Abstract
Digital twins (DTs) are digital representations of assets, capturing their attributes and behavior. They are one of the cornerstones of Industry 4.0. Current DT standards are still under development, and so far, they typically allow for representing DTs only by attributes. Yet, knowledge about the behavior of assets is essential to properly control and interact with them, especially in the context of industrial production. This behavior is typically represented by multiple different models, making integration and orchestration within a DT difficult to manage. In this paper, we propose a new approach for hybrid DTs by intertwining different DT models. We also show how to realize this approach by combining the Fraunhofer Asset Administration Shell (AAS) Tools for Digital Twins (FAST) to create Industry 4.0-compliant DTs with Apache StreamPipes to implement and manage multiple DT models. Our prototype implementation is limited to a subset of the AAS metamodel and pull-based communication between FAST and an external Apache StreamPipes instance. Future work should provide full support for the AAS metamodel, publish/subscribe-based communication, and other execution environments as well as deployment strategies. We also present how this approach has been applied to a real-world use case in the steel production industry.
Highlights
In recent years, digital twin (DT) technology has rapidly gained acceptance in various industries
Since DTs should not be proprietary solutions but rather be developed based on standards, we propose to develop DTs based on the Asset Administration Shell (AAS) [8]
We investigated and compared the following state-of-the-art digital twin and IoT standards and specifications: Asset Administration Shell, W3C Web of Things, Digital Twin Definition Language, NGSI-LD, OData, and OGC SensorThings API
Summary
Digital twin (DT) technology has rapidly gained acceptance in various industries. Some situations are not observable (as there are no sensors for real-time monitoring) and the data needed to build prediction/classification models are missing [4] In such cases, only models with lower accuracy or based on other phenomena can be developed. We propose a new approach for modeling digital replicas of assets given the fact that accurate models and data are not available and traditional DTs cannot be applied. To demonstrate the necessity and applicability of the proposed approach, we have explained how it was applied to realize a complex tool wear scenario in steel production We argue that such an approach brings several benefits, from creating accurate orchestrated models for the tool wear process to helping operators to gain a deep understanding of the wear process and improving the decision-making process (e.g., optimal decision on when to repair a piece of equipment).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
