Digital Twins in the Built Environment

In manufacturing, digital twins are digital representations of physical products that store all relevant information over a product’s life-cycle. Digital twins allow meaningful integration and structuring of data collected about products and leverage advanced sensing and analytics to provide real-time insights about a product’s status and performance. Digital twins may also provide input data to models so that the behaviour of a product at any stage in its life-cycle can be predicted. With these features, it is not surprising that many engineering products are now managed, and often operated, through their digital twins.

In the built environment, advanced sensing technologies and the Internet of Things (IoT), contribute to making buildings and infrastructure assets major sources of data. Like other domains, data can only be processed to yield insights only when relevant contextual information exists. Unlike other domains, with most of the building stock and infrastructure in Europe predating modern digital design, contextual information is missing or unusable. This makes the cost of generating, exploiting and sharing digital twins of buildings and infrastructure at a mass scale prohibitive.


Building Information Modelling

Building Information Modelling (BIM) can provide the context for representing infrastructure assets, including buildings, as digital twins. BIM involves the generation and management of digital representations of physical and functional characteristics of physical infrastructure assets, such as buildings, plants, roads, bridges, ports, tunnels, and utility networks. At present, BIM allows stakeholders to extract and exchange information for networked decision-making; individuals, businesses and government agencies can use BIM to plan, design, construct, and operate physical infrastructure for individual projects. Hence, BIM offers infrastructure owners the potential to compile digital twins bringing all types of data together and allowing their cross-platform exploitation by Architecture, Engineering, Construction and Facility Management (AEC/FM) stakeholders.

BIM usage to design, engineer, and deliver new facilities in the AEC/FM sector rose from 13% in 2010 to 74% in 2018, and its use is spreading inexorably. However, the way BIM is currently understood and used by AEC/FM sector professionals is entirely constrained by the narrow view imposed by local commercial imperatives and the extent of the current imagination. For most infrastructure owners and practitioners, BIM is still an information process and a set of technologies that can be used to optimise the design and delivery of individual construction projects. The idea that BIM models can be thoroughly verified through engineering analyses and operational performance simulations before construction, and thus provide the AEC/FM sector – which has never enjoyed the benefits of rigorous prototyping – with rich virtual counterparts, is starting to be talked about but barely understood. The notion of compiling multi-dimensional models that can support the entire asset lifecycle, and that such models can be mined for data and used for machine-learning, requires a leap of imagination.

BIM is still rapidly evolving both conceptually and in practice. Models are often proprietary and not widely available, design models are seldom updated during construction, interoperability problems are abundant, and only ~2% of the existing building stock has been designed with BIM. The slow pace of manual, incomplete and poor-quality digitisation can be accelerated and improved through the methods proposed here. This complex task requires an understanding of the multiple application requirements, overcoming the limitations of evolving BIM standards, and development of new technologies to generate, enrich and exploit models in the cloud.