We introduced the principles underlying the National Data Strategy on the Data Impact blog in late 2020. Here we introduce the ideas behind digital twins and share some examples.
Last year, the UK government published its National Data Strategy, which it hopes will drive the UK economy forward, by enabling access to new data sources, the use of standards and by fostering skills development in the data arena. The government have also plans to modernise the access and interoperability of data held within its remit.
One of the items the National Data Strategy discusses is the use of digital twins.
Digital twins are virtual representations of a physical object or process. They are used to provide insights that support good decisions in the physical world.
A brief history
Digital twins are not a new concept.
In the 1970s, physical duplicates or ‘twins’ of spacecraft were used by NASA to duplicate what was happening to the real craft. Such a twin helped scientists identify issues and potential solutions when Apollo 13 and its crew were in peril.
Once computer-aided design was invented over 30 years ago, modelling of ‘twins’ developed further, initially existing to visualise what an object could be, once manufactured.
As technology and data capture has advanced, so has the capability to take real world objects and processes and model how they would behave in the real world.
The term ‘digital twin’ was introduced in 2022 by Michael Grieves, who applied the concept to the manufacturing product lifecycle.
More recently, this concept has been extended to allow for recording of information from sensors and other sources.
The data from these sources are then attached back onto the original designs, helping users to understand how the object or process is behaving in the real world. This further allows monitoring of whether the object or process is responding as expected.
Early uses of digital twins included the building and infrastructure communities.
Digital twins are now also used to
- enable modelling and predictions of events
- provide a basis to monitor the built environment
- store the history of an environment
Where are digital twins used?
While there are many applications of digital twins in existence, a few examples indicate the breadth of possible applications.
Social distancing
During earlier stages of the Covid-19 pandemic, a digital twin was created of London’s St Pancras International station.
The digital twin used camera, data and computer vision software with the aim of identifying quieter areas to direct people to, as well as providing information on when lifts and escalators needed to be redirected to support social distancing.
Cars
The electric car company Tesla creates a digital twin of every vehicle it builds.
Sensors in the car feed back to the digital twin and AI is used to analyse that data, determining whether there are maintenance or other issues. As part of an iterative and reciprocal process, decisions arising from the digital twin and associated AI can then feed into decisions around software or temporary configuration updates which can be uploaded back into the car itself.
The Living Heart Human Model
In the United States, a high-fidelity model of a healthy human heart has been produced.
The digital twin can be modified to explore the effects of potential defects and diseases, as well as to explore potential treatments and how the heart would respond to the insertion of medical devices.
Digital twins at city scale
Carson City (Nevada, US) has long been modelling drought levels to help manage demand at peak times.
More recently, a digital twin of the city’s water needs and usage has been built, allowing the city to model different scenarios of peak usage by different sectors and identify ways to rebalance supply levels.
Singapore is developing its own twin. Their aspiration is that
Virtual Singapore will be the authoritative 3D digital platform intended for use by the public, private, people and research sectors. It will enable users from different sectors to develop sophisticated tools and applications for test-bedding concepts and services, planning and decision-making, and research on technologies to solve emerging and complex challenges for Singapore.
In a similar manner, the Shanghai Urban Operations and Management Center has created a digital twin of the city, including 100,000 different elements and with data updates supplied by satellites and drones.
Helsinki has developed a twin, but has also added an additional slant to the concept.
As well as allowing modelling of interactions within the city, they are also using the twin to allow virtual tourists the opportunity to explore the city, without physically being there. This includes the opportunity to visit a souvenir shop and have purchases shipped to your home address.
Data needs of digital twins
The UK Data Service holds a collection of data which tell us about society at any one time.
The biggest of these, of course is the census, which covers almost the whole population of the UK and is a unique snapshot of the constituent countries every ten years. As a snapshot, it is however limited in providing the sort of regularly updated data which would a larger digital twin would need.
Some efforts are made, including annual population updates and estimates.
The census may also change in the coming years. Professor David Martin, Deputy Director of the UK Data Service, recently wrote for the Data Impact blog about what some of those changes might look like – some of the possible solutions may begin to overcome that gap.
The UK Data Service is also home to a range of surveys and longitudinal studies – the British Social Attitudes Survey, Understanding Society and National Diet and Nutrition Survey to name a few. These data may also prove useful in the development of future digital twins, used within the strict conditions needed to maintain people’s anonymity.
Again these data provide snapshots which may be useful but cannot provide the coverage that a digital twin might need.
But how do we get a picture of what the population is doing over the course of a day?
Population24/7 is a project developed by researchers at the University of Southampton (including David Martin) which explored this issue. The researchers looked at the different sources of data that could be used to estimates of the population at specific time periods of the day, such as the night time, the work day or weekends.
They have produced models in the form of maps which can be used to understand the distribution of the population over the day. The outputs are available as open data from the UK Data Service.
Looking forward
It’s still relatively early days for the developing concept of digital twins. No doubt, technology will develop further and, alongside it, the requirements for a range of data to underpin and enhance the model.
It will be interesting to see what new ideas are developed and how data collection will evolve to support these models.