Delmia Interview Part 2: Dassault and the Virtual Twin

JD: While doing the research for this interview, I often hear people talk about Delmia trying to “connect design to the real world”. One term that I read a lot about regarding to Dassault Systems is this concept of the “virtual twin”.

JD: Obviously, we do not necessarily have a lot of time, but if you could give us a brief rundown as to this virtual twin concept.

Adrian: So, as I said before, I started my career working on and building digital models of mainly manufacturing operations.

Adrian: So, we have programming style languages, which allowed us to code a model of, say, a shoe packaging line or a fast-food restaurant or a manufacturing operation and we then developed this model where you could simulate what would happen; but, at that time, those models were very much offline.

Adrian: Back in those days, it took a long time to build those models. We fed them with probabilistic and stochastic information to see what might happen in the real world; and then, we would run it for a day in real time to simulate, perhaps, a year of simulated time. After 24 hours, the operations people would use that to make better decisions about how to lay out their line or how to do it. That was the birth of digital twins, and they were static offline. They were fed with statistical data versus real world data.

Adrian: So, if you think now about today’s version of that – which we call virtual twins or the virtual twin experience – it’s largely trying to solve the same problems. We’re trying to make mistakes in the virtual world, so we don’t make them in costly, expensive, or unsafe ways in the real world. It’s the same concept of building a model of our product to represent our operations or our supply chain, but the difference is now we’re not building it in a programming language.

Adrian: We are taking objects such as robotic arms, machining centers, test, and assembly operations or, whatever that might be, and we’re dropping them into a model. Then, those objects are not only rich with visual representation of those operations, but all of the details of the speed, the flow rate, the preventative maintenance, and all the rest of it.

Adrian: So, we’re building it from an object-based point of view and we’re connecting it with real time data. So, if we build a virtual twin model of, let’s say, a packaging line. As we go through and we collect information about what’s happening in our factories, we can feed that information back into the virtual model so that we can refine it, make changes to it, see how we can do things better, see how we can react or reconfigure the line based upon more breakdowns than we expected or supplier changes or material changes.

Adrian: And so, the differences between the digital twins of the early years, which was part of my legacy and the virtual twin of today, is that virtual twin model is something which sort of lives and breathes. It’s actually connected; it’s in being informed by the real world.

Adrian: You can make changes to it. It’s modular. It’s reusable. Once you’ve built the model of your production line and made your decision about, perhaps, a line layout change, you don’t throw it away. It continues to live and breathe as a working virtual model of that environment. You can continue to make changes to it and – more importantly – you can connect it to other parts of upstream and downstream processes.

Adrian: So, if you have a digital thread from design and engineering – say, if you have your CAD models of your products – that model-based information persists; and then, you can use that design as part of the virtual twin of your manufacturing operations to determine “if I make a design change, what does that mean to the engineering bill of materials? What does that mean to the manufacturing bill of materials And therefore, how will my model respond to that – if we change Part A to part B, or if we make a modification or a design change?

Adrian: So, virtual twins are dramatically different from, what previously were called digital twins. What we see as a virtual twin is something which is in constant use and something that can be used very easily and rapidly for lots of different types of innovation or reacting to disruption – which is, of course, the flavor of the day for most manufacturing companies.

JD: During your last answer, you talked a lot about making changes in real time and, just out of curiosity, because it seems to me that the flavor of the day these days is industry 4.0, Out of curiosity, does, is Delmia making changes or adapting this virtual twin concept in a way that benefits industry 4.0? And if so, how?

Adrian: I think virtual twin technology is very much a core part of Industry 4. 0. There are lots of different elements of technology and connectivity, IOT connectivity between what’s happening on the shop floor, analytics, and data sciences; that’s all part and parcel of Industry 4.0.

Adrian: I think we’ve been an enabler of key parts of that technology for a long time; and, we’re constantly adding to it. If you look at other new immerging technology and industry trends – like augmented reality – as part of the new workforce of the future, we have capabilities we’ve added to our portfolio products that support that as well.

Adrian: And so, it’s all part and parcel of that, but do you call it industry 4.0, smart factory – there’s a lot of different terminology to discuss the same thing. II think it’s sometimes confusing to manufacturers is, what do we do? What do we do first in terms of getting that next level of digitalization of the factory or reaching some of those industry 4.0 standards? But, yeah, we’ve been strong proponents of supporting that and other areas for quite a while now.