Podcast: How to Blend Tech and Shop Floor Experience for Successful Predictive Maintenance
What You'll Learn
- Why technology alone isn’t enough and how pairing condition monitoring with root-cause problem solving drives actual reliability gains.
- Where plants should start when launching predictive maintenance programs, using KPIs, historian data, and early wins to build credibility.
- How to bridge generational gaps by combining tech-savvy approaches with the deep knowledge of experienced technicians.
- The danger of overplanning, and why Kuhn recommends acting instead of falling into analysis paralysis.
Predictive maintenance is reshaping plant reliability—but only when technology and human experience work together. In the monthly podcast miniseries, Ask a Plant Manager, Joe Kuhn—industry veteran, author, and former plant manager—unpacks how companies can avoid over-relying on dashboards and smartphones while ignoring what's happening on the shop floor.
Joe Kuhn, CMRP, former plant manager, engineer, and global reliability consultant, is now president of Lean Driven Reliability LLC. He is the author of the book “Zero to Hero: How to Jumpstart Your Reliability Journey Given Today’s Business Challenges" and the creator of the Joe Kuhn YouTube Channel, which offers content on starting your reliability journey and achieving financial independence.
Below is an excerpt from the podcast:
PS: Well, today we're going to talk about technology, specifically predictive maintenance technology. But before I get to that question, I just wanted to draw attention to a column that you did a few months ago, and you were warning people against what you called a hubris contagion, and not really a failure of hubris or overconfidence in general, but you were talking specifically about an over-reliance on technology.
And you were clear that the message is not that tech is bad. It can be a game changer when it's used correctly, but there is this dangerous gap that can form between virtual data and shop floor reality. For example, too many people managing from their cell phones instead of really knowing what's going on out there. So how should managers and the plant floor in general be using this kind of technology, and specifically when it comes to predictive maintenance technology, what's your advice for plants that are interested but don't know where to start?
JK: Great, great questions, and I've actually been through this, so it's very vivid in my mind, what you can do wrong and what works. The temptation for people is to, like you said, manage from your computer, manage from your cell phone. And while technology can be great or is great, I'll say is great for early indication of a potential problem, it doesn't tell you what to do. It doesn't tell you what to do. So you've got this anomaly that say, shows up on a motor. What do you do? Well, you can just change out the motor, but you’ve got to get out of your office, talk to the electrician, talk to the mechanic, on whatever the device is. Go and see how it's operating. What's going on right now? Is operations, are they using it outside of the design parameters? So they're running the motor at 120% capacity? Are they beating and banging the pump around that you're using? What is going on on the shop floor.
That's where you transition from, ‘Hey, I found an anomaly’ to ‘What are we going to do about it?’ To fix the repair and then also problem solve so it doesn't happen again. If you do not add problem solving, this is very, very important, if you do not add problem solving to your condition monitoring program, you're going to be horribly disappointed. All you're going to do is have more problems come at you, and you've already got your unplanned downtime, and all the chaos that can be the life of a maintenance leader, but adding problem solving is how you get better every single day.
So technology is great. And what I've been seeing is technology is taking over 90% of the effort, and I think it ought to be no more than 40, where you're finding the anomaly, you're finding things fast. And what's critical about condition monitoring is you find the pump failure when it's starting to have an anomaly for vibration. You don't find the pump failure when it's burned on the ground in a pile of ashes. Well, try to problem solve what went wrong when it's just this molten ball of rubble and you're stressed around getting that production center back up and running. So the whole focus is on restore and flow, where, with condition monitoring, finding problems early, you schedule that downtime, then you take that pump and you autopsy it. What happened? Oh, it was our lubrication practices, or it was our packing practices, or it was a bearing problem, and then you root cause that and say, ‘Hey, let's not use style A bearing. Let's use style B bearing from this manufacturer.’ And then that problem solving occurs, and then it doesn't happen again. So you have to blend both. It can't just be technology. And I see too many people managing, just from the red and green on their cell phones. An anomaly is saying, hey, let's just change out that pump. Change out that motor. You are going to be disappointed if you don't add problem solving and going and seeing adds that dimension of ‘Okay, now what are we going to do next?’ Okay, very critical.
The other part of your question was, okay, we think the condition monitoring is the direction we want to go. Where do we start? Well, it depends a little bit on where you're at. If you're 90% unplanned maintenance and you're just reacting, every single day is a brand-new day. That's a tough place to begin, but condition monitoring can be the right solution there. The first place I would start, and I'm going to trip the audience up here, if they've been longtime listeners, I'm going to say, look at the KPIs, and talk to the maintenance people, talk to the historians at your site, and say,’ Where are most of our problems?’ And they may be, ‘Oh, well, gosh, if we can ever figure out these motors. These motors are supposed to last 10 years and are lasting 10 months. Or could be a pump, it could be a gearbox. Get an indication of where your problems are, and then that targets your observation. Okay, so go out and see go talk to a mechanic, watch them do a PM, on a on a pump, on a gearbox, and compare that to what the standard is. And you may say, Okay, I think lubrication is the place to start. After you looked at KPIs, and after you talk to some people, you have this gut feel this is the right place to start. You may think it's IR because it's electrical problems. IR detects heat very early, and you can, like I said, perform the preventative maintenance to fix the circuit board, or whatever, the fuse, whatever's hot, and then you can troubleshoot that. So what you're you want to start a condition monitoring program where you have problems. You don't want to start them on, if you got 10 production centers, and the first five are critically important and the last five aren't, you don't want to start on the last five. Who cares if you're successful, right? Nobody really cares if it doesn't save money and doesn't improve downtime. So you want to pick an area where you can make some improvement because people are watching. You've got the leadership team watching. They're investing in people and in technology and in training. Also that's a given, but also they're investing in downtime. You're going to tell them that this gearbox has a vibration in it, but to them, it's working fine. And you're going to tell them, I want to take an eight hour outage on Thursday to replace a perfectly good gearbox. And there's a lot of faith that comes with that, okay? So you've got to work on problems that are going to make an impact, okay? And then you're going to dive into that gearbox. You may change it out, or change out a bearing or gear, whatever. But then you need to take that gearbox, put it on an autopsy table and find out what's wrong with it. And that's the problem solving that I said you have to add. You have to add that problem solving.
So I would say where to start, with your historians. Start with your KPIs. Point you in a direction. Whether you want to start with IR on some electrical equipment, some vibration on pumps and rotating equipment, you may even start with UE. UE, ultrasonic emissions is a great place to start, because it's really the technology that finds anomalies the earliest and gives you more time to change things out, do the repairs and problem solve.
But one more thing I want to add is, there's, errors you can make in this. You didn't ask this question, but the pitfalls, one is not adding problem solving. And I'm saying this because we started a condition monitoring program at my plant, my first one, and we were pounding our head, ‘We've invested so much into this, and six months later, we've not seen any improvement, nothing.’ And then I reread a book by Ron Moore, “Making Common Sense, Common Practice.” It's on my shelf. Sometimes I point to it, and it says you had to add problem solving, and it's on page 220. I remember the page number. That's how dramatic this was. I'm going back to 2000, and I remember the page number of the book, and it said you have to add problem solving.
But there are some regrets you're going to have going down this path of condition monitoring and problem solving. You're going to wait until your program is perfect. You want to wait until you have a complex program. You're going to study your pump, look at failure modes, look at the technologies, UE, vibration and lube, and you're going to study how you can apply it to that one pump. And it's going to take you six or eight months to get a good program in, or one pump. I used to have 20,000 assets.
Imagine how long it's going to take to do root cause failure analysis, failure mode analysis on every single piece of equipment. You can't so you just start. That's my message: is to just start. If you think it's a lube problem, start with lube. If we think it's going to be electronic, start with IR. Walking around the plant with an IR camera is a great place to start. You start to see things. That's not your best on a lot of rotating equipment. IR is not the best because it's kind of late. You may be hours away from an event, but waiting until you have a very comprehensive plan to start predictive maintenance will be a regret you have, because you just have to start. You just got to start and learn by doing.
The second failure I see people have, and I've had, is they're slightly understaffed. Say you have 20 mechanics or technicians on your team, and you're down to 18, you're going, ‘Oh, we can't wait to do this condition monitoring stuff, but I got to get to full staff.’ You'll regret that, because condition monitoring will dramatically change the number of work hours you need out of your team, because you're going to find problems early. You're going to problem solve them. So those are the two regrets I predict that you'll have. Ty not to have them. Learn from Joe's mistakes.
About the Podcast
Great Question: A Manufacturing Podcast offers news and information for the people who make, store and move things and those who manage and maintain the facilities where that work gets done. Manufacturers from chemical producers to automakers to machine shops can listen for critical insights into the technologies, economic conditions and best practices that can influence how to best run facilities to reach operational excellence.
Listen to another episode and subscribe on your favorite podcast app
About the Author
Anna Townshend
Managing Editor of Control Design and Plant Services – Endeavor Business Media
Anna Townshend has been in B2B magazine publishing for more than a decade. She took her love for business journalism to a new industry in June 2020, where she is the managing editor of Control Design and Plant Services, published by Endeavor Business Media.
Plant Services helps plant operations make the most of smart manufacturing processes and asset management.
Control Design serves industrial machine builders and the OEM market for industrial controls.