2021_moddit_presentation_final.pdf
0 likes24 views
The document discusses the Christian Doppler Laboratory's exemplar for model-driven digital twin engineering focused on air quality management, highlighting the challenges and solutions in creating cost-effective digital twins. It outlines a digital twin implementation for a room, which includes various sensors and applications to monitor and visualize air quality. The future work focuses on enhancing the system's heterogeneity, integration, and collaboration opportunities for further development.
![Digital Twin
• Industry-Scale Implementation using Microsoft Azure [1]
• Open-Source Implementations [2]
• Eclipse Hono/Ditto/Vorto
• Runtime Monitoring Infrastructure from Academia [3]
• Challenge 1: Platform-neutral modeling
[1] https://github.com/derlehner/IndoorAirQuality_DigitalTwin_Exemplar/tree/main/digital_twin/azure
[2] https://github.com/derlehner/IndoorAirQuality_DigitalTwin_Exemplar/tree/main/digital_twin/open-source
[3] M. Vierhauser, H. Marah, A. Garmendia, J. Cleland-Huang, and M. Wimmer, “Towards a model-integrated runtime monitoring infrastructure for cyber-physical
systems,” in Proc. of ICSE-NIER, pp. 96–100, IEEE, 2021.
time
temp.
System Structure
Digital Twin Model Historical Data
8](https://image.slidesharecdn.com/2021modditpresentationfinal-220728121113-660ec73e/85/2021_moddit_presentation_final-pdf-8-320.jpg)
![Digital Twin
Challenge 2: Dealing with model evolution [1]
9
[1] UML/AML Models for Evolution Cases available via: https://github.com/derlehner/IndoorAirQuality_DigitalTwin_Exemplar/tree/main/digital_twin/models
Building
Room
id: String
aqLevel: Int
[0..*] rooms
[0..*]
controllers
Controller
type: String
value: String
unit: String
Sensor
[0..*]
sensors
id: String
id: String
isActive: Bool
Alarm
[0..*]
alarms
Room
Ventilation
System
level: int
[0..*]
ventilationSystems
Building
Room
Controller
[0..*]
robots Robot
[1..1]
controller
[1..1]
current_room
id: String](https://image.slidesharecdn.com/2021modditpresentationfinal-220728121113-660ec73e/85/2021_moddit_presentation_final-pdf-9-320.jpg)
2021_moddit_presentation_final.pdf
- 1. Christian Doppler Laboratory for Model-Integrated Smart Production Institute of Business Informatics – Software Engineering Johannes Kepler University Linz Altenberger Straße 69, Science Park 3 4040 Linz Christian Doppler Laboratory for Model-Integrated Smart Production CDL-MINT Air Quality Management: An Exemplar for Model-Driven Digital Twin Engineering Hari Shankar Govindasamy, Ramya Jayaraman, Burcu Taspinar, Daniel Lehner, Manuel Wimmer
- 2. Motivation Creating and Maintaining Digital Twins is expensive • MDE techniques can help • But how (expensive to showcase applications)? Contribution: Cost-Effective Exemplar to • Identify challenges • Showcase future solutions Exemplar provides • Automation device (< 100 €) + historical data • Different Digital Twin implementations • 3 applications that make use of this DT 2 Room 101
- 3. Digital Twin of a Room for Air Quality Management 3 Physical Twin Digital Twin Applications Communication Middleware DT Repository DT Service Room101 CO2 Sensor Temp. Sensor Hum. Sensor Raspberry 1 Prediction Service Simulation Service Visualization App exchange_data get_data send_data interact_with Room101 temp. CO2 LED change_color hum. History Service History Repository
- 4. Digital Twin of a Room for Air Quality Management 4 Physical Twin Digital Twin Applications Communication Middleware DT Repository DT Service Room101 CO2 Sensor Temp. Sensor Hum. Sensor Raspberry 1 Prediction Service Simulation Service Visualization App exchange_data get_data send_data interact_with Room101 temp. CO2 LED change_color hum. History Service History Repository
- 5. Automation Device 5 Controller Raspberry Pi Alarms LED Sensors CCS811 Sensor DHT11 Sensor Buzzer Breadboard
- 7. Digital Twin of a Room for Air Quality Management 7 Physical Twin Digital Twin Applications Communication Middleware DT Repository DT Service Room101 CO2 Sensor Temp. Sensor Hum. Sensor Raspberry 1 Prediction Service Simulation Service Visualization App exchange_data get_data send_data interact_with Room101 temp. CO2 LED change_color hum. History Service History Repository
- 8. Digital Twin • Industry-Scale Implementation using Microsoft Azure [1] • Open-Source Implementations [2] • Eclipse Hono/Ditto/Vorto • Runtime Monitoring Infrastructure from Academia [3] • Challenge 1: Platform-neutral modeling [1] https://github.com/derlehner/IndoorAirQuality_DigitalTwin_Exemplar/tree/main/digital_twin/azure [2] https://github.com/derlehner/IndoorAirQuality_DigitalTwin_Exemplar/tree/main/digital_twin/open-source [3] M. Vierhauser, H. Marah, A. Garmendia, J. Cleland-Huang, and M. Wimmer, “Towards a model-integrated runtime monitoring infrastructure for cyber-physical systems,” in Proc. of ICSE-NIER, pp. 96–100, IEEE, 2021. time temp. System Structure Digital Twin Model Historical Data 8
- 9. Digital Twin Challenge 2: Dealing with model evolution [1] 9 [1] UML/AML Models for Evolution Cases available via: https://github.com/derlehner/IndoorAirQuality_DigitalTwin_Exemplar/tree/main/digital_twin/models Building Room id: String aqLevel: Int [0..*] rooms [0..*] controllers Controller type: String value: String unit: String Sensor [0..*] sensors id: String id: String isActive: Bool Alarm [0..*] alarms Room Ventilation System level: int [0..*] ventilationSystems Building Room Controller [0..*] robots Robot [1..1] controller [1..1] current_room id: String
- 10. Digital Twin of a Room for Air Quality Management 10 Physical Twin Digital Twin Applications Communication Middleware DT Repository DT Service Room101 CO2 Sensor Temp. Sensor Hum. Sensor Raspberry 1 Prediction Service Simulation Service Visualization App exchange_data get_data send_data interact_with Room101 temp. CO2 LED change_color hum. History Service History Repository
- 11. Advanced Visualization Application Analyze air quality in a building • Derive Actions for improving air quality • Analyze the effect of actions on actual air quality values Challenge 3: Runtime-Integrated Model Visualizations • Effort for adapting a general Dashboard to specific user needs • Visualization model and configuration model for auto-generation 11 Visualization Dashboards
- 12. Physical Simulation Service 3D physical modelling of the flow of co2 molecules • Application: simulating different scenarios • How does a ventilation system affect co2 values? Challenge 4: Integrating Physical Simulations • Effort for creating simulation model • Reuse information from existing models • Generate simulation for new DTs 12 3D model of a room flow of co2 molecules for single source point
- 13. Prediction Service Neural Network to prediction future Air Quality values • Proactively influence values Challenge 5: AI-integrated temporal models • Integrate data from past and present with predictions for the future • Query future values 13 Historical Values n - inputs n - outputs Trained Model Future Values Neural Network with 2 hidden layers
- 14. Summary and Future Work Summary • Digital Twin Exemplar for Indoor Air Quality Measurement • Setup details for Physical, Digital Twin + Applications available on Github Next Steps for Exemplar • Vendor-Neutral Meta-Model • More Digital Twin Implementations • Adding heterogeniety on hardware side • Providing implementation details for evolution cases We‘re open for collaborations/contributions! 14
- 15. CDL-MINT https://cdl-mint.se.jku.at/ Christian Doppler Laboratory for Model-Integrated Smart Production CDL-MINT Thank You! Comments? Questions? Feedback? Daniel Lehner, Hari Govindasamy daniel.lehner@jku.at http://github.com/derlehner hari.govindasamy@jku.at Use Case https://github.com/derlehner/IndoorAirQuality_DigitalTwin_Exemplar