Every year, thousands of US wildfires start with a spark from a downed power line. NREL thinks AI can catch those sparks before they spread. (NREL)
With funding from the US Army Construction Engineering Research Laboratory, researchers trained neural networks on simulated fault data—everything from wet grass to dry gravel—to spot the tiny shorts that often slip past today’s detection systems.
Eaton supplied real-world testing, while NREL’s supercomputers scaled it into a tool that utilities can use to flag risks and respond faster. The lab is now working with providers nationwide to bring the tech onto the grid.