Vibration Analysis 101: Unlocking the Power of Predictive Maintenance - with Chris Garrison

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.

This episode covers:

The Evolution of Vibration Analysis: Discover how this crucial predictive maintenance technique has evolved from manual listening with a screwdriver to sophisticated instrumentation, enabling early detection of equipment issues weeks in advance.

Best Practices for Implementation: Understand the importance of careful planning, selecting the right sensors for specific assets, and conducting criticality assessments to avoid common pitfalls like "whack-a-mole" problems and ensure a strong return on investment.

Why Vibration Analysis is a Foundational PDM Technology: Explore its broad applicability across various industrial equipment and environments, making it one of the most comprehensive methods for identifying a wide range of fault modes compared to more niche predictive maintenance technologies.

The Game-Changing Impact of AI and Cloud: Learn how these advanced technologies have revolutionized vibration analysis by enabling rapid data interpretation, providing remote access to expert insights, offering scalable monitoring solutions for any facility size, and ensuring continuous software updates.

Achieving Actionable, Contextualized Insights: Find out why cross-departmental cooperation is vital for successful implementation and how AI-driven platforms like Senseye provide tailored recommendations by understanding a facility's unique operational context, maintenance history, and risk tolerance.

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

They might have 3 or 4 vibration analysts on staff, people that they've developed from, from engineering and for maintenance over time. Um, and send them off to get certifications and, and, uh, maybe they would be in kind of an apprenticeship mode until they can move into a, to a senior analyst role. But long story short, there, it takes several years to develop an an analyst. Um, you can, you know, confidently look at vibration data in its raw forms and, um. And accurately, you know, identify faults. 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. I'm really pleased to be joined today, um, by Chris Garrison, um, who has been on the podcast before. Um, it's been a, a trend recently, um, where we've been discussing how many times, uh, the guest has actually been on the podcast. I don't know if you remember Chris off the top of your head, how many times, cause I, I tend to forget these things, but yeah, 3 or 43 or 4. This may be the 4th time. Yeah, there we go. There we go. And many more to come, I'm sure as well. Um, I'd like to think. But, um, but yeah, so today, today we're talking about, um, vibration analysis, a topic you're very deeply involved in, I guess, I mean, we could talk, go into sort of your background, Chris, and And how you'll talk about this topic with customers as well is going to be of great interest and of course how it relates to predictive maintenance, of course, because that's why we're here, right? So, um, but I guess before we get started, Chris, maybe a quick intro from your side for people who may not have listened to your previous episodes before, just what you do at Siemens and that kind of thing. Yeah, absolutely. Been working with um Siemens and the Sensi product for roughly 3 years in total now, but have been in the. Uh, predictive maintenance and condition monitoring space for, uh, 22 or 23 years now, um, which feels like a lot to say, but, uh, yeah, working with vibration analysis, instrumentation, and of course, um, Um, you know, platforms, uh, such as, as Sensei. So yeah, 2022 years of, uh, And predictive maintenance and condition monitoring background. And I guess where did you first come across this term vibration analysis? Much earlier on in your career, I can only presume. Yeah, yeah, um, in the beginning, it was, uh, lugging around, uh, kind of heavy recording equipment, data collectors, uh, a lot of personnel going to take surveys on US aircraft carriers. They would collect, uh, every machinery, uh, or, excuse me, collect data from every piece of machinery on, on an aircraft carrier, uh, once per quarter. It's called an MCA survey, um, and then they would bring all the, the recorded data back and analyze it and issue a report to, uh, help ensure readiness of the US Navy. And we've come a long way since then and how we gather that data and how fluid and how quickly we can respond and provide those types of details. But yeah, yeah, back in about 2004, I guess now, um, it really started with manual data acquisition and going back and, and digging into that data for, for weeks on end and, and providing manual reports. And then what led you into sort of the world of predictive maintenance as well then, just to bring the two together, I guess early on. Yeah, yeah, um, I think, you know, working with with machinery and and maintenance and then um. Joining an organization way, way back, uh, ABB in the very beginning. Um, You know, just, just the interest of problem solving and You know, providing good insightful information from the very beginning. I think the mission is, is very much the same today. Um, you know, can we, uh, prolong asset life? Can we be more proactive in planning? Um, be less reactive, and, and I think that, um, the goals are much the same and they've evolved over time and gotten much more fluid and efficient. But, um, yeah, yeah, it's, it's always been gratifying work, um, to be able to identify a potential or pending problem, um, head it off and, uh, save everybody some time, headache, money, all of the above. No, exactly, exactly. But I, I guess to, you know, move back towards the, the vibration analysis side, um, so I get this question might be obvious to many people listening out there, so apologies for this. So I'm asking it from my perspective and my understanding as a non-technical person, I guess, but, um, what actually is vibration analysis and how does it work in reality? Sure, sure. So, obviously, every, every rotating piece of equipment, um, Produces, you know, vibration, movement, uh, sound measurements. And uh you know, from, from the shafting of the equipment spinning up and rotating within all of its confining components such as bearings. Um, every moving part within, within an asset, you know, provides some excitation, some movement, a, a signal that can be monitored. Um, maybe we've, you know, touched on the history lessons in past podcasts, but It, it started with, uh, with guys with a screwdriver to their ear, you know, walking up to a machine and, and, uh, and listening and they could tell by the whine or, or, or sound of the machinery if it might have a bearing issue and I think transferring that, that human ear, uh, magnifying it and moving into, to providing some instrumentation to actually. Um, cover more or all of the frequencies within a piece of equipment is really where that, uh, where that started. And that's a, that's an interesting point as well. I mean, how, maybe we go into it deeper a little later, but how do you bring those two sort of close together? So the technology side and the person who, like I said, it's almost instinct, right? It's someone who's known, known these machines for many, many years. And, so how do you bring those together in reality? Yeah, I, I think in the beginning, um, there was maybe some, some cautionary feelings about predictive maintenance, about, um, you know, how effective it could be. I think you had a lot of career, um, engineers and, and maintenance managers out there that knew their equipment inside and out, and if they could hear it, that was good enough. Um, and I think breaking, breaking through that barrier and showing them that with some acoustics equipment with accelerometers and so on. You know, you can, you can learn that a little bit before your human ear, ear can, can pick that up and maybe we, we're 2 weeks out, 3 weeks out from, from what you might later, you know, pick up with your own ear. And I think that, um, Slowly you build that trust and and people realize that this equipment can can obviously detect beyond. Um, but even maybe your best, uh, mechanic out there can hear with a screwdriver or with their, you know, with their ear alone. And yeah, yeah, I think it's a, it's a, it's a process. I think today, hopefully, um, vibration analysis is much more of a known and trusted entity, but maybe not always the case. So I guess what are the sort of, you sort of touched on it then. So what are the sort of misconceptions around vibration analysis? What's the, yeah, what presumptions are made, let's say around, around it. Yeah, I, I think that um. There has to be some care taken about the application, the instrumentation. I think that, you know, today, maybe more than ever, there are so many choices out there. Um, and it's not really a, a, uh, you know, one for one, pick, pick any sensor, put it on any machine, you know, there's some solution in there and there's some considerations that have to take place, and I, and I think that There have been times where um sensors have been maybe just kind of thrown out en masse, put on equipment without too much thought or planning. And, and you know that could produce false positives, false negatives. Uh Old, older engineer that I used to work with, uh, referred to it as whack a mole, um, having problems erupt all over the plant and, and, you know, now this new vibration equipment is, is telling you you've got more problems and, and how much of it is actionable. Uh, I think Those are the cautionary tales. But the good news is with, with proper planning, um, with proper solutioning, matching the right sensors to the right assets, um, you know, we can absolutely avoid all of that and uh You know, I, I guess as with most things, some of these lessons are learned the hard way. No, exactly, exactly. And you, you mentioned one thing, you know, an important part of the puzzle as well as, you know, selecting vibration sensors. So, you know, I mean, how to, how can someone sort of go about choosing the right sensors for their particular use case, the right vibration sensor, I should say. Yeah, yeah, um, just, just preceding our podcast here, we're spending a good hour or two on this very subject and, and, um, you know. There's a lot to decipher and different spec sheets and different things out there. We have a lot of Siemens and Sensei approved partner solutions where we kind of vet them and match them to different asset classes. So there are certainly. Folks on, on our team that, that can delve in and, and help with that process. Um, we also, you know, try to provide some, some straightforward material to that end to, to help you make some decisions about sensors and instrumentation, but, um, Yeah, I think that uh it's something everybody who is in the position of, of, you know, choosing censors should, should educate themselves a bit and, and certainly I hope we can, we can help with that. Uh, where possible, And then I guess, yeah, so again, selection, the first thing, but you, you mentioned also about sort of the rollout process and, you know, whack a mole. Well, I do love that. I'm gonna take that one away, I think we'll be it's a future podcast, but, um, but that sort of strategy. So what, you know, you know, I guess why, why do a lot of organizations approach it that way, just roll it out as far and wide as they can and what can they, what should they be thinking about instead? Yeah, uh, great question. So, um, ROI I think is, is top of mind, right? Maintenance budgets can, can be some of the first under threat, um. If you're going to have a Capital expense or an operational, you know, opex expense for, for these sensors, um, where are you're going to get the greatest payback, where you're going to get the greatest justification, um, so looking at things like machine criticality. Um, you know, what assets within your facility are going to stop the show, if you will. Um, you know, that, that might be one of the best places to start to, to get that return, um, and, and then kind of moving outward from there. So, We're starting a brand new discussion with the, with the facility. We might say, you know, um, have you done uh criticality assessment just yet? Um, you know, surely maintenance and operations and everyone knows what's most critical in there, um, you know, to their operations and what can shut down the plant. But, uh, you know, seeing if maybe they formalize that process and uh we can get a look at what's most critical and what challenges might exist around those assets. And if I guess if we talk about vibration analysis again, and industrial equipment, so what, why is it so important? You know, because there's lots of different types of analysis that could be a lot done and lots of different data types. So why, why does vibration stand out with certain use cases? Yeah, there's lots of great PDM technologies, um, vibration just seems to stand the test of time of being the most encompassing. If you have to choose one technology that can identify, you know, the most fault modes within a piece of equipment, it's generally vibration. If you have hyper low speed uh equipment, um, or, or, or different types of assets, you know, you know, you might be able to use thermography, um, Uh, you know, Oil analysis for some of these gearboxes and things, but most of the other predictive maintenance technologies are rather niche, if you will, you know, they're targeted to specific assets or specific. Um, problem statements out there, ultra low speed equipment and so on. Vibration just kind of seems to be the overall. I don't want to say one size fits all, but it's the most encompassing of the technologies. Yeah, and, and comes in maybe the right words in the, in the sense, I mean, does it cover certain machine types or certain industries, but I guess what you're saying, Chris, is no, it's, it's very sort of broad can be broadly applied. Yeah, yeah, um, there's a lot of commonalities in, in equipment. Certainly there are considerations from, um, different verticals, whether it's food and beverage, oil and gas. Um, auto manufacturing. But in the end, we've got, you know, typically motors, um, driving different pieces of equipment. And um the considerations really are are in how variable that equipment might be in, in its operational environment. Does it change speeds a lot? Um, does it run at ultra low speeds, um. Uh, is the environment, you know. Particularly Um, harsh, you know, are there really high temperatures, really low temperatures, and so on. So, yeah, I think you're, you're, you're considering all these things. We're putting, um, calibrated. Sensors and things out into these industrial environments and we've got to be able to stand up to to the elements, so that's one consideration. And do you, I don't know if it's the right way to phrase, but do you, do you often, do you get sort of questioned about why vibration analysis versus another type of analysis? I know it's always, you know, a discussion with the customer and finding out what's best, and obviously there's a reason why they work with vendors like Siemens to provide that sort of advice. But is there any sort of question marks? Yeah, yeah, I think, um. Especially today, uh, you know, you've got these, these kind of inherent data points that are available from the OEM, um, maybe you have, uh, some kind of vibration data already provided as part of the package from the motor, you know, uh, manufacturer. Maybe you have, um, current and temperature and some of these other things. So what we really do is, is look at, um, failure mode or fault mode coverage, um. If we've got a small, highly efficient servo motor, um, maybe criticality or, or replacement cost is rather low, and we've got, um, you know, current and temperature, and we can calculate torque and, and Things of that nature, you know, maybe you don't need vibration, right? That there is a, you know, an expense associated with instrumenting. You don't have to put a vibration sensor on everything, you know, particularly in our world, um. Thankfully, for, for Sensei, we can work with the data that is available um from those motor manufacturers or from those equipment manufacturers in general. But, um, as you move on to, to larger equipment or as you look to be more encompassing with, with fault mode coverage, um, It's not, uh, it's not really a stretch to where you start to look at vibration analysis to, to fill in those gaps, uh, to cover more failure modes. And when, when implementing sort of either vibration sensors or vibration analysis, is there, what are some, some of the sort of challenges or things organizations sort of need to consider before, you know, sort of going ahead with that? Oh, loaded question, very loaded question. Um, I don't try to. No, no, I, I think, um. I think cooperation across different um. Different departments and different stakeholders within an organization is very important, um, having IT and those involved with data security involved from the very beginning, um, understanding and assessing, you know, whether data can be moved to the cloud fluidly, um, or whether it needs to be stored on the edge and, and, and maybe moved to different ways. So, so data architecture, in short. Um, and then of course, uh, you know, machinery, um, there is, uh. Data interpretation and analysis, right? Who, who's going to take action on, on the results, you know, they're um put forth by a platform like Sensei. Um, How, how might you feed information back into the platform, um. To kind of close that loop, right? Um, let's say a fault has been identified, maintenance has gone out and they've corrected it, um, you know, closing that loop might be, uh, having the work order. Information fed back into a platform like Sensei where you kind of close the loop and say, yeah, that that problem was confirmed, we, we changed the bearing or we swapped the motor. Um And we're off and running. So maybe a long winded answer there, but um. Going back to your original question, having maintenance, plant management, having everyone on board, I think from the beginning is, is very important. Um, Because it really takes that kind of cross-department cooperation, I think, to, to, to be successful. Yeah, no, absolutely, absolutely, it's, it's something we often, we often talk about on the, on this podcast as well, the importance of alignment. Don't, don't try and, you know, start these projects or continue these projects without alignment with other stakeholders. Um, but maybe you could sort of talk us through sort of a typical implementation of vibration analysis, Chris. Um, you know, what, if there is a typical, and I know, no one, no project is the same, right, but maybe there's an average you could, you could take. Um, I really think it starts with an asset list, you know. What, what equipment is in scope, what equipment is already being considered, um, is that asset list really, you know, encompassing all of the equipment? That, that could be monitored or should be monitored. Um, again, back to criticality, but what's most important to the facility and the operations. Um, you know, I want to clarify, I guess that vibration analysis and, and predictive maintenance in general, certainly doesn't have to just be applied to your most critical assets. I think we talked about it as a starting point. Um. And when you're doing things like putting on vibration sensors on the equipment, you know, you do have some expenditures there to consider. Um, But even, even the, the least critical type of equipment, you know, is a potential candidate for predictive maintenance. I think you're, you're just at that point thinking about spend, right? Do you want to, um, add additional instrumentation? Um, what can be done with the data environment you have today, I think we talked a lot about that, um, going back to You know, whatever data might already be there from the OEM that's being fed into a DCS system and so on. Um, But yeah, you can do a lot with the data you currently have and then go to vibration from there. So back to the asset list, criticality. Um, Return on investment, maybe, you know, cost versus benefit. For vibration and then digging into the characteristics of that equipment. Um, how does it operate? Does it change, um, during its operational characteristics frequently? Um, does it ramp up or ramp down in speed? Um, you know, again, maybe kind of repetitive, but, um, the environment, how harsh is it? Um, is it, uh, maybe an Ax hazard rated environment? Do we need to have special equipment that, uh, is explosion proof, um, or spark proof? All those things, so, so, so much there. Yeah, that is a lot of detail and that's why there's scope workshops to, to go through all this in, in great detail, right? Absolutely. And in I mean you talked about right at the beginning of the podcast you talked about, you know, the the person walking around, you know, getting a feel for the, for the machine and things like that and of course we're in an age of AI or cloud as well, but AI and cloud. But how, how have those two, technologies or elements, how have they sort of transformed vibration analysis? Oh, yeah, some fantastic changes there, right? …

Full transcript (WebVTT)