DigitalituM Podcast Episode 13 - Michal Ukropec - twinzo Visualization with Digital Twin

Episode Description: DigitalituM Podcast – Exploring Digital Twins with Michal Ukropec from twinzo

Topics Covered

• Michal’s Journey: From a background in taxation and IT to founding Twinzo, a company revolutionizing Digital Twin technology.
• Defining Digital Twins: Understanding what a digital twin is and the diverse interpretations across industries.
• Evolution of Digital Twins: From indoor positioning and real-time location services (RTLS) to full-scale digital twin applications in manufacturing and beyond.
• Technology Behind Digital Twins: Integration of IoT, data analytics, visualization, and real-time tracking.
• Use Cases: Digital twins in logistics, production, construction, and facility management.
• Challenges and ROI: Identifying cost-saving opportunities, improving process efficiency, and overcoming integration challenges.
• Future of Digital Twins: The role of AI, predictive analytics, and advanced visualization.

Key Takeaways

1. Digital Twin Definition Varies: While commonly understood as a digital replica of a physical asset, its scope changes depending on the industry and application.
2. Digital Twin Evolution: Twinzo started with real-time location tracking and evolved into an advanced digital twin platform.
3. Practical Applications: Companies use digital twins to optimize logistics, track production efficiency, improve facility management, and enhance safety.
4. Technology Integration: IoT sensors, ERP systems, real-time tracking, and data visualization enhance the effectiveness of digital twins.
5. ROI from Digital Twins: Real-world examples show significant cost savings, such as reducing unnecessary forklifts, streamlining logistics, and improving maintenance operations.
6. Adoption Strategy: Companies can either have a clear use case before implementation or explore broader applications to find the most impactful areas.
7. Future Trends: AI and predictive analytics will increasingly enhance digital twin capabilities, though businesses are still in early adoption phases.

Resources and Links

• twinzo Website: www.twinzo.com
• Michal Ukropec on LinkedIn: https://www.linkedin.com/in/michalukropec/
• Twinzo on LinkedIn : https://www.linkedin.com/company/twinzo/
• Host Markus Rimmele on LinkedIn: https://www.linkedin.com/in/markusrimmele/

🚀 Call to Action Visit https://www.twinzo.com/ to explore more about twinzo and digital twin applications.

Stay tuned for more inspiring conversations about manufacturing and digital transformation. Also, remember to follow and subscribe to the DigitalituM Podcast for exclusive insights from industry leaders and innovators.

We appreciate your likes and comments. If you feel you can add value to this podcast series and want to be our guest, send an email to Sales@DigitalituM.com

Transcript

Markus Rimmele (00:09)
Welcome to the Digitaliton podcast at the intersection of manufacturing and digital transformation. Today we have a special guest, Michal Ucjopek from Bratislava, Slovakia, and our topic today is digital twin. It's a long journey for Michal, starting with a master's degree in taxation at the University of Bratislava, founding his first company and now already

in his second startup with TwinSow focusing on digital twin. This is a super interesting topic. Let's welcome Michal to the show. you so much, Markus. Yes, it is quite quite unexpected that someone who studies bookkeeping and taxes goes into the IT. But in my case, it was a coincidence that I was not in IT, but in the accounting because I'm an IT guy.

I was programming since my 16 and then I went on the University of Economics is just a coincidence. So there was no plan, but well, that's how it started. And it's always good to have financial knowledge and background because it's easier to be an IT guy with financial education than an accountant with IT education. I agree on that. And so

As you said, you started with programming already in your youth. What were the things you programmed back then? Yeah. So it started in 1997 when we got the internet at the university. We were like, what internet? What's that? Okay. Let me check the email, you know, and then we realized, there is this worldwide web and okay, let me check. then I discovered, ah, I can do those websites. So there is this HTML.

At that time it was HTML 3.2 I think. I bought myself this thick JavaScript book. What is right now this small portion of JavaScript possibilities. And then I created my first website and in the Slovak, let's say Yahoo, I had the website and one day came an email from Switzerland. Imagine that. Hello, hello, we are looking for Slovak designer, la la la, we are in Slovakia now. Do you want to meet?

And as you know, I can speak German. So I pick up the phone. Hello, hello. May I talk to Mr. Rosenberg? Bam, bam. And I had the first deal. Two years later, I was able to thanks their funding, but not in the venture capital meeting, but they paid all the people and everything. So I started a business and it went up until the dotcom crisis because then it. Yeah. Let's start first in your point of view.

What is the definition of a digital twin? Whatever you want. I left all these discussions about what it is, what it isn't, and what's the digital thread and digital this and digital that. It's quite annoying because of course everyone who wants to become interesting needs his own definition. They need to find their own business case for their living. The definition and the scope of the digital twin topic can be very, very different to

cover everything and every aspect in one definition is for me impossible. First of all, I would like to show you how it actually began. So because it wasn't at the very, very beginning about digital twins, it was about indoor positioning. And that's what you will see also see on the other slides. That's where you kind of came from. So.

Just to get that definition of digital twin, in my point of view, digital twin is the digital replica or twin of a physical thing. Do you agree on that? Yes, but once you dive in details, the mess begins. So we didn't begin with the digital twins, we begin and for those who are not watching, but just listening. So

Now I'm showing Markus my phone with the video from 2014. It's not a screenshot actually, but it's a footage from another phone. It was at that time KitKat 4.4 Android. And you can see here my fingers, how we are trying to track someone within a building. So we see a layout of an office floor and a lot of moving dots. And the moving dots are the

people in this case, which get tracked of where they are. And back in the days, what type of tracking technology was used for that? We used the phones for that because to track Android devices with a special application was at that time the easiest way. These technologies are called RTLS. So it's a real time location service. And there are plenty of technologies, usually RFID, ultra-vibe band, Bluetooth.

Wi-Fi camera vision so there are many options how to track something or someone within the building because outside is easy you have GPS you can use GPS to Wi-Fi connectors so even though we don't have the the GPS signal you can Use the the Wi-Fi you are connected to or you are seeing for for location But within the building it's much more complicated and you need a special technology for that. So that's how we started

And at that time we didn't know that there is something like a digital twin. But I said at that time, so 10 years and four months ago, guys, but I want to sit on the phone and I want to sit in 3D because all those buildings are three-dimensional, so I don't want to use it on a floor plan or something. And then we developed the technology, OK, but

We didn't do it for manufacturing businesses, but we wanted to sell it to event companies. Because if you have an event and there are hundreds of waiters and staff working on that particular site, you want to know where these guys are. Because there's thousands and thousands of visitors and you want to track your own people, but there is no business case in reality. So then we contacted one of our customers at that time with another project.

Do you think that this could be useful somehow for you? Bring it here was the reaction. And that was actually our first industrial project with RTLS and with the Digital Twin. They saved money because they had a track between two buildings in the plant and they could send the track away because they were able to simplify processes, logistics processes and save, I don't know, like 80,000 euros.

yearly just by, my guys don't work that much as I thought they do. So it was very, very simple. Okay. Now back to the digital twins. So let me share with you my screen now. And for the people who are not watching, I will try to describe what to see. All right. We'll see your screen. Okay. Cool. So what did you

could be? Well, it depends as we began with the discussion. Whom do you ask? Here you can see a structure architecture of how Microsoft sees the digital twins. So there is of course in the middle is Microsoft Azure gathering all the data from different sensors and systems, tables and databases and all the stuff. However, there is no visual representation. They see it more like data. Is it valid?

No, yes, but then you talk to the architects and construction guys for them is the digital twin their BIM, so building information management software where to plan the wall construction and okay, we need 27 doors and we need these windows and this piping and electricity and sprinklers and everything. Is it a digital twin? Yes, but again, if you need to sell something.

look at this video. So here you can see how a digital twin can be used also in a sales process. this is by coincidence another Slovak company called VAU Studio and they created this platform where the developers of residential buildings can sell their apartments and buildings and you can configure, okay, I want this mirror and I want this and that and this, not this toilet, but that toilet.

And I don't want to have white doors, but brown doors or whatever. So you can design your apartment through that digital platform and have a real time view in how it could look like. Exactly. If you want to use the standard the developer is providing. And so you can configure everything and in the end in their CRM and you place the order, you pay the

the initiation fee, get the invoice, bam, bam, ciao. Then you have the order and this is going to be your new apartment one day once the developer is finished. Then, for example, during the construction process, you can use technologies like Goshen splitting. Look at that. This is a 3D model. This 3D model is photorealistic and it was created from a video or drone footage.

So you take pictures from the drone or a video and for people not looking, this is a construction site. There is a crane, a lot of material, some started construction. It kind of looks like Google Earth animation, but it's not from a satellite point of view like on Google Earth. It's actually from a drone in a hundred meters, 300 feet kind of height.

So therefore, all the details are there. And we see it in kind of a 3D, and we can see it from different point of views. Exactly. And the logic is that you send the drone guys every week to the construction site. They fly all the time the same route. And then you can compare the progress week by week.

Therefore, you see your construction process, how everything was built and when what part was actually built. Exactly. From outside, from inside you can't use it. You need to use another technologies for interior, but for the outside works, this is perfectly fine. Yeah. For the inside, can use, for example, a 360 camera and then have someone walk.

through the property and and capture there for everything. I saw that use case before. Then if you ask Siemens guys, what is a digital twin? Then you get this one. So this is the plan simulation where you can plan and simulate the production process. So you have workstations, have conveyors, you have people, you can say what process takes how long.

And you see, okay, this is the input of the material. is the output. is this, is this something what I'm expecting or not? Should I do the relay out of the plant and so on? However, it's like 1990s. From a graphics point of view. Yeah. Like Prince of Persia at that time, the first one. Because when you look at this simulation, this is

Nvidia Omniverse.

Again, not perfect because the complexity of the software and documentation, are challenges. However, if you look at the level of the simulation and the visual quality, look at that. now we can see an engine, how it explodes. And seeing each individual nut and bolt on that level. Exactly. So we can get to this level. Then...

If you have an older building, old facility of any kind and you need to digitize it, then again, you need something. And for example, this is somewhere, this is a building on the first side, let's say 18th, 19th century. And of course you have zero documentation to that building. Yeah. So you can use technology like Navis in this case, German guys, they have a scanner.

you put it on your shoulders, you have one LIDAR scanner above you, one in front of you, and you see in front of you the tablet, you see what you have already scanned, and it creates so-called point cloud. So dots in the air and it can have some shape, and then there are automated tools being able to convert these point clouds to meshes through some documentation, then you just fix the things you need to fix.

Another use case. This is from the movie. Markus, you recognize it? Not particularly, but I guess they explore a cave or something like that. So it's from the movie Prometheus.

It's 2013. Okay, so 12 years ago and the guy threw in the air some balls, but let's call them drones and they fly through the air and scan the cave and here you can see how it creates the digital twin. You can see the position of the people, the position of those drones, the shape of the cave and so on. So this was 12 years ago.

And when we saw it in the movie, it was like, this will never ever happen to see something like this. Sure. 2023, 24, 25, you can buy it. Okay. Of course, it's not a ball without propellers you just throw in the air. No, it needs a pilot and it's a proper drone and so on. However, you can scan also this hard to access locations like...

caves and tunnels and things like that. And the drone flies, you pilot it, there's cameras and lighters, scanners and everything, and you can get the documentation. Also, you see a lot of dust there. So we can buy it. Yeah, it's possible and you can do it on a price point which you can afford. Yes. And also when we talk about safety,

send someone to do the documentation of such a location. Two dangers. So this is what the digital twins could be. However, where is twins zone now?

So what do we do is that we take the 3D data, the live data from information systems, IoT sensors, whatever it is, and we glue these two things together. Now I will show you a few videos from our projects. All these projects were publicly presented somehow, so there is no problem with that. This is an

office building in Budapest, a large oil and gas corporation. So we say, okay, this is the 3D model, fine. But then I click on the people and all these orange dots are people at that time. Now, but you have this live view on site. Then they said, Michal, but we need to integrate the room booking systems, plenty of meeting rooms.

You click on one of them, OK, it's excellent, it's free, let's go there. They have plenty of focus rooms. you click, and you see temperature, humidity, CO2 level. And this app runs on such totems. So imagine the touch screen. And you go there, OK, where is Marcus? And for facility managers, you see the temperatures and humidity and CO2 levels of every single room in the building.

And this project is 2019. OK, so it's six years old. And once we delivered this project to Bratislava, to London, to Dubai, to Budapest, said, hold on. We have to build something more robust, with better insights, something more generic. And that's how we created WinZo. So here you can see all the WinZo.

Everything starts with the earth because our customers are global companies and they have plenty of locations Usually we start with logistics as there is the best business case here You can see the heat maps the spaghetti charts. I can click on the on the play button So you see what the guy is doing and then you can begin to ask. Okay, what did you do in this part of the plant? And you see the time is running 20 minutes and he began to work after 20 minutes

What did you do those 20 minutes? Can you visualize all the coffee machines on the way there and see where they made all the stop? No, because they are delivered by sub suppliers, so we can't. But, so this is the typical case. Please tell me what the logistics is doing. But then they say, okay, but we don't have just logistics. We have also production and we have

Some data in SAP. We have some data here some data over there in Excel files Google Sheets and so on so we have open API's and you can integrate the data To to twins. and you see okay. This is how the production is running 99 % Please show me this is the paint shop 93. Uh-huh, but there was a drop but okay, it's Still green if I click on this production line. I see that the production line is exposed now

but it was running on 100%. And as it is multi-site, multi-floor, multi-tenant, multi whatever technology, if you have 10 plants, you can jump from one plant to another and compare how they perform. And this is, for example, the project at Skoda. The plant is like two and a half square kilometers, 25,000 people were there, but this particular building is 70,000 square meters. You see different floors.

you see the machines, the stamping forms. see if you click on the production, you see the temperatures and the performance of the production line. For example, what they need to manage is the garbage because they produce a lot of waste daily and they need to know, okay, here we have three bins, there is five, there is for this type of garbage, for that type of garbage.

of course, logistics. So you can say where the AGV or the forklift spend how much time. And you can see the heating cooling because of course, these old information systems don't have information about the location. Yeah. And I think that's the big value what this visualization brings. And a lot of

IOT visualization, you see the data kind of in a spreadsheet or in a dashboard. The issue is specifically in engineering and manufacturing that a lot of things are called in a different way. So everybody talks about the same thing, but the one group calls it this way and the other group is calling it this way. So if you put it just in a dashboard,

Data is getting misinterpreted by the label on the data. What I see as a big value is if you add the graphic or the layout, the model to it, it doesn't really matter what is the label on that data. Just by looking at it, I can identify what it is.

and it even doesn't matter what language it is. You can show me this in any language because I understand the layout of the factory. I understand the data, so it doesn't really matter in what language that variable tag is, right? Yep. And not just that, it's also about the

and the technical stuff about the data integration because right now the information systems are very separated. Of course, the suppliers don't talk to each other. They don't want to give the data. They don't want to provide the APIs and so on. And also these reports are created by engineers. They understand it perfectly because they know that the machine 27 is this one.

But even the maintenance guys don't know which machine is which one. They know the problems of those machines, but if this is 17 or 437, they don't know. And the 18 has a problem. Which one is 18? So... Yeah, go again in your visualization in animation where we see things. That's pretty fascinating. So also the management and also the shop floor guys understand the things.

much easier and for remote and especially central teams is this important because they don't know the layouts of all their plans. Specifically, if you then look in another factory where you might be not familiar with how the specific line looks like and what's the naming convention on that part

Exactly. And then that brings us actually to how this is done. Because the most common question is, okay, Michal, but how do we create those 3D models? Everyone is scared of it. And in the end, the 3D models are the easiest part of the project. Because what takes most of the time are the security assessments,

the data integrations to get the data to identify the use cases and this stuff because to create the 3D models with the digitization technologies being in place right now, there is no big deal with that. In one to three weeks, you have all the layouts and you can use it. And of course, it is useful also not just for in our case, mainly logistics or production, but also for facility management, in many cases for safety or security.

because also there are also dangerous places. For example, in chemical industry, there are places where you don't want to be not even one minute longer than necessary. Yeah, there are many places in manufacturing and industry which are hazardous for chemical radiation and other things where you don't want a human inside. Yes, yes. So everything starts with the...

connecting all the data points together because what we see here are the three steps. We are now able to deliver step number one and number two. So we can collect the data, we can create the 3D models and then we can create the digital twin. And I like your picture with the Legos. So on the first step connecting all the data.

you just have a pile of different Lego bricks. And then on step two, the global digital twin, you already put a static display together like a house or a building. And then in the last step, you have that model animated. Yeah. you, so in the third step, it's about combining

the prediction, the planning or simulation and the reality. When we put all those things on top of each other, then we can begin to find discrepancies and, I thought that this is this way, but it's actually that way and managers can begin to see where the problems are usually coming from because

Our typical cases, and then also the customers ask about the ROI. That's what we tell them all the time. Guys, let's start with logistics because I guarantee you that the biggest issue is exactly there. Because to have the material on time on the production line is always an issue. The speed of those forklifts can be a problem because we saw incredible speeds.

allowed in the plants because they needed to be fast. But then they realized, aha, but my team is wasting time here and there even without knowing that. Because the people were used to do it this way. And they didn't confront their daily reality with another options. And then when the logistic managers saw all the spaghetti chart and hit them, they're like, what?

Why is he driving through the wall, planned 20 minutes, somewhere to pick something? We should have a buffer here. in, say, three times, you just feed the buffer with all the material, and then we have time. And there are plenty of places where to save money. And it always reminds me on the best ROI ever we had, now.

It was like 1 million euros yearly saving in OPEX that they had for a 30,000 square meters big plant, 37 forklifts. We're like 37 forklifts for 30,000 square meters. It's a little bit too much, would say, because let's say 2025, but not 37.

and the issue was the plant manager because when the plant manager went to the shop floor and he saw someone not driving, number 9, okay, number 9 will not get bonuses this month because he is not working because for the plant manager driving means working. Is it true? Not that often.

It depends if you work productive or unproductive. If you just take your forklift and drive circles in the plant, it looks like you're busy. So they increased the fleet and new guys and more guys and more forklifts and still the productivity was low. And why is that? Well, because the plant manager explained him very well. I want to see you driving. So we were driving.

But then a new logistics manager came and he told him, you know what, let it be and I will manage it somehow. And after two months, he was able to cut 17 from those 37 forklifts. And in Eastern Europe, the cost for the forklift and three guys for three shifts is yearly 60,000. In the US or in Germany, it would be like 180.

And 60,000 times 17 is more than 1 million euros just because the people began to work instead of work simulation, so to speak. that's quite a good ROI. OK, this was the best one, but it is still a good example how far the companies can get with poor management.

I agree. Pretty impressive in how you do that from connecting the data to getting into the digital twin and then animated with the, how you call it, MASEp, the manage, analyze, compare, evaluate, and predict. Can you give a few more examples on that? So with prediction, it's complicated.

And I was at multiple conferences when the high-ranking managers talked about AI implementations. However, all the time it was about these large language models. So we have these documents, know, and we will ask it and... Okay, fine.

but it is actually not used in the production itself or not even in logistics because the development of such a technology is, the cost is high. The complexity of input is even maybe higher. You need in the team someone who understands the things. So right now we are giving the information. The prediction is more about using common sense.

because the companies are not that far. And we are working for world's largest and most advanced companies ever. But if you take a large business with 50 billion in revenues and they tell you, Michal, you know we have 6,000 applications.

with 6,000 whatever, it is a problem. And if you have to define the correlations between the systems and this is good and this is bad and this is what they expected and this is what they didn't expect.

It's complicated. So I can imagine predictions on, let's say, one to maybe three parameters combined in a chart. We can use more or less simple statistics for that, like flowing mean or these basic predictions. if someone expects that

that an information system will tell him what the logistics, example, in our case, will do tomorrow and in two days in three days. This won't happen and it will take a while to get there. We're not there yet, but kind of an outlook in the future. My understanding is that's where a lot of development goes under the umbrella of AI and large language models and

bringing all this unstructured data together and get something out of that? It will come, but it needs time and it needs a good business case. Indeed, it needs pioneers because if someone has to invest with not that confirmed vision of the result, all these large companies are

are stuck with that because, okay, Michal, when we will get the money back? I don't know. I can tell you with logistics, with this kind of stuff, where is the ROI. But when we talk about AI and how the data integration combined with predictions will save you money in energy, safety, let's say,

a furbishment or rework, I don't know. You have the data somewhere, but even you don't know what exactly it means and how to evaluate the data and what can we do with this level of information quality. In my point of view, specifically manufacturers know where are their problems and where, in what areas they think they...

lose a lot of money. And by understanding these problems, in my point of view, your platform and the digital twin technology can be used to then solve specific problems and then have out of this, of course, the ROI. Instead of saying, hey, I want now a digital twin of my factory, the approach should be

Yes, I want a digital twin of my factory, but I want to solve this and that problem first. This is funny because we have actually these two kinds of customers. The first kind, first type, they know exactly what they want. Okay, Michal, this is my use case. This is my problem and I expect to solve it by, by Twinzo. But we have also the second type of customers who know, okay.

We know that we will find the business case, but we don't know it yet. And we think that it could be these 10 things. Logistics, cybersecurity is also a case. Employee onboarding, for example, and this type of stuff. But they don't know exactly where the money will lie in the end.

but they have the budget for it. Michal, want to evaluate it. And once we see, so we tested these 10 scenarios, three are nothing, five are more or less, but these two will have the payback for all the stuff we need because that's the biggest pain point and the otherwise just sort thing and we have it.

So both approaches are valid, also very strict, very concrete, just one topic, and also this broader one for more visionary companies. I agree. There are always more than one passes to Rome as a saying says.

Yeah, quite interesting what's all possible in the digital twin space and thanks for explaining that all what you and the TwinZo do there to add value. How can people get in contact with you if they're interested in that whole topic of digital twin and the TwinZo? Yeah, so...

It's quite easy. If the people are on LinkedIn, just drop me a message. Yeah. And we will post your LinkedIn profile in the show notes. Thank you. Thank you so much. The other option is twinzo.com. And here I would say for the enthusiasts and with people with some technical background or at least very early adopters, you can also go to the App Store or Google Play and download the Twinzo application.

There is a big demo zone. And in the demo zone, there are plenty of use cases, so you don't need to register, you don't need to do anything, you just click, okay, show me the demos. And you don't need to be a rocket science engineer to do that? Yes, but there is also the ISS, so the International Space Station as well, but there is a plant, construction site, hospital, stadium, smart city.

like more than 10 use cases so everyone can find his own taste. Very good. That is quite amazing and we put the link to that in the show notes as well. Thank you, Markus, so much for the invitation. Hopefully it helps. Yes. In digitalization, it's all about spreading the word, talking about use cases and applications as

This topic has overall some complexities and a lot of people are still in my point of view in what I call the digital fog where they do not really identify what is in for them. And through that podcast and things we share on LinkedIn.

That's also a learning and education. That's that's how I see it and that's why I do this and thanks again for joining this episode of the digital tone podcast Thank you so much one more time to the to our listeners all the best and enjoy the rest of the week. Bye. Bye guys