D.14: Process Digital Twin

• Short Description: The Process Digital Twin (DT) is a virtual representation of a process that incorporates real-time data together with other forms of AI to analyze the system behaviour, performance, and outcomes. Therefore, the Process DT can be seen as a higher-level AI-enabled AAS that ingests from other AASs deployed in the production line and other DTs (Product and Person/Human). This “newer” AAS is especially designed to collect, preprocess data and execute AI/ML models for computing answers, insights, optimizations, while solving “circular manufacturing” problems. The Process DT has been developed by using the AAS as technological background. This means that it provides the same REST API, the "data image" of the process is built by using the AAS's metamodel. And event-based communication (using MQTT) is supported. Since it needs to collect and send data to other AAS that are part of the process, then an orchestrator has been embedded. The orchestrator is using the behaviour tree mathematical model for creating, defining, managing and executing complex tasks. Finally, an embedded AI engine has been developed to allow the exectution of Neural Networks developed using TensorFlow.

• Reference, URL: https://www.uninova.pt/

• Applicable Business Category: Manufacturing

• Application in relevant Projects/Initiatives: Yes, Circular TwAIn

• Asset Type: Executable

• AI Breadth: ML, Neural Networks

• Learning Ability:

• Related technologies: Tensorflow

• Applicable Research Area:

• Applicable Technical Category:

• License Information: Open Source license

• Related to circularity and sustainability:

• Audience:

Go back to the AI portfolio of reference implementations