Move on down the reliability road: Transitioning from a predictive mindset to a reliability-focused approach
Editors’ note: Plant Services met veteran millwright and newly minted predictive maintenance technician Michael Macsisak several years ago, and Mike has checked in with the editors as he began work in mining and then again late last year to look ahead at the future of our industry.
In this new interview with Editor-in-Chief Thomas Wilk, Macsisak explains how to move from a predictive mindset to a more reliability-focused approach, along with some observations on how the sport hunting and reliability mindsets are surprisingly similar.
PS: What were your first steps down the road to predictive maintenance?
MM: In the beginning, I needed someone with the experience to show me and train me. To learn you have to go with someone and shadow them, watch what they’re doing and learn from that. After you shadow them and figure out what you’re doing, then you take control of the PdM equipment and start gathering your own data.
It’s good to get a win because when you find your first problem, that will inspire you that predictive maintenance works, and then you’ll get more into it – you’ll learn all the other techniques, and then you’ll learn all the other tools.
PS: At what point with PdM did you not only think, “Yeah, this works,” but that earning certain certifications was going to be the next step on the journey?
MM: With any company, they need to see certifications before they take you on. At the plant I worked at a few years ago, what we did in the beginning was that we brought infrared and ultrasound into our plant, and we had about 15 people get certified.
Basically it’s a week of studying, and learning all the terminology. What you’re doing is great but when you have to explain it to people, you need to know all the terminology involved with it so people can understand you. For example, I could take an IR picture with you, and you could look at it, and if I don’t tell you what we’re looking at or show you how it all works, you just see an infrared picture. With ultrasound you have to learn from training the right angles, how to set up the ultrasonic gun properly and then know what you’re listening to. And with vibration, it’s basically first steps first: You have to learn how to collect vibration data, because if you don’t know how to collect those data, nothing will work; if you’re not collecting good data, there’s no point to even doing it because the data will give you all false readings and you can’t trend anything.
The whole point of using all the tools is to have good reports, so you could trend data on a monthly basis, and you could watch as the asset either proceeds to get worse or holds its own. If you trend data every three months, for example, there’s really no point – there’s too many things that go wrong in a couple months, and if you’re not on top of it, and then you’re behind, and then while you’re catching up you may or may not get it before it fails.
PS: What are the most useful PdM tools in your experience? And are those also your favorites to use?
MM: I use all of them – vibration, IR, ultrasound, and oil analysis – but my most favorite one is from UE Systems. With an Ultraprobe 15,000 with a dual-channel, besides hearing, I can watch a waveform, and I can trend the data. I can see all the impacts, so I can catch problems on the spot. With vibration, I can see the problem if I have the vibration sensor set up right, but I still have to go load the data in the computer for analysis.
With the Ultraprobe, I’m watching the data come in as I’m listening, so even though I’m collecting it, I see it faster than I can with anything else; with vibration it takes a while until you understand what every line means. Basically, I can find a fault with the Ultraprobe faster than vibration; when I use vibration and ultrasound together, it’s all-win, no-lose.
PS: We’ve got the context set for what you’ve been doing with PdM, so let’s talk about the reliability side. What were your first steps toward reliability-focused maintenance?
MM: The reliability approach is different from PdM because it goes past finding issues before the asset fails, into doing more work after it fails, exploring why did it fail and what could I do so it never fails again.
When I first started doing this, I started finding assets going bad and then I started to watch the assets actually go bad after six months. I said to myself, “So this is great, I’m finding it going before it goes, but what can I do about it so it doesn’t fail again?” I started looking at assets in the mixer-cooker area of my plant, in the high washdown area, where the data showed we would lose something in six months.
When I was looking at some bearings, I brought engineers in from the bearing vendor, and I told them what I needed and what I wanted to do. There’s a lot of bearing companies out there, but certain companies have certain bearings that do certain things, like high washdown, or dirt, or dust. I’m not an expert on these, so I’ve learned to bring in engineers from all the companies – they show you what you’ve got, and then you can make a better decision. They’re more than glad to help you, and they showed me how to look for issues I didn’t even think about, but that’s what they do for a living.
In all honesty, those mixer-cooker bearings that I worked on are still running three years later, and they used to last no more than seven months. I’ve been gone from my previous job three years, and I’m told that the bearings look today like they were put on last week. The vendor engineers showed everybody how to grease them and what to do. The bearings are running in a high washdown, high heat area, and they’re running like there’s nothing wrong. In fact, for the silent chain I put in, the Lambda Chain, they tell me the gears are not even scratched yet on the sprockets. You can’t get any better than that – that was a win, that was a big win.
Another example is gearboxes: if you’re looking at bearings, you’re looking in the gearboxes. Are they the right gears? Is it the right gearbox? Are we using the right oil? Are we lubing it right? And all the details that everybody takes for granted really matter, because if you put the wrong oil or grease in the machine then it’s going to fail, and there’s nothing you’re going to do to prevent a failure because you’ve already killed it before it began.
There’s also people who say that oil is oil and grease is grease. And I watch how that works, and it doesn’t ever work that way. When someone says to use a specific oil for a specific reason, and not “just because”, then follow them and don’t be scared to call out to the people that sold you the gearbox or the company itself for help. I do that, and they answer immediately – they’ll make sure you got it right and are doing it right, because it doesn’t make them look good if their product fails.
I’ve seen over-greasing kill assets too, where someone just greased it, and greased it, and greased it, and popped the seals off. And you knew there was a bearing there but you couldn’t even see it. Now, when that happens, all the dirt gets into the grease and the grease will suck it back into the bearing. And once it gets into the bearing, you’re on your way to failure.
PS: You mentioned there was one big difference was that with PdM, you’ve detected a likely fault before it happens. And with reliability, you take the situation apart, understand where the fault is, and fix it so that fault doesn’t occur again. What other differences have you noticed between predictive maintenance and reliability?
MM: Well, some people will take the reliability part of it halfway and stop, or they’ll just apply it for a couple months and then stop. When you go through a reliability journey, there’s no point in getting there and then quitting. You’ve got to follow through, and you’ve got to keep doing techniques like the Five Whys – I love to use the Five Whys – and there’s also root cause and breakdown analysis.
If you do that every time, at some point you’ll finally have it all figured out, and you’ll be on to wins and not just catching up. Everything is changing constantly – it’s not like the old days, and you have to keep up with technology as it advances, have to keep advancing with the technology because otherwise you’ll fall behind.
PS: What surprised you about reliability, and was it what you expected it would be?
MM: One of my big things was the engineers would put a machine together and buy it, and the company that’s selling it is selling based on price, not reliability, so they would put stuff on the machine because it’s cheaper. Then a year later, you’re redoing it all because you saved money in the beginning, but then you lost it tenfold when you lost production and you started incurring downtime.
So I would follow behind and make the machine reliable, and then I would tell them, “Why don’t you do this in the beginning? Instead of bringing a machine in where elements fail, why don’t we make it not fail from the beginning? Are you buying reliability, or are you buying price?”
Now maintenance gets involved with the engineers, because the engineers are the ones with the hands-on influence on purchase and design. If you put them both together, it’s a win. Maintenance should be involved when they’re putting a machine together, because maintenance are the ones that touch it, and they know what’s going on. They’re not looking at it on a CAD drawing; I’ve learned, everything on a computer looks beautiful, but, in the real world it doesn’t work that way.
We’ve started looking at each new machine and putting in elements we knew would work. I wasn’t worried about price; if the machine breaks down in three months because someone wanted to save money, well now you’ve lost money due to downtime. That’s when you start adding up, was that $200 extra worth it? And most of the time, it’s worth it.
PS: What else surprised you about the reliability side?
MM: I learned that I could make the assets work right before they even got to the plant. Now the attitude is, “why would we want to put something on that we know is not going to do the job and then change it out? Why won’t we just put the right thing on before we even get the machine, and then we’re three-quarters of the way home?”
Of course, there’s tweaking and everything involved, too, but as long as we get the basics right, you can work around all that, the little stuff, that’s easy. But putting on a wrong bearing or a wrong gearbox because of price will cost you in the end.
PS: When it comes to the digital world, the world of the internet of things, the world of looking at cloud-based remote data collection services and analytics – in many places, that’s not maintenance’s job, because they’re too busy fixing things. That’s also not an operations job, because they’re too busy running the machines. So in these places, digital initiatives seem to be falling in the lap of reliability. Are you seeing that?
MM: I don’t even think about digital, it’s just...it’s the way it is, and I don’t even dwell on it anymore. I do save a lot of data in the cloud, and in the beginning of my reliability steps, I realized that I could look on my phone and see stuff happening. So I take it more as it’s part of the system now.
The best thing now is once you make it part of your system, you don’t even really dwell on it because you’re used to using it. Getting used to it, in the beginning, is the hard part. Once I had some new digital sensors and we used them, and I’m like, “Wow, this is awesome. I can watch vibration from my house. This is great.” And then after a while, I don’t even dwell on it, I just use it. It just becomes part of your system after a while and you don’t dwell on it anymore. It’s just part of the whole PdM deal.
I’m going to say this, too, also, it’s about that one word I always talk about: change. I believe in change, so to me, it’s very easy to change to do what I know is right, and I don’t dwell on it. But in the beginning, change is hard. Change like this does often come from the reliability guys because maintenance engineers are usually fighting through their backlog, and operations is too busy operating.
Change is good, but it’s only good if you don’t get crazy and you do it step-by-step. If you try to implement a whole pile of changes at once, it usually self-destructs because it’s too much, too soon, and nobody wants to run with it. You take it from here to there, from there to there, and then it runs, and they see how it works, and it just becomes part of your system.
PS: You’re an avid sports hunter, and you’ve mentioned to me before that you see similarities in reliability and hunting. Can you elaborate on that?
MM: The reliability of hunting, with all the guns and everything that are out there...in the old days, you had to spend $800 or $900 to get a gun that was a good gun. Now, with all the computerized everything, you’re taking human element out of it. I can get guns out of the box for $250 that shoot just as good as my thousand-dollar gun. The way it’s going now: I don’t need to spend a thousand dollars any more for one gun. In fact, one of my newer guns actually shoots better than one I paid $1,400 for. It really makes you cry.
What’s happening is that vendors are taking the human element of those guns, and that’s just like what industrial machine vendors are doing. That machine does not care if it’s Monday morning, Wednesday afternoon, or Friday night. It just keeps doing what it does. There is no lunch time, no this, no that, and it makes product the same way every time. Once in a great while you’ll get a screw up, but it’s getting far and few in-between anymore.
PS: And just to clarify, when we’re talking about plant assets in this regard, we’re talking about the fact they can be engineered to such good tolerances. I don’t think what you’re saying is that cheaper is always better. But it sounds like what you’re saying is, don’t get nervous if something is less than you thought it would be, because these days, prices are coming down for highly reliable machines.
MM: Yes. For hunting, the trade magazines are all saying that these guns are coming out of the box with a .5, which is a half inch MOA. That means on a five-shot group, you’re a half inch off of bull with five shots. Now the cost comes down with all the computerized design, and you can retool that design quickly now too.
In the old days you needed expert gunsmiths who knew what they were doing with the gun, or else you were stung. But these days many gun-makers have taken the human element out of it. When you’re taking a shot, in the back of your mind all this stuff is going through your head, which can get in the way of your focus. Now, you don’t factor that in anymore; I know that if I squeeze the trigger, I’m right on the money most of the time.
There’s one more parallel with hunting and reliability. Remember, when you are collecting live data or ultrasound or infrared data, make sure you have all the proper PPE on. Make sure you keep your distance. And, as you’re gathering data, be aware of your surroundings, so you don’t walk into anything that’s bad.
Because you’re not looking at where you’re going, you’re so engrossed on you’re looking on the screen. I still do this once and a while, I look up and go, “Whoa.” I’m two inches away from a pipe at my head. It’s like, one more step and it would have been a head banger. You forgot about, when I move from here to here, am I looking at where I’m walking? Am I looking at what’s ahead of me? Because, that always hurts bad. It doesn’t hurt easy, let me tell you.