The Digital Leap: An Engineer's Journey from Shopfloor to Senseye - with Tom Jacques
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.
This episode covers:
Tom's unique career journey: Discover how his path from hands-on agricultural maintenance and shop floor experience led him to become a solution engineer at Siemens, showcasing a fascinating blend of mechanical and digital expertise.
The transformative power of diverse experience: Understand how having a broad perspective, including practical insights into various industries and machine failure modes, is crucial for developing effective, customer-centric solutions.
Insights into the evolution of Predictive Maintenance (PM): Explore how AI and intelligent software are moving PM beyond basic condition monitoring, making it more accessible, accurate, and capable of anticipating failures across all types of assets.
How Siemens' Senseye tackles real-world industrial challenges: Learn about its agnostic solution that quickly baselines machines, digitalizes expert knowledge, and provides tangible insights to optimize operations, reduce downtime, and enhance safety.
A glimpse into the future of industrial maintenance and valuable career advice:
Hear about exciting upcoming trends like simpler wireless sensors and advanced AI integration, along with practical tips on becoming an "expert generalist" and fostering continuous learning in a rapidly evolving field.
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
Welcome to the trend detection podcast powered by Siemens, and now your host Nel Sullivan. Hello and welcome to the latest episode of the Trend detection podcast. Um, I'm really pleased to be joined by a very familiar face on the podcast, Tom. Welcome back. But we're actually in person. You've probably heard us quite a few times talking about many fantastic, um, subjects online, but we're actually in person, so it's a pleasure to do this in person to be here. No, fantastic. Um, and I think towards the end of our last podcast we did, we sort of, or maybe it was after we stopped recording, but we. Came up with this idea cos your career is, has been so interesting and you've seen sort of both sides of the, the table, both sides of the table, both sides of the coin in terms of, you know, working on a shop floor to now working at Siemens and with the um deploying Sensai with our customers, so we sort of wanted to talk about that journey today but I guess maybe the best way to sort of open the conversation is, um, people might be familiar with you already listening, but. Just give us a brief overview of your career so far, where you started from, and that'll just open it up to wherever yeah, with pleasure. Um, so, yeah, hi, my name's Tom, as you know very well, and we've done these things a few times, I think when it comes to my career, I mean, I could start when I was 3 or 4. Sounds bizarre, but that comes from a, a, a history of agricultural maintenance. So my dad's got his own business, um, which is very heavily focused on maintaining machines, maintaining assets in automotive space, anything to do with agriculture. So, you've gotta think, as a 3 year old, that's where I first started to see things like bearings and sprockets and chains, and that's then had a big impact on, on the direction my career's gone in. Now it, it could have continued down that route of maintenance or of, of mechanical processes. But actually I ended up doing an apprenticeship, uh, with Siemens in, in Germany, uh, and that was really focused on automation, so Control and automation, Mechatronics was the degree, and that's taken me down the route of digitalization, what you can do with sensors, what you can do with PLCs and control systems, and now, of course, what we can do with Artificial Intelligence. So, you, you literally just combine those, those two things that I really enjoy doing, um, and that's my career, so I'm very fortunate. And I think you're right to start with that as well, like because I, even I didn't know that, so it started all the way, all the way back, way, way back when. And I guess I mean if we take it from that point, from the point of childhood and having an interest in that to where you sort of started to work on shop floors and stuff, how, how did you sort of, what was the transition to that? How did that? Um, really the transition came from, you remember at school you have to do work placements. So my first work placement was when I was maybe sort of 1516 years old if I remember rightly. And that was doing control and automation, uh, on lots of machinery in a small, um, effectively OEM manufacturer where they made custom projects for different customers, and some of it was, was defense customers, some of it was, you know, other machinery. Uh, one of the famous things they created was a, a small, uh, machine that could put the springs into the barrel of a lock. That was all patented technology, really clever stuff, and, and that's kind of where I first saw, you know, wiring systems, PLCs linked together. And obviously, I came back from this experience bouncing off the walls just thinking this is, this is what I wanna do, this is really, um, it makes me passionate about engineering. It also makes me wanna see where I can take this in my career, so let's, let's, you know, in many ways I've, I've been fortunate that it's continuously gone in this direction, not straight down the middle. Sometimes it's gone off in a, in a control and automation direction with electronics and so on, and equally it's gone off in the hydraulics and lubrication direction, uh, we can, we can go into that. Yeah, um, I guess how have all these sort of, again, we could go into maybe more, more of the specifics, but how to going all these different directions and having experiences in this, in these different areas, how has that sort of shaped what you do today at Siemens? That's a great question. So what that's allowed me to do is see things from lots of different perspectives. If I walk into a customer, I can understand pretty well the kind of problems that they're gonna face, um, as a solutions engineer, pre-sales engineer, that's really what we need to be able to do very well, so. That's been able to understand which assets are critical, what does downtime mean in a certain facility, so you have to have a good understanding of, you know, mechanical degradation, how does, how does a, a, a bearing fail, why would it fail, um, is it due to the, the, the oil that you use, the lubrication system, the pump that supplies that lubrication oil? There's lots and lots of different failure modes in the facility, and then you start to look at entire production lines, so there's loads and loads of different assets. You start to look at, you know, entirely different processes. Famously, we're agnostic at Sensei, so that's everything from a steel mill, a paper mill, um, a food and beverage facility, a pharmaceutical process. So, I've seen lots and lots of different machines. I've seen how they fail, and then crucially, you know how to, how to offer solutions and offer suggestions to, to fix those things, which we have done now very successfully. Yeah, I guess that was the question I was gonna ask. What sort of, maybe you answered it a little bit there about how, you know, what sparked the interest in predictive maintenance. When did you first come across it maybe as a, as a concept in your own sort of day to day? I think, I think predictive maintenance for me really became, uh, my main topic of focus being the head of an Industry 4 development division, um. This is for another German um equipment manufacturer and parts supplier and, and really there what I was able to do was, was lead a team, lead the division of, you know, what are we trying to achieve? We're trying to digitalize hydraulic lubrication systems and that's where I first saw, you know, with things like Tribology and what you can do by, by analyzing a fluid. How that impacts the way the machine's gonna degrade from that point onwards, so it's not in a failing state at the moment, but it's going to get into a failing state pretty quickly, so you need to be able to anticipate, um, and there's a lot of money to be made as well in, in if, if you can anticipate when that's, when that failure's gonna take place. So my, my sort of focus there, I think was probably the best word to use, is how do we do this? Which sensors should we use, which, which equipment do we need to make that happen, you know, most effectively. And as I've then gone into it, you start to realize that time-based maintenance is a, is a certain element of it. But really with predictive maintenance, you can remove standard time-based intervals and convert it entirely to a, an intelligent software that's just looking at the health of a system, um, you know, exactly, I often say this as, as like an individual patient in a hospital or in a surgery, is effectively how you should look at every machine, so we see this with a, with a filling line, let's say, every piece of equipment there has its own, its own ailments, it's own, it's own history. If you can document that in a piece of software, you've got some clever AI that can monitor it, suddenly you can anticipate pretty accurately as well, when is it gonna fail, how is it gonna fail, what do we need to do to, to anticipate that and to mitigate that downtime or over-maintenance that we might be doing. Um, so it's, it's a real, it's a real fascination from my perspective to see how you can solve mechanical problems with a piece of, a piece of software. And, and the fact you've seen sort of both sides, does that, that must make it easy when you're speaking to customers today that you're, you know, you're not just, you know, I've only worked in software and I know this and that, you've actually seen what they've seen, experienced what they've experienced, that must help greatly, right? I think it really helps to do, to do with, um, respect from a certain degree because if the customer can understand that I understand their problem, I've surprised a few customers as well in the past, um, because I think typically solutions engineers have to have been through the mill. Ironically, steel mill, paper mill, you name it, but you end up having to have seen lots and lots of different things, and I've been fortunate enough to have seen loads, um, at the age of 30, so, um, and also what I've seen very often as well is, is how intelligent, and I really mean this, the, the standard maintenance engineer is becoming. It's no longer someone who just uses spanners and, and certain mechanical tools to fix things. They know how to diagnose problems, they know how to interact with software and, and maybe do some controls and automation stuff on the side to interrogate a failure code on a PLC or something like that. So, we're starting to see a little bit of a shift in industry, um, which, which aligns well to my skill set. And, and actually that, that, that leads very nicely onto this, this question about may, cos on the face of it, working on the shop floor and working now as a solution engineer at Siemens are completely different things, but by the sound of it, they're getting closer and closer together. They are, they are, they're getting a, they're getting much closer, um, because of how intelligent the software's getting. So for many, many years we've had the ability to do condition monitoring, so condition-based maintenance where you, you can just effectively see in a software if. Temperature equals 80 degrees, um, then we send an alert, and that's that's been possible for many, many years, you know, decades, but now with the, with the, um, the dawn of Artificial intelligence and how intelligent it's now getting, we end up in a situation where, The software can anticipate based on live telemetry, historical, um, you know, data from that machine, looking at the PDF documents, any other um document that we've got, a a manual, uh, a user spec sheet, anything like that, it can now read it and look down through the document, see what you, what oil do you need, what problem is, is there to be fixed, exactly the steps you should take to change a filter or to change a bearing. And that's really fascinating from my side, because. You know, we're gonna, we're gonna see this, this get much, much closer, and we'll end up, I'm sure in about 5 or 10 years where we, we can't even imagine maintenance without having this intelligent software bolted onto the side of it. And I guess to that point, do you, would you say that PDM or predictive maintenance, give it its full title, is, is becoming more accessible now compared to when you, you know, first start to, you know, working on the shop floor, for example, it's more accessible to maintenance engineers and, and organizations as a whole. It is, yeah, it is. That's a great question because, because what you need to try and do is understand how do we build the algorithm, you know, in, in a classic example of, um, monitoring turbines. You know, you've got quite a few PhD Engineers to put together a mathematic model of what that turbine is gonna do next. That asset is worth multi-millions, downtime is effectively unavoidable, you need it to do its job. How do you then translate that across to a smaller asset which has less of an impact but equally still a big impact? Now Sense I does that with, with effectively an agnostic solution, um, building a, a tool that can learn baseline of an asset in 2 to 5 days, normal operation. Now when I first learned that, it made me wanna be part of the team because that's a breakthrough in this sector, uh, and being able to do that on any asset, be that a, a, a critical conveyor or, or a filling line or, or any other asset that you see in the industry. To understand and be able to build a mathematic model, um, obviously there's a lot more to it with, with patented technology in the background, how do we, how do we, you know, understand what the fingerprint is of that asset? And then map that across an entire factory of rotating, non-rotating assets, you know, HVAC on the roof and something else, so utilities are also included. It's amazing, it really is incredible. And how have you, how do you find translating that, you know, cos, you know, you know, maybe someone like me maybe, you know, joined Sensi a few years ago and was like, oh this is very technical anomaly detection and all these buzzwords and stuff, but you, you, you must find a way to, well, I mean that's your job really is to translate that into tangible results. I think in terms of translating it, what we, what we often see is a customer who knows the problem and then they've got sensors and data and all the rest of it to, to try and anticipate what's going wrong. So they say, well, we've got all the data. We've, we've built all this stuff together, we've linked the SCAA into a Historian, but we, we now don't know what to do with that data, we don't know what's next, we don't know how to, how do we get some sort of tangible insight from that? The answer is you need a software that's got the capability to, to do something very useful with that software, and that's what Sensei is. So, this is one of the reasons we're winning big contracts because we're effectively the only product, uh, product on the market that can do that. And it's very hard, um, we could go into the specifics of, of how the software works, but I'll leave James to, to do that, you know, very well, our head of Research and Development. And there's, there's other podcasts that he covers it really, really well, so I won't try and paraphrase it, but effectively if you can monitor all those different machines, it gives the customer like this, this hope I guess that we can now do something with what they've worked so hard to put into place, um. So to come back to what we talked about earlier, um, you know, why does this job make so much, uh, you know, why is it so enjoyable? It's because you, you're working to solve customers' problems. And that feels really good, it feels good to help a customer who, who'd done everything they possibly could, except for that last piece that we then fill in the blanks and we, we sort of complete the puzzle, um. And I, I think when it comes to what do we see most, so when, when a customer really has a problem. You know, is it because they're over maintaining the, the asset? Is it because they're, they keep reacting to downtime, they don't understand why it keeps happening? It might be because they're on a shoestring budget and they don't have the resources that they actually need. So in many ways we're, we're kind of, we're not solving one problem, we're helping them solve a whole host of different things, um, included within that of course is also safety, we've done an entire topic of this. Yeah we um. And it's, it's then a case of are the assets in a, in a sustainable, um, operation, like if something's about to critically fail, is someone's life at risk? So suddenly, you know, you can put a price to downtime, you can put a price to lots of other things, but then, but then there's even more important things to dis to discuss, you know, you could lose a leg if a, if a machine goes wrong. Um, I've seen some of the things that go wrong in hydraulics, for example, um. And then you have to just make sure those assets are healthy, uh, and that's what we're doing. And I, I guess another sort of angle to it, because, of course, there's people who know they've got a problem, they need software or an approach like predicting maintenance to solve that, but you must come across, um, you know, different organizations that maybe take a different view. They've got their own traditional practices, and they're like, you know, why should we change? And, you know, coming from that angle. And I'd imagine your your past experience can again, can help see it from the other side again, and that's a good, that's a, that's a great question, because if you think. You know, we've done everything we can, um. The one thing you don't wanna do is offend anybody and say, well, wait, what you've got here, it's not good enough or whatever, it's like what you've got here, you've done as, as well with that as you can. What we can do now is we can help you get to the next level. Um, that's what Sensi's doing. We're doing that with, you know, big multinational corporations. Equally, we're doing that with smaller companies. We, we've got some good examples now in, in pharmaceutical space that are smaller, but the downtime impact is massive. You know, you lose a batch, you lose 2 million pounds worth of product. If you can start to anticipate when a a compressor is about to break or a cooler or any kind of other system. And you really can see that as well, so if you can start to see in the data, um. You know, you start to get an upward trend or some sort of step change in the data. You and I could look at that with our own eyes and see something within it, but since I can do so much more with the same piece of information, plus what it says in the user spec sheet, plus what it says, you know, in the last case closure that was, that was had. What exactly is, is this going to look like as a failure and, and what do we need to do next to, to solve it? So let's get onto that later on, but where, where does Sens I add value? For me, it's 10% of it is anticipating what's gonna happen next. 90% of it now, especially with co-pilot, is how do we then fix that? So understanding that, you know, you can see the problem sooner, but then being able to solve it much faster is amazing because you can, you can get the spare parts into stock, you can anticipate what maintenance is, is needed to be scheduled, um. And I've maybe far more than answered your question, but it's, well, I mean, I mean just to, uh, you know, you definitely have, I'll just maybe we could draw more on some of the objections to some of that, cos even, you know, even if you present it as like, you know, this is your solution, there's still gonna be. You know, skeptics, let's say, in terms of, you know, you know, how it can be deployed, what sort of return on investment it can provide and that kind of thing. So how do you sort of use your perspective to. No, to shape theirs, and, you know, without, you know, like, to your point, without saying everything you're doing is a waste of time. Yeah, yeah, exactly, it's, well, you, you're exactly right. What you need to do first of all, what we typically try and do is understand what is currently going wrong. Do we already have an understanding of what are the exact costs of downtime? If you lose production for an hour or for a day, what does that actually do to impact, um, your process? More often than not, we talk about it as a maintenance problem, but what it actually is, is a, is a production problem. So you've got the production staff standing around saying, right guys, we need to make sure this machine is capable, maybe a press in a car plant, we need to make sure that that can output panels for, you know, doors, all the rest of it. And what you then have to try and do is make sure you're, you're fulfilling the KPIs of different people across the business. …