Skip to main content
Search
Digital Twin 101

Digital Twin 101

  • 05 June 2020
  • Augmented Intelligence Digital Twin Automation Connected Customer Experience

When oxygen tanks exploded soon after the launch of Apollo 13 in April 1970, the world held its breath as NASA engineers scrambled to save the lives of three astronauts. Their success hinged on the local team’s ability to resolve technical issues via a digital twin from 200,000 miles away.

Digital twins have historically been leveraged in manufacturing, describing virtual replicas of physical assets or machines, but at Kin + Carta we believe the application of digital twin thinking, technology, and data analytics is much broader and includes the enhancement or eventual replacement of existing analog—and often heavily manual—business processes such as collecting and verifying data manually.

Watch Deena McKay moderate the conversation between speakers from Microsoft and Arrow as they dig into Digital Twins 101. They highlight the two types of digital twins—enhancement (IoT, IIoT, etc.) and augmentation (RPA, eCommerce, Convo AI, etc.). The latest technologies that make digital twins possible include immediate applications of digital twin technology to a wide array of business initiatives in virtually all industries.

Return to all on-demand sessions

Go back

Speakers

Mark Carlton, EMEA Solution Architect IoT & Edge at Arrow
Maarten Struys, EcoSystem IoT Solution Architect at Microsoft
Deena McKay, Functional Consultant, Kin + Carta

Deena:
For the next 30 minutes, we're gonna be highlighting the latest technologies that make Digital Twins possible. And that's gonna be from immediate applications of Digital Twin technology to a wide array of business initiatives in virtually all industries. We'll also be doing a Q&A at the very end. So if you have any questions, please make sure to send those and we'll answer the questions live. So to get started, as Dan mentioned we have two representatives, Mark Carlton of Arrow and Maarten Struys of Microsoft, Mark and Maarten could you please take a moment to share a brief introduction with the people watching today.

Mark:
It's no problem. My name is Mark Carlton, I work for Arrow and I'm an EMEA IoT Solutions Architect. My role is to work with businesses and partners across Europe and across the globe really to understand and how we deliver IoT solutions from end to end. Part of that feeding into the Digital Twin conversation which is now relevant within our portfolio.

Maarten:
My name is Maarten Struys, and I'm working with Microsoft as an IoT Solution Architect. It means in my day to day business I'm helping our partners to build complex IoT solutions but I'm also responsible for educating partners and our own staff. And I'm super excited to be here today in this panel of discussion.

(00:02:56.06)

Deena:
This is going to be a great discussion. And to get started, we're gonna talk about the origin and definition of digital twin. So if anyone is familiar with Apollo 13, that was back in April 1970, they could not have predicted what was going to happen during that mission. And what happened was there was a fight for survival as the oxygen tanks exploded early into their mission. And so one of the key ways that the rescue mission was helped was with the help of a digital twin model of Apollo 13. So the engineers were able to test possible solutions from the ground level. And so I wanna go to the next slide and start talking about what a digital twin actually is. So from a Kin + Carta perspective, digital twin is leveraging the application of digital twin thinking. So we take technology and data analytics and bring them together. And Mark and Maarten I want to begin with you both giving your definition of what digital twin is.

Mark:
So for me, a digital twin is not something that's new. Okay, so digital twin is a new term that we're coming to use today in conversations but also that we're looking and working with businesses to start up on the journey to build these digital twin models. Well, digital twin has been around for a long time as you said with the Apollo solution and how we started to change what we can do and how we can consume those digital twins today. So, how did they deliver? And one of the big things for me is how the digital twin has started to move into all the processes side of businesses. So it's core of that physical element where we're looking at physical assets, physical devices, and processes now we're able to actually start to move those and using digital twins it starts to move into those business processes and how you can start to change those within a business. So you're gathering the data into these large, I would say, data databases. As such, if you think about it from that context where you've got lots of data there's lots of data that comes in and you're able to push that across. To be able to change the business outcome.

Maarten:
I’d like to add to that actually like let's do yet another definition. And again like everybody's talking about digital twins but it's not something new. It has been around indeed, Apollo 13 let's say around, like since the 70s, or something. But if you look at a formal definition, you could say that the digital twin is an IoT service these days and it helps you to create models of physical environments. And that's kinda cool because what we did with IoT so far was like measuring maybe some telemetry from a single sensor or something. But now with digital twins we are modeling a physical environment and it means that we are describing that environment and instead of measuring an individual sensor or something we collect lots and lots of data, seemingly unrelated data. But we're bringing hierarchy into that data. And then we can, you know, use those multiples, to maybe predict the future.

(00:06:09.05)

Deena:
Great, and speaking of popularity and how the digital twin has been around since about the 1970s, we even had a mention of digital twins in 1998 from a professor from University of Michigan. Why do you believe that digital twins are gaining popularity and traction now?

Mark:
When you look at the digital twin today I would say a lot of it is down to the data we can actually collect. Both from a cost point of view, as well as the amounts of data we collected and how much we can actually process that data and the ability to process the amounts of data we have. About 15 years ago, I worked in the steel industry and we used basically a form of digital twin, but it was very expensive to build because you were building models around conveyor belts so you could understand how they worked. And you could start to have that prediction around those. So it was very expensive to implement, and only got implemented in specific spaces. Now I see the change in IoT in the adoption of IoT within the industries and things like sensors and connectivity now becoming more consumable as well as more affordable. Now having that connectivity of those devices and being able to bring that data in, as well as then move it and use that data and the modeling capabilities and the performance we have around that, I feel that's a reason why we'd start to talk more and more about the digital twin capabilities.

Maarten:
Don't forget, I mean, digital twins are like everybody's talking about right now. And indeed, what Mark was saying, we are collecting tremendous amounts of data and as a matter of fact. I’ve frequently asked customers and partners, what data should I really store? Like it's almost that today we answer, well you should store probably everything you can because there might be relations that today you don't understand. But by collecting more data and trying to find those connections, those relationships, you will get better insights. And that's where digital twins will help. They will not just help by collecting and storing your data but also be part of a feedback loop feeding information back into your system to either improve your production process or maybe the experience of a customer in a store or anything you wish basically.

(00:08:38.03)

Deena:
Yes, and I wanna talk about a point that Mark and Maarten have already brought up and that is IoT and how that's enabled digital twins to become cost effective. So it's enabling us to be able to put that into our businesses today. And we've kind of touched on this already but I do wanna ask the question, and point out even, with digital twin and the evolution of IoT there's some challenges, digital twin solves certain challenges within the business. I wanna talk about what those challenges are and how we're solving the challenges with digital twin.

Mark:
So for me, when we look at these two digital twins we've got the physical enhancements. So we look at the challenges that we're having today in the conversations we're having with customers around this. I would say when we have that conversation straightaway around digital twin, it's not only for a case of understanding how to create a model. So if you think digital twin used to also be known as virtual models, or virtual twins. Where you could really start to look at 3D simulations and those sorts of ways. Now it's starting to build into processes. We're now seeing that challenge starting to change because rather than having to look at the actual device and can we create a model of that device? Can we test against it, can we create new devices can we model how those devices are going to move forward? We're now starting to move into what you see there as the experience of augmentation. So how we're actually using those to now moves into business systems. So we're taking those processes and those enhancements and now taking it to that next level.

Maarten:
One of the other things and this is a challenge basically with the internet of things itself. I mean, if you have things gathering data and storing that data on a massive scale in the cloud there's always one fear there. That fear is like how do I know the identity of a particular device or thing if you wish. A digital twin can help giving that device unique identity. The other thing which is a challenge if you connect things to the cloud is that all things are different and they use different ways to connect to the cloud. And it's kind of cumbersome to write software to make all those connections. Digital twins bring a very big promise as well for something called IoT plug and play. Where a device exposes its capabilities to the cloud. And the cloud knows how to deal with that. And with that either if virtual or not, it's a complex or a simple thing if it has the capability of exposing its telemetry and maybe its commands you can send to it and its properties then it's becoming very easy to connect those things to the cloud. So that's another area where digital twins definitely will help us.

(00:11:51.02)

Deena:
Yes and talking about the future and where we can go. If you look at the bottom of our slide that's being presented right now we have Experience Augmentation. And one of the examples that we provide is conversational AI. And I know that is a space that is growing in traction and let's just speak to that and your experience with conversational AI and digital twin. Maarten could you begin with your experience?

Maarten:
Absolutely and this is a great question actually Deena. Because one of the things we see happening is that we are also democratizing AI. We're making it simpler to use AI without worrying too much about what kind of complex underlying models we are utilizing. So one of my examples here in the augmented reality sphere or actually in the, let's say in the conversational sphere, is that typically, we have customers who have a customer service center and people are calling that customer service. Well, one of the things that's happening frequently today is that these customers are being answered by a robot and the robot understands. The bot service understands things like the sentiment of the speaker. So ifthe person asking questions has a negative sentiment so they might be impatient or they might have a certain tone or something, we go faster to connect them to a real person. But if somebody is just telling you how happy they are with your servers, then the bot itself can handle that conversation. And this is a scenario where there's this whole augmentation or cognitive services as we call them at Microsoft plays an important role.

Deena:
Mark, you have experience and experiments Experience Augmentation as well. Do you wanna tell the audience about your experience?

Mark:
So rather than conversational AI my experience is more from an implementation of solutions around challenges. So when we look at how customers are now starting to use the data they're collecting and pushing to things like the retail sector so we look at retail and we're now starting to build this digital image of a supply chain within businesses. So this isn't down to where stock is, this is down to pricing, we start to look at discount levels, how's this store have this discount on them, we're able to then store that data and be able to build that up. And the augmented side of the solution is how you start to then replicate it or build this into different solutions. So you can start that automated ordering process. So if we look at some stalls at the moment, there was a company I was working with, they were able to start to understand how their current sales flow works right. Whether they've got stock in that store or whether they're able to take that in from another element. Now this is an all automated process. There's no manual side of having to actually ask, “Do we have that kit in any more, place an order, find the PO, do that.” We're now able to start making those decisions early and being able to build that digital model understanding using business processes as well as the technical physical devices that are there. Businesses are providing different customer experiences because they've got more things on the shelves as such, as well as provide that, I won't say a step above competitive stalls. So they're able to then start to be able to take it forward how they use digital twins to be an enhancement to their business.

(00:15:31.09)

Deena:
Great, and we've talked about two different types of digital twins, how we can implement and how they can be implemented within a company. But we haven't gone into the full discussion around implementation. So I wanna ask, what makes a company a good candidate for digital twin?

Mark:
It's said as a wide open question. So a company for digital twin, is it really comes down to what are they trying to achieve and the type of business. What is the business challenge. So it's understanding what the challenge is and where we want to go and what the future is of that business. So this is where we start the conversation. I would say I never start the conversation with what sensors you regard or anything like that or how the process is, it's understanding that challenge first and that's how a business needs to understand whether they need a digital twin. Not every company is going to need one. But they may in the future. And this is where we start to look at solutions today. Because it's all part of a journey to a digital twin. You're not just going to turn it on and have one. You've got to start building up to being able to have this model. And I'd say that's where I start to uncover that conversation. To be able to understand whether a business needs integration, one of their biggest things. So integration to applications being able to take the data and being able to provide automation. You've gotta think about solutions that are more into that predictive side of things. So anything that's more in the predictive modeling where we're looking at trying to look at sales or we're looking in industrial, we're looking at machinery. And we might be looking at the predictive maintenance side of things. So we're looking at models of machines and are they going to fail, what's the downtime? Those are candidates where I would say we really want to be focused in looking at that digital twin opportunity. Because that's what's going to enhance and build that for the business.

Maarten:
I completely agree there Mark, and it's kind of great that you took the answer first here because if I would have answered as the first person here on this question I would probably have said like, hey, we are a platform company. We are implementing digital twins and I have a fantastic implementation of a digital twin for you. You can use it right away out of the box. Go ahead and get started. And I would forget that we are trying to solve business problems with these digital twins. It's just a piece of technology. And the other thing there is that it's important that piece of technology needs to be standardized as well. Because what Mark was saying so far, it's like, hey, you've got different companies that might do something like in the predictive manner or something. But what we see with digital twins is that you also get hierarchies and collections of companies working together to solve a larger problem. So let's say like in the public sector a smart city could consist of a number of smart buildings. Now, a smart building owner is not the owner of the city because the public sector is, but they can make use of the same technology and can use each other's data. So the data generated through these technologies is the important thing and also called the new gold.

(00:19:00.01)

Deena:
Great, thank you. So we've talked about the two types of digital twins. But if we could go to the next slide we also have talked about this the next slide please. So we've talked about the advancement but we've also talked about different verticals that digital twin has successfully been implemented in. And we've mentioned business. We've mentioned the plug and play by Maarten and we mentioned retail and I wanna ask what do you foresee for digital twin being used in the future? And in what industries do you think it will be most helpful in?

Mark:
For me I think again, digital twin has a place in probably a lot of verticals at the moment. It just comes down to that business reason why they need it in there. We look at things like smart buildings Maarten mentioned, smart city, smart building, smart workspace. And we look at how people are starting to use them within those areas. Again, all those would be feeding into a smart cities digital twin. But the smart building may have a digital twin of its own. The workspace element will have a digital twin of its own. And that's to understand things like employees.

So you're looking at utilization within an office you're looking at things like the energy consumption within an office. You're able to use all these different data that you're collecting, as well as the processes within the actual building itself. Things like the BMS system. So you're looking at the business management system and able to use data to be able to compare that data between the visual management system and the digital twin. Do you know you're running the business in the most efficient way? And all these different tweets and that's one area that I would see. But if you think about the agriculture industry the agriculture industry has started to move more and more into the technology sector. The IoT, how they're using IoT, they're using drone technologies, they're using autonomous vehicles, they're using chips in cattle and animals to be able to track them. Not just track them physically, but also tracking things like their health and being able to build that model up over how they're starting to use that. And then how that feeds into things like the food chain.

So I talked to a number of companies around how we start to look at the actual data in the agricultural industry, how we look at livestock and actually be able to store that data and build that data up. So we're building that in a digital image of the animal all the way from origins all the way through the food chain. So we know where it's gone, how it's been, where it is, what field it has eaten out of, everything like that. So these are the areas where I see digital twin really starting to push into, other than what's already being traditionally used. Also, if you look for instance, at health care, like a description or a modeling way of an operating theater for instance. And then all of a sudden lots of different techniques come together as well and they can all be modeled through different digital twins. But one of the things for instance in an operating theater right now, like if as a doctor, if I need some assistance, an experienced other doctor, if I can use something like augmented reality, I could actually have the other person helping me actively performing an operation on a patient. Another thing in healthcare could be preventive health care, maybe for elderly people, like what if I could build a crest detector that a wearable device can predict before somebody is falling, that something is going wrong. You need to describe like a lot of physical characteristics of a person in order to do that and digital twins can actually help you making that description.

Let's say and just to add to that, Maarten had touched on the healthcare side of things using augmented reality and that's everything. You start to go into worker safety, you start to look at that side of things and how customers are starting to use worker safety. Worker safety is a huge thing from a health and safety standpoint, especially if you're remote workers. If you're out by yourself, you need to be able to do that. But do you need four people to go do one job if you can start to use augmented reality to be able to help it back. That person may be skilled in specific ways. And if something needs fixing, say you're out at one of these wind farms or something like that. You go out today, you don't want to be sending an engineer out each time. You want to be able to then start once you've got a person that fits what you can. Again we go back to the digital twin and having that data and having that model then to be able to then scale up and use someone that might be onshore to be able to fix something that's offshore by an engineer that's not skilled but able to pull a specific set of goggles on. Glasses to be able to know what the engineer is able to see to then tell them what to fix. So it's another area that we're seeing it's starting to move into.

(00:24:35.01)

Deena:
Great, and thank you Mark and Martin for that information. We have about five minutes left in the session. So I do want to open it up for Q&A. So anyone who is participating, if you have questions make sure to drop it in the chat for us. To begin Mark, Maarten we have a question from Zack Dispro. I'm still trying to understand the difference between a digital twin and a digital model of a process experience. Are they the same?

Maarten:
They are more or less related to each other. We use digital twins more if you wanna like describe hierarchies of seemingly unrelated let's say processes and still try to connect them together we'll try to connect them together. For instance, my example of a smart city consisting of smart buildings but also of streets and traffic lights, et cetera. It's kind of easy to model that with a digital twin. Because the the ownership is, like shared amongst multiple companies. And it might be a little bit of process will be a subset of it in my opinion.

Mark:
I agree with that so that's it.

(00:25:33.01)

Deena:
All right we have another question. Can you give a few examples of successfully implemented digital twins?

Mark:
I can talk about a workspace solution, how this has been implemented. Of course, I can't talk widely about the solution itself, how this is really starting to be successful more than anything, especially in that workspace side, is the transferable model into buildings so office spaces. So we're starting to look at utilizations and power consumptions. So not only are we looking at the building itself, we're looking at things like the employee satisfaction. So can you get office space, can you not get office space? Have you got hot desking? Have you not been able to map that and have the internal maps? Been able to then relate how many people you've got in the office to what air conditioning? What heating systems need to be turned on? What's the CO2 levels, how you can start to then build that into those building management systems to be able to automate those sorts of. Make sure you've got the right amount of people in, or the right surroundings and environment for those people. The other way that I've seen it start to grow is where it starts to actually look at not just the environment, but also the cost element. So do you need the space? And then that clear modeling of how the digital twin has been moved between building to building so if you've got a similar sort of setup or application. So if you've got rented office spaces that you're just going to have repeated over and over again you can very clearly start to use that as a model to how you move forward and how you set out those offices and be able to budget for those in the future. So that's one of the areas where I've seen digital twins really start to take off as a successful case.

Maarten:
Yeah, for instance also if you look at our campus it's kind of similar or the same example. Most buildings are equipped with digital twins these days and it's indeed for occupancy meeting room availability. But it could also be used to guide traffic and an emergency situation. Like if everybody needs to go to the emergency exit what's the most efficient way and you could even have digital signs pointing there. Another one that I hear more and more about in my country in the Netherlands these days is for the electrical grid to measure or to control energy requirement and that has a lot to do with the fact that as we get more and more electrical cars, you know, and one of the studies we're doing right now is using cars or using the grid not to charge cars but also using cars to temporarily store energy. So that's a very interesting juice. It's where digital twins play an important role as well.

(00:28:28.08)

Deena:
I want to wrap up with a final question. So any question that was not answered during the session we'll make sure that it is answered for you and you get that answer via email or possibly a chat. So the final question for the session is for anyone who is listening today to our conversation and is interested in starting a digital twin where do they start?

Mark:
Okay, so one of the things I would say is digital twin is a very complex deployment. It's a very complex thing to start to build. So partnerships are important. Being able to look at the different elements and break those down. The other side that I would say is once you need the skill sets to build a digital twin, you wanna focus on that business side of things, the challenges. If you understand the challenge and what the customer is trying to get to or what you're trying to get to as a customer, whether you're deploying for someone else or you're deploying for yourself, it gives you a set of questions you can ask yourself. So what are the processes that's needed? What do I need to collect to be able to build this to get to it, so you can start to then spread out and kind of look at the different components to then start to build that digital twin. So that's how I approach it when I'm talking and working with solutions is to start at the end and what the end goal will be and then start to work backwards to what we need to implement.

Maarten:
Yeah and again, I totally agree there. First, we start with the problem you need to solve. And then you need some technology in order to do that. And then you can start asking yourself questions like, do I want to master this technology myself? Does it even make sense? Do I need to become a digital trained expert, for instance? Or am I gonna partner with companies who already proved that? And that's typically how we approach that market? I mean, for instance, if we provide the platform, the digital twin platform, we still don't know anything about your business logic. So we need to partner together in order to come to a solution.

Want to know more?

Click here to get in touch