from virtual prototype to digital twin · 2020. 8. 6. · 13 analytic models used during - analysis...
TRANSCRIPT
Examples from the Smart Lighting domain
Jacques VerrietApril 9, 2019
FROM VIRTUAL PROTOTYPETO DIGITAL TWIN
2
Analytic models used during
- Analysis and design ➔ Virtual prototype
- Realization, testing, deployment, operation ➔ Digital twin
- Operation (vision) ➔ Adaptive system
OUTLINE
09-04-2019From Virtual Prototype to Digital Twin 2
Analysis
Design
RealizationV&V / Testing
Deployment
Operation
3
Analytic models used during
- Analysis and design ➔ Virtual prototype
- Realization, testing, deployment, operation ➔ Digital twin
- Operation (vision) ➔ Adaptive system
OUTLINE
09-04-2019From Virtual Prototype to Digital Twin 3
Analysis
Design
RealizationV&V / Testing
Deployment
Operation
4
VIRTUAL PROTOTYPE
09-04-2019From Virtual Prototype to Digital Twin 409-04-2019Digital Twin 4
Analyze &Optimize
Informalsystem
requirements
Virtualprototype
Formalize
Physicalsystem
Develop
5
• Witte Dame building
- Landmark historical building
- Former lighting factory
• Smart lighting system
- Complete 5th floor
- 367 light points
- More than 1,300 behaviors
SMART LIGHTING CASE
09-04-2019From Virtual Prototype to Digital Twin 5
6
WITTE DAME FLOOR PLAN
09-04-2019From Virtual Prototype to Digital Twin 6
7
WITTE DAME FLOOR PLAN
09-04-2019From Virtual Prototype to Digital Twin 7
8
WITTE DAME FLOOR PLAN
09-04-2019From Virtual Prototype to Digital Twin 8
9
VIRTUAL PROTOTYPE: INTERACTIVE SIMULATION
09-04-2019From Virtual Prototype to Digital Twin 9
10
• Challenge- 1000’s of components (sensors, buttons,
light points, …)
- 1000’s of interdependent controllers(occupancy, daylight harvesting, …)
• But, few types of components and controllers
DESIGN SPACE EXPLORATION
09-04-2019From Virtual Prototype to Digital Twin 10
Analyze &Optimize
Virtualprototype
11
DOMAIN MODEL FOR DISTRIBUTION CONTROL APPLICATIONS
09-04-2019From Virtual Prototype to Digital Twin 11
Control
Experiment
Requirement
Domain DSLs Validation DSLsSystem DSLs
Topology Scenario
Event
Structure
Behavior
Modify
Generate
Modify
Virtualprototype
12
• Executable virtual prototypes are used to optimize system behavior in early development phases
• Domain models can be used to handle system variability
• Design space exploration is used to find a suitable system configuration
VIRTUAL PROTOTYPE SUMMARY
09-04-2019From Virtual Prototype to Digital Twin 12
13
Analytic models used during
- Analysis and design ➔ Virtual prototype
- Realization, testing, deployment, operation ➔ Digital twin
- Operation (vision) ➔ Adaptive system
OUTLINE
09-04-2019From Virtual Prototype to Digital Twin 13
Analysis
Design
RealizationV&V / Testing
Deployment
Operation
14
Analytic models used during
- Analysis and design ➔ Virtual prototype
- Realization, testing, deployment, operation ➔ Digital twin
- Operation (vision) ➔ Adaptive system
OUTLINE
09-04-2019From Virtual Prototype to Digital Twin 14
Analysis
Design
RealizationV&V / Testing
Deployment
Operation
15
DIGITAL TWIN
09-04-2019From Virtual Prototype to Digital Twin 1509-04-2019 15
Monitor
Modify
Analyze &Optimize
Digital Thread
Physicalsystem
Digitaltwin
16
• Witte Dame building
- Landmark historical building
- Former lighting factory
• Smart lighting system
- Complete 5th floor
- 367 light points
- More than 1,300 behaviors
SMART LIGHTING CASE
09-04-2019From Virtual Prototype to Digital Twin 16
17
• Challenge
- 1000’s of components (sensors, buttons, light points, …)
- 1000’s of interdependent controllers (occupancy, daylight harvesting, …)
• Humanly impossible to keep overview of the system behavior
• Automated monitoring required
- Reuse of virtual prototype model as digital twin
SMART LIGHTING SYSTEMS
09-04-2019From Virtual Prototype to Digital Twin 17
18
FROM VIRTUAL PROTOTYPE TO DIGITAL TWIN
09-04-2019From Virtual Prototype to Digital Twin 18
Monitor
Modify
Analyze &Optimize
Digital Thread
Virtualprototype
Digitaltwin
Input (sensor) eventsOutput (light point) events
Physicalsystem
Differences
19
ROOT CAUSE ANALYSIS FOR SMART LIGHTING
09-04-2019From Virtual Prototype to Digital Twin 19
Collect Detect Analyze Resolve
20
ROOT CAUSE ANALYSIS FOR SMART LIGHTING
09-04-2019From Virtual Prototype to Digital Twin 20
Digital twin
Physical system
Comparator
System monitor
Rule-based diagnostics
Collect Detect Analyze Resolve
21
ANOMALY DETECTION
09-04-2019From Virtual Prototype to Digital Twin 21
22
DIAGNOSTICS: RULE 1
09-04-2019From Virtual Prototype to Digital Twin 22
1 Correctness of digital twin behavior
Is a sensor in the group reporting occupancy?
2 3 Is the digital twin behaving correctly?
23
DIAGNOSTICS: RULE 2
09-04-2019From Virtual Prototype to Digital Twin 23
2 Group problem Do other light points in the same group behave identically?
4 5 Is it an individual or group problem?
24
DIAGNOSTICS: RULE 3
09-04-2019From Virtual Prototype to Digital Twin 24
4 Commissioning error Have all behaviors of group 2002 been commissioned?
5 Commissioning error
25
• An executable virtual prototype specifies the desired system behavior
• The virtual prototype can be coupled to the physical system to become a digital twin
• The digital twin can be used to detect and analyze differences between specification and realization
• After analyzing the difference, one needs to modify the physical system and/or the digital twin
DIGITAL TWIN SUMMARY
09-04-2019From Virtual Prototype to Digital Twin 25
26
Analytic models used during
- Analysis and design ➔ Virtual prototype
- Realization, testing, deployment, operation ➔ Digital twin
- Operation (vision) ➔ Adaptive system
OUTLINE
09-04-2019From Virtual Prototype to Digital Twin 26
Analysis
Design
RealizationV&V / Testing
Deployment
Operation
27
Analytic models used during
- Analysis and design ➔ Virtual prototype
- Realization, testing, deployment, operation ➔ Digital twin
- Operation (vision) ➔ Adaptive system
OUTLINE
09-04-2019From Virtual Prototype to Digital Twin 27
Analysis
Design
RealizationV&V / Testing
Deployment
Operation
28
Adaptive system
ADAPTIVE SYSTEMS
09-04-2019From Virtual Prototype to Digital Twin 2809-04-2019Digital Twin 28
Monitor
Modify
Analyze &Optimize
Digital Thread
Physicalsystem
Digitaltwin
29
SMART LIGHTING ADAPTATION EXAMPLE
09-04-2019From Virtual Prototype to Digital Twin 29
30
• Detect
- Lights switch on shortly after they have switched off
• Analyze
- Occupancy sensor is not sensitive enough to detect all occupancy
- Occupancy sensor is not operating at maximum sensitivity
• Resolve
- Increase sensor sensitivity to detect smaller motions
ADAPTIVE LIGHTING CONTROLLER: SCENARIO 1
09-04-2019From Virtual Prototype to Digital Twin 30
31
• Detect
- Lights switch on shortly after they have switched off
• Analyze
- Occupancy sensor is not sensitive enough to detect all occupancy
- Occupancy sensor is operating at maximum sensitivity
• Resolve
- Increase time between last occupancy and switching off
- Issue service call to replace/move/add occupancy sensor
ADAPTIVE LIGHTING CONTROLLER: SCENARIO 2
09-04-2019From Virtual Prototype to Digital Twin 31
32
• Analysis based on digital twins is currently done manually
• Integration of the digital twin in the physical system gives a self-aware system
• Self-aware systems may adapt themselves or initiate service actions
ADAPTIVE SYSTEMS SUMMARY
09-04-2019From Virtual Prototype to Digital Twin 32
33
Analytic models used during
- Analysis and design ➔ Virtual prototype
- Realization, testing, deployment, operation ➔ Digital twin
- Operation (vision) ➔ Adaptive system
OUTLINE
09-04-2019From Virtual Prototype to Digital Twin 33
Analysis
Design
RealizationV&V / Testing
Deployment
Operation
34
• Executable virtual prototypes can be used to formalize system behavior specifications
• These virtual prototypes can become digital twins,which are used e.g. to detect and diagnose anomalous system behavior
• By incorporating the digital twins in the system,the system become self-aware or even self-adapting
For more information, visit our diagnostics booth on the demo market
TAKEAWAYS
09-04-2019From Virtual Prototype to Digital Twin 34