S2:E3 The Importance of Maintenance Data to Unlock the Capabilities of Automated PdM with Robert Russell
In this week’s episode of Masterminds in Maintenance, we have Robert Russell, CTO and Co-Founder of Senseye, on the show! Ryan and Rob discuss harnessing technologies like Industry 4.0 and AI/Machine Learning to do the heavy lifting of predictive maintenance. Listen today!
Episode Show Notes
- Why is data such an important parameter in maintenance?
- What are some practical ways we can take data from sensors/programs and turn it into actionable insights to keep our machines up and running?
- Advantages of automated monitoring over manual monitoring
00:02 Ryan Chan: Welcome to “Masterminds and Maintenance,” a podcast for those with new ideas in maintenance. I’m your host, Ryan. I’m the CEO and founder of UpKeep. Each week, I’ll be meeting with a guest who’s had an idea for how to shake things up in the maintenance and reliability industry. Sometimes the idea failed, sometimes it made their business more successful, and other times, their idea revolutionized an entire industry. Today I’m super excited, we’ve got Rob Russell here on the show. Rob is the CTO and co-founder of Senseye, a cloud platform providing predictive maintenance. It’s used by maintenance teams to decrease unplanned downtime and increase maintenance efficiency. Senseye is also trusted by several Fortune 500 companies such as Nissan, Tata Steel, Alcoa, Siemens, Schneider Electric, and more. Wow, you got quite the impressive customer base, Rob.
00:48 Rob Russell: Yeah.
00:48 RC: Welcome to the show. I’m really, really excited to be speaking with you about predictive maintenance, Senseye, and the future of everything maintenance and reliability. Welcome.
00:58 RR: Yeah, thank you. I appreciate the invitation.
01:01 RC: Alright. Well, you know, I always love starting these off. I know a little bit about your background, Rob, but could you share with our listeners about your background, how you got started in this field of maintenance and reliability?
01:16 RR: My background as an individual from an educational standpoint is as a Mechanical Systems Engineer. After graduating, it was back in the ’90s, and you couldn’t be too selective with jobs at that time. I ended up, through good fortune, working in a helicopter company in the South of England. You can maybe tell, that’s a long way from home for me. I’m originally from Scotland, but there I ended up… I was working within the area of avionic systems design and there was a natural fit into the role of condition monitoring, which in those type of vehicles, it was described as health and usage monitoring. And it was all about monitoring the mechanical aspects of the aircraft, and that gave natural links from the utilization of that data through into the maintenance world.
02:04 RR: As we moved and the on-board systems were all developed, naturally started to tend to work on the ground-based systems. So, this is the software that’s receiving all that complex information from aircraft and processing that and trying to make it into something that’s super meaningful and useful to the maintenance team, but there was always an involvement of the OEM as well as the operator. So, at that time, we were doing maybe what you would call now cloud platforms, but it was all within constrained, secure military networks. But we were bringing data back from many different military operators, back to the OEM environment to provide that level of support. That led us to introduce data warehousing and web-based applications for people, and there was a natural sense there of that capability. Back in the later 2000s there, there was an emergence of connectivity that was coming out of the industrial IoT space and it was just too good an opportunity for us to miss. We realized that we’d done this in high value assets, but there was areas that could greatly benefit from this type of capability as we moved into an industry 4.0 world.
03:20 RC: The whole IoT in predictive maintenance has kind of been this buzzword in our industry and some people have said that it isn’t widely adopted, it hasn’t widely spread to every single industry, every single company, the way that a lot of companies have predicted by the year 2020. I’m curious, is that similar to what you are seeing, Rob? And if it is, I’m curious, why do you think there has been a slower adoption of this IoT 4.0, Industrial 4.0 Revolution?
03:57 RR: We see that reflected very much in the inbound marketing that we have. If we go back to when we first started Senseye back in 2014, there was very little. We were going out, trying to find those early adopters in the market and the people that didn’t quite understand what it was that they were looking for, but we had this vision for the product very much in the last two years or three years that traffic and that inbound has started to ramp up, so we can see the market is moving on and the message is getting out there, but there’s some hype to overcome. There’s people that relate a lot of the capabilities that we provide with other buzzwords like machine learning and artificial intelligence, and there ends up a lot of confusion in the market. We find that having that sort of focus on that specific use case of predictive maintenance helps us rather than some people that approach this by trying to be like a general analytics platform that you can apply to lots of use cases. We just focus down on the specific use case and use the right technology behind that.
05:03 RR: But back to the point about the reason that maybe the speed of the market adoption relates to some other technologies that we rely on, so we’re very much a cloud application, so we need that level of connectivity within the factory. And some people talk about the roll out of the Industrial IoT. Well, that’s quite different to the way that we would roll things out in the commercial world. Some of these machines, they’re bought on very long procurement cycles, and they’re expected to last for decades. So, if you’re buying a machine, you want it to be there 20 years. So that level connectivity that you need to accelerate, in some cases, has to come from retrofitting because you can’t rely on the whole market adopting the newest and latest and the greatest connected CNC machine or robot or conveying system. So, there’s a mix there and it’s that retrofitting capability which is super important. And we’ve got hardware partners that we talk to regularly that help to fill those gaps.
06:08 RC: Right. So, Rob, what I’m hearing is kind of like this is the natural adoption phase for companies purchasing multi-million dollar assets, it’s not like we’re gonna buy the newest, latest and greatest, like you mentioned, CNC machine when we’ve got one that’s only halfway into its life cycle. So what it sounds like retrofitting is really, really important and able to connect these very expensive assets to the cloud, which is kind of the crux of enabling predictive maintenance. We talked about the difference between predictive maintenance, big data, machine learning, artificial intelligence, how do you define the difference between predictive maintenance versus all the other buzzwords out there going on and floating in big marketing messaging platforms?
07:02 RR: Fundamentally, every type of maintenance really is reactive. You’re reacting to some piece of information and whether that is an advisory coming up to say that you’ve got a scheduled piece of maintenance to do on this machine right up to with all the things that happen in between with condition monitoring solutions. For us, it’s about providing as much lead time and a warning on those failures as possible, to enabling companies to have the right type of ability to react, but also being able to present that alongside supporting evidence and information that helps you to diagnose and understand the cause of the problem. We’re very much moved from the world of people building very physical, analytic models of machines. And this is one of the differentiators within Senseye that sometimes it’s harder for us to get across. We won’t be using those traditional approaches that I experienced in the aviation sector where you understand everything about the design of that component. We have to have a far more data-driven approach and also utilize as much context from other systems like CMMS systems that you’re more experienced with than I am, really. It’s as well as having that understanding of the normal behavior of the machine is to be able to build up that evidence set of how machines have failed in the past and then automatically build the models irrespective of knowing the physical characteristics, making it highly data-driven.
08:35 RC: Absolutely. So what I’m hearing is, data is the most important piece, enabling better… It’s almost like reactive maintenance. Data is the crux that enables us to have a bigger lead time and that longer lead time is, it sounds like the difference between reactive and what marketing defines as predictive maintenance. You’ve helped large Fortune 500 companies adopt some parts and pieces of the Industrial 4.0 Revolution IoT. What have been the most common mistakes that companies make when trying to adopt these IoT solutions?
09:15 RR: Not having quite the right level of appreciation of the right type of data that you need to gather and how to gather that data. Now, we’ve talked about these legacy machines. There’s a lot of information that exists within those control systems that can be utilized, where it’s typically quite noisy process data and it’s influenced by a lot of other parameters. One of the things that we work very hard with our customers is showing them how they can exploit that data, but there’s levels of processing that has to take place and that appreciation… The data fundamentally has to have information within it related to the condition of the machine. There’s no magic here within the algorithms of machine learning. You can only learn from what you, of the information that’s available, what you tell them. So that’s an important aspect, is being prepared and ready and having an understanding of good condition monitoring practice.
10:12 RR: Other aspects as well that we would relate to lessons learned, is being very aware of the cultural shift that’s gonna have to happen in your organization. Some people feel threatened by software analytics and think that jobs are going to go and we’re all going to be replaced by automated services. That’s not necessarily the case because the demand is increasing, margins are reducing, and people have to just work smarter. And that’s really, really key for us, understanding or having the maintenance team understand that we’re there to help, there to give them more information and help them move from that reactive position to the proactive. But that takes support right from the top of an organization all the way down to get that cultural change. It’s definitely not bottom up.
11:05 RC: Absolutely. We actually hear something very common as well where it’s like almost a fear of embracing technology, but at the end of the day, what we see so much is that technology enables people, maintenance technicians, facility managers to do their jobs better and be more efficient and ultimately run their facilities much smarter. It’s still people at the end of the day turning wrenches. I’m excited for the day that machines, robots automatically repair themselves, change the filt… I’m excited for that day, but I think it’s a little bit farther out. Rob, I’m curious, obviously we’re in this COVID world right now. How do you think IoT is gonna affect businesses, Fortune 500 manufacturing businesses. How is IoT gonna affect these types of businesses in the post-COVID-19 world?
12:02 RR: Even before the situation we are now, we always envisaged this concept of the dark factory. So, the reason it’s dark is it’s unmanned. That’s an area where you won’t have personnel walking around, it will be picking up on strange nuances and machines and maybe picking up the early stages of faults. So in that area you need something that’s able to monitor those high numbers of assets and fundamentally do it remotely and then bring in humans to the decision-making process at the appropriate time. With the current global situation, my feeling is that those strategies and thought processes about remote work and not having so many people in and around factory environments will probably be accelerated and people will be looking for ways to mitigate against a future scenario like we’re in now. These will be in business contingency plans. So we will end up with the ability to remotely access the data from machines, but that will require some level of monitoring to make sure you get the right attention to the right machine.
13:16 RC: Absolutely. I definitely see the role of remote condition monitoring accelerated in this post-COVID-19 world because, like we mentioned, companies, businesses are gonna have to start doing more with less, and obviously the level of personnel inside of a factory is gonna be limited because of fear of spread of COVID and the more that we can monitor equipment from the seats of our homes, the better we’re gonna be to protect all of our workers that are working inside of a factory. I think that’s one big benefit of obviously automated remote condition monitoring over the technician walking up to a meter and recording its value. I’m curious, what other value props, benefits have you seen from folks adopting IoT?
14:10 RR: That just increased level of insight into the machine. Even in some simple cases before we’re even seeing the advanced end outputs from a systems like ours, people are just having a much better understanding of the way that those machines are operating and access to the information is fundamental. But more importantly than that is the level of coverage. Typically within a condition monitoring or predictive maintenance space, you’ve been looking at high criticality assets, things that have the severest impact on your business. Typically these are things that are over-monitored, over-maintained and potentially over-engineered to mitigate in all factors. But we talk about the expression the “balance of plant,” so what about all the other things in there?
14:57 RR: Now, we’ve seen experience with our customer base that it’s those other assets that can be just as impactful, but it’s things that you don’t predict and foresee. And I relate that back to issues we experienced in the helicopter sector, the type of failures that we saw that grounded entire aircraft fleets were not things that had been envisaged or embedded into the health and usage monitoring systems and their capability of monitoring and these were unforeseen, your failure modes. So you can… There’s a lot of really excellent and experienced asset managers out there following very good asset management practice and looking at failure modes effects and considering that in their maintenance policies, but there will always be a surprise and having that, some level of connectivity into those machines and looking for unusual behavior can be very insightful.
16:00 RC: Yeah. What I’m hearing Rob, is IoT really decreases the cost to get better coverage across all of your assets and ultimately give us better insight so that it’s not just the 3-5 high value assets that we are continuously monitoring. Now that we have a lower cost IoT option, it gives us an easier way to have a much more broad coverage of secondary tertiary components that could ultimately affect our entire production line. I definitely learned a ton today, Rob, about predictive maintenance, the future of IoT. I’m curious from you, what’s something you wish more people knew about in the maintenance and reliability industry?
16:46 RR: Even though we’ve been talking about the need to retrofit and using newer sensors on legacy equipment, that shouldn’t block you because people should be looking into the data that’s already available within their network and trying to exploit that. And just thinking about taking steps into this, a lot of cases it doesn’t need to be a massive investment and I talk to people about the ability to exploit capability that they have and use that to drive the business case for future enhancements where you might want wider coverage.
17:22 RC: Any other resources that you find yourself still learning from? Where do you go for new ideas?
17:28 RR: Because of my background within aviation, I’m always desperate and hungry for information within the industrial sector, so there’s a lot of great online reading material from various, like the automotive sector, and you find yourself reading Mining Weekly, this is quite bizarre but these are excellent sources. But also there’s a sort of community that we have, we produce a lot of white papers and I love reading those type of things from other companies, especially some of the hardware providers that are out there as well with their new innovations and sensing.
18:01 RC: Thank you so much, Rob. How can all of our listeners connect with you, follow you along your journey?
18:08 RR: Yeah, so you’ll find me on LinkedIn. I’ll keep people up to date there. We run regular webinars and do blog posts at Senseye. Visit our website. If you’re interested, sign up for a demo and then we’ll let you have a look at the app.
18:26 RC: Awesome. Thank you so much again, Rob, for joining us and thank you to all of our listeners for tuning in to today’s Masterminds and Maintenance. My name is Ryan Chan. I’m the CEO and founder of UpKeep, again. You can reach me and contact me either on LinkedIn or also directly at [email protected] Until next time. Thank you so much, Rob. Have a great rest of the day.
18:45 RR: Bye.
Join the Masterminds in Maintenance Podcast!
Are you an industry leader in the fields of maintenance and reliability? We want to hear from you!
Please sign up to be a featured guest on our podcast here!
Stay tuned for more inspiring guests to come in future episodes!