Multiple Modeling Languages for Digital Twins

MAster assignment

Type: Master CS

Period: Start date: as soon as possible

Student: Unassigned

If you are interested please contact:

Background:
The concept of Digital Twin (DT) has gained popularity as a digital representation of physical entities that interact with their real world counterparts in (near) real-time through sensors and actuators. DTs can be applied across different sectors, offering benefits like simulation, remote monitoring, and predictive maintenance, which are relevant capabilities of smart systems. However, achieving the full potential of DTs requires addressing interoperability challenges posed by the complex networks of devices and systems that play different roles in DTs. Recently, we published a paper [1] that presents a research agenda aimed at enhancing DT interoperability grounded in four computer/information science disciplines: architecture of distributed systems, model-based system engineering, ontology-driven conceptual modeling, and linked data with semantic web. This paper highlights how leveraging on existing standards, such as modelling languages and ontologies, is important for improved DT interoperability.

Interoperability refers to the capability of multiple systems or components to exchange and effectively utilize shared information. Therefore, interoperability defines the way of interconnection between sensors, devices, manufacturing systems, and people, including exchange of products and materials among facilities. In particular, semantic interoperability is the most challenging because it is about the “interpretation of shared data in an unambiguously way, ensuring that the understanding of the information is the same for senders and receivers”. Establishing automatic semantic interoperability for seamless systems’ integration is an arduous task.

The core function of a DT relies on merging the virtual model with sensor data that are collected with support of IoT technologies. DT data are formalized in diverse ways, gathered from various sensors and must be integrated with other data that rely on different languages and their serialization syntax. These can vary according to the different domains and purposes, and this complexity elevates integration and interoperability challenges at both syntactic and semantic aspects, and in all interoperability levels: legal, organizational, semantic and technical. We have been working with the concepts of digital thread, digital model and digital shadow within the DT research [2], which involves various representations of a target system adapted for specific purposes. These representations can include digital models for static analysis or simulation of different system versions. Advances in IoT technology enable the creation of digital shadows, using real-time data for visualization. DTs take this further through a bidirectional connection to the real system, utilizing real-time data to mimic and influence the actual behavior of the system, facilitating analysis, prediction, and rapid corrective actions by integrating models from the digital thread with sensor data and system actuators. In this context, the specification artifacts covered by the architecture of distributed systems play an important role for digital threads, since they prescribe the structural and behavioral elements of the systems, such as components, data sources, and services.

Master Project assignment:

The ontological analysis approach is relevant for improving the interoperability of general purpose modelling languages used in well-established fields like Enterprise Architecture (e.g., through ArchiMate), Business Process (e.g., BPMN), Business Modelling (e.g., through e3-value), System Engineering (eg., through SysML), Data Warehousing (e.g., through MDX), and domain specific ones (but still general within the domain), such as in Construction Engineering (e.g., through BIM). All these modeling languages have their specific purposes for representing systems’ characteristics, and therefore, play a pivotal role as modeling languages of the artefacts that compose digital threads. Using multiple languages in a systematic and consistent way for the specification of information systems is an old challenge, which is still open in the MBSE field, and is even challenging within MDSE, i.e., for automatically generating the technology-specific implementation artefacts. Research has shown that this consistency among modelling languages can be supported through the ODCM approach where a foundational ontology (in particular UFO) serves as a “single-source of truth”.

We have been working with the SEMIoTICS framework [3], which guides the use of three specific languages for IoT-based Early Warning System specification: OntoUML for context modeling, the Situation Modeling Language (SML) for situation identification modeling, and BPMN for situation reaction modeling. These are agnostic from the implementation technology. Models represented with these three languages can then be used to generate operational languages through MDSE transformations, like gUFO/OWL, Java ESPER (event processing language), and jBPM (business processes automation) respectively. The implementation of architectural decisions can also be automated, for example from the EA, DDD and OAS models, which can then be transformed into microservices (REST endpoints) that expose the data interfaces with JSON-LD serialization of gUFO/OWL. Future work comprises the improvement and further validation of SML by implementing and comparing new MDSE transformations for different Complex Event Processing (CEP) technologies, such as cloud- and fog-based stream analytics.

References:

Moreira, João. 2024. The Role of Interoperability for Digital Twins. EDOC.
Pessoa, M.V.P., Pires, L.F., Moreira, J.L.R., Wu, C.: Model-based digital threads for socio-technical systems. In: Marques, G., Gonzalez-Briones, A., Molina Lopez, J.M. (eds.) Machine Learning for Smart Environments/Cities. Intelligent Systems Reference Library, vol. 121, pp. 27–52. Springer, Cham (2022). ISBN 978-3-030-97516-6. https://doi.org/10.1007/978-3-030-97516-6 2
Moreira, J., Pires, L.F., Van Sinderen, M., Daniele, L., Girod-Genet, M.: Saref4health: towards IoT standard-based ontology-driven cardiac e-health systems. Appl. Ontol. 15(3), 385–410 (2020). ISSN 1570–5838. https://doi.org/10.3233/AO-200232