Greater cross-agency organisational and educational support will be needed, if digital twin technologies are to realise their full potential across all areas of the economy and society: that’s the conclusion of a recently published report from the US National Academies of Sciences, Engineering, and Medicine.

The report, Foundational Research Gaps and Future Directions for Digital Twins, looks at the environment for fostering digital twins in the US and finds that federal agencies, such as the US Department of Defense, Department of Energy, National Institutes of Health, and National Science Foundation, need to develop new crosscutting programs to advance mathematical, statistical, and computational foundations underlying the technologies.

Karen Willcox, chair of the committee that wrote the report, highlighted the promise of digital twins in creating value across areas of science and technology, including engineering, the natural world, and medicine.

“Our report makes clear that there is a real opportunity here to bring together domains and disciplines in new, valuable ways, but to achieve that value requires investment in interdisciplinary foundations,” said Willcox, who is director of the Oden Institute for Computational Engineering and Sciences and professor of aerospace engineering and engineering mechanics at the University of Texas at Austin.

The report emphasises the role of agencies in ensuring trust in the operation of digital twins. Verification, validation, and uncertainty quantification (VVUQ) is an area that requires collaborative and interdisciplinary investment to advance responsible development, implementation, monitoring, and sustainability of digital twins, according to the authors.

They see an important role for agency collaboration on development of the ethical, privacy and security environments for digital twins. The report says: “While security issues with digital twins share common challenges with cybersecurity issues in other settings, the close relationship between cyber and physical in digital twins could make cybersecurity more challenging.” 

In some cases, digital twin deployments may be held up by a lack of access to the considerable computational resources needed for them to function effectively. The report recommends that federal agencies could ease this shortage by creating mechanisms, which provide researchers with more computational resources.

According to the report, federal agencies and other organizations could also help advance the digital twin ecosystem in several other ways, including:

• Identifying areas where collaboration with industry helps advance research—the Department of Energy could initially work with partners on energy infrastructure security and improved emergency preparedness, for example. 

• Conducting assessments of data and model sustainment needs to ensure that digital twins can be sustained and maintained beyond their initial creation. 

• Helping advance the interdisciplinary skills of the current and future workforce through federally funded research and development centres, institutes, and interdisciplinary research programs that incorporate training, as well as by seeding discussions across industry and academic sectors to explore the creation of interdisciplinary degrees. 

Defining a digital twin

The report also provides what the authors describe as a cohesive definition of the term “digital twin” – an acknowledgement that the term still means different things to different people. 

A digital twin uses modelling and simulation to create a virtual representation that mimics the structure, context, and behaviour of a physical counterpart. But the authors say that, in addition to modelling and simulation, a digital twin should also have two-way interaction between the virtual and physical worlds. 

That means the digital twin should be capable of absorbing data from the physical world it is modelling to update itself, and also have predictive capability in order to inform decision-making for the physical counterpart it is twinning. Those decisions could be fully automated, used as recommendations to inform a human’s decision, or constitute a mix of the two.