skip to content

Centre for Digital Built Britain completed its five-year mission and closed its doors at the end of September 2022

This website remains as a legacy of the achievements of our five-year foundational journey towards a digital built Britain

From 2018 to March 2022, the Centre for Digital Built Britain led the National Digital Twin Programme on behalf of the UK Government. Securely sharing infrastructure information through a digital ecosystem can support better outcomes for all, as illustrated in the poignant short film Tomorrow Today (produced ahead of COP26). 

Early on, a CDBB task group developed the Gemini Principles as the ‘conscience’ of the National Digital Twin: nine guiding values, to ensure that these digital twins have purpose, maintain trust and function effectively.  While the uptake of the Digital Twin concept grew rapidly, CDBB wanted a show how these criteria could be met in reality. 

An interdisciplinary team based at the University’s West Cambridge campus took on the challenge, with CDBB funding researchers at the Institute for Manufacturing (IfM) and the Department of Computer Science and Technology (Computer Lab) to develop a digital twin demonstrator and research facility.  The project enabled practical in-the-wild exploration of the challenges associated with production of a National Digital Twin. 


Researchers needed to gather building information from the range of systems used on campus, and to ascertain what additional information would be needed. Three buildings were identified as potential case studies. These included the Civil Engineering Building (opened in 2019), built with five sensor packages integrated to monitor the whole-life performance of the asset as part of research carried out by the Centre for Smart Infrastructure and Construction (CSIC). The pandemic impacted what was possible and some sensor solutions and processes were necessarily developed in remote domestic labs (researchers’ home offices), with the Alan Reece and William Gates Buildings functioning as the final campus demonstrators.

Researchers met with estates, facilities and building managers, to understand their work and how better information about buildings and assets might help them perform their roles more efficiently. Could common problems be anticipated, issues spotted in real-time? How could human expertise be accounted for within the DT systems? Interviews yielded data about recurring issues, and the typical processes to detect, diagnose, and repair issues arising. Taking note of known historic events, researchers were later able to compare chronological data records. Referring back to specific events enabled them to apply Bayesian changepoint detection algorithms and automatically discern which types of data could help raise an alarm if similar events occurred in future. 

For the computational work, carefully-designed infrastructure was an essential starting point. On the West Cambridge campus, a pair of roof-top aerials combine with judiciously-placed compact gateways indoors to capture sensor data. Low-cost, low-energy, and finessed to overcome minor challenges posed by layout and electrical interference, the results were persuasive enough that the University Information Service agreed to support a campus-wide WiFi network for the devices.

Embedding machine-learning was necessary because building information systems easily record too much data, or the wrong data. The researchers also tackled issues with data size by reprioritising what happened to fresh data. The first thing that most building systems do with data is send it to storage. In storage, queries are applied. This means the time between data capture and query answer is automatically slower than if the data had been made available immediately. If digital twins are to provide real-time information, being up-to-date matters. Many of the decisions researchers made were guided by the principle of timeliness. 

Another key variable is space. When drawings produced for construction fell short of asset management needs, researchers had to gather spatial information independently (surveying). Designing a system that could be fed into an inter-operable national system, all locations were captured with appropriate building-relevant coordinate systems, and a digital architecture that allowed each of these to be consistently normalised and related via the global latitude, longitude, altitude system (WGS84). Integrating accurate spatial datapoints made it possible to construct complex queries (e.g., interrogating neighbouring sensors).


Throughout, the work met significant support and interest from external stakeholders, some of whom became collaborators. For example, via SmartCambridge the Greater Cambridge Partnership of regional authorities have been close observers of the West Cambridge Digital Twin platform and the LoRaWAN sensor data network, with Computer Lab learning now informing plans for smarter regional infrastructure.

Bentley, GeoSLAM, and Topcon assisted with scanning processes, generating clear accurate information about the Institute for Manufacturing building and drone scans of the wider West Cambridge campus. Bentley also collaborated with CDBB researchers on development of the Digital Twin platform. RedBite assisted with deployment of the asset registry network, enabling building end-users to submit feedback via QR codes. 

Outcomes and outputs

The Adaptive City Platform (ACP) built by researchers focuses on processing real-time data, taking a stream-processing approach. Stream-processing differs from other approaches because it reacts to events (indicating some change of state) rather than merely collecting and storing messages; and because it prioritises real-time processing of the data (and resulting actions) rather than other tasks (e.g., archiving the data). Archiving still happens, but as a secondary task. 

A significant part of the ACP is the Watcher system, providing end-to-end complex event detection, with reference to spatial and temporal information. An anomalous report from one sensor may trigger a task to check and compare data from neighbouring sensors, testing hypotheses about the source of the anomaly (a blocked radiator, an open window, etc.). The complex event detection can also take account of variables such as time of year: high room temperatures may be natural in summer but the sign of a broken heating system in winter. 

The researchers did not neglect the DT commitment to uphold Gemini Principles. Camera-based sensors coupled a Raspberry Pi processor with other off-the-shelf technology to create a custom machine-learning computer vision platform that transforms video footage into anonymised data about the movement of people and vehicles in public spaces in real-time. Because all the data-processing occurs within the sensor, the raw video footage can be thrown away. The DeepDish camera can be applied to tasks such as calculating occupancy without exposing the identity of the occupants.

The researchers also devised algorithms in response to specific challenges identified by building managers. These allow for automated detection of anomalies and faults, with capacity for fault diagnosis. A strategy for granular building energy metering has also been developed. This combines two standard taxonomies (IFC and Brick Schema) to associate submeter readings with the relevant spaces. The result is a more comprehensible account of what energy is being used where.  

All the above feed into a live web-based application that enables end users (facilities managers, those with offices in the case study buildings, etc.) to view appropriate levels of information. By simply integrating basic employment records and up-to-date information about space allocation (whose office is whose), the system can ensure people can retrieve their own data.

All these processes are documented in 12 articles and working papers available in the CDBB Knowledge Database Navigator. An accessible account of the strand of research led by the Computer Lab team is available separately as CDBB West Cambridge Digital Twin: Lessons Learned.

So what?

Connecting digital twins requires new technologies and processes to curate, manage and enable interoperability of data from various sources. CDBB research has conclusively demonstrated possibilities for automatic asset anomaly detection by combining data collection with data analytics and new algorithms. Approaching the challenges of building data management systems and frameworks with the Gemini Principles in view, innovations such as the DeepDish camera serve the public good and can enhance the public good of all digital twins. 

As Cambridge building managers have been discovering, perhaps especially as an outworking of the COVID-19 pandemic, having accurate data and analytics about the real performance of assets allows better decision-making in the operation of buildings. The knowledge obtainable through the stream-processing platform should also influence design, construction and maintenance of future buildings to support productivity, efficiency and resilience, drive decarbonisation and address resource constraint.  

Throughout the research process, academics have liaised closely with external stakeholders.  

Industry impact

Researchers developing the West Cambridge Digital Twin Research Facility were integrated across the Construction Innovation Hub (CIH), CDBB and the Digital Framework Task Group. This integration has meant that learning and findings were shared iteratively in collaboration with industry. The application of new technologies and methods to specific sites at the University of Cambridge means the University Estates service and those responsible for managing buildings across the West Cambridge campus can already speak persuasively about the benefits for built asset management. Application of sensors and CDBB processes to inhibit the spread of airborne infection is a case in point (see below).

As different factors combine to increase the cost of energy, it is easy to imagine fresh applications for the sensors to understand how, where and when energy is being consumed and identify opportunities to reduce energy use. The agnosticism of the framework and interface—designed for flexibility and dynamism—means that such queries can be developed and refined rapidly, informed by real-time data. The whole infrastructure was deliberately developed to have a low-energy footprint. For instance, the low-power aerials that receive information from the network of sensors are both battery-powered, requiring a new battery once a decade. 

The built-in flexibility and dynamism required at the outset mean that the potential impact of this work extends beyond what we can yet imagine. The most significant known achievement of this project is probably the shift in latencies: where instrumented buildings relying on periodic data updates may take perhaps 5 minutes to detect a change in room occupancy, the Long-range Wide Area Network built for CDBB’s demonstrator takes up to 1 second to register the same change. This is the real time data that building managers desire.

The West Cambridge Digital Twin was the first twin to be registered on the DT Hub. Its developments have been shared and will continue to be shared across the digital twin community, with hundreds of industry leaders and organisations keen to learn from these experiments.    

“Decision-making, determining what is worth doing and when is at the heart of asset management. Having appropriate asset information with the correct attributes and of the correct quality and available to the right people at the right time is an extremely important factor in decision-making. When managing an estate as diverse and complex as that of the University of Cambridge, the difficulties associated with asset information are compounded. The introduction of Digital Twins will facilitate the provision of the correct data at the correct time and in the correct format, improving the effectiveness and efficiency of the decision-making process.  Ultimately, it will allow us to derive the best trade-off between cost, risk and performance over the life of the Estate, thereby delivering optimal value for the University and its myriad stakeholders. The West Cambridge Digital Twin project is an important step forward for Estate Management and provides the foundation upon which to deliver maximum value, from assets and facilities, for the University, the City and wider society.”  
Christine Leonard, Facilities Asset Strategy & Compliance Manager, University of Cambridge 

Wider benefits

For the full promise of this research to be realised, it is necessary for industry (construction, estates and asset management) to commit significant resources to change. 

The COVID-19 pandemic created many challenges for the researchers. It also created a unique opportunity to show how sensors are deployed promptly and inexpensively to address a specific need: ventilation data. 

The case study sites in West Cambridge host thousands of sensors. Challenged by the airborne virus threat, some of those sensors rapidly became a proxy for detecting unsafe occupancy levels. Dr Ian Lewis (Director of the Adaptive Cities Programme in the Computer Laboratory) explains: 

“It was genuinely surprising how much the formative Digital Twin platform allowed us to pivot to dense monitoring of CO2. Characteristics we’d designed in to give us highly time-responsive information related to meaningful building assets (i.e. spaces, offices, lecture theatres) meant we could maximise our use of the space remaining within safe limits. 

“We were able to demonstrate what ‘great’ could look like and derive immediate advantage from that in the Computer Lab. Those techniques couldn’t possibly be common today, but we’re setting the direction for techniques that will be considered normal in the future.”

Informed by the Lab’s example, other departments were required to obtain and use CO2 sensors. While none shared the sensor density of the two case study sites, educated guesswork about safe levels across campus was informed by West Cambridge DT parameters. Its baseline information helped real-time decisions about safety. Such rapid deployment attests to the economy and applicability of the solution and to institutional capacity to adapt when there is sufficient motivation. 

Building manager Ali Digby reflects on the significance of the research for her work:

“Over 450 sensors have been deployed in the William Gates Building since March 2021. The sensors measure CO2, humidity and temperature in their immediate surroundings and have been placed in offices, teaching spaces, meeting rooms and common areas. 

“As Building Services Manager at WGB, access to the real-time data produced by these sensors has been invaluable, especially whilst navigating the reintroduction of in-person teaching and reoccupation of the building during and following the COVID-19 pandemic. 

“Immediate response and actions, such as increasing ventilation to rooms, could be taken when CO2 levels being monitored started to rise towards the ‘safe’ thresholds, set by University policy, within multi-occupied spaces. Different mitigation strategies, to prevent levels getting too high, could be trialled for effectiveness in different spaces with the benefits measured and recorded in real-time. Gathering this information so quickly meant that best practice for multi-occupied offices and teaching spaces could be established, with data to back up the strategy, and presented to the wider Department. This has guided the introduction of policies relating to lowering the risk of communicable diseases within the building. 

“I have found the real-time Building Information Management platform intuitive to navigate, allowing me to get a quick overview of the whole building, spaces within the building, or the area surrounding a specific sensor. This, in turn, means I can make preventative changes to the Building Management System before large events and reactive changes when needed.”

It is a highly specific example, but one with global relevance. Digital twin systems using the West Cambridge framework are versatile and interoperable, with unlimited potential.

The researchers have delivered a demonstrator of the technical competence and created significant innovations in terms of machine-learning and the handling of real-time data from the digitalised built environment. The conditions for realising that promise through the National Digital Twin must also overcome socio-technical obstacles. As the researchers outline in their account of Lessons learned, to reap the greatest benefits, asset owners will need to be willing to survey and (re)instrument their assets and not rely on building information from the (pre)construction phase.

Data from digital twins informs better policy, planning, and management decision-making on the interaction between the built environment and the economy, society and the natural world. Digital twins are already improving organisational safety, productivity and efficiency. 

The researchers have developed digital twin-prototypes that address real situational issues. For example, a paper delivered by Dr Xiang Xie in October 2021 highlighted how spatial and semantic data can be combined to understand energy use in different parts of a building. Acting on these recommendations can help those creating and managing smart buildings to reduce energy use, saving costs and cutting carbon emissions.

Digital twins also provide the foundation for integrating city-scale data to optimise city services to provide better outcomes for people. The NDT can release value for society, the economy, business and the environment at scale, with the Adaptive City Platform providing a low-energy cost-effective framework with an integral commitment to privacy.  

Looking ahead

Researchers are carrying the learning into further funded projects with other collaborators, ensuring the research is applied and extended:

TwinAIR (Digital Twins Enabled Indoor Air Quality Management for Healthy Living) is a Horizon Europe project, collaborating with 22 academic and industrial partners across the UK and EU. It aims to introduce a technological solutions system to improve air quality in a wide spectrum of indoor living activities (residents, workplaces, transportation, hospitals etc.). The asset management team at Cambridge has three key work packages, contributing to (a) indoor air quality monitoring through the development of physical, digital sensing, air filtration and wearable devices; (b) development of the Building Assets Operation, Control and Maintenance strategy for improved air quality and energy efficiency; and (c) development of the Data Management Platform as well as the Digital Twin Platform.

Digital Hospital: Funded by Cambridge University Hospitals NHS Foundation Trust, the Digital Hospital project seeks to exploit the potential of digital technologies to address key challenges in the delivery of effective and efficient healthcare provision at Cambridge University Hospitals (CUH). By devising, implementing, validating and exploiting Digital Hospitals as an integrated concept, we intend to combine patient records, processes, equipment monitoring, occupancy and infrastructure use to improve hospital performance.  

TwinBAS: The project aims to demonstrate a combination of novel technologies in the form of a Digital Twins enabled Building Automation System for Improving Indoor Environmental Quality (IEQ) parameters and optimizing comfort conditions and energy performance. The conventional “one-size-fits-all” comfort management is changed for good and personalized comfort goals are delivered for current and future office buildings.


  • Estates Division, University of Cambridge 
  • University Information Service, University of Cambridge
  • Smart Cambridge 
  • Bentley Systems
  • GeoSLAM
  • Topcon
  • RedBite

Project Status