ComEd uses drone-imagery AI to rank pole health and cut storm outage risk

Title

ComEd uses drone-imagery AI to rank pole health and cut storm outage risk

Content

ComEd deployed a drone-imagery-based computer vision model that scores the structural health of wooden utility poles, reaching 86% accuracy versus human expert agreement and giving the utility an actionable replacement queue across its network of more than 1.3 million power distribution poles in northern Illinois. “More utility companies are using image detection models to identify damaged equipment,” said Joey Martinez, principal data scientist for ComEd’s grid analytics team. “Our pole top AI health ranking model goes a step further by giving us a more actionable list of poles to prioritize for restoration, which improves budgeting and scheduling efficiency.” On the reliability angle he added: “Identifying these poles is huge for reliability. By being proactive, we can replace poles that pose a risk of falling or breaking during a storm and prevent outages for our customers.” Traditional pole inspection is a ground-based, manual measurement: field crews measure the circumference of the pole’s base and run engineering calculations to estimate remaining strength. The critical indicator — the amount of decay at the very top of the pole — is only visible from above and was effectively unmeasurable at scale. ComEd’s grid analytics team conceived of a new system to close that gap. Field crews began capturing drone imagery of pole tops in 2024, and the grid analytics team completed the computer vision models in early 2025. The model was trained on thousands of expert image comparisons — engineers labeling “which pole is worse?” — so the AI effectively thinks like the best pole inspectors. Rather than only flagging damaged or decayed tops, it assigns a “health score” that ranks poles, letting ComEd prioritize which to replace or reinforce. For aging poles with less damage, a C-Truss reinforcement can be installed to extend lifespan; six retired poles identified by the model were sent to the ComEd forensics lab to be sectioned and analyzed so future reinforcement decisions are grounded in measured decay patterns rather than visual estimates alone. “We believe that we are the only utility in the industry currently utilizing this new approach, and we are excited to share it,” said Tom Mahar, principal data scientist for ComEd’s grid analytics team. Jim Ortega, director of grid analytics for ComEd, framed the broader goal: “At ComEd, we’re always looking for better ways to run our operations—more efficiently, more effectively, and with our customers in mind. That means bringing new technologies into how we manage and maintain the grid, including automation and predictive analytics. This pole-top AI health project shows what’s possible through collaboration across our employees and teams, using smart tools to spot issues earlier and keep power flowing reliably for the people who count on us every day.” The project will be further enhanced as more pole-top imagery is collected and combined with forensics insights, strengthening the AI models and improving the prioritization of pole replacements — and reducing potential outages for ComEd customers.