 Uncertainty Bounds

8.8.7 Deterministic Calculations

8.9 Summary

8.10 References

Chapter 9 PHM Cost and Return on Investment

9.1 Return on Investment

9.1.1 PHM ROI Analysis

9.1.2 Financial Costs

9.2 PHM Cost–Modeling Terminology and Definitions

9.3 PHM Implementation Costs

9.3.1 Nonrecurring Costs

9.3.2 Recurring Costs

9.3.3 Infrastructure Costs

9.3.4 Nonmonetary Consideration and Maintenance Culture

9.4 Cost Avoidance

9.4.1 Maintenance Planning Cost Avoidance

9.4.2 Discrete–Event Simulation Maintenance Planning Model

9.4.3 Fixed–Schedule Maintenance Interval

9.4.4 Data–Driven (Precursor to Failure Monitoring) Methods

9.4.5 Model–Based (LRU–Independent) Methods

9.4.6 Discrete–Event Simulation Implementation Details

9.4.7 Operational Profile

9.5 Example PHM Cost Analysis

9.5.1 Single–Socket Model Results

9.5.2 Multiple–Socket Model Results

9.6 Example Business Case Construction: Analysis for ROI

9.7 Summary

9.8 References

Chapter 10 Valuation and Optimization of PHM–Enabled Maintenance Decisions

10.1 Valuation and Optimization of PHM–Enabled Maintenance Decisions for an Individual System

10.1.1 A PHM–Enabled Predictive Maintenance Optimization Model for an Individual System

10.1.2 Case Study: Optimization of PHM–Enabled Maintenance Decisions for an Individual

System (Wind Turbine)

10.2 Availability

10.2.1 The Business of Availability: Outcome–Based Contracts

10.2.2 Incorporating Contract Terms into Maintenance Decisions

10.2.2 Case Study: Optimization of PHM–Enabled Maintenance Decisions for Systems (Wind Farm)

10.3 Future Directions

10.3.1 Design for Availability

10.3.2 Prognostics–Based Warranties

10.3.3 Contract Engineering

10.4 References

Chapter 11 Health and Remaining Useful Life Estimation of Electronic Circuits

11.1 Introduction

11.2 Related Work

11.2.1 Component–Centric Approach

11.2.2 Circuit–Centric Approach

11.3 Electronic Circuit Health Estimation Through Kernel Learning

11.3.1 Kernel–Based Learning

11.3.2 Health Estimation Method

11.3.3 Implementation Results

11.4 RUL Prediction Using Model–Based Filtering

11.4.1 Prognostics Problem Formulation

11.4.2 Circuit Degradation Modeling

11.4.3 Model–Based Prognostic Methodology

11.4.4 Implementation Results

11.5 Summary

11.6 References

Chapter 12 PHM–based Qualification of Electronics

12.1 Why is Product Qualification Important?

12.2 Considerations for Product Qualification

12.3 Review of Current Qualification Methodologies

12.3.1 Standards–Based Qualification

12.3.2 Knowledge–Based or Physics–of–Failure–Based Qualification

12.3.3 Prognostics and Health Management–Based Qualification

12.4 Summary

12.3 References

Chapter 13PHM of Li–ion Batteries

13.1 Introduction

13.2 State of Charge Estimation

13.2.1 Case Study I: SOC Estimation

13.2.2 Case Study II: SOC Estimation

13.3 State of Health Estimation and Prognostics

13.3.1 Case Study for Li–ion Battery Prognostics

13.4 Summary

13.5 References

Chapter 14 PHM of Light–Emitting Diodes

14.1 Introduction

14.2 Review of PHM Methodologies for LEDs

14.2.1 Overview of Available Prognostic Methods

14.2.2 Data–Driven Methods

14.2.3 Physics–Based Methods

14.2.4 LED System–Level Prognostics

14.3 Simulation–Based Modeling and Failure Analysis for LEDs

14.3.1 LED Chip–Level Modeling and Failure Analysis

14.3.2 LED Package–Level Modeling and Failure Analysis

14.3.3 LED System–Level Modeling and Failure Analysis

14.4 Return–on–Investment Analysis of Applying Health Monitoring to LED Lighting Systems

14.4.1 ROI Methodology

14.4.2 ROI Analysis of Applying System Health Monitoring to LED Lighting Systems

14.5 Summary

14.6 References

Chapter 15 PHM of Healthcare

15.1 Healthcare in the U.S.

15.2 Considerations in Healthcare

15.2.1 Clinical Considerations in Implantable Medical Devices

15.2.2 Considerations in Care Bots

15.3 Benefits of PHM

15.3.1 Safety Increase

15.3.2 Operational Reliability Improvement

15.3.3 Mission Availability Increase

15.3.4 System s Service Life Extension

15.3.5 Maintenance Effectiveness Increase

15.4 PHM of Implantable Medical Devices

15.5 PHM of Care Bots

15.6 Canary–based Prognostics of Healthcare Devices

15.7 Summary

15.8 References

Chapter 16 PHM of Subsea Cables

16.1 Subsea Cable Market

16.2 Subsea Cables

16.3 Cable Failures

16.3.1 Internal Failures

16.3.2 Early–Stage Failures

16.3.3 External Failures

16.3.4 Environmental Conditions

16.3.5 Third–Party Damage

16.4 State–of–the–Art Monitoring

16.5 Qualifying and Maintaining Subsea Cables

16.5.1 Qualifying Subsea Cables

16.5.2 Mechanical Tests

16.5.3 Maintaining Subsea Cables

16.6 Data–Gathering Techniques

16.7 Measuring the Wear Behavior of Cable Materials

16.8 Predicting Cable Movement

16.8.1 Sliding Distance Deviation

16.8.2 Scouring Depth Calculations

16.9 Predicting Cable Degradation

16.9.1 Volume Loss Due to Abrasion

16.9.2 Volume Loss Due to Corrosion

16.10 Predicting Remaining Useful Life

16.11 Case Study

16.12 Future Challenges

16.12.1 Data–Driven Approach for Random Failures

16.12.2 Model–Driven Approach for Environmental Failures

16.13 Summary

16.14 References

Chapter 17 Connected Vehicle Diagnostics and Prognostics

17.1 Introduction

17.2 Design of an Automatic Field Data Analyzer

17.2.1 Data Collection Subsystem

17.2.2 Information Abstraction Subsystem

17.2.3 Root Cause Analysis Subsystem

17.3 Case Study: CVDP for Vehicle Batteries

17.3.1 Brief Background of Vehicle Batteries

17.3.2 Applying AFDA for Vehicle Batteries

17.3.3 Experimental Results

17.4 Summary

17.5 References

Chapter 18 The Role of PHM at Commercial Airlines

18.1 Evolution of Aviation Maintenance

18.2 Stakeholder Expectations for PHM

18.2.1 Passenger Expectations

18.2.1 Airline/Operator/Owner Expectations

18.2.2 Airframe Manufacturer Expectations

18.2.3 Engine Manufacturer Expectations

18.2.4 System and Component Supplier Expectations

18.2.5 MRO Organization Expectations

18.3 PHM Implementations

18.3.1 SATAA

18.4 PHM Applications

18.4.1 Engine Health Management

18.4.2 Auxiliary Power Unit (APU) Health Management

18.4.3 Environmental Control System (ECS) and Air Distribution Health Monitoring

18.4.4 Landing Systems Health Monitoring

18.4.5 Liquid Cooling Systems Health Monitoring

18.4.6 Nitrogen Generation System (NGS) Health Monitoring

18.4.7 Fuel Consumption Monitoring

18.4.8 Flight Control Systems Health Monitoring

18.4.9 Electric Power Systems Health Monitoring

18.4.10 Structural Health Monitoring (SHM)

18.4.11 Fuel and Hydraulic Systems Health Management

18.5 Summary

18.6 References

Chapter 19 PHM Software of Electronics

19.1 PHM Software: CALCE Simulation Assisted Reliability Assessment

19.2 PHM Software: Data–Driven

19.1.1 Data Flow

19.1.2 Master Options

19.1.3 Data Pre–processing

19.1.4 Feature Discovery

19.1.5 Anomaly Detection

19.1.6 Diagnostics/Classification

19.1.7 Prognostics/Modeling

19.1.8 Challenges in Data–Driven PHM Software Development

19.3 Summary

Chapter 20 eMaintenance

20.1 From Reactive to Proactive Maintenance

20.2 The Onset of eMaintenance

20.3 Maintenance Management System

20.3.1 Lifecycle Management

20.3.2 eMaintenance Architecture

20.4 Sensor Systems

20.5 Data Analysis

20.6 Predictive Maintenance

20.7 Maintenance Analytics

20.7.1 Maintenance Descriptive Analytics

20.7.2 Maintenance Analytics and eMaintenance

20.7.3 Maintenance Analytics and Big Data

20.8 Knowledge Discovery

20.9 Integrated Knowledge Discovery

20.10 User Interface for Decision Support

20.11 Applications of eMaintenance

20.11.1 eMaintenance in Railways

20.11.2 eMaintenance in Manufacturing

20.11.3 MEMS Sensors for Bearing Vibration Measurement

20.11.4 Wireless Sensors for Temperature Measurement

20.11.5 Monitoring Systems

20.11.6 eMaintenance Cloud and Servers

20.11.7 Dashboard Managers

20.11.8 Alarm Servers

20.11.9 Cloud Services

20.11.10  Graphic User Interfaces

20.12 Internet Technology and Optimizing Technology

20.13 References

Chapter 21 Predictive Maintenance in the IoT Era

21.1 Background

21.1.1 Challenges of a Maintenance Program

21.1.2 Evolution of Maintenance Paradigms

21.1.3 Preventive vs. Predictive Maintenance

21.1.4 P–F Curve

21.1.5 Bathtub Curve

21.2 Benefits of a Predictive Maintenance Program

21.3 Prognostic Model Selection for Predictive Maintenance

21.4 Internet of Things

21.4.1 Industrial IoT

21.5 Predictive Maintenance Based on IoT

21.6 Predictive Maintenance Usage Cases

21.7 Machine Learning Techniques for Data–Driven Predictive Maintenance

21.7.1 Supervised Learning

21.7.2 Unsupervised Learning

21.7.3 Anomaly Detection

21.7.4 Multiclass and Binary Classification Models

21.7.5 Regression Models

21.7.6 Survival Models

21.8 Best Practices

21.8.1 Define Business Problem and Quantitative Metrics

21.8.2 Identify Assets and Data Sources

21.8.3 Data Acquisition and Transformation

21.8.4 Build Models

21.8.5 Model Selection

21.8.6 Predict Outcomes and Transform into Process Insights

21.8.7 Operationalize and Deploy

21.8.8 Continuous Monitoring

21.9 Challenges in a Successful Predictive Maintenance Program

21.9.1 Predictive Maintenance Management Success Key Performance Indicators (KPIs)

21.10 Summary

21.11 References

Chapter 22 Analysis of PHM Patents for Electronics

22.1 Introduction

22.2 Analysis of PHM Patents for Electrical Systems

22.2.1 Sources of PHM Patents

22.2.2 Analysis of PHM Patents

22.3 Trend of PHM Activities for Electrical Systems in Industries

22.2.1 Sources of PHM Patents

22.2.2 Batteries

22.2.3 Electric Motors

22.2.4 Circuits and Systems

22.2.5 Electrical Devices in Automobiles and Airplanes

22.2.6 Networks and Communication Facilities

22.2.7 Others

22.4 Trend of PHM Activities for Electrical Systems in Academia

22.5 Gap of Viewpoint on PHM between Industries and Academia

22.6 Summary

22.7 References

Chapter 23 A PHM Roadmap for Electronics–Rich Systems

23.1 Introduction

23.2 Roadmap Classifications

23.2.1 Component–level PHM

23.2.2 System–level PHM

23.3 Methodology Development

23.3.1 Best Algorithms

23.3.2 Verification and Validation

23.3.3 Log–Term PHM Studies

23.4 PHM for Storage

23.5 PHM for No–Fault Found/Intermittent Failures

23.6 PHM for Products Subjected to Indeterminate Operating Conditions

23.7 Nontechnical Barriers

23.7.1 Cost, ROI, Business Case Development

23.7.2 Liability and Litigation

23.7.3 Maintenance Culture

23.7.4 Contract Structure

23.7.5 Role of Standards Organizations

23.7.6 Licensing and Entitlement Management

23.8 References

Appendix ACommercially Available Sensor Systems for PHM

Appendix BJournals and Conference Proceedings Related to PHM

Appendix C Glossary of Terms and Definitions


Note: Product cover images may vary from those shown
3 of 4


4 of 4

Michael G. Pecht, PhD, is Chair Professor in Mechanical Engineering and Professor in Applied Mathematics, Statistics and Scientific Computation at the University of Maryland. He is the Founder and Director of the Center for Advanced Life Cycle Engineering (CALCE) at the University of Maryland, which is funded by more than 150 leading electronics companies. Dr. Pecht is an IEEE, ASME, SAE, and IMAPS Fellow and serves as editor–in–chief of IEEE Access. He has written more than 30 books, 700 technical articles, and has 8 patents.

Myeongsu Kang, PhD, is currently a Research Associate at the Center for Advanced Life Cycle Engineering (CALCE), University of Maryland, College Park, MD, USA. His expertise is in data analytics, machine learning, system modeling, and statistics for prognostics and systems health management. He has authored/coauthored more than 60 publications in leading journals and conference proceedings.

Note: Product cover images may vary from those shown
5 of 4
Note: Product cover images may vary from those shown