Digital Twin

Definition and differentiation

The digital twin is a virtual representation of a product, a process or a production system that collects and processes data by means of automated data synchronisation and can influence its physical counterpart via control commands. The digital twin is a core element of digital transformation and many other digitalisation efforts. Despite its widespread use, there is no standardised understanding of the term digital twin. This is due to the different areas of tension that the digital twin covers. A clear definition is therefore not yet available in the current reference works. Nevertheless, a delimitation of the concept can be derived from the definition approaches.

The core of the definitions includes three dimensions: the physical element, the virtual element and the data flow that connects both elements ^[1-3]. The virtual element in particular is defined more specifically. For example, the digital twin must contain an ultra-realistic representation of the physical element, which is suitable for carrying out simulations with a high degree of fidelity and synchronised data ^[4]. In addition, the digital twin accompanies the physical element throughout its entire life cycle and develops accordingly ^[5]. The original three-dimensional definition was recently extended by two further dimensions. The digital twin therefore continues to consist of the physical and virtual element as well as the data connections between the elements. The generated, recorded and stored data sets and the services offered by the digital twin were added ^[6,7].

Further elements are described in recent taxonomic and typological studies. Jones et al. define a total of 13 elements of a digital twin. In addition to the already known five dimensions, they add the physical and virtual environment of the elements, the element states, the processes of state measurement and state change, the synchronisation rate and the consideration of physical and virtual processes that influence the digital twin ^[8]. Based on the various characteristics of an instantiated digital twin, five archetypes can also be derived ^[9]. Based on a maturity model, there is initially the basic twin, which only contains the basic characteristics, such as synchronisation. The other enriched twins build on this, which offer more comprehensive communication options or deeper interoperability, among other things. Finally, the fifth archetype is the most comprehensive development stage and includes an extensive range of data processing, data transfer and data storage options.

However, it should be noted that the development of the digital twin, and with it the final definition, is not yet complete.

History

The history of the digital twin dates back to the 1960s, when NASA began working with physical twins as part of the Apollo programme ^[10,11]. Specifically, the failed Apollo 13 mission is often cited as the first test, where the physical twin on the ground provided a real simulation environment to test alternatives and recreate real flight conditions ^[10,12]. At the same time, Airbus began developing its first aircraft model, the Airbus A300, at European level. Here, Airbus relied on so-called Iron Birds, which physically simulate the aircraft systems and thus provide real-world data even before the first flight. This is still the case today with the latest new developments despite the wide range of digital possibilities ^[13]. In 2002, the basic concept of the Digital Twin was first presented by Michael Grieves as part of seminar lectures on “Product Lifecycle Management (PLM)” ^[2]. Although the concept at that time already contained the core elements of today’s Digital Twin concept, it was referred to as the “Conceptual Ideal for PLM”. In the following years, Grieves in particular continued to promote the concept, but changed the name several times. The term Digital Twin was first introduced in 2010 during a collaboration between Grieves and NASA engineers ^[14,15]. The concept of the digital twin has been expanded. The Digital Twin is primarily understood as a virtual representative of the aerospace fleets of NASA and the US Air Force ^[16,14,5]. Grieves himself finally described his ideal PLM concept as a digital twin in 2014 ^[12]. Various literature analyses show that the digital twin has been widely researched since 2015 ^[1,17,18].

Today, the digital twin is a widely used term in a variety of different domains and application contexts, e.g. in engineering, computer science, mathematics, social sciences, medicine and many more.

Application and examples

Digital twins are an integral part of industrial processes. Of the ten largest German companies by market value alone, all are already developing or using digital twins: from the monitoring and control of production and logistics systems (cf. VW Group, Mercedes-Benz, BMW, Deutsche Post, Merck), to facility management and plant maintenance (Bayer), to the provision of services based on digital twins (SAP, Siemens, Allianz). Almost every major company outside of the top ten is also working on this topic.

Lighthouse projects in this area include the Industrial Digital Twin Association IDTA ^[19]which designs and offers digital twins on the basis of the administration shell, and the Digital Twin Registry, which is part of the Catena-X ^[20] is being developed within the automotive industry.

Criticism and problems

The non-exhaustive definition of the digital twin results in a proliferation of interpretations of digital twins. This is particularly evident in two aspects: Firstly, the clear distinction between digital twin and simulation applications is very often blurred in practical application. Secondly, there are other related concepts that are similar to the digital twin.

Although simulation is one of the most important applications of the digital twin [9]the two technologies are not the same. The simulation application is merely a subset of the services offered by the digital twin ^[21]. However, the close link between the two terms is not surprising given the simulation-orientated definitions of the digital twin, including as the next evolutionary step in simulation [11]. Studies show that many projects that call their virtual artefact digital twin have not developed a digital twin by definition, but have carried out classic simulations ^[22]. On the other hand, the simulation landscape continues to develop convergently with the digital twin, so that a simulation study of recent years offers many characteristics and capabilities of a digital twin, which can be attributed not least to the rapid further development of simulation software ^[23].

Another challenge is the differentiation of the digital twin from related concepts such as the digital shadow, the digital model or the digital life record. Various differentiation approaches exist, which focus in particular on the different types of data transfer ^[24]. The term “virtual twin” is used less frequently in connection with the digital twin. For the most part, this term is used in the literature for the virtual part of the digital twin or as a synonym for the overall concept of the digital twin itself (cf. ^[25]). However, there are also endeavours to integrate the digital twin within a metaverse as a virtual living environment and thus enable the user to have virtual experiences ^[26]. Dassault Systems, for example, is playing a pioneering role here with its virtual twin within the 3D experience.

Research

Research into digital twins is currently focussing on the aspect of greater differentiation from related concepts. Furthermore, the findings from production technology are being adapted and pursued further, particularly in the areas of smart cities, logistics and healthcare. There are various research projects at universities, colleges, Fraunhofer Institutes and in industry that are developing specific sector-relevant solutions:

• CogniTwin
• FA³ST
• Digest
• Co-Twin

Further research activities concern the secure handling of so-called shared digital twins, which are subject to special requirements analogous to data trustees.

---