Digital Drivetrains & Predictive Maintenance: Turning Motion Data into Action - with Louis Mahlau

Trend Detection Podcast ·

Show notes

Welcome to the Trend Detection podcast, brought to you by Senseye Predictive Maintenance – which gives you visibility and insights into all your assets, from single machines to full plants to help you reduce downtime, increase knowledge sharing and accelerate digital transformation across your organization.

In this episode, we’re joined by Louis Mahlau, Product & Portfolio Manager - IoT & Analytics at Siemens, who explores how digital drivetrains are transforming the way industrial assets are monitored and maintained and how combining IoT, AI, and domain expertise is unlocking a new generation of predictive maintenance.

What a digital drivetrain is and why it underpins so much of modern industrial operations

How predictive maintenance shifts organizations from reactive and preventive approaches to truly predictive insights

How sensors, connectivity, cloud computing, and digital twins come together to turn raw machine data into actionable intelligence

A real-world example of how connecting a single motor enabled early detection of issues before production downtime occurred Where the true value lies beyond technology — including data quality, scalability, and change management

Why many pilots fail to scale, and what successful organizations do differently from the start

How industrial AI and copilots are making complex machine data easier to understand and act on

What the future looks like — from prescriptive maintenance to autonomous, self-optimizing systems

You can find out more about how Senseye Predictive Maintenance can reduce unplanned downtime and contribute towards improved sustainability within your manufacturing plants, by visiting: www.siemens.com/senseye-predictive-maintenance

Transcript

Yeah, I mean, we are looking at normally for our customers into certain areas, which is around reliability and as you mentioned, cost in terms of preventing the downtime, downtime cost, and there's a huge potential to address those two main drivers or KPIs. Welcome to the trend detection podcast powered by Siemens. Hello, my name is Pate Perina and I'm the host of today's episode of the Trend detection podcast. I have here with me a special guest, Louis Marlow, who is the product and portfolio manager for IoT and Analytics within the digital drive train at Siemens, and he's also leading the cloud portfolio there. Hello and welcome, Lewis. Hi Matte, thanks for having me. Thanks for coming. To kick things off, could you please explain to our listeners what it is that you do as product and portfolio manager and how you came about it? Yeah, sure. So I'm within the digital industry motion Control in the digital drivechain organization leading the cloud portfolio, and we are looking at digitalization solutions around the drivetrain technology within the field of IoT and analytics, as we call it. So this is kind of the value stream, and the product and portfolio manager role, we often refer to as the entrepreneur of, of the kind of product, solution. Driving the success in the market. And yeah, I mean, I'm within theigchain team and family since the beginning, so roughly around 3.5, 4 years ago, the organization was formed to centralize and bring the power of our digital solutions in different segments together and really driving the future in terms of digitalization technology around a drivetrain. Together forward and yeah, since then it happens a lot when I think about it, 3, 3.5 years ago, how we shaped the portfolio, how we incorporated new, new trend, technology like AI and very, very glad and, and honored to, to be part of this great team. Bryland, could you shed more light on how you became a portfolio manager? What, what's your background? What, what did you study? Yeah, sure. I mean, of course, you're not leaving university and saying, yeah, I want to lead a cloud portfolio of IoT Analytics in in the digital driver organization, even though that back then it didn't even exist. Therefore, I mean, I started after my engineering degree, so I studied mechanical engineering at the Technology University in Munich and Siemens with different areas. Before, maybe a bit to my background, I, I mean, in the mechanical engineering studies, you, you do math, physics, you really design gearboxes and really understand the machining part of it and really. The physical world around it is, is straightforward, being explained, and you, you're driving that and digging deeper into it. And in the end of my studies, I focused then more on what then comes up with industry 4.0. So, um, industry fear punk null and the digitalization was there. The, the new trend which is coming up and So I had a chance to be part of a project which then came also to my master thesis to incorporate digital twins in the area of medium sized companies, family run companies, the hidden champions of Germany, so to say, and that was a very, very interesting project where at first, the time that they're coming really from this physical understanding machine side of thing and saw the digitalization in action. By really understanding how this change also the work, yeah, by introducing back then digital twins were were kind of a new concept coming from more the aerospace engineering side of things, really bringing this new technology in and having really an impact also how people work, yeah, so the control system engineering department I was in. Needed to link them to other departments more and this impact also shaped me to, to say, OK, it's not only interesting to design certain parts of a machine and optimize performance, but it's really also interesting to be on the software side, on the digital side of things, and, and connecting this somehow, yeah. So the physical and, and the digital world coming together. Yeah, as I said, back then it was all around industry punk. Then later on, when I joined Siemens, that was also my, my drive, uh, within the different roles I, I could take on. So, I went through a management program. I had the chance to work then in mechanical engineering department in the low voltage motor area. So drive technology was very close to my heart at the beginning of my career at Siemens. Then I was within the digital manufacturing, so driving digitalization. Within our own factories and had also the chance to be part of a market push of digitalization in the US so I spent some time in the States and looking at the North American market. Which is, by the way, very differently functioning when you, when you are looking at Europe or Asia, and we see it there actually right now that it is completely working different also for Siemens and yeah, then coming back, worked then in the product and portfolio management area for some time and then joined the digital Drive organization. So overall ari right through, so to say, coming from physics and ending up in, in the digitalization area. Wow, cool. That's pretty diverse and extensive experience that you have. So would you help our listeners understand what exactly a digital drivetrain entails and how it connects with predictive maintenance? So what's the fundamental idea behind bringing these two powerful concepts together? Yeah, so I think digital drivetrain. As the name already says, digitalization of the drive technology, first of all, it's probably good to understand what is actually a drive train or what we understand it. So most of the time consists of a motor, kind of the muscle frequency converter, which controls the speed. And a certain application, so it could be continuous motion pump fan, compressor, but also dynamic motions around a press. So there are many different applications which relying on drive trench in general. So we refer to them as kind of the muscle and brain of motion in the industry. So everything which moves somehow. In the discrete or process industry, you have a drivetrain involved and this already leads to the question, OK, so if drivetrains really bringing this, this kind of foundation of the industry, it makes also sense in terms of control the entire what we call drivetrain value chain, and that's what we do with our digitalization approach coming really from The early stages drive train design, dimensioning, simulating the drive train to be accurate in motion, to be fast in motion, that is one part of the side of things we are looking at. And then also being what we call drive chain health area, so connecting the drive trains which are in the field and optimizing them to be better in motion, to be accurate in motion in terms of also. Capturing the data and understanding what is actually happening with the physical drive train components in the field. So this is the, what we say digital drive train value chain. We want to kind of bring value along these phases. In each and every part of it. And when we're looking now at the predictive maintenance use case and concept, we are then at the driveframe health cluster, as I mentioned, connecting the drives and optimizing the drives in the field. All right. So, we touched upon it a bit, but what are the core technologies that make the digital drivetrains and predictive maintenance possible? So, are we talking about sensors, IOT, AI, cloud computing? How do these pieces fit together to create a a cohesive system? Yeah. So, I mean, first of all, um, digital drive train technology in the context of the predictive maintenance is really driving or want to come to the predictive part, right? So, just to understand before that, I mean, there's the different phases, probably also often discussed here in this podcast. So, coming from reactive to prevent. to predictive and predictive is then where we say, OK, we identify problems before they cause any damage, any downtime in the operation. And as I mentioned before, driveres then play a lot, a lot of industries is a really key role in driving the industry and therefore that is the target. Yeah, so not when something is already damaged and you repair it. Maybe sometimes you're in the field of preventive where you say, OK, I do that frequently. And I deliver at least transparency with the data of the digital drivetrain in this field, but we want to get to the predictive part. That's been said, you already mentioned a couple of, of, yeah, components, I would say, which need to come together. And let me make maybe one example. So we, we have a motor in one of our own factories which is driving a F application. So it's a motor direct online driving. The fan and the fan is part of a painting system, and this painting system is overall the last step of a process in the factory, and in the past this fan made trouble, it stopped working at a certain time. As suddenly, and the whole painting system stopped and the production stopped. So this is the part where you have downtime. So how can you prevent this now? So first of all, you need to connect the drive. In this case, you need to connect this motor with the fan to understand, OK, what and how is the, the, the, the motor behaving or at at first understanding, getting transparency, the right data out of it. So we have a smart sensor connection motor IoT. Which you plug and play on the ripping cool on, on the motor, so that is a key, that's a very important part of it. We want to make it easy for our customers, so to mount it on the outside, you have then vibration, magnetic flux, temperature, so different parameters which this sensor measures, and this sensor is then sent it to the cloud directly. So this is one part of it where you connect. Now the motor. Now comes the second part where, as you mentioned, cloud computing. We have a digital twin model of this motor which the customer in this case in our factory is kind of configuring, so giving mechanical and electrical parameters of the motor based on the rating plate information into it. Then the digital twin model is created and with this. Parameters you measure from the smart sensor, the multi-sensor inputs you have like vibration, like magnetic field, we are calculating all the relevant KPIs. And now comes the trending part which you mentioned before to have a complete system and to be able to do predictive maintenance at the end. You need then to understand how this motor is behaving over time. In this case with the fan. So the fan has certain particles which from the painting comes, which is kind of the weight of the blades are different. You have a misalignment and you have a different pattern in the vibration signals. So we do all this analysis and all this intelligence in the cloud to analyze and then say in this case for our factory here, look, there is a problem. With your motor or with the misalignment, so please check it. And that is happening before then really the motor breaks down and that is a simple example where you can see already a lot of pieces need to come together from the connectivity in this case with our smart sensors, so much connectivity of bringing the data into the cloud, bringing this. Data into context analyzing it and then being able to feed our models to predict a failure and that is kind of the chain which we are offering to our customers and in this case with our in our own factory we could shift kind of the phases from the preventive maintenance approach to a predictive part. So the, at the end, the maintenance team doesn't need to do the weekly checks they have scheduled before. Now they look into. What we say digital drive train in our case, in this case a drive and analyzer application and sees if there is a misalignment, a bearing problem, or any vibration patterns which are kind of critical. And yeah, so it changed also at the end. The way how the maintenance team now is working. All right, so could you maybe sum it up for us to understand what the general benefits are? I, I see that it's all about preventing bad things before they happen, but I guess there's, there's a bit more to it. Yeah, I mean, we are looking at normally for our customers. Into certain areas which is around reliability and as you mentioned, cost in terms of preventing the downtime, downtime cost and there is a huge potential to to address those two main drivers or KPIs. The other part of the challenge we see is this topic of sustainability, so half of the electrical wire consumption is kind of caused or connected with drive trains, so there's a huge potential to reduce that as well. And with our data we get from the drive and also with our smart digital motor model, we can calculate energy consumptions and understand where to start in order to reduce the consumption in your fleet. So especially when we look at customers in the water and wastewater industry, we have a great example where we can see. We are reducing cost of repairing and maintenance immediately by 10 to 15%. We are extending the lifetime of the asset up to 15% and as I said, also the energy consumption is reduced dramatically and overall that is the that is the goal, yeah, so to reduce the risk and the OPEC for our customers and yeah, so there are many different And so to say, industries and examples where we see we have an immediate impact and return on investment. That sounds pretty impressive, but adopting new technologies may come with its hurdles. What would be the biggest challenges companies face when implementing digital drive trains and predictive maintenance? Would you have any tips or suggestions for overcoming them? Yeah, I, I think what I see out of our customer projects, and of course, not, not every project has this return on invest to be honest, and, and sometimes they're stuck in, in certain phases and, and you don't, can leverage this value and What, what I see out of many different customer projects and, and also pilot projects, first of all, there are probably 3 areas. The first is that the customer underestimates or the partner underestimates all this chain we have looked at, so the complexity of turning really data into value and There are many different stages where a customer can be in there, right? So, you have sometimes customers say, oh I heard about this AI I want to use it as well, let's go. And you say, OK, where is your, your data of the drive or the motors or the complete drive train. And so this is probably the base ground to say, OK, first you need to connect your assets in order to get data or the right data. So sometimes you have data but not the right one, missing, for instance, vibration inputs to, to really detect problems in the overall mechanical setup. But, so what I want to say is, you see, That from connecting the assets, getting the data, interpreting the data, and then having insights which are then used accordingly, changing also the way how how maintenance is done, that is a long process, a step, and sometimes Uh, customer underestimates. So they say, OK, um, we had one example of a customer which started to come to us and say, OK, your, your analytics software, I don't need, I, I just need the, the data. So give me the connectors for, for my drives or the CMIOT which is, by the way, open MQTT. You can get the data. And I have my own guys uh interpreting the data. The same customer came back half a year later and said, OK, well, Uh, I see you have a point in interpreting data and knowing when a bearing fails is not so easy, so you need to interpret and combined vibration with speed, knowing what the FFT is really saying to you, and also bringing this into context, how the motor is operated. So this is complex and therefore this domain knowledge we have, we incorporate in our software. And I see often in these processes from I capture the data, I have the right data, and bring the insights or I derive insights out of it is sometimes more complex than you think. So that is one core challenge I see or heard so often in the customer projects. Secondly, I think, yeah, I mean, often you start with a pilot and the pilots stuck somewhere because it is not looked at the scalability at an early stage. So of course, I understand you first and want to understand what is the value for you and how is the return on invest. So you said, OK, we have certain area where we want to test your solution to see if it brings value. But you can see in customer successful stories where they already fought at the early stage on two scale, and you see others which are maybe just stuck in this pile of phase because nobody takes on it and really want to drive it. So one example of a very successful story was in the waterwayste industry again. Where they looked at, OK, what are my all my assets, my fleet, what is critical, what is not so critical, which processes I need to do, which kind of system, and with that early on they started on one side and extended it across many multiple sites today of hundreds of different assets. So they started small. But thought at the beginning big, and I think that is one core challenge or hurdle I see that often pilots just stuck because they don't look at the scalability, um, and the overall return on the vest they want to reach. And I think last part of it, I touched it maybe already a bit is, is us, as the people, right? So you can have the best system, the best insights. If you not also take that in terms of changing, and this brings me maybe also back to my story with the master thesis where I could really see digital twin change the way how people work, that's the same thing here, yeah, so if you have the right insights, the people need to take that on and you also need to change the process of how you do maintenance maybe from today on. And that, yeah, is, is an important part of it, so it's also part of skills and change, change management and, and involving the right people into, into these projects. Otherwise it's stuck somewhere. Would you say that a lack of talent nowadays would fill into this, this part as well or into this challenge as well. Yeah, I, I think. That is definitely one part of it, especially when we look at digital drivetrain or the drivetrain technology itself. There are less and less experts out there to understand really what's what's going on with the drivetrain. And also, even you have these experts, they are not everywhere in your different production sites, right? And therefore I think AI is really helping us in a way or combining the knowledge which we have from the industry with AI is one part of it. So that's what we're trying to address also with our industrial co-pilot and, and what we have with our operations co-pilot for Bright and Analyzer, because here we are looking at a solution. Where we incorporate our domain know-how and the, the Industrial-Copilot is aware of the data and analytics we have on the software and bring this then to the customer. So we have a very big change in how we can bring in really specific industry knowledge towards our customers and combining it with our analytics we have there and, Analysis itself is impacted by, by AI, by our industrial co-pilots. So I think that is definitely one part of the hurdles, but also something we want to address now with the, the new technology, which is, and the new possibility which brings AI or what we say industrial AI. Hm. I think we are hearing nowadays a lot of about co-pilots, industrial AI, AI in general. But could you perhaps explain how it actually works in connection with Drivetrain analyzer? So you have the co-pilot, you have Drivetrain analyzer, and how does it bring value to the customer? Yeah, sure. I mean, at the moment, it is an integrated part of our solution. …

Full transcript (WebVTT)