Industry leaders discuss the pandemic and digital transformation trends

This past July, reliability and maintenance professionals met for a virtual conference, Leading Reliability. The event included a closing keynote panel discussion led by Maureen Gribble of UE Systems Inc. She asked the panelists – Kevin Clark of Fluke; Jeff Hay of RDI Technologies; Blair Fraser of UE Systems; and Shon Isenhour of Eruditio – about what’s impacting maintenance and reliability related to COVID-19, and technology and digital transformation. This article presents selected highlights from the closing panel; you can view the entire panel at https://info.eruditiollc.com/leading-reliability-virtual-conference-recap, as well as register for the next virtual Leading Reliability event coming up in January 2021.

MG: With respect to COVID-19 and the reliability industry, what did you find interesting throughout this time during the pandemic?

JH: There’s a lot to be learned from the reliability industry. One of the things we talk about, for example, is the P-F curve and things like early detection, and being able to diagnose. I think it’s really interesting how we’re starting to implement that in a parallel track with how we look at machines and then how we’re implementing that with COVID, so rapid testing and being able to know that early enough so we can actually respond to it. I think we often think of it as downtime with our machines. Now, it’s shifting into humans, which are the important factor here, and it’s really hard to measure that.

MG: What maintenance and reliability technology do you think has been most impactful during COVID-19 and how do you see that playing out post-pandemic?

KC: From a maintenance and reliability standpoint, actually a number of things stood out, but I would probably put it more in a category than one particular technology, and I would say remote was the big thing.

What we did notice was systems like EAMs, and CMMSs, and mobility, and the capability to predict failures, and we got a tremendous amount of feedback on how those that were prepared with those systems, were able to go remote and be able to run their operations with limited staff.

• Read "COVID-19 and maintenance planning: The importance of planning during a crisis"

From our perspective, the most impactful technology out there was the capability to quickly go remote. And then going forward, a lot of these clients have already seen the big success in being able to go remote and have been looking at more technologies that put them in an even better place to go remote.

MG: What concepts or processes have been impacted by COVID-19? Are people having to innovate faster, or are they having to take their time to think through potential pitfalls or problems?

BF: I would include innovation on the remote side, and what was interesting when I looked at it with my friends and family outside of the manufacturing industry, accountants and things like that, they just brought their laptops home. For those in manufacturing, how are you going to bring a 250-horsepower engine home? You just can’t.

A lot of companies were saying, “We realize we need IoT. We don’t have any money to spend.” And that’s where we’re seeing new business models come up in terms of being able to lease equipment, rent equipment, all of these SaaS models that are really driving new business. I think those that invested are going to accelerate and those that did not are a little behind the gun but have an opportunity to catch up to the rest of the other organizations in terms of achieving Industry 4.0.

SI: I think one of the things we’ve definitely seen is a lot of folks that had put off education or put off sharpening the saw, if you will, they’ve taken the time to do that.

I’ve heard a lot of folks talk about the fact that they read this new book or that new book and sharing articles and things on LinkedIn. I’ve certainly noticed with people being out of the day to day, they’ve definitely taken the time to learn, to dive into understanding some new technologies.

BF: To me, it’s all about change, so it’s all about how you adapt to that change. Those are the companies that are going to survive: the ones that can adapt to change. It’s not necessarily about implementing the right technology; it’s those that can adapt to change.

And when we look at it, there are a few standouts within our manufacturing ecosystem. Those that quickly adapted their processes, from making beer and liquor to making hand sanitizer, and how quickly they changed course. You could not have done that in a pre-COVID world.

MG: How is the camera being talked about as a sensor for the future, and what do you think about technology moving forward in our industry as a whole?

JH: I think sensors and technology and data collection, a lot of it revolves around access to it, and I think we’re getting more and more access to it. If you take, for example, a camera, there’s a lot of data that comes into that.

I think when you break it down, with the camera, you get one data point, such as a pixel on a camera, and that has value. It tells you something, right? And then you get a second one, and you put those together, and that really has more value than just the sum of the two because you can kind of compare and you can pull more information out of that. And I really think that that’s where we’re going. That’s been a very big push in our industry, in lots of industries.

I like to think of it as actionable information. There’s really no value in data, unless you can take action from it. You put all the pixels together, and then you can make sense of all those data points together in a visualization and connect that back to where we’re going with Industry 4.0 and the cloud.

MG: Why do predictive and IIoT data-collecting methods and technologies seem to be an afterthought, and not built into the design phase?

BF: I think it comes down to cost, and the other factor is the consumer, whoever’s buying that equipment.

No longer is there going to be a single winner in the IoT ecosystem. There are going to be components and people that do parts really well, but they’re all going to feed together. That’s truly what a connected system or an IoT system is. I think we collectively as vendors, need to make sure that we build systems that are interoperable moving forward and that they communicate with other systems.

KC: We experience, from a predictive standpoint, that some of the basics haven’t been done, and that’s established in condition monitoring. That’s established in methods. Some of the reasons why predictive feels like an afterthought is the fact that we just don’t know what’s available to us.

So, we’ve got data sitting all over the place and nobody does anything with it. We’ve started seeing organizations take some of that data that’s just sitting there and drop it into an AI system and all of a sudden, it tells them all this crazy new information that they never knew for the last 20 years that they’ve been collecting that data. But that data all of a sudden makes sense. So, sometimes we don’t deploy things because we just don’t know that they’re available to us until we see it.

SI: I think one reason a lot of these technologies are an afterthought, honestly, is because we’re still not thinking about lifecycle costing. In general, we’re still purchasing the lowest cost component, the lowest cost asset. And if you’re doing that, you’re not going to get that technology built in. So, if your organization is not used to that lifecycle costing approach, where you’re looking at the cost of that asset over its whole entire span, then it would be very easy to make decisions to not spend that extra money.

BF: From an OEM perspective of designing and building the machines, they really have to take some ownership in the IoT ecosystem. OEMs hold the data. Specifically, if it’s a repeatable solution, that’s where they can start to add value, in terms of analytics and IoT, not just on the end-user’s end, but also on the OEM end because they should be able to track all the failures. I think they have a better chance at solving a lot of our problems from an OEM perspective on what commonly fails, rather than waiting for the end-user to supply that information.

JH: In a lot of industries, the breakdown happens any time there is a transfer process. One of the things that we are trying to encourage more and more with sensors is to get more on the commissioning, the design, and the installation phase – that critical time when something is coming in – to be able to get as much information during that process, and also to ensure through your technologies that everything is going according to plan.

MG: What area do you think is the most neglected in the maintenance or reliability field, and what impact do you see that having?

SI: I think a lot of the facilities out there today are neglecting the planning and scheduling side. It’s really easy to do some things, maybe implementing root cause analysis or maybe scrubbing some preventive maintenance task and cleaning them up a little bit, but I think more often than not, facilities aren’t getting the planning and scheduling right.

You need the ability to get the work done that you identified with the technologies. And if you’re just going to do the repair reactive, you’re not going to take precision maintenance into account.

JH: We’ve heard a pretty consistent message, especially the way people use the motion amplification tool. As a communication piece, often the people who are collecting the data that are on the technology side, they know about problems for quite a long time before they get fixed.

MG: Maintenance, reliability, and operations functions – which if any of these do you see as bearing the primary responsibility for being an IIoT/digital champion in their plants or facilities? Are you seeing any pattern among plant teams that you’ve worked with?

BF: It really depends on the organization, which functional silo is really driving it. For smaller organizations, often engineering maintenance can be one team or that can be separate. I do believe the IIoT value has to get outside of maintenance to reach its pure potential. It’s not just a predictive maintenance tool. It’s the ability to increase operations or truly get to operational excellence.

MG: How do you think machine learning and AI will contribute to Industry 4.0 and 5.0 over the next few years?

KC: The one thing that I’ve noticed over these industrial revolutions is that the first three were about 100 years apart from each other. Once we hit four, there was only about maybe a 50-year gap between Industry 3.0 and 4.0. And there is already talk of Industry 5.0 coming on. So how quick is that going to come? I don’t think it’s going to take 50 years to get to Industry 5.0. I think there’s going to be a massive shift in the way we look at industrial revolutions as well. But AI and machine learning, whether you believe it or not, is really in its infancy right now.

AI and machine learning are a big part of Industry 4.0 in that we’re starting to look at data, massive data in a different way than we looked at it before. And now we have the bandwidth to actually do it.

• Read "Do more than bounce back from COVID-19, bounce forward"

What they’re saying is Industry 5.0 looks more like the direct interaction between robot systems and people, automation systems and people actually interacting seamlessly. I think that’s going to come on much faster than 50 years.

BF: Generally speaking, algorithms have been out there since the 1970s. They haven’t changed. The AI model is when you apply data in the training to an algorithm, then you have an actual AI model. We’re using algorithms that are typically 20, 30, 40 years old. What’s different about that? It’s how we’re training those models and the data we have. So, first of all, we always have a lot of data. Chances are we’ve always had that data, but how we’re training that model is different.

We still need human intervention, and we still need subject matter expertise. This idea that you can just throw data at an AI model, and you’re going to come up with information is one of the most dangerous things I have heard in the asset management space. The statement, “Give me your data, and I will find you answers to questions you didn’t know to ask,” is just as dangerous as saying, “I have always done it that way.”

SI: The struggle I see is with a lot of these startups that come onto the market, small companies that are attacking artificial intelligence, they’re attacking machine learning, but the issue that I see is that they don’t understand the failure modes of the assets. So, it gets to be this situation where they’re making judgments based off of things that really have nothing to do with the performance of the asset.

Know the failure modes of that asset so that you’re actually looking at those predominant failure modes moving forward.