Introduction to Machine Health Monitoring Webinar
In this hour-long video, Peter Woodman, Technical Sales Lead for Everactive, and Applications Engineering Manager Tom Ross discuss the firm’s groundbreaking approach to machine health monitoring through IoT and hardware-enabled SaaS.
PETER WOODMAN: Welcome everybody. Thanks for joining us today. The subject of today’s webinar is machine health monitoring [and] batteryless sensors. [We’re] premiering a new product today. So we’re going to take a little time to go through that, and kind of go through our company as well. It’s an exciting day for us here to formally introduce this to you all. Thanks for making the time to be here with us.
Our agenda for today: We’ll do our introductions, talk a little bit about our company in general. Then we’re going to take a deep dive into that new product, today. We have new hardware here in the room, too. So I’ll switch on some video and show it to you guys in person. We’ll talk through the solution itself. Then next steps: how we get this into your environment and help you guys transform your environments to pervasive sensing for all your motors, pumps, fans, rotating, and vibrating equipment. We do have QA here as part of the Zoom meeting, so if you look inside your Zoom toolbar there, you’ll see a Q&A box. Feel free to type any questions in you have as we go along. Since this is a webinar and there are a bunch of attendees, we’re not going to stop for every question during the presentation, but we’ve reserved time at the end to make sure we get those addressed. So please go ahead and drop those in as you go through.
Okay first off introductions. My name is PETER WOODMAN I’m the Technical Sales Lead here, a Sales Engineer for Everactive. I’m a lucky guy. I got into this company before we launched our first product. It’s been really fun for me to watch all of our products, but specifically this one evolved from just an idea to something that you can hold in your hand. So, super excited to share that with you guys today. My co-presenter is one of Everactive subject matter experts in motors and machines TOM ROSS. Hi, Tom!
TOM ROSS: Hey Peter, good morning. And good morning to everyone. I am recently with Everactive about six or seven months now. It’s been an exciting ride as we launch this product. And I come to Everactive with 25+ years of a variety of reliability and maintenance.
PETER: Awesome. So glad to have you on the team, Tom. And yeah as Tom said, there good morning. He’s working out of the west coast. Tom works out of our California headquarters in Santa Clara. I work out of the Midwest in our Michigan office, in Ann Arbor. That’s the afternoon for me.
But as you can see here in these photos, Tom and I are usually most comfortable out in the field in our customer environments. We’re outdoor cats right? But typically we’d run a premiere product like this at a trade show, or bring them directly to a customer site. Given what’s been going on with Coronavirus I think we all have kind of a different understanding of you know what the ‘new normal’ is around process digitization.
So some of our customers have been thinking about this for a long time: digitally transforming how they do business. For those who haven’t and have been putting it off, now more than ever, I think people are sitting up and thinking about how they can you know get through this unique moment in history, and find solutions that allow them to be more efficient or have a better understanding of their physical assets remotely. Just to cut down on the routes that people are running or people having to be there in person to detect how a process is operating. But to understand how we got to where we are today we’re gonna go back to 2012.
The Internet of Things
So the Internet of Things has not lived up to the hype. And why we say that is, back in 2012 the IBM Watson team predicted that we’d have 1 trillion IoT devices out there in the world by 2015, which would mean three years later. When we got to 2015, we were nowhere near that and you can see all the future projections for how many IOT devices are coming in the future – always several years away – they keep falling lower and lower. And today there’s, you know, under 20 billion IoT devices out there in the world. Only a fraction of that are industrial IoT devices right. That go after the kinds of things that people are typically inspecting manually, and would really, really benefit from IoT sensing to bring those data streams.
Here at Everactive, we think there’s two reasons for that. Two reasons why we’re not hitting these projected numbers. The first reason is batteries. Logistically, it’s just very difficult to stay on top of a fleet of aging batteries out there in the world. We’ve met a lot of maintenance planners in the world, and a lot of technicians, but we’ve never met somebody who’s a battery changer or a battery changing scheduler. So if you’re gonna adopt a fleet of battery-powered sensors into your environment, you’re taking those existing people and now adding a new task. Those maintenance planners have to make sure they have all the batteries they need in stock, and the people to go out there and change them, and you know usually they are not considering that when they thinking about getting these new data streams.
Environmental Impact and Battery Power for IoT
On the environmental side you know many of these batteries can be recycled but if you think about having to just store the good batteries before you deploy them and then take the bad batteries and warehouse them for some amount of time before you can take them to be properly disposed of it’s a big headache I have a huge jar here that my wife fills with our dead batteries and once a year so I try to figure out what to do with them you know find someplace that will take them if you think about thousands or tens of thousands of batteries in an industrial environment where you have to worry about keeping on top of that stock it’s a difficult thing to do so instead of you know disposing of them properly sometimes we see them thrown away which is a really really difficult thing to wrap your head around there.
But even if you had excellent batteries a 10 year battery life which we have never seen a product that lived up to a 10 year battery life in industrial sensing at birth but even if we were able to get tenure batteries to hit that IBM Watson number we need 274 million replacements of batteries every single day. It’s a lot to stomach so everybody wants all the good things that come with IOT solutions. The whole idea is to run more efficiently, have better insight in your facility and not have to dedicate as much time to thinking about these things until they they pop up and become a real issue.
But if you add the human cost and the actual cost of buying those batteries, it just means that we’re getting fewer data points than we should. So we see smaller deployments of sensors where people are picking the top 5 or 10% of their you know assets and putting sensors online, but the sensors themselves are also rationing out the data they send to try to milk as much life as possible out of those batteries. So instead of transmitting continuously, which would cause them to run out of battery within a couple of weeks to a couple of months, they’re gonna transfer it a couple of times an hour or a couple of times a day to try to live up to those battery projections. To drive the point home one more time here about human capital involves a little projection. If you had a million data points coming in through battery-powered sensors, as those batteries aged out you would need 120 full-time employees or full-time equivalent employees to go around just changing batteries, which is a huge team. So that makes it pretty clear me why people are not willing to scatter these things everywhere and subscribe to this new lifetime of repeat battery and replacement.
Steam Trap Monitoring
So the other thesis we have for not hitting those projected numbers is the fragmentation of what’s out there and the integration of them into your environment. So for each one of these rows here most solutions today have a different provider for each one. With our first product steam trap monitoring we had a customer come to us who’s looking at us and a competitor. And for each one of these rows in their RFP response, they had a different vendor show up under one umbrella for the bid. So they had four companies that they had to wrangle to get the mounting to the asset, the sensor itself, the connectivity in the plant back, all of the cloud and then the analytics to make sense of that data. All from four different companies that would have to work together. And it turned this particular maintenance professional into a project manager just trying to figure out how to get these external vendors managed, but we’re not the only ones who’ve seen that example. Our customers did a study on this in 2016. They went specifically to predictive maintenance professionals about industrial IOT and said ‘Are you excited about this in 2016?’ and the enthusiasm was 90% of people said ‘Yeah we want to get industrial answers.’ In 2018, they went back and they recalled the same people who responded to that survey. I think enthusiasm had dropped way down like more than half. Like 30% of people were excited about it for predictive maintenance
And so they commissioned a study that came out last year and they boiled down the reasons why these people were not as enthusiastic anymore:
These four core reasons: IT infrastructure is not ready for this amount of data. The build-it-yourself method is hard to do, there’s no guarantee of success and it can cost a lot of money. If you grab things that are pre-built off-the-shelf and try to put them into a corporate or industrial environment, a university, government facility you are opening yourself up to security exploits and flaws. Most of these solutions don’t have a connection to energy use or energy consumption on how to reduce it so you get some kind of raw data without analysis as to what you should go fix to improve your operating conditions.
So those all make sense to me. We’ve heard from our customers too. So enter Everactive, our company, we make sensors that sit out at the edge on steam traps, pumps, and fans rotating the vibrating equipment.
Machine Health Monitors
Today we’re looking at the machine health monitors. There’s a picture of a steam trap there but we’re not going to spend too much time on that today. These sensors sit out there at the edge and they gather data continuously. So they’re always on and always measuring and then transmitting back wirelessly. And they can do this perpetually with very low levels of harvested energy, like the amount of light in the room you’re in right now. Or a small difference in temperature, they’re designed to be out in industrial environments. So that IP66 rating is for water or dust ingress into the enclosure. They’re fine to be outside, we have them in process environments where they’re getting hosed down by customers today. So we’ve seen in those environments, caustic spray down that sort of thing, sensors rated to handle that. They’re also intrinsically safe at a rating of class 1 division 2. So if you have electrical classification in your facility to make sure you’re explosion-proof, our sensors are designed to work in those environments as well. I say wide operating range here, what I mean is a wide temperature operating range. These sensors can work everything from an outdoor process in the Midwest winter, where we saw 40 below, up to summer in the south in a steam tunnel where we see continuous operation at 160 degrees Fahrenheit.
So once those sensors digitized that process they take measurements around those systems. They send them up to the cloud through a pretty traditional IOT gateway, where we think of them, once they have the cloud as new data streams. And the reason why we say that is if you’re getting quarterly vibration analysis data before on a motor and you install our sensor which will report back every 15 seconds, in the first minute you’re going to get a year’s worth of data. In the first hour, you’re gonna have 60 years worth of vibration analysis data on that motor. So that’s a ton of data to wrap your head around. A huge data lake and a new data stream, but we do not expect you to parse all that yourself it would be too much for you to understand. So that’s why we have analytics and a notification engine where we can send you alerts when something moves out of spec. So your attention is called only to the things that need fixing, which means you’re not going around running predictive analysis with handheld tools visiting machines that most of the time are good. Most of these manual audit runs we see 90% plus of the assets they’re checking in honor are behaving just fine. So in our view that’s a wasted step to go and visit all those things to check on and only see that they’re good. With continuous sensing you can take that same person and redirect them to only things that are failed. They’ll never check on a good one again but we don’t just use alerts to do that. Today we’re doing that with text messaging and email but we can also use cross-platform API’s to get into our customer systems too. So later this year we’ll be able to integrate this product into worker systems which could tip your staff off to go check on an asset without even having the acknowledgment in an email.
One thing we should note about these solutions: everything from the low power chips that allow our sensors to be batteryless, all the way up to the web UI and the analytics are all done by Everactive employees. This means you don’t have to bring separate integrators together and we can design prototype support to troubleshoot these things all in-house. So you don’t have to worry about that integrator issue or getting stuck. As you know, with one of your support people being a middleman we can call our VP of Software right now, our help desk has his phone number. So there’s nothing we can encounter here in the commissioning of this system that we can’t help. All of that without ever changing a battery.
So let’s talk a little bit about that next fabulous operation, our core technology. We’re own election appropriate earlier differentiator is this batteryless technology. We were not university researchers, our co-founders met at MIT. I went back to their home University of Michigan and the University of Virginia, one of them with a specialty in low-power radios and the other with a specialty in super low-power processing. When you bring those things together you can build these smart sensors that sit out at the edge and can run off of a very small amount of temperature differential with a thermoelectric generator. That’s what’s down here in the lower left-hand corner, indoor/outdoor solar light radio frequencies or a vibration. In the case of the applications we’re doing today, temperature differential and light have been so plentiful we haven’t needed to productize radio frequency or vibration. So in the case of this product, machine health monitor, it can use either of these two indoor/outdoor light or a temperature differential to power itself in perpetuity. We provide everything you need to get up and running any solution. So we have some sensors here on the left-hand side photos of those and then in the middle is one of our iOS e gateways. This is a very standard kind of industrial PC in a box. It’s also waterproof and rated for extreme conditions in temperature or spray down. We have an intrinsically safe model as well. This one pictured here is about the size of a shoebox. So we place these around the facility and then they transmit directly over LTE to the cloud where you can access the results of that sensor data and our analytics on cross platform software. And it’s truly universal, you do not need to install an app. Everything is driven through a web browser; the notifications come via email or text so you know you don’t have to install a special app on your computer or commission a server on your site in order to use our technology. It’s completely free standing. As I talked about earlier, we are working on integrations into other platforms as well. So you can take data up and out from there but you don’t need to bring anything to get started, just bring your motors or your steam traps.
So for our customers this means right out of the box they have a solution that they know is gonna work. You take sensors out, you take the Gateway out, you apply those assets to the motor and within a few minutes really you’re seeing data flowing to the cloud. You haven’t added the maintenance problem of dealing with batteries, you’re not going to go back and visit those assets until you know something is wrong with them. So you’re not going to make this trip just to check on a battery or replace a battery when there’s nothing wrong with the underlying motor. And finally with our web dashboard we’re able to call attention to things that need fixing. So it’s very clear to see how we’ve added value in your environment by showing you those faults that we’ve detected and displaying them in a linear straightforward way.
So next, let’s take a closer look at the machine health monitor itself. This is the moment I’ve been waiting for. I’m sure some of you have to think of our company, here’s an example of an install out in one of our beta install environments. You can see here our Eversensor sitting up on the motor, a thermoelectric generator, and a photovoltaic solar harvester. Take a closer look at each individual component here, but this is designed to fit motors of all sizes from small horsepower motors: 200 plus 300 plus horsepower motors. That’s kind of the rear see these sensors being most valuable to be deployed ever since itself this small. Like I said, it fits in the palm of your hand. I’ll show you one here live in a moment this is what transmits data back to your gateway. It does have a supercapacitor on it for some energy storage. We can use magnet screw mounts or you know, kind of perch mounts on the motor stud mounts or epoxy mounts for that. This unit is smart, it has a processor in it and can generate a Fast Fourier Transform right there on the unit at the edge. That’s what transfers back to the Gateway and that’s water and dust tight to our IP66 ratings.
So this little sensor is capable of perpetual transmission every 15 seconds back to the gateway at just a 15-degree difference in temperature from the skin temp of the motor. So if it feels warm to you, that’s more than enough for us to train your remarkable little device. Inside there’s a tri-axial accelerometer. Now we have a magnet in turn that helps us measure the cycles per minute or calculate the FD slip. If the motor is not running at the rate you think it should it also takes an ambient temperature measurement and ambient relative humidity measurement. But what’s really neat about this to me versus our previous product is that it’s capable of taking over their updates. So we can update the software on this remotely and change parameters on it remotely. Once it’s placed in the location you wouldn’t have to go back and make any changes in it even if you wanted to alter the software settings. It’s a pretty cool development in our new radio protocol that we’re able to do that.
Taking a closer look at our thermoelectric generator, these are magnetic but they can also be epoxy down. Just find a warm spot on the motor if it feels warm to the touch on your hand that’s probably more than enough for us to power. It’s that temperature differential that generates the power. So in the core of this is something called a Peltier Device. We did not invent that, it’s a technology that’s been around a for while. What’s really remarkable about us is that we can use such a small harvester and still power our sensors. So it’s inside that Peltier device, or two dissimilar pieces of metal; if one is warmer than the other then it starts to emit a trace amount of current that’s what those blue fins are for. That heat sink allows us to exaggerate that temperature differential from the skin temp of the motor in the air there. So that’s what gets the electricity flowing right there it’s wired back to our Eversensor and you can actually place this anywhere on the motor, body or driven equipment. You could even go on a neighboring machine if you needed to but it does track the temperature where you spot it. So if you put it somewhere where you want a temperature measurement that’ll be reported as well. Another view of our thermoelectric generator here. The neat thing about this is it does daisy chain out to that solar harvester so you can add an additional harvester there. That’s new for this product and platform for us. The solar harvester comes in two different form factors: they’re for indoor or outdoor life. It’s used in conjunction with the thermoelectric generator it could run on as little as 200 lux which is really dim. Lux is a measurement of light and 200 Lux is almost certainly the amount of light in the room you’re in now.
Even if the lights are off it’s probably light enough for us to power this sensor. When we back your light modules available to and essentially any amount of outdoor light with fairness perpetuity without their let you get IR to red light and it’s a much more plentiful source of energy for us. So yeah dissimilar harvester is a really handy new way to power one of our sensors super excited the CDs get hung out into the world. So we started with a photo of a big motor and a still motor which is a smaller one to give you an idea of the size. For a complete install there with the sensor thermoelectric generator and their solar harvester so our turkey for this is kind of the balance of the planet.
What we mean by that is if you have a thousand horsepower motor it probably came with a suite of built-in analytics like onboard sensing. And you probably already have somebody looking into that data regularly, but there’s just a huge number of motors. A big percentage of the motors in your plants today or in your facilities that have been deemed maybe not worth putting expensive instrumentation on. In order to understand how they’re behaving we can break through that. This will retrofit pretty much any existing motor. As I said, if it’s not a magnetic motor we can epoxy down the small and large motors, both we can measure. I mean the idea here is to help you understand sooner when they’re in danger feeling so they can go be serviced. We collect your vibration levels in as I said it’s a tri-axial accelerometer, so three axes. We are going to give you in the software side the magnitudes of those nine highest peaks, an interpretation of the VFD output from that magnetic sensor, the surface temperature of the motor, ambient temperature in the air and we’re working on calculating the machine runtime, and starts and stops as well in our software.
So what I’m really excited about on this slide is this data transmission. If you’ve seen our previous product with our Gen 1 radio the range on this one, 800 feet, is a big leap ahead. So we design all of our sensors to fit in industrial environments. In fact, I talked to our lead RF engineer about what he had in mind when he designed the protocol and he said to think of the windows pipe screensaver. I might be dating myself a little bit there but think about that dense environment, really they’re just kind of a jungle of pipes in your way. So all of our testing is done in industrial environments where it’s dense. We know we’re not going to get a line of sight so that NLOS is non line of sight. We also support, on all of our products, tremendous density sensors in the gateway. So we’ve tested a thousand plus sensors reporting to a single gateway, no problem. From that gateway, you can use LTE, Wi-Fi, or Ethernet to get that data up to the cloud.
Look it’s the cloud, you can mouse through all of that data at any time. Now we have threshold alarms that we can configure for those parameters and then send you an email so you can jump right into the dashboard from there. And we’ll escalate the ones that we think need the most attention and it’s really just simple straightforward easy to get going. You can edit the metadata for the sensors right there on the cloud to pull it up and have it match your facility. Straightforward, really linear, easy to grasp!
Next on our agenda here I want to show you the hardware. So I’m gonna stop screen sharing for a moment and we’re gonna toggle over to our camera. So here in the room with me I have a small motor, it’s like a seven and a half horsepower motor. I’ve put the can of coke there just as a reference point, you get a sense of the scale, both of the motor and of our sensor itself.
Tom and I were talking earlier about this motor. It’s an old marathon electric industrial motor salvaged in a factory here in Detroit. I asked Tom ‘What’s a typical application for a motor like this?’ And he said there’s dozens to hundreds of potential applications from mark. Probably see a lot of them that look like this in your facility. Just starting off, I just wanted to talk about our mount. So the mount itself here is a small about two by two inch cube that you can epoxy down or magnet down. You can also see in the middle there there’s a hole for that stud mount in Britain and you know find a spot along the border. Could be they’re taking their pre-existing measurements at top dead center ,so we’ll put that there. This is the sensor itself Breen’s of the outfit so it has all of our low power chips in there for the radio transmission, the processing, creating that FFT. There’s a little window on the front home. You can see there’s a gore-tex logo on there because that’s how we can insulate and waterproof the sensor and still get an ambient humidity and temperature reading.
And it’s made into this base. We’ve got a couple other different bases too that could get you a stand off from the motor. This is the one that we’re shipping on day one. So get into there and then it screws down and you just place that on the motor body. this one has kind of a satisfying magnetic lock you’re going on, but once again there’s a host of different ways that you could mount those. Next is our thermoelectric generator, so this contains that Peltier device I talked about. Essentially from the base here which will apply to the surface of the motor to the tip of these fins. If we have like a 15 degree fahrenheit temperature differential that’s enough for us to power the sensor. So that’s our thermoelectric generator. If you’ve seen our previous product, we used to have one —we still have one actually– that mates to a pipe for steam applications. But this one was redesigned in order to meet the motors. So we have some arms here that’ll help our tank from drifting. Once we’re applied through a little stability arms, we can bend out of this to find a spot on the motor. I’m gonna put it right here over the bearing because that’s a spot I’d like to monitor: the temperature on the connector between the sensors.
And our hardest errs is a USB C type connector. We feed them so they fit in exactly and can be screwed in. There’s also a waterproof gasket around it, so this means that the cable won’t be pulled out once you’ve screwed it in. You think it’s gonna be in there tight and it’s gonna be waterproof, that’s why we have one of our solar harvesters here for. The daisy chain, do small solar cells. It’s funny when I think about solar cells. Like if you go to Harbor Freight or if you drive them down the side of the highway you’re used to seeing these huge solar cells the size of your car. Just the roadside light or a sign you know that kind of thing. So to be able to use a really small discrete harvester like this it’s a breakthrough for us, for both of these harvesters.
Bluetooth Low Energy (BLE)
You know, if you tried to power something like my Apple watch it would not even boot because the Apple watch has four very power-hungry things. The top four energy consumers are LTE, the screen and Bluetooth Low Energy (BLE). Bluetooth Low Energy is kind of the most humble of those at 50 microwatts. Our wake up radio is at 200 nano watts, so it’s an order of magnitude lower. We have individual components we’re working on that operate in the picowatts. So when you think about just that small amount of energy, that’s really what allows us to be different from our competitors. So we’re going to screw this in with the same connector here: waterproof gasket rated USB C connection. And we just find a spot where we have our best access to light. I’ll pop it up and it’s also magnetic. So that’s in position and there you go, simple. It’s all in the amount of time I’ve been talking if we had warmth or you know if we had light we’d already be up and communicating. So again that’s the hardware side of the solution.
Next up we have our software. We built this from the ground up once again to call your attention to the things that are most important. So each time we install one of these motors, those motor monitors will gather information off the nameplate of the motor. It will put into the cloud that’ll help inform our analytics, as well as just making it easy for you to understand which motor we’re talking about on call. So here’s an example of our vibration level measurement in the velocity area. The measure here IPS peak and for each little spot on this graph is an individual measurement in time. so you can drag your mouse across them highlight them and see them compare them to one another.
A neat thing about our cloud platform is you can go back and see all these measurements. At any time, you can go back and see the history of this motor. So I could come see measurements that are months or even years old in order to see if the data we’re looking at today, in context, if it makes sense or not. You know, if it’s a radical innovation from expected behavior. So that level view is where a lot of our customers start looking for a peak or something off. But we also have a waterfall view of our FFTs. So this is a three-dimensional graph which allows us to use this waterfall plot to look for things that might look a little bit off. We see a couple things here that may be noise, may be another outlier, or two sample data. Basically sort of look a little bit different depending on your application environment. We can also zoom down to ranges inside this measurement if we wanted to take a look at a specific part of the data. So it just gives you a quick way to understand why you got an alert and what you should look for when you go and check on the motor.
So in the case of our threshold around Ruiz that we measure, we can set thresholds around those. Obviously, the vibration level is the one that most of our customers are interested in. So we can set vibration levels there and then get a notification that’s per asset you know. Tom I were talking about these thresholds and how you know they’re not typically one-size-fits-all. You can’t apply a general rule to all these. Tell me, I think you’re telling me a story about what happens when you get trained in vibration analysis.
TOM: Yeah Peter, the old rule of thumb is that you take your first round of readings coming back from your first training class and if you applied the broad brush recommended alert and alarm settings you’d end up shutting down after your plant. Something I think many of us have learned the hard way. Yeah it’s funny to think about right. Like the textbook says it one way but then when you get there and you’re into actual making and you’ve applied it, it might operate a little bit differently. So you know we do not try to assume that a universal alert threshold is going to work for everybody. We can change that to fit your environment.
PETER: Good right. So our product is set to be released commercially in July. We do have some pre-release customers who are using the sensors today. So we’ve been able to be out there in the world gathering data on them and so we can speak to these specifications as being valid, it’s just that our official commercial release is coming next month. So we’re accepting customers today to sign up and look at applications in their environments. We’ll be shipping them out the door in July.
Electromagnetic Field Harvesting
I talked a little bit about that thermoelectric generator that you know 15 degree temperature differentials so small between the motor surface, but you can also in combination with the TEG use solar. So if you have light that’ll help as well to power the sensor. We’ve talked about EMF electromagnetic field harvesting and frankly in all of the applications we’ve looked at so far, we’ve been fine with either temperature or solar. EMF is on our roadmap so later this year or next year you may see EMF harvesting join the fray as another way for us to power our sensors. Our default interval for transmission you know is 60 seconds on these but we are testing the data every 15 seconds for applications where that’s necessary. That’s you know for those vibration levels for each of those three axes, the temperature readings, both the ambient and the skin of the machine, and VFD output.
Over-the-Air Software Updates
In the case of the runtime and the stop start counter that would be coming later this year, one really neat thing about this platform is since we can make software updates both to the cloud and to the sensor itself over-the-air we do not need to be locked into a software feature set. When we ship we’ll be able to introduce additional features after the products are already out there in the world. So you could buy one in July and as we continue to evolve the software for those new features you’ll be able to get those automatically applied via over-the-air updates. The sensors themselves can also report their stored power level and their signal strength, which can help us in troubleshooting. If there’s an obstruction we can see that maybe they’re on the fringe or have a weak signal but rain just becomes much less of a problem for us since developing our new radio standard.
What’s New With Our Software
So Evernet – this will be the first product we ship. With that we’re seeing an 800 foot range in industrial environments. We’re standards-compliant so we have our FCC IDs that’s in the sub gigahertz range, super low power with great range. Which are two things that are usually at odds with one another. And that over air configurability is new for us with this product at the sensory level. The gateways themselves haven’t changed a whole lot. I mean I’m having our new generation radio in them. They still support LTE Wi-Fi or Ethernet to back all up to the cloud and we can use AC mains, you know like a traditional wall transformer or power over Ethernet in order to run those gateways. So that’s what our customers are doing today. On the cloud side we continue to advance the cloud product, so we’ll be able to do complete pairing and provisioning from the cloud side when this product is released, which is new for us. So if you’ve added new sensors it’ll be easy and simple for you to drive through our web interface.
You will get early warning notifications for you that are reviewed both by our analytics and also vibration analysts and we have PDM experts in the house that are looking at these alerts who can help you put them in context. Obviously, we’re also getting threshold alerts that are user configurable. They can come in Evercloud as a dashboard email or SMS. Later this year updates to the cloud will give us data analysis to do deeper things like alarming on the rate of change or a standard deviation change from our mean for measured parameters. So another example of something that we hoped to have on day one, but we can still add later this year via over-the-air updates.
We’re also working on machine learning and deep learning analytics around those assets to give you greater insight as to how they’re behaving. And finally, we’re working on a calculator to give you a business impact analysis about the down time consumption of these aging motors. You know there’s a thesis where there’s a time in a motor’s life as it ages and becomes less efficient, where instead of running it’s a failure. From an electrical consumption standpoint it’s cheaper and makes more sense to just replace it outright or service it. so that’s what we’re after with those analytics and those are something we’re continuing to work on.
You saw the physical install just now, it’s pretty simple and straightforward to get going, applies pretty easily to motors, pumps, fans, gearboxes, really any rotating and vibrating equipment that you would like to have notifications from. Wake a sensor up and you know right out of the box usually these things go on in under five minutes. So the magnetic mounts are the easiest way to get going but we also can give you stud mounts if you need to get up and away from the motor body a little bit. Or you can epoxy it down if you have a high vibration environment. Do you think the magnet could walk? And from the time we shipped these they’ll be IP66 and electrically classified.
Our Solutions in Your Environment
Alright so let’s talk about next steps: what this looks like in your environment. That first ejection as to why we’re not seeing more predictive maintenance devices, a few devices out there, in the world that IP is not scalable. I think that’s still true in a lot of these environments. Luckily, Evernet supports thousands of sensors reporting to just a handful of gateways. You can do over a thousand sensors on a single gateway most of our integrations out there in the world today don’t require a single network connection. And none of our pre-release customers for this product have used in-house networking connections. So all of our pre-release customers have been able to use LTE to get this information up to the cloud. This build-it-yourself method where you’re the integrator is difficult, expensive and risky.
So our sensors are made to retrofit to equipment just like yours without having to turn anything off. You don’t have to buy a smart motor, you don’t have to stop your process, and unless these applications go in a couple minutes there are no big upfront capital expenses to get started. We’re service based so you pay a flat service fee and that is you know simple and easy to calculate for us to quote ahead of time and that way you’re not stuck you know buying into this huge platform and then waiting for years for it to pay off. You’ll know right away if you’re getting returns or not.
And since we are a service we’ve really de-risked it for you. If you do find for some reason that monitoring these machines is not providing money to you, you can turn the service off and we have to come and get our ball and go home.
That third objection attempting to integrate off-the-shelf devices introduces security flaws. That’s true, anybody is like a Ring Doorbell can tell you you know those files go up to Amazon and who knows what happens from there. Our Everactive products, and here they’re really strict security standards for both hardware and software encryption. Our solutions are already trusted by manufacturers, government agencies here in the US, and big universities. So we’ve had to pass a bunch of very stringent security audits. In fact, we have self-imposed sought to audit our company. So that SOC 2 level data on it means that we are perpetually being audited to make sure we’re being sanitary in our data handling, introducing security issues into our environment, or our customer environments. That sought to audit standard is the same standard that the government uses for their NIST and IST standards. So that is the beef standard for data security in private industry.
Monitoring Energy Consumption
Last of all, here most IOT solutions don’t have a clear connection to energy consumption. They give you a temperature measurement, under which they give you one piece of data but they don’t have a translation as to where you need to go in your facility to fix it or reduce this energy consumption. Evercloud has a summary rundown of alarm parameters. So it’ll call your attention to the machines that are not operating over your thresholds and that are furthest away from optimal, so you can figure out which regimes are wasteful and need attention. So when we think about our network architecture, we’re all sensors out in the world, these Eversensors. And the idea is to get many of them out there pervasively sensing your environment to cut down on these manual rounds that you’re doing on foot. That low-powered Evernet reports back to our gateways and once the gateways are in, we establish this network sending data up to the cloud. You may start today with something like steam traps or motors and then you’ll have a single dashboard that reports you know all that data out.So as we work on future products the Machine health monitor which is coming now, or like pressure safety valves which are another application we’re looking at, if you’ve already got the Gateway established and you’ve blanketed your facility in Evernet, trying new products and adding them in is simple. You just pop them on and get started with those simple retrofitting stalls like the one we looked at.
What’s Next for Everactive
So when we look at our short-term roadmap a lot of people ask what’s next? And we are very application-focused. We try to think about things that we know are wasteful and causing operations issues, uptime issues in our customer environments. So we started with steam traps. From there, we knew we needed to measure temperatures: remote temperatures around the trap, ambient temperatures, you also get humidity and Lux. So these top four sensors with a thermoelectric generator for the hot pipe, that’s our steam trap monitor.
For vibrating equipment, we’re using ambient remote temperature measurements, humidity and Lux. In addition, we have tri-axis accelerometer and that magnetometer to give us a sense of VFD slip. All six of those sensors are in a single package that can use either a thermoelectric or photovoltaic to power itself. So if you start thinking about future applications for us, we think about what sensors you can support and how we’re going to power those applications. So we’ve already tested pressure, acoustics, and ultrasound gaskets. action and we’ve been able to power those for some amount of time off of our harvested energy budget transmitted data back. So we’re working on productizing those into other things that you may see around your facility that could benefit from pervasive sensing. Where you don’t want to have to go and change a battery EMF and vibration are also on our roadmap for harvesting. A lot of people think hey if we know we have a vibrating machine why don’t we use vibration as our primary harvester and that’s it ?
TOM: That’s an interesting question. The reason why we don’t is for harvesting from vibration, you have to tune the crystal to a very specific range of vibration. And as motors age, even before they go out and parameter, they may change the harmonic they’re vibrating at. Which means you could walk away from what it takes to power the sensor while the machine is still operating within specified parameters. So we’re continuing to do research around that and maybe other applications or that make sense but for all the applications we’ve seen. So far in our pre-release, we’ve been able to power them off of temperature differential or a solar cell. Finally, here we are still accepting feedback about this solution. As I said, it’s malleable. Even once the hardware is deployed we can continue to add features via software so that’s something we’re working on.
PETER: This giant QR code on the right-hand side of your screen if you hold your phone up and open your camera app you’ll be able to scan it and it’ll take you to a survey where we’d love to get your feedback about what you’ve seen today and ways we could improve the product. If there are must-haves in there for you, we’d love to hear them obviously. We’ll take a few questions today as well so if you take a look at the bottom of your screen there’s a Q&A box. Tom and I will take a look through those together here and I’ll read a few of those out and you can see what folks are interested in talking about here.
Question and Answer
Question: Alright Tom, I feel like that was a whirlwind. You know you kind of went around the world and back. Ok, first question: do you make your own tags or do you simply source them?
Answer: Great question. We source them and we wrap the mechanical engineering around them to make them work, but the actual core itself, the Peltier device, was two dissimilar pieces of metal. That’s kind of a commodity that’s out there and can be had. So we take tags that we buy from a supplier and wrap them inside a little mechanical engineering in order to get them to work.
Question: Ok, what does low levels of harvested energy mean exactly?
Answer: This is early in the presentation. I think we’ve talked about what we mean by low levels of harvested energy. So just a low level of solar light like a diamond or lighting 200 Lux. And actually, they’re apps on your smartphone, or you can measure Lux using the built-in cameras. So if you have questions about what exactly that means, you know we can talk you through what 200 Lux looks like. But in the case of temperature differential, that’s a change in the difference between the air and the thing that our tag is on and that starts at 15 degrees Fahrenheit or above, that’s enough for us.
Question: What is an FFT and why should I care? Tom, will you talk through the Fast Fourier Transforms and how we convert data from our accelerometer?
Answer: Yep happy to do that Peter. I think anybody that’s been exposed to vibration analysis understands FFT. It stands for Fast Fourier Transform, it is a series of calculations that takes a complex wave that’s not truly a sine wave and breaks the individual frequencies of vibration and the magnitudes of vibration. And that really helps you zero in on what the cause of the high vibration is and therefore what maintenance steps should be taken to correct it that makes good sense
Question: My next question is, is the solar harvester optional or required solar?
Answer: It’s optional. We can power off a temperature differential alone. You can put it downstream from the tag, but if you don’t need it you don’t have to take it.
Question: Another question, the kiln between the sensor and the tag is pretty short. Do you have big sizes?
Answer: Yes, great question. We have all kinds of sizes we can ship out for these. I think one of the advantages of having you know these components be modular and hardwired to one another is that you can and I build it kind of suit to fit. So we start every integration now with a walkthrough; talking to our customers about where they’d like to place the sensors, where they’d like to place the harvesters, and if we need longer cabling in order to permit that, we can do that. We’ve seen a few examples already where the only part of the application had light shining down on it or only part of it was warm. Like the driven equipment was warm, but the motor was using longer cable lengths to extend off of the sensor, which allows us to power ourselves even in environments like that.
Question: Ok, can you elaborate about cost effectiveness ROI ?
Answer: Yeah. So we you know we take a look at for our applications how we can pay off by looking at you know where you’re being wasteful today in energy consumption? And also where you’re being wasteful in manpower right? If you think about the cost of doing routine manual audits on all these motors by taking those people and redirecting them to only things that need to be fixed, you’re going to run much more efficiently. So we are working out some ROI calculators about implementing our product. What that means for savings at the facility level, I’m happy to sync up with you after the meeting we can talk about what those models look like.
Question: Is the solar harvester shockproof? Is it protected against physical damage due to debris?
Answer: It is. There’s a rubberized casing around it and it’s pretty tough. Obviously, we had a vibration spec for it and I can run that down from you. It’s on our spec sheet, but yes it won’t be shaken to death by the motor for sure. Obviously, if there’s impact damage to it, like if you drop a hammer on it, it could break. But one of the benefits of us being a service model is you know we cover our hardware. So we’ll be happy to work with you on replacement parts. If you have something like a failed unit due to impact damage, the first couple of times it happens it’s on us. If you start shipping them back to us with boot prints and tire marks on them we can do to make sure they’re better protected.
Question: Do we use a piezoelectric crystal for measuring?
Answer: No, we do not use piezoelectric for measuring time. You can correct me if I’m wrong there, I think we’re just using a three axis accelerometer for measurements today on vibration analysis, and a magnetic censoring magnetometer for the VFD measurements. The accelerometers are piezoelectric based, but they are digital accelerometers.
Question: I see another question about accelerometer sensitivity being typically a hundred millivolts per G?
Answer: Again we use a digital it’s under try a steel those being there the conversion time I I lost your idea for just a moment their digital accelerometer I got that for sure if you wouldn’t mind repeating yourself I’m gonna jump to the next question time will circle back around for your answer okay I’m having a few issues there I’m gonna jump over to our next question not sure if you coming, sorry.
Question: How are the sensors connected to the edge gateway?
Answer: Yeah so we use a proprietary low-power radio standard to do that because the protocols that were out in the world before we designed our products were just too power hungry. So if you’re trying to, for example, use a harvested energy budget to transmit via LTE, you’d only be able to transmit maybe once a month or something along those lines. Just because it requires so much energy to do so. From the gateway out to the edge sensor that is proprietary super low power with great range, like I said up to 800 feet, off of just using our energy harvesters.
Question: Can you set a threshold based on FFT?
Answer: Yes you can. I showed a screenshot on that and we can set thresholds for FFT values. So you can figure out if your overall energy vibration level is too high, you can get a notification based on that and that can be configured per asset per parameter. So if you have two motors right next to one another and the one you want your axial vibration set to be different from the one next to it, we can be that specific in setting our thresholds.
Question: Alright, how does the customer decide the threshold? How can you support it?
Answer: So we work together at the time of install to set some ground rules for thresholds. So if you don’t know where to get started, we’re happy to do that. Obviously, since there’s notifications involved we want to make sure that we’re choosing a reasonable threshold so you don’t get too many notifications early in an integration. But for our team, obviously we have people who will be on site with you to help you get installed and we have a team of PDM experts with deep experience and vibration analysis that can help you come up with reasonable thresholds to set when you’re getting started.
After you’ve gathered data you know if you’ve had the accelerometer data out there for some period of time you can always reevaluate and say alright, this is where I’ve seen it operating and in my application and then we can make changes as we see fit. Those are editable at all times so if you have a motor that’s too chatty you know it’s running okay, but the threshold was set though so you can go and state that yourself anytime.
Great, some more questions around price. I’m happy to sync up with you outside of here so we can talk a little bit about what it looks like at your facility but essentially, we price you know per asset. So we’ll look at the cost of censoring up your motor, will determine how many accelerometers we need, how many measurement points we need, and there will be a monthly or annual fee involved in that that will extrapolate that out to your facility.
Okay oh man lots of questions here. Some of them are probably more time specific so with Tom’s audio off at the moment there are a few of these that are specific about frequencies in the spectrum range. I’m gonna have Tom speak to those and we will sync up with you outside of the meeting. On their questions about load and shaft deflection I will sync up with you on that one outside.
Question: What plans in the time frame are you looking at distributing to Australia?
Answer: So we are. Our radios can be tuned to fit international spec standards, so we will sync up with you about that outside of the meeting as well. But we are currently testing in countries outside of the US, it’s just a matter of getting those certifications in order. So I can talk to our partnerships team and see if they’ve started work already on Australia certifications. I don’t know off the top of my head but there are a number of countries outside of the US we’ve already started the process of getting those certifications in. So it may be as soon as later this year just depending on all the hoops that are involved that we have to jump through in order to get the equivalent of an FCC certification out there.
Alright.well we wanted to thank everybody for taking the time to join us here today. For the questions that we weren’t able to get to in the group Q&A, I’m gonna make sure we sync up together outside. We have your names here on that list in the Q&A and we’ll reach out to you individually to make sure your questions get answered. Thank you so much everyone for making the time to view this today. We’re super excited to introduce this product to you and looking forward to getting out into your environments, putting these on your motors, and getting that data gathered. I think it’s gonna be really fun to see this technology kind of take flight here into the world. So thanks again for being part of it with us today!
For more information on the IoT, and how Everactive is trying to overcome the battery obstacles click here.