Improve Reliability in 100 days with Machine Learning, AI & Whiteboards

Matt (Host): All right, everyone. Welcome back to the Comparesoft Podcast. Great to have you here as it always is. Today, our guest is James. James is the operations manager at AJS Asset Care. James has worked with companies including Musk Process Services, Crown Holdings, and now is with AJS Asset Care.

James has been a mechanical engineer, an engineering team leader, and an operations manager with a drive to improve reliability in 100 days, so it’d be really interesting to hear his take on maintenance management.

Hey, James, how’s it going, mate? Thanks for being on the show.

James Maile: Hi, Matt. Thanks for having me.

Matt: No worries at all. Let’s jump straight in. Can you tell the listeners what you do? And a little bit more about your job, specifically?

James: Yes, so I’m the operations manager for a business that helps create and deliver a maintenance strategy within their organisation. That can include a variety of different tasks, and it’s very specific to what that end user’s requirement is. That could be a fast-track approach and, as you mentioned, the goal could be to drive improvement and reliability within 100 days, or it could be a more specific approach to address a certain area that is perceived to be deficient. We can really highlight that through what we call an asset care assessment.

People often ask; why 100 days? Well, it’s 100 days, because to effectively deliver that new strategy, we need enough time to engage with the engineering team and to drive the cultural development and change that comes part and parcel with developing and implementing a new maintenance strategy because they’ve always done it a certain way.

There needs to be some education. They need to be on board with what we’re trying to do. And, being part of their engineering team for 100 days, helps us drive that change within the business.

Matt: Okay, well, how would you set a good maintenance culture?

James: That’s a really interesting question, and a bit of a hobby horse of mine, if you like. Because, fundamentally, I don’t believe that you can set a culture, but what I do believe is that you can control the environment, which in turn should influence the culture. So, people’s behaviour is influenced by what leaders measure and what they react to. So, KPIs for me are really fundamental when considering a maintenance culture and a really good example of what I mean by that is going to a lot of organisations where they will measure the mean time to repair as a metric.

I often say to them, “Well, what behaviour are you actually trying to drive with that metric?” Because to me, you are at that point telling the engineering team that they are being measured on how quickly they can repair something. In engineering terms, that is the wrong mentality to have. Because a much more effective measure would be something like, “How much root cause analysis have you done this month?”

Because then when an engineer goes to that breakdown, they’re not going to look at it and say, “How can I repair this as quickly as possible because that is what the leaders of this organisation want me to do?” They might ask themselves, “Why has this failed? How do I repair it properly? How do I stop it happening again? What information do I need to take now to be able to make an informed decision later on?”

Matt: Okay, excellent advice there. And then, so, how would you recommend to plan maintenance activities?

James: So, for me, maintenance planning always has to be a team sport. They’re maintenance activities, but they’re also business activities because your maintenance strategy should feed into your overall business strategy.

So, production planning and engineering all have a role to play in the planning process. And that should be a formal meeting, which is documented and people are held accountable for; Okay, this production line is going to go off between these hours, these are the activities that we’re going to plan in at this time on this date. It needs to be robust from that point forward.

And that plan really needs to reflect the maintenance cycle within the computerised maintenance management system. It’s been scheduled based on the factory’s requirements and should be based on a criticality assessment. So we’re focusing on the maintenance that should be done. So it needs to feed directly into that schedule.

Matt: How do you recommend implementing an effective maintenance plan?

James: So the plan must be both visual and accountable. So, put [it] on notice boards. Maybe you can plan through a tablet, because you’ve got a tablet-based CMMS system, and the engineers can go on and see it and production can go on and see it. But it has to be visible to the people in the organisation so that they know what’s happening and when it’s happening. You also have to be accountable for what is planned. So, has it been done? If it hasn’t been done, why hasn’t it been done?

Again, I do refer back to key performance indicators a lot. But if you’re using metrics, such as plan adherence, and corrective work order backlog, then you can really use that to measure how effective your planning is, and how effectively it’s being used to make the decisions within your planning process.

Matt: Do you think software tools are useful for managing maintenance activities?

James: Yes, I think that software has actually changed the way that we conduct maintenance within the UK, for sure. If you think back to the good old days, then it would have been completely unmanageable to deliver an effective maintenance programme without the help of software. And software’s really grown from the kind of infant days when it was a scheduling tool. It’s now an analytical tool. It’s now a tool that feeds other avenues of the business.

I know often, if we’re talking about a CMMS system, as an example, as a software tool, that is the heart of your maintenance strategy. Everything stems from there. Absolutely, everything stems from there. That’s where your decisions are made. That’s where your KPIs are drawn from. That’s where your schedule exists. Not all organisations are there yet, but really, it’s more of a mandatory thing for a functioning engineering department to have in this day and age.

I mean, could you imagine running a factory with over 500 assets, as an example on a paper-based system or without something that enabled you to schedule effectively, and then manage your maintenance tasks and input your maintenance tasks? It blows my mind just thinking about it!

Matt: That sounds like a nightmare. Doesn’t it, basically? The not-so-good old days, as it could be said. So, how do you see connected technologies changing the way we work in the future?

James: So, technology is moving very quickly at the minute, especially coming out of the pandemic. Industry 4.0 seems to be the words on everybody’s lips. There’s some really impressive stuff going on out there. I think the area that’s really piqued my interest the most has to be machine learning, which is using various operating conditions to inform artificial intelligence and ultimately do some maintenance based on data.

You know, that may be vibration from the line, it may be work orders raised in the CMMS or the maintenance schedule that’s feeding into that tool coming from a motor. But that data that’s feeding an algorithm that tells you all of that information about a machine, while it’s operational and doesn’t require the intervention of an engineer is really where I think the big change within industries going to be.

That being said, I’m seeing a lot of instances where industry 3.0 hasn’t yet been adopted, and manufacturers want to talk about 4.0. They don’t actually have an effective maintenance planning regime in place at the moment, or they aren’t using condition monitoring. Or they aren’t doing PMs that add value.

But to answer the question, I do see connected technologies reducing human maintenance activities through this machine learning process and using the engineers to – in real layman’s terms – put things right rather than to use their engineering judgement to make a decision based on a calendar maintenance task.

Matt: Yeah. Okay, and then so going back to those good old days that we were talking about. What’s your take on using spreadsheets for maintenance management?

James: It’s chaotic. It’s never ideal, but they do have a place. They have a place if there is an absence of a robust maintenance management system because you have to use the tools in your arsenal. If there’s nothing there, it’s a base point. It’s somewhere to start collating some information and driving some change as a very short-term stopgap.

But it isn’t a database. It has its limitations. It’s limited in how it can inform your decision-making. Essentially, something is always going to get missed. Whereas, with a computerised maintenance management system, it does the thinking for you. That’s a complete manual process when you relate back to the spreadsheet, in my opinion.

Matt: Interesting take and view on that. So, how would you use maintenance as a competitive advantage for the business?

James: So, in my opinion, maintenance is an investment rather than a cost. It’s often seen as a cost to the business. But while good maintenance does cost money, it actually realises savings in indirect costs to the business. Whether that be downtime, whether that be labour, that’s not used reactively as an example. But what maintenance allows you to do in terms of a competitive advantage is, it allows you to take control of your assets. It allows you to manage your customers’ expectations better as a result of that, because you are in control, more in control.

You can never be fully in control. Because things fail [and] some breakdowns are unavoidable. But what you can do is you can plan things ineffectively [and] manage your customers’ expectations through managing when you are going to switch those assets off. If you don’t have a robust system in place that does capture degradation and captures corrective actions before something fails, then you’re always at the whim of the asset telling you when it’s going to switch itself off.

So you can manage your customers’ expectations better through that process. You can also start planning your budgets more effectively because you understand the performance, and you can make better-informed decisions on when is it potentially better to replace.

Matt: Okay, James, well, you’ve given us some amazing information and advice there already. But now, I’m going to ask for three more pieces of advice. So what are your top three tips on effective maintenance for our listeners?

James: So, first one, I’m going to say [is] act on the data, don’t act on emotion. Engineers will always migrate to the pain that they felt. I spent a lot of time repairing that breakdown, that must have been a big deal. I’m going to put it right. Actually, what you can do is you can go into your management system and you can say, “all right, okay, I’m going to have a look at the data.” These five minutes stop that’s happening every two days, that nobody particularly cares about, is cumulatively causing more downtime than those one-off events that everybody attaches themselves to because it was a big deal and it costs a lot of money. But actually, that’s where the cost sits and I know that’s where the cost sits because I’m looking at the data. And my engineers are putting that data into my CMMS. I can use that analytically to make decisions. But that goes so unnoticed in organisations, and those five-minute stops are an absolute killer for productivity. And then there are nobody’s radar. So act on that data, not on the pain you felt on the emotion.

Well, my second point would be; have a process for learning from your failures, or typically use root cause analysis. That’s failed. We want to do something about it. Use [the] root cause analysis five why’s; why did it fail? Why did it fail? Until you get to the answer and the actual reason why it failed so that you can mitigate that happening again. Because typically what we say is you experience the same breakdowns because you never actually repair them properly because you never actually find out why it failed in the first place. You believe you’ve put the issue right, but you haven’t. You’ve just put a plaster on it. And those same failures are happening time and time again within the factory, but then they’re never fully addressed.

And the final tip would be; corrective maintenance is the maintenance that actually makes the difference. If you don’t plan your correctives, you’re never going to stop reactive failures. You could have a very effective plan maintenance regime in place, and the engineers are picking up bearings before they fail raising it in the CMMS. But until you plan that proactively before it fails, you’re actually not ever going to make a difference to the downtime that you’re experiencing. All you will, in turn, do is be aware that that failure was going to happen and then you experience it anyway.

Matt: Okay. And then so wrapping it up in a bow, what’s your favourite saying or quote on maintenance?

James: Well, if you were to ask my team, they would say that I have a few, maybe two. I think my favourite one which I use to myself really a lot for calibration, is when you’re thinking about maintenance and maintenance activities is this; the consequence of the failure justifies the maintenance effort.

What I take away from that quote is that as engineers, we focus a lot on what maintenance we could do, but not what maintenance we should do. I see a lot of wasted efforts in industry today. Actually, sometimes running an asset to failure is the right maintenance approach.

Matt: Well, there we have it guys and girls. You’ve been a fantastic guest, James, and dropped some really concise, precise information for us there. So thanks very much for being on the show.

James: Thanks for having me, Matt. I’m just happy someone had listened to me for once.

Matt: Well, I’m sure you’ve got lots of people listening to you now. So there we go. As always, it’s great having you guys back here listening with us on the show. So stick around and we’ll see you on the next episode. Cheers.