California utilities are turning to drones and AI to find equipment problems that could spark wildfires. Using drone cameras, utilities can take photos and videos of equipment, such as distribution poles and transmission towers, and spot problems like corrosion using computer vision and machine learning.
- The approach can't manage the current wildfires in California, but can monitor and spot problems before they lead to fires.
- California's largest utility, PG&E, has already used machine learning to estimate how its equipment could fare in high winds. This allows it to take steps like cut off power to certain areas.
- PG&E, as well as San Diego Gas & Electric and Southern California Edison, also use computer vision to enhance inspections of distribution and transmission infrastructure in areas most prone to fires. This cuts down on the time and opens the door for inspectors to focus more on other ignition risks.
- Wildfires have destroyed almost 3.1 million acres in California, 900,000 acres in Oregon, and 600,000 acres in Washington state. Experts say that climate change is responsible for heatwaves and severe dryness that are making wildfires more likely and destructive in western states.