UK urged to build 'digital twin' of its infrastructure

Using the power of big data and artificial intelligence (AI), it may be possible in future to determine what changes need to be made to the way the public uses transport if the population on London were to increase by 50% by 2050, for example, the NIC said in a new report (76-page / 2.94MB PDF).

"A digital twin model can provide insights beyond what is currently seen with existing infrastructure models and can be used as a tool to aid decision making," the NIC said. "This twin will capture data on infrastructure, which can be augmented with models, predictive asset management approaches and advanced data science to optimise the performance of infrastructure systems."

The NIC said a digital twin model could facilitate the planning of maintenance works on UK roads and railways in a way which minimised disruption to consumers, or allow those planning to install new infrastructure to virtually "overlay" their plans onto existing infrastructure "to show interdependencies and evaluate the optimal timing of investment". The 'digital twin' could also be used to "identify efficiencies in energy use which can be tested across different sectors through simulations", it said.

"Through the input of verified data to these digital twin models it will be possible to develop a richer understanding of the way the infrastructure system works," the NIC said. "Applying deep learning techniques may offer extended predictive capabilities."

A six month pilot initiative should be carried out "to explore and experiment with the benefits of building a digital twin of a specific geographical area", the NIC said. The Centre for Digital-Built Britain (CDBB), Alan Turing Institute (ATI) and the UK Infrastructure Transitions Research Consortium (ITRC) should all have roles in the pilot, it said.

A review of the pilot should be completed by October 2018 and lessons from it used to inform "any future development of larger-scale or more complex digital twins, and the most effective institutional structures to support continuing progress in this area", it said.

The pilot scheme should look at issues such as security and data protection, and comprise collaboration between the public sector, businesses and academia, the NIC said.

"A successful digital twin pilot project will demonstrate the value of better coordinating and sharing data about infrastructure assets," the NIC said. "A digital twin will be able to offer solutions to difficult decisions arising from population growth, congestion, climate change and the development of new technologies. Collaboration across academia, the technology industry, network operators, utilities, relevant public sector bodies and consumers will be key to making a digital twin pilot project work."

"The task of modelling existing infrastructure in the UK is arguably more difficult than in other countries with newer infrastructure systems, yet ultimately all national digital twins will need to encompass existing infrastructure and the UK has the combination of academic and private sector data science and AI research and development capability to lead in this area," it said.

In its report, the NIC described data as "as much a critical component of national infrastructure as steel, bricks and mortar", and said real-time data can be used to "inform how infrastructure is operated on a second-to-second basis".

A report by Pinsent Masons, the law firm behind Out-Law.com, earlier this year highlighted a mix of views within the infrastructure and technology industries about who typically owns data in 'infratech' projects and who they believe should own that data.

The survey found that 91% of respondents wish for an 'open access' approach to data among collaboration partners, but that in reality just 62% of projects operate such a model. In other projects, the owners of infrastructure, infrastructure companies such as engineering or construction companies, operations and maintenance providers, technology companies, procurers and regulators all typically own the data generated, it said.

This complex picture was reflected in other results from the survey which showed that 38% of respondents from technology companies and 34% of respondents from infrastructure businesses identified agreeing data requirements or standards in infratech projects as a challenge.