Experts: David Atienza (EPFL), Daniel Schmid (ZHAW)
The concept of the digital twin connects physical products with digital representations of them and ensures an exchange of data between the two entities. Today, digital twins are mainly used in industry, but they have promising potential for use in other domains, such as smart cities or digital health, where there are already areas of initial research. If the transfer of knowledge from research to tangible applications succeeds, Swiss SMEs could set the tone for special applications, thanks to their know-how and good reputation, backed by Switzerland’s economic and political stability.
The concept of the digital twin typically comprises three components:
Thanks to the digital twin, the physical object can either be controlled or else embedded more comprehensive approaches, such as predictive maintenance or smart maintenance. For companies in the manufacturing industry, digital twins offer an opportunity to complement their products with services. This makes the concept of the digital twin an elementary component of what is referred to as Industry 4.0.
Digital twins are primarily used in the manufacturing industry. Firstly on the manufacturer side, for carrying out simulations of digital prototypes during product development, but also in quality assurance, for monitoring the state of machines and production facilities. Secondly, digital twins are used to monitor the life cycle of products, so as to incorporate any necessary services or updates, or to trigger maintenance.
The general thrust of efforts in research and development worldwide is to try to make the concept of the digital twin more productive outside industry as well. In the healthcare sector, for example, there are attempts in the context of digital health applications. Another research field is that of implementations in smart cities, modelling the volume of traffic caused by flows of people and goods, for instance. Such models can aid urban development or be used to control electronic traffic guidance systems. Models of structures in the context of building information modelling also constitute a form of digital twin.
In Switzerland, there are numerous research activities in the field of digital twins. However, only a few commercial applications are emerging from these efforts. Combining digital twins with sensors and artificial intelligence offers major potential for new applications. In recent years, there has been considerable progress in the field of artificial intelligence; however, many sensors are very expensive, as they are often custom-made one-off items.
The main impetus in the development of digital twins is coming from large international system suppliers. Nevertheless, due to Switzerland’s excellent reputation as a location, there are opportunities for Swiss SMEs to provide innovative adapted solutions. On the application side, it is not easy to include Switzerland in any comparison regarding the implementation of digital twins as it is a country of SMEs and the majority of digital twins already implemented are used in large companies. Although many SMEs from the manufacturing industry have not yet installed any digital twins at all, it is worth assessing relevant applications and observing developments in the field.
Implementation of digital twins and corresponding applications not only entails technical challenges, but also requires related skills in the handling of data and models. SMEs in particular often produce relatively small batches, which can make modelling problematic. If models are based on small datasets, it can happen that data going into the model is more likely to represent one-off episodes, raising the question of whether, and to what extent, it is representative of the entire population.
When it comes to applications that have possible safety consequences for users, the legal situation has not yet been fully clarified. For instance, much like the liability issue with autonomous vehicles, it is still unclear who is liable if damage occurs as a result of a model being followed. Moreover, in contrast to cars, damage in a production facility or involving a model might only occur after a delay, which makes it much more difficult to trace the chain of causality.
Another challenge with digital twins is that of cybersecurity issues. Although this challenge applies equally to all digital products, it is clear that security issues should not be left as something to deal with later on, but must be incorporated as a central factor during product development.
Although there are some activities at Swiss universities, there is a lack of effective funding for digital twin projects, as they often encompass various disciplines and quickly become very staff- and cost-intensive. In addition, there is a need for stronger networking among the players.