Podcast: Play in new window | Download
SkySpecs CEO Danny Ellis joined us to talk about the fully automated wind farm, and his vision for using data and autonomous drone inspections to revolutionize operations and maintenance in the wind energy industry. How can Skyspecs’ automated, high-tech wind turbine drone inspections impact the wind industry? And how can their proprietary Horizon software use data analytics to make operations and maintenance easier and more cost-efficient than ever? He shared his big vision with us in Uptime Episode 13.
Learn more about Skyspecs and what they can do for wind farm operators.
Learn more about Weather Guard Lightning Tech’s StrikeTape Wind Turbine LPS retrofit. Follow the show on YouTube, Twitter, Linkedin and visit Weather Guard on the web. Have a question we can answer on the Uptime Wind Energy Podcast? Email us!Â
Watch the video version on Youtube.
Full Transcript: Uptime EP13 – SkySpecs CEO Danny Ellis Wants to Create the Automated Wind Farm
Dan Blewett: [00:00:00] Alright. Welcome back to Uptime. So we’ve got an awesome guest today. Danny Ellis, the CEO of SkySpecs is here with us. Allen. So having already recorded this conversation, what were some of your takeaways from, from Danny Ellis?
Allen Hall: [00:00:21] To me. It’s about the technology, Danny and SkySpecs are bringing it to the wind turbine market.
They are really pushing the envelope in terms of technology. It’s not a drone company anymore. It is now a data company and a, a management company to make sure that you wind turbines are operating at their optimum and you can prevent. Really severe damage by using the technology in their horizon software.
It’s an amazing piece of equipment they have in an amazing piece of software. So a very interesting interview with Danny.
Dan Blewett: [00:00:57] Yeah. I, just hearing the way they can do these models of, of the wind turbine, the way their drones, you know, their operators basically take them out, hit star and the drone.
Automatically flies, maps out the, you know, all three blades, you know, scans for damage, and then they can give that all that data back to their customers. And the horizon software is pretty intense. I, I mean, some of the things you just wouldn’t think about, like, Oh yeah, I could get my own drone and, and inspect my own wind turbines, but.
He rings up a lot of really good points, which is you’re going to get that close to the blades, even when they’re not spinning without crashing that thing over and over and over to do 15 wind turbines in a day, which is, which is what they can do. And to have them doing that in an automated way. Cause that’s, I think that was my big takeaway is how they were engineering for scale.
Not just for like, yeah. Anyone could grab a drone and just go survey a drone or a turbine for one day, but can you scale that? Right? That’s why it’s gotta be automated. Right. And, yeah, that’s that’s right. And then that’s not the point. Anyway, the point like you said, is that they need to, so these wind turbine operators need that data to say, okay, well we know we have cracks in these blades.
These ones need repairs. These won’t have some minor damage that we don’t need to repair or Hey, SkySpecs. Like what do we, what do we do with this data? Like, do we need to repair every wind turbine blade? No, this one has a little bit of damage. Should we do that one now? Or should we put that off for the future?
You know, what’s the best bang for our buck. and that seems like the big selling point for their software, which is, Hey, we’ll inspect it. We can do it really efficiently. And. We’re going to give you all the data and we’re going to help you interpret it and say, and give you some suggestions about what you should do with it.
Allen Hall: [00:02:38] Yeah. And it’s also the ease of use of the software. It’s, it’s very interactive. Things are on the cloud. you can access it at any time and it’s in the data is secure from the second. The data gets into the drone to when it gets up to the cloud to when, people look at the data, all that’s a secure link.
So they’ve, they’ve really thought through the process and made it really customer friendly.
Dan Blewett: [00:02:59] So without further ado, we’re going to turn to our conversation with Danny Ellis CEO of SkySpecs. So SkySpecs, is often referred to as a drone company, but, is that exactly how you see your company?
Danny Ellis: [00:03:11] No. We actually see SkySpecs as a wind company.
we’re really trying to work to optimize operations and maintenance of wind farms. We just happen to use pretty advanced robotics and other data analytics, tools and capabilities to do that. And really the drone is the component that’s in the field and kind of seeing first. So we are often viewed as a drone company, but all of our innovation comes in the software, the data analytics, the predictive maintenance, helping to prioritize repairs.
It’s really in what we find with the drone, where we’re really advancing innovation.
Dan Blewett: [00:03:43] So if there was a category that you said, like, this is where we fit in. Is there a, currently like an existing product category service category? Or do you guys kind of define your own? Like where do you feel like you fit in there?
Danny Ellis: [00:03:55] Yeah, that’s a great question. We, often feel like we fit in operations and maintenance, but we’re also trying to speak with, you know, Customers that are in the data science side of the world, in the project planning in the longterm maintenance. And often those are not the same groups of people. And so it is somewhat a new category where predictive maintenance is relatively new in the wind industry.
the wind industry is not all that old, you know, particularly the 1.5 megawatt machines and above, and there’s a lot that’s unknown about them. And so it’s really a combination of multiple groups that are working on this. And so we don’t squarely fit in operations and maintenance. Oh, that is what we call ourselves.
Cause that’s what we’re really trying to optimize.
Dan Blewett: [00:04:39] Okay. And so you guys have actually like a pretty big new software, push, right? So your horizon software, what what’s, what’s the driving force behind that? Like what was the pain point that you identified in the market that, that spurred a, your development with horizon?
Danny Ellis: [00:04:52] Yeah, the, the pain point really came once we started dumping thousands and hundreds of thousands of images of defects in damages of blades. On customers. we needed a way to organize that data, to sift through it, to identify damages. And then we needed a way to help predict what was going to happen next and how to actually prioritize the, the repair campaign.
And so we realized, by trying to solve one problem, taking a picture of the blades, we created many more problems of going through that data. And that’s when we decided to go, you know, build horizon. Is this really central repository of. Understanding, you know, the health of the blades and what to do about it.
And that’s, that’s really started to take off and become the center of the conversation for repair companies, owners, operators, OEMs, warranty, claims, you know, those kinds of capabilities.
Dan Blewett: [00:05:41] So you’re saying companies didn’t want to just say like, Oh, Hey, SkySpecs came to our site and they said, we have this huge pile of problems.
Here you go. Like, see you later. So you guys were trying to say, okay, here’s your pile of issues, but. This is what we think you can do with it. I mean, is that the kind of the general point?
Danny Ellis: [00:05:58] Yeah, we realized we had to actually bring some expertise to our customers. Many of them don’t have teams upon teams of engineers.
There are small teams and none of them are really focused on one particular aspect, like blades. The owners rightfully have teams of engineers focused on the entire farm. but not nearly as specific as a crack on, you know, certain blade type. And that’s where we realized we had to bring that expertise, both from the data analysis, but also from composites and manufacturing and really understanding what’s the root cause.
And how do we plan a fix and, and really sort through the data. That was really the biggest challenge. Once we started presenting all these images. And Danny. So
Allen Hall: [00:06:41] how does the customer interface with that software? Is it a cloud base software? Is it local to their computers on their desks? How does, how do they interface
with the software?
Danny Ellis: [00:06:50] Yeah, horizon is entirely cloud based. so therefore it’s accessible from anywhere. it really helps streamline the data capture analysis and reporting process. So once we’ve gotten the data from the field, the drones at the end of the day, upload everything to the cloud and runs through our analysis process.
And then the, the customer’s able to see the fully interactive report. They log in. They can go to the static reports of, you know, telling them what’s wrong or they can dive in dynamically and see all the images and measurements and notes we’ve taken and put onto the, the reports themselves. And then they can start scheduling followup and seeing trends over time and filtering the images.
And, you know, it’s super interactive and that’s giving them the tools they need. But not necessarily forcing him to look at every single square inch of each plate.
So it’s, it’s an accumulation of information. So if, if I’m the, if I’m the user of that information, I don’t have to go look through every possible file to dig out this information.
You’re, you’re accumulating it and providing some oversight on it so that I don’t have to go dig every single blade.
Right. Yeah. We don’t want customers digging through every single blade. We want them to have the ability to, you know, how the data is there and they can go filter it and sort it, and we’ve built a lot of analytics tools in, so they can start to get macroscopic views.
but we definitely want to give them the action items. You know, it’s not about the picture. It’s not about the drone. It’s about what are the next steps and how do we help reduce costs over time and extend, extend the longevity. of the wind turbines themselves.
Allen Hall: [00:08:20] That’s just brilliant. Brilliant. Let’s
Dan Blewett: [00:08:24] talk a little bit about the damage that you’re finding.
So obviously you go open and you do all these inspections and you get all this data. where’s. Where’s that damage coming from? Is it from defects? Is it, is it solely from the environment from wear and tear or what, what has the, has that big, I guess pie breakdown. Yeah,
Danny Ellis: [00:08:41] there’s a wide variety of damages and defects.
We’re looking for. many of them are wear and tear and you’re going to have wear and tear rain, pollution, ice. It’s it’s going to erode the blades. You still want to keep track of that. UNC, how quickly it’s happening and then schedule the repair at the proper time. You definitely don’t want to fix it at the first sight of it.
And then every turbine will beat down every few months and it just would not be optimal. So we’re definitely not telling customers fix everything that we find. that’s that’s not the right strategy at the same time. You know, looking for something is also not the right strategy. And then you’re just waiting for a catastrophic failure.
And that kind of strategic failure, it could be, you know, due to a lightning strike that ends up splitting a blade, but it was more likely due to a manufacturing defect. And you know, we’re not also not saying that every blade has been made poorly, but there’s always going to be kind of a balance between making them affordably, making them quickly.
Yeah. And also making them impervious to damage. And we’re really saying. You know, it’s probably at a pretty good balancing point, but you’re going to have to have ongoing maintenance and that is going to be the optimal solution.
Dan Blewett: [00:09:51] So where do, so where do you find the balance? So obviously there’s, there’s, different categories of damage, you know, from one to five and I’ve heard those aren’t exactly standardized, right?
I mean, it’s, it’s not exactly like a utterly precise there’s some interpretation. Is that
Danny Ellis: [00:10:05] right? Yeah. I don’t think any two of our customers agree on the category. One to five scale. And, you know, I
Dan Blewett: [00:10:12] don’t think boxing, boxing, gloves. No, this is the category, right? So three,
Danny Ellis: [00:10:16] I swear. Yeah. I think what it really comes down to is understanding their risk tolerance, their budgets, their strategy.
some customers want to take advantage of fixing as much up there as possible. If they stop the turbine. Others, they want to, you know, minimize their cost on materials. And so it’s not that any one strategy is right, but that does lead to disagreements over what is it? Category three or what is a category four.
And so we’re really trying to help customers, kind of even that discussion out a little bit. we’ve built what we call a, an algorithm. We call Oscar that is our translation mechanism. And so we put in our categorizations and fairly objectively as objective as we can be. We’ll say it’s a crack of lengths X, or it’s a blemish, or it’s a Rojan, but then we allow the customers to adjust their rule set.
based on their strategy and their, their tolerances, and then that translates it to their categorizations, which then fits into their prioritizations. so I don’t know if there’s really a right or wrong, but there’s definitely a translation that needs to happen.
Allen Hall: [00:11:22] So in this Oscar, evaluation and that’s sort of our Homebrew thing, you’ve come up with that based upon all the manufacturers over time.
Does that show up in horizon two? So that Oscar classification of one through five that pops up on a S on a display that says, Hey, this blade is a five. We need to do something about it today.
Dan Blewett: [00:11:40] Yeah. And is that an acronym or where’s that term derived from Oscar? Yeah, it’s great. It’s a grouchy test or
Danny Ellis: [00:11:47] we turned it into an acronym.
But it was originally started because our team was sketching. What this algorithm was going to be on the board. And the rules engine kept being a rectangle. And one of our guys called it a magic trashcan where everything went in and then he started calling it
Dan Blewett: [00:12:06] off
Danny Ellis: [00:12:08] and was nailed it. Oh, that is the origin of that term.
But it does stand Sanford. Something that I do not call anymore. but we’ve branded it in colonoscopy. yeah, that is, that is a module on top of horizon. So not all customers have that. Wow. Customers are able to purchase that as an upgrade. and so customers that have custom rule sets and different strategies, they will take advantage of that.
Allen Hall: [00:12:32] So the horizon software is not forcing anybody in, into a particular box. You’re allowing each of the OEMs to, to go ahead and design the thing that they want using the data. That’s that’s that’s. Brilliant. And it it’s a smart way to go about it because different part of the world have different parts of the world have different levels of things they want to address at different times.
If you’re on, if you’re on the sea, you want to take care of things that are salt, water related, and it can do a lot of damage. If you’re on a mountain top, you’re gonna have different kinds of requirements, just access, you know, and doing everything at one time, probably when we were there. Well, that, that makes a lot of sense from a software standpoint.
Danny Ellis: [00:13:08] Yeah. And it’s, it’s really, you know, we’re not trying to force any strategy. We’re just trying to produce more tools, more data, more direction that can help them make these decisions. And there’s, you know, the, the idea of optimizing is going to be different for every situation. if we can make those decisions faster though, then they will be better off.
Dan Blewett: [00:13:27] So speaking of development, you guys raised money recently, so you raised $17 million. where, what sectors of your business did that get funneled into? I mean, is that like a big push for horizon or what new markets are you guys going into or where as CEO are you using those new financial resources?
Danny Ellis: [00:13:49] Yeah, our goal is to really run a robotically run, wind farm. And so as much innovation as we can pour into automation, that’s both into physical robots for data collection, maybe doing minor repairs someday, but it’s also automation on the data and decision making. And so a lot of our investment dollars to go into more R and D and more innovation, a lot of that’s on software.
A lot of that is on, on data services and helping customers understand how to manage their data. but it also does go into more geographic expansion. We have an office in Ann Arbor, Michigan, and another one in Amsterdam. Well, we are also operating in 23 countries right now and looking to expand further.
So there’s definitely an operating component, that, that the money is going towards. But most of the venture dollars that we raise are for innovation.
Allen Hall: [00:14:39] So when you’re talking about making a, sort of a drone based inspection system, that’s at a higher level than what we even know today, because I think what we’ve done so far is remarkable.
What is that next? Push let’s select the Elon Musk question. What is the next puzzle? Ilan would tell you, Hey, we’re going to Mars, baby. Let’s go in your particular world. What does that look like? Where does this, where does this go?
Danny Ellis: [00:15:04] Yeah, there’s a couple of directions on the robotic front. one that we’re working on right now is reducing the complexity of the drone so that it can be deployed by anybody, not just a trained certified operator.
right now the drone is completely automated. The operator just plugs in the battery and hits go, but they still need to know how to assemble it. When they get to site, they still need to watch for any sort of air signs, if they need to take over. or reducing that complexity and then eventually working towards completely unmanned operations, meaning that nobody is actually on site.
So what that looks like it could be in the back of an autonomous ground vehicle or not autonomous truck. It could be on an autonomous vessel. You know, we’re working towards building, completely automated offshore inspections, which are, you know, many hours of boat ride hours away. Making it, so you don’t have to send a crewed vessel out there to just get a pictures of blades or in the cells or, you know, anything like that.
And so that’s the future of the, of that aspect of robotic innovation. We’re also looking at how can we use other sensors to look below the surface, capture other, other data that will be influential in decision making and predictions. And like I mentioned earlier, eventually minor repairs. We don’t expect major repairs to be done through robot anytime soon.
But we would love to be able to quickly throw up a bandaid, quickly, throw a patch in place on a crack that we find, and then you can keep running it, waiting for the repair crew to come. So, you know, that’s the, you know, the innovations we’re looking towards robotic front, away from the, the software front of the data analysis.
Dan Blewett: [00:16:40] So with the drones, so you say, you know, An operator can go there. Just hit go. So that’s to be some pretty high tech software. So the drone essentially just knows how to identify that that’s a wind turbine. And then it, me take me through that process. I mean, what, that seems crazy that this drone just knows.
Okay, here’s my target. I gotta send you here. Here’s the blade. I mean, walk me through this. It’s kind of mind
Danny Ellis: [00:17:03] blowing. Yeah. When we go to site, the drone has no concept of which make model size of tower. It’s going to look for. It doesn’t know what position the bleeds are in. There are no preprogrammed points.
The operator doesn’t need to do anything special when they get to site. They literally just put it on the ground and turn it on and hit one button. The drone then takes off and it’s completely on its own. It builds a three dimensional model of the turbine. And then from there it plans its own flight path.
And so if you set it on one side of the tower or the other side, it’s going to do a different flight path each time. But the data it collects is going to be identical because it’s actually lining up the images with that three dimensional model and only capturing the images that it needs, based on the gaps in the model.
it’s also automatically calibrating and, and resetting all the camera settings to get the right white balance, right. Focal length. On every single picture. And so it’s not just the automation of the flight, but it’s also the automation of the data capture itself.
Dan Blewett: [00:18:04] And that’s just crazy. how long was that?
Let’s go back to the, sort of the origins of your business. I mean, If that’s today, and obviously that’s going to continue like with your tech to improve and improve, but what were your first drone flights like? Cause they weren’t like that. I’m sure.
Danny Ellis: [00:18:19] Yeah. No, the, the very first drone flights, way back when in the, you know, early 2000 tens, we were building collision avoidance technology to start.
And so the concept was, if you could put the controller into an operator’s hands and have them fly it. And as long as they couldn’t crash, you wouldn’t have to worry about how long an inspection took. You’d eventually capture all the data. That was, that was where we started. And, I was actually the first operator to go test this collision avoidance against a real turbine blade.
And the, the, you know, operations plan was to try to crash it into the blade. And when it stopped, then we could get a good picture. And that was the first time we went out to ever do anything. we then quickly realized like this isn’t going to scale commercially. If we have to rely on someone kind of haphazardly flying around and we decided in the beginning of 2016, we had to automate the entire process.
And we had to automate all three blades, not just a single blade, which is what we had done before. automating all three blades allows you to stop it in any orientation and not worry about wind speeds that day, or, you know, actually touching the tower at all. And that was really important in getting a scalable operation.
Danny,
Allen Hall: [00:19:33] the technology is amazing. It’s it’s all autonomous. What are the constraints on the OEM side? Do the turbine has to be stopped. The blades have to be stopped. Is there any, any other thing or limitation? Wind speeds. Day night. What are the restraints there?
Danny Ellis: [00:19:50] Yeah, they don’t really have to do a whole lot.
when we go to site, our technicians are trained to stop the towers. And on some sites, if they don’t have direct access, it’s usually a simple phone call to stop the tower. We don’t need it stopped in any particular position. The drone will function, you know, in any orientation. we try to stop one blade close to vertical, you know, standing up in the 12 o’clock position, but that’s simply so we can get a good angle on all the blades for a picture.
the drone itself doesn’t physically fit between the blade and the tower. So if it’s in alignment with the tower, we’ll just miss part of the trailing edge of one blade, which is not ideal. So we, we try to stop close to vertical, but it doesn’t need to be precise in any way. we don’t fly at night mainly just due to FAA restrictions right now for night operations.
and our, our technicians know livelihood. we don’t want them running 24 hours a day right now, but we have talked about, you know, someday expand, extending to night operations, for certain applications. We don’t usually fly in rain. but that’s also just because water gets on the lens and then the picture looks bad.
the drone is fine, but, we don’t need to waste time if we’re going to get bad pictures and then wind speed limitations. we can fly in 15 meter per second, which is 25, 30 miles an hour of wind. And we can go up until about 18 meter per second gust. And so it’s not often we get stopped for wind speed.
Usually if it’s that windy, there are bad storms in there anyway, so we’re not onsite, but, you know, we don’t have too many wind limitations.
Dan Blewett: [00:21:20] And then, so you mentioned potentially in the future of being able to reduce some of these long boat rides to offshore wind, what do you guys do now to access some of these really remote?
either offshore or just like. Very crazy, like out in the remote mountains
Danny Ellis: [00:21:33] kind of areas. Yeah. The offshore ones are sometimes really crazy. I’ve been out there a few times and we send our teams on vessels, for the ones that are close to shore, they take the boat out in the morning. They inspect a handful of towers.
recently we’ve been doing between 16 and 18 offer inspections in a single day, which is just in crazy it’s crazy numbers. And, but the ones that are really far off shore, we’ve done a job that was an eight hour boat ride. And so our team had actually, you know, boat out one day, stay on the flow tells, which are these huge, like, you know, basically cruise ship, like, you know, hotels that are on water.
And they stay there for two weeks straight, each day going out to inspect some towers and come back. And then the end of the two weeks they take the eight hour boat ride back. And so it’s a, it’s pretty challenging.
Dan Blewett: [00:22:23] So say that it’s only two hours away or say they’re in that flow tell, and they’ll go out and inspect a bunch in a day.
Do they do it from the boat itself? Or do they always sort of dock and get out on like the base of the, I know a lot of these turbines don’t have a base where they could really stand what’s the procedure. Like they’re just out of curiosity.
Danny Ellis: [00:22:39] Yeah. They, they don’t have to touch the tower at all. which is a great, concept for us, you know, because once you touch the tower, you need significantly more training.
If you don’t have to touch the tower and dock your training is, is just simple offer, training. It’s, it’s not the docking training. They do launch and land from the deck of the boat. And so we clear out the deck, we put, some fake grass down, which is really just a way for a soft landing. and then they take off and land right there.
Dan Blewett: [00:23:08] And even with these like really high, I mean, you see some of these offshore wind farms are like crazy high seas and. Oh, are these like big vessels they’re taken out there? I mean, what does this, I know this is just like a fascinating picture for me because it’s such a harsh environment. I mean, that’s why it’s, so they need to obviously to do these drone inspections, but how do they navigate some of those, those high seas?
Danny Ellis: [00:23:29] The seas can definitely be high. So there is an additional limitation when we’re off shore is wave height. So we don’t sail on the smaller vessels when it’s about one and a half meters of, of wave height, that makes it extremely challenging to sail and extremely challenging to land a drone. so we do usually need some, you know, calmer seas then what you might see typically, the weather windows are harder to come by.
but we. We definitely don’t operate on large enough vessels. Most of the time it is a, you know, they get tossed and turned for sure.
Allen Hall: [00:24:00] Lack of seasickness is one of the qualifications to go do this job.
Danny Ellis: [00:24:04] Oh, a hundred percent. And I do not qualify
Allen Hall: [00:24:10] well, you know, it’s one of the things about the off shore.
Turbines is the towers and obviously all the salt and the corrosion that goes on there are you inspecting the towers too? And those highly corrosive environments,
Danny Ellis: [00:24:22] we’ve done a few, one-off inspections at the towers. it’s something we’re looking to automate in the future. it’s definitely, you know, getting hit by the elements and we’re also looking at what can be done below the surface of the seat.
there’s a lot going on down there that needs to be monitored as well. And if we’re going to go out there and we might as well go underwater too, probably not something we’re going to innovate ourselves, but it might be something we do in collaboration with another company, to have, a conjoint, operation, you know, and just sail once if possible.
So we’re looking at those kinds of applications, to really, like I said, get to that robotically operated, wind farm, and be fully comprehensive of. all the data we need to capture.
Allen Hall: [00:25:01] So are you talking about an above water, drone and a below water drone? Or are you talking about in above air drone
Danny Ellis: [00:25:07] above, above
Allen Hall: [00:25:08] the water drone that can actually see
Danny Ellis: [00:25:09] underwater?
No, it would be a below level, a below sea level drone, I guess you can call it. They call them ROV is usually a, and those have been around long before aerial drones were, so that’s a more of a well known field that has done some pretty cool things with their technology. we just don’t have any of them to deploy.
but I would look to, to work with someone who’s doing that right now and see if we can make everyone’s operation cheaper.
Dan Blewett: [00:25:34] It’s funny. It reminds me of this, this movie, the life aquatic with Steve Zissou. They had these dolphins following the boat and they had like these things on their heads. I’m like, you know, these, these dolphins, they they’re supposed to like follow the boat and like kind of protect us, but we never found any evidence.
They’re actually intelligent. I was like bill, bill Murray’s characters, a quote in that movie, doubling back to a. To some of the technology that’s out there, obviously you’re trying to push the envelope with software and your own drone tech and all the autonomy. What are some of the new emerging technologies?
and are any of them, you feel like, man, this has a lot of promise, maybe for your company or even just, just in the industry in general, doesn’t have to be in the drone industry.
Danny Ellis: [00:26:15] Yeah. I mean, there’s a lot of innovation, across the entire wind industry and it’s, it’s amazing to watch each and every year.
some of the things that we are looking toward is just other repair techniques and technologies, a ways we can fix things earlier, more affordably, whether that’s a new coding, a new leading edge protection, or whether that is a robot doing the repairs and likely that robot isn’t flying the robot is probably clamping on in some way.
and we’ve seen some of those in, in early tests and they’re still quite expensive and time consuming to deploy, but I’m really excited about where they’re headed. because I think when you can choose to repair something faster and cheaper, you’re going to choose to do it more often and therefore extend the life of the blade.
And that will just go into our repair planning and predictive maintenance models. That you can capture this stuff earlier. So I’m pretty excited about that.
Dan Blewett: [00:27:11] So yeah, this, that kind of, that’s a good segue where I want to go back, which is to some of your customers and who tends to buy from you guys and who doesn’t so real quick, who is your, your main customer and who’s, who do you hope to capture more of in your, in your market going forward?
Yeah.
Danny Ellis: [00:27:27] We, work with customers across the entire value chain. So we work with large owners, small owners. We work with the OEMs and we even work with some of the independent service providers. the way we view it, we’re trying to bring like truth of what is going on with the structures. If that is when it’s under the OEM’s a warranty period, they want to know they want to be able to fix things before they get worse.
If it is after the warranty has expired, then we’re usually working with the owners on a longterm maintenance plan. And if it’s in the middle of a larger repair campaign, we might be working with the repair companies directly, either providing, inspection capabilities or the software itself. We have a handful of repair companies now that are using horizon to log their repairs and track their entire campaigns.
So now the data on the blades is all encompassed in horizon. we like to view horizon as a complete health record of the blade. And so it’s not just what’s wrong today, but it is what has been wrong over time. What has happened from the repair companies who has fixed it? How did that go? Did that hold up?
Was that a good decision? all of that being in, in one central place is, is our goal. So we do work with everyone in the industry, in, in that value chain. well, we are looking to expand it further and get beyond blades and start connecting other types of data. The best way to understand whether our decisions on repairing the bleed has been optimized.
Is, has it affected power production? Has it, is it affected uptime or minimize unplanned downtime? And really it’s connecting it to all the other services, in the wind farm that will help answer those questions.
Allen Hall: [00:29:11] How do you get that information back into something like horizon? Do you track power production?
Can you monitor that remotely? Say, Hey, we identified these issues with these blades that got repaired early, and therefore there’s been an uptick in uptime. Is that something you can track?
Danny Ellis: [00:29:27] That is something we are working to track. we have a few early customers that are connecting that data themselves, but we’re working to help automate that connection.
so it’s just a part of the analysis. It’s not there yet. Right now. It’s a very manual process with the customers that are willing to sign up for it. But we really think that any data you can bring on, on the entire life cycle of the turbine. Well help determine, did you make an optimal decision on any of the maintenance, whether it’s the blades, the gearbox or otherwise?
all of it is connected and we think the data should be connected as well,
Allen Hall: [00:30:00] so that the uptime percentage. It was, it was where we’re all trying to get to is to have more operational time. That seems like the critical factor in all of this. So your company is all about getting to uptime. That’s what it seems like to us is that all the focus on, Hey, let’s, let’s look at maintenance.
Let’s look at repairs. Let’s look at lifetimes. Let’s look at predicting what happens in the future, but the whole, the whole thrust of this is to get us to have more power generation in the future. Does that seem like the right approach?
Danny Ellis: [00:30:30] Yeah, you’re pretty spot on. we also phrase it as trying to minimize unplanned downtime because there’s always going to be planned downtime to run the turbine optimally.
And so we don’t want that to kind of count against the optimized numbers or shooting for. We also can’t control the weather or we can’t control the grid demand. And so until there is good energy storage, there’s going to be planned downtime. but if we can help minimize unplanned downtime, that’s I think in the near term where the optimization can come from.
Dan Blewett: [00:31:01] So I want, I want to talk about warranties a little bit, cause I know you guys have kind of been, I don’t know if disrupt disruptor is the exact right term, but you guys have, I know provide a lot of companies. With the opportunity to really get a lot of their wind turbines fixed before their warranty expired.
Can you speak to that a little bit?
Danny Ellis: [00:31:18] Yeah. We, entered the market, largely doing end of warranty inspections. And initially we spoke to customers who had either missed their warranty, or they only inspected maybe 5% of their towers and kind of extrapolated the health of their farm from there. And a lot of that was just due to time constraints and cost constraints.
If you were doing the inspection from a rope crew, it could easily take four to eight hours for a single turbine. And so it just wasn’t feasible to get a comprehensive, a inspection of everything before the end of the warranty. So we, entered the market and started producing these reports of every blade and allowing them to actually file these warranty claims really trying to help them get the best performance out of the towers once the warranty expired.
And what we were finding is that if they let it run because they thought everything was okay, one, two years after the warranty expire, they realized then they have either a series of, catastrophic defects. Or serial damage. And now, you know, they can’t do anything about it. The OEM is it’s off their plate, and now it’s entirely up to the owner to repair.
And it was really causing issues on both sides of the equation, because now the owner’s frustrated with the OEM. They don’t want to buy from them in the future. The OEM’s not really that happy that the owner’s frustrated, but at the same time, they don’t have a financial or legal obligation to fix it.
And so we’ve really tried to help make that conversation more transparent and happened earlier so that everybody that is playing a role in that warranty, understands what the optimal approach is.
Dan Blewett: [00:32:53] Yeah. That makes a lot of sense. So when customers don’t go with you guys, what it seems like this is pretty cut and dry.
Like, Hey, let’s get your stuff inspected and fixed before the warranty runs out. what, what common reasons do you have that people don’t choose you?
Danny Ellis: [00:33:09] Yeah, it’s a, it doesn’t happen all that often, but, either they’ve chosen to skip doing a comprehensive inspection altogether. If they’re really running on a tight budget, they may not schedule the inspection, or they’re taking pictures from the ground, which is another feasible method of doing an inspection.
But. While it is cheaper. It’s not nearly as comprehensive or nearly as measurable. And so when it gets to the warranty, a lot of it comes down to what is the length of a crack or the length of erosion, and you need an accurate measurement and an accurate position on the blade. And, you know, we’re able to provide that.
So really high degree of accuracy that a ground-based inspection might not be able to capture, might be able to get the right angles and the right lighting to really capture it. And, and often it can miss the minor cracks that we’re able to see by being so close to the blade. other customers, this was more, many years ago compared to now, they used to view us as just the drone flyer and no added technology beyond that.
And so I would look at that and say, you know what? Yeah, I’m going to go buy a drone and I’m going to fly it myself. And we saw that a lot and we did a call a few weeks later, say, yeah, I crashed it. Or it’s going to slower. The video is not all that good. And, you know, we then explained to them why that is not the best method for capturing this information.
Dan Blewett: [00:34:26] Yeah. Especially, I mean, I’ve flown drones a little bit. Like I’m a very novice, very, very beginner, but to get really close to it, Turbine blade, like, like you said, it’s really, really difficult to get where you need to go and have a spatial awareness, even though they have the point of view, you know, in the camera, you know, at a D a DJI model.
and I also crashed mine too, so yeah, over and over
Danny Ellis: [00:34:49] and over. Well, yeah. And then imagine doing that and then imagine doing that on repeat for 33 turbines in a day, and you’re going to get tired. You’re next going to hurt. And, you know, it’s not repeatable, it might work once or twice, but it’s just not repeatable.
Dan Blewett: [00:35:06] So what, what makes your drones different? Cause you guys aren’t manufacturing your own drums anymore. but I’m sure you have a lot of really intense specs that you demand out of them. So what makes your like hardware. Unique.
Danny Ellis: [00:35:18] Yeah, we, take advantage of as much, off the shelf components as possible.
so we do use a cinematography great drone. It, it’s got six propellers. It’s, it’s quite large. It’s got a heavy payload. And then we customize the payload that goes on top that provides the autonomous navigation, all the sensors and software. And then we also have a custom inspection camera that goes on that payload.
And so that payload is really the core of our innovation on the robotic side, not the drone itself. And then from there, you know, the, the technology, in the, in the payload is transferable to other, you know, heavy lift systems. We’ve just found one that we really like, we just stuck with it and it’s, it’s been highly durable, built in the USA.
you know, it’s a company in Seattle called free fly that, has been awesome to work with. And that has allowed us to scale globally very quickly by not also running a manufacturing line for groans. Obviously,
Allen Hall: [00:36:15] if you’ve missed that future catastrophic blade defect, the expense of that just totally wiped out any possible savings you were going to have in going to a lesser drone company.
But I think on the second part of this, which is the whole industry is driving itself towards. Better safety record. Anytime you put someone out on a rope to do an inspection, they’re at risk. I don’t know that. And the record has been really good in the, in that area, but it’s still a risk. And in terms of mitigating risk, it’s just seems like.
Don’t want to have people on ropes and spectrum blades in these environments, particularly off shore that it forget about the cost for a minute. You just can’t have someone get seriously hurt or killed, inspecting a play that we were beyond those times, right?
Danny Ellis: [00:37:04] Yeah. We’re, we’re pretty much beyond those as a standard method for inspection.
the only reason to send someone on ropes today for an inspection, not for a repair, but just for an inspection. Is if we’ve already identified a defect that we now need a subsurface inspection. So if they need to go up there and scan it with an ultrasonic or start to grind it away, to see what’s going on below the surface, that is, is when we need the rope crews.
And, you know, we’ve worked with some of the road crews and they definitely have a very good safety record. but they also recognize like the fewer times they go up, the better their safety record is going to be. And so we really only want to send them what we have to, not just to take a picture. And I think, you know, we’re, we’re beyond that now.
the other, reason that rope crews are not great at doing the inspection is the data it’s just not digestible and something you can put into any sort of big analytics. it’s requiring them to take a picture and measure where they are in the blade. And. And that doesn’t scale either, right? So you miss out on a lot of information doing it that way.
Allen Hall: [00:38:08] Well, and your information you provide to the technicians that get out on the ropes. You’re identifying exactly where on the blade they need to go, so they can bring the tools and equipment with them. So they don’t, they’re not going out there to do a pre-inspection to figure out what’s going to happen next.
They know they’re loaded to go. They’re going to do it one time and one time only, and then move on, which. Obviously shorter amount of time is on the rope, but also makes the repair much quicker because you have provided them with the data.
Danny Ellis: [00:38:32] Yeah. And they’re, they’re not logging into horizon themselves to quote the jobs in the first place.
And then when they go to the field, they look at, at horizon to figure out which blade, where on the blade, what material do they need to bring. And we’re starting to see an effect of this, where we are minimizing. Having to redeploy crews cause they either showed up to the wrong tower or they repelled down the wrong blade or brought the wrong material.
you know, trying to cut back on those inefficiencies is a, an area where horizons playing a big role. And it’s all because we’ve been able to consolidate and analyze the data in the first place because
Allen Hall: [00:39:08] it’s the whole package comes together. It just seems like you finally reached that precipice where the industry needed to get, which is to have.
Data to make the efficiency increase substantially to the, to the point where one, we’re not putting people at risk to the winters. Our uptime is starting to get incredibly high in three. The reliability of just the, the systems are getting so much better. And I know there’s a recent report talking about.
More modern wind turbines, more direct drive, wind sort of bins lasts longer, have fewer defects, but that’s, that’s a result of some of the stuff that you have done that you have provided that data back to the industry to say, Hey, these kinds of wind turbines, these kind of geared systems are less reliable, quote, unquote, reliable than other ones because that data which you’re accumulating.
Provides so much knowledge and insight to the rest of the world. It’s a remarkable thing that you’re doing right. Is, is where is that data stream going from here on out? Where does that data stream go
Danny Ellis: [00:40:08] from here? Yeah, it’s a, it’s a really good question. And you know, there’s a lot, that’s playing a role in that increased longevity.
We’ve seen a dramatic shift in customers, mindsets around maintenance five, six, seven years ago. It was a reactive maintenance plan. It was, you know, decided to run until failure and then choose what to do about it. Then it moved over to proactive, which is time-based. So do something once a year or do it once every other year, we’re really helping to push customers to a data driven approach, which is what we’re calling predictive maintenance.
And it’s not necessarily saying we’re going to know the exact date something’s going to fail, but if we can help predict when to do another inspection, when to take action. That data-driven approach is now going to feed into the entire system. And you’re going to see more uptime and more longevity because of it.
We’re seeing similar trends on the gearbox, and yes, there are pros and cons of direct drive over gear. But I think just keeping up on the maintenance is helping to make some, some turbines look better than others. Because of the, approach, of how they’re going to actually do the maintenance punches of data security.
And so the data is only going to the customers that are paying for that data. but they are then, you know, using that in, in warranties and in discussions over future innovations themselves to figure out, you know, how can they maximize the longevity of these towers, you know, down the road. And this is going to become an even more important conversation when tax credits cease to exist.
And we’re, we’re at that transition point now where there are fewer and fewer tax credits and they’re phasing out. And if the turbines don’t last, the, their expected lifespan, then owners are gonna have a hard time keeping them profitable. And so it’s something that you’re making a small change. Now on that approach is going to have a lasting effect for decades to come.
So Danny,
Allen Hall: [00:42:06] one of the key ingredients that we hear about all the time is sort of data security. All the OEMs are very tight about sharing information. And, obviously as an operator, you don’t want to be sharing information to the rest of the world, but you have a, a huge database. How do you handle data security, from customer to customer?
How does that work?
Danny Ellis: [00:42:28] Yeah, we are very careful about the data security. So we do not share the data between customers. none of our customers can see the specific images and rule sets that the other customers are using, that is entirely on their version of a horizon. everything being in the cloud, you know, we get to take advantage of all the advancements of data security in the cloud.
horizon is entirely built on AWS. And so once the data goes there, that is where the data security is stored. It’s not sitting on individual hard drives and being mailed between people and downloaded to a local machine. And then the local machine is lost. none of that is taking place. Now we can’t confirm that our customers don’t download it themselves and put it on their own machines, but sure.
That’s that’s up to them at that point. So we are very careful about data security. We know it is of critical importance and, you know, the, the data of the customers that paid for is. Is a, is theirs to, to share amongst themselves. And if they want to take it somewhere else, to share it with the customer, then they can choose to do so.
So
Allen Hall: [00:43:34] it’s, it’s essentially a secure interface system, much like things like a simplistically saying like a Dropbox or in any kind of AWS situation where we’re handling data from one to another is basically secure from beginning to end.
Danny Ellis: [00:43:49] Yeah, exactly. The, the web interface of horizon is just like any other web interface, like, like you said, in Dropbox.
And so the data is stored in the cloud and those are the companies that are helping to secure that data. And, you know, we, we try to make sure we’re using the best in class, for this kind of data as well. additionally, you know, the data on the drone. when it goes from the drone to the, the, horizon interface, it is not really going through the hands of the pilots, that data’s not readable by the pilots.
It’s not readable until it gets to horizon. So we’re also not worried about, you know, any of the field crews having access to that data, unnecessarily.
Allen Hall: [00:44:28] Yeah, that’s brilliant. So the second that data actually gets acquired by the drone. It’s secure,
Danny Ellis: [00:44:33] correct? Yeah. That, that data is stored on the drone until the end of the day.
And at the end of every day, it is st with the cloud and it’s taken off the drone and it is now up in horizon to be analyzed and shared with the customer. And obviously our team has access to it because we’re the ones doing the analysis and providing the report. but we’re working, very one-on-one with the customer, on that data.
And not using it as an example to share with the different customer on why they should perform a certain type of repair.
Dan Blewett: [00:45:06] So in, in, in nonspecific terms, obviously not related to any given company, but given that you’ve gone through so much data and done so many inspections, what are some trends you see?
Whether it’s, you know, common types of damage or just like, is there any where the industry’s going or that they need to go to really fix a major problems that your is identified?
Danny Ellis: [00:45:25] Yeah. I mean, there are many issues that we’ve seen trends on both getting worse or getting better, in, in both directions.
Erosion is a huge issue. Every blade we find has some form of erosion. The different level of erosion, you know, that, that varies depending on the size of the tower, the length of the blades, the environment that it’s in. but it’s definitely an issue that every blade faces. we also see that there are, are, different types of cracks on new blades that may have not been completely tested in the field yet, as they get longer, they’re undergoing much larger stresses and strains, and that can lead to a more catastrophic cracking.
And so we’re, you know, paying close attention as the industry is getting larger and turbines are getting, yeah, just these massive sizes where the blades are longer than a football field. And, and they, you know, need some close attention in the, in the first few years of life. but we’re also seeing that there is more attention to, quality of, of manufacturing.
And so, you know, as we’re getting bigger, there has to be more attention to the process of, of how they’re being made. And, even though we don’t have access to that data directly, you know, we’re hoping that feeding this, field data back to the, the OEMs will help make better blades in the future.
Dan Blewett: [00:46:43] And what about lightening? So I know lightening is still a persistent issue. what effects of lighting do you see on a, on a consistent basis?
Danny Ellis: [00:46:50] Yeah, so we’re not testing the lightning protection systems directly, but we see evidence of lightning strikes all the time. we see charring on the blades and near the lightning production systems.
And for the most part, it’s just evidence that it’s been struck and it’s okay. But we’ve also seen lightning strikes leading to split split, trailing edges split ends, you know, and that becomes far more catastrophic and needs a Up-tower repair or replacement. And so it’s something that is very hard to control.
and we are looking towards future innovations to test the lightning protection systems. And if we can, validate that they’re working or not working. We’re hoping that catching the not working systems will help prevent a catastrophic lightning strike damage in the future. but that’s a future innovation, another test that we’d like to get to.
it’s pretty hard to stop them from getting struck.
Dan Blewett: [00:47:40] Yeah. I mean, and, and this is Alan, this is some of the stuff that your friends at university of Nottingham or are potentially going to be assessed.
Allen Hall: [00:47:48] So we’re doing some modeling of, of the efficiency of some of the lightning protection systems. And that’s one of the things that, some of the early work has been done is these protection systems are.
You know, somewhere between 80 and 90, sometimes 95% efficient, but they’re not a hundred percent efficient. So as the age, over time, the blades just like with the leading edges, you’re seeing as, an aging effect on the blades that we can some from lightening protection standpoint, and then you get.
10 years in 15 years in you getting this major catastrophic lightning strikes occur. And then when we got to call Danny to come in and take an inspection of the blade, it just seems like we don’t have a really good handle on what’s happening in the field. Part of the horizon effort. I think our Danny’s companies is brilliant in the way they’re accumulating data.
Is that in the lightening protection world? We also don’t have a lot of data. We don’t have a cumulative knowledge of what happens out in the field. And until we get to that big data point, we’re not going to be able to help reduce the effects of lightening. And in a case like Danny, where he’s got this ability to do these great things with drones, you know, what things to standing will look for, what should he be looking for on this particular blade to make sure that it’s lightning protection system is still up and functioning?
Yeah. That’s where we’re at today.
Danny Ellis: [00:49:03] Yeah, we, we would love to come into contact with the blade and be able to physically test the continuity. no, that’s a concept we’ve, we’ve talked about for awhile. we’re not quite there yet, and even starting to pursue it. but it’s definitely a concept of how can we robotically test the LPs systems and help prevent that
Allen Hall: [00:49:19] we got to get there.
Right. If we can get rid of, if we, if we’re working on leading edge erosion and there’s a huge industry pushing towards that and getting the data is the first step in that process, same thing, enlightening. Right. We’ve got to get to the data so we can then put the engineers on it to make it better. Yeah.
Dan Blewett: [00:49:35] Yeah. So I got a couple more topics to cover. before we wrap today, but, number one, so you guys are all over the world, but like, what are some of the, what are some of the unique characteristics about some of the markets that you’re in? I mean, do you see, your inspections go wildly differently in say Japan or Brazil versus the United States?
I mean, what are, how does the U S look, from an inspection standpoint, how does it differ from the rest of the world?
Danny Ellis: [00:49:59] Yeah. I think the, the main difference is probably in the drone side of it actually is just getting aviation clearance to be on wind farms. Some countries are really easy there, so easier than the United States, others are extremely challenging and time consuming.
And so we take care of all of that for our customers. Yeah. We don’t want them to worry about the aviation regulations of the drone whatsoever. So we manage all of that before we get on site. And sometimes that can delay jobs. We’ve also run into issues, and customs getting stuff stuck in customs for weeks on end.
Brazil is a particular problem. but once we were able to get a working relationship, with the customs offices in Brazil, we were able to get in and out and do the job. So, you know, we were sent to do there. so yeah, we’ve been on five continents all over the world, everywhere except Africa. And we, we expect to do some work in Africa this year.
onshore and offshore. And I’d say from an inspection perspective, the United States is kind of in the middle of the road. I would say Europe being a bit of an older industry has done more about caring for the blades, over the longterm. And some of the newer markets are where the U S was five years ago.
And they’re only focused on installation and they’re waiting for maintenance, you know, for a few years from now. So it’s kind of a trend that we’ve seen in, in the life cycle of the wind industry in each country. And, we’re trying to get in earlier and earlier and, you know, start to do construction inspections and not wait until the end of the warranty.
Dan Blewett: [00:51:25] Gotcha. Gotcha. And as a CEO, obviously you don’t really live in the present as much as you know, many of your employees and other executives. Do you have to live, you know, two, three, four, five years ahead. So. Where are, where are you in the future? I mean, where is SkySpecs headed? What is your product roadmap?
Where, where do you hope to gain market share? what can you share as far as the future of your company?
Danny Ellis: [00:51:47] Yeah, the future. We want SkySpecs to be known for the robotically run, wind farm. And that’s not to say there won’t be technicians. There will definitely still be technicians. We’re just simply hoping to optimize the use of their time.
So when they do grow up tower or they do go onsite in the first place, they’re there to do a very specific job and they know exactly what they need to do and do it well. And we want to take away some of the more mundane data collection and understanding of what’s going on kind of getting an, a, you know, a pulse of the wind farm itself.
And what is the status of its health? And if we can do that in a way using robotics and onsite, you know, technology, whether it’s IOT devices or something else. And if we can have it be deployed all the time, we can now get a much better understanding of exactly when wind turbines are failing and then prevent those in the future.
And so that’s, that’s our goal. We’re looking to do it through partnerships and bringing in, you know, other technologies that we’re currently not investing in, to, you know, come onto the horizon ecosystem, work with the repair companies to help facilitate their, you know, their data and data management and an understanding of how to do a repair.
And if we can help improve that conversation between the owners, ISP and OEMs. You know, we think we can get a, the wind farm to a point where it is by far the cheapest form of electricity. We can play our part in that we will feel pretty good about it at the end of the day.
Dan Blewett: [00:53:15] Yeah. That’s what it’s all about.
I mean, especially being in the renewable energy sector, like, I mean, if you’re not looking to change the world and make the environment and the earth better than we found it, I mean, I think that’s, that’s the goal for everybody. So.
Allen Hall: [00:53:29] You know, there’s, there’s one thing to see the technology and use and to watch it on the web on YouTube.
And it’s another thing to actually hear what goes behind it and how much work and effort and the team you’ve built are incredible. so all the best. Yeah.
Danny Ellis: [00:53:45] Yeah. The team is world-class, they’ve done a phenomenal job and, we would never be here at all without the amazing team we’ve put around us. And so.
You know, they, they deserve all the credit and, you know, and our customers have been awesome as well. And, looking forward to the next phase of, of the wind
Dan Blewett: [00:54:01] industry. Yeah. I appreciate you being on the shows up to great conversation and a really interesting look, been behind the scenes of a really high tech company who, and you’re, you’re dealing with a lot of companies that are very.
Like you said, sensitive to data security, not sharing trade secrets. So there’s not as much known about your company and a lot of other companies because of the secrecy aspect in the industry. And so we really appreciate you being here to kind of shed some light on, on what you’re doing. And also some of the things that are just going on in the industry that.
Other people aren’t necessarily talking about. So we appreciate it.
Danny Ellis: [00:54:32] Yeah. Thanks very much. Thanks for having me.
