73. Improving Asset Inspections with Drones and AI

Improving Asset Inspections with Drones and AI The U.S. bulk electric system consists of more than 360,000 miles of transmission lines, including approximately 180,000 miles of high-voltage lines, according to the U.S. Department of Energy. That means power companies have a lot on their plates when it comes to inspecting and maintaining those assets. Furthermore, the stakes are high. As wildfires in California and elsewhere have shown, billions of dollars can quickly go up in smoke when equipment fails. San Francisco-based Pacific Gas and Electric (PG&E), for example, faced as much as $30 billion in potential liabilities for the role its equipment played in causing wildfires in 2017 and 2018. PG&E filed for Chapter 11 bankruptcy in January 2019, at least partly as a result. Therefore, finding and fixing problems before components fail is worth the effort. “For the most part, utilities are inspecting somewhere between 10% and 12% of their assets each year,” Chris Beaufait, CEO of Sharper Shape, said as a guest on The POWER Podcast. “They’re always trying to figure out how to prioritize in the best way, but then they also run into human resource issues, and just how much they can accomplish in a single year.” That’s where automated drone technology and artificial intelligence (AI) algorithms can help. “We can do detailed inspection of various assets using drones, and we actually have specific software packages that allow parts of that to be automated so the utility can get a repeatable process on their inspection,” Beaufait said. “Then, we basically process all that data. If we do a LIDAR [light detection and ranging] data collection, as an example, we can create data assets of everything that they have in the physical world, and then we can use AI algorithms and human intelligence to really come up with actions that they can do around that so they prioritize what they're going to do in the field, around maintenance, around vegetation management, around asset replacement, or asset improvements,” he explained. Utilizing drones and AI technology allows a much larger area to be covered in a shorter amount of time. “We're talking tens of thousands of miles, and using this combination of artificial intelligence, we can tell them where they may have risks,” said Beaufait. The inspections generate an enormous amount of data. According to Beaufait, Sharper Shape has completed more than 35,000 miles of collection, which includes more than 800,000 different structures and assets, equating to about 5 petabytes of data (5,000,000 GB). It would be impossible for a human to thoroughly process all the information in that amount of data, but computers with AI algorithms can look for anomalies and prioritize the problems identified. And while the inspection capabilities available today are remarkable, the technology is constantly improving, which means future asset assessments could be even more advanced.