Towards a Family of Digital Model/Shadow/Twin Workflows and Architectures Randy Paredis 1 a , Cl´ audio Gomes 2 b and Hans Vangheluwe 1,3 c 1 University of Antwerp, Department of Computer Science, Middelheimlaan 1, Antwerp, Belgium 2 Aarhus University, DIGIT, Department of Electrical and Computer Engineering, ˚ Abogade 34, Aarhus N, Denmark 3 Flanders Make@UAntwerp, Belgium Keywords: Digital Model, Digital Shadow, Digital Twin, Architecture, Workflow, VariabilityModeling. Abstract: Digital Twins (DTs) can be used for optimization, analysis and adaptation of complex engineered systems, in particular after these systems have been deployed. DTs make full use of both historical knowledge and of streaming data from sensors. DTs have been given numerous (distinct) definitions and descriptions in the liter- ature. There is no consensus on terminology, nor a comprehensive description of workflows nor architectures. Following Multi-Paradigm Modelling principles, this paper proposes to explicitly model construction and use workflows of DTs as well as their architectures. We apply the concepts of variability (also known as product family) modeling, in particular to DT workflow and architecture. This allows for the de-/re-construction of the different DT variants in a principled, reproducible and partially automatable manner. To illustrate our ideas, two small use cases are discussed: a line-following robot (representative for an Automated Guided Vehicle) and an incubator (representative for an Industrial Convection Oven). The use cases focus on important systems in an industrial context. 1 INTRODUCTION Digital Twins (DTs) are increasingly used Industry 4.0 and industrial processes for many purposes such as monitoring, analysis, optimization. While their definition has changed throughout the years, the con- cept stayed somewhat the same: there exists a dig- ital counterpart of a real-world system that provides information about this system. Usually, this infor- mation concerns itself with optimizations and correc- tional behavior of this system. Academic and industrial interest in DTs is grow- ing, as they allow the acceleration through digitization that is at the heart of Industry 4.0. Digital Twins are made possible by technologies such as the Internet of Things (IoT), Augmented Reality (AR), Product Life- cycle Management (PLM) and many more. Despite the many surveys on the topic (Rosen et al., 2015; Negri et al., 2017; Kritzinger et al., 2018; Cheng et al., 2018; Park et al., 2019; Zhang et al., 2019; Aivaliotis et al., 2019; Kutin et al., 2019; a https://orcid.org/0000-0003-0069-1975 b https://orcid.org/0000-0003-2692-9742 c https://orcid.org/0000-0003-2079-6643 Bradac et al., 2019; Cimino et al., 2019; Lu et al., 2020; Tao et al., 2019), there is no general consensus on what characterizes a Digital Twin, let alone how it is constructed. There is no one-definition-fits-all and therefore anyone in need of a DT starts from scratch to build (what they believe to be) a DT. The main con- cern is to create “value” for the user and to ensure minimal re-use of workflows and architectures. To some, a DT defines the virtual counterpart of the sys- tem, while for others, it encompasses the full concept of having both virtual and real systems at the same time. For example, Lin and Low (2019) define DT as “a virtual representation of the physical objects, processes and real-time data involved throughout a product life-cycle”, whereas Park et al. (2019) define DT as “an ultra-realistic virtual counterpart of a real- world object”. This paper attempts to unify the most common definitions and viewpoints in the form of a family of problems solved by DTs as well as variant workflows and architectures. Several of the above cited surveys propose refer- ence architectures for DTs. For instance, Bevilacqua et al. (2020) proposes one that shares many common- 174 Paredis, R., Gomes, C. and Vangheluwe, H. Towards a Family of Digital Model/Shadow/Twin Workflows and Architectures. DOI: 10.5220/0010717600003062 In Proceedings of the 2nd International Conference on Innovative Intelligent Industrial Production and Logistics (IN4PL 2021), pages 174-182 ISBN: 978-989-758-535-7 Copyright c  2021 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved