Home / Blog / DT evolution in Manufacturing – I

DT evolution in Manufacturing – I

Digital Twins used to be a straightforward, well defined concept. As they evolve it becomes more difficult to have a precise definition that is agreed by all. Image credit: IEEE DRI

In these last few weeks I have been preparing for the webinar I’ll be giving tomorrow, as part of the EIT Manufacturing Tuesday from Future. I came up with a syntheses of the work that has been done within the Digital Reality Initiative that I feel can be of interest to a wider audience, hence my decision to post this new series. Your comments, as always, are welcome.

Everything evolves, right? Why shouldn’t Digital Twins evolve as well?  Indeed, they have been evolving and as I look at what is happening around the world in many sectors they will be evolving even more. I guess one should take for granted a widespread knowledge on what a Digital Twin is, particularly in an audience that comes to listen to the evolution of Digital Twins.

As a matter of fact the concept of Digital Twin used to be straightforward just 5 years ago. It is a digital copy of a physical entity. Yet, when I discuss this with different people I get a variety of nuances and when I think about DT today in different sectors and how they are evolving the definition becomes fuzzier and fuzzier. In a way this is a proof that there is a significant evolution under way and at the same time the fuzziness supports further evolution (if something remains well defined it is also constrained by its definition, i.e. does not change, nor evolve!).

Hence, the very first point to address is to look at the DT definition (for the record, last year I participated in a discussion by the group of authors engaged on writing a book on DT -it should be out in a few months- and then again, more recently, in a discussion within the EU expert group on personal digital twins. In both cases different opinions on what DTs are -should be- emerged and it was not possible to come to a single, unanimously agreed, definition).

So let’s look at the “old” definition: a digital copy of a physical entity. The digital copy:

  • is a mirror of the physical entity, i.e. its digital model
  • keeps track of its current status, it shadows the physical entity, and
  • keeps a record, thread, of the evolution of the physical entity

The first point that is “implicit” in this definition is that a Digital Twin is not, and never was, a “copy” of the physical entity. First because of the “thread”, the physical entity is “as is” at this precise moment, it does not have “memory” of its past (not necessarily at least). Second, even disregarding the thread, the digital model updated to the present status (through the shadow) is “always” a partial model of the physical entity (to be extreme, we would never be able to model each individual molecules making up the physical entity). This partial model is fine, as long as it represents what matters from the point of view of using the digital twin.

Here we come to the second point: as the use of the DT changes, so we need to change the digital model. I’ll show in subsequent post what this means from an evolution point of view.

When dealing with the digital model an important aspect is how we can create (and we create) it. Historically the digital model of a product in manufacturing has been created (and by far still is) using the output of CAD, Computer Aided Design, the tool used in the design phase. Hence, most of the time, the Digital Model precedes the existence of the physical entity. In other areas, like building construction, the digital model can be the result of the BIM, Building Information Modelling, a tool -and standard- used in that sector. Here again the digital model precedes the physical entity. In other cases, like in healthcare, the digital model comes after the physical entity and can start from the EHR, Electronic Health Record. The digital model can be “generic” or specific to a physical entity. In the end we will always need a specific digital model that in case of a generic one (like the one produced by the CAD) requires an instantiation taking up the identity of the physical entity (the manufacturing process produces many “pieces” all alike, each one associated to a specific instance. All those instances will share the same digital model but will have different shadows and different threads.

There is, however, another way to create a digital model: through observation of the physical entity interaction. This is what happens, as an example, with Alexa. Through its interaction with the user it can (I am not saying that it does this today, just that it “might”) create a digital model of the user for what concerns her behaviour. It can develop a digital signature of her voice, sufficiently accurate to be able to distinguish among different users and therefore interacting using different “instances”, different Digital Twins. The voice digital signature can of course show pattern alteration (still the same person speaking but with a different intonation…) and Alexa can derive information about the mood of the user (and react accordingly). Of course there is plenty of information in the interactions (what the user wants, when she wants it…) able to create a model of the user “habits”, interests, …

I will show how this way of creating, expanding, a digital twin may become a crucial assets in manufacturing in the framework of Industry 4.0.

More to come.

About Roberto Saracco

Roberto Saracco fell in love with technology and its implications long time ago. His background is in math and computer science. Until April 2017 he led the EIT Digital Italian Node and then was head of the Industrial Doctoral School of EIT Digital up to September 2018. Previously, up to December 2011 he was the Director of the Telecom Italia Future Centre in Venice, looking at the interplay of technology evolution, economics and society. At the turn of the century he led a World Bank-Infodev project to stimulate entrepreneurship in Latin America. He is a senior member of IEEE where he leads the New Initiative Committee and co-chairs the Digital Reality Initiative. He is a member of the IEEE in 2050 Ad Hoc Committee. He teaches a Master course on Technology Forecasting and Market impact at the University of Trento. He has published over 100 papers in journals and magazines and 14 books.