INDUSTRIAL

Christina Cardoza: Hello, and welcome to the IoT Chat, where we explore the latest developments in the Internet of Things. I’m your host, Christina Cardoza, Associate Editorial Director of insight.tech, and today we’re talking about industrial trends and transformations with Rainer Brehm. But before we get started, let’s get to know our guest a little bit more. Rainer, thanks for joining the podcast.

Rainer Brehm: Oh, first of all, thanks, Christina, that I’m part of the podcast. Yeah, and I’m really excited on the topic.

Christina Cardoza: Yeah, me too. So, let’s get to know you first though. You’re from Siemens. What can you tell us about yourself and the company?

Rainer Brehm: Well, first of all, I started at Siemens in about ‘99, and I was starting in the business I’m responsible now for, it’s the business around factory automation. So I started there in ‘99. Programmed my first PLC. It was a little bit odd because I studied computer science and it was a different language; it was more a language for an electrician, which is very useful, but it was not the IT language I was used to.

I was working there in Princeton at that time, where we have our corporate research department, really going beyond what exists today. And it’s a very interesting business, because, with Siemens, we are by far market leaders, so imagine every third machine or every third line globally is controlled by a Siemens logic controller, which means this is somehow the hidden champion behind the factory door, optimizing not only factories, optimizing a lot of processes. This could be vertical farming, that could be the metro line in New York, that could be the package claim in the airport. So it’s really a quite broad usage of the PLC.

And it’s also one important topic which really helps going forward on aspects of sustainability, because we strongly believe there will be no sustainable future without automation, electrification, and digitalization.

Christina Cardoza: Yeah, absolutely. And I’m looking forward to dig into those topics a little bit further. Wow, since ‘99 at the company. So I’m sure you’ve seen quite a bit of evolution, not just the language that you were talking about, but, like you said, adding all of these new advanced capabilities and new ways to do operations within the factory.

So, to start off the conversation, I’m curious if we can go over how you’ve seen this space evolve, and how it is going to continue to evolve next year—what sort of industry trends can we expect in 2023 and beyond?

Rainer Brehm: You know, the topic of Industry 4.0, I think, is probably well known in that community. And it was starting more than 10 years ago, I think even 11 years ago it started. And they were first ideas, but, due to corona [virus], due to supply chain constraints, that really accelerated significantly. And we see that it’s now really kicking in and it’s getting to be a reality.

So the trends we are seeing that are combining the digital and the real world—kind of digital, which is the simulation module, the digital twins, but then the real operation combining those, it gets more and more reality. So you simulate basically everything up front, and then you implement it. What also becomes now more reality is that you have a feedback loop. So basically when you have a simulation module and you implement it, but then you get the real-time data out of the operations and feed it back to the digital twin, then you can further optimize it.

The leveraging of data is significantly important because with that, now you can feed AI. Because AI isn’t yet really a big thing on the shop floor, but it will become, as all data is going to be more and more available. What we also see is, we call it a software-defined control, software-defined automation: where currently everything is very much bundled and tied with hardware, it’s going to be more decoupled, it’s going to be more virtualized. I think these are trends which we see.

And last but not least, I think, which is very important, especially when we look at the shop floor, the users of those more complex technologies, they are still the persons operating machines. These are no IT experts—the people maintaining the machines in remote locations somewhere in the world—they need to be capable to operate and to maintain those lines, those machine, those infrastructure plants, and therefore the topic of human-centric automation. So, how can we make it as easy as possible? That going to be a very important topic for the future.

Christina Cardoza: Yeah, absolutely. Certainly a lot of changes happening in the factory space right now, especially, like you mentioned, in the last couple of years. And you said in your intro that the future of sustainability really isn’t going to be reached without this factory automation. Also, I think the success of these digital transformations that are ongoing in the Industry 4.0 landscape right now.

So I think we know, like you said, all of these benefits and opportunities to these changes, but it can be difficult actually implementing them and deploying them. What are some of the challenges you’re seeing on the factory floor when it comes to trying to reach these trends and goals you just talked about?

Rainer Brehm: First of all, I think a lot of technologies are there. The topic, why doesn’t—or why didn’t—they scale and start scaling is that OT and IT people they simply speak different languages. I experience that within our organization. I’m more the OT guy, yeah? Even I studied computer science, but we have also very big software business, where we have the PLM software. When I talked about connectivity, we have problems with connectivity, I thought the connectivity to the real world, to the equipment, to the sensors, to their drives and so on. The IT person, when he talked about connectivity, he was thinking about connectivity to databases, to cloud, to data lakes. So it was even that word, “connectivity,” was completely differently interpreted.

And what we experience in our company, in Siemens, when I talk with colleagues, we experience that in our customers as well. So there is still a gap between the IT department—even the factory IT department—and the OT persons who are defining how you’re going to automate something, how you set up the equipment, how you set up the lines, how you maintain it to optimize it.

So there is a big topic on the languages. How do you bring the languages together? This could be terms, but this could be also how you, for example now, program the OT landscape, which I said was very much on the mindset of an electrician—not so much of an IT person. I think that is one main topic, how we can do that.

And, for example, we have now introduced a new programming environment called Automation AX Extension. It’s called Extension because it makes the OT world more accessible to the IT people, number one. Number two, the landscape is very, very heterogeneous. So even though the people don’t speak the same language, a lot of the machines doesn’t speak the same language because they’re also from a different vendor. They don’t have standards. So standardization is still missing, that you really can’t scale, you need somehow a standard.

And that also applies even to new machines, to greenfield. But it applies even more to a brownfield, because a factory normally runs, I means some factories run a minimum of 10 years, most are 20 years, 30 years. If you go to the energy sector, it might—or at chemical sector—it might run 40 years. So, you have a lot of brownfield, and that brownfield doesn’t speak the language which you might need to scale up. So I think these are the topics: how you standardize on your greenfield, on brownfield, in order to scale it up.

Christina Cardoza: Yeah, this idea of the IT/OT convergence is something that we’ve seen on insight.tech becoming more prevalent over the last year. And I’m excited, as we go into 2023, I think it’s just going to become even more important. And I’m excited to see how companies like Siemens are going to try to bridge those two worlds together. Because, like you said, there are things that need to happen now that just weren’t possible, or you couldn’t even think about when you started in the company. So now that we’re here 20-something years later, there’s a lot to think about. And especially, like you mentioned, the AI capabilities. There’s so many new devices and connectivity and just advanced features and technologies that you can utilize on the factory floor, and how do you now match that up with 20 years of technology or legacy infrastructure?

So I’m curious if we could talk a little bit more about that emergence of these edge devices and AI capabilities—how that’s complicating things, but also benefiting and adding new opportunities in the industrial space.

Rainer Brehm: Exactly. So if you look on an edge device from an IT perspective, that’s something a little bit different than maybe you look from an OT perspective. First of all, I already kind of elaborated a little bit on which people are operating it. I mean that was one of the main topics. So, how can you make it easy? And we brought a lot of cybersecurity aspects in. So, you need to have key management because to onboard a device, for example, believe it or not, that’s already a big, big hurdle. In IT, somebody managing keys is normal. On the OT side, probably most people when they buy our PLCs, for example, they probably disable that functionality because they are too complicated. So how you make this, which is normal on a IT side, accessible?

If you look further, there are some necessities, if you take edge computing. When you talk about edge computing on a shop floor, I think it’s very important that you understand that edge computing has some more requirements. So, for example, it should be in a lot of cases real-time capable. And if I talk about real time, maybe we talk about a chitter of microseconds. Because if you imagine a very fast process, in a microsecond a lot of things could happen already. And, if you’re not reacting fast enough then you might question a machine, or you might get to different results. So the topic of real time is very important.

Secondly, if you then want to deploy AI workload, for example, on a shop floor, and you want to react very fast, it’s important that this AI workload has an inference close to the machine. Simply because of the speed of light you shouldn’t put it far away. This is one aspect. The other aspect is also you want the AI to interact frequently with your real process. So basically you—so you’re going to interfere with the process, so you want to have that kind of close allocation, close to the machine or to the line. On the other side, you also want to take data out of the process and feed it back into the AI. So you also have—and these are a huge, huge amount of data which is produced.

I can give you one example. In our factory in Hamburg, we produce about 10 terabytes of data. So you don’t want to send the 10 terabytes of data into a cloud. You rather want to have them executed directly there where the source of this data is. So that is different, maybe, to a classical IT landscape. Furthermore, we have an industrial edge platform, which is DOCA-based. But we need to add, not only real-time capabilities, we need to add also the topic of safety. Because, you know, it’s a little bit like autonomous driving. Safety is a very important aspect. And you could imagine when you want to do autonomous driving in the car industry, you don’t want that the cloud is now defining whether you stop or not if a child is running on the street. You want that being executed as fast as possible directly in the car. So the same is on a machine. If somebody crosses, or a press is going down and somebody has his finger there, it should stop immediately. So you need to have that kind of fast reaction as one of the assets.

And another topic is, why not thinking ahead? Leveraging AI not only for optimizing processes, but also thinking about, couldn’t we use AI for a more autonomous factory? So also there we think, how can we use AI, that a machine, a robot could decide itself what to do? And then it means AI is not only optimizing processes, optimizing, enhancing engineering, but it’s really steering the robot, the machine, and the line. And that application for AI is really, really exciting, because it opens up really new fields for automation.

Because when I started in ‘99, what you automated basically was you automated very repetitive tasks. And mass production was perfect, because mass production, a lot of repetitive tasks. Or you automated something which is predictable. You couldn’t write a program “if-then-else” if you don’t know what is the “if,” and the “then,” and the “else.” So you basically can only automate what you know. If you’re now leveraging AI in automation, you suddenly could automate something which is maybe a “Lot Size One” and which might be unpredictable, so you automate the unknown, which is not possible today. So therefore AI in automation could open completely new application fields.

Christina Cardoza: I definitely agree. And I think we’re only scratching the surface of what’s available or possible out there. There’s going to be new ideas that companies like yourself and the people you work with are going to come up with together. And so you mentioned a couple of different solutions or efforts going on at the company, and I would love to learn a little bit more about how Siemens is working to make this all possible—if you can go over some of the solutions that your guys are using or providing the customers to make it a little easier. Or any use cases or customer examples you could provide us with.

Rainer Brehm: Several ones. I mean, let’s first maybe start with when we apply—because we have our own factories. I mean, so we apply something, what we apply to our customers, we apply it to ourselves here. So, one example of a use case IT/OT leveraging AI is in our plant in Hamburg, where we produce every second product that is going out of the factory, even more in the meanwhile. So it’s a very high throughput, and we have PCB lines, which in the past we—there’s a complex process how you put the components on the circuit board: how are you soldering it, and so on and so on. And at the end we normally did an X-ray in inspection of the PCB. You can’t do it with vision systems, because somehow you need to have soldering points below the chips, so you do X-ray.

In the past, where we had the X-ray machine was a bottleneck. So, leveraging AI, we basically now predict whether this PCB, this individual PCB, has a high quality or not. So everything with a very, very high probability that there is no quality issue, we don’t send to the X-ray machine anymore. We send it then directly, bypassing the X-ray machine, going to the final assembly. With that, we save the X-ray machine, for example. So we’re using data out of the process.

Another topic—optimizing processes—we see currently in the battery industry. You know, this is a big, big investment also in the US with the Inflation [Reduction] Act, a lot of battery manufacturing is produced. Currently it’s hard to scale them up, scale them up on the right quality level of batteries. So, how much material comes in? How many batteries come out? Still that is not a process which is mature enough and optimized. We see—and we’re working with customers—that we need to get the data of the complete processes from the beginning—mixing the slurry—at the end, kind of doing the aging information of the batteries, getting data from the different process steps, looking at those data and optimizing the process end-to-end, which is not done today, but we are working with customers on that.

Another example could be in infrastructure. So, we are using—we are doing tunnel automation. So, if you drive through a tunnel in, I don’t know, in the Alps, or in the Rocky Mountains, or somewhere in the world, there’s a high probability that those tunnels are automated and controlled by our PLCs. What we do, we now using AI more and more in order to detect an emergency situation in the tunnel—if there’s a traffic jam, if there’s fire, whatever. So you need to react fast—how do you evacuate the tunnel? How do you switch on or switch off vents, lights, and so on. So we are using, even in infrastructure now, an AI workload aiming to optimize it.

And maybe, last but not least, going back to the factory again, to automate the unknown. We have an interesting application where we’re doing real-time flexible grasping. So, a robot is not programmed, but an AI tells the robot where to grasp an aspect. So, you can see that on a fulfillment center in the logistic area. So we take something out of a box—we can do that without training the robot on the thing which needs to be picked up. We train the robot on the skill to pick up. So, basically the robot can pick up everything—if the gripper allows it—that is and that needs to be necessary. But, with the skill of grasping, we can automate something, we can grasp something unknown, unpredictable.

And my last use case, which is not reality, but where I invest currently money, because I believe that’s really something interesting, is: can you in the future automate repair? Because if you talk about cycle economy, the one topic is, how do you recycle things? And a more interesting thing is, can you repair in the future something? And we know that currently there’s a lack of people capable to repair. And if you take, for example, a car battery in the future, it consists of cells, so can you maybe in the future take a car, take a cell from a Ford in the United States, you go to a to a workshop, it takes out the batteries. There’s a defect, and a system can automatically detect where is the problem and autonomously repair the battery cell. If you do that, you automate the unknown, because every battery is a unique thing, it has a different lifetime. And can you automate that leveraging AI? So, some of the use cases where I’m really excited, that’s IT/OT convergence leveraging AI, leveraging new technology, really will make a difference in the future.

Christina Cardoza: Yeah, absolutely. A lot of exciting use cases and things to look forward to. I love one of the first ones you talked about, which was actually applying these things to your own factory, because it shows that you guys not only are solving the pain points, but you felt the pain points also, and you have the experience working within your own factory to remove some of these, so that’s great.

And listening to you talk about some of these, they sound like huge undertakings, and I should mention that insight.tech and the IoT Chat, we are produced by Intel^®. But I think a lot of these things require collaboration and partnership throughout the ecosystem to make some of these a reality, or to make some of these possible. So I’d love to learn a little bit more about the partnership you have with Intel, and how that’s been valuable to your solution and the company.

Rainer Brehm: First of all, we work with Intel probably, I don’t know, since four decades, way before I started with Siemens. But I know very much that we started in 2012 with the Technology Accelerator Program, the TAP program, where we said, “Hey, if you have an OT workload, the topic, especially on low latency, is a very, very important one.” So we worked very closely with Intel to enable the processes of having a low-latency functionality, especially for those workloads where you need to act in microseconds. So that was very, very fruitful and helped us to use the Intel chips in our controllers, number one. And I think it also—that helped Intel in order to have the processes capable for having more OT, or more real-time, real real-time workload. I mean, that is one important topic.

On the other side, we’re working with Intel currently, I mean it’s really the supply chain crisis, and, also thanks to Intel, I think we were capable to fulfill not all demand of our customers, but thanks to Intel I think we were quite capable to produce as much as possible. And also we’re capable to react fast on changes. And, basically with the digital, our digital product, similar to all the product, we’re also capable, if Intel said, “Hey, that product is not available, but I have a slightly different version of the product available.” We were quite capable of redesigning our product quickly in order to then build in the different product. And also there, thanks to Intel, we had a very, very close collaboration of finding out what fits and what doesn’t fit.

Christina Cardoza: I love that longstanding relationship, that you guys have seen these evolutions throughout the last couple of decades and worked together. And, of course, Intel, every year they’re just releasing new capabilities, new features that are helping you guys solve some of these real-world challenges in use cases you’d mentioned earlier.

So I’m wondering, especially the recent advancements that Intel is making, how the new updates or features being added to Intel^® Xeon^® processors, for example, how those play a role in Siemens and helping you guys reach some of the goals and trends and transformations we’ve been talking about?

Rainer Brehm: Absolutely. I mean, first of all, on the embedded side, we are now releveraging this kind of low latency. On the other side, as you said, now the new Xeon family, the 4th generation, what we see is—and I mentioned that number—we are producing every day 10 terabytes of data. And now we need to—and that data isn’t really used, it’s used partly, as I said, maybe on the X-ray of our PCB lines—but I think we can do much, much more leveraging of this data. But this data—no controller which is controlling the process was made, ever, for handling this data, computing this data, storing this data. So, but you see a lot of customers which say, “Well, I don’t want to move all the data in the cloud because it doesn’t make sense. I want to use it on premise. Yeah, I want to use it in the factory.” So, for that we see the trend of micro-industrialized data centers, which are not in a room, but maybe even close on a cabinet, close to the line, to the machine, which can compute an immense amount of data.

So that was a reason why we expanded our portfolio, which was currently more on the PLC side, on the industrial PLC side, now really two kinds of data center–like equipment for that high workload on AI, on digital twin, on simulation. And we see that immense—for machine-vision application is also another workload which consumes a lot of compute power. And for that we came to the conclusion, we will bring out a new portfolio leveraging the 4th Gen Xeon^® Scalable platform. Looking forward to introducing that in the market pretty soon, in the middle of 2023. So, very excited having that new portfolio element, addressing exactly that need we see on the shop floor.

Christina Cardoza: Exciting stuff. I’m excited for when that release comes out to see what else you and your customers come up with, and how these use cases are going to expand and just advance over the next couple of years. It’s been a great conversation, and unfortunately we are running out of time, but before we go, I’m just wondering if you have any final key thoughts or takeaways you want to leave our listeners with today.

Rainer Brehm: First of all, I strongly believe that for a sustainable future you need to electrify, automate, and digitalize. And, therefore, what we do together with Intel really is a significant contribution for our future. So, number one. Number two, I believe the area of automation will expand more and more while we automate workload, which is unpredictable and maybe a “Lot Size One,” very individualized. And, thirdly, we need to make this technology accessible, available for, I wouldn’t say unskilled people, but make it as user-friendly as possible that OT persons can handle this complex technology. And these are for me the main three topics, and I’m very happy for further collaboration and working on this vision together with Intel.

Christina Cardoza: Absolutely. And we’ll be on a lookout for that new portfolio or solution you just mentioned that leverages some of these Intel^® Xeon^® processors, some of the new releases coming out, because I think that’s just going to be so huge for the industry and solving these pain points and trends.

But it’s been a great and insightful conversation, Rainer. Thank you again for joining us, and thank you to our listeners for tuning in. If you liked this episode, please like, subscribe, rate, review, all of the above on your favorite streaming platform. And, until next time, this has been the IoT Chat.

The preceding transcript is provided to ensure accessibility and is intended to accurately capture an informal conversation. The transcript may contain improper uses of trademarked terms and as such should not be used for any other purposes. For more information, please see the Intel^® trademark information.

This transcript was edited by Erin Noble, copy editor.