Stop Chasing Leaks with Embedded Acoustic Leak Detection

Hello and welcome to this WaterWorld presentation entitled "Stop Chasing Leaks," sponsored by Kamstrup. I'm Mandy Crispin, Editor -in-Chief of Waterworld with Endeavor Business Media. Let me explain how you can participate today.

First, if you have any technical difficulties, you can submit feedback to the questions window on the side of your screen and one of our technical experts will help you. We recommend disabling any pop-up blocking software extensions in your browsers.

These cause issues with the webinar player. Also, we will be recording today and the archive will be available on Waterworld.com within 24 hours. You'll be notified when that's available. Again, we are live, so submit any questions that you have at any point during the presentation, and we will be answering those after the main presentation.

Again, the questions window on the side of your screen is how you submit a question. I would like to introduce our speaker for today's discussion. His name is Graham Mattison. He is a Solutions Manager for Kamstrup North America.

He has over 16 years of water industry experience and more than a decade in acoustic leak detection. He has helped utilities across North America to eliminate more than a billion gallons a year in non-revenue water.

Welcome Graham and please go ahead whenever you're ready.

Hi, everyone. Like she said, my name is Graham Mattison. I am a Solutions Manager with Kamstrup. I'm going to be going through a little discussion here today, essentially taking you through kind of how we manage leaks, how we go through leak detection, how we actually manage NRW, reduce water loss, and then how we actually keep it low and reduce it even further.

So first thing I'm going to do is I'm just going to get into kind of a brief introduction. For those of you who don't know, Kamstrup is a Danish company. We are headquartered in Denmark. We've been in business for 75 plus years.

Here in North America, Kamstrup's focus on empowering water utilities with reliable, cutting-edge ultrasonic metering solutions, intelligent network communications, and intuitive software applications.

Best part is it's made right here in the USA. Everything I'm going to be talking about is made in our brand-new facility in Cumming, Georgia, down in the southeastern U .S. When it comes to metering and what we actually do here at Kamstrup, in North America, all the meters that we produce are ultrasonic.

We've been doing it for 30 -plus years. We're not new to it. I know in the past two, three, maybe five years, we've seen a lot of other metering companies start introducing ultrasonics to the marketplace. We've been doing it for quite a while.

That means that we are able to innovate on top of the meters and bring you things like integrated acoustic leak detection. One of the big things we want to talk about here when we're, when we're talking about the manufacturing process, what we do, all the meters that we produce in North America, are ultrasonic meters, all of the meters are going to be discussed in day are ultrasonic meters that are produced in our factory in North America.

The biggest thing is that when we talk about our meters going out the door and coming back, we have a return rate of less than a quarter of 1% here in the US globally of the 14 million ultrasonic meters we've shipped in the past 30 years, our return rate is less than half of 1% and I want to point out just a wording there.

The reason we use return rate. Is because we're counting everything that leaves the factory. If it comes back for any reason other than physical damage or ordering the wrong component. If it comes back for any reason other than that within the 20-year warranty period, it gets counted here and even at that we're still at less than one quarter of 1% return rate across the entire spectrum of our product lines.

As I mentioned, all the meters that we use here in the US, all the meters we manufacture are ultrasonic. Just give you a brief background on how ultrasonic technology works. Essentially, the meters have a couple of sensors in them, two ultrasonic sensors.

Ultrasonic just means that it's sound in a range that's outside of the normal range of human hearing. So it's a frequency that's higher than you can hear, but it is still sound. We essentially have two acoustic sensors in that are passing a signal back and forth.

When flow is going through the meter, any of the signal that's going with the direction of flow speeds up a little bit. The signal that's going against the direction of flow slows down a little bit. We're just measuring that difference in transit time.

So essentially, it's that same equation that everyone loved in high school where one train leaves the station heading one direction, another train leaves the station heading to the other direction. When do they arrive on time?

If they don't arrive at their expected times, we're just measuring what time did they get there and how much does that shift? That's how we can measure the flow. The nice thing about that is we don't have to wait for anything internal inside the meter.

There are no moving parts. It's a static meter. It's just math. All we're doing is measuring the speed of sound through water. And if it's different than it should be, it's because there's movement.

If there's no difference in the speed of sound between the two sensors, it means there is no flow. What that allows us to achieve is unlike a mechanical meter where starting flow rates are between one eighth of a gallon and a quarter of a gallon, with these new ultrasonic meters, we're able to get down to one one-hundredth of a gallon per minute as far as our flow rates and our starting flows.

When it comes to water loss, first thing we need to talk about is kind of why it matters to your specific utility. So there's a lot of different drivers for water loss. Maybe revenue is the big one. That's pretty universal across utilities.

Some states have more regulations than others. There may be water scarcity issues in your area. You might be affected by drought. You may just be looking to have a more efficient system. You might be looking to reduce your environmental impacts and create customer service.

The big thing here, when we look at all of these, we had gone out and we did a survey of 30 utilities across the U.S. We asked them kind of what their drivers were, what the different things were that were actually driving their non -revenue water, and what are the things that they are looking for in a system to be able to address those challenges and help their customers.

When it came to customer satisfaction, we were kind of surprised at some of the responses we got because when we talked to the customer service folks or the personnel out in the field that are actually doing repairs and interacting with customers on a daily basis, the biggest question that they were getting from customers is, what is the utility doing to conserve water on their side?

We've seen rebate programs and other things to help improve customer efficiency and usage on the customer side beyond the meter. They want to know: what is the utility doing to actually reduce water loss on the utility side.

And now we have an answer for that, and we'll go through it here in just a little bit. When we talk about water loss, we need to break down what we're actually referring to. So when we talk about non -revenue water, that's a combination of apparent losses, which would be inaccurate metering or unauthorized consumption theft, essentially meters that are older mechanical meters that lose accuracy over time, that's the water loss we're talking about. It's not actually lost. It has gone through a meter. It's just the meter did not register it, did not capture all of the flow going through it. When we talk about real losses, what we're talking about is leaks on the distribution main or on the service lines before the meter.

So losses where the water is escaping the network before it actually goes past a meter. With this solution that we're going to be discussing here today, it's the first time that you have a single piece of equipment that allows your utility to address both your apparent losses and real losses at the exact same time.

When we talk about non -revenue water, there's multiple sources of non -revenue water. You may have overflow of storage tanks or theft. You may have some areas that are non-metered. You may have things like hydrant flushing programs, and if you don't meter those, those would all fall into your non -revenue water bucket.

The biggest thing we're going to be focusing on here today is water loss through your distribution leaks and your leaks on service connections before the meter. Typically, when we look at non -revenue water, the breakdown on that, typically about 51% of your water loss tends to come from leaks, and a little more than 50% tends to come from leaks on your mains as opposed to your service connections.

So, if 51% of water loss is attributed to leaks, how do we find the leaks? Well, before we can actually go out and find the leaks, we need to know what types of leaks we're actually looking for. So, when we talk about leaks, hopefully we're not talking about this one, where a geyser is destroying a car dealership.

The whole goal with leak detection is to identify those leaks before they become a catastrophic failure, a catastrophic event bursting through the ground. causing standing water, sinkholes, things like that.

That's what the general public pictures when they think about leaks and water utilities is that geyser does that sinkhole. Our whole goal with leak detection is to catch it before it ever gets to that point and reduce your water loss by being proactive about monitoring.

When it comes to background leakage, those are the types of leaks where we're talking about a seep or a weep might be a drip every few minutes or so. It's something that's not putting enough energy into the system.

It's not generating enough energy to actually produce a noise. No one can find those because they're not producing a signature that allows you to locate them. What we're looking for is leakage that does not surface but is generating enough energy that's producing an audible noise that can be heard or detected remotely.

When it comes to the different types of leaks that you can find that you actually go out and repair, most people, like I said, they picture that large geyser, that sinkhole, that big large main rupture.

And yes, those lose a ton of water. In this case, it's a circumferential break, essentially the pipe snapping off and just pouring water out the end of it on an eight -inch pane. It ran for three hours because it reported itself.

It was spotted immediately. You get out there, it gets repaired. Yes, you lose a lot of water for three hours, but it's repaired fairly quickly and shut down fairly quickly. When it comes to a reported utility side leak, if it's a service leak, a lot of times we'll send someone out to look at it.

If it's serious and it's on the utility side, it gets repaired fairly quickly. If there are other higher priority leaks to be dealing with at that time and it is not a bad leak when they investigate it, a lot of times those will run for a little bit.

The longer we let the small leaks run, the more they lose and they lose it continuously. And you have a number of small leaks and they add up. This particular example, it's four of a half-gallon leak and a running for 16 days, it's 104 ,000 gallons, which is three times as much as is an eight inch full main brake running for three hours.

Now, when you talk about customer side leak, it's kind of at the mercy of whenever the customer gets around to actually repairing it. So it tends to run a lot longer, assuming they're paying their bill and you're not in a drought condition, it's probably not that big of an issue for the utility, but we do want to be proactive and make sure that we're notifying customers and being as proactive with conserving water as we can possibly be.

When it comes to leak detection and actually finding the leaks that are generating our water loss numbers, regardless of what technology you use, regardless of what type of methodology you use, survey method, the process is going to follow this general outline.

You're going to do a system assessment. You're essentially going to try and figure out where in my distribution network is most likely to have leaks. Then you're going to localize it. You just try to narrow down where in that area, am I going to start my investigation?

You send someone to investigate. They're going to confirm and pinpoint if they hear something. Once you've pinpointed and confirmed it, you're going to go out, you're going to have a crew repair and then report on it.

Essentially take down all the information about that leak so that you can do your year -end water loss totals and reporting. And then the cycle repeats itself. So it doesn't matter if you're doing a manual survey, if you're hiring people to come in, it doesn't matter if you're putting loggers out, it doesn't matter if you're using meters with built in acoustic leak detection if process, regardless of the technology or the methodology that's used.

When it comes to acoustic leak detection, many of you have probably seen some of these things at your utilities, even if you're not familiar with them as far as how they work or using them. Most common leak detection equipment that we find at utilities is just your common ground mic or listening stick.

Might be called geophones where you're at. Essentially just a headphone with an amplifier and a probe for listening to the ground. Sometimes it might be an elephant's foot, which is just a microphone with a cover around it.

So we don't hear a lot of noise from around us. Either way, that's the most common thing. It's just listening directly below you, and that's what gets used for pinpointing leaks out the field after they've been identified.

When it comes to kind of the next advancement away from that, it was lift and shift noise loggers. So these would be your drive -by loggers, essentially something you set out on valves or any other type of connection that you can get to, the main or the services, and just have them listen overnight.

You come back, you download the data. If there's a leak, you repair it. Great. Then you move on and you start sweeping through another area of your system. You're essentially moving these devices systematically throughout your system in order to get through all of it and identify all the leaks in each zone.

Then you have things like your in -pipe surveys, your smart balls. So these would be for your larger diameter mains. Things that are larger than 14 inches or large diameter PVC, HDPE. These smart balls work fairly well for those in -pipe surveys.

One caveat there is you do have to make sure that you are in place to catch it at the other end or they do like to get lost in the network if you miss them with the net. When it comes to correlators, that technology is essentially just for narrowing down where to focus your investigation.

Once you determine that there is likely a leak, say in this block so you have loggers on either end telling you there's a leak, this correlator allows you to put a sensor on either end and then it'll tell you within three feet where you should dig or where you should start your investigation for pinpointing with a ground microphone.

It's great for that, it's not really a survey tool, it's more of a pinpointing tool. When we talk about cellular correlating leak noise loggers, that's the type of equipment that I worked on for 10 years and helped develop and deploy.

The nice thing about that is it's going to give you built -in the ability to record audio files so if two devices here leak at the exact same time, it's going to record an audio file, it's the same time of the same night you'll be able to compare those audio files it'll tell you roughly in between them where you should be going to start your investigation and do your pinpointing to identify is it a leak and where is it?

Saddling leak detection is a top -down kind of survey snapshot. Essentially it's the same technology that we've used to use and we still use for oil and gas exploration or mineral exploration. It's just been repurposed to look for residual chlorine in the soil.

So and throughout the U .S. essentially what it would do is it would take a snapshot top down and give you a radius around each spot that it's seeing residual chlorine in the soil. One thing to keep in mind the northern half of the U.S.

we tend to salt the roads in the winter. It does have issues with that just because salt is made out of chlorine so we're going to see it everywhere every road is going to light up like it's leaking during the winter if you're salting.

Same thing can happen near the coast if you get salt spray from the ocean it can look like there's residual chlorine in the soil just keep that in mind if you're looking at that technology. All of those technologies are what we would typically refer to as a one -time leak survey tool.

It allows you to go out find your leaks repair them and move on. It allows you to sweep the system allows you to identify where your leaks are at currently it allows you some of them allow you to sit there for two three up to five years at a time and listen for leaks.

When it comes to things like your satellite leak detection again top down one -time snapshot same thing when it goes for a leak noise correlator you're out there you're identifying the leak you pinpoint it and then you move on to another one it's a one -time snapshot what you don't end up with is visibility on what's going on after that leak has been repaired and you don't have the ability to see what's going on around that device separate from the leak detection equipment itself.

One other thing that we did want to bring up just because we were getting lots of questions about it in the past few months is in -meter pressure monitoring specifically for leak detection and it does work for leak detection but it's really good at identifying those catastrophic leaks like that geyser that was destroying the car dealership earlier.

It's good at identifying leaks where the system can't physically recover the pressure behind the leak. Basically a leak that's so bad that our pressure drops and we can't get back to where we should be with the pressure.

We don't typically need to locate those, they tend to be, like I said, your geysers, your sinkholes, your standing pools of water or the stream that's running down a sidewalk. Those are fairly easy to identify. What in -meter pressure monitor is really designed for is verifying a hydraulic model.

So giving you the ability to, if you're putting the other hydraulic model or you already have one, being able to check it and then correct any errors that you might find in that model. Biggest issue with any kind of pressure monitoring equipment is it is analog.

So, the analog sensors will fatigue over time and they can drift. With most leak detection devices, where it's a standalone device, or sorry, pressure monitor devices, where it's a standalone device, you have the ability to disconnect that, re -zero to atmosphere, put it back on, you have the ability to change your sample rate to make it sample faster.

Within meter pressure monitoring, in order to re -zero that sensor to atmosphere, you'd have to physically remove it from service, and then it's only going to sample as fast as that radio that's connected to it is paying it.

So it doesn't have the ability to log multiple different configurations, and if we're trying to find water hammer, we really need it super fast, like more than once a second. So it does have its uses, it is useful for certain things, it's not an ideal fit for leak detection, but it is great for verifying hydraulic models.

We've covered kind of the other technologies you can use for leak detection, how you use it, where you would use it. What is built -in lead detection? What is built -in acoustic lead detection when we're talking about having it built into a meter?

Well, when it comes to ultrasonic meters, like I mentioned before, what we're attempting to do is pass a signal back and forth between two different sensors, and then we're measuring the shift in that signal, either faster or slower, based on whether or not there's flow going through it in which direction.

In order to measure that flow, we have to account for all the acoustic interference that's going on around that meter. Essentially what that means is we're listening all the time to see if there's any noise around that meter and adjusting our signal path to account for it.

Well, one of our engineers thought, well hey, if we're listening for all the noise that's going on around our meter currently anyways, why don't we put a little module in here that tells the meter to be quiet for a little bit, just listen to what's going on around you, and then pass that value on.

In the past, we've just been throwing it away, using it for our calculations. Was it useful to anyone who realized, hey, we can listen for leaks with this? So now that's what we're doing. We're essentially taking the noise value of exactly what's going on around the meter and we're passing that information along, allowing you to see what's going on acoustically around the meter.

And the beauty of it is, it's the same sensor that we're using to measure the flow. It's the exact same process that we're using to measure the flow. We don't have to do anything different, which means there's zero effect on battery life.

That means we can have a 20-year lead detection solution now and not have to worry about the batteries dying three, five years down the line. It's 20 years now. So what types of leaks can actually be identified with a meter that has built -in acoustic leak detection?

So we're not no longer just talking about constant consumption. We're not talking about like customer side lead detection. We could do that. Everyone's been able to do that for 25 years. That's what we've been calling leak detection in the metering industry for more than two decades.

Well, we can do that, but what we're really focusing on is what's going on before that meter. All the stuff that's not going through the meter, not being captured, not being recorded or billed, what's going on before the meter, and how to identify where we should be sending our crews to go out and pinpoint, find, and repair these things.

Well, typically if it's between the meter and the main on the service connection, it's going to have one meter that lights up, or maybe the adjacent meter that's really close by, and main reason being, excuse me, main reason being is that the service connections don't actually weigh very much, so they don't transfer a ton of energy to the water main.

What this means for you is that if you do have a leak on the actual main itself, as opposed to on a service line, instead of one meter or maybe the adjacent meter lighting up, the water main weighs a ton more than the smaller service line connections that are attached to it.

It's going to vibrate and shake everything that's attached to it, so if we have a leak on the water main, it tends to vibrate everything around it, so all the meters around it light up, and then we're just looking in between the two loudest meters.

If it's a service connection leak before the meter, it tends to be that meter, maybe an adjacent meter, and then we're just going to the loudest service connection, but this allows us to know where we're going to be going before we ever roll a truck.

We know what the meter's doing, we know what the sensor's doing, that means we can tell is this leak alarm for a customer side leak, that could be an email, that could be a phone call, a text message, or is this actually on the utility side, and we should roll a truck and go investigate it.

Now it's a question of are we investigating the service, or are we investigating the main that these services are connected to, and in between which two services. Hopefully that makes sense. When it comes to built -in leak detection, why is that a big deal?

Why do we want to have it built -in there? Is that better than having a standalone leak detection device that only does leak detection? Well, it can be. So one of the things that we always get asked when it comes to leak detection is how far can your sensors hear the leak?

How far away can they hear? How far apart can I put these sensors and still hear the leak? And it's a good question because it's the question we've always had to ask when it came to leak detection. How far apart can we space these things and still hear the leak?

And it's a good question. The issue is it's not the question we need to be asking if it comes to trying to find all the leaks. The real question is, how far can the sound transfer from where it's originating?

So where that leak is, how far away can the sound transfer, either through the pipe wall or through the water column, in order to hit a sensor that might be hearing it? One of the nice things is we're listening through the water column.

So we're passing that signal off to reflectors through the water column and we're accounting for noise in the water column. So we're actually listening through the water. But we're also part of the pipes.

So when it came to sensors that were sitting on top of a valve via magnet, those are actually using weight and the force of the magnet to help the sound couple and transfer to that data logger. We don't have the issue of is it on, is it connected, is it coupled all the way, is it perfectly centered, because literally the meter is the pipe and the pipe is the meter.

And the meter is the noise sensor. So it's all in one. But because noise travels further and faster through the water column, and because we're spaced so much closer together, there's literally almost nowhere for the leaks to hide.

And we'll go through and explain kind of how that works and why. One of the other things when it comes to leak detection, all of the technology that we're talking about here today is built into the meter.

It's under the glass. That means that there are no extra wires. There's not a standalone sensor. There's not anything separate from the meter, literally the radio or modem module, the noise acoustic lead section, everything is built into the meter under the glass.

What that does for you as a utility, you no longer have three different warranties to worry about. There isn't multiple batteries anymore. The sensor isn't acting independently from the meter. And the meter isn't notifying you independently from the sensor.

And what I mean by that is you're no longer going to get alarms from a sensor that is sitting right next to a meter that has constant consumption and constant usage because that meter knows what's going on.

And now the sensor itself knows what's going on in the meter. And having it built in there gives you the ability to see in two directions. Is this a leak? And which direction is it at? If the meter's spinning, we know it's on the customer's side because it's not going to zero.

And if it's not spinning, we know that this noise is being generated from in front of the meter. It's not due to consumption down the line. just makes it a lot easier for us to actually manage this and not have to maintain the hardware constantly, not replace batteries all the time, not worry about wires getting chewed up or cut, not worrying about the the end point radios actually being ripped out of the ground or hit by lawn mowers,

things like that. It's one single unit under the glass, no extra wires, no extra sensors, no extra hardware. And I say all that because everything that we've been doing up until this point when it comes to leak detection is what's referred to as a minimum viable survey deployment.

Literally, we were asking the question, how few of these sensors can we possibly put out in hopes of catching at least some of the leaks? We don't have that issue anymore. You're literally getting 10 times as many sensors.

And because it is the meter, and it's the same sensor element we were using before, there's zero hardware cost now. So not only are you getting 10 times as many sensors per mile, there's no cost for those sensors other than the meter itself.

If you already have meters in your system and you're putting meters in, there's no additional hardware cost for a leak detection sensor. There's no additional infrastructure. Biggest thing here though, is instead of spacing of 500, 1,000 or 1 ,500 feet, now we're spaced in town 50-100 feet apart.

In the past, a lot of what we were doing was innovating in order to overcome the fact that the sensors are spaced 500, 1 ,000, 1 ,500 feet apart. And the reason they were spaced so far apart is because they're expensive, unfortunately.

So the nice thing here is, like I said, zero hardware costs, you get 10 times as many service connections per mile. And it's very easy to manage because you're literally so granular as far as how tightly you can get down and narrow down your focus to your investigation, to your pinpointing.

We're literally sending crews out to do pinpointing, not surveying anymore. When it comes to leak detection software, regardless of what technology you use, leak detection software is gonna give you three basic things.

It's gonna give you a map with your sensor locations. It's gonna give you an audio graph so you can see what the noise was like at that location before, during, and after you repair the meter. So we should see it go back down and stay down.

If it does not, if it does not go all the way back down, we know that there's another leak there so that it just won't. And if it goes down and comes back up, then we know that a new leak is popping off and that's the situation we need to manage.

One of the biggest questions that we get at that point, similarly, well, I'm getting 10 ,000 times as much data as I ever got before. I have sensors everywhere. We never had sensors before. Maybe we were hiring someone to come in and sweep our system, maybe a third of our system every year, half of our system every year.

I know I've dealt with utilities in the past where they're doing 20%. So, essentially, they get through their system once every five years. How do you manage all that? Well, one of the ways that we do it, it's kind of hard to see, but behind the little white arrow with the pulse sign there, there's a red slider.

What that allows us to do is actually sort and categorize our dot colors, basically our central locations on the map based on how loud it is. What this allows you to do is if you're taking full -time meter readers and you're going to be repurposing them or retraining them to become expert pin pointers or expert leak detection personnel, we want to be able to train them on your system so they can learn what leaks sound like in your system because everyone's system is going to sound a little bit different.

You have a different mix of pipes, different types of soil, different pressures. It's going to be a little bit different everywhere you go. Nice thing is we allow you to sort it based on how loud the leak is, which means that when your crews go out there and they start working with the system.

They're getting trained on the easiest to find leaks first, and it's going to get progressively quieter and progressively harder. But they're being trained on what a leak sounds like in your system, what it should sound like.

If there's pump noise in the area, they're getting an opportunity to learn all of that on the easiest to find leaks first, and it's systematically progressing and sweeping through your system, getting harder and harder, but more confident as they go along.

So before we get into case study here, we actually see what a customer that's used this, what they're actually getting out of it, and how it actually works out in the real world, as opposed to just me talking about it.

Just want to go back through some of the main features just to reiterate. So the acoustic leak detection that's built in the meter, you're getting acoustic leak detection across your entire network. That's 7,300 distribution -wide acoustic leak surveys over 20 years.

There's no additional hardware. There's no additional infrastructure to collect the information. And most importantly, there's no additional manpower required to set up, maintain, and utilize the system.

It's out there. It's in the meters, and as long as your meter reading comes in every day like it should, you're getting your leak detection every single day on time, just like you should. What that gives you is the ability to get closer and actually focus and prioritize your leak detection efforts and your NRW reduction efforts.

What I mean by that is you're getting 10 times as many sensors per mile surveying constantly every day. And we do it proactively. So we're listening every 55 minutes, 26 times a day, 365 days a year for 20 years.

And we listen every 55 minutes, because what we're trying to do is eliminate our false positive noise sources. So intermittent noise, like pumps running or people irrigating. Traditional leak detection equipment listens at around 2 AM.

And it does that for a couple of reasons. It's listening at 2 a.m., one, because hopefully most people are asleep, which means they're going to have their taps shut off, and everything that is spraying for a leak should have a little bit more pressure behind it and spray just a little bit louder.

The other thing is they're hoping that because everyone's asleep, it's going to be a little bit quieter, so we're going to get less ambient noise in the area. But the biggest reason that we're listening at 2 a.m., or most traditional leak detection equipment is listening at 2 a.m., is because we're trying to conserve battery.

The sensor takes power to run, transferring that data takes power to transfer. It's additional in addition to just being out there listening for the leak. It has to do everything. Nice thing with this, we're listening every 55 minutes.

It's 26 times a day. We're taking the lowest value because we're not using any additional battery in order to do this. This is the same value that we would typically be using to do our normal flow calculations.

Thank you. We just have an extra module in there that's telling our meter not to write to long -term memory, be quiet, and only listen to what's going around you. We do that every day, 26 times a day, and we take the lowest value, and we do that because it eliminates all those false positive noise sources.

If it's a leak, if you see additional noise and it doesn't go away, it's most likely a leak because if it wasn't there before and it is there now and it's not going away, that's what leaks do. If it goes away, if it wasn't there before, it got loud, and then it stops being loud, it's not a leak.

Leaks don't typically repair themselves, so the noise typically only gets louder, not quieter. So what are customers that are actually utilizing this? What do they get out of it? What do they get out of the system? How does it work? Let's see it in the real world.

First case study here is the town of Oneida, Tennessee. They were one of our first users that deployed a full AMI system with all meters with built -in acoustic leak protection.

They started out with 4,620 mechanical meters in their system that they needed to replace. They were looking at doing a six -month deployment that required them to install 15 data collectors in order to monitor their 118 -square -mile service territory and 322 miles of mainline pipe.

Six -month change out is fairly aggressive, considering the size of the utility, but they did it. They changed out all 4 ,620 mechanical meters in six months. And the reason that they were so aggressive about it and the reason that they did try to accelerate it in six months is because at the time, they were at 51% water loss.

The first three months after getting the meters deployed, that went from 51% down to 28% non -revenue water. So 10 .7% of that was just from the initial meter changeout, so that was just the savings and the gains that you get from going from old mechanical meters to brand -new ultrasonic static meters that never lose accuracy.

The difference was 10 .7%. Everything else was from the acoustic leak detection. When they turned on the system at initial startup, Oneida had 77 meters that had acoustic noise level over 100. Essentially, it's loud enough to indicate that there's a potential leak there.

So what did Oneida do? They took their full time, their two full time meter readers and they had them retrained to become expert pin pointers. Not necessarily surveyors. What we want is people that can get to the site, identify where the leak is, pinpoint it and put an X on the ground so that we can repair it.

We no longer are just kind of wandering trying to find the leaks. The system's going to tell us where they're at. We just need to be able to identify them and locate them once we get there. So they took their two full time meter readers.

They had them retrained as expert pin pointers, essentially got the equipment to be able to do the leak detection themselves. It's like your ground mics and your correlators to be able to pinpoint on site after the leaks are identified.

That first year they repaired 70 plus leaks. Two years later now, they're expected to be at or below 15% water loss, and that should save them an additional $140 ,000 in loss revenue and gain an additional 36 working hours, which is a fairly big deal.

The even cooler thing, though, not only were they able to reduce their water loss 51% down to 28%, but they were able to keep it low and then drive it lower. A good example of that, when it comes to their water treatment plant, their treatment plant had typically been running 15 to 14 hours a day.

And since deploying the system, since repairing those leaks, that's gone from down from 15 to 14 hours down to 11 hours per day. So a three hour per day reduction in water treatment plant runtime. That doesn't sound like a ton, but that's 45 operating days a year.

It's a month and a half less operating time on their water treatment plant every single year, going forward for the next 20 years. So what types of leaks were they able to find with it? Well, again, they did have 77 leaks, and about half of the leaks that you're able to find are going to be service leaks, because about half of the leaks that are out there tend to be service leaks.

The thing is, this one ran for four and a half months. It's not a big leak. It's four gallons per minute. But because they had so many other leaks to get to, by the time they've gotten to this one, it's estimated a little over three quarters of a million gallons had leaked out of it in 135 days.

It was non-surfacing. It was in a ditch line, just rocks rubbing against the poly service line caused a split. It was well -drained soil, and there's a creek on the other side of it. Did not surface, would not have known about it without this.

The big thing is they're able to repair it, and they're able to see that the noise level has gone back down, and now they're able to monitor for any additional leaks in that system. But we're able to tighten up the zone and keep it tight.

First case study here is the town of Oneida, Tennessee. They were one of our first users that deployed a full AMI system with all meters with built -in acoustic leak protection.

They started out with 4,620 mechanical meters in their system that they needed to replace. They were looking at doing a six -month deployment that required them to install 15 data collectors in order to monitor their 118 -square -mile service territory and 322 miles of mainline pipe.

Six -month change out is fairly aggressive, considering the size of the utility, but they did it. They changed out all 4 ,620 mechanical meters in six months. And the reason that they were so aggressive about it and the reason that they did try to accelerate it in six months is because at the time, they were at 51% water loss.

The first three months after getting the meters deployed, that went from 51% down to 28% non -revenue water. So 10 .7% of that was just from the initial meter changeout, so that was just the savings and the gains that you get from going from old mechanical meters to brand -new ultrasonic static meters that never lose accuracy.

The difference was 10 .7%. Everything else was from the acoustic leak detection. When they turned on the system at initial startup, Oneida had 77 meters that had acoustic noise level over 100. Essentially, it's loud enough to indicate that there's a potential leak there.

So what did Oneida do? They took their full time, their two full time meter readers and they had them retrained to become expert pin pointers. Not necessarily surveyors. What we want is people that can get to the site, identify where the leak is, pinpoint it and put an X on the ground so that we can repair it.

We no longer are just kind of wandering trying to find the leaks. The system's going to tell us where they're at. We just need to be able to identify them and locate them once we get there. So they took their two full time meter readers.

They had them retrained as expert pin pointers, essentially got the equipment to be able to do the leak detection themselves. It's like your ground mics and your correlators to be able to pinpoint on site after the leaks are identified.

That first year they repaired 70 plus leaks. Two years later now, they're expected to be at or below 15% water loss, and that should save them an additional $140 ,000 in loss revenue and gain an additional 36 working hours, which is a fairly big deal.

The even cooler thing, though, not only were they able to reduce their water loss 51% down to 28%, but they were able to keep it low and then drive it lower. A good example of that, when it comes to their water treatment plant, their treatment plant had typically been running 15 to 14 hours a day.

And since deploying the system, since repairing those leaks, that's gone from down from 15 to 14 hours down to 11 hours per day. So a three hour per day reduction in water treatment plant runtime. That doesn't sound like a ton, but that's 45 operating days a year.

It's a month and a half less operating time on their water treatment plant every single year, going forward for the next 20 years. So what types of leaks were they able to find with it? Well, again, they did have 77 leaks, and about half of the leaks that you're able to find are going to be service leaks, because about half of the leaks that are out there tend to be service leaks.

The thing is, this one ran for four and a half months. It's not a big leak. It's four gallons per minute. But because they had so many other leaks to get to, by the time they've gotten to this one, it's estimated a little over three quarters of a million gallons had leaked out of it in 135 days.

It was non-surfacing. It was in a ditch line, just rocks rubbing against the poly service line caused a split. It was well -drained soil, and there's a creek on the other side of it. Did not surface, would not have known about it without this.

The big thing is they're able to repair it, and they're able to see that the noise level has gone back down, and now they're able to monitor for any additional leaks in that system. But we're able to tighten up the zone and keep it tight.

Here's a really good example why most utilities feel like with leak detection they've been chasing a dragon they're never going to catch. Essentially why is it that when we go out, we hit survey, we fix our leaks, we repair all the leaks, we come back and we look at our outflows from our treatment plant or what we're pumping out into the system two months later and it looks exactly the same, maybe we've gotten slightly worse.

And this is exactly why that happens. So typically you go out there, you'll identify the leak in this scenario, a leak had started, it plateaued, it got worse, plateaued again, they were able to get out there, repair it.

But two days after the repair, you see a little green line on the right here creeping up. Essentially what's happened is they've made the repair on the mainline leak. It was a big leak, 180 gallon per minute leak, it's on a six-inch AC main.

For this logger away, it was about 1500 feet that heard that initial leak. They repaired it. two days later that same logger that was the furthest one away before is now hearing a new leak starting at that location.

The difference is we're able to see this leak at the beginning in its infancy. What that means is you're able to schedule the repair. It's no longer an emergency situation. You're not having to send someone out at 3 a.m. on New Year's Eve hoping that you know they can identify this leak or dig it up or get to it repaired in a timely fashion while everyone's kind of eyes are on you. You could do this in perfect weather conditions now.

You can make sure you have all the materials and you could do it on your time because we're finding it early and we can monitor in real time.

Here's another example from Oregon. This is going to be my last example of the slideshow but this is for the people that really want to know how far away you can hear the leak.

So this example is a 14-inch main. It's a 30 gallon per minute leak. We were able to hear it up to half a mile away on both ends. Now granted this had galvanized service connections. Galvanized service connections tend to ring like a tuning fork so if you have a leak they're going to transfer the sound from that main if there is any very well.

But again it's a 14-inch main. We're able to have our meters on either side of it hearing it from up to a half mile away. Just to reiterate again 51% of water loss is attributed to leaks. We now have a good way of doing something about it.

It's not going to cost you an arm and a leg. You get 10 sizes of made sensors without any additional hardware, infrastructure, or manpower. And it's a 20 -year solution now. We no longer have to worry about replacing batteries every three to five years.

We've put it out there with the meter. It is the meter and it's going to last you 20 years. Essentially why is it that when we go out, we hit survey, we fix our leaks, we repair all the leaks, we come back and we look at our outflows from our treatment plant or what we're pumping out into the system two months later and it looks exactly the same, maybe we've gotten slightly worse.

And this is exactly why that happens. So typically you go out there, you'll identify the leak in this scenario, a leak had started, it plateaued, it got worse, plateaued again, they were able to get out there, repair it.

But two days after the repair, you see a little green line on the right here creeping up. Essentially what's happened is they've made the repair on the mainline leak. It was a big leak, 180 gallon per minute leak, it's on a six-inch AC main.

For this logger away, it was about 1500 feet that heard that initial leak. They repaired it. two days later that same logger that was the furthest one away before is now hearing a new leak starting at that location.

The difference is we're able to see this leak at the beginning in its infancy. What that means is you're able to schedule the repair. It's no longer an emergency situation. You're not having to send someone out at 3 a.m. on New Year's Eve hoping that you know they can identify this leak or dig it up or get to it repaired in a timely fashion while everyone's kind of eyes are on you. You could do this in perfect weather conditions now.

You can make sure you have all the materials and you could do it on your time because we're finding it early and we can monitor in real time.

Here's another example from Oregon. This is going to be my last example of the slideshow but this is for the people that really want to know how far away you can hear the leak.

So this example is a 14-inch main. It's a 30 gallon per minute leak. We were able to hear it up to half a mile away on both ends. Now granted this had galvanized service connections. Galvanized service connections tend to ring like a tuning fork so if you have a leak they're going to transfer the sound from that main if there is any very well.

But again it's a 14-inch main. We're able to have our meters on either side of it hearing it from up to a half mile away. Just to reiterate again 51% of water loss is attributed to leaks. We now have a good way of doing something about it.

It's not going to cost you an arm and a leg. You get 10 sizes of made sensors without any additional hardware, infrastructure, or manpower. And it's a 20 -year solution now. We no longer have to worry about replacing batteries every three to five years.

We've put it out there with the meter. It is the meter and it's going to last you 20 years.

So with that let's open it up see if we got any questions. So go ahead and type your question, click submit. Any questions we don't get to, we can email you.

So, you know, don't be afraid to just flood them in. Okay. So we do have some to look at right now, though. So, well, actually, well, there's one I'm really interested in, but we can start with this one.

Someone wants to know what size meters you go up to.

So for leak detection, it would be all of your residential meters, one inch and smaller, are going to have built -in leak detection. Up until this past year, we were still producing some meter lines that did not have the leak detection built in from one inch smaller.

We've made the decision earlier this year to consolidate that and just put. leak detection in all the meters, one is smaller, and then reduce the price so that it's the same price as it was before we added leak detection.

And then, so you said this is the first time apparent losses and real losses can be measured at the same time, I think that's what you said.

So does anyone does anyone else have this solution?

No, so this is a patented solution that was developed by Kamstrup specifically for our ultrasonic metering technology, and because of our ultrasonic metering technology, we've been doing it for 30 years, so at this point, we're not in the trial and error phase, we're in the let's perfect it, let's see what else we can do with it phase, as long as we're not going to affect battery life and reduce the service life for the equipment.

So we spent about four years developing this, testing it, making sure that it wasn't going to have any effect on battery and that it was going to work, and it does both of those things spectacularly.

That's pretty impressive. Okay, this one is one that I'm curious about as well.

So someone wants to know what about other sources of noise, and I was thinking about like noise contamination, and you mentioned ambient noise. So can you tell me how that affects the reporting data, and would also like to know what kind of ambient noise are we talking about? Are we talking about dogs barking, or are we talking about fireworks going off?

So when we're talking about ambient noise, typically, when we're referring to leak detection technology or equipment, we're talking about noises that may occur between like that 2am, 3am hour, just because that's when everything else that's out there is designed to listen.

So usually it's things like pumps that are running or irrigation systems, but it could be literally any noise sources. We're listening on the side of a hydrant, it could be something like rain or wind. If you're listening in a valve, it could be something like, well, now it's raining and we hear rain hitting on the top of the valve box lid or we're having flooding in the pit and you know other issues like that.

So I mean essentially what we're doing in order to counteract that is just listening all the time and then taking the lowest noise value of the day and the whole point behind that is if it's a leak it should add additional energy.

So even if it's already noisy there if there's a leak in the area it's gonna add additional energy in the form of acoustic noise to that typical background noise level. All we're doing is trying to flag it when we see that uptick and change that does not go away.

If it's a leak it should get louder it should stay loud and it shouldn't go away until you're repaired. So that's that's how we address that.

Someone wants to know does this require a radio read or can you still do this? I think they mean a manual read. Does it require a radio read or can you still do it with a manual read?

So visually, it does require data to be transferred. That being said, you can read it either drive by or with fixed network.

The only caveat there is with drive by, you're only going to collect the data as often as you drive by to collect the data. What we do there to kind of make it a little bit easier is if you go out there and say, collect the data on a Wednesday, we're going to give you this Wednesday and the noise values from the previous three Wednesdays prior to that.

Essentially, we want to give you the ability to see four weeks' worth of history. So we can see, OK, is this noise value that we're seeing today, is that unusual? Or is that what it has been all the time?

Or is this something where suddenly we've seen a spike up in the last two times that we've read it? It's been high. And we need to send someone out to investigate it. When it comes to having to come back over fixed network, over AMI, something like that, you're getting that data every time it calls in.

Every 55 minutes, it's sampling. Every day, it's logging the lowest value. And you're getting that as soon as the meter read calls in. So it's not like just when you go out into the field with a truck. Does that make sense?

OK, so yeah, and then there was another question, which I feel like you just answered. They said, have the meters with acoustic sensors been used as real -time condition assessment of pipelines? But I think that's what you're saying. It's every 55 minutes.

So it's every 55 minutes, it's going to sample. And we're taking the lowest noise value. So this isn't something for doing like pipe walk addition assessment, where you'd be physically putting a signal into the pipe wall and cycling it so that you can measure how fast that signal is going through.

What we're doing is we're listening through the water column. And one of the main reasons we want to do that is because the water doesn't rust. It doesn't degrade. It doesn't slow down. So the speed change is always the same.

So when we're listening for. for it, it's actually transferring regardless of if there's been a repair made or we have a material that absorbs the sound more than another. Does that make sense?

And then someone wants to know what more alarms has the, are there on the meters that can only detect leaks?

Oh no, so it's gonna do all your standard like tamper alarms, burst alarms, so like a high rate of flow for a steady period of time, it's gonna allow you to know about that. It's also gonna do your customer side leak detection.

It's gonna give you high and low ambient and water temperature. So it gives you the ability to monitor for a lot of things and alarm on a lot of things. We're just talking about the leak detection aspect of it today because it's a unique feature that pretty much flips leak detection on its head.

Let's see, this one is fun. Everyone's interested in AI nowadays. So does the software have any AI, does it learn as it goes or notice patterns, for example, is what they say.

So yeah, exactly. So one of the unique features that we've actually introduced with the Leak detection software for this equipment is machine learning and pattern recognition.

So every time someone enters in their leak information for their year end reporting, essentially they're taking down their pipe size, their pipe material, your pressure at that location, the estimated gallons per minute being lost and then kind of your start and end dates.

That's essentially giving you the ability to record all that information and report on it. Sorry, excuse me, can you repeat that? I just lost my train of thought again. I need your, I need to repeat on that.

That's okay. Let me see if I can.

That was a different question and now I'm losing what we talked about.

Well, it was just about artificial intelligence. So does it learn as it goes and...

Sorry, so we're putting all that information in, but every time that we do that, it's compared to the audio graph. So we're literally taking an audio graph, we're entering information and we're confirming, yes, this is a leak. This is the type of leak it was. This is the size. This is how far it was from each sensor.

And we're able to compare that in the future to future leaks. We're also able to do that for any false positive noise sources. So we want you to be tracking not just your leaks, but anytime that you've gone out there and found something that wasn't a leak because that actually makes our ability to determine is it a leak or isn't it and alert you to it more quickly, it makes it a lot stronger.

At this point, when we talk about the software, you have a dashboard, main bucket is gonna be your leak bucket. So everything that's in there has been compared to previous audio graphs and machine learning.

If it's in that bucket on a dashboard, there's a 90% chance that when you get out there in the field and you go to the site, you're gonna... find a leak at that location. The next box right next to that one is just showing you all the meters that are loud enough to potentially be a leak but haven't been consistent enough to be classified as a leak yet.

Essentially it's going to say okay all of these are loud today. Were they loud yesterday? If so, great. We may want to flag it. We may want to track it. If it's something where this is the first day it's been loud we kind of want to see if that's going to go away tomorrow.

If it doesn't great we're going to flag it. If it does though it's not a leak it's just some intermittent noise source that happened to last the entire day. Maybe there was a tank that got drained and we're refilling it something along those lines but essentially we want to sort all that out.

Does that make sense? What can the AI as long as to do is really hone in on okay is this a leak or is it not? Where should we investigate and what are we investigating before we get there?

So, someone wants to know, does acoustic leak detection in small units, can it be combined with other softwares or does it only work with camp strip software?

And what do you mean small units?

I think what they're trying to get at is, like, say you don't have, say all of your meters are not Kamstrup meters, can the information from the Kamstrup meters be integrated into, like, their SCADA system or something.

I think that's what they're asking, I'm not sure. We do have the ability to, like, get the data out of the back end of our software and feed it into another analytics platform or dashboard, something like that, and we do have the ability to provide the leak detection information.

The only caveat there is if you're using a third party software for or third party meter reading system to actually read the meters, that software has got to be set up to do something with it. We can provide the data, it's going to, but if their software isn't designed to do anything with it, there's nowhere to display it and there's nowhere to see it until they do that.

That being said, we do make it available. We are working with a few different companies to make sure that they can display it properly, and we are providing the data to other, like, third party dashboard providers and things like that as well, so they can display it and analyze it a little differently.

I don't know if that answers the question, but the answer is yes, kind of, sort of, as long as the provider you're using wants to do it. Does that make sense?

Then someone wants to know what data do I get from the meter besides volume? I mean, I think you've started, and then someone else wanted to know, can it detect pressure? So I think we've already kind of talked about this, but if you could talk specifically about those two items.

What is detecting is volume, I mean, is meter. The only other thing is going to be detecting is, what does that volume mean? So is it is the thing where it continues consumption, customer side leak?

Is it a high rate of flow above a certain amount that it could be a burst pipe beyond the meter? Is it something where we're feeding the data into a dashboard and it's determining is it a toilet? Is it a faucet?

Whatever it's likely to be based on the flow rate of the consumption. All that stuff is doable, can be done, has been done or something that we are doing. As far as the data goes though, it's fairly standard across the board.

It's still a meter when it comes to pressure. At this point, we haven't been doing pressure in the meters. I know we've looked into it. I don't know if that's something that we're planning on doing because, again, with having the pressure sensor in the meter, there's biggest caveats are if You want to sample fast enough, it's going to affect your battery life in order to catch things like transients.

And if you need to re -zero the sensors and you need to expose the atmosphere to do that, which means you've got to pull the meter out of service, open the atmosphere and then put it back into service, which isn't really the most practical thing.

In the past, for about 10 years, aside from doing lead to section, I was actually doing pressure, active pressure management, where you're putting controllers on PRVs and things like that. And you really only need maybe two to five sensors per zone at most, where you're trying to monitor your critical points of the zone to be able to figure out, is the guy on the top of the hill that's going to lose pressure first when everyone else has their taps open?

Is he affected? Can we lower the pressure or not? It's great to have the pressure sensors everywhere, but if you only need it in two to five locations per zone, it's kind of overkill. So, that's the thing we're balancing right now is, is there a compelling reason to do it?

And is it going to affect anything else as far as performance of the meters in the system going forward that could potentially affect the life of the system? Make sense.

Okay, and then we kind of started to talk about this, but someone wanted to know, have you done partial deployment? Like, 20% of total meters with acoustics along with another system?

That's actually happened quite a bit. So, we've had a lot of customers that had started purchasing meters before we came out with the acoustic lead detection and interspersed these in there.

We also have some utilities that are running one brand of meters and one metering system, and they're gradually kind of converting over to these where they're running two systems side by side. The building system doesn't care.

We're just matching that how it sees it normally right now and we're putting the data into it seamlessly. So you don't, you don't notice the difference on the back end, but the data is just coming in from a different source. Does that answer the question or is there is there anything else you think we I'm missing there.

I'm not a classic leak detection expert so you'll have to tell me if you answer the question or why I'm just like, is there anything else in there that I'm missing. We are live, so if you want to add to your question, I'll happily, I'll have me happily ask Graham to expound so someone did ask is isn't having acoustic meter for each customer in the same street overkill.

So you start to talk about overkill, but this one specifically about each customer on the same street.

It would be if you had to pay for the handwear. So, here's, here's the thing with when it comes to lead detection. Really, it's a question of, are we building a better mouse trap because we're trying to catch all of the mice or are we building a better mouse trap because we're trying to justify the cost of the mousetrap.

When it comes to leak detection, we know the solution for finding all the leaks. Put a sensor everywhere. Get your ears on everything. Listen all the time. The problem in the past has always been, how do we do that cost effectively?

Because a standalone logger costs more than a meter ever would. That those standalone leak detection devices are a lot of times double, triple, four times the cost of what one of these meters would be.

So even if we're talking, is it overkill on sensors? Maybe. But if you don't have to pay anything more for the hardware, it doesn't hurt to have an additional sensor. Because if there's a repair, or there's a fitting, or there's something else in the system that could potentially make it harder to hear that leak, the closer we can get to the source of the leak, the easier it is to hear it and identify it.

Also, when we get out into the field, the closer together we have the sensors, the smaller our search area is. That's the whole reason we needed correlation in loggers that are spaced 1 ,000, 1 ,500 feet apart.

You need the correlation to narrow down your search area. Now we have in town meters that are 50 feet, 100 feet apart. It's already narrowed down. It's just literally go between the two loudest ones and pinpoint right below you where is that leak.

We don't have to go block to block or hydrant to hydrant anymore and survey once we get there. We can literally go to the location that it should be in and pinpoint. Does that make sense?

Yeah, OK, so we have time for like one more. And this one is just super interesting. And I think we should have some fun. So this person said, how can you detect a leak underwater, let's say in the sea? Can you do that?

Yeah, the way that, well, anything that's going on underwater, if you're trying to detect sound, it would be with a hydrophone sensor essentially listening through the water or you can listen if it's a pool you can listen to the surface of the side wall of the pool. That being said when we're talking about in a pipe if you're listening through the water column as opposed to through the pipe wall when it's going through the pipe wall it's literally spiraling along that pipe wall trying to find the path of least resistance essentially the sound is trying to travel along the best part of that pipe so hopefully there is a best part of that pipe but as the pipe degrades it the speed of that sound is going to slow down a little bit and that's kind of how a correlator works is it gives you the ability to adjust for those differences in pipe size and pipe material.

We're not really necessarily worried about those differences because we're listening through the water column and that's the same way that you would identify anything acoustically in water is to listen through the water column.

Sound travels through the water column about as fast and as easily as it would through like a four-to-six-inch duct layer pipe. it rings like a belt, like it transfers the sound really well, water doesn't compress, so there's no loss of the sound.

There's just a slight shift the further you get away from it, the lower the tone is going to be, the closer you get to it, the higher the pitch is going to be.

Okay, so, Graham, where can people learn more?

We can either learn more on our website. We do have distributors all throughout the country for this equipment. And if anyone wants a demonstration of the actual software, we can reach out to you after the event and show you how you would actually use the system on the field, what it looks like when you're actually trying to verify a leak and narrow down where is this leak at and how it works.

But we can certainly set that up for you, either reaching out to our distributors or regional sales managers, or just directly after this.

Great, and that's so it's Kamstrup.com. And obviously today's program was sponsored by Kamstrup. So thank you so much for sponsoring and getting this good information out to people. Thank you, Graham, for your time today and for speaking and all of you for joining. And I hope everyone has a great rest of your day.