Expert Vision: How a GE Vernova Engineer Uses AI to See the Invisible

As the global energy transition accelerates, maintaining the long-term stamina of critical infrastructure like wind turbines requires being able to spot hidden risks before they can become failures. Proactively protecting these systems means uninterrupted energy for neighborhoods, smarter and safer environments for inspectors, and lower operational costs. Yet inspecting a wind-turbine blade 100 meters in length - longer than the wingspan of a commercial airplane - to find tiny flaws or stress points that could weaken it is a daunting task. Where would you begin?

Until recently, this job was done manually, by technicians relying on intense attention to detail. One person would walk through the interior of the blade with a high-powered flashlight, recording detailed images. Others would painstakingly scroll through the resulting data, checking for any deviation from exacting requirements. The concentration needed was so demanding that a technician would have to take a break every 25 minutes to avoid losing focus. From start to finish, an inspection could take up to 15 hours.

There was a clear opportunity to improve the process, making it more efficient while heightening accuracy. Sivaramanivas Ramaswamy specializes in such challenges. For more than two decades, Ramaswamy, a senior engineer with GE Vernova's Advanced Research Center in Bengaluru, India, has been developing techniques to see hidden problems before they become disruptions. He is an expert in nondestructive evaluation - technologies that detect hidden damage without cutting equipment open. Using ultrasound, terahertz waves, and other tools that send signals into materials and read the echoes, his work gives engineers entirely new ways to see inside complex machinery without taking it apart.

"I say to my friends that I work on failures," says Ramaswamy, whose team received a tech award in 2024 from GE Vernova for the new inspection process, known as the Digital Blade Certificate. "When you image a crack with all its details, you don't get the same joy as when you see the face of an unborn baby on ultrasonic images. But it does still give me a high when I'm able to see something that's not possible with your naked eyes."

Working with the global team in Bengaluru and Niskayuna, New York, Ramaswamy helped develop this digital inspection protocol for wind turbine blades. Today, small machines known as crawlers capture images in areas inside the blades that are inaccessible to human inspectors. An AI-powered model scans the images for anomalies, flagging features such as cracks, gaps, or weak bonds. Human technicians then review the most critical areas, balancing automation with human judgment. Ramaswamy now is improving the process by experimenting with an AI model trained on ultrasound images.

Adapting Cross-Industry Technologies for Advanced Inspection

As a university student, Ramaswamy developed expertise in nondestructive evaluation by learning fields such as electronics, computer science, and mechanical engineering. After working in aerospace for the Indian government, the opportunity to tackle broad research questions drew him to GE Vernova.

In addition to the wind industry, early in Ramaswamy's career, there was a "continuous back-and-forth between different teams" in the gas power sector as well as aerospace and healthcare, Ramaswamy says. "At first, it looks like these domains don't have anything in common, but at the heart of the technology there is a commonality. The multidisciplinary element is core to nondestructive evaluation, and working at GE Vernova keeps me motivated."

Ramaswamy frequently adapts technologies from one industry to solve problems in another. For example, he has helped modify laser-based terahertz inspection tools, already used in automotive and aerospace, for the extreme conditions of gas turbines. These systems use femtosecond laser-generated signals to probe a material's subsurface, measuring microscopic coating thicknesses and detecting hidden defects.

Ramaswamy uses tools like this robotic arm to drive improvements in nondestructive inspections that apply across multiple fields.

His team also is exploring millimeter-wave technologies - high-frequency radar systems similar to the sensors used in autonomous vehicles, that can effectively see through solid structures. Early field trials in wind energy suggest these tools could expand how engineers inspect critical infrastructure, allowing them to spot failures before they occur.

Stopping Failures Before They Start: AI and Digital Twins

The next challenge for Ramaswamy is to not just uncover problems that have already surfaced, but to predict potential failure points before they happen. He is working on building digital twins - virtual models that combine sensor data with AI to show how parts wear down over time and forecast where failures are most likely to occur. He and his team have deployed this approach in Southeast Asia at two power plants that contain up to 350 miles of tubes, which are time-consuming to inspect and are subject to significant heat and pressure. If a tube ruptures anywhere, the resulting outage can take weeks to repair. Predicting where this might happen makes preventive maintenance possible.

While his interests are wide-ranging, Ramaswamy's innovations help build equipment that lasts longer and is easier to maintain, ultimately bringing down the cost of power generation and helping to keep the lights on for households around the world. While most people will never see his work behind the scenes, his work helps pave the way for fewer unexpected outages, less downtime, safer inspections, and more reliable power for homes, hospitals, businesses, and grids.

"I still see myself as a student," he says. "Even now, if I get an opportunity to meet a new customer or travel to a new site, I keep doing it as part of my learning exercise. There can be a million ideas on the table, but there are only a few that really help the customer. Working at GE Vernova helps me to identify those few that are critical."