Editors’ note: Three years ago, Plant Services met veteran millwright and newly minted predictive maintenance technician Michael Macsisak (https://plnt.sv/1810-MM), and PS then caught up with him a year ago as he began work in mining (https://plnt.sv/1910-MM). Macsisak sat down recently with Plant Services Editor-in-Chief Thomas Wilk to look back at his work as a PdM specialist and look ahead at what’s next for our industry.
PS: The past couple of years, as you’ve matured as a reliability specialist, you’ve had increasingly bigger wins. What’s an especially memorable one?
MM: During my work in the mining industry, it was (one related to) the hoist, which is a main piece of equipment that transports ore up to get crushed. The bearing was failing and massively out of alignment. I can’t give you numbers, but I’d say I prevented four to five days of total unplanned downtime by finding the faults early. And it could have been longer depending on what all happened when something let loose.
PS: What combination of technologies did you use? What kind of data were you looking at to find that potential failure?
MM: I don’t believe in one; I use all of them. Oil analysis comes first, then I track and trend with vibration, ultrasound, and infrared. The thought process being that you find it first with ultrasound, and then vibration comes into play, and the third is IR. After IR, which detects heat, the next step is failure. The challenge is that failure could happen later or it could happen sooner. There’s no set way to tell with 100% accuracy: “This is really hot. It’s going to make it two days to failure.” That’s when you have to act. However, when a fault can be picked up with IR, something will fail, so you have to act sooner.
PS: In the case of the hoist, did IR pick up a fault whose emergence was imminent?
MM: Yes, the bearing was starting to heat up, but I caught it more with the vibration and the ultrasonic, and that put me on the path to track and trend. The whole point of this is to track and trend at least monthly, and in critical equipment, biweekly.
Without trending, you have no idea what happened or what’s going to happen, because you can’t watch stuff as it increasingly goes to fail. I set it up; I’m looking for bearing faults; I’m looking for an imbalance, alignment, I’m looking for all those things. Most of the equipment that fails, I’ll say 80% of the time it’s due to misalignment.
PS: When you saw this hoist failing, what was the data that caught your manager’s attention? Was it the IR data? Vibration data?
MM: It was a combination of both, plus management has to have to trust in what you’re doing. If they trust you, they will follow your lead, and it will work. And my maintenance manager agreed with me, and we moved forward, and we caught it before it went in total failure.
Also, with these kinds of technologies, keep in mind that some of them may not work to detect the fault. Only the correct one or two will work. You can’t use all technologies on every asset, because they often won’t work in the situation that that machine is in. For example, if a machine is designed to shake, it’s very hard to track it with technology that measures shaking.
I’m the guy with boots on the ground, and I confer with the engineers and then we make a plan and move forward. That manager was a predictive maintenance guru, so it was really good for me because we combined our knowledge and that’s how we made our plan. It’s always good when you have someone with predictive maintenance experience as your maintenance manager or reliability manager, because they understand what I do.
PS: When you found that failing gearbox, was it using the same set of tools you described? The combination of ultrasound and vibration and IR? Or was it something different?
MM: It was ultrasound first, and then I backed it up with vibration. Remember, the chain of command, the first thing you do for PdM is oil analysis. Without the right oil and the right grease, no matter what you do it will fail. Once you get past that part and you’re properly greasing, you track with ultrasonics and vibration, and you compare notes. And when they come together, then you know pretty much you’re dead on the money.
Then the question becomes, how long? This is the biggest question I’m ever asked: “How long do you think it will last before failing?” I’ve learned over the years that you have to learn it and figure it out on your own – there’s no set procedure. That bearing could last six months after you find it, or it could last a day. You just don’t know. The only way to know is to turn it off and take it apart. So my thoughts are: Catch it early; fix it now; and keep going, instead of (dealing with) the uncertainly of, “Can we make it last? Could it last a couple days?” Because betting on that bearing to last a couple of days doesn’t work all the time.
Everything fails when you don’t want it to fail. It never fails on your terms. When you don’t want it to break, guess what happens? It goes “boom” is what happens.