What makes up a digital twin and how can manufacturers best utilise their capabilities? Rab Scott, Professor of Industrial Digitalisation at the AMRC, explains more.
Open any manufacturing magazine these days and you are likely to encounter the term ‘digital twin’. To some people this may be a new term or concept, but actually digital twins have been around in manufacturing since the last century, it is just that in those days, they weren’t called digital twins. The computing power that was needed to create them was very expensive, the data that these twins ingested wasn’t easily available and the business case for spending time and effort on them wasn’t necessarily well understood.
Those days are gone. Since then, the reduction in the cost of the compute predicted by Moore’s Law has happened; miniaturisation and mass production of sensors has allowed the commoditisation of data capture devices; and the availability of connectivity through any number of communication protocols all lead to the case for the adoption of digital twins now making economic and business sense.
But there is still a problem in that many manufacturers don’t necessarily understand what a digital twin actually is or where the value of a digital twin lies.
Digital twin being created at the AMRC’s Factory 2050
Context and connectivity
In a report published by the High Value Manufacturing Catapult in 2018, a survey of engineers showed that there were a number of components of a digital twin that they considered essential, some nice to have, and some not necessary. Since then the common understanding of what makes up a digital twin has matured but it still reflects what that report in 2018 highlighted.
The main thing that a digital twin has which a simulation model doesn’t is real-time connectivity to the physical asset, process or system. And there lies the heart of the matter – unless you have a real asset, you cannot have a digital twin. You can have a high-fidelity simulation model of a virtual product process or system, but you have no way of validating it in the context of the real world.
In fact, a recent publication from the American Institute of Aviation went so far as to state: ‘No physical asset equals no digital twin’ – mirroring the opinion we published in our report entitled Untangling the Requirements of a Digital Twin in October 2020.
Where next for the digital twin?
In 2020, Gartner suggested that within five years, more than 50% of products would come with a digital twin and that, in the same timeframe, 75% of composite twins will be created from interoperable individual twins.
The National Manufacturing Institute Scotland, Advanced Forming Research Centre, has used digital twinning to help the Scotch whisky industry reduce inventory waste when filling casks
But therein lies a problem – as yet, there are no standards for digital twins either in the format or their architecture. In a recent report by the Centre for Digitally Built Britain (CDBB) – The Pathway Towards an Information Management Framework – they look towards a Foundation Data Model. The clarity required for this model is a step forward for digital twins, as only by having a coherent framework for a digital twin can we hope to merge individual twins into a composite twin.
There is, however, a contrary opinion to merging individual twins into a composite twin. Some companies believe that assets may have many twins of their own developed for different use cases and that these could have different access permission levels that can then be utilised by other twins. This is sometimes referred to as the ‘Russian doll effect’ of information that is available by an asset, for example, the deeper you can go with the right permissions, the more you can find out.
The challenges ahead
Whichever strategy for the development and deployment is adopted, there will be challenges associated with the monetisation of digital twins; if the digital assets are passed along the value chain, the actor in the ecosystem which extracts value from the digital twin may not be the same entity that invested in the original creation.
As well as value extraction, there are questions about liabilities with inherited digital twins. These challenges will come into greater focus as the adoption of digital twins increases.
Irrespective of which architecture and approach is taken, the one thing that remains consistent is that a digital twin mindset is not about a specific technology, the mindset is about a methodology or an approach which helps to answer the question for manufacturers of how they can make better informed decisions to improve their products or processes and thus improve their bottom line.
If you need help to think through how digital tools could help your business, come and talk to us at the High Value Manufacturing Catapult.
More information www.hvm.catapult.org.uk
Rab Scott is Professor of Industrial Digitalisation and Head of Digital at the University of Sheffield Advanced Manufacturing Research Centre (AMRC)