Step into the realm of proactive machinery care with the #reliabilitygang as we reveal how condition-based maintenance strategies can revolutionise the way you approach maintenance. Unlock the power of letting your equipment's condition prompt maintenance actions, mitigating the risk of 'infant mortality' in new parts and steering clear of the pitfalls inherent in traditional preventative practices. Our conversation is rich with insight on the unpredictable nature of failure modes—bearings being a prime example—showcasing the necessity of vigilant monitoring to ensure longevity and reliability in your machinery.
Will Crane and Will Ocean talks about the dance of vibration analysis and ultrasound monitoring. We dissect the complexities of machine 'heartbeats,' learning to interpret these vital signs accurately with the different technologies available and how to use them with precision for measuring different parameters . Our discourse isn't complete without delving into the heat of things; thermal imaging stands as a pillar of our discussion, where we unpack the nuances of emissivity values essential for precise temperature assessments. Armed with a multitude of monitoring strategies, from oil sampling to motor current testing, this episode is your guide to a comprehensive approach to fault detection and the maintenance of industrial equipment. Join us for a deep dive into the technologies that maintain the pulse of our machines.
Step into the realm of proactive machinery care with the #reliabilitygang as we reveal how condition-based maintenance strategies can revolutionise the way you approach maintenance. Unlock the power of letting your equipment's condition prompt maintenance actions, mitigating the risk of 'infant mortality' in new parts and steering clear of the pitfalls inherent in traditional preventative practices. Our conversation is rich with insight on the unpredictable nature of failure modes—bearings being a prime example—showcasing the necessity of vigilant monitoring to ensure longevity and reliability in your machinery.
Will Crane and Will Ocean talks about the dance of vibration analysis and ultrasound monitoring. We dissect the complexities of machine 'heartbeats,' learning to interpret these vital signs accurately with the different technologies available and how to use them with precision for measuring different parameters . Our discourse isn't complete without delving into the heat of things; thermal imaging stands as a pillar of our discussion, where we unpack the nuances of emissivity values essential for precise temperature assessments. Armed with a multitude of monitoring strategies, from oil sampling to motor current testing, this episode is your guide to a comprehensive approach to fault detection and the maintenance of industrial equipment. Join us for a deep dive into the technologies that maintain the pulse of our machines.
Hello, welcome back to another episode of the Reliability Gang Podcast. Hope everyone's having a great day. We are back with another podcast. I'm here with my right-hand man, will Crane. How are we keeping buddy Doing very well? Thank you very much.
Speaker 1:We had a great response on the last podcast. We talked about preventative maintenance, different maintenance and practices and with a lot of messages actually kind of on LinkedIn as well, just kind of just you know, telling us about all the different types and how that's kind of helped them kind of try to figure out what route is best for them in their journey. It kind of makes us follow into kind of today's podcast and we're going to follow on from it because I think we had such a good response from it. We're going to go kind of go into a free piece kind of discussion today.
Speaker 1:The first discussion we're going to talk about condition-based maintenance. How does that kind of tie into this? We're going to go into a little bit more depth on that because we were itching to in the last podcast but we had to stay on track. We're going to talk about condition monitoring techniques as well and kind of how that ties in obviously into the condition-based maintenance as well, and then we're going to understand the importance of failure modes and also how they affect different technologies. And why do we choose the technologies Like, why would we use VA? And instead of just having a blanket approach and saying this is what we're going to do, we're going to start to look at the failure modes of a machine and start to understand what techniques are the best to be used as well. So Will I'll let you dive deep into condition-based maintenance a bit.
Speaker 1:I kind of let you have a free run last week and it worked quite well, so explain to us what is condition-based maintenance and the benefits.
Speaker 2:So condition-based maintenance is a maintenance strategy, just like we discussed in the last podcast. We had preventative maintenance and a run-to-fail type maintenance. These are strategies that we decide on an asset that we may use as part of a maintenance plan. All condition-based maintenance is that it is a strategy that we employ on certain assets, depending on their failure modes, where we decide that we're only going to conduct some form of maintenance on that asset when the condition changes and we use different techniques to determine when that condition change happens and that might invoke some sort of response, whether that's a maintenance action or a task to go and investigate.
Speaker 1:What I love about condition-based maintenance is kind of it's the efficient way of doing maintenance. It's almost kind of saying well, if it's not broke, don't fix something, don't try to fix something that's not broke. I quite like that analogy as well, because essentially a lot of the time when we didn't have the technologies that we have now preventative maintenance and we see that now on this on a lot of sites, don't we? They still have these PMs in place to be able to do these things that haven't changed. And how many times have we changed something and there's been nothing wrong with that moment that has been changed. So we're kind of now looking at well, let's leave it, leave it alone until there's a condition change that warrants for it to come out, and whether that be a bearing defect, whether it be housing problems or whether it be a coupling problem and all the rest, of it and we let the technology tell us when.
Speaker 2:That issue is the when to act. And condition-based maintenance really came about off the basis that people were finding that preventative maintenance was causing them poorer reliability, more failures, because when we delve more deeply into the reliability type training and stuff we look at something called infant mortality. So I just want to touch on it because I think it's quite important to understand the benefits of condition-based maintenance and that strategy is that fundamentally, when we put a new piece of equipment in or a new machine and we do some form of repair, there is a period of time that that machine is more likely to fail. It has a high likelihood of failure and we call that the infant mortality period. Once it's out of it then, depending on the failure mode, we can determine that failure is fairly random at that point and this all forms under a complete study that was done by two people called Nolan and Heap. You can go and have a look more into that we might do another podcast on it later.
Speaker 1:I think we'd probably do a separate podcast on it because it's quite in depth, but when we do the Mobius training, it's obviously a piece that kind of allows us to be able to understand well, why do we do condition-based maintenance and what is the difference between that and preventive maintenance. And the idea is, as well with this type of study, is that is understanding. There was another slide within the Mobius and I'm not going to get too deep into it, but there was a slide that kind of had a look at I think it was a hundred bearings, was it, and how and when they failed in terms of run hours, and you could see the difference in terms of so random as some of them were fading after extremely short periods of time.
Speaker 1:Some were going for a very long time, some were in the middle. So this random element of when they were going to fail just kind of proved that you can't just choose a certain run hour to change something, a time interval to say are we going to save all hundred bearings if we change it a hundred hours?
Speaker 2:Yes. So, like when we look at failing modes later on in this episode, you'll understand that some failing modes are random in nature, like bearings. This is where condition-based maintenance can be highly effective, because a time-based or an interval-based maintenance wouldn't be very efficient in doing so. And then that infant mortality piece is there, because when you are changing something just for a run interval and there's nothing necessarily wrong with it what you're actually doing is you're putting that machine in a period of more likelihood of failure every time.
Speaker 1:And it might not seem like it's oh, you've got brand new shiny mower that has been installed and you think, oh, it's brand new, it should be fine, it should be robust and it should be reliable. But the idea is there's still an element of infant mortality when anything that gets manufactured, whether that's down to a mower or it's a product that we buy or electrical piece of equipment as well. That it kind of applies for most things and as well. Another thing as well that I do, you know, condition-based maintenance and a lot of people have to understand this as well is that when we're doing it every single month, the reason why we're doing it is because failure is random, and just because it hasn't failed in the last two years and it's been green doesn't mean we can start to change the frequencies on it now because it's green, Because Thod's law, when you start doing that there's still the random element, that the randomness doesn't change, whether it's been green for the last five years, 10 years or whatever the defect like could actually happen at any point here.
Speaker 1:It's random. This is where we have to start to understand this. So when we have been doing these condition-running programs and some customers, they're like oh well, it's okay, guys, because just don't worry about the greens, because they've been green for ages, just start doing them less. It's like no, guys, you're not getting a point. Failure is random. And Thod's law when you start, I tell you as well, thod's law as well if you start to change them frequencies, there will be a defect. That happens on it. I just guarantee it because that's just the way things go as well, that's it.
Speaker 2:And that's the thing. Some failure modes are very constant. Some get increased the likelihood over time, some decrease over time, some do all sorts of kind of things and that all forms into that research that Nolan and Heap done. But that's the reason why we do condition-based maintenance is because we don't want to be unnecessarily changing an asset when it doesn't need to be, because we actually increase the likelihood of failure for a period of time and we want to make sure that we're doing a condition -based strategy so that we can pick up and trend that information, because we know failure could happen at any point and we want to try and identify it before it happens. Of course, yeah.
Speaker 1:We will touch as well later on in terms of how do we understand what types to use and as well that's really important as well to understand how frequently should we be doing that test. So, whether that's all, sampling or all. But that kind of brings us to our next discussion point the types of condition monitoring that are available, and assessing the condition can be done in so many different ways.
Speaker 1:I think when people say condition monitoring, a lot of people do say, oh, vibration analysis. They kind of default to that because it probably is the most commonly widespread used, widespread, and in my personal opinion it's probably one of the most effective techniques in terms of what you can do. It can't save everything and it can't detect absolutely every defect, but the majority of problems that you'd find on a motor or a piece of running machinery you know it is quite effective. If you put the if you're doing it correctly, that is, and you've set up the parameters and you've got the right readings to be able to detect a lot of different type of failure modes, it can be very effective. So vibration analysis will kind of just go through a quick brief summary of what it is and how it can detect different types of failure modes.
Speaker 2:Okay, so vibration analysis monitoring the vibration on a piece of equipment. I don't want to go in too much detail and I'm holding myself back here because hold back will, because we are in a time limit.
Speaker 2:The most important thing I think to take away from it is that vibration analysis isn't singular. I find that we find the talk to a lot of people and they think of vibration as being binary, one thing moving in one direction. When we monitor the vibration on a piece of equipment, we're monitoring it at multiple different frequencies and different signatures are coming from the bearings, from the if it's a pump, from the impeller, from the motor, from the electrical line frequency lots of different things. We're getting in the vibration data. But from that vibration signature, which is like the heartbeat of the machine, we can determine the different faults that might be going on within it. So there's a variety of different faults that we can detect with vibration analysis, which I'll let you, you can go through.
Speaker 1:So yeah, obviously, depending on on what frequency range is we're testing, this is very important as well because people you know can assume that all vibration analysis, you know you can get these tests that will give overall values, but a lot of the time they'll only measure up to a thousand hertz. So, that being said, then there's a lot big frequency range of fault conditions that you're not going to be able to detect and a lot of the time the thing is, when it comes out of vibration testing, if you haven't done the relevant courses and have the awareness, you can pick up one of these devices and think, oh, I'm testing vibration and it's kind of just more of a uniform test, when it isn't. There's a lot of different fault conditions that you can get from different frequency ranges and now, especially us using our falcons, we can measure up to 20 kilohertz. 20 kilohertz. So we're now starting to detect poor lubrication ranges and really early stage like poor lube, so we can do poor lube, so poor lubrication is a great way of understanding. If you have an analyzer that can measure that.
Speaker 1:High Bearing defects is probably the number one thing. That, being used with different filtering methods and different technologies, each different collector has got their own kind of way of doing it, but essentially that's demodulation or envelope readings as well, which is more of a filter applied to the vibration signal to purely look at the impact-related events. So for Emerson Kit that would be peak view Envelope is what we use within our software in terms of the falcon as well, hd envelope and the stuff that SBM uses. So all of these technologies are very similar and how they kind of get to their result. They are a little bit different in how they do that, but essentially they're looking for bearing defects. I say bearing defects, they're looking for impact-related events, so you can see gear faults with them as well, can't we Exactly? So gear faults, that's impact-related event, housin impacts as well, shaft peak impacts as well. So that could be coupling impacts with a coupling that could be wear, could be kind of these type of different fault conditions as well.
Speaker 2:And we can do running conditions as well, right.
Speaker 1:Running conditions. Yeah. So the way that I like to see it is, running conditions is kind of a lower frequency events and then you've got the more defect-related higher frequency events. So in terms of running condition, you can detect looseness. It's another one. Invalence is very effectively as well.
Speaker 2:So if you've got fans on your site, fans VA can be great for anything VA can be incredible for identification of poor balance conditions as well.
Speaker 1:Misalignment another really good, very common problem. Misalignment Very common, but again I say that the high frequency events can pick up misalignment as well.
Speaker 2:We've seen that as well. It puts the bearing under stress, doesn't it it?
Speaker 1:does and we've seen different things and we've done different case studies on bearing stress and envelope readings as well and raised floor readings with that as well Can pick up resonance issues as well. Some complex vibration.
Speaker 2:Some complex vibration.
Speaker 1:So the identification of these things can be quite good.
Speaker 2:But I think this is why vibration analysis is quite effective, is it does have a very broad range of failure modes that we can identify Bearing gears.
Speaker 1:This but that's why, as well, you do need to be the thing about vibration analysis. You need to be trained quite well to understand how to deduce these problems and where to look for them when to look. So, in terms of having the experience, ok, go on, you could probably pick up a device that would give you overall reading and you'll be able to understand and trend between oh, is it high, is it low? But the understanding of trying to pick out the different frequencies in terms of different ranges can be quite challenging and you do need to have a level of experience and some theory-based knowledge should be able to do so as well.
Speaker 2:What would you say will for vibration analysis as some of the limitations? We've talked about a lot of the great things that we can do, but what about the limitations?
Speaker 1:So limitations we've found this as well. So with vibration analysis you have to remember that you're looking for periodic frequency events that are happening in synchronous ways or non-synchronous ways. If you look at the bearing defects In terms of other failure modes as well, you have to look at, for example. Say, for example, you're analyzing a conveyor, for example, where you get impact from the chain or sprocket and a conveyor depending if it's a dry conveyor or a screw you can get a lot of noise come through the conveyor. So where VA can be really difficult at times is when, if you're measuring a device that is naturally noisy and has these obviously impact-related events, it can drown out the other vibration signature that you've got. So where it can be really difficult sometimes is trying to detect problems or issues and very noisy, loud equipment and trying to filter them. Things out can be quite difficult. We do a lot of gas engines.
Speaker 2:They're quite tricky to monitor with the impact events. What about slow speed stuff? I know vibration can do slow speed, but it also is a bit of a challenge as well as a refresh.
Speaker 1:It can be the introduction now of obviously analyzers that able to sample extremely high rates now it is a lot more effective now, but I mean that's a time balance isn't it, it is a time balance as well, because remember, when you do measure slower speed events, you have to measure for longer periods of time as well.
Speaker 1:So if you've got a very, very slow moving bearings, anything under kind of 10 RPM, it can be very difficult, even though we have done it. But this is where ultrasound can be a little bit more effective, which leads quite nicely into our next technology, doesn't it? Exactly so. Ultrasound, again, is another technology, and ultrasonic ranges are above 20 kilohertz and above.
Speaker 2:So we're moving out of vibration Now we're listening to more type of ultrasonic range, which we can do via airborne or structural or structural concept born.
Speaker 1:So obviously, when Chris on here, he went into this quite well, and so we're going to try to do him some good justice. He's watching us, we know you're watching us, chris. So, yeah, and again, like ultrasound, in terms of airborne kind of sounds, can be extremely effective for picking up air leaks. Yep, humatic air is something that is a huge energy cost in terms of energy production. Humatic air is something that is used across multiple industries to be able to create products and services, but the idea of that, you know, if you don't manage or have a look at how much you're losing per year, it can be a huge cost as well, just as a bit of a background.
Speaker 2:so ultrasound is generated through two methods. So it's generated through turbulence and friction, which is why, also, ultrasound is very highly effective for lubrication, because that is the first kind of stage that we have, that isn't it?
Speaker 1:Yeah, that friction level does happen at extremely high frequencies and again, we talked about our falcons being able to detect that, that poor lube. But remember that is at a later stage. Ultrasound will be able to detect that far a lot earlier, because you're looking at the higher, first initial stages of that initial friction. So the idea of using ultrasound for lubrication is even better in the sense that if you can detect it extremely early and you can detect that slight bit of friction, you can make an impact a lot quicker, can't you Exactly? And the idea of that is if you can lubricate before there is friction, as soon as the friction starts to occur, you are then essentially at very early stages damaging the bearing. Yeah, do you know what I mean? It might be at a very slow rate initially and a small rate, but the idea is, if you're looking for premium, top class, you know reliability. The idea is you want to lubricate that before we even get to the friction state as well.
Speaker 2:So with ultrasound we've got kind of air leaks, very effective for lubrication. It's quite a good tool for slow moving as well. We've touched on with the vibration, being that there's a lot of time involved in doing that. But with ultrasound, because the bearing is so slow, why is ultrasound so effective?
Speaker 1:in that slow speed kind of range it's so effective because when you get a difference in terms of friction level impacts, right, when you're listening to an ultrasound signal, because it is such a high frequency, the noise floor compared to where that defect happens is quite significant. So when we're looking at ultrasound, when we're looking at time kind of time analysis with the ultrasound signals especially using some of the UE stuff that kind of works. You know, chris has sent me some case studies in terms of looking at that, just recording the time wave and looking at the difference between impact related events and noise floor. It's quite definitive and generally that's the way that we now approach. The VA is looking at time wave form analysis and then using something like kurtosis to give us the alarm that there is a difference or crest factor, the difference between the peak level noise or the impact of what's happening and the normal noise.
Speaker 2:Because obviously on a slow speed, bearing the actual energy behind that impact is much less than say something doing 3000rpm.
Speaker 1:Yeah, who. So there's not a lot of energy when you're looking at very slow moving events, and that's always been a bit of a difficult thing to be able to articulate within a vibration signature, a spectra, because when that event happens there's not a lot behind it to be able to make it significant. And as well, sampling rates now have increased massively for the new analyzers, so when that event does happen, it's not missed between the samples, whereas now that can be captured by ultrasound because it's such a high frequency as well and we're measuring that as well. It allows us to be able to, in the time wave form, identify that quite effectively as well. So for really slow moving stuff, yeah, ultrasound has always been extremely powerful and I think that has been for many years, and now vibration is kind of getting to a point now where it can do that as well. But again, all the pins on devices, sample rates and how you set that up to be able to detect it.
Speaker 2:What about limitations for ultrasound? What do you think? I've got a couple that I can think of, so if ultrasound is really good as well.
Speaker 1:and again, a lot of now ultrasound devices are able to give it FFT, so you can now pinpoint frequencies out. But obviously some of the stuff that we used to use in the more basic ultrasound stuff was more of an indication by just listening or gave a decibel level. So trying to pick out the fault condition with it can be quite difficult.
Speaker 1:And again it's looking at higher frequency events, is not looking at kind of imbalance or any of the lower frequency events as well, so you're not going to be able to really tell with ultrasound If it's misaligned yeah exactly when, as vibration allows you to look at the kind of the bigger picture as a whole. But again, the advantages for it as well is for slower moving things and as well you can do early surveys with it, which is very effective as well. Look at poor lubrication and as well you know the new introduction to technology with it, that on track stuff from the camera as well, the camera as well.
Speaker 1:Incredible technology where you know, we showed on the on the UE podcast that you know each and every single little hole within that camera is an actual sensor. So now what we can do is pick up loads of different sources in one picture, and it just allows us to be able to pick up them air leaks 10 times as quick and as well, give a picture to it as well, which is extremely good technology. Again, very good advantage for ultrasound as well.
Speaker 2:So we've got vibration analysis, we've got ultrasound. Should we delve into a little bit on thermography?
Speaker 1:Yeah, so thermal imaging again is an incredible technology and it looks at thermal radiation right from different components.
Speaker 2:It's not actually the temperature measurement, isn't it? I think a lot of people get confused with that.
Speaker 1:No, so obviously it gives a temperature measurement when you're looking at the device, but we're not measuring actual temperature. Well, what we're actually measuring is radiation, so what it is giving off from that particular object, I suppose that's quite important, then, for if you've got an object that is maybe highly reflective, Exactly because, remember, when you look at emissivity and again, we could probably go into another whole podcast. We're trying to keep this brief.
Speaker 2:What is emissivity for guys that are listening?
Speaker 1:So emissivity, in a very short form, is basically how much energy can be emitted from an object. Yeah Right, and there's three different things from emissivity. So you've got what has been given off, what has been absorbed and what's been reflected. So when you've got something that has high emissivity, there's not a lot of absorption and not a lot of reflection. So can you imagine a black body? They say A bit like your black radio.
Speaker 2:Like my black radio is sitting there giving off energy.
Speaker 1:The idea is that a black body is supposed to have an emissivity value of one, and when it is one, that means that everything, from a heat source point of view, is given off. 100% of that has been given off, depending on what is reflective and absorbed with. That as well can then obviously affect the emissivity of an object. So the lower the emissivity of the objects means that there's less heat given off and there might be more heat being reflected, for example. So if you could imagine the system as well, what you've got and this is another one, this is what really kind of used to really baffle me when I was doing my thermal training is a radiator, so a towel radiator that has a reflective surface Okay, so the more reflective the surface, usually the lower the emissivity of the object. Usually a rough surface right has a very good emissivity Okay. So towel radiators the one that are reflective okay, because they're very low in emissivity and highly reflective, they're good at keeping heat in. Okay, but that means if you keep in the heat in, you're not giving it off, so you're not radiating it. All right, so they would only emit if you put a towel on it, because then by conduction, the towel would be heated up by conduction from the towel radiator because it's touching it and then, because the towel has a high emissivity, right, the towel then give off the heat, okay. So, yeah, really, really, when I was doing my thermal I thought what? So all these towel radiators that everyone had in their houses weren't giving any.
Speaker 1:But this is an example. So if you put your thermal camera to try to detect the radiation on it, right, there's not a lot of radiating, radiant energy coming off the towel radiator. So your thermal camera is going to only read right, but you might touch it Exactly. So the thermal camera might read 30 degrees, but then if you touch it you're like, okay, it's not hot, but the idea of it is still hot, it's just retaining its temperatures, not giving it off. So when we look at thermal imaging, we have to understand the emissivity. We have to understand there's a lot of science behind it and radiated energy, different gases, conduction, convection all these different ways that heat can transfer, and it is extremely, extremely light.
Speaker 2:So what can we? What is thermography really good for?
Speaker 1:So it's very good for looking at electrical faults and panels Extremely effective for that as well. But you have to be very careful about depending on the surface structures of different components as well.
Speaker 2:Yeah, because a lot of high voltage has, like perspex covers, exactly. We can't go through that Again. You can't go through glass, as well.
Speaker 1:So remember, a perspex cover will completely eliminate any radiation straight away. So when we do look at these things, like obviously IR windows are specifically used and they use a special kind of material that allows us to be able to see that radiation through that. But again, it's really good as well for just picking up kind of mechanical temperatures on surface temperatures, surface temperatures on bearings as well, motors as well, looking at kind of winding conditions as well. I've seen it.
Speaker 2:We've used it quite regularly and noticing all that, motors particularly hot, and then we find that the fan cowl is completely blocked. Yeah, exactly, and the particular clean, but the thing is VA.
Speaker 1:when you're doing vibration analysis, you know the motor could be completely fine in the vibration point of view, but until you touch it or you get some form of indication as hot, especially if you've got, maybe, remote sensors collecting vibration data.
Speaker 2:It's not necessarily going to indicate, and this is why maybe having a walk around is still quite important.
Speaker 1:It is as well and not even that you know some of these online systems that we've got. So the Falcon Sparrow is able to take temperatures as well, so that would be able to identify both particular failure modes as well. So, again, that's an uncommon failure mode, but it does happen, we've seen it, and the idea is, you know you can't be looking at every single asset on your plan, you know, in terms of temperature.
Speaker 2:That kind of falls quite nicely. I mean, we've covered vibration, we've covered ultrasound and covered thermography. How are we doing on time, because last episode we were very good, are we doing?
Speaker 1:well, yeah, we're doing good. We've got about five minutes mate.
Speaker 2:So on that, then failure mode we talk about. It's an uncommon failure mode, having a fan cowl. Why is understanding the failure modes of the plan particularly important, and how does that enable us to decide what technique we need to use?
Speaker 1:It's very important because the idea sometimes, especially with vibration analysis, it is used as a bit of a blanket approach, so when a lot of people do condition monitoring they'll just throw a VA at it Because you know predominantly 80% of things can be caught with it. But the idea of actually really understanding again the criticality of the plan, then understanding the real critical items and doing for me because on them is understanding well, how can this asset actually fail and is vibration analysis capable of picking them defects up to even identify that failure mode?
Speaker 2:Yeah, because we even are now reviewing another one of our newer customers. We're going through this process with them now, looking at the different failure modes of their assets, and we're now looking at even more condition based condition monitoring strategies than we've had time to talk about today. So we're looking at oil sampling, we're looking at all the motor current testing, we're looking at where particle analysis and they've got some extremely large gearboxes on extruders and we do vibration analysis on them and we do them very effectively. But we don't want to miss out on the other failure modes, like the oil degrading.
Speaker 1:Yeah as well. You mentioned all analysis.
Speaker 2:We haven't mentioned the PNF curve and understanding exactly what it's probably quite important to put that in. Can you put that in the video as well? We can put that in the video as well.
Speaker 1:I can put that in now. But if you have a look at this PNF curve you can understand. Now, looking at down the kind of as it goes down the scale to point of failure, you've got different techniques that identify the different stages along the point of defect. You know, initiation to the point of actual failure. And if you have a look across there, the different technologies allow us to be able to identify these defects a lot sooner than other technologies. So when we do our kind of you know the failure mode analysis, we can look at kind of well what techniques can identify these things more effectively.
Speaker 1:So for gearboxes, for example, oil analysis is a really good early detector of any issues before VA picks it up. And the thing about usually when you do pick up with VA there's damage there. That means that it's too late, it's irreversible. You've done some damage to that, so it's difficult to be able to get that back in the condition as it as where you'd want it to be. So the idea is when we do our failure mode analysis we can look at gearboxes, for example.
Speaker 2:We have to start thinking about you know all analysis because it's a very effective way of identifying you know oil condition, even oil condition as well before, and we can do the worst particle analysis as well, which is even that next level step where we can really understand what actually part of the gearbox is giving off this pit. Whatever's in the oil? Is it from the?
Speaker 1:bearing. Is it iron from a bearing? Is it copper from actual gear? There's a lot of different, you know elements that we can look at, you know, across all analysis itself as well. So yeah, looking at failure modes and looking at kind of how something can fail, we have to have a look at especially the real critical assets and make sure we're not missing anything.
Speaker 2:And to like round it off really with what we talked about yesterday, you're going to have certain things like, for example, belts. You know a belt might fail due to it being degraded over time. We know belts start to fracture and crack over a period of time. Now, naturally, that is actually quite tricky to detect using vibration vibration ultrasound You're not going to detect that belt cracking. So unless you go in and do a actual visual inspection.
Speaker 1:Yeah, you're right.
Speaker 2:Yeah, so this is then where we move away. Although a visual inspection can become under a condition based type of approach when reviewing the condition and visually a lot of the time with belts we actually can through CMS data and information from belt manufacturers, we usually can look at that as being a more of a preventative maintenance task.
Speaker 1:It is and it still needs to be done. But once you look at kind of how things can fail, then you start to look at every element of the function of a fan, for example, that has pulleys and a belt, and you can say, well, how can the belts fail?
Speaker 2:And it's a combination of these kind of activities, because when those belts start to fatigue and we start to look at the data around it, we can sort of say, right, well, based on our data and based on the manufacturer's recommendations we know that this is a challenge for a condition based task we're going to put in a two year preventative maintenance task to check the belts. But what we're going to do is we're going to do a condition based task at a year just to make sure they haven't randomly.
Speaker 1:Exactly. And that's the thing. You kind of choose your condition, you choose your battles and you fight your battles effectively with using condition monitoring, a little bit of plant maintenance, made potential preventative maintenance, and you do a good strategy for the asset, but only once you've understood exactly how it can fail Exactly and you cover yourself. So, guys, I think we've kind of wrapped up quite quickly. We've got to the half an hour mark. So you know, we try to keep quite strict on these times because we're doing them weekly. But thank you all for tuning in Again what we're going to probably do now. I want to put some more polls out. So I'm going to put a poll out now. We've got some good ideas and we're going to put four options out. Guys, pick what you guys think we should talk about next. So, guys, have a great week, take care, peace.