Limited functionality available
Digital twins can mirror specific products or illuminate entire systems. Tanya talks with Airservices Australia's Mick Snell and Deloitte's Scott Buchholz and Aaron Parrott about the possibilities this opens for business.
Mick Snell: We look after about four million movements, aircraft movements, and more than a 160 million passengers annually.
Tanya: That’s Mick Snell. He leads the Strategy practice at Airservices Australia.
Mick: Our job is to safely manage all of the airspace from the ground through to space for about 11% of the world’s airspace. On a map, it’s roughly halfway to the African continent, two-thirds of the way to New Zealand, all the way from Antarctica up to the archipelago of Indonesia and Papua New Guinea.
Read the Digital twins chapter
Explore the Tech Trends 2020 podcast collection
View Tech Trends 2020
Learn about Deloitte’s services
Go straight to smart. Get the Deloitte Insights app.
Tanya: That’s a lot of territory to manage. And it’s challenging. But technology—specifically, a technology called the digital twin—could make that job easier.
Tanya: I’m Tanya Ott and I’ll be playing ground control in the conversation. Our pilots today are Scott Buchholz
Scott Buchholz: I have the privilege of being Deloitte’s Emerging Trends research director, which means I get to lead our Tech Trends research on an annual basis.
Tanya: Aaron Parrott
Aaron Parrott: I’m a managing director and leader in our Supply Chain Network Operations practice. I work with clients in manufacturing all across the globe, helping them transform their supply chains and their operations, specifically focusing on digital solutions.
Tanya: And Mick Snell, from Airservices Australia.
Mick: My job is to try and make sure that we’re unlocking all of the potential in technology to deliver the services to our customers.
Tanya: I’d love to have you talk a little bit about delivery drones and aerial taxis and how that sort of thing further complicates the already challenging job that Airservices Australia has.
Mick: The air-traffic system or the aviation system at large is amazing. Over several decades now, with a heavy focus on safety, it’s been able to maintain a level of service that’s second to none in terms of transport modes. The interesting problem for us, when you introduce new airspace uses, particularly at low levels in and around urban areas, is that they are going to operate in a completely different way. So, as we think about new delivery drones or urban air mobility type activities, we need to look to different solutions in order to be able to scale it effectively.
The focus on how do we create simulation and optimization opportunities is really at the forefront of how we need to shift the service that we provide now, because at the moment we have air-traffic controllers talking to pilots. That’s not going to be the delivery model 10 years from now for all of the drones that might be flying around cities, delivering various packages or delivering people from place to place. It’s going to be a whole line of new systems and technologies that are going to need to be tested and trialed. So that’s a pretty big challenge for us over this next decade.
Tanya: Speaking of testing and trialing, you have been looking at this idea of digital twinning. I want you to start with what that looks like for you, the tests that you’ve been doing.
Mick: We’ve got a pretty rich history in innovation in Australia over several decades, but one of the things that was really important to me through this process was to try and bring it back into small, manageable-sized pieces to deliver incremental value. In today’s system, we rely heavily on the highly trained professionals, whether they be the pilots in the cockpit or the air-traffic controllers on the ground. But the system is fragile. Disruption in one airdrome creates a ripple effect right across the network. For us, the common problem here is how do we enhance our flow and network management capabilities. How do we actually create a digital replica of an entire ecosystem? Our capabilities are in managing a network in real time. That’s literally thousands of aircraft moving through the airspace. How do we make sure that we enable an airline operator to make the right decision for their fleet in and amongst civil airline operators that are all making optimal decisions for themselves? For us, the ability to step into an environment where we can simulate and optimize those decisions in a collaborative way with our customers was where we saw the initial focal point.
Tanya: When you were looking at doing that simulation, you were using historic air-traffic data in order to create the scenarios that you’d have to model. Am I understanding that correctly?
Mick: Yes. We’ve got access to quite large data sets, historical data sets that gave us the ability to, build out some proof of concepts to allow us to build some confidence that we were on the right track. Certainly, over the last 12 to 15 months, we’ve been very pleased with those results. The proof of concepts have delivered more than what we anticipated in terms of options and being able to expose some of the gaps in our thinking [on] what’s evolved for us. Particularly the last four or five months have been a focus on how do we take the proof of concept and validate that it can be deployed operationally? We’re really keen to try and introduce the capability in real time. So the focus of the team’s work over the last four or five months has been, how do we make sure that we’ve got the right performance standards to be able to validate that it can be deployed operationally and support our decision-making.
Tanya: So it’s got to be more predictive than reflective, I guess. You’ve done these four proofs of concept and now you move into the piloting stage. What does that entail?
Mick: The key focus for us is being able to run multiple applications off the top of the digital twin. We see the ability for network decision-making being really important. How do we make sure that we create a common operating view so that we can manage delay effectively across the network? If there are four or five aircraft that are delayed at one airdrome, what are the implications as those aircraft moves through the system? What was really important to us in this approach, in being customer-centric, is to create a collection of applications that help the airline operators set their own priorities for their fleets. Airlines make decisions to move their own aircraft in and around their network based on a whole range of factors running an airline and that’s really complex. What helps us deliver the optimal service is having an insight as to what their priorities are.
Tanya: Aaron, I’m going to lean on you to explain what a digital twin is, because we hear that word a lot. And for a lot of people, though, it’s still a little bit of a mystery.
Aaron: When you think about a digital twin, what we think about is it is the digital representation of the physical activities and actions that are taking place at a manufacturing facility, from a manufacturing perspective. So we look at it, there are two different types of digital twins. One, I can do a digital twin of a process, which basically means, I’m digitizing the entire process that a project is going through and understanding how the machine’s performing, tracking that information. But I also could do a digital twin of a product so that I’m providing key information about that product as it’s being built, which would include some aspects of the process. That would include, what the machine was running. What were the key outcomes? What parts were used in building the product? How was it actually built?
And more importantly, when we think about a digital twin, and expanding this, it’s not just about what happens inside manufacturing’s four walls and when you sell the product to the customer, but also creating that digital twin of how the machine is actually being used in the field and operating and performing in the field. But most importantly, another aspect of this, is how it’s being maintained. So when you think about an aircraft or you think about a large piece of construction product or even a car, some of the most difficult questions to answer are, I know how it was built, but [with] what parts, and what’s been changed, and what are the maintenance and services that have been done on it since then? And that digital twin really provides a significant amount of value for the manufacturers, as well as for the people that are servicing those products.
Tanya: What I think I hear you saying is that, you could build a model of something in a specific setting, based on location and all that kind of stuff. But when you move that same asset into another space or you give it different treatment or different maintenance and that sort of thing, it’s going to perform in a completely different way, even though it’s the same asset that is somewhere else. That, digital twin allows you to model it in different scenarios.
Aaron: Right. And it’s the actual—meaning, as a user product, I have to service it. And I have warranty and I have recalls. And there’s, what parts, and was it in [a particular] lot, meaning that it’s now impacted by this warranty recall? And if you have a strong digital twin aspect of the products that you built, you can very quickly determine which of your products and what you put on to marketplace that may be impacted by warranty. Or you can very quickly understand, when I go out to maintenance and service this machine, I already know what parts have been moved over or maybe been upgraded, so I understand how I do the maintenance and service the product as well.
Tanya: I’m going to ask you a question. It’s going to be the stupid question. But some of our listeners are going to be wondering, those that particularly aren’t familiar with this concept, is this digital twin a real physical thing or is this something that exists inside some massive computer somewhere?
Aaron: From a standpoint of is it a real physical thing, it’s providing and giving real outputs and information that are needed and necessary. Yes, it sits inside a computer. It is a representation of a process or a product. But the information that’s being gathered is extremely valuable to be used and leveraged, as even Mick was talking about in real time, to make real-time decisions about how my client is operating. So, it is a digital representation of a physical process or product, but the outputs of it makes it much more a real thing that people can touch and understand how to leverage and utilize this information, either with real-time decision-making or for a longer-term strategy decision-making.
Tanya: Mick, Aaron was just talking about a car or something like that that might need to be serviced. But you’re looking at very complex systems. When you’re taking this idea of the digital twin and applying it in the very complex systems, how do you think about it a little bit differently, maybe?
Mick: It’s an interesting concept. We didn’t start out with the language of a digital twin. It’s an interesting communications challenge even within our business. One of the interesting things from a decision-making perspective is that we rely heavily on the experience of highly trained professionals and there can be quite some variation for us across any given day, depending on who’s making those decisions. One of the advantages of moving into the digital twin space is the ability to do the trial and error, the what-if scenario planning based on up-to-date information and be able to have the processing speed to be able to present decisions in real time or insights to decision-makers in real time was the really important factor for us. Having a simulation capability, that takes three days to load the data and a day to crunch the results to then review what you think the scenario might be isn’t going to fit with us.
Tanya: Scott, how does this idea of a digital twin fit into the overarching concept of tech and where tech is going right now? Because that’s something you’re looking at pretty deeply.
Scott: It’s really interesting, because what we see is people are doing experiments at the edge of what they can do. And as you heard Mick say, not everybody is calling it a digital twin, but a lot of what we’re talking about is this idea of advanced computer simulations, doing things that mimic the real world, that interact with the real world, that simulate the real world, that help optimize the real world. One of the reasons when we were looking at this trend that we felt it was really interesting is because we’re starting to see it get used across industries. Clearly there have been use cases in manufacturing for decades. Those are really interesting as they become more advanced. Mick is talking about a use case where technology that’s used to optimize Formula One vehicles is actually now being used to look at airspace management. We have other use cases where we have customers using things like some of the massive multiplayer online game physics engines to do county simulation for urban planning. So the really interesting thing is the degree of creativity that we’re seeing in this space, whether or not people call it urban planning or they call it optimization or they call it something else.
Tanya: That’s really fascinating. So it’s much more than just big data. It’s creating something that didn’t exist before or taking something from one industry and creatively using it and applying it in another industry.
Scott: Aaron can go on for hours talking about really interesting use cases in manufacturing. When you listen to some of those stories, you start realizing, oh, that’s a problem that other industries have. They might not describe it that way. They might not think about it that way, but managing the flow of components through a manufacturing line is not always incredibly different from managing the flow of people through a retail store. And that’s not always all that different at the simulation level than managing the flow of people in a transportation network like public transportation or highways. And now we actually are starting to see not just the ability to do all of those things in similar ways, but people actually trying to figure out how to take advantage of those capabilities.
Tanya: Aaron, I don’t want you to go on for hours, as Scott suggested that you could, but if you’d like to give us a concrete example or two of what digital twinning has allowed in the physical space, in manufacturing, for instance.
Aaron: One real example that is where a lot of our clients are starting is creating a digital twin of the factory. We had one client where we worked with them to enhance what they already had for sensor information and PLC (programmable logic controller) information that was coming off their complex machines, but they weren’t really leveraging the information that they were getting off these machines. It was just spitting out data and being stored away. So we added some additional information and additional sensors that were going to be impactful for them to really understand how their line was performing and start to proactively detect that there were issues. And specifically, what we’ve been able to do then is create the digital representation of how that assembly line actually performs and help them understand what are the right settings for each of their complex machines that are connected on this factory floor to significantly improve their efficiency and their production. But at the same time, enabling them to very quickly react when we sense that there is a problem, whether machine is going out of vibration, whether the machine is not performing or acting [as expected] or we’re sensing a temperature issue now all of a sudden with this machine. Not only are they getting a representation of the factory so that they can understand what are the right settings based on a lot of different factors that have to be incorporated, but the second part of that is we’re giving them real-time information to tell them that you have a problem that needs to be addressed immediately.
That’s helped them become much more efficient. So much so that once we did one line, we ended up doing several hundred lines across several factories for them to help improve their performance of their factory lines and then create a digital twin of their operations across the globe.
Tanya: And that predictive ability is really important for business, but there are some businesses where it’s even more important, I would argue. Something like power plants and energy generation. If you could have a digital twin of a power plant, that would then be able to notify immediately that there’s some sort of issue, that’s pretty important from a public health or safety perspective as well.
Aaron: And there’s not even a human in the loop. The digital twin detects that there’s an issue and knows a series of actions that it needs to take automatically or by itself to address and adjust the problem.
Tanya: Mick, I’m wondering, as you’ve been on this journey, what sorts of challenges have you faced? Whether it’s technological or even at a more basic level, getting the team onboard with this idea that they can trust a digital twin?
Mick: It’s a really good question, and we certainly started out this journey with a focus that it wasn’t just about the tech. For us, the focus was as much a cultural shift in trying to move from what has been a systems engineering, high reliability, safety-focused approach, to shifting into more of an innovative, minimum-viable product, test-and-learn environment, [which] creates a number of procedural and operational challenges for us. We spent a lot of time at the front end of this project thinking about how do we help the team embrace these different ways of working and making sure that we didn’t create cultural clashes, in terms of how we were going to go about trying to solve this problem. The way that we started thinking about this problem evolved within the first three sprints. What we discovered in those first six weeks have fundamentally pivoted the next six months’ worth of work. And we needed to try to make sure that we were very mindful to the fact that we were disrupting people’s normal methods and normal workloads. We were also careful to select partners who didn’t just bring to the table the deep technical experience. We recognized that this is part of a wider business transformation activity, and so we wanted to be able to translate the tech into business outcomes.
Scott: We in the IT space have to learn time and time again that technology for its own sake is not something that anybody cares about, other than technologists. If you look at the journeys that people go on, in a lot of cases it’s really about, how do we bring people along? Even more importantly, in many cases, when you think about the description of Mick’s journey, when you think about the cases Aaron was talking about, we’re talking about a bunch of people who have deep expertise in what they do and [we’re] trying to make the systems support them so they can do their jobs better. Starting with that in mind and thinking about, how are we going to create technology that enables humans to do their jobs better, is actually really important. Really important to making sure that by the time you get done, it’s useable. And then certainly, as Mick talked about, the shift to showing people things early and minimum viable products and focusing on where the value is, all of that’s really important because that allows you to show people what’s going on and bring them in along the journey. I’ve spent many years in technology and we spent many of those years going into dark corners, building things, and showing up months or years later —surprise! Showing the users and having them react poorly because we had forgotten to bring them along, because we’ve gotten so caught up in what we were doing. So, what Mick’s talking about is really important in terms of being successful when you’re doing things that are new and different.
Tanya: Which leads me to the question: What is it about technology right now in the tech space that allows us to be looking in a meaningful way at this idea of digital twinning?
Scott: It’s an evolution. What we’ve seen is the increase in computational capability, the increased power of machine learning, the improvements in the tools, the interoperability of the tools with sensors and systems. It’s a combination of all of these things, where 30 years ago when people were talking about what we today called digital twins, what they were talking about was designing an aircraft on a computer before you built the model. Today, what we’re seeing is, instead of doing one design and then having to build a physical model, people can do millions of designs with millions of simulations. They can use machine learning to do optimization. The capabilities have evolved to the point where it’s almost a different feel. Again, when I come back to why now for digital twins, that was really what we were saying. It’s people using tools across industries, people being creative about the tools they’re using, and the dramatic increase in the capabilities of the tools themselves.
Aaron: I would agree with that. The capabilities of the technology are really enabling manufacturing clients to do things that frankly were a pipe dream even just a few years ago. The other part of this is, when you think about the continued pressure to reduce your costs, improve your efficiency, and the pressure that manufacturing clients are on to be much more centered around their customer base, it means making a lot more different options and configurations of your product. Your demand cycles are much shorter and you have to have much greater flexibility to respond to what’s happened in the marketplace. That’s creating these new issues and things that they have to figure out how do they address them in a way that they just can’t do in their traditional operating model today. We’ve become very lean as manufacturing. We’ve looked at our cost structure. We’ve looked at how we’re buying our product and how we’re buying our assemblies and material that we need. The digital twin capability is really allowing us to take that next evolution of manufacturing to address the challenges that, frankly, for 25 years we’ve been trying to resolve to a certain extent, but now being able to take this to a whole new level.
Tanya: So you’re talking, Aaron, about taking it to a whole new level and moving it forward. I would love to just close out with, not your crystal ball analysis, but what you all think is next in this arena. Mick, maybe we could start with you. What do you see as being the next big thing that we’ll be talking about when we’re talking about the concept of digital twins, whether you call it a digital twin or not?
Mick: I The ability to create the foundations for being able to bring more and more data sets into the digital twin to optimize the network further will be the key focus for us. What we have at the moment is a pretty clear vision of the pathway that we need to follow, but if we were to imagine ourselves 15 years from now, we see potential scale of airborne vehicles increasing in order of magnitude by sort of some 50-odd times to what we have today. The array of data sets and the ability for the system to cope with that is not something that the aviation system has been built for up until now. So for us, it’s going to require quite radical transformation of most of the endpoints across the ecosystem; whether that be the airborne aircraft themselves [or] all of the underlying technologies and equipment that support that, we’re going to see this decade, for us, be probably the most disruptive that we’ve seen in the last 100 years. And the backdrop for that is technology disruption creating quite a lot of that momentum for us to move forward.
Aaron: There are two steps for what’s next in manufacturing. We’ve talked about this, how you can design in the digital world and then run multiple simulations. What truly is going to happen with the digital twin is that you’re going to be able to visually design—not just simulate—you’re actually going to be [able to] physically, in a digital way, run your product through what your operations are expected to be, very quickly adjusting, redesigning things so that you know where there’s an operational inefficiency or there’s a problem with how the part is being produced on one of your machines. To rapidly be available to address those issues, to create a first design that you’re releasing of the product, or that will run much more efficiently, enabling them to get to market much quicker with their designs. So that’s true connection between a digital design through a digital production. The second piece of this will then be large manufacturing of complex products will connect not only within the factory walls with a digital twin, but they can connect the supply base for their key products, so now they’re creating a digital twin across their subassemblies that are being assembled with their supply base all the way through what they’re manufacturing.
Scott: If you want the far-flung future, a few decades hence, what we may find is we have increasingly high-quality digital twins of human beings and hopefully that will help medical science do an even better job of helping us optimize our health and well-being. And I look forward to that day.
Tanya: Scott, I love how you leave us on digital twins of human beings.
Scott: Where else would you leave off? (laughter)
Tanya: That’s for another podcast episode, right? (laughter). Aaron, Scott, Mick, thank you so much for the conversation today. It’s fascinating stuff and you’ve helped us understand it a little bit better.
Aaron, Scott, and Mick: It was a pleasure. Thank you.
Tanya: Digital twins of humans … we are so not there yet. But when we are, you can bet we’ll be talking about it. At our website, you can find all kinds of interesting conversations and resources on topics ranging from the ethics of technology to how your finance and IT department can work together more effectively. That’s at deloitte.com/insights.
This podcast is produced by Deloitte. The views and opinions expressed by podcast speakers and guests are solely their own and do not reflect the opinions of Deloitte. This podcast provides general information only and is not intended to constitute advice or services of any kind. For additional information about Deloitte, go to Deloitte.com/about.