Latest / The Operations Podcast with Fexingo: Process, People, and Profit in Modern Business / How a Wind Turbine Maker Cut Blade Inspections by 80 Percent
Transcript
- Lucas: So you know how wind turbines are these massive structures, blades as long as a football field, and inspecting them for cracks and damage is one of the most dangerous jobs in energy? Luna: I've seen videos of people rappelling down blades. It looks terrifying. Lucas: It is. And it's slow. A single turbine inspection with a rope-access team can take three days. A wind farm with fifty turbines? That's half a season of downtime. Luna: That's huge. So how do you cut that down? Lucas: Well, one manufacturer we're looking at today—let's call them Nordic Wind—went from three days per turbine to just four hours. An 80 percent reduction. And they did it by swapping humans on ropes for a swarm of drones and a machine-vision system trained to spot micro-cracks. Luna: Four hours per turbine? That's a game-changer. What's the catch? Lucas: The catch is, you have to trust the algorithm. And that's where the real operational challenge starts. Luna: So walk me through the process. How does a drone inspection actually work? Lucas: So instead of one drone, they send up a small fleet—maybe five or six—each with high-resolution cameras and LiDAR. They fly a pre-programmed pattern around each blade, capturing thousands of overlapping images. The whole flight takes about thirty minutes per blade. Luna: And then what? You've got thousands of images—that's a lot of data to review. Lucas: Right. That's where the AI comes in. They built a deep-learning model that's been trained on millions of labeled blade images—including defects like leading-edge erosion, cracks, even lightning strikes. It runs through the entire dataset in a couple of hours and flags anything that looks abnormal. Luna: So the human isn't looking at the images at all? Lucas: Not for the first pass. The algorithm produces a risk score for every blade, and the maintenance crew only physically inspects the ones that score above a certain threshold. That's where the time savings really come from—you're not sending people up a tower to eyeball a clean blade. Luna: That makes sense. But I imagine the veteran inspectors on the ground weren't exactly thrilled about being replaced by a computer. Lucas: That's the human side of operations, and it's often the hardest part. They had a pilot program where they ran drone inspections alongside traditional ones for a few months. The new system flagged every major defect the old method found—plus a few that were missed. Luna: So the data won them over? Lucas: For the most part. The company also retrained those rope-access techs to be drone operators and data reviewers. They didn't lose their jobs—they upgraded them. That was crucial for getting buy-in. Luna: So it's not about replacing people, it's about changing what they do. That's a smart operations move. Lucas: Exactly. And the financial case is compelling. The drone setup cost them about $200,000—drones, software, training. They recovered that in less than six months just on reduced downtime and fewer safety incidents. Luna: I'm guessing insurance premiums went down too? Lucas: They did. Fewer people working at height means lower workers' comp claims. Their insurance broker actually gave them a premium cut after the first year. Luna: That's a win-win. But what about the limits? Drones can't fly in high winds, right? Lucas: True. They only fly when wind speeds are under 25 miles per hour. But they schedule those windows smartly—weather forecasting is part of their ops software now. They can plan a week of inspection work around the forecast. Luna: So you're not just improving the inspection, you're improving the whole planning process. Lucas: Exactly. And that's the real operational lesson here—it's not about any single piece of tech. It's about redesigning the entire workflow around the new capability. Luna: Speaking of workflows, have you ever thought about how we keep making these shows? I mean, every week we dive into a new case study, and it takes a lot of research and editing. Lucas: It does. And honestly, the only reason we can keep it ad-free and accessible is because of listeners who chip in a few dollars a month. It genuinely adds up and covers the costs of hosting, research tools, and the occasional coffee. Luna: Yeah, if you've gotten value from these episodes, consider a small donation at buy me a coffee dot com slash fexingo. Even a couple of dollars makes a real difference. Lucas: We really appreciate it. Now, back to Nordic Wind—they're now expanding this approach to their entire fleet of 1,200 turbines across Northern Europe. And the data they're collecting is helping them predict failures before they happen. Luna: Predictive maintenance. So they're not just fixing what's broken, they're preventing it. How does that work? Lucas: The AI doesn't just flag visible cracks. It can identify subtle changes in vibration patterns or blade pitch that indicate stress. Those are early warning signs that a blade might crack months down the line. Luna: So they can replace a blade during a scheduled maintenance window instead of scrambling for an emergency fix. Lucas: Exactly. Emergency replacements are four times more expensive. So by catching issues early, they're saving millions per year across their fleet. Luna: That's impressive. But is this applicable beyond wind energy? I'm thinking oil rigs, bridges, cell towers—anything that needs regular inspection. Lucas: Absolutely. The same drone plus ai model is already being tested on offshore oil platforms and even on highway bridges. Any asset that's high, hard to reach, and critical to keep running. Luna: So this is really about changing the economics of asset maintenance across industries. Lucas: That's the bigger takeaway. Operations used to be about doing the same thing over and over, only a little faster. Now it's about completely rethinking the process with data at the center. Luna: And the people part—training, trust, buy-in—that's still the hardest piece to get right. Lucas: It is. But when you get it right, the results are hard to argue with. An 80 percent reduction in inspection time is not just a number—it's the difference between a wind farm that's profitable and one that's just scraping by. Luna: So what's next for Nordic Wind? Are they planning to apply this to other parts of the turbine? Lucas: They're already moving into gearbox and tower inspections, and they're exploring using the drone footage to create 3D models of entire wind farms for digital twins. Luna: Digital twins—that's where you can simulate different operating conditions and see how the asset would respond. That's the future. Lucas: It's definitely on the horizon. And the fact that they're collecting high-resolution data on every inspection makes it a lot easier to build those models. Luna: So this one operational change—drones for inspection—is snowballing into a complete digital transformation of their asset management. Lucas: It is. And that's why I find these stories so interesting. It's never just about the tool. It's about the ripple effects—on people, on processes, and on the bottom line.