What Rosie Learned in Europe’s Wind Turbine Test Labs

Rosie returns from Denmark and shares her experiences seeing DTU blade testing, a lightning lab, and a 15 MW Vestas test turbine. The Uptime Wind Energy Podcast is brought to you by Weather Guard Lightning Tech, creators of the StrikeTape Ultra LPS retrofit. Subscribe to Uptime’s Substack newsletter. And check out Rosemary’s "Engineering with Rosie" Youtube channel. Have a question we can answer on the show? Email us! Episode Transcript What Rosie Learned in Denmark's Blade and Lightning LabsIntro and Risø blade testingAnnouncer: The Uptime Wind Energy Podcast, brought to you by StrikeTape, protecting thousands of wind turbines from lightning damage worldwide. Visit StrikeTape.com. And now your hosts.Allen Hall: Welcome to the Uptime Wind Energy Podcast. I'm Allen Hall, and I'm here with Rosemary Barnes and Yolanda Padron. Rosemary just came back from her long trip to Germany, and a good part of it was also spent up in Denmark. Rosemary, you visited a number of wind-related companies on your travels, so give us a little update.Rosemary Barnes: I started in Copenhagen and then went over to Roskilde, mostly to see the rain erosion test facilities. But I also got a tour of the rest of the testing, mostly blade testing. There's a really interesting history to that facility. It started out as nuclear power research and had a reactor there. I didn't even realize that.Yolanda Padron: Yeah.Rosemary Barnes: And incidentally, the reason it turned into a wind facility is that they already had all of these amazing wind measurements for the site. That's super important if you've got nuclear power. If there's a meltdown, you need to be able to track where the fallout is going to go. So they knew the wind speeds in the region really well. I was told by my friendly guide that that's the reason it was seen as a really good place to get started on researching wind energy in Denmark. They could put a turbine up knowing what was going on with the wind resource there.Allen Hall: Niels Bohr worked there, right? Is that part of the history?Rosemary Barnes: He worked there. I think it was his facility. I think he set it up. I didn't plan to talk about this, and I haven't actually verified any of it. I'm just remembering what I was told weeks ago. My understanding is that he set it up to do nuclear power research. They don't do that anymore, but they do a bunch of energy work. Mostly it's wind energy, and then a bit with the grid. They're doing some interesting trials with EV charging and how that can integrate into a grid powered by variable renewables. They've got their own little microgrid going on there, and if you participate in the trial, I think you can charge your car for free. All of the businesses I visited in Denmark had rows and rows of EV chargers, though mostly not free.Allen Hall: So you went to the research labs at Risø, the structural labs there. Did you walk through where they were doing blade testing and some of the structural testing?Rosemary Barnes: Yeah. There's actually an American guy there, Derek. He was working at GE in America before that and had worked for TPI for a while, so he brought that expertise to Risø. They've got a bunch of test rigs in there for small blades. I can't remember the exact length, but maybe 20 meters.Allen Hall: Twenty. Yeah.Rosemary Barnes: Which was once a full-scale wind turbine blade, but now it's small scale. It's mostly for research, trying out various new things. They produce blades for other research facilities as well and test them, and they're trying out some new manufacturing techniques.They're also trying new testing techniques. Normally when you test a wind turbine blade, you start with a static test. You basically get some ropes and pull on the blade hard enough to represent the largest gust that blade should ever experience in its lifetime. After that, you do a fatigue test, where you flap it up and down and side to side, one after the other. That takes months. For really long blades it can take a year, because you can only flap as fast as the natural frequency of the blade. A longer blade has a slower natural frequency, so it takes longer to get through the hundreds of millions of bending cycles it sees in its lifetime.They've got new test methods where they're trying to do both directions at once. Flapwise is the direction where a gust of wind blows the blade toward the tower. Edgewise is the direction where gravity bends it back and forth as it rotates. Normally you test those one after the other. You put it in one direction and flap it for millions of cycles, then you swap it and flap it the other way for millions of cycles. But now they've got actuators that set it off so it traces out an ellipse with every flap. It's doing flapwise and edgewise at the same time, so you can get through the test faster.People also want combined loading, because in the field you've got flapwise and edgewise loads in some combination at the same time. I get the point of that. But just because it's two loads at the same time doesn't mean it represents how those two loads actually co-occur in reality. So I'm not entirely sold on it being a more realistic test. But it is good if they can do those two fatigue tests at the same time and save some months on the testing.Another thing is that they're looking at better ways to test torsional rigidity, because that's more and more important as blades get longer. I believe torsional stiffness varies with the fourth power of the blade length. When blades were very small, it wasn't remotely a problem. It was kind of like centrifugal loads in a blade. You know they're there, but they're not a big deal. If you design a blade that can withstand all the other loads, it's definitely going to be strong enough for centrifugal loads. That used to be true for torsional stiffness as well, but it's less and less true as blades get longer.It's very hard to test and hard to verify. It's also hard to model, because the torsional rigidity of a blade has a lot to do with how the adhesive and the glue joints deflect under real loads. That's probably one of the weakest parts of a finite element model. It's quite hard to accurately model those thick adhesive joints, because they're less predictable than composite structures.The challenging thing is that every time you test something in the lab, some parts of the design will be under-tested and some parts will be over-tested. In the static test, for example, you can do it as simply as putting a single clamp on the tip and pulling on that. But a real aerodynamic load is applied over the whole surface. You can apply several clamps and get a more realistic load, but every clamp stiffens that area. You're never going to get a failure exactly where the clamp is, because the clamp prevents it from buckling. So you're distorting things. If you get more accurate in one sense, you're probably getting further from accuracy in another.That was one of the threads through all of the testing I watched in Denmark: how much can you accelerate the test without getting away from reality?LEP rain erosion testing and field trialsRosemary Barnes: I do think that's something the industry struggles with. Later on, at Wind Energy Hamburg, I visited a lot of leading edge protection companies, and a lot of them had little samples showing how amazing their material was in the rain erosion tester compared to their competitor's. And it's like, yeah, but what did you test? It doesn't match reality. I hate those little comparison tests, because everybody knows there are all sorts of little things you can do to make a product fail faster. If you just apply your competitor's product badly and then test it, you can easily crush them. But what's the point in showing that? When I see companies dissing their competitors, I find it an indication that they don't have confidence in their own product. So, to go off on a tangent, that's my hot tip.Allen Hall: No, that's a good tangent.Rosemary Barnes: It's not the message you think it's sending when you show your competitor's product.Allen Hall: Well, the V-N curve, which is what you're trying to achieve with rain erosion testing, does give you some predictive capability for what a coating would do at any particular wind farm. I know you're going to throw a bunch of variables at me, but as a baseline, you have this V-N curve, which is roughly the velocity versus the number of water droplet impacts. What I haven't seen is people publishing their product's V-N curve to the engineering world. That doesn't happen, which is weird, because they spend so much money to get it. Why would you not communicate it to someone like Rosemary, who can then use it to see whether that product will work on a wind turbine somewhere in Australia? I'm really struggling with that, because people spend a lot of time in the rain erosion world trying to get those numbers.Rosemary Barnes: I feel like I've already had a rant about this on an episode.Allen Hall: I don't think so.Rosemary Barnes: Maybe it was how I got into some robust discussions, which some might have interpreted as arguments, with a company complaining that their product is fully validated, that it doesn't need any more trials, and that everyone should just purchase it for their entire wind farm all at once. And I was saying that you haven't validated it for their wind farm, so they don't know.Allen Hall: Oh, yeah.Rosemary Barnes: But okay, that cuts both ways. If you had two side-by-side wind farms and one of them had trialed it successfully, then the neighboring wind farm would likely feel like, okay, that's great, they had good success, we'll put it in. But I was getting examples like, "We tested this in Texas and it's perfect." Okay, good for Texas-adjacent wind farms. They can rejoice that this product works well there after one year, or whatever it was.But with all the testing you do in the lab, and even the testing you do on site,...