The Engineering of Digital Twins

Preface
Contents
List of Contributing Authors
Acronyms
Part I Foundations
	Chapter 1 Engineering Digital Twins for Cyber-Physical Systems
		1.1 Introduction
		1.2 Cyber-Physical Systems and Digital Twins
			1.2.1 Cyber-Physical Systems
			1.2.2 DTs of Cyber-Physical Systems
		1.3 Aspects of DT Engineering
			1.3.1 Digital Twins are Model-Centric
			1.3.2 Inside the Digital Twin
			1.3.3 Fields Related to Digital Twins
			1.3.4 Emerging Digital Twin Standards
		1.4 The Transition to Digital Twins
			1.4.1 Digital Twins for Existing Physical Products
			1.4.2 A Federated Future
		References
	Chapter 2 The Potential of Digital Twins: Four Industry Perspectives
		2.1 Round Table Discussion Structure
		2.2 Introductions
		2.3 Businesses
		2.4 Where are you thinking of targeting DT technology?
		2.5 What does success look like?
		2.6 Why Digital Twins?
		2.7 Stakeholders, Developers and Users
		2.8 How would you expect to develop DTs?
		2.9 Do DTs help Dependability?
		2.10 Themes
			2.10.1 The DT Life Cycle
			2.10.2 Model-Centric Digital Twins
			2.10.3 Getting Data from the Physical Twin
			2.10.4 Services Supported in Digital Twins
			2.10.5 Further Research in needed on Digital Twins
		References
	Chapter 3 Foundational Concepts for Digital Twins of Cyber-Physical Systems
		3.1 Introduction
		3.2 Running Example: the Tempeh Incubator System
			3.2.1 Tempeh and how to Make it
			3.2.2 Tempeh Production
			3.2.3 DT-enhanced Tempeh Incubation
		3.3 Basic System Concepts
		3.4 Models & Data
		3.5 Digital Twin Services
		3.6 Digital Twin Assets and Management
			3.6.1 Physical Entities
			3.6.2 Data and Data Management
			3.6.3 Models and Model Management
			3.6.4 DT Services
		References
	Chapter 4 Digital Twin Engineering Processes
		4.1 Introduction
		4.2 DT Engineering as Systems Engineering
		4.3 Stakeholders’ Expectations, Needs and Requirements Processes
			4.3.1 Business or Mission Analysis
			4.3.2 Defining Stakeholder Needs and Requirements
		4.4 System Requirements and Architecture Processes
			4.4.1 Defining System Requirements
			4.4.2 System Architecture Definition
		4.5 Realisation Processes
			4.5.1 Design, Systems Analysis and Implementation
			4.5.2 Integration
			4.5.3 Verification
			4.5.4 Transition
			4.5.5 Validation
		4.6 The DT-Enabled System in Operation
			4.6.1 Operation
			4.6.2 Maintenance and Disposal
		4.7 Tailoring Processes and Teams
		4.8 Processes and Competencies
			4.8.1 Agreement, Organisation and Management Processes
			4.8.2 Competencies and Roles in DT Engineering
		References
Part II Models and Data
	Chapter 5 Modelling for Digital Twins
		5.1 Introduction
		5.2 Overview of Modelling Formalisms
		5.3 Models for the Incubator Example
		5.4 Physics-based Models
			5.4.1 Ordinary Differential Equations
			5.4.2 Partial Differential Equations
			5.4.3 Model Order Reduction
		5.5 Data-driven Models
			5.5.1 Requirements of Data-driven Modelling
			5.5.2 Artificial Neural Networks
			5.5.3 Learning Methods
				5.5.3.1 Supervised Learning
				5.5.3.2 Unsupervised Learning
				5.5.3.3 Reinforcement Learning
		5.6 Models for Computer-Based Systems
			5.6.1 Finite State Machines
			5.6.2 Vienna Development Method
			5.6.3 Verification Methods
		5.7 Coupling of Heterogeneous Models
			5.7.1 Co-Simulation
			5.7.2 Hybrid Automata
		References
	Chapter 6 Calibration of Models for Digital Twins
		6.1 Introduction
		6.2 What is Calibration?
		6.3 Calibration of Linear Algebraic Models
			6.3.1 Calibration Under Noisy Measurements
		6.4 Calibration of Non-Linear Algebraic Models
			6.4.1 Gradient descent
			6.4.2 Alternative Non-Linear Optimisation Methods
				6.4.2.1 Newton Method
				6.4.2.2 The Gauss-Newton Method
			6.4.3 Calibration of Differential Equations
			6.4.4 Genetic Algorithms and Design Space Exploration
		6.5 Practical Considerations
		References
	Chapter 7 Sensing and Communication of Data from the Physical Twin
		7.1 Introduction
		7.2 Sensors and Their Limits
			7.2.1 Limited Sampling Frequency
			7.2.2 Quantisation
				7.2.2.1 Analog Signal Compression as Non-Uniform Quantisation
				7.2.2.2 Dynamic Quantisation
			7.2.3 Immeasurable Quantities
			7.2.4 Noise
			7.2.5 Clock Drift
		7.3 Network Communication
			7.3.1 Data Link – Medium Access Control Protocols
			7.3.2 Protocols and their Tradeoffs
			7.3.3 Mitigating the Effects of Network Delays and Drops
				7.3.3.1 Network Degradation
				7.3.3.2 Network Drop
				7.3.3.3 Requirements on the DT
				7.3.3.4 Simulation of Network Degradation and Drop
			7.3.4 Data Compression
		7.4 Message-Based Communication
		7.5 Storing Data in Time-Series Databases
		7.6 Software Sensing
			7.6.1 Sensor Fusion
			7.6.2 Deep Learning Perception
		References
Part III Services for Digital Twins
	Chapter 8 Visualisation in a Digital Twin Context
		8.1 Introduction
		8.2 Visualisation
		8.3 Visualisation Services in a Digital Twin
			8.3.1 Visualisation Techniques
		8.4 Frameworks used for DT Visualisation
			8.4.1 Dashboards
			8.4.2 4D Visualisation
		8.5 Visualisation Examples
			8.5.1 Dashboards: Incubator Prototype
			8.5.2 Augmented Reality: Incubator Prototype
		References
	Chapter 9 System Monitoring through a Digital Twin
		9.1 Introduction
		9.2 Describing Desirable Properties
			9.2.1 Temporal Logic in General
			9.2.2 Linear Temporal Logic
			9.2.3 Signal Temporal Logic
		9.3 Monitoring using Runtime Verification
		9.4 Data-driven Anomaly Detection
		References
	Chapter 10 Advanced Digital Twin Services
		10.1 Introduction
		10.2 What-if Simulations
			10.2.1 Design Space Exploration with Conflicting Objectives
			10.2.2 Fault-injection enabled Digital Twins
			10.2.3 Runtime Verification of What-if Simulations
		10.3 Fault Diagnosis and Resilience
		10.4 Predictive Maintenance
		10.5 Re-configuration, Robustness and Optimisation
		References
Part IV Realising Digital Twins
	Chapter 11 Realising Digital Twins
		11.1 Introduction
		11.2 Digital Twin Frameworks
		11.3 Cloud and Virtualisation Technologies
		11.4 Digital Twin Composition
		11.5 Digital Twin and Physical Twin Configuration
			11.5.1 Requirements
			11.5.2 Existing Digital Twin Configuration Formats
			11.5.3 Digital Twin Configuration Template
		11.6 Digital Twin Class and Instances
		11.7 DTaaS: Reference Architecture for Digital Twin Platforms
		11.8 DTaaS: the DT Execution Manager
			11.8.1 On-demand Execution
			11.8.2 Execution Isolation
			11.8.3 Distributed Execution Manager
		11.9 Prototype Implementation
		11.10 Support for DT Services
		11.11 Fleet Analysis
		References
	Chapter 12 Case Studies in Digital Twins
		12.1 Introduction
		12.2 Summary of Characteristics
		12.3 The Tempeh Incubator
			12.3.1 Incubator DT Overview
			12.3.2 Fundamental Characteristics
				12.3.2.1 C1: System-under-study
				12.3.2.2 C2: Acting Components
				12.3.2.3 C3: Sensing Components
				12.3.2.4 C4: Multiplicities
				12.3.2.5 C5: Data Transmitted
				12.3.2.6 C6: Insights/Actions
				12.3.2.7 C7: Services
				12.3.2.8 C8: Enablers
				12.3.2.9 C9: Models and Data
				12.3.2.10 C10: Constellation
				12.3.2.11 C11: Time-Scale
				12.3.2.12 C12: Fidelity Considerations
				12.3.2.13 C13: Life-cycle Stages
				12.3.2.14 C14: Evolution
			12.3.3 Summary and FutureWork
		12.4 The (Desktop) Robotti
			12.4.1 Desktop Robotti DT Overview
			12.4.2 Fundamental Characteristics
				12.4.2.1 C1: System-under-study
				12.4.2.2 C2: Acting Components
				12.4.2.3 C3: Sensing Components
				12.4.2.4 C4: Multiplicities
				12.4.2.5 C5: Data Transmitted
				12.4.2.6 C6: Insights/Actions
				12.4.2.7 C7: Services
				12.4.2.8 C8: Enablers
				12.4.2.9 C9: Models and Data
				12.4.2.10 C10: Constellation
				12.4.2.11 C11: Time-Scale
				12.4.2.12 C12: Fidelity Considerations
				12.4.2.13 C13: Life-cycle Stages
				12.4.2.14 C14: Evolution
			12.4.3 Summary and FutureWork
		12.5 The Flex-cell
			12.5.1 Flex-cell DT Overview
			12.5.2 Fundamental Characteristics
				12.5.2.1 C1: System-under-study
				12.5.2.2 C2: Acting Components
				12.5.2.3 C3: Sensing Components
				12.5.2.4 C4: Multiplicities
				12.5.2.5 C5: Data Transmitted
				12.5.2.6 C6: Insights/Actions
				12.5.2.7 C7: Services
				12.5.2.8 C8: Enablers
				12.5.2.9 C9: Models and Data
				12.5.2.10 C10: Constellation
				12.5.2.11 C11: Time-Scale
				12.5.2.12 C12: Fidelity Considerations
				12.5.2.13 C13: Life-cycle Stages
				12.5.2.14 C14: Evolution
			12.5.3 Summary and FutureWork
		12.6 The Research Vessel Gunnerus
			12.6.1 Fundamental Characteristics
				12.6.1.1 C1: System-under-study
				12.6.1.2 C2: Acting Components
				12.6.1.3 C3: Sensing Components
				12.6.1.4 C4: Multiplicities
				12.6.1.5 C5: Data Transmitted
				12.6.1.6 C6: Insights/Actions
				12.6.1.7 C7: Services
				12.6.1.8 C8: Enablers
				12.6.1.9 C9: Models and Data
				12.6.1.10 C10: Constellation
				12.6.1.11 C11: Time-Scale
				12.6.1.12 C12: Fidelity Considerations
				12.6.1.13 C13: Life-cycle Stages
				12.6.1.14 C14: Evolution
			12.6.2 Summary and FutureWork
		References
Part V Advanced Topics
	Chapter 13 Security and Privacy-related Issues in a Digital Twin Context
		13.1 Introduction
		13.2 DT Security Architecture
		13.3 Approaches to a DT Security and Privacy
			13.3.1 Standard approaches to cyber security in a DT context
			13.3.2 Formal Methods-Based Approaches to Cyber Security for DTs
			13.3.3 Attack mitigations
			13.3.4 Attack Detection in DTs
				13.3.4.1 Design considerations for DT attack detectors
				13.3.4.2 Motivational Examples Showing Significance of Attack Detection
				13.3.4.3 Attack Detection Strategies
		13.4 Intellectual Property Protection
		13.5 Security in the Real World
		References
	Chapter 14 Autonomous Reconfiguration Enabled by Digital Twins
		14.1 Introduction
		14.2 Autonomous Systems and DTs
		14.3 Self-* properties
		14.4 Goals
		14.5 Collaboration between Systems
		14.6 Safety and uncertainty in reconfiguration
		14.7 Roadmap
		References
	Chapter 15 Future Directions and Challenges
		15.1 Introduction
		15.2 Firm Foundations for Digital Twin Engineering
			15.2.1 Understanding the Limits of Predictions
			15.2.2 Uncertainty: Quantification and Propagation
			15.2.3 Towards Verified Digital Twins
			15.2.4 Protection against Security and Privacy Attacks
			15.2.5 Synthesising Safe Scenarios
			15.2.6 Mutual Calibration
		15.3 Digital Twin Platforms
			15.3.1 Automating Digital Twin Production
			15.3.2 Modelling Languages for Digital Twins
			15.3.3 Incorporation of Ontologies and Knowledge Graphs
			15.3.4 Distributed Simulation and Workload Distribution
			15.3.5 Bi-directional synchronisation with the actual system
			15.3.6 Full Life Cycle Management
			15.3.7 Leveraging Multiple Levels of Abstraction
			15.3.8 Certification of Digital Twins
		15.4 Increasing the Level of Autonomy for Digital Twins
			15.4.1 Reducing Human Supervision
			15.4.2 The Cognitive Digital Twin
			15.4.3 Awareness of the Reality Gap
			15.4.4 Capture and Representation of Causal Relations
			15.4.5 Increased Robustness
		15.5 Supporting Composition of Digital Twins
			15.5.1 Need for Standardised DT Interfaces
			15.5.2 Collaborative Digital Twins
			15.5.3 Openness of DTs
		15.6 Novel Applications of Digital Twins
		15.7 Concluding Remarks
		References