Digital twins – dynamic models that respond to real-time data

DOI: https://doi.org/10.58248/RR82

Overview

This briefing gives an overview of digital twins, outlining:

• how the technology works
• applications for digital twins across sectors
• challenges and opportunities in their development and use
• research investments in the technology.

What is a digital twin?

A digital twin is an accurate computer generated virtual model that replicates an system, object or process from the physical world, using real time data. Examples could include a jet engine, a sports stadium, a wind farm or an organ such as a heart. There is no universally agreed definition, but this description applies broadly to the technologies discussed here.

Unlike traditional virtual models or simulations, a digital twin is continuously connected with data from its physical counterpart. This data can come from sensors and creates a dynamic two-way flow of information.

Insights from the digital twin can be used to understand the design, function and operation of the physical object or system, and how its characteristics and performance can be changed.

How do digital twins work?

A digital twin has three core components, illustrated in the diagram:

• the physical entity (the object, process, or system) that exists prior to the digital representation, and can function without a digital twin
• the digital representation
• a continuous data feedback loop that connects them, collecting data from the physical entity

Real-time data from the physical object is transmitted to the digital twin and used to optimise and simulate conditions. By linking them, the digital twin enables a more accurate understanding of the physical system. For example, a digital twin of a heart could be slowed, stopped, and restarted to explore its functions. For some digital twins, the physical and digital systems can evolve and optimise together.

Digital twins are more complex than traditional simulations because they can represent a virtual environment composed of multiple simulations, each modelling various processes. This allows a more accurate and dynamic representation of real systems, and improved detail and realism.

Advances in artificial intelligence (AI) and machine learning have contributed to the growing interest in digital twins. This is partly because AI could help to develop new types of understanding, generate new insights or help to write code faster for data analysis.

Types of digital twins:

• Object: a digital twin of an aircraft could use real-time data from physical sensors (such as atmospheric data) for analysis and decision-making. Digital twins can consider multiple factors like the internal cabin environment and its potential impact on passengers. A twin can be used to assess the performance of aircraft features, to inform maintenance practices.
• Process: digital twins can help to optimise processes, such as warehouse management and database analysis. This allows retailers to deliver more efficient, streamlined services tailored to their customers. This is seen as an important way to maximise productivity.
• Complex system: biological organs are highly complex, comprising different cell and tissue types. Surgeons can use digital twins of organs (for example, the heart) to practice techniques, to reduce the risks associated with surgery.

Which sectors are using digital twins?

Estimates vary, but the global digital twin market could grow by up to 45% annually between 2023 ($13-16 bn) and 2030 ($138-195 bn). Rapid growth is anticipated, driven by the integration of digital technologies into manufacturing and industrial processes (Industry 4.0).

In recognition of the importance of this technology, the Government Office for Science published a technology assessment on digital twins in November 2023, noting applications across several sectors including:

• water
• energy
• transport
• health

The assessment noted that as the technology develops it will have applications in defence and national security. The need for national capability in digital twinning was outlined in the 2021 Integrated Review of Security, Defence, Development and Foreign Policy. In common with many technologies it can be dual use, with applications in civilian and defence contexts across both the private and public sector. An emerging technologies review for Government noted that that digital twins produce “relatively large projected economic impacts relative to investment”.

Global tech companies, such as Meta, are contributing to growth, by creating a virtual environment called the metaverse, with digital twins playing a key role.

In 2018, the Department of Business and Trade launched the National Digital Twins Programme (NDTP), to foster research and applications, standards and processes, and a marketplace. This initiative involves a wide range of stakeholders, from central and local government, universities and the private sector. One example is the collaboration on the Isle of Wight to deliver energy autonomy on the island through a Virtual Power Network (VPN).

Digital twin applications

• Healthcare: encompasses a wide range of applications, from drug development and clinical training tools to optimising healthcare system performance. The NHS recognises the potential of personalised medicine, such as treatments tailored to each individual’s unique biology. Digital twins offer potential benefits in drug development or to enhance the safety and effectiveness of surgeries, enabling practitioners to refine their skills using computational and robotic tools. According to the Government Office for Science, digital twins could be used to improve efficiencies in managing waiting lists, by enhancing decision-making processes.
• Infrastructure: digital technology in modern construction is reshaping transport systems and the development of smart cities. The Digital Roads project, funded by the Engineering and Physical Sciences Research Council (EPSRC), can assist with maintenance and repairs, and offer benefits for road users, such as accurate traffic alerts. The UK Parliament building has a digital twin, modelled for the Restoration and Renewal programme.
• Service industries and commerce: Retailers are exploring digital twin technology to enhance customer experiences, optimise operations, and improve decision-making. In retail, digital twins can create virtual models of stores to analyse customer behaviour, optimise product placement, and streamline inventory management. For example, Amazon uses digital twins to predict customer spending habits based on real-time browsing and past purchases. In other contexts, digital twins are being applied to improve safety and customer experience, particularly in Smart Stadiums.
• Defence: Digital twins are important for the UK’s defence as they enable the simulation of real-world systems in a virtual environment, allowing for safe testing without real-world risks. This technology can reduce development times and costs. By facilitating the development of advanced military assets, digital twins contribute to enhancing the UK’s defence capabilities.

Challenges

The potential capability of digital twins is not yet matched by the maturity of the technology. Developing and integrating this technology presents challenges, including:

• Technological limitations: A key challenge in the implementation of digital twins is the dependency on reliable internet of things (IoT) connectivity and the complexity of simulations and modelling. The digital twin industry is heavily reliant on advances in fields such as machine learning and AI to enhance model complexity and improve the accuracy of digital representations of the physical world. For instance, AI systems must be reliable, minimising deviations from real-world data to ensure that a digital twin is an accurate representation of the physical entity and to minimise uncertainties within the simulation.
• Cyber security: Digital twins use and create large amounts of data, some of which may be sensitive. A twin built on data from multiple systems (such as a transport network) may require higher security than its constituent components. Data protection is a critical consideration as is the potential for their susceptibility to a range of cyber threats. Attacks on digital twins, or on sensors in the physical object, could lead to effects in the real world and safety of systems. The National Cyber Security Centre has published guidance on their secure development and operation.
• Cost: Due to the vast amount of data, there is a significant computational cost, including data storage and processing power. Additionally, the time required for deployment and integration with physical systems, such as sensors and software, is often a limiting factor. However, these costs are expected to decrease as the technologies develop and become more efficient.
• Ethics & legality: Digital twins raise ethical challenges, related to their accuracy and representativeness, and any bias. Decisions based on flawed digital twins could disproportionately affect certain groups. The transparency of algorithms, informed consent, and accountability in the development and use of digital twins also present ongoing ethical challenges that stakeholders note require careful consideration and regulation, especially in relation to a digital human and personalised medicine. From a legal perspective, digital twins raise concerns related to data privacy, accountability of decision making, algorithmic integrity, and intellectual property.
• Accessibility and collaboration: The complexity and diversity of digital twins, and the approach required to develop them requires:
  • • collaboration across multiple projects and teams
    • work between academia and the private sector
    • involving people with different skills and expertise.

  This can pose challenges for smaller groups to access the technology and may lead to communication issues between collaborating teams on digital twin projects. Programs such as the NDTP seek to address this through investments to facilitate effective collaboration.

Research investments

Research investments come from both the private and public sector, the latter through UK Research and Innovation (UKRI) and its constituent bodies.

UKRI’s Digital Twinning Network Plus seeks to transform the UK’s national capability. Research councils have numerous investments, including:

• Innovate UK, as part of Connected Places catapult project is working to support collaborative development, innovation, and improve access to tools for creating digital twins, in partnership with UK companies such as Artemis Technologies and Thales.
• The Natural Environment Research Council will allocate £2.8 m to several upcoming projects within the TWINE Earth Observation Investment Programme, which focuses on using digital twins for environmental applications, including:
  • SyncED-Ocean, a digital twin of ocean ecosystems
  • FlOODTWIN, a project that forecasts water-related hazards
• The Engineering and Physical Sciences Research Council (EPSRC) invested £100 m in 2024 to the Alan Turing Institute for data analysis and AI. The Turing Research and Innovation Cluster in Digital Twins was created, for projects like Neural-symbolic AI for digital twins and mathematics behind digital twins.

In addition to government investments, tech companies such as Microsoft’s Azure and Google’s Supply Chain Twins are making significant investments in digital twin programs. This is driving growth in digital startups specialising in digital twins projects.

Further reading 

• National Digital Twin Programme – Department for Business and Trade
• Digital Twins Network – DTNet+ – UKRI funded research network
• Digital twins: secure design and development – National Cyber Security Centre
• Digitally rebuilding the Houses of Parliament – TechInformed
• The rise of digital twins – Nature
• Navigating modern business digital Twin vs simulation – Program ACE

Acknowledgements

POST would like to thank the following peer reviewers for kindly giving up their time during the preparation of this article:

• Professor Varuna De Silva, Loughborough University and Thematic Research Lead in Digital Technologies and AI, UK Parliament
• Dr Charith Perera, Cardiff University
• Andrew Peck, Loughborough University

Photo by: Gorodenkoff, via Adobe Stock