Minimizing Cost of Ownership with Simulation and Digital Twins
Carl Wouters, Atlas Copco
Atlas Copco is constantly looking for ways to improve their products and services throughout the whole product lifecycle of their machines in all aspects. To reduce cost of ownership, they integrate simulation and data analytics from engineering to production, sales, and services. One of their objectives is to enable thousands of sales engineers to do reliable performance simulations during customer engagements while interacting with customers. Another objective is to allow the services division to set up different customer-specific maintenance strategies based on real-time data collection from more than 100,000 machines in the field.
Connecting colleagues, customers, and the supply chain using a single source of truth results in shorter manufacturing throughput times and improved workplace safety. This is achieved by establishing an enterprise-wide data acquisition and analytics platform. The importance of a model-based engineering approach is key.
This presentation showcases how this approach leads to better products and services throughout the whole product lifecycle.
Recorded: 20 May 2019
To be honest, I'm really happy to stand here and be able to show some of the really cool stuff that we've been doing at Atlas Copco. And but I want to give a very big disclaimer. I will speak about some buzzwords today. You will hear about digital twin Industry 4.0 IoT, but I hope at the end of the presentation it will be clear for everyone here that it's a full journey that we're taking. We still feel that we are at the beginning stages of this journey, but I want to bring you all a bit up to speed on how we are approaching this step by step.
I loved in the previous presentation the focus of engineering. And that's why we're really happy to stand here before this engineering community because we hear a lot of software discussions on all of these buzzwords popping up, and typically they're looking for data scientists-- data scientists, everyone needs hundreds and hundreds of data scientists. But very often, these aren't engineers. They don't know anything about assets, they don't know anything about products and how to develop these products themselves.
And the approach we're taking at Atlas Copco call is completely different. We are really coming from an engineering background. We are very much engineering company, and we are really looking for us the mechanical component, the electrical components, the software components are really core-- the full essence of it.
Now, maybe before going into the real-- the meat of the presentation, who knows about Atlas Copco and the products that we have? Please put your hands up if you know Atlas Copco. It's really not a lot, so let me first give a small introduction on what we do.
So this is a typical compressor room, but let me go a bit further on. Now I think-- I guess most of you came by car today. More than 1/3 of the cars worldwide are produced with our tightening equipment, with our assembly systems, and also using our compressed air installations.
Now, I drove today from Belgium. It took around two hours drive, so it was still OK. I already had a lot of coffee in me just to stay awake. And a lot of the coffee processing is done with also oil-free air, so that's really a key. You don't want any oil in your air or in your coffee or your cookies or in your food, so it's really to focus on oil-free air. And don't forget the mugs-- oil-free compressed air.
Now I-- for me I'm a Belgian, so we really love our beer. I'm a big dubbel fan, so for me this is one of the most important slides that I will show today. But it's really impressive to see that more than 50% of the industrially produced beer is using Atlas Copco oil-free compressors in the whole process. And I think as Belgians we are pretty proud of that because 50% in the world, that's a lot.
So what is Atlas Copco? Atlas Copco we are around 9 billion revenue. It's a company almost 150 years old, which is, of course, pretty impressive track record, and it's a Swedish company. We are split up in 4 key business areas.
So the first business area, this will be the focus of my presentation today, is compressor technique. It's around 45% of the revenue-- of the total revenue of Atlas Copco, and it's really focused on compressed air installations. But not only products, a big part is also adding service to it, adding our SMARTLINK platform, which is really connecting the devices to the customer infrastructure.
Now another business area is vacuum technique. It's one of the newer business areas, it's founded four years ago. And there we focus on vacuum technology, so semiconductor industry if you're food and packaging applications. Another business area-- industrial technique. This is really assembly systems, so the real tightening equipment to operate a whole production line.
And the next time you drive by a big construction work, if you're-- for example, they're doing a lot of construction works on the roads-- try to look for the colored boxes next to the road. That's what we call power technique. Inside of Atlas Copco we call it everything on wheels, so that's power technique. It's compressors, generators, lighting equipment, but everything on wheels. So the focus of today will be compressor technique, and I will try to give you some insights on what we're doing there.
But let's start off with one of our key products. This is a new product released last year. We call it the ZR 160 VSD+. I know it's not a fancy name-- we are not in the consumer market, so it doesn't really matter how we call the product itself. But we completely redesigned almost every component of this product.
And it's important to note-- so we have two compressor elements we will walk through next. These compressor elements, these screws, are running at micrometers from each other, so that's what you see here. These screws are really running at 10 to 20 micrometers from each other. They can't touch each other or you break your machine. Both of the screw elements have permanent magnet motors, completely engineered and produced by Atlas Copco.
If you look into the cooling systems-- cooling systems, again, fully engineered and done by Atlas Copco. And you could ask yourself the question, OK, why does Atlas Copco-- is it really core? Well, we are moving step by step away from pure component optimization. For us to keep on increasing the efficiency and keep on increasing the reliability of our products, it's really full system optimization that we want to do.
And actually, not only system optimization of the compressor itself, but actually of the whole compressor room. It's not only taking in scope in certain components. And that's why for us, core is really the whole compressed air system. And you will see also in further on in the presentation, it's not only the single product, it's really we want to assist our customers in whole optimizations of their vehicle.
So here at the end you see the final water separator and then to the end. But we need to be honest. For me it's a really sexy product, but I guess for you it's just a gray box. But there is a lot of engineering effort and production and marketing effort going into it, so this is our ZR 160 VSD+.
Now in the rest of the presentation, I will use this product to show how we use simulations and this digital twin, how we see it in Atlas Copco to integrate this into the whole process. So it is I need to say the most efficient compressor, but my marketing guy said don't put it on a slide because else we could get attacked by anyone worldwide. But we are really happy that this-- we increase the efficiency with more than 10% across the full range, which is a huge leap in a mature market like compressed air industry.
And the development of this kind of product-- like I said, we are developing every component, it's almost completely engineered and produced by Atlas Copco. So it means we have a huge differentiation in product teams that all need to work together. It's hundreds and hundreds of Atlas Copco employees that are involved in the development, production, marketing, and sales of this whole system. And that will also bring me back to, later on, on why we need this digital twin in our company.
Now, it also has more than 50 sensors. A lot of them are used for control operation, safety, and reliability, but we're already preparing this product for future predictive maintenance. So we're already integrating sensors to learn on how these machines are running and [INAUDIBLE].
I don't think that a lot of people understand this, but more than 10% of the electricity consumption worldwide are compressed air systems, so that's huge. And so you think about having this compressed air system here, 75% of the total life cycle cost of a compressor, it's just electricity-- it's your electricity bill. So if this compressor runs for 10 years and it should be able to run for 10 years long 24/7, you're spending more than a million euros just purely electricity.
And there is just taking $0.10 a kilowatt hour if you're thinking in China. Belgium, typically, we are much higher than that. So it's a huge cost, and we're not taking account yet maintenance, installation, and so on. So that's why, of course, at Atlas Copco it's really focused on energy efficiency next to reliability because if a compressor fails, production plant fails, so it's really important.
Now, what are the challenges in developing this kind of product? So as I said earlier, it is a mature market. Compressed air systems are already there for a long time and Atlas Copco was there. So Atlas Copco already was from the early '90s developing products, and our first generation of oil-free screws was in 1956.
Now the industry is really pushing us more and more into the shorter time to market. And I know every presentation it's again the same story-- shorter time to market, what do we need. But it's really a big focus because we are a global company, which also means that we have competition worldwide and it also-- industry also wants to keep on pushing, pushing everyone further and further and we are really in this. Let's say, it is a challenge.
Now as I said, hundreds of Atlas Copco employees-- how do you make these employees work together? How do you have this completely cross divisional worldwide collaboration and still be able to reuse as much as possible? So I already covered a bit the components. And for us, this product was also a completely new way of developing our controller algorithms.
So in the past, compressed air installations they had a gearbox, the elements are running at a certain gear ratio from each other. and those are the most applications worldwide. Now these two elements now, the compressor elements are running with permanent magnet motors, and we can completely control each and every motor separately. They all have their own converter, which means that suddenly it opens up a full complexity and, of course, opportunity on doing complete modifications to a control system on the fly.
So we integrated a full self adaptive control algorithm, which takes in account ambient conditions, takes in account operations conditions during the whole life cycle of this product to really guarantee that this product keeps on running. Now, an important note, I always made a comparison with a compressor and a car, and it will pop up a little bit later also. Now if you select a car, you select your seating, you select your lights, you select whatever selection and configuration you want to do. In a compressor it's the same.
This variant has thousands and thousands of different possibilities and variants that you as a customer can select. And it's, of course, our customers want to have optimized systems for their application. They don't want to have an off the shelf application, which is more or less OK. They really want to have this full optimization for their system.
But this also is a big challenge for us, because it takes care to have a huge product reliability. We're thinking of hundreds and hundreds of [INAUDIBLE] variants across our product ranges, which is of course in development and production a huge challenge. And never forget high product reliability-- 60,000 training hours, it's a lot for something which is running at micrometers from each other, I can tell you that.
So, how do we tackle these challenges? How do you envision how to step by step integrate it? Now we use the buzzword, digital twin, there. It is a buzzword, but we really envision this as the roadmap for taking. And for us, what is a digital twin?
For us, a digital twin is the representation of your product during the full product life cycle, acting as a single source of truth. And that's really important, because often if you hear people speak about digital twins, it's typically the IoT story. It's how you link machines running in the field, linking that data with AI or with models combined to it. For us, it's really the combination of the full product life cycle.
So in the rest of this presentation, I will try to highlight these four segments on what we are doing with this digital twin and how we integrate it. So as designed, I guess that most people here are engineers, so that's for us the as designed phase. Now also important for us is-- and I heard it also in the previous presentation-- is the integration of these physical and the insights that you have inside of the whole implementation life cycle. And that's what we're trying to do with our digital twin.
So also looking in the as configured phase, how do you configure a machine? How do you sell it to the market? How do you take care of that these thousands of sales engineers use this correct data and make the correct optimization for a certain customer?
Next phase is as built. This is a typical Industry 4.0 story, right? It's gaining insights of how your production flow is working, and then trying to optimize your production flow for your specific components. And then the final one as maintained, the typical IoT story where I will show some of the nice stuff that we've been doing in the last years. Well, in my opinion it's nice. I hope that you agree.
So, let's start off with our core platform. And that's something we already started more than 70 years ago. And this is really-- it's been growing a lot, but we call it the Atlas Copco Model Based Engineering Platform. And the goal there is really to build a full community. And I will come back to it later, but it's really the foundation where we're building everything on top.
And in a company like Atlas Copco, worldwide we have a lot of different data sources-- we call them external data sources. This can be ERP systems, PLM systems, MES, it can be whatever system and application that it has. And typically, these data sources are very application specific. So what we want to do is an easy interaction with them. We want to build a platform which can easily take data and send data wherever possible.
Now in the last couple of years, we've putting these digital twin on top. And this is really what we see as this single source of truth. And for me, a digital twin is a combination of data together with real physical insights and physical representation, here defined as processing. It's not just like some companies try to list it down as Pure Data. No, it's really there's a full physical component next to it. And that's why we, from an engineering company, that's where we put a lot of effort on. And all the application on top, will walk through a couple of these applications in the next couple of slides where we really use this digital twin as a single source of truth, and then build applications on top that everyone has access to the same data, same information.
So the as designed phase-- and that's something that more than 70 years ago we started-- it's really our model based engineering community of excellence. We are a very, let's say, distributed company. We have a lot of different product teams in engineering spread across the world, and our key here is to have complete open information sharing between these product teams. And we use there the MATLAB tool chain as really a platform-- to standardize on our tools, to develop standardized modules that everyone can use, and we don't do duplication of effort.
Now an important part here is that we want to be fully integrated in every product team. It's not at this moment already like that, so we still have a lot of product teams who are not yet heavily involved. But step by step, we're integrating our presence inside of the different product teams, and it's really this knowledge sharing. It's not different product teams working on themself, it's really step by step we try to bring everyone together.
And in the beginning we had a lot of discussions, will we have an external team, just a separate team who does all the calculations, simulations as a sort of internal consultancy? But clearly we said, this is not what we want to achieve. We really want to move into a full community way of thinking. We really want to have the people with the knowledge-- and I heard it, again, from the previous presentations-- people with the knowledge in the product teams that know everything about their product-- they should be able to integrate their products and simulations. And that's really the key of it, and that's why we are really pushing this full community.
Of course, we have source control. There's also a lot of non-MATLAB code and they're fully integrated into the MATLAB platform to take care of it. We can build a complete skilled workforce. That's really a lot of people, a lot of engineers, have the same skills and again we share the calculations and simulations as much as possible.
So let's move into the ZR 160 VSD+. I know, not a sexy name, but I will say it still a couple of times. So we define the core framework as the typical calculation engineer. You will use a lot of MATLAB libraries developed by more than 60, 70 people worldwide to really do standard calculations. Simulations on, OK, what would be the performance of this compressor element if we change the tolerances? What would be the impact of a gear selection on our compressor elements? And if you do measurements in the lab, how do we efficiently take measurements and try to map it into our physical models? That's really the typical calculation task.
Now next to that, because we have more and more complex controller algorithms-- the time where a compressor is just on, off, does the control algorithm-- those times are far behind us. So there we really, with this complete new concept, we used Simulink tool chain code generation to really test already algorithms completely from nothing. And to really be able to test it in a reliable way, for us, that's there the physical framework. It's really doing transient modeling of our compressed air system, really bringing the compressor into complete transient behavior, and try to analyze and simulate this and couple it together, which are control written to really have reliable testing infrastructure.
Now this is all as designed engineering task. I think these are stories that already a lot of people have been hearing for a long time. Now for us, one of the key aspects is also full integration into other applications. And with other applications, I mean sales tools, I mean IoT platforms, I mean production systems. And that's where the interface framework, where we really integrate into our digital twin, will pop up.
So, what is this digital twin platform how we see it? So again, it's a combination of storage together, with this processing power together, with these physical models and physical representations. So how do we do it? We do our MATLAB models in the community of excellence. Everything what we have there we can deploy to our cloud's cluster.
So we have a MATLAB production service running on our cloud infrastructure because we really see this as the single source of truth. We really want to integrate into sales use, we want to integrate into production, want to integrate into IoT platforms. That's also where we are looking into cloud infrastructure to have the scalability, because it means the higher the requirements, the higher processing power you will need. So it needs to be easily scalable up and down-- the down is typically not really done a lot.
So an important part is also reliability of your models. It's not only an engineer anymore working at his desk, doing a simulation, nice [INAUDIBLE] simulation is done. No, we are really integrating these engineering models into a sales tool. OK, then of course, you need to have full reliability of your models in this deployment process because first, this digital twin, it's a living thing. It's something that's continuously modified, contains every information we get, we integrate it in there, which also means that you need to take care of your whole supply chain of your models.
So let's move into one of these applications, and that's where I move into the as configured. Now we have these nice engineering models. How can we now really use them to optimize a customer, to really optimize an application at a customer? This sales tool now-- this is one of the latest additions to the digital twin family, let's say in Atlas Copco-- is used now more than thousands of sales engineers worldwide. And the nice thing is, it's exactly the same MATLAB models from our community. Exactly the same data is coming out of this sales tool as from the engineering community.
So that's, of course-- it can sound a bit silly, but for me that's really impressive. If you have all these product teams collaborating in a platform and then being able to share it in sales engineering, this is really a key for us. How it was done in the past-- a lot of customer centers doing a lot of manual labor to try to get the right data for the right customer in specific. Now we're really optimizing the selection procedure.
Of course, I removed some of the performance values to not show what the real values are of the compressor itself, but this is really the engineering data running in the back end. And the sales tool, it's just the front end because all processing is done on our MATLAB production server, our digital twin area. And it's of course a combination of static data with algorithms and performance data.
Now the moment you select it then-- the final product, OK, this is a product we want to sell to our customer because we believe this is the one he needs. Then you go into the quotation process. So then you need to put sales figures there, you need to put more technical data, how will you connect the machine, how will you install it, which is then, of course, a whole quotation procedure.
How was it done in the past? Manual, manual, manual. For every sales engineer, we had a guy in the customer center or a girl in the customer center just making quotations-- gathering data, making Word files, Excel files to get together all the technical data required for a sales process. At this moment-- fully automatic. So for all the engineering data, it's coming straight from our digital twin. For all the marketing data, it's coming straight from their databases, and there we are also looking into final integration of it.
So now it's just a push of a button, you make your configuration of your ZR 160 VSD+, you select what pressure, variants, frequencies, you make your selection, you press a button, you get your full technical documentation next to it automatically after one second. And again, same models running in the back end. Not storing huge amount of data, of all your performance data, because we have hundreds of thousands of different variants. If you need to store all the data and you need to keep track of it, it will be a very big part.
So let's move in-- what you're doing in as produced. And there, I said I like comparisons with cars. For the Formula 1 lovers here, please don't shoot me down, OK? It's just-- I'm not minimizing the incredible effort that these engineers are doing. I just try to put everything within perspective on what we are doing.
So I'm making the comparison between an oil-free screw element and then a Formula 1 car engine. So our oil-free screw elements, they run between 3,000 and 35,000 rpm depending on the capacity that our customer is requiring and depending on the efficiency that we want to gain. But they can reach speeds continuously of 35,000, so they can run at 35,000 rpm for the long activity which is required. A Formula 1 car engine, there we are looking at smaller than 18,000 rpm, which is, of course, limited by the Formula 1 organization.
Now if we think of the tolerances of these both, both of them are also in micrometers. If you look at the pistons within a Formula 1 car engine, it's in micrometers. I couldn't find the final version of it because I guess it's also very confidential what they're doing there.
But for our compressors, compressed air screws, they're really running at 10 to 20 micrometers from each other. And I hope that you have to think a bit about it because 10 to 20 micrometers running from each other with pressure ratios of around 4, a temperature higher than 200 degrees Celsius, and there's no lubrication whatsoever inside of the compartment here-- no lubrication. There's only lubrication on the bearings itself and the gears, of course. So this really can't touch. The moment that they touch each other continuously, it's done. Your screw element is completely over.
If you look at power density, of course, their Formula 1 engine is a clear winner, and it's really impressive what they can get of power out of this kind of lightweight engine. But there is one key difference in both of them. Formula 1 car engine after 20 hours of running, they completely overhaul the system-- they Completely change the pistons, they completely overhaul it.
Our screw elements need to be able to run more than 60,000 running hours-- so not just standing standby, not just running on idle-- real running hours, loaded hours. And that's, of course, a big challenge for our production flow. So you can make this kind of element, you can make it in prototypes, you have the perfect tolerances, and you can make it that it works perfectly. But how do you make hundreds of these? How do you guarantee your production flow that all of them coming at the back end is equally as efficient and equally as reliable? And that's really where we're looking into typical Industry 4.0 story.
Now the first part-- connectivity. And I will not go too detailed into it, but for me it's also an important aspect on how will you connect machines. Step by step here we have more than 150 CNC machines only in Belgium, and step by step we're connecting all of these machines.
But first, the key there is also similar where you need to have, as soon as possible, adding information insights in there. Most communication protocols like, let's take Modbus TCP, you really need to know what a certain data register means. Is it the cooling temperature? Is it in degrees Celsius or Kelvin? You really need the full specification on these data registers on where is it, what is it, what does it mean? And that is the information behind it.
And if you want to automate your process, if you want to develop AI, it's really the data cleansing, the gleaning of the-- filtering of data it's a huge amount of effort. So the sooner as you can enter information inside for data sets, the better it is for your whole process. And that's why we in Atlas Copco, we are really step by step moving into OPCOA.
I'm not going to go too detailed into the technology, but the idea is that this information around what is this temperature, what does it mean, is it in degrees Celsius, the engineering unit, it's all included in the data protocol itself-- completely included. And that's really how you can get, as soon as possible, in your production plant the information available. And, of course, then the next steps are adding insights, knowledge, and then really moving into, OK, how can we create wisdom, how can we improve our process, how can we optimize the whole system behavior?
So what are we doing there? If you have, again, the compressor element-- so we are really getting all data sets at this moment, all different applications, different databases, because someone chose for a certain PLM system, someone chose for a certain ERP system-- step by step you want to integrate them all. So from the community, we really have are pulling all this data in and then start combining all of the data together. And the key, again, is not just looking into machine data.
Typical story, again, I'm sorry to say, but you take machine data, you put some AI on top, and then it gives you a result. And that's what I really liked about, again, the previous presentation. You get rubbish if you don't understand what's coming out of it, you really get rubbish. It's really the insights that you need to include, and that's also why we combine all the data. We combine quality data, we combine even tactile measurements, measurements sometimes done by hand we need to go in, and edit them together with a performance model with a simulation-- edit with a real measurement data at the end. And that's really something we're now heavily ongoing in process to step by step get all these data connected.
Now this is something that I think that Atlas Copco can be very proud of. A lot of companies are discussing IoT. A lot of companies are saying, OK, we want to connect 1,000 machines, 100 machines. Atlas Copco has been doing this now already from 2012.
In 2012 they started with putting a 2G chip in every compressor that we sell. It moved then into a 3G chip leader on, and we start with the bigger machines then step by step we were scaling it down. And now, every machine coming from a production line in China and from Airpower, which is Belgium, every compressor gets this chip implanted. We have already more than 120,000 machines connected worldwide, which are sending sensor data, operational data continuously to our data warehouse.
And most of you will probably be engineers. I can tell you this is a bit of heaven on earth if you start analyzing the data, it's really a lot of fun. But I really like this graph. I had to check with, of course, our management-- can I show it, yes or no? But I'm happy to show it, that we are really globally, and it's really everywhere in the world that we are sending data continuously to our data warehouse.
And the insights that you get from that information, the insights are incredible. We see now that our competitors are often completely misused, completely installed in the wrong way, and that's really, of course, where we can gain efficiency to our customers. We can go to that customer, run it in this operation condition, run it in those areas, maybe buy a new product, whatever, but we really can optimize the system level.
And we split up in four main segments. So first of all, we sell these as a product, of course. So we do condition monitoring, we sell to our customers. We can sell that we follow their compressors, that we do the full service, that we analyze how they can optimize their system. But they can also buy the product for themselves if they want to do the full condition monitoring themself. So they can access then completely the data what we installed.
But that's only the beginning step. That's what our first goal was to achieve and, of course, get a return investment on those first investments that you're doing. But all of the data is also easily accessible if you have access, because there is of course some access priorities on it. But if you have access to it, you can just with two lines of MATLAB code get all the data you want.
You can say, give me all the data from the ZR 160 VSD+ on all machines worldwide. After 10 minutes, you get all your data structured in MATLAB, and you can start your analysis. And this is now what continuously has been doing in engineering-- the moment we make new products, the moment we see some quality issues popping up, that's the first thing we do. We start analyzing, how are all the rest of these machines running in the field? In this, we are already gaining knowledge and gaining insights in how our products are actually being used.
Of course, this data is used for optimization of our maintenance strategies. So we have a lot of contracts. So there's a lot of contracts that we sell to our customers where we take over the full compressor room, and we will optimize their compressed air plant, and we will take care of the maintenance. Again, this is very cost efficient where you can really optimize, OK, that filter is completely not needed to replace them. Instead of the $1,000 and $2,000 that we see, you can really start optimizing your whole maintenance strategy through really cutting costs there.
Now the things that you're doing next is trying to integrate our engineering models. So again, our digital twin integrated into the whole IoT platform to really do predictive maintenance on it, and that's something that we're working on now. Also, together with MathWorks to try to fully integrate our digital twin platform inside of our IoT platform.
So although that I think that you're only still at the beginning stages, I hope that it was a bit clear that you are already on this pathway, that you're already through trying to do these things. And it's trying things, not working, and trying something else-- it's continuously doing new things. But there are already some things that we are very proud of.
And the things that we already achieved, like a company-wide platform that everyone can start building in their own applications, it's a really full collaboration. And it's really the start of this digital twin throughout the whole life cycle. And that's for me, really key. I hope that if you go home today that that's the bit-- the key message that Atlas Copco wants to bring-- it's digital twin for us is really the whole product life cycle in account on how you produce your machines, how do you develop them, how do you sell them to your customers, and really taking all that information in account.
Now, of course, we are a bit lucky. We have access to more 120,000 machines in the field. That's, of course, as an engineer, you're very lucky to be able to use that kind of data, but it really gives us a lot of value. And by showing this ZR 160 VSD+ where we really tried to incorporate all of these components in, we really show how value can be created already on the pathway to digital twin, because I will be the first one to tell you we don't have a perfect digital twin or whatever you want to call it. It's not perfect, it's still at the beginning stages, but already regenerating return investment by just doing it, by just implementation.
So what are the big challenges that we see? It's a lot of challenges are in the data sector. There's a lot of different data sources. Gleaning of the data, filtering of the data, bringing the data into a decent structured data set with data models-- it's not an easy task and it's something that's from a company, which really starts from legacy. If you're 150 years old, almost 150 years old, your data has been growing significantly. And it's not a greenfield anymore, it's really a brownfield, and we really need to plug into very old data systems. It's not that easy to already put the mentality of a real digital twin there.
Now there is one point I want to make, and that is we see that the investments in cloud is important, especially for scalability. It's really for the full integration. And it's, of course, also after discussing with MathWorks that we want to keep them pushing in this direction to really strengthen their presence there. And that's for us really key, that it's not just an engineer sitting at his desk doing a simulation. That's all nice, but if you really want to move forward with this, it needs to be all cloud systems that you can easily deploy these models and algorithms into the Cloud.
Now, for us, it's really the integration of the digital twin is key and it's successful. And I hope that today with the presentation, I could give you some insights what Atlas Copco has been doing, and I will be completely open for questions if you have them. Thank you.
选择中国网站（中文或英文）以获得最佳网站性能。其他 MathWorks 国家/地区网站并未针对您所在位置的访问进行优化。
- América Latina (Español)
- Canada (English)
- United States (English)
- Belgium (English)
- Denmark (English)
- Deutschland (Deutsch)
- España (Español)
- Finland (English)
- France (Français)
- Ireland (English)
- Italia (Italiano)
- Luxembourg (English)
- Netherlands (English)
- Norway (English)
- Österreich (Deutsch)
- Portugal (English)
- Sweden (English)
- United Kingdom (English)
- Australia (English)
- India (English)
- New Zealand (English)
- 日本Japanese (日本語)
- 한국Korean (한국어)