While IoT has become a buzzword for gas turbine lubricant reliability programs, not all IoT sensors are created equal. Atten2 develops optical particle counting sensors that give rotating machinery vision and voice so that plant personnel and asset managers can have peace of mind in knowing their equipment can talk to them to stop a low-cost lubricant failure from becoming a catastrophic machine failure.
Video Transcript
Bryan Kneipp:
Hello everyone. My name is Bryan Kneipp. Thank you all for joining us today. I know these virtual conferences are a little strange in this day and age, but that’s the hand we were dealt. So thank you again for being here. I’m a certified lubrication specialist and a certified machinery lubrication technician. And I’ve been in the lubrication business for about five years. Ever since I got into this business, I’ve just automatically been drawn to the research and the technical side of everything that goes into reliability programs. So, I really have a passion for enhancing and optimizing people’s reliability programs. So let’s get into the presentation.
Bryan Kneipp:
I work for a company called Reliability 360, which is actually a division of Delta Fuel Company. It’s based out of Faraday, Louisiana. Reliability 360 is just a general reliability company. We offer filtration services, education and training, storage and handling solutions, and a lot more stuff like that. But today’s presentation is on the Atten2 S120 Optical Particle Counter. Reliability 360 and Delta Fuel are the sole distributors, in the United States, for Atten2 products. Atten2 is a sensor manufacturer that’s based out of Spain. So they have spent many, many years developing this specific product. They’ve achieved multiple patents. And as you’ll see, as we get into the presentation, it’s very tech forward-thinking, very progressive. And I think it’s the most accurate particle counter on the market today.
Bryan Kneipp:
Just a quick visual of some of the customers throughout the world using this technology. Some of you are probably wondering what optical technology is, or what digital imaging is, when it comes to particle counting. Basically, we use a digital image processor, shape recognition technology, an LED light, and a very high-resolution camera. So here’s an illustration of what the inside of the sensor looks like. You’ll see an LED light backlit like most particle counters these days. The oil flows through a high definition lens, the digital image processor, and then your high-resolution camera.
Bryan Kneipp:
So, this is what the oil looks like as it goes through the sensor. There’s two frames every second, and it’s taking pictures. And as you’ll see these brown circles here, that make up one big circle on the lens, that’s actually a self-calibrating tool for the sensor itself. So the technology knows how big those etchings are on the camera. So that’s how it’s able to tell how large each particle is, as it’s taking pictures. So you never have to recalibrate a sensor like this as compared to some laser technology or light refract in technology.
Bryan Kneipp:
So, that’s a really nice feature of optical particle counting. So again, this is just a picture here. You can see with the different images we’re looking at, in the particles that we’re looking at, it’s able to actually recognize air bubbles, which are perfectly round, water droplets, solid particulates. And we’re also able to recognize and quantify how many particles of fatigue wear, sliding wear, and cutting wear are present in your oil. And I’ll show you how we do that here shortly. So, this is a good visual of what the brain of this sensor is actually looking at when it takes a picture. You can see, it’s identifying every particle on the screen. Each color represents whether it’s just a solid particulate eliminating an air bubble from your count or identifying a different wear mode particle.
Bryan Kneipp:
So the Atten2 S120, the data that you’re gaining from this sensor, which can be done remotely through Wi-Fi, cell phone signal, a local display screen, Bluetooth. And really, you can get as creative as you want to, as how you want to view the data. But what we’re looking at is your standard three-digit ISO particle count. Also, you have the ability to look at your NAS cleanliness code. And here’s the root cause analysis through the digital imaging analysis. So, we’re actually using what’s similar to facial recognition technology to identify wear particles and categorize them in their proper setting. And here’s the bubble discrimination, as I mentioned. And the oil degradation and varnish risk assessment is a topic for later. But Atten2 has some really good technology to do that as well.
Bryan Kneipp:
So here’s some of the examples of the library of thousands of wear particles that are stored in the software. So, every time the sensor takes a picture, it’s actually comparing the particles to those in the library. And as you can see, here’s an example of sliding wear. Sliding wear typically has really jagged edges. Fatigue wear is more rounded. So, even if there’s not a good match in the library, it’s still able to tell you what type of wear particle it is. And here’s a cutting wear particle. There’s two of them. You can see the hook that comes off here. It’s an elongated particle, which means it’s cutting wear. So as it cuts, it almost has a curl here like a metal shaving.
Bryan Kneipp:
And here’s a real-life example of each particle on the right-hand side, under a microscope. So again, with the sliding wear, you can see the jagged edges. The cutting wear, you can see the curls here in the elongated particles. And then your fatigue wear, it’s more smooth and rounded edges. And then you have your bubbles here. So all of this is important, especially when identifying the different wear particles.
Bryan Kneipp:
As you can see, this is the evolution of the concentration of particles, and the size of particles, and where that lands you in your different wear modes, whether it’s severe or catastrophic. So, you have your fatigue wear in the green [inaudible 00:08:11] to get a really good picture of what’s going on. And here’s just an example of a success story from a customer. They were able to catch a failure that was starting to happen early on. It cost them $17,500 because they caught it early, as opposed to over half a million dollars for a fatality.
Bryan Kneipp:
All right. Just some additional unique features of this sensor. As I mentioned before, the sensor is totally self-calibrating. It calibrates itself every time it takes a picture. Most of the maintenance can be done remotely with just software revision. So, as the sensor is hooked to, whether it’s a cell phone or Wi-Fi, whatever it may be, we can actually update that with just software updates remotely. It has the capabilities for all your typical communication protocols. And we provide all integration start-up support. And then one more unique feature that’s important is, this sensor is compatible with just about any fluid. So, water, fuel, glycols, esters, all the exotic chemistry oils, this sensor is compatible with all of that.
Bryan Kneipp:
Just to mention, as you do monitor sensors remotely, just in general, from time to time, you’ll lose your service on how you’re monitoring that. So this sensor has the capability to store the last 1000 measurements. And as soon as your sensor comes back online, it basically just does a massive data dump back into your system. So you’ll be able to have that data available to you. Again, ISO 4406, your three-digit code, the NAS cleanliness codes, and your AS 4059, which I don’t think anybody in the States really uses.
Bryan Kneipp:
You can also set the sensor to read every 30 seconds, up to every one hour. So you don’t have to have data constantly coming in. The sensor’s fully customizable. And how you set your alarm limits for how many wear particles are present, what your cleanliness code is, how many air bubbles are present, and water droplets. So, that’s fully customizable based on your application. I know in the gas turbine world, we’re looking at some pretty tight tolerances. So, that’s something you really want to look at when setting those parameters.
Bryan Kneipp:
Here’s just a chart that differentiates the different technologies out there. This is the optical technology that’s produced by Atten2, laser imaging, which is LaserNet Fines, for most of you familiar with that. And then the light extension, which is basically counting shadows. As the oil passes through, there’s a light shining on it, and it counts the shadows in the back to count particles. And these two are magnetic technology. So, they’re looking specifically for the presence of wear metals. The stars mean it’s very good. Blue means it’s acceptable or average. And the red means it does not have that capability.
Bryan Kneipp:
So this goes back to that graph I was showing you with the concentration of particles, and then also the sizes. They used to use magnetic fields, or magnets, to determine wear particles. The drawback to that is, if it’s a non-ferrous material, it will not pick up the presence of that particle. So, specific to the gas turbine world, with the use of all these superalloys, we see a lot that are nickel-based, chromium-based, and different non-ferrous materials. So, if the concentration of ferrous metals along with those nickel bases is not high enough, you will not detect the presence of that wear particle. And the smallest particle, you can see here, even if it is a ferrous material, starts, this says 70 micron, but some of them will do down to 50 micron. Whereas, these start at four micron.
Bryan Kneipp:
And why that’s important. I’m going to go back to this slide. There’s 50 micron, and you go up to this arrow and you see that you’re already in severe wear mode. And at 70 micron, you’re already in catastrophic wear mode. So, is that tool handy? Yes, it is. The only issue is that you’re not catching a problem until it’s already really too late. So it only allows you just to basically plan a quick shutdown because you know something bad is fixing to happen. Being able to see these particles trending from four microns and on, allows you to take a lot more proactive approach.
Bryan Kneipp:
And here’s something that’s really interesting to me. It’s actually a recent development. And this sensor was not designed to actually do this. It just happened almost by accident. I’m sure a lot of you are familiar with the anti-foam additives that are put into the gas turbine oils. And basically, what those anti-foam additives do is, they adhere themselves to the inner wall of an air bubble. And when they do that, they form an insoluble micelle. Basically, what that means is it’s just forming an insoluble particle that takes the shape of a sphere. And it actually looks like a water droplet to this sensor. So, as it takes pictures, and it counts the particles, it will eliminate the presence of this anti-foam additive because of its spherical shape.
Bryan Kneipp:
And I know a lot of you may wonder why that’s important, but there’s been a lot of studies recently that show that once your oil is in service, and your anti-foam is being put to work, and these insoluble particles are formed, you can have anywhere from 4,000 to 120,000 particles per milliliter, which is going to send your cleanliness code up by three, or much more than that. So, a lot of people start to panic. If you are running an inline particle counter at this time, you see a massive jump in your cleanliness, and you’ll go put some type of fine filtration on it. You’re actually filtering out your anti-foam additives when your cleanliness code just wasn’t accurate. So we’re actually providing the most accurate cleanliness code in the market because of this. It’s a recent phenomenon. A lot of people call it a phantom particle. And the laser technology, and the light extension technology, it will cause a refraction in the light or the laser, and it counts it as a particle.
Bryan Kneipp:
So that’s something that’s pretty recent. There was just an article released about that. So if you’re interested in it, I’d be happy to send it to you. And here’s just some examples of what the software looks like on the back end. So you monitor this sensor remotely, you log in with your username and password, and here’s some of the data that you will see. It’s pretty customizable, how you want to see it. We actually are able to retrieve the pictures that the sensor is taking so you can actually look at them yourselves. If you want to, it’s really helpful to see how many air bubbles are present, and you can actually see water droplets on here. I don’t recommend this for your sole way of detecting water, but it has the capability of doing that if you look at the pictures.
Bryan Kneipp:
And it has the capabilities of generating some custom reports. And let’s get into some of the gas turbine stuff. I’m running a little behind here. So, here is a project where we put the S120 on the gas turbine, the hydraulic starting system, the generator bearings, and the generator gearbox. And what that looked like is, we ran all the sensors to a central gateway, sent through the encrypted tunnel, and into the cloud. So, these are the different things we look at on the different components. Obviously, we want to look at cleanliness all the time. The bubble level monitoring is important, especially in gas turbines, because we want to look for any type of gas leaks or air leaks or anything like that, how are these bubbles getting into your oil? And the other thing we looked hard at was the start-stop events, and how that affects the wear of your components.
Bryan Kneipp:
So we were able to use the ISO 14 count as our lagging indicator of imminent failures, and the ISO 6 count and below as our trending tool leading up to a failure. This graph on the left is showing generator speed, and then the ISO 14 count. And you can see, as the speed goes up, your count goes up. As it comes down, it goes down.
Bryan Kneipp:
But what’s interesting is the effect of a start-stop event. As you can see here, it stopped. And here’s your start. The particles went through the roof. And here’s another one. So the start-stop events are really, really rough on the components of the gas turbine. Here’s another graph. The blue is ambient temperature. And the orange, again, is your 14 count. And the black is speed. So here’s your start-stop event. And we were actually able to see that, in the height of the day, the hottest part of the day, is when the speed was fastest and the ISO 14 count was generating the most. And then as we came into nighttime, the temperature goes down and your particle count goes down. Just some interesting overlays.
Speaker 3:
Just a time check. You have about five minutes left, Bryan.
Bryan Kneipp:
Okay. I’ll hurry up. The red here is generator load and gas turbine speed. Again, looking at start-stop events. Every time there was a start, we have a high generation of particles. And here’s just an example of trending here. The red is the total particles. And the fatigue wear is the brown here. We typically see more particles of fatigue wear than any, in the turbine applications. So this was a trend that led to an eventual failure.
Bryan Kneipp:
And I was talking about the water earlier. So here’s just a film. I’ll get to the bad part. Those little bubbles you see going through the screen are all water droplets. And there’s a lot of them. So the customer knew at that point that they had a ton of water, and were able to perform a corrective action. So basically, the reason to use technology like this, and any technology when it comes to predictive maintenance, we’re trying to maximize the life of your equipment. So we want to achieve savings of between 5 and 20% of your total cost of ownership. And we want to minimize downtime as much as possible. So we think that you can reduce your downtime by at least 10% by monitoring your equipment with this technology.
Bryan Kneipp:
And lastly, we want to extend the life of your oil. A lot of turbine operators are really maximizing the life of their oil so that, you guys do a really good job of that, but it does have that capability of extending drain intervals. And here’s my contact information. If anybody wants to see more or hear more about some of these case studies, feel free to reach out to me.
Speaker 3:
Thank you, Bryan. Very, very nice presentation. Looks like a product we should all have. It’d certainly save a lot of time and effort pulling manual samples and gives a real-time indication of what’s going on. At this point, we have about two minutes left. If anybody has any verbal questions, please unmute yourself and feel free to ask Bryan your question.
Speaker 3:
All right. There is one question on the chat for you, Bryan. Question is, what is the typical installation? Is it on the supply line to ensure the protection of the equipment monitored, or on the drain line to determine the condition of the equipment monitored, or maybe both?
Bryan Kneipp:
That’s a really good question, actually. And my answer is, what are you trying to monitor? I would say both. But some people will use this to monitor actually the useful life of their filter. So they’ll have one before and after the filter. And they don’t use pressure drops to know when to change the filter. So, you can get your beta ratio that way. But typically, we want to see it on the return side, downstream from your actual components. Ultimately, we’re looking at your cleanliness code, but we also want to monitor the wear particles being generated.
