Data-driven invariant modelling patterns for digital twin design - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Journal of Industrial Information Integration Année : 2023

Data-driven invariant modelling patterns for digital twin design

Résumé

The Digital Twin (DT) is one of the most promising technologies in the digital transformation market. A digital twin is a virtual copy of a physical system that emulates its behaviour to predict failures and opportunities for change, prescribe actions in real-time, and optimise and/or mitigate unexpected events. Modelling the virtual copy of a physical system is a rather complex task and requires the availability of a large amount of information and a set of accurate models that adequately represent the reality to model. At present, the modelling depends on the specific use case. Hence, the need to design a modelling solution suitable for virtual reality modelling in the context of a digital twin. The paper proposes a new approach to design a DT by endeavouring the concept of "modelling patterns" and their invariance property. Modelling patterns are here thought of as data-driven, as they can be derived autonomously from data using a specific approach devised to reach an invariance feature, to allows these to be used (and re-used) in modelling situations and/or problems with any given degree of similarity. The potentialities of invariance modelling patterns are proved here by the grace of a real industrial application, where a dedicated DT has been built using the approach here proposed.
Fichier principal
Vignette du fichier
Semeraro et al.pdf (4.48 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03888725 , version 1 (07-12-2022)

Identifiants

Citer

Concetta Semeraro, Mario Lezoche, Hervé Panetto, Michele Dassisti. Data-driven invariant modelling patterns for digital twin design. Journal of Industrial Information Integration, 2023, 31, pp.100424. ⟨10.1016/j.jii.2022.100424⟩. ⟨hal-03888725⟩
57 Consultations
151 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More