Fluke and Boston Dynamics have developed a partnership where the SV600 Fixed Acoustic Imager payload from Fluke Process Instruments is connected with the Boston Dynamics Spot robot, enabling remote and/or automated plant inspection routes. Plant Services Editor in Chief Thomas Wilk spoke about this technology partnership with Martin Budweg, Product Marketing Manager for Fluke Process Instruments, and Tim Dykstra, the Director of Strategic Partnerships and Channel Sales for Boston Dynamics.
PS: I’ve followed both your companies for so long. Plant Services and Fluke have gone way back in terms of working together to promote new technologies and new innovations for our readers. This is the first time, Tim, that I think I've talked to someone from Boston Dynamics, although I've seen the videos of your robots as they've grown and evolved over time.
Specifically today, we're here to talk about the partnership, where the SV600 Fixed Acoustic Imager payload from Fluke Process Instruments is connected with the Boston Dynamics Spot robot. We've already got a link to the video in our podcast comments, guys, so, I'll just tell readers right now, if you want to see this robot in action and the instruments in action, go ahead and click the video, but not until we're done talking, okay?
Tim and Martin, tell us a little bit about yourselves, and then we'll talk about the technology.
MB: My name is Martin Budweg. I'm located in Berlin, in Germany. So, I'm the Product Marketing Manager or Product Manager for the Fixed Acoustic Portfolio, and also responsible for the SV600. I've worked for 10½ years at Fluke or Fluke Process Instruments, and my background is physics. I started physics, have a PhD in seismology, geophysics. So, background is a bit acoustic also in the past, but working really for Sensor Technology now for more than 20 years.
I started the project here with Fluke about two years ago. And so, because acoustic was new for us, because our big brother, the industrial group has this portable ii900 imager, we found out that customers needed not just a handheld solution, but also a fixed solution. That's why we created the SV600with our partner Sorama, and introduced this last year in October. Since then we are ramping up and starting to bring this on the market.
I think it's a year now Tim, that we closely started our cooperation, so it came in that we found out there is actually a new vertical for us. We call it “mount and move”. Probably, Tim calls it a bit different. So, it's pretty interesting for us because it is new. It gives us a lot of possibilities now to go differently into verticals, where we could not go before with a fixed solution.
PS: Tim, tell us about how you got into Boston Dynamics, and what you're working on.
TD: So I'm the director of Strategic Partnerships here at Boston Dynamics, predominantly focused on the Spot side of our business. I came to Boston Dynamics a little over two years ago, almost 2½ years ago, about the time when we started to bring Spot to market. It's still a relatively new product for us. Prior to coming into Boston Dynamics, I did a lot of work with robotic forklifts, as well as other six-axis robots. I have been in the robotics business for about 15 years.
With Boston Dynamics and Spot, it's a relatively new product, only been on the market for about two years, but Boston Dynamics has been around for a while. As many people know from our YouTube videos, we focus on legged robotics, and really the reason for that is a problem with mobile robots in the past has been they struggle in areas that are built for humans. A lot of the plants around the world are built for humans to get around, and so a wheeled, or a tracked robot, or even a drone may have difficulty getting around those areas, and that's really where Spot excels. As Martin said, people are putting sensors on top of Spot and it walks around a facility to inspect assets.
PS: It's funny, Plant Services has covered a whole lot of portable condition monitoring tools and technologies, but this takes it to a whole new level where the robot carries the monitoring device around by itself. Tell us about the collaboration between these two companies. How did we get here, and where does the collaboration stand these days?
TD: Maybe I can start with that and give a little context. As I mentioned, with Spot, I call it a walking IoT device, right? It can get around the factory pretty easily and you can mount any piece of hardware onto it. In the early days, we were focused a lot on thermal and RGB or regular photo information, capturing data on assets, maybe it's a thermal scan of a motor, or looking at an electrical cabinet. That was a big, big focus of ours. On what was probably one of my first site visits during COVID when people started go back on sites, I had a manufacturing customer that said, "Hey, Tim, it'd be great if Spot could carry around this. It is a ii900, this is a Fluke device that my maintenance people walk around with to look for compressed air leaks."
This company was losing about $75,000 a month in compressed air. And they said, "we've purchased this device to go around and monitor leaks, but we lost a few maintenance people, and it's really difficult to find new maintenance people with the labor shortage, so we're not able to walk around with this as much as we want to. It'd be great if Spot could do it for us." So, I said, "Okay, that's interesting."
We looked at this acoustic imager by Fluke. The next week, we had a person in the food and beverage industry come to us and say the same thing: "Hey, I have this ii900 that it'd be great if you could integrate with." The following week, I had a nuclear customer say the same thing: "Hey, we do these inspections inside a containment area. It'd be great to have this device."
So I scrambled and said, "Who do I know at Fluke? Who can I find at Fluke?" I found Sorama, and found Martin at the Fluke team, and this was probably a little bit over a year ago. I think it was in December of 2020 that we started talking, and that's really where this relationship and this partnership formed, and we've been working closely with them. Thankfully, as Martin said, they were starting this fixed sensor, and they said, "Hey, the ii900 is a portable device, a handheld device, but we actually have this fixed sensor that will be perfect to mount on top of Spot." And I'll let Martin talk a little bit more about that.
MB: It's roughly a year ago now, so as Fluke continuously innovates to keep business industry around the globe up and running, and there seems to be new merge between handheld and fixed. This is a classical fixed application because you need to integrate this into a system here, it is definitely a robot, but it is more because you have also the factory behind, you have the people behind, so the protocols behind, you need to do alarming functionality, you need to do reporting functionality. This is the classical fixed sensor, where Fluke Process Instruments definitely has key knowledge in. And so, that's why it came actually through our industry group and Tim, and we already started to develop the fixed sensor, and then to say, okay, let's develop this payload for Spot because we already had some customers who were very interested, in the beverage and food industry, utility industry, power generation, mining is getting pretty big. All the assets where you need to monitor, for example, leaks, partial discharge, or going into areas where humans cannot really go because it is too dangerous, or it's too small, or as Tim is saying, to unload their maintenance workers because they need them for other work.
To collect data, Spot is perfect to use, and we wanted to give Spot an ear, better than a human ear, because we can hear beyond the human. Friendly speaking, a human can hear roughly up to 20 kilohertz (that's for young people, probably not for me anymore). We can hear up into the ultrasonic range, and this is where you can easily find leaks. Also the key point is to make this visible for humans. With Spot then and all the other payloads, you bring all this knowledge together, all this data together. This is a big new industry step, I would say, and I think we see more in the future. We see already that in many cases, customers are interested in these solutions, independent solutions, especially also more and more in the petrochemical industry where again, also safety is a reason. That's why we decided together with Boston Dynamics to go this way and start this collaboration a year ago. We started the combined launches, the combined video material, and we hope we can show more applications.
PS: It's amazing how quickly the acoustic imager technology has caught on just by itself. But then you get an application like this partnership between Fluke and Boston Dynamics to solve the variety of problems you both talked about – limited staffing or staffing shortages, difficult places to go in a plant. If we can focus on safety for a second, I'm curious about this: do you see this as fitting into current safety best practices, or is this sort of breaking new safety ground for industry?
MB: I think it breaks (new ground) because the robot can do more than their human alternate based on specific maintenance work. Of course, we need to go step forward out to fulfill some stuff, especially in the petrochemical industry here, because there we have hazardous environments. Definitely this solution on a robot or here on Spot, because Spot can go through the sites and stay safe, for example in a power plant or nuclear power plant, because we are working out here together with human beings, right? It can go inside and collect data where humans cannot stay long, or it is too small (for people) to go there; Spot is not that small, but it's also not as big as a human. You could also go in the dark, especially in the mining industry where it's also very unsafe sometimes, so they want to monitor for leak detection, for example, they want to send Spot inside there. The maintenance worker is beyond the screen outside, or they can automate (data collection with Spot), so it's definitely safer for human being.
To go in another direction, (Spot) is also to save the product, let's say in a clean room, for example, where normally a human should not go because of all the other stuff we are bringing inside with ourselves. A robot can be a good solution for maintenance control for monitoring for leak detection, because you can make a robot pretty clean, you can keep it clean, right? You don't need to change it, unlike when a human goes in a clean room; a robot can be cleaned up and can be used for a specific time. This is also a safety factor, not for the human but for the product itself.
PS: I'm struck by how precise the robot's movements are. In the videos that we've got, you see the robot legs moving very carefully, and clearly it's designed to avoid obstructions, get over obstructions. For those of our readers who might not have seen the robots from how they work, are these robots guided by humans in a control room? Or are there sensors in the robot that help the robot make self-correcting movements as it moves through a passageway or a hallway?
TD: That's a great question. First, on the smoothness, the walking abilities have really been perfected over the last 28 years as we've been researching developing these legged robotics and so that's a big factor of what plays into how easily they are able to locomote across a facility. But Spot can be driven in a couple of different ways. It can be driven in a manual fashion. And a lot of that is the safety applications that Martin described, going into a nuclear containment area, going into a high voltage substation at a utility plant, or some of these dangerous hazardous areas in a petrochemical facility.
A human can manually control it, but really all they're doing is giving high-level commands like telling Spot go forward, turn right, go to the side, and Spot is taking care of everything else on its own. It's using stereo cameras that are embedded into its body to create a 3D map of the world, and then it can understand, is there an obstacle there? Is there a step there? How do I adjust my gait to move over or around whatever it may be facing in its way? That's how it moves so gracefully over let's say a catwalk or a staircase, as well as avoid obstacles that are in the way.
The other aspect is those same stereo cameras can actually be used to map the environment, or you can put on a LIDAR for extended range and mapping environment. Then you can perform autonomous missions. A lot of our maintenance and reliability folks that are using Spot, that's really the mode that they're using Spot in, where it's performing an autonomous round, and basically doing a rounds and readings type application, or a route-based monitoring application. It's going around collecting that same data over and over, maybe every hour or every day, or every week, and collecting that data on a routine basis, just as a maintenance worker would perform a round and reading. So, there's two different ways to operate it.
Listen to the entire interview
PS: That’s fascinating, literally it has eyes in the back of its head, and in its elbows, and on its knee joints, and other places to help them map the environment. You mentioned that autonomous inspections and mapping the environment. Are we already at the point to looking forward a bit where the robot would memorize its own routes, and use algorithms to adjust this route if it sees obstructions in the way?
TD: That's very much in the capabilities of Spot. If it's walking around a food and beverage plant, and there's a pallet all of a sudden in its pathway, it can dynamically go around that. Even furthermore let's say, construction sites and other industry where Spot's used in, if a wall is all of a sudden put up in its route, it can even say, "Hey, I need to find an alternative way to my destination." And it's very capable of doing that fully autonomously.
PS: That's amazing. A lot of our listeners and our readers love nothing more than a good case study story. Is there a story or two that you both can share about the way that the robot has used the acoustic imager and achieved some real cost savings, perhaps in terms of finding some compressed air leaks?
MB: Sure, we have some applications in the mining industry, they have a lot of compressed leaks, or for the food and beverage industry, they're using also a lot of compressed air. If you know your weak points in your factory or in your plant, in the first step you would configure Spot together with the SV600 and point to the spot, take a picture, and make the analysis. The SV600, because of the beamforming structure and the data they're collecting, you can really locate the leak, you can see it based on the volume on the sound pressure, let's say, so the dB level, and then determine the throughput, how big the leak is. You can create an alarm, which alerts a maintenance worker to go in there and do something to fix that. Then you go to the next critical area and you find another leak or not, then you may find a leak outside this critical defined area. With the camera’s beamforming array, you can really locate it.
This definitely saves money, first let’s say for beverage and food company because I think they are saving a lot of dollars per year if they know where the leaks are. Some of them you cannot hear, you cannot see, but the camera sees the leaks and the ultrasonic range which we cannot hear finds them before they harm you, right? Normally they get bigger in the future, so you have to find them in an early stage and you can then definitely fix them. For natural gas, that's actually more critical because we're not talking about compressed air, we were talking about dangerous gases sometimes here, so the earlier you find them, the better it is for these companies.
You could theoretically also do this autonomously, so you can program it and it goes through, and then at the end you get an alarm out, or you get a report out, and then your maintenance workers afterwards can go there and fix this. Or if it is a serious let's say leak, you probably know it is on a dangerous line, then you get a direct alarm out so you can react.
TD: I can give a few examples as well if you'd like. One of the more recent examples that is a favorite of mine are our mutual partners, Intuitive Robots in France. They were visiting Total Energies for a proof of concept, really a couple-day event, where they were showcasing some of the capabilities of Spot and the SV600 Acoustic Imager. And what was exciting about that is within the first three minutes of recording a mission, an Autowalk mission, which is the autonomous mission, Spot was able to identify a gas leak, as Martin mentioned. That was really exciting to see this technology in just really a short POC-style deployment proved to be beneficial for Total Energies in this case.
We have another example as Martin mentioned in the food and beverage space, and that's way beyond proof of concept at this point. I think it's going on about a 9-month almost 12-month pilot with the SV600, as well as they're doing thermal scans of equipment on a beverage line, on a canning line. Right now they're starting to move into the next phase, and the next phase is to not only identify those leaks or identify those thermal anomalies, but now start to report on them. They're tying it into maintenance systems like IBM Maximo or SAP, so work tickets can automatically be created. Those companies can be most efficient with fixing whatever that problem is, whether it's a thermal issue with a motor or a pump, or maybe an acoustic gas compressed air leak.
Those are two examples where people are deploying successfully and that beverage manufacturer is now looking at, all right, this has been successfully (deployed), I have this canning line, let's move it to the next five or six canning lines in the facility and really start to use this technology across an entire plant.
PS: We've heard from our readers that part of the challenge is getting data – like the data being collected by the imager – into your CMMS to get the work order triggered, and to hear an example where that process is automated is really interesting.
MB: I can see this also. This was also in the beginning of a plan of the different phases we have. So, step by step, we want to put this into the CMMS system, so then the maintenance workers can do something with this one, right, or you get a direct safety alarm. But as Tim was saying, you get the ticket, the workers go out there and can fix it, they know where they are, they probably know also is it a big leak or is it not a big leak. Because of the location, they know how big it is, and where it is, and so now we see more and more they want to really automate this, get it into the system, so the company's workflow will be new here, because in the past, you go through this round, then you collect the data, you download, you analyze this, and then probably a day is already gone. But here you collect the data automatically, autonomously, so then you could do something in the future. This is definitely the plan, especially with the food and beverage company, because there we have really stepped into this direction to get this maintenance process completely automated with the SV600 and Spot.
PS: I've got two more questions. One is sort of light-hearted. I've got a dog named Hunter, and Hunter is not nearly as nimble or dexterous as Spot is. When I get frustrated with him, I'll simply call him “the dog” because I don't want him to hear me call him by his own name in anger, right? I want to make sure that there's some affection there. I'm curious to know when it comes to technology adoption, have you found that plant teams look at the robot plus the acoustic imager and these technologies and say, "Aww, I have to get one of these in my plant. This is really cool and sweet, and it looks like my own dog."
TD: Yeah, we see that a lot, and we call Spot an “it.” It's a tool that is used to help people, really is what it is. But yeah, plants love to take ownership of it. Oftentimes, they have naming contests, name it a cute name for their specific facility, and we see people fall in love with it. What's most exciting to us is at this beverage place, for instance, we saw maintenance workers from other parts of the plant, say, "Hey, I need one of those." And to us more than being cute or more than being a cool piece of technology, that's really exciting because it's actually helping with efficiency, and people are saying, "hey, I need this. This is going to help my day-to-day job," which is really exciting to see.
MB: We see the same, of course, more from the payload perspective, because we give the dog at least the ears, some eyes, but definitely the ears, right? So, the customer can do something with this information, collecting the data and for quality control, for alarming functionality. Of course, mostly we go where a Spot already is, but it comes more and more together because we see more and more customers in all verticals want to see sound, so Fluke makes that visible.
PS: Last question then today, guys. How has this project, in your opinion, pointed the way towards what else is needed in industry? Do you see what might come next in terms of autonomous inspections or application of this technology?
TD: That's a really interesting question, I'd answer it in a couple of ways. One, it points to the collaboration that is needed, right? So Spot, as I described, is just a mobility platform, and it really takes a broad partner ecosystem to make it most useful. The partnership with Fluke has been great, because we are able to give Spot ears, we are able to give it thermal eyes, which is a real huge benefit to our users, and we wouldn't have been able to do that without the collaboration and the partnership with Fluke. That's one thing it points to is the importance of collaboration and partnership, whether it's adding sensors to it, or partnerships with companies like IBM to drive that data directly into Maximo, is really crucial in delivering an end-to-end solution.
The other interesting trend I think it points to is this idea of retirement of a large part of the labor force. There's a lot of expertise that is leaving plants; there's people that they just hear something, you know. I was walking through a plant the other day, and a maintenance guy, older person said to me, "Hey, do you hear that? I can hear something's wrong with this machine." That's something that's difficult to replace, so I think Spot can be a part of that combined with the SV600. It will allow for that expertise to live on at a facility, and our hope is, it also allows these plants to be more efficient. No one wants to just walk around and inspect things. Even a maintenance person, they got into that profession to fix things, their least favorite thing to do is go around and inspect things. It's giving those types of professionals a higher purpose role, and more time to focus on actually fixing equipment, and making the plant more efficient.
MB: So, yeah, I agree with Tim here. As we are committed to improve the safety building and supporting a skilled workforce, alleviating workforce shortages, and advancing sustainability, so Fluke really needs this partnership, right? We need our customers, we need good partners, and Spot here is new, because it's going into a part of industry to make it more safe. And as Tim was saying, Spot is a mobility platform, so Fluke has a lot of sensors so, at the moment, we have definitely given Spot the ear. There is definitely more potential with all the sensors we have in Fluke, and I think this could be a very good, big partnership in the future.
In the end, the customer decides what they want to measure and where they put that in, so more and more data collection goes into more and more CMMS systems. They decide what they're going to do with this data. Is it just reporting? Is it quality control? Is it maintenance? Is it asset monitoring? Or whatever comes in the future, what could be on production side that the robot can help? Because at the moment it’s really more the maintenance and the asset monitoring side. So we are working together to develop the SV600 so that Spot can get operational work and can do something there.