More frequent and extreme weather events are beginning to overwhelm utilities’ incident management teams with widespread outages and vast amounts of data such as weather reports, customer calls, and damage assessments. In fact, there is so much data of late that it’s increasingly difficult to analyze the situation and devise an action plan. During an outage, every minute counts. AI can help utilities improve planning and staging by improving data quality, and shave crucial minutes off their restoration times by quickly identifying hazards and locating damaged equipment. This will allow incident management teams to prioritize the replacement work and focus on restoring customers.
Over the past week, more than 12,000 lightning strikes have hit northern California, sparking hundreds of wildfires and burning over 1.1 million acres of land. Two of the largest fires - the SCU and LNU Lightning Complexes - are the 2nd and 3rd largest in California’s history, respectively. On the storm front, Tropical Storm Marco is the 13th named storm for the year. The average hurricane season, which runs from June to November, generates about 12 storms a year, but experts are expecting up to 19 this year.
Extreme weather events causing billion-dollar disasters are increasing. In 2019, the U.S. experienced a total of 14 billion-dollar disasters, stemming from 3 floods, 8 storms, 2 tropical cycles, and 1 wildfire event. In the most recent decade, from 2010-2019, the U.S. experienced 11.9 such events on average per year. In the prior 3 decades, the U.S. experienced an average of 5.9, 5.2, and 2.8 events per year.
As utilities face more extreme, more complex, and potentially more expensive weather events, they need to implement measures that will enable them to more quickly recover as the events happen. To recover from an outage and restore customers, they need to remove hazards, rebuild damaged or destroyed assets, and re-energize the grid. AI can be used to augment and speed up this manual process so that back-office storm teams, as well as field crews and mutual assistance crews, can focus on the important work of rebuilding the grid.
Below are some ways in which AI can help.
Before the event, utilities will often run models that predict the path of the storm or the wildfire. They will combine that prediction with internal asset databases to evaluate how the event could impact their assets so that they can plan their staging and response. Noteworthy’s AI models, for example, can identify equipment tags, the types of assets that are attached to the pole, as well as the asset’s location. Accurate information can help utilities more accurately predict which assets may be impacted, enabling them to better prepare and deploy their storm response.
After the extreme weather event has passed, utilities take ground surveys (by sending out trucks and crews) and aerial surveys (through helicopters, airplanes or drones) to assess the situation and make the area safe. Noteworthy’s AI models can augment the hazard identification process by quickly sorting through aerial and ground images to highlight hazards such as downed lines and broken tree limbs. And since the imagery is geotagged, storm teams can immediately improve their situational awareness by understanding where and what the hazards are; ultimately enabling them to prioritize restoration efforts and safely direct resources.
During restoration efforts, a significant amount of data is generated such as imagery, video, weather information, hazard locations and resource deployment. Back-office teams sort through these data to develop action plans. For small events, this manual process works fine. However, as the events get larger and more complex, these back-office teams can get overwhelmed, and restoration ends up taking longer. By leveraging AI to help sort through the images and video of damaged assets, and superimposing that information on a map, restoration teams can focus less on sorting the data – and more on the action plan to get the grid re-energized.
More and more extreme weather events are happening around the world. After each event, utilities and communities begin the process of restoration. During an outage, every minute counts toward the utility’s SAIDI (System Average Interruption Duration Index), a common reliability metric that measures the duration of the outage per customer. By leveraging A.I., utilities can speed up their analysis by quickly filtering for the most critical data – hazards in the road, fallen trees, damaged equipment – and utilize this intelligence to shorten their restoration times, improve their SAIDI, maintain their customer satisfaction, and satisfy their regulators.