Podcast: Play in new window | Download
We’re revisiting a great episode with Nicholas Gaudern, CTO of PowerCurve, discussing their new AeroVista software. AeroVista uses drone inspection data to evaluate wind turbine blade damage and power production potential. Allen and Nicholas discuss how this technology enables strategic repair planning to maximize power recovery while avoiding unnecessary costs.
Sign up now for Uptime Tech News, our weekly email update on all things wind technology. This episode is sponsored by Weather Guard Lightning Tech. Learn more about Weather Guard’s StrikeTape Wind Turbine LPS retrofit. Follow the show on Facebook, YouTube, Twitter, Linkedin and visit Weather Guard on the web. And subscribe to Rosemary Barnes’ YouTube channel here. Have a question we can answer on the show? Email us!
PowerCurve – https://powercurve.dk
Pardalote Consulting – https://www.pardaloteconsulting.com
Weather Guard Lightning Tech – www.weatherguardwind.com
Intelstor – https://www.intelstor.com
Allen Hall: Welcome back to the special edition of the Uptime Wind Energy Podcast. I’m your host, Allen Hall, and this week, we’re going to go back in time to September of last year, where I sat down with Nicholas Gaudern, CTO of PowerCurve. And Nicholas and I discuss a new piece of software that PowerCurve has developed called AeroVista.
And AeroVista is a unique tool. It takes your existing drone images and then predicts the AEP for the turbine or the particular blade. A lot of operators that Joel and I have run across recently are interested to know what blades to repair based on the amount of damage. And we see damage from all over the world.
And there are blades that have very minor damage that you probably leave alone. There are some with very major damage and those you should obviously fix. It’s the ones in between where you’re not really sure. And this AeroVista piece of software is a predictive tool. It will help you design your campaign to repair blades during the warmer months.
So it’s a very powerful tool and a and a well needed tool for the industry. So I thought it was time to revisit this episode with Nicholas Gaudern of PowerCurve. Enjoy.
Leading edge erosion is a massive power losing problem for most wind farms. Almost every wind turbine blade that has been in service more than two years has some level of leading edge erosion. I’m your host, Allen Hall, and I’m here with our guest, Nicholas Gaudern, CTO of PowerCurve.
In this episode, we are discussing leading edge erosion, how it reduces average energy production, and when to address it for maximum revenue generation. And just a brief background on PowerCurve. PowerCurve designs, manufactures and installs power upgrades for wind turbine blades that help their clients make their wind projects more profitable.
PowerCurve’s Technology has been thoroughly tested and validated, and they continue to work closely with universities to refine it even further. And the upgrades have been installed on blades worldwide. Nicholas, welcome to the program.
Nicholas Gaudern: Hi Allen. Really nice to be back talking to you.
Allen Hall: So you have some new software tools at your disposal, and anybody that knows PowerCurve knows you guys are really good at aerodynamics to understand how blades produce power.
You wanna, you wanna describe what this little software breakthrough
Nicholas Gaudern: is? Yeah, I’d love to. So, so what I want to talk about today is, is our new tool that we’re calling a. It’s it’s an a p i, you can call it, it does something and you’ll get some really insightful data back. Maybe just take a, a step backwards.
It’s all about taking a, a data driven and an engineering driven approach to understanding the performance losses that you will get from damages and particularly leading edge erosion on a blade. So, We’re about modeling those losses and telling you how you can deal with it. Yeah,
Allen Hall: because there’s a lot of information on the internet today link.
You see a lot of it on LinkedIn talking about leaning edge erosion and, and how you should repair it and it should be repaired and how quickly should be repaired. Those are really interesting data points. Right. But I think the real critical decision is if you should repair it and how. How, how far how many years can you wait?
Right? Because it’s all about spending money and spending money wisely on your turbines to keep your production power up. But there’s really is not a tool out there today that tells you, Hey, we need to repair this turbine, but not that turbine. Yeah, exactly. And I think what
Nicholas Gaudern: we’re addressing here is, is this is not a new problem.
Leading edge erosion blade damages, and they’ve been around for a long time. They also think there’s been a lot of acceptance that they cause a power loss. How you go about quantifying that in, in a reliable manner in the field consistently is something that from what we see hasn’t, hasn’t been achieved.
So we set out with the goal of, of kind of cracking this problem, doing it in a way that users common data and existing data. We didn’t want to go out and put this big new data requirement into the field to be able to do these kind of calculations. It’s happened quite quickly, I think in the last few years, but I would say, you know, a huge percentage of the installed base of wind turbines stay are inspected annually, at least annually, by a drone.
And the primary focus of that drone inspection has been to assess the structural condition of the blades to look at whether there’s any critical cranks or when you should fix them so you don’t have a catastrophic failure, and that’s super important. There’s no link to aerodynamic performance and it’s simply not good enough to say, oh, well I’ve got the Cat five structural damage.
Therefore, that’s bad aerodynamically. It might not be, you know, you might have a crack in a certain place. These kind of almost invisible aerodynamically, but as very bad structurally. So what we do is we take existing damage metadata. So all of the data is captured by the drone companies, the drone inspection companies.
Then we map that onto an aerodynamic model of the Pacific turbine in question. So for example, if we’re out analyzing a a GE 1.5 or a Vestas V80, whatever, we will have built an aerodynamic model of that turbine. So we are taking as many assumptions outta the system as we can. So we look at the real data, so the real damages that were on the turbine at that time.
It’s mapped onto an ergonomic model of that real turbine, not some generic fence.
Allen Hall: So most wind turbine operators have a bunch of drone images and data they’ve accumulated over the last couple of years. So in that dataset, I just wanna understand this clearly here. So in that dataset that the operators already have, there is metadata that describes the damages that they have on the blades, the leading edge, erosion damage, maybe some lightning damage, maybe impact damage or a crack even.
All that damage is metadata already in their existing data that they have purchased.
Nicholas Gaudern: Yes. So that data exists and that’s an important thing with, with the system that we, we can go back historically, so we’ll, we’ll go forward. There’s new inspections are made, but if you want to see what has happened in the last three years in terms of your AEP loss from leading edge damage, we can do that because that metadata.
Will exist. Obviously there’s differences between drone inspection providers and the level of detail, et cetera, but fundamentally, it, it exists.
Allen Hall: So I have this metadata and I put it through the PowerCurve, API which you’re calling AeroVista. So I take this existing data set already purchased, I sent it to PowerCurve.
You process it and tell me how much power that blade is losing and what, what that means in terms of revenue generally. And you can do that over, over a trend line over the last couple of years. So I can kind of see how that blade is doing. Is it basically being the same or is it really dramatically dropping off?
Can you can tell those sort of things just from the metadata.
Nicholas Gaudern: Yeah, exactly. And, and this is the great thing about using a real turbine model. What has been really enlightening when we’ve been adding more and more models to the system is that you could have a hundred meter diameter rotor. Designed by one manufacturing with some erosion on, and you could have another one that’s designed by a different manufacturer that basically looks the same, you know, on the photographs.
But when you actually run the calculation, you may find that one of those rotors loses way more energy than the other. And that’s because it’s not just about how the erosion looks, it’s about what is the aerodynamic shape of the blade. What do the airfoils look like? What’s their performance? What’s the store margin?
What r p m is this turbine spinning at? So again, we’re taking all this into the, into the model that is just cutting out all these assumptions that you’d otherwise have to make if you didn’t build a model. So what you’ll find is that even on the same wind storm, it’s notionally seen the same weather, and it’s been maintained notionally in the same way.
The spread of AEP losses across the site can be quite dramatic. You may have half a percent loss on some turbines over here, but you might have a 2% loss on some turbines over here. So again, by using the real data of each turbine combined with this neurodynamic model, it allows you to get some incredibly valuable insight as to how much money you could be losing due to blade damages.
Then it can guide you to say, well, which turbine should I do something about first? Because if you just looked at the structural categorization, say you needed to pick half your fleet to repair in one year, you may say, oh, well all of them have cat three damages, so I’ll just go and I’ll repair some of them because you know, I can’t tell the difference.
The reality is though, that. That’s not a full picture of the aerodynamic behavior. It matters where that damage is, what type of damage is, and these are all the things that we’re modeling. So you can say, right, I’m gonna go and fix these five turbines first, which they were killing me in terms of performance and spend your money wisely because you can’t fix everything all the time.
We have to prioritize, and it’s a tool for prioritization.
Allen Hall: So when I hook, hook up your a p i and I, I go look for these losses that I’ve. Most operators have right now this is a pretty quick response time, right? You’re not sitting around waiting for months for the, for PowerCurve to come back. It’s like put the data in outcomes, the information, you’ll have it in a day or so
Nicholas Gaudern: under an hour, typically.
Wow. It’s, it’s really fast. So we, we built this in a very flexible and scalable manner. So, you know, we’re running it in, in the cloud on servers that can scale. So if necessary, you know, we can take hundreds of calls into the system simultaneously and deal with them really, really quickly. There’s no, there’s no handholding.
We’re not there checking, clicking buttons. This has been set up as a robust, scalable system, so the results come through incredibly quickly. So, yeah, it’s, it’s, it’s very, very robust. Got a pretty high headroom and hopefully as we grow and people start to use the system, We can scale with it. That’s the beauty of these kind of modern cloud setups.
Allen Hall: So if I’m doing a drone inspection at the end of the summer period, which, where we’re at, right? So a lot of companies like to take them at the end of the summer or at the beginning of the springtime just to get a status on where there are with their, with their, their damage. With, I get that, all that drone information, I get the metadata, I send it to PowerCurve and boom, within an hour I know what I need to do.
Or leading edge repairs. I, I can start scheduling people for next year now and tell them what they’re gonna be
Nicholas Gaudern: working on. So instead of having to, to guess right now, is it worth putting on LEP or is it worth re cleaning the blade or is it worth putting on bgs? You know, there’s all these options available to a customer and we’re hoping to just provide some insight into what is most valuable because.
You know, LEP is not some silver bullet. LEP itself changes the leading edge shape and therefore aerodynamic behavior of as section. And depending on the LEP that may be next to nothing, but it might not be. So if your erosion is, is really mild, you could use our tool. You might see, oh, I’m only losing 0.3% AEP from erosion.
Well, it’s probably not worth putting on LEP from an aerodynamic perspective because the LEP is probably in the same order of magnitude for losses, whereas m losing 3%, well, I should probably do something, but you know, you can then have a conversation with us about vortex generators, about LEP, about blade repair and will help to guide you as to what you should, what you should do.
So I think that’s the thing here. Once you have. A piece of data that is come from an engineering driven approach, you can start to make much more confident decisions about how you go about maintaining your fleet and carrying out more effective asset management. Does it
Allen Hall: also provide some information, like you could get another half a percentage point in AEP by putting on VGs?
Does it provide that sort of additional bonus information besides just the leading edge erosion issue?
Nicholas Gaudern: So as of today, you know, you’ll get the loss, you’ll get the loss broken down by blade as well. You’ll get a heat map that shows you which of the damages are contributing most to that loss. And yeah, as we get more customers, using the system will enhance the functionality, but driven by what the customers want.
I think, you know, we want to get it out into the market and used, we don’t want to. Assume what an operator wants to know. We want to have dialogues, get people using it, get the data flowing, and then we’ll start adding the functionality that that is most useful to the operator.
Allen Hall: So there, Avista a p I has been out for a little while now and you’ve, you’ve had some initial customers use it.
What’s the feedback been?
Nicholas Gaudern: Very positive. I mean, I think this system as, as a whole started life a couple of years ago. When we really started developing the engineering and the process behind how you do this. And then last year at keen Power in the U S A, we launched a, a version of the system.
It wasn’t called Avis at the time, but it was with sky Specs. So Skys specs had been a great partner in this service from a very early stage. So we launched this service with Sky Specs. So that was, yeah, over a year ago now. We’ve run a lot of turbines through the system and what we’ve seen is that it’s given operators the ability to really start looking at that prioritization of prepare and help to drive internal conversation about loss, because without a number, very difficult to talk about budgeting our air VGs or LEP whatever.
So the feedback has been strong in, yes, it is a valuable tool to prioritize. But we’ve also heard some interesting things that we maybe didn’t latch onto straight away, and that was, well, if you have historical data, you’ve got three years of inspection data, you’ll start to see a trend. You can then start to project that forward to plan future O&M campaigns on the expectation of erosion.
But then you could even take it one step further and say, well, I’m planning a brand new windstorm of this turbine type. What did I learn over the last few years in my other wind farms? And then you start to look at, you know, very important financial decisions about, you know, write downs and assets and depreciation, all these kind of things.
So once you’ve got that data set, there’s a whole world of decision making that it can open up. And I feel at the moment we’re scratching the surface. Yeah. And again, the more operators we engage with, I think more we’ll, we’ll learn that. Yeah, so the
Allen Hall: data analytics becomes really interesting there because it’s an unexplored area of leading edge erosion.
The progression on a sort of a national, even global scale with this a Vista a p i tool. Yeah, you can, you could then, theoretically you start projecting when a farm does certain turbines, maybe they don’t do some turbines, right? Because it may not be worth it. But it may give a, be the industry a better sense of.
What is the proper timing and regimen to do leading edge erosion repair?
Nicholas Gaudern: Yeah. And we’ll start seeing patterns. We’ll start seeing, you know are there particular turbine models that seem to suffer more than others, for example, you know, and that, and that would be very valuable. So for people making investment decisions.
So I think we’ve, we’ve got a really nice grounding in the tool. As I say, we’ve, we’ve done a lot of work with Skyspace. There’s a lot of calls gone through the system and that’s helped us to really tune it over the last year. Now it is open and open to the world where we can take a p I calls from, from anyone as long as they’ve had the right setup process with us.
So we’re in dialogue with a number of major drone operators around the world. I think to me, in the near term, it’s more likely we’ll be taking calls from the inspection providers because they’re typically the ones in control of the database and the tagging. And all that kind of stuff. Of course, if an operator themselves has a, has all that data to hand and they wanna go direct to us.
Sure. I just feel that, from what I see in the industry, it’s generally the, the inspection providers that are, that are best placed to call us, but it, it’s pretty operator specific I think. As wind
Allen Hall: farms start to hibernate in the Northern Hemisphere, now’s the time to get that drone data, start taking a look at it and figure out what the plan is for next season there.
I know we have seen a number of operators already planning for next year trying to get resources lined up because there’s only so many resources out there. ’cause you need to. Figure out who’s going to do some of these repairs and get them booked. And now’s the time that, that that happens. Using the ar vista a p I would be a quick way to, to help organize that for an operator.
Nicholas Gaudern: For sure. The data’s there. The data’s there. You’ve just got to use it. Where we are adding another stream to the decision making process and. It’s simply not good enough. You know, in these days of challenged energy prices and contracts and, you know, wind farm deployment, you’ve got to squeeze everything you can out of your assets, and that means using the data you have available.
Don’t throw away your two 3% ap. Do something about it, but it’s very hard to do something about it if it’s not visible to you, and that’s what I hope this tool is, is going to bring this. Consistent, reliable methodology to, to give that insight. The Arrow
Allen Hall: Vista, a p i is a result of years of collaborations with local universities.
You that help to basically define how this process works. You wanna describe what happened behind the scenes there? Yeah, so
Nicholas Gaudern: I think, you know, PowerCurve, we’re, we’re a small company, but we, we leverage a lot of things we’ve done over the years and we’re, we’re very proud to have been awarded, you know, EU funding, so, so r and d projects into TEX generators and leading edge erosion and, and other aerodynamic topics over the years.
And our current E U D P funder project is called lca, or Leading Edge Roughness Categorization. So this is a project that’s being led by D T U, the Danish Tank University, and then there’s a number of o e M partners. So you’ve got investors. Siemens, gaa, Len, Sue Long and then PowerCurve, part of the, the consortium as well.
So we are really proud to be part of that project with, with all of these big players in the industry and led by such a, I would say, a world renowned organization of D T U when it comes to wind energy research. So the whole idea of this project is to come up with a universal. Categorization system for leading roughness that is considering aerodynamic and air, air acoustic penalties.
So I think there’s this problem at the moment, like there is with structural tagging of images that individual companies individual operators, they have their own tagging criteria. What makes it to level three or a level four? And I think that can only go on so long before. The industry has to kind of come to a decision about what is the standard for tagging.
’cause then that allows much more effective conversations to happen between OEMs, operators and service providers, third parties such as ourselves. So over the next couple of years, Leah Cat is gonna be working to do a lot of wind tunnel testing and engineering analysis such as C F D, computational fluid dynamics.
To help move us towards this universal acceptance of a classification system for damages. So leading edge erosion. So I think this is a great project. I think it’s a lot overdue and with the partners involved and sort of the financial muscle of the OEMs, I think we can get a long way towards towards the goal and having that classification.
So, Keep watching this space. I’m sure you’re gonna see LinkedIn Post and other media from the OEMs, from D T U, from Power Codes explaining what we’re doing and, and why and how we’re contributing to, to the knowledge base around Leady and
Allen Hall: George. Right. And, and how do the operators connect up with the AeroVista?
A p i like would they just reach out to their local drone company? Do they contact PowerCurve specifically? Do they hook up with you on LinkedIn? What’s, what’s the process here?
Nicholas Gaudern: So I would, I would say if if your drone inspector hasn’t already spoken to you about it, then just contact Power Cove directly and we’ll, we’ll make it happen.
So if you want to go through your drone inspection provider, we’ll then talk to that provider. We’ll have an integration process with them to make sure they know how to talk to the a p i, you know, they send the, the right style format, et cetera, et cetera. The, the boring stuff. And then, yeah, they’ll get the data pipeline plumbed in and where.
We’re good to go. So it’s, it’s a simple process, you know, it’s, it’s designed to be light on, on input effort, but, but heavy on output value and just turn it around really quickly. As we were talking around earlier.
Allen Hall: So Nicholas, I really appreciate you being back on the podcast. We love having, you know, there’s just so many cool things happening in Blade Aerodynamics inefficiency, so it’s, it’s good to get the download.
