Acoustic Vibration Detection for Predictive Maintenance

- Thank you all for joining today's webinar, Good Vibes, Acoustic Vibration Detection for Predictive Maintenance. My name is Vatche Arabian, application expert for Boston Dynamics. We've got a packed webinar today, but before we get started, I just want to run through some quick housekeeping. We'll be sending out a recording of today's presentation following the webinar. If you have any questions during the webinar, please use the Q&A button, and we'll be going through those towards the end of the presentation. Let's kick things off with some intros. Bob, you want to start us off? - Sure. Thanks, everyone, for joining us today. My name is Bob Ochiai. I'm a product manager for Spot. And I'm focused on the industrial verticals, and also work closely with partners like Fluke. So I've had a chance to talk to many of our customers in that and excited to talk more about what we're doing with Spot there. With that, I'll hand it over to Martin. - Hi, everybody. So my name is Martin Budweg. I'm a process instrument. I'm the global product manager for fixed imaging portfolio, so mostly acoustic and thermal imaging here. And I'm located in Berlin, Germany. And working more than 10 years now in Fluke and the industrial application space. - Awesome. Right. Let's get started. Spot is being used today by hundreds of customers around the world for industrial inspection. Walking around your facilities, collecting data, and mapping performance trends via our desktop software, Orbit. As we speak to customers about the types of inspection they were looking to automate, overwhelmingly we heard vibration, which is why we're here today. If you're new to Spot for industrial inspection, I'm going to give you a really quick overview. Spot moves around your facility collecting real time data on your legacy equipment. It captures this information with onboard sensors in a structured and repeatable way, providing you with clean data that helps you predict failures early. This frees your staff to focus on high value tasks, with a robot doing the dull and dirty work of data collection and your team acting on the results. And finally, a lot of the area Spot can work in are dangerous, hard to reach, or may need to be de-energized for inspection by people. Spot can enter these areas for inspection more frequently, all while your people stay out of harm's way. On their journey to scale, many of our customers started with thermal inspections, and proved that Spot can safely and efficiently operate in their facility. They validated their site-specific ROI that we help them calculate, and then scale the number of inspection points as we added additional inspection modalities. As we reach the end of 2024, we can proudly say that we're seeing customers in total perform over 400,000 inspections a quarter. Today, you can use Spot to perform visual inspection of gauges, collect thermal data, detect leaks using the fluke SV600, and now vibration. I'm going to pass it over to Bob to speak more about vibration detection with Spot. - Thanks, Vatche. In talking to a lot of our customers in the industrial space, the biggest thing that really stood out to me was the challenges that they have to overcome to efficiently and effectively run their facilities. And this really existed in many verticals, where talking to customers in a place like oil and gas, to utilities, all the way to manufacturing as well. And the message that I keep hearing is that everything has to go right for them to effectively run their business. And so in order to do that, they're really challenged to do that effectively with the resources at hand. And as many of you already-- probably are doing, you have maintenance and reliability strategies and programs around some of the high risk and critical areas, and are always looking for solutions to help with that. The interesting thing that's been happening and why we're here today is that Spot is part of that strategy with a lot of our customers. And so by mobilizing a lot of different sensing modalities, we're able to bring sensing to parts of the plant that may be in a blind spot right now, because instrumentation is an expensive thing to perform. And so in working with Fluke, we started with things like the air leak detection, and from our side, we're also building things like thermal inspections. We were looking for other sensing modalities that can really bring value to our customers, and like what Vatche mentioned, some of the vibration techniques that you're installing today was something that was of interest to a lot of our customers. And so if you go to the next slide, Vatche, we started the partnership with Fluke in performing air leak detection with their SV600 sensor. This is an ultrasonic payload that can listen to noise that's being generated beyond our human abilities for listening. And by doing that, we can actually find things before we can intuitively detect them. And so we're kind of using that same kind of technique for things like mechanical inspections. And so today, Martin is going to get into more detail here, but we're going to be going into the details of what this new capability is and how it works. But before I do that, I just wanted to mention-- Vatche, next slide, please, that as I mentioned, the biggest thing that Spot is performing today for a lot of our customers is the ability to sense things before they become problems, and by putting this variety and suite of sensors on top of Spot, we have an opportunity to do that. And what this opportunity gives us is the ability, again, to detect problems early, so that we can reduce the risk of unscheduled or unplanned downtime to things like making really data-driven decisions, based on not just things that you know about, but also things that Spot can catch and detect for you. So to get into the details of this new inspection modality, I'm going to pass the baton over to Martin, and he can, yeah, provide a deeper dive into the technology here. - OK. Yeah. So just giving us a short intro about Fluke. Who doesn't know Fluke. Fluke is founded in 1948, so it's 76 years old. Last year, we had our 75 celebration. So the Fluke corporation is the world leader in compact professional electronic test tools and software for measuring and condition monitoring. It is used by, let's say, a lot of our customers are technicians, engineers, electricians, maintenance process and quality managers and metrologists here. They are troubleshooting with our instruments here. And so with our, let's say, product series, we have more than 1,000 here, not just acoustic imaging here, also thermal imaging, power quality and a lot of calibration products here. So we are worldwide spread out with more than 1,000 patents here. And also, Fluke is very active in training sessions and also posting about the application. So Fluke is leader and expert in industrial application. So today, next slide, we want to talk about the acoustic technology. So Fluke brought out 2019 a handheld acoustic imager series, very successful for leak detection. So with this handheld tool, let's say up to 20% of energy costs could be reduced to finding mostly compact air leaks through then the years. Of course, some other gases also could be leak seen. So based on this one, we got asked by our customer if this can be put on a tripod or could be mounted on a machine, because a lot of stuff should be independently done and not, let's say, with a person behind. So that's why the fixed sensor-- the fixed sensor was developed. So that's the SV600 we are talking about today. The sensor is not just developed for Spot, it was originally developed really for fixed installation, for autonomous inspection, for process control. So it hits all the standards here and also the protocols and fieldbus communication here. You can set up events and alarming here to continuous monitor 24/7 and for maintenance control or for product control, for example. Then as Vatche was saying, I think three years ago now we came in contact here, because we thought it was a perfect instrument also to install on a robot, and robotics autonomous inspection is increasing a lot. So with various projects, we developed the payload, the 600 payload here to worry, to actually independently mostly leak detection first, because the sensor is done for it, because we just can say that the robot is for the sensor, a PLC as you have in a fixed environment. And it was very successful leak detection, and I want to give a bit overview about the acoustic imaging itself. So what is acoustic? So acoustic images. So sound waves are pressure waves, actually. So it needs a medium here, and also we are non-contact technology here, so our medium is definitely the ear, because we're looking through the ear, so what is happening here when a sound occurs? It put out a pressure waves. So the articles in the ear get compressed and uncompressed. And as you more, you get this compressed air and as high as the amplitude, also this is related then definitely to the volume here. So this is what you see on the left side. If somebody is whispering, then the amplitudes are not that high. And if it's very loud, then the amplitude is high. So it means like, the higher the amplitude, the more the articles are compressed. So this is how it moves, let's say, through a medium. Of course, acoustic waves are not just in air, but for us, it is the medium at the moment because the sensor is a non-contact. Acoustic imaging is a pretty new technology for industrial applications, and so just four or five years old. It is well known with sniffers, so probably have to say the beamforming technology is pretty new. So coming now to the next slide here. So regarding the amplitudes here, everybody knows this. So when we are talking about the conversation, so we are between 55 and 60 dB. As I'm saying, the acoustic is measured in decibels, so the volume here, this is logarithm. And as louder it gets, more energy is inside there. And potentially, I put this examples here. We know if you hear a helicopter, for example, a jet engine, it could even hurt your ears. And this is how much energy is inside there. And this is what we are looking for. But this is in the hearable range. So when we come to the next slide, I want to give you an overview about the frequencies, which is the second part, which is important. So we're talking-- this was the audible range here. So the human is hearing at about 20 hertz to 20 kilohertz. This is where we are in. So we are used to this one. We are used to this frequency ranges. So all our life, we are learning how it sounds like and how we act with sound signatures. So I'm just saying, let's say if you're walking through a court and somebody is dropping a bottle behind you, you know it's a drop and who's breaking. You learn this in your life. And this is now this information, the sound signatures now, we need to give, let's say, to the instrument to make it independently. But there is much more than in the audible ranges, especially in the industry. We have a lot of frequencies in the ultrasonic range, which is above 20 kilohertz. And so as I'm saying, the acoustic imaging, the beamforming technology mostly starts with typical leaks at 40 kilohertz. So this is where a small leak when it breaks, it emits the sound. And this is what gets easily detected, let's say, in this frequency range. And then we have another part, mostly partial discharges from interesting here. It starts at 32 kilohertz and goes up to higher frequencies here. And now what we want to focus today, we want to focus in the so-called industrial ranges, so this includes the mechanical sounds and focusing more or less here on bearings today. So these are about one to 28, sometimes also a bit higher, especially when I'm coming to this one. So mechanical inspection, why? Today, 59% is not inspected here, if we talk about, especially about bearings. So if you're looking into the inspection today, it's all analytics, fixed sensors, handheld and human sensing. Human sensing is still pretty big. And so today it's done, let's say with your ears and your eyes, sometimes with your nose, depends on-- its contact temperature control, non-contact, so thermography contact vibration and airborne ultrasound. So this is used today. So there are various technologies today, but still 59 is not inspected, and there is a lot of potential for failures here. And what we are doing now with the acoustic beamforming imaging technology-- next slide. So we are adding another technology to this one. It's called-- it's acoustic imaging. So because it's easy, because it's non-contact, it's user safe because you can make it from away. If you're handheld with a robot, you can go into areas actually, also, where it's not safe, or you can mount it on a fixed side for 24/7 monitoring, it makes it safe. It's accessible to measurement point is pretty high. It's easy to use. And another very important point is with the DIPF Curve, the cheaper of what Bob was saying is. We bring it up from the reaction to the predictive or to the predictive zone. So we want to see now with acoustic imaging earlier potential failure. So with this technology now, we want to help our customer earlier, let's say, to fix it, or earlier, let's say, or later to start maintenance cycles, and also to easy up that work. So the acoustic imaging technology for mechanical is an add on to the other tools here. A bit about acoustic imaging. So what is acoustic imaging or what are we doing here. So we have an acoustic camera which use the beamforming technology. This is what you see there in the gray box. So this is the SV6000. It has 64 microphones in a beamforming area array. So what does it also mean? So every array mostly has the same distance to its neighbor. And this helps us, from the beamforming perspective, to increase the signal to noise ratio. So everything what we are doing is to increase the signal to noise ratio. So if we have a sound source here, for example, a four bearing or a leak, so it emits a sound wave, a pressure wave, so a sound signature. So this reaches the sensor. So the sensor-- so this pressure waves go through all the mic, 64 microphones, and we have a time signal, so we know exactly what time the wave enters the sensor and exits the sensor, and also what time every microphones hear. And with the beamforming era, then, what we can do is we can do through the timing, the travel time correction and we do filtering and process the data. So we gain and stack it. What does it mean? So we get out the noise and we get out the signal. And because we have the geometry and we know we have x, y, z, so the distance and let's say left and right coordinates, we know exactly where it comes from. So we can tell our customer, it comes from the upper left, from the upper right with a frequency value and also with a sound pressure means with the dB value. So the sensor can, at the moment, record up to frequencies up to 54khz. And so this helps the customer then, especially in the ultrasonic range here to locate problems where we cannot hear. That's the main point here. So as I'm saying, we are looking to the mechanical part here, which is not the new part in. So before we did leak detection. So, now we want to help our customers, especially here with bearings, because bearings can be very long, miles and/or kilometers long. And let's say sometimes a very work intensive to monitor them, to take the data and to get a result out of this one. And besides the other technologies now and including with the payload on Spot, we want to help now the customer to collect this data and build up a database to earlier detect a mechanical issues or mechanical sound signatures here. So what we are doing is actually we are looking from the sound perspective in the second and third and the fourth harmonics. This is based on the physics and based on, let's say, on different velocities of the acoustic wavelength. So what we are mostly interested in the ultrasonic ranges here, what you cannot hear in with our ears. So it's mostly up to 28 and up frequencies here. So normally, let's say a bearing is between five and 12 kilohertz if you hear it. So you put that frequency in. And then if you hear something already in the ultrasonic ranges here in the second and third harmonics, this could give you a hint that there is a problem. Also, the sensor gives you a possibility to adjust the signal to noise ratio and also say, OK, when is it interesting for me when it comes out of the noise, and this is important here. It is not an ad hoc technology, let's say, measurement like leak detection, because the crack just occurs and then it emits a frequency mostly at 40 kilohertz here, mostly in the mechanical ranges here, it builds up, let's say, the problem. So with Spot now, also with the A 100. So you build up a database, you monitor your bearings or your mechanical equipment and you build up-- you see then also that the signal is changing. Let's say it comes out of the signal to noise ratio and then you get an alarm from the system here from Orbit or also here from the handheld system. And this is important also, the difference to a leak detection. Yes, of course, you will be able to hear something immediately, but this, I would say, it's a one hit, because your engineers knows mostly when your bearing or the technical equipment does not sound good, and if you already hear it, then it's too late. You want to hear it before, in the ultrasonic range. And this is what you want to build up here with Spot and Orbit. So Spot now enables you really to collect the data from-- with the SV6000 and to build up the database and on the upper right, on the lower right, you see actually, so the signal comes out, and then you have a threshold and alarming giving you which automatically gives you then a hint you have to do something here. So this will enable your maintenance people, less data collection and more really going into the problem and see the problem here. And so next slide. So with the system now, we're talking here about imaging is-- Fluke gives you the complete solution together with Boston Dynamics. So it starts mostly on the lower end side with the handheld, with your engineers go into the factory collecting the data, can make it visible. But this is a very heavy workload. So we are now bringing it to the next step with the autonomous inspection on Spot. So Spot takes the data here, and so you can use your maintenance worker for more experienced work. And so collects data. And then if you have a problem, you get an alarm through the Orbit system. And also if you can put it 24/7 system in here for fixed installed, 24/7 monitoring in your critical areas here when you know I have an equipment there, I need to monitor, get 24/7 monitored, you can set that up with alarm settings, thresholds and so on. And also this gets reported then into your data management system. And then your maintenance engineer gets the information back. OK, I have to go there, because you know where it is, and have to deeper look into this one, or repair it, or have to do, for example, maintenance work, for example, for bearings lubrication or something else. - Yeah. Thank you so much, Martin. That was a great deep dive into what we're doing together with the SV6000 and really excited to continue that work together. Before we wrap up, I just wanted to give our guests here kind of an overview of how we fit into part of your strategies of digital transformation going forward. And so as both Vatche and Martin mentioned, our customers today are using Spot to really offset some of the needs to collect data from your site. And a lot of our customers are starting with one robot on site, but some have multiple robots on site, and they all communicate with a platform called Orbits. This is a Boston Dynamics platform which is acting as a fleet management service, as well as a data management device as well. And that is the system that all of the data that all of these robots collect ultimately end up in, and then can get digested into some of your other systems as well. Next slide, please. And so Orbit is really the window into Spot's world as we talk about here. The day to day is really Spot autonomously performing these inspections, and you coming in to look at the data that Spot collects. And so we've made a lot of enhancements into our system here so that we're more readily able to consume the information that all of the Spot's on site are doing. And furthermore, this is kind of again, the place where a lot of our customers interact with the robot. So you can remotely operate spot within Orbit, as well as plan out the autonomous mission schedules as well, and then ultimately consume the data in this platform. And just to wrap it up, again, we hit on a lot of these points today. And there's actually another webinar that goes into a lot of the way our customers are really getting value out of the system. But I wanted to emphasize that, together with the challenges that's involved in upkeeping a facility, really, our Spot's coming into place as a solution for things like sustainability goals that go into the kind of-- separately from the plant upkeep. But also the biggest message that we're hearing is Spot is really helping some of the labor shortages that our customers go through today. We're often hearing about a lot of the tribal knowledge that every company is losing every year because of these long term engineers and technicians on site are retiring and not re-entering the workforce. And so while a lot of our customers are able to-- the best case scenario here is to keep that headcount the same, ultimately, they're losing a lot of that know how that is retiring. And so using human labor, human resources that you have in the most effective ways is what our customers want to do. And some of these observational inspections that Spot can help with is really helping offset those needs. And so together with partners like Fluke, we're trying to help fill that gap with new technologies into the space as an option for you to consider for your gaps there. And so yeah, we're excited to introduce this new capability with Spot. This is something that we're going to continuously do to bring in more capabilities with Spot. And thank you again for joining us today to have this discussion. So turn it back over to Vatche. - I'm excited, because we have so many questions, so I hope you guys are ready. - We're ready. - I'm going to kick things off with a question that came up in a couple different ways. It was asked in a couple of different ways. But the basic question was around the capabilities of detection range regarding distance with the SV600 for this inspection. Is there a best practice? What do you suggest, what do you recommend, Martin, people consider when they're thinking about distance to target with SV600? - Well, so we know we can-- we can, with leak detection, we can see something which is 200 meters, which is 600 feet away. But of course, the resolution goes down. So let's say normally, from the industry perspective, I would say everything, what is about two meters, what is it, six feet, and I would say, 10 to 15 meters, so up to 30 to 45 feet. This is, for the industrial range, pretty good, because then you're not losing, let's say, most of the resolution. Also, you don't have the damping factor in the frequency and also in, let's say, sound pressure-- or less, let's say, because it takes-- what is it? Square R, it goes down with the damping factor. But the thing is like, from my experience is mostly-- most of the applications, you are pretty close because you go there, and not there more than I would say larger than 10 meters to 30 feet. So this is what I see, everything there, you will be safe. And if somebody wants to do everything 200 meters away, yeah, you can see it's something, but I think from the analytics perspective, this is critical. This is to give you a good overview. For example, when you go there with a handheld, you see something there, but then you go closer, because then you have a better resolution. So that's the best practice for me. - Great. Someone asked, can Spot be loaded with vibration ultrasonic thermal vision to replace expensive condition monitoring sensors on large plants. If so, do you have any examples. So I think going back a couple slides where you had the ecosystem of workflow, Martin, and we've got a lot of other examples. Spot doesn't necessarily replace other sensors. It works within an ecosystem of sensors. It's really the person walking around, taking rounds, right. We hear from our customers that they don't have time to go do those manual inspections. So what the robot's doing is rather than getting fixed sensors for all these slow to fail assets, it's a suite of sensors that goes to those assets and collects that data in a nutshell. And that's really industrial application for Spot. That's the story. We've got plenty of examples and great customer case studies, most recently with Michelin. So I highly recommend checking out how customers are pulling in Spot for this application, how they're using it. Shifting to a question that was pre-submitted. Any implementation bottlenecks? I think this includes SV600 and/or Spot. So I don't know, Bob, maybe you can go, you can kick it off and Martin can talk about SV600? - Sure. Yeah. Depends on, I guess, the definition of bottleneck here. But there is, I guess, a journey to take to get Spot into your facilities. First of all, there's a lot to consider around the mobility of Spot. Where does Spot going to go, where is it going to perform the inspections? And we actually have a group of really good sales engineers that scopes out a project with you, walks around the site with you, and to figure out where Spot can really make an impact on your facilities. But beyond the inspections, there's also other things to consider, like who's going to be managing and maintaining Spot, as well as what IT requirements are there, what security implications are there. So there are few things like that that is an early conversation that we like to have so that they're not roadblocks to success when Spot actually gets to your site after you place the order. And so that's something that we've put a lot of work and effort to really understand what our customers need ahead of time so that again, we tackle those problems together early and then work together before Spot is actually fully deployed and perform these autonomous rounds. - Just adding to what Bob is saying is for the acoustic imaging, I mean, it's also a training thing, because if we want to train you and give you all the information together here with Boston Dynamics, how you do a good acoustic measurement. It's a bit of learning curve. So how you position, how you do this, if you use area of interest or not, if you move a bit to identify reflections, how you understand a bit about the frequencies here, because it is about-- also for the engineers, if you are not a acoustic expert, we are used to our frequency range here, but we are looking into ranges we are not hearing, too. So this is a bit of a learning curve, also, this is I will say. Also, how you set up the thresholds, how you set the alarm things up, this is a bit of training thing. But if your engineers getting used to it, and this is the same actually also for other sensors, then I think it gets easier and easier, it will help a lot. So in the beginning, of course, we're having a FAQs, we have documents. Of course, we're doing also trainings here also together with Boston Dynamics to train our end customers, really, how is the best practice to use acoustic imaging. - Great. There's a question, has there any-- has there been any traction availability for specific region of interest selection on the acoustic system like the thermal camera system offers. Bob, do you want to take that one? - Yeah. So in developing this with Fluke, we did find that there are situations where things like background noise or other noise in the field of view of the SV600 impacted the readings that we were getting back. And so in response to that, the teams here worked together to build out being able to draw a region of interest within the field of view of the SV600, to then only use that region for the analysis of the inspection. This is something that we are releasing as part of the acoustic mechanical inspection, but also for the air leak detection as well. So in the new software, you're going to be able to get the region of interest for both of these inspection types. And furthermore, it is something that you can draw through the tablet as you're recording these inspections. So the folks who are familiar with it, you can record these inspection missions through a tablet controller that ships with Spot, and there's an interaction to draw that, but Orbit also allows you to edit and add those in after the fact as well. - Great. There's been a couple questions, both pre-submitted and in the chat about how do you calculate the business case for a Spot system or for vibration, et cetera. I know, Fluke, you've got some great resources for calculating our return on investment on the different acoustic technologies, but we do the same with Spot. We are more than happy to work with you and walk you through our calculator as part of a conversation. So there's also some great resources on our website of how customers are thinking about ROI when it comes to automating inspection with Spot. So definitely invite you to check that out as well as the webinar we did, I think a few months ago. A lot of material out there, and we're more than happy to dive deeper with your specific numbers to create something really custom for you as you think about the business case for Spot. This one's a great one. I was wondering if Fluke offers training courses, charts to understand acoustic sound frequency results. - That's a good point. Yes, somehow through our channels. But this is then definitely for the fixed sensor. I think with this solution, this is a combination between Boston Dynamics and Fluke, so this is what we're offering. Let's say if we're talking about the payload, then this goes here together with Boston Dynamics. This is either done by, I think, by Boston Dynamics and us. So we just do-- clear the sensor, and also, what about acoustic imaging, and Boston Dynamics is doing then all the integration with Orbit. But this works all together. So for acoustic imaging itself, let's say for the A100 or for the SV600 as a single sensor, yes, we provide through our services and also through our channel partners. - And that's a good chance for us to plug our training services as well. We offer really hands-on training. We come to you or you can come to our training center here. We teach you the basics of Spot and how to leverage all the payloads to get the results that you're looking for. This is a good question, Bob, this one's for you. Is it possible to integrate the Fluke SV600 onto a Spot robot that's already equipped with an arm payload? - Yeah, that's a great question. So there are a few different positions where you can mount the SV600 into. Of course, the considerations to make here are how much space you have left in the back of the robot with the arm in place, but we should be able to mount it above the EAP2 if you have one, that's one of the optional configurations. So it'll be facing backwards, but that'll fit right alongside the AV spot arm. - Great. There is a question that I don't totally understand, but how can the camera be used in general facilities, not necessarily industrial. Obviously, if you're doing thermal or acoustic inspections in a non-industrial facility, we'd love to learn more about that use case and we'd be more than happy to talk to you. So if you want to raise your hand at the end of the webinar, we'd love to hear more about how you plan on using it or what you're thinking about. This is a good question. How do you trend the measurement over time, and what metrics do you use to determine whether a machine needs repair? Martin, you want to take that one? - Well, you do it pretty easily. You print it, and you can use the SPL, let's say, the threshold when you think this is a good, let's say, a signal which is differentiate from the noise. Let's say normally, I'd say bearing is doing specific noises, but this sounds normal. That's a normal noise, I would say. And you correct it, and over time, then you do some measurements and then you see that the sound signature is changing. And this gives you definitely a message something is wrong. We can at the moment not directly tell you if you have to repair it or not, you have to check it. This is definitely hopefully coming in the future of more AI coming in. Let's say if you lubricate it and so on, because there are a lot of physics about bearings, different states with different frequencies here, this is what we want to use. So today, it's mostly really, we will tell you something is wrong. And let's say especially in big factories or in big conveyor belts, because Spot can easily collect the data, and based on your way path you defined, you know exactly where your biggest problems are or where you have already problems, and you take into account this and you have to go there very often back with your people. So this saves you the money. Let's say this saves you the resources, also. Or you can use your experience resources definitely for the analytics and other stuff. We are working also together with BD to analyze this if for the next steps, let's say if you're here there, do you have to lubricate this or not. But keep in mind a bearing is not just a bearing, a motor is not just a motor, you have a lot of factors here. It could be covered, so the frequency is changing, the damping factor is there, or you have a metal bearing, or a ceramic bearing, or plastic bearing. So you have a lot of different sounds. So I would say this tool helps you at the moment a lot to say this is a good or a bad sound, and we have quite good, I think, experience with this already. - If we already use acoustic imaging at our site, how easy would the conveyor inspection be to set up? Bob, do you want to take a stab at that one? - Sure. Yeah. Martin mentioned this earlier, and I think a lot of our viewers here today are already installing some kind of vibration or maybe acoustic imaging already in a handheld form. Where Spot fits in is really turning this into a more frequent cadence for you, where it's collecting data across time with less-- with more data points and less blind spots to having a time-based data set that you can look into. And so the way we're imagining our customers using this, and actually, a lot of our customers are already starting to use it this way, is that they have handheld sensors that only-- they might sit on the shelf or they only use when things go wrong, where that might be fine because Spot is now performing these inspections across time in more spots within your facility, and can give you a heads up when it starts noticing something sounds different. And so that's where the combination between this autonomous system with more frequent inspection points pave the way for manual inspections being done through these handheld devices, which the handheld devices and the operator behind it can use their intuition, their experience, their knowledge about the facility to do a deep dive into diagnosing the problem further, which is something that the robot is not really a good fit for today. And so I think like Martin mentioned, it's going to be really a situation where you're doing both, but maybe at different times throughout the life cycle of the facility. - Yeah, there's been a couple questions on work order generation and connectivity to facilities management systems and ticketing systems. Bob, you want to talk a little bit about how that works today with Orbit? - Yeah. So as many of you on the call here probably already have, and utilize EM systems and other plant management systems that give you a melting pot of all that data, Spot is starting to connect to a lot of those systems as well. As far as the journey goes, Spot has-- it's been designed as a standalone system so that those weren't requirements going into deploying a system. And so Orbit gives you a certain level of visibility into what the robot's doing, but not necessarily what the entire plant is maintaining and managing. But what we've done over the last couple releases is actually built systems that are able to infer external systems that you already have to say, hey, there was an alert, inspection alert that Spot caught. And so we call the system Webhooks, which is a externally facing messaging system from Orbit that can then communicate to these other systems, say, hey, you should take a look at this. And so when that happens, we also have an API that you can pull data from Orbit to then ingest into the system so that you can collect data as needed, but also, you can utilize that for really, a deeply connected system as well. So we do have flexibility in how you utilize the API that's built into Orbit depending on how you want these systems to communicate together. - Great. How does Spot deal with humans and traffic? Like cars, bikes, forklifts, if expected to navigate in a busy factory setting? - Yeah. I've only seen a few bikes in facilities, but I have seen some. Yeah, this is actually a really great question, because when we initially released Spot, Spot was this mobile walking sensor. And so some of the considerations that we had after we launched Spot and brought it into real facilities or customers facilities is that there's still a lot of people walking around the site, and some of the initial challenges that we had was trust in Spot, right. So Spot previously, initially, looked at people the same way as they would look at any other obstacle that it encounters. It didn't recognize that it was anything more than something it couldn't walk over or walk past. So over the last release, we actually launched a system called the entity detection system, which is able to recognize people and moving objects like cars and forklifts and things like that as things that the robot has to be careful around. And so some of the behaviors that came out of that is as simple as the robot recognizing you, pausing, making sure you walk past the robot and then continuing on its autonomous mission path. And so we're incorporating a lot of these things into the robot so that it can really merge well with the existing facilities that our customers operate in. - Great. When implementing Spot, is the identification of findings automated? Is it AI or is the client responsible to develop that? - Yeah. Depends on, like Martin said, I think the vision here is a really full system where it knows a lot about the plant and it takes in more data and uses that to make determinations of what the operators or the maintenance and reliability folks should do. We're still at the first entry point there by collecting more data. And Spot's really good at that, because it can have a diverse set of sensing that it can move around the site as needed. And so as we go forward into this, we are looking at opportunities to incorporate some of those technologies into the system so that it's not just saying, hey, I found a very specific problem based on this very specific inspection that you asked me to do. It's looking at data across time, it's looking at other data points, it's looking at other plant conditions. So that's the future that we're going towards, but we're just getting started with that journey. - And just to add to that, like today, you saw the Michelin video, or if you haven't seen the Michelin video, spoiler, they generate 72 work orders with their system with Spot. And today, the way it works is Spot reports that a threshold has been crossed that they have set, knowing the baseline for their equipment, and then that equipment gets scheduled for downtime or a repair or whatever. And what ends up happening is a person goes and looks deeper and diagnoses the issue. The end state, the future state that we're empowering people with is you're collecting all this data, and to Bob's point, you're going to have this historical data for all of this equipment so that you can better map out and better predict and get to that future vision of truly predictive maintenance. What are the biggest challenges in integrating vibration analysis systems with existing equipment, particularly for companies with older machinery? I think, Bob, you want to take that one? - Sure, and Martin, feel free to jump in, too. Yeah. So this is an interesting question, because while futuristic and amazing Spot is, really Spot is there to help with a lot of these legacy factories that were designed with different things in mind. And so the ability to get Spot into these places is really the opportunity that we've had to help our customers operate these facilities better. And so thinking about the age of the facility is not necessarily the way to think about it. It's really, what can a mobile sensing device like Spot do with any of my facilities. And so thinking about this specific question, it really depends on how much inspections that you're performing today, maybe through your technicians, and also how much sensing you're doing with installed sensors and seeing where the blind spots are today, and that's where Spot fits into your day to day picture. - Just can I add for the, let's say, for the acoustic imager. I mean, I showed the slide. It's an add on to existing technology, because we have a lot of, let's say, white space there. As I said, 59% is not inspected or it's human sensing, also. So there's exactly we want to add on. So to get the white space, let's call it smaller, to help the end customer. So it will not replace, potentially, other technologies. It will help to understand the failure better and earlier. Yes, potentially in the future could be that you don't need some of these anymore, but would not really see this at the moment because there is a lot of things which is working in parallel. You're just adding another data point. And if you can feed this then into your fleet management system, this is another piece of puzzle to give you a better and more information. - Great. There's just a follow up. Does Spot work with SAP? Yes, we've had customers that have connected their Orbit instance with their SAP work order system. Just to wrap things up, I guess, how does-- if someone wants to obtain a license for this feature on Spot, how do they go do it? And I think that this answers a follow up question, too, of how can I get this and how can I add this inspection modality. So the best way is to please reach out. Please raise a hand on the follow up to this webinar and let us know that this is something you're interested in getting or adding to your-- either Spot deployment or you're starting from scratch. We're more than happy to talk to all. So please raise a hand in that follow up or use the contact sales form on our website. We'd be more than happy to talk more about this and other ways you can use Spot. So thank you for joining us today. Thank you to Bob and Martin. Thank you to our presenters. And we'll see you next time. - Thank you. - Thank you.