Hawaiian Electric on July 16 announced that it has begun deploying a network of high-resolution video cameras using artificial intelligence technology to provide enhanced situational awareness and early detection of ignitions in elevated fire risk areas near company infrastructure.
The investor-owned utility’s announcement marks the latest example of utilities tapping AI to track and rapidly respond to wildfires.
Hawaiian Electric recently installed the first camera station in Lahaina and has plans to deploy a total of 78 stations in elevated fire risk areas on the five islands served by the company, with each location having two cameras to provide a full 360-degree view.
The camera feeds will be monitored 24 hours a day, seven days a week.
The $14 million project is the latest step in Hawaiian Electric’s ongoing effort to reduce the risk of wildfires associated with company equipment, it said.
Hawaiian Electric signed a five-year contract with California-based ALERTWest, which will install and maintain the camera stations as well as provide around-the-clock monitoring for potential ignitions by experienced wildfire safety professionals.
Approximately 50% of the project costs will be covered by federal funds allocated under the federal Infrastructure Investment and Jobs Act estimated at $90 million in grant funding covering various costs related to Hawaiian Electric’s resiliency and wildfire mitigation work.
Hawaiian Electric also will be able to achieve cost savings by leveraging its existing telecom network to provide communications support for the project.
ALERTWest’s software platform, which is widely used in fire-prone areas through the Western U.S., is assisted by AI to detect smoke and other early indications of fire in real-time.
It was developed by ALERTWest in collaboration with UC SanDiego’s ALERTCalifornia team, California utilities, and the California Department of Forestry and Fire Protection (CAL FIRE).
ALERTWest also recently partnered with the University of Oregon and the University of Nevada, Reno, creating the largest publicly available, interoperable, AI wildfire detection system in the United States. The system is assisted by AI to detect smoke and other early indications of fire in real time, day and night. This allows first responders and emergency personnel to cross state boundaries in mutual aid situations and have access to the tool when needed. The AI Solution was recognized as one of TIME magazine’s best inventions of 2023.
The software platform’s pan, tilt, zoom cameras constantly scan their surroundings, completing one 360-degree sweep every two minutes. Utilizing AI, the system detects changes from previous images and highlights them with a red rectangular box on the screen. This, coupled with 24/7 human verification, helps eliminate false alerts caused by mist or dust. If a suspected ignition is detected, the ALERTWest Operations Center staff reviews the camera imagery to ensure there is sufficient visual evidence before notifying Hawaiian Electric and emergency response agencies.
The public will be able to access the live feeds from any of the cameras on the ALERTWest website at www.alertwest.org. Half of the video camera stations are expected to be operational by September 2024, with the remainder to be installed in the first half of 2025.
In early August 2023, wildfires swept across Maui’s Lahaina community. More than one hundred residents of the community died as a result of the wildfires.
In the wake of wildfires on Maui, Hawaii, President Biden on Aug. 30, 2023, announced that the Department of Energy would provide $95 million through the Bipartisan Infrastructure Law to harden Hawaii’s electric grid, improve service, limit damage during future events, and help prevent failures in the future that could lead to severe events.
Austin Energy Utilizing AI Technology as Tool in Wildfire Mitigation Efforts
In a recent episode of APPA’s Public Power Now podcast, Chris Vetromile, Wildfire Manager for Texas public power utility Austin Energy, details how Austin Energy is proactively working to mitigate the threat of wildfires including utilizing artificial intelligence technology.
He noted that the utility has partnered with Pano AI to deploy high definition cameras “that are put out in locations around Travis County and our service territory, the city of Austin, and they actually detect and look for smoke.”
The podcast episode and the transcript for the episode are available on APPA’s website.
Salt River Project
Arizona public power utility Salt River Project in May 2023 noted that in order to support early detection of wildfires near the power system, it had invested in a pilot to install solar-powered infrared cameras that use AI technology to detect smoke from up to 10 miles away.
The “SmokeD” cameras learn surrounding environment and provide alerts to SRP as well as the U.S. Forest Service and emergency response crews.
At the time, SRP said that it was setting up two test areas with 12 cameras to capture images of a portion of its 500-kV transmission towers. The cameras can work night and day, and capture images every 10 minutes with a 360-degree view.
“SRP is still running this program and is looking at incorporating additional weather functions at these sites to continue our work to mitigate wildfires,” Elaina Steingard, Senior Media Relations Representative at SRP, told Public Power Current.
Pano AI Also Works with Other Utilities
Along with Austin Energy, Pano AI also works with a number of other utilities including:
- Portland General Electric
- Puget Sound Electric
- Xcel Energy Colorado
- PacifiCorp
- Rocky Mountain Power
- Holy Cross Energy
- CORE Electric Cooperative
- Avista
Colorado Utilities Tap AI Technology to Detect Recent Fire
Xcel Energy, CORE Electric Cooperative and Pano AI on July 17 shared details about the use of AI-powered cameras in detecting and aiding the successful containment of the Bear Creek Fire in Douglas County, Colo.
On June 16, 2024, at approximately 9:30 a.m. CT, the U.S. Forest Service and the Pano camera system simultaneously detected wildfire smoke caused by a lightning strike in the Bennett Mountain area above Strontia Springs Reservoir in Douglas County. The Pano system provided triangulated coordinates and a video of the smoke, confirming it as wildfire.
Given the conditions of the day, a Red Flag Warning with high temperatures and low humidity, Emergency Services determined this was a high-risk fire with the potential to spread quickly. By 9:54, the department launched a helitack response with two 5TA helicopters when the fire was at 0.25 acres, burning on the remote north side of Bennett Mountain, near Bear Creek. An emergency services crew of nearly 40 hiked 2.5 hours from Waterton Canyon to reach the site, Xcel noted.
The helitack lasted 5.5 hours, dropping 56 tanks of water totaling 17,765 gallons. By 7:00 p.m., the fire was fully contained at three acres. This rapid initial attack prevented the fire from progressing to a large-scale, hard-to-contain fire.
The Pano AI cameras, funded by CORE and Xcel Energy, “are strategically located near utility infrastructure or on third-party assets, like telecommunications towers, and were instrumental in the early detection and response to this potentially devastating wildfire,” Xcel said.
The detection and location identification of the Bear Creek Fire resulted from triangulation between two Pano AI camera stations—one installed for CORE and one installed for Xcel Energy. Having the specific coordinates helped crews save valuable time when hiking for more than two hours to the remote location.
To date, CORE has invested in six advanced camera stations using technology by Pano AI. Xcel Energy currently has 28 cameras operational in its service area statewide and plans to activate 14 more by the end of this year. Together, CORE and Xcel Energy will have 48 stations live in Colorado in the coming months along with a dozen more from other local governments and enterprises for a total of 60.
In June 2024, Xcel Energy submitted its 2025-2027 Wildfire Mitigation Plan to the Colorado Public Utilities Commission.
Proposed investments and improvements in the 2025-2027 Wildfire Mitigation Plan include, among other things, adding hundreds of weather stations for detailed information near power lines and equipment and tripling the number of AI cameras for early smoke detection.
Oregon Utility Shares Feeds from Wildfire Cameras that Use AI Platform
Oregon investor-owned utility Portland General Electric notes on its website that its Wildfire Mitigation and Resiliency team shares the feeds from wildfire cameras with a number of fire and emergency services agencies including the U.S. Forest Service, Oregon Department of Forestry, Oregon State Fire Marshal, tribes, state and local dispatch centers, as well as county and city emergency managers.
The remote cameras, which cover PGE’s High Fire Risk Zones, use an artificial intelligence platform to detect wildfire smoke, increase accuracy in identifying the location of smoke, and provide real-time notifications, including text messages and emails, it noted.
“In the past, fire crews had to drive around looking for the source of the smoke. Now, in areas with camera installations, these cameras can share a Google Map pin that takes crews directly to the location,” PGE said.
State Agencies Also Tap AI
State agencies in wildfire-prone regions are also utilizing AI to detect and fight wildfires.
In July 2023, CAL FIRE and UC San Diego’s ALERTCalifornia program announced an ongoing partnership aimed at revolutionizing firefighting through AI.
The collaboration involved a trial of the AI system executed by several CAL FIRE units, including San Bernardino and Sonoma-Lake-Napa, and utilizing ALERTCalifornia’s then 1,032 high-definition, pan-tilt-zoom cameras.
These cameras, with 208 funded by CAL FIRE, offer 24-hour surveillance and can sweep 360 degrees every two minutes, providing high resolution imagery.
This new AI tool was rolled out to all units in the late summer of 2023 and now analyzes camera feeds across California for anomalies, alerting Emergency Command Centers and first responders to potential fires, sometimes even before 911 is notified. Since the rollout, the network has seen significant growth and innovation.
In March 2023, Washington State Commissioner of Public Lands Hilary Franz announced a pilot partnership using Pano AI technology to enable early detection and monitoring tied to wildfires. The Washington Department of Natural Resources and Pano AI signed a contract to begin the AI pilot project.
“We have 21 stations deployed with them,” Jocelyn Johnson, Pano AI’s Director of Communications, told Public Power Current in reference to the Washington Department of Natural Resources.
University of Southern California Details New Tool to Predict Wildfire Spread
Meanwhile, researchers at the University of Southern California have developed a new method to accurately predict wildfire spread, USC reported on July 22.
“By combining satellite imagery and artificial intelligence, their model offers a potential breakthrough in wildfire management and emergency response,” it said.
Detailed in an early study proof published in Artificial Intelligence for the Earth Systems, the USC model uses satellite data to track a wildfire’s progression in real time, then feeds this information into a sophisticated computer algorithm that can accurately forecast the fire’s likely path, intensity and growth rate, USC said.
Biden-Harris Administration Invests $250,000 to Develop AI Tool
The federal government is also supporting efforts to tap AI to detect and respond to wildfires.
In late May 2024, the Department of Commerce and NOAA announced $250,000 in funding to support the development of a new artificial intelligence model aimed at improving fire weather forecasts through better lightning prediction.
With the funding, the Center for Satellite Applications and Research, part of NOAA’s National Environmental Satellite, Data, and Information Service, and the University of Wisconsin Cooperative Institute for Meteorological Satellite Studies are creating a customized version of LightningCastoffsite link.
This AI model uses imagery and data from the Advanced Baseline Imager and the Geostationary Lightning Mapper -- both of which fly on NOAA’s geostationary satellites -- to predict lightning strikes in the next hour at any given location, including those around wildland fires.
This version of LightningCast includes a new interface, additional machine learning technology and fire-related data to increase its value for decision-making.
NOAA continues to develop products to help first responders stay safe during wildland fires by testing an improved LightningCast tool and related research on early detection of dangerous fire-generated thunderstorms.
In addition to the risk of lightning strikes, wildland firefighters also contend with new blazes ignited by lightning strikes. Incident meteorologists at NOAA’s National Weather Service use LightningCast to provide real-time information to wildland fire incident teams, which improves the safety and effectiveness of their wildland fire response.
The wildland fire-focused version of LightningCast began testing in August 2023, within the newly formed NOAA Fire Weather Testbed. Users will continue to explore the tool and provide feedback for this year’s wildland fire season. NOAA plans to transition LightningCast to an operational status in 2025.
APPA has a webpage that offers member utilities resources related to wildfire mitigation planning.
