What to do when the twins start multiplying?

[Ian Gordon]

When asked by a relatively senior member of staff here what the Digital Twin is all about, and why they should care, I pulled together some SmartArt (pictured) to try to explain the component parts of an infrastructure organisation's twin. Keen to get the wider community's thoughts on this approach.

Digital Twins are having a bit of moment here at Highways England, to the extent that our principle risk is not a lack of twins, but a surfeit of incompatible twins. I'm beginning to think that the ‘Digital Twin’ of a complex organisation such as HE will actually need to function as a hierarchical system of systems.

We need to understand how our organisation functions and what it looks like from a conceptual data perspective (the Schema), we then need a single source of truth, preferably one structured as a graph to reflect the Ontology (the Data), and finally there will be the specific manifestations of the above for different parts of the business (e.g. BIM, digital product catalogues, design, porfolio management etc. etc.) which should be united by the common schema and data above.

[Samuel A Chorlton]

@Ian Gordon really interesting post and a conversation that I'm sure has been had in more than one organisation. My personal feeling is that although in a ideal world we would have a single twin that represents the entirety of an organisation the reality of the situation is that is just not feasible and therefore the hierarchical system is much more likely to reflect what we deliver. This is the approach that we are intending to take with the National Digital Twin ( @Miranda Sharp and @James Harris ). Structuring it in a hierarchical manner also presents an opportunity to limit the data that is exposed through the system.  

One question that I always struggle to answer is how do we ensure that assumptions and errors are properly propagated through the system so that the decisions making process is able to correctly recognise these. 

[Peter El Hajj]

@Ian Gordon great topic.   I voted for 'As many as it likes, as long as they talk to each other'.  Although my preferred option would've been 'as many as you like, as long as they have a clear purpose and they talk to each other.' 

The IMF pathway used the term "composite twin" to refer to individual twins coming together to fulfil a higher purpose. 

I came across this blog recently, which had a good representation of discrete twins and composite twins:

 

 

[Ian Gordon]

How useful do you think that the parallels in sectors such as manufacturing and aerospace are likely to prove in infrastructure @Peter El Hajj?

It feels like the established use cases for Digital Twins are usually twins for 'things' rather than 'systems'.  Does the twin approach for a factory (e.g. a system of defined things) translate to the twin approach for an infrastructure management organisation (e.g. a system of people who manage things)?

Edited June 16, 2020 by Ian Gordon

[DRossiter87]

17 hours ago, Ian Gordon said:

It feels like the established use cases for Digital Twins are usually twins for 'things' rather than 'systems'. 

Yes, it does feel like that @Ian Gordon, especially when the messaging is about a digital representation of physical assets.  I don't know if you've seen, but we have been doing a bit of research on use cases via a conversation starter Useful UseCases.  Please feel free to comment on the post or the linked document ☺️

I like that image @Peter El Hajj, when I talk about digital twins I speak about them in a similar vein expect I use the less eloquent "thin twin" and "fat twin" to explain the same idea.

[Peter El Hajj]

 

On 16/06/2020 at 15:20, Ian Gordon said:

How useful do you think that the parallels in sectors such as manufacturing and aerospace are likely to prove in infrastructure @Peter El Hajj?

@Ian Gordon I think there is a lot to uncover and learn from those sectors.  Check this out from O&G:  http://15926.org/topics/plant-lifecycle-model/index.htm#DataModel

I imagine it would be interesting (& fun) to map a potential hierarchy of individual and composite twins (aka fat twins ?) against the life-cycle of the physical system of things.  One way to start this could be by mapping the life-cycles of the physical system and overlay the interventions required on top of it.  The aim would be to have the minimum number of digital twins (maybe measured by data and cloud cost?) to capture all required information and enable decisions and interventions to be made safely and in time.