Welcome to the Trend Detection podcast, powered by Senseye, an industry leader in using AI to drive scalable and sustainable asset performance and reliability. This is a new publication designed to help you go away with ideas on how to achieve maintenance efficiencies.
For this 3-part series, I’m joined by James Bond, an individual with over 25 years of experience in maintenance, condition based monitoring and predictive maintenance.
In the second episode of this series, we discuss how James uses a “toolbox” approach to maintain his assets and his advice for building a business case and deploying a predictive maintenance solution. You can listen to episode one here.
Niall Sullivan, Senseye: That's interesting. So, let's dig into that a little bit more. So what actually makes up the toolbox? I like the toolbox analogy, because you use a hammer for this and use a screwdriver for that. So that's quite interesting. What are the different technologies you use and why do you use them, I guess, for that purpose?
James Bond: Each technology that I have, that technology has its sole purpose, say for example, the vibration analyzer, it measures vibration. Okay. What if we're looking for a sound? Okay, well, the sound we're looking at, and plus more higher frequencies than a vibration analyzer will pick up, the ultrasound will pick up an early bearing detection, but it's most commonly used for air leak detection, then commonly used for proper lubrication techniques. And then say for example thermography, well, you can't determine temperature with a vibration analyzer. So there you go.
I can take an infrared camera, go out and do a scan of maybe an electrical box or something like that. There may be a loose connection or something. And then that way, I can be able to go out there and say, hey, here you go, you got a loose connection or something, you might need to take a look at that.
Then oil analysis, of course, you can't look at the oil with vibration or ultrasound or thermography. So your only case is doing oil analysis. Let's see what condition this oil is in, causing it to run this way or let's see what kind of wear particles or see if there's anything going on inside of that gear box to be able to determine the cause of failure.
Niall Sullivan, Senseye: And then with all those different tools, I guess, do you feed them into a predictive maintenance tool and then use the AI machine learning to bring out those insights? So it brings those data feeds together to bring you an overall picture of the asset.
James Bond: Well, when you're using, say for example, I use CBM, of course you already know that, condition based monitoring, I have the capability of being able to pull up a screen, which is actually including everything that we can possible to incorporate everything into that one window. So measuring condition of oil, measuring what kind of vibration's going on. Okay. What kind of torque is going on on this motor and so on and so forth.
Niall Sullivan, Senseye: That's interesting. I wanted to talk about condition monitoring and, I guess, predictive maintenance more specifically. So when did you exactly start using predictive maintenance tools? And then, I guess, maybe summarize some of the benefits you've seen from using those in your day to day job.
James Bond: Okay. Well, back when I started, it was amazing after I got the proper training, more formal training, and all my certifications, all these classes opened my eyes on stuff that I just had tunnel vision to.
So say for example, vibration analysis, whenever I went to classes for this and getting my certifications, what was totally amazing is that I was actually seeing stuff, not really realizing up to two to three years in advance. But it also taught me that I did not have to act on it right away, but just to monitor it and see if the condition progressively got worse.
And that was basically the same way with say oil analysis. Oil analysis, I was able to ... I have a microscope that I can look through things, and then being able to determine the shape, the size, everything else, and be able to determine what kind of wear was going on on this machine. Which may have explained, okay, that's why it's vibrating so bad.
Then ultrasound, like I said, ultrasound was good for the lubrication side of things. But it was also teaching me that we have to be consistent with how we do things. So being able to make sure the ultrasound gun per se was consistently being calibrated through a period of time. If we're not careful about that, our ratings can vary as well.
Niall Sullivan, Senseye: So, I just wanted to pick out a point you said earlier about being able to see signs of failure, I think you said years in advance. I mean, how helpful has that been from a practical level overall?
James Bond: You can detect stuff years in advance, but it's something that you don't necessarily have to act on right away. That's the reason you have to have good data and you have to have trendable data. That way you can extend the maintenance time out a little bit longer.
And then plus it gives a scheduler planner time to plan for these repairs so they can get everything kitted and on the shelf and be well ahead of the P-F curve on that. So you're taking advantage of that. And then whenever the next available downtime is, or when we deem that it needs to go ahead and be addressed, everything's already ready there. And by doing that, you're also reducing wrench time as well.
Niall Sullivan, Senseye: Okay. And wrench time, I guess, means repair time, does it?
James Bond: Yes. Or what you're calling corrective maintenance.
Niall Sullivan, Senseye: Okay. Okay. That's interesting. I quite like that phrase actually. But again, along that line, you said about planning. So how do you approach planning and scheduling alongside a predictive maintenance tool? How does that help? Yeah, I guess, what's the process of using the two? Or using predictive maintenance to help with scheduling and planning, I guess, would be a good question.
James Bond: Okay. It's pretty much all about prioritizing things. What the criticality is of this machine that needs to be addressed and being able to prioritize that. And one that's going to just shut your whole facility down versus, okay, we got something that's just blowing air on a ... have something going on with an air handler, say, and it's just adding comfort to a technician. That's something that can be addressed down the road. But something that's going shut your facility down is something that really needs to be addressed right then. All about prioritizing.
Niall Sullivan, Senseye: Yeah. And it's about, I guess, prioritizing resources in the right area?
James Bond: Yes
Niall Sullivan, Senseye: Actually that's an interesting point as well. So in order to, when using a predictive maintenance tool is, do you need a big team in order to deploy that? But what you talk about there, about the use of resources, I guess, it actually reduces the burden on resources, so you can actually focus more with a smaller team. You don't need a bigger team necessarily. But happy for you to correct me on that.
James Bond: No, you're correct on that. Every technology is very unique in its own ways. So by using that technology, like I said, prioritizing and everything, as long as you have all your ducks in a row, I hope you don't mind all my analogies, but as long as you have all your ducks in a row, everything should go smooth as long as everything is planned properly. And not every little tool that you need, part and everything like that is put into place. So everything should go fairly smooth.
Niall Sullivan, Senseye: Yeah. And I guess on the spare parts side, inventory side, I guess, it allows you to plan a lot better on that side too. So you don't have to be so, again, reactionary to, you're having to bring in parts at the last minute to do a quick fix kind of thing. You can plan. Again, it's all about planning, I guess, in advance.
James Bond: Yeah, that's correct.
Niall Sullivan, Senseye: I also wanted to ask, in terms of the types of assets, because I guess, I think our audience be interested to know the types of assets, from your experience, have you monitored in the past? Yeah, I guess, that's a good starting question. What types of assets do you monitor currently?
James Bond: Well, some days when I'm working, I feel like I am monitoring just about everything I can think of imaginable. Like I said, I just attack the critical ones and they could range from doing pumps, air handlers, just a wide variety of things, so you would be surprised, even conveyor drives.
Niall Sullivan, Senseye: Is there a reason why you wouldn't focus just on the key critical assets in your plant and go wider than that essentially. Because I think that's the temptation, especially with predictive maintenance, is to focus just on the small number of critical assets. But actually the real benefits come from expanding that out. Do you agree with that?
James Bond: Yes. But I work with a facility where there are lots of assets, and each one of these, depending on its criticality, has to be checked. And I have lots and lots of assets to check. And I also go through a criticality matrix, which determines, okay, the safety aspect of it, okay, how much downtime could it cause if this machine goes down. And I don't know, I think I mentioned the safety part of it.
Yeah. A lot of things to factor in. And then whatever that final score is is how we do that. It also varies on, okay, this is a critical asset, but how smooth has it been running over a period of time? Maybe we can increase the interval that way in order to help us utilize our times where in other areas that needs to be addressed.
Niall Sullivan, Senseye: And how many assets do you monitor using predictive maintenance? Maybe not an exact number of the top of your ... That would be testing you, wouldn't it?
James Bond: Okay. I'll put it to you this way. I currently monitor a whole lot more than I can count on my fingers and toes.
Niall Sullivan, Senseye: Interesting. Interesting. Yeah, because I mean, the next thing I was going to ask really was actually around deployment. And obviously the more machines you're monitoring, I mean, how easy is a predictive maintenance tool to deploy from your experience?
James Bond: As long as you have, like I said, as long as you have all your ducks in a row, everything should go fairly smooth. Not all the time, but most of the time, yeah, everything should go fairly smooth. And still has to go through approvals and stuff like that.
Niall Sullivan, Senseye: And I mean, could you detail some of the steps you'd take in order to successfully deploy a PDM tool?
James Bond: Let's use ultrasound for an example. Say you have a facility that's trying to get started up with predictive maintenance. So the first thing you're going to do, you're going to use stuff, so you can gather the low hanging fruit. Say for example ultrasound, the most common use for ultrasound is early detection. Ultrasound, from what I have seen, is not as expensive of some of your other predictive technologies, but you can still get a big bang for your buck with it.
So you go out here and, okay, let's train this individual up and tell them, okay, we want to make a case just for how much compressed air we are losing through our lines. So you go out there and find several issues where there's leaks, how bad the leak was and everything, and build a case on that.
And then once you figure out let's say a dollar amount, and go out there and then make your case, okay, here's what we have found just with this, but we need to do, well, we feel like we need to do more.
Now, why not take this dollar amount that we are detecting that you are losing, and once these corrections are made, look at that dollar amount. Now why not just take this dollar amount, let's reinvest it in something that will help us out just a little bit more.
And it's like just building one on top of the other until before you know it, you have a great maintenance team. You've got great technologies. So basically just building that case.
It's like I said before, when you start showing dollar amounts and how it's benefiting the company, they're willing to spend more money on you because you're showing the initiative, okay, I'm trying to help you out here.
Niall Sullivan, Senseye: And how long does it take to get to that point? Maybe there's not a traditional timeframe. But yeah, just as a rough idea.
James Bond: I guess it would depend on your facility. And then who they have in place to say, yay or nay to what you want to do.
Niall Sullivan, Senseye: Yeah. Again, it's feeding into what we talked about before, about culture as well.
James Bond: Yes
Niall Sullivan, Senseye: The culture in order to do that.
James Bond: You have to get that culture changed before any of this stuff will fall into place.
Niall Sullivan, Senseye: Yeah. Actually yeah, thinking back to my question about how long does it take, because if you've got a culture in place that's ready and open to this, then things can move much quicker than one where you're trying to push through a wall kind of thing. And trying to get things through that you believe are correct or maybe even know are correct, but it's about convincing others in the organization about the value or showing the value let's say.
James Bond: Correct.
Niall Sullivan, Senseye: And I think we've gone over it a little bit, but in terms of building a business case around that, what types of information would you collate or would you recommend collating in order to show management that this is worthwhile continuing? I know you said downtime, I guess, is one area.
James Bond: Yeah, we call it downtime cost avoidance. Is basically the main tool that I used to use. Cost avoidance is a big thing. And especially if you have a facility that just runs right around the clock, you have very little time to put a wrench on anything except for the weekend. And then you have to pay overtime for people to come in and take care of the issue. And there's another thing, you're saving the overtime by avoiding this downtime. And then plus being able to do all this, the time it takes to, say, write that work order or have to do a PM when that PM could probably be stretched out or eliminated.
Niall Sullivan, Senseye: It's interesting about downtime actually, because it is a cost avoider, but it doesn't put money necessarily straight back in your pocket that you can go and spend elsewhere if you see what I mean. So, it's a saving, you don't have the expenditure, but it's not something you have tangible in your hand like, oh, I've just been given this money I can now spend. Do you agree with that viewpoint? Or do you have any thoughts on that side of things?
James Bond: Yes. That was a struggle for me for a little bit, determining that, okay, how much money have I saved you on the downtime that was avoided? So downtime cost would have to be calculated in with that. Like I said, also how much time did they spend entering in stuff they did on a PM, the technician itself? And then also say something did break down, it takes so much time to do an analysis on that. Okay, why did it break and what can we do to correct that in the future? And how can we make it better? It takes time to do all of that.
But by letting them know all of this and being able to increase the longevity of these machines, that decreases that labor needed, which is actually a saving to the company as well.
Niall Sullivan: Yeah. Absolutely. Absolutely. And actually on that note, so from your personal experience, what results have you achieved using predictive maintenance? I mean, even just in general terms.
James Bond: My main goal when I go to work every day is to leave it in a lot better shape than it was when I first entered into it. It may be something small, it may be something grand. But each day just strive to go in and make that one improvement. That way you can work on improving something else on the next day.
Niall Sullivan, Senseye: Yeah, that's good. And I guess, you have to have that attitude when you're working with so many different assets. It might be hard to achieve everything in one day on one asset. But like you say, it is the incremental benefits, I guess, which is key.
James Bond: Well, I've also been considered as a dinosaur, but I'm a dinosaur.
Improving PdM With Maintenance Data
Moving from a reactive to a Predictive Maintenance approach requires a change of mindset and structure within an organization. From the top down, the focus must shift to one of continuity and efficiency through the use of data and technology.
Technology is a core enabler of Predictive Maintenance, whether collecting, analyzing, transferring or responding to machine data to keep machinery running; as such, the role of the IT team has never been more critical.
Download our free white paper to learn:
- The role IT plays in enabling Predictive Maintenance.
- The importance of maintenance data.
- Best practices for data collection and transfer.