A new artificial intelligence platform designed to detect the electrical faults that can ignite wildfires has received a 2026 R&D 100 Market Disruptor Award, placing a tool developed by the U.S. Department of Energy’s Oak Ridge National Laboratory and Southern California Edison among the year’s most notable innovations. The technology, known as the Multi-Event Grid Intelligence Platform for Wildfire Prevention and Resilience, or AWARE, is intended to give electric utilities a faster and more detailed view of dangerous conditions developing across the power grid. Its developers say the platform could help prevent fires, limit equipment damage, reduce outages and improve decisions during fast-moving grid emergencies.
The award-winning system addresses a problem that has become increasingly urgent as electrical infrastructure operates in hotter, drier and more heavily vegetated regions. Power lines, transformers and other equipment can experience abnormal conditions that are difficult to identify using conventional monitoring systems. In particular, low-current electrical arcing can occur when electricity jumps across a damaged conductor, contaminated insulator or narrow gap. The resulting discharge may be intermittent and too weak to trigger traditional protective devices, yet hot enough to ignite surrounding vegetation. Because the signal can be brief, irregular and hidden within ordinary fluctuations in electrical demand, detecting it rapidly has historically been a major challenge for utilities.
AWARE is designed to analyze electrical waveforms, the time-varying signals that describe how voltage and current behave on the grid. Rather than treating the grid as a collection of simple threshold measurements, the platform examines the shape, timing and characteristics of waveform disturbances. Artificial intelligence and machine-learning models assess those patterns to determine whether an event is occurring, classify the type of fault and estimate its seriousness. The system can also identify when an event began and distinguish separate disturbances that appear close together in time, converting complex electrical measurements into information that operators can act on.
That capability is important because real grid events rarely occur in isolation. A single waveform may contain several overlapping or repeated disturbances caused by equipment failure, conductor contact, switching operations or arcing. Conventional monitoring tools often issue an alert when a signal crosses a predetermined limit, but they may provide little context about what happened before or after the alarm. AWARE is intended to go further by performing multi-event analysis. It can separate multiple abnormalities within one captured signal and produce structured event reports that describe the disturbance in terms useful to utility engineers, control-room operators and field crews.
The platform can automatically notify utilities when it detects potentially dangerous behavior, including the subtle signatures associated with low-current arcing. Rapid notification could give operators an opportunity to isolate a section of the network, inspect equipment or alter operating conditions before a fault escalates. In a wildfire-prone region, even a short reduction in response time may matter. A damaged line that remains energized near dry grass or brush can transform a localized electrical problem into a fast-spreading fire, especially during periods of strong winds. By identifying abnormalities that might otherwise remain undetected, AWARE seeks to turn waveform data into an early-warning system for both infrastructure failures and public-safety risks.
Researchers at Oak Ridge National Laboratory have been validating the technology using five years of field-collected data from Southern California Edison. The long-term data set allows the team to test the algorithm against the irregular conditions found in an operating utility network rather than relying only on laboratory simulations. Southern California Edison has also contributed feedback on how the platform should function within utility workflows and how its alerts and reports can be integrated with existing systems. This collaboration has helped shift the project from a fault-detection algorithm toward a broader decision-support tool intended for practical use by utilities, grid operators, sensor manufacturers and organizations responsible for critical infrastructure.
Ali Ekti, who leads the project and the ORNL Grid Communications and Security Group, said the recognition was particularly meaningful because the system is moving from publicly funded research toward potential adoption by utilities. The platform combines signal processing, machine learning, event classification and reporting rather than relying on a single artificial-intelligence technique. That combination is designed to address a central problem in modern grid management: enormous quantities of continuous sensor data are collected, but operators need concise and reliable explanations of what the data means. AWARE is intended to bridge that gap by identifying patterns that may be invisible to conventional monitoring and presenting them as actionable events.
Support for the award nomination came from Southern California Edison and GridVisibility, a company that develops grid sensors. GridVisibility CEO Scott Caruso described the system as a practical advancement in the way waveform information can be used to improve reliability and resilience. The platform is designed to work with scalable sensing architectures, allowing data from edge devices to be analyzed and converted into standardized information about grid behavior. As utilities install more sensors throughout distribution networks, systems capable of interpreting high-frequency measurements could become increasingly important. Without automated analysis, the volume and complexity of those measurements would make it difficult for human operators to examine every disturbance in real time.
Future versions of AWARE may also help locate where a disturbance occurred. The proposed approach would compare measurements from sensors installed at multiple points on the grid. Because an electrical event reaches different sensors at slightly different times and may produce different waveform signatures at each location, those variations can be used to estimate the source of the disturbance. Pinpointing the likely location could help utilities narrow field searches, prioritize inspections and restore service more quickly. It could also be valuable during severe weather or wildfire conditions, when crews must make rapid decisions while roads, visibility and access may be limited.
The ORNL team behind AWARE includes Ali Ekti, Ozgur Alaca, Ali Boyaci and Bruce Warmack. Southern California Edison contributors include Kyle Chang, Hamed Valizadehhaghi, Vik Trehan, Michael Balestrieri and Kevin Richardson. The project is managed at Oak Ridge National Laboratory by UT-Battelle for the Department of Energy’s Office of Science, the largest U.S. supporter of basic research in the physical sciences. If deployed broadly, AWARE could represent a shift in how utilities respond to electrical threats: from waiting for obvious threshold violations to interpreting subtle, evolving patterns before they become outages, equipment failures or wildfire ignition events. Its Market Disruptor award recognizes not only the underlying algorithm, but the possibility that intelligent waveform analysis could make increasingly vulnerable power networks safer and more resilient.
Subject of Research: Artificial intelligence for electric-grid fault detection, wildfire prevention and grid resilience
Article Title: AI Platform Detects Hidden Grid Faults Before They Can Ignite Wildfires
Web References: U.S. Department of Energy Office of Science
Image Credits: Morgan Manning / Oak Ridge National Laboratory, U.S. Department of Energy
Keywords
Artificial intelligence, electrical arcing, wildfire prevention, electric grid, grid resilience, machine learning, waveform analysis, fault detection, Oak Ridge National Laboratory, Southern California Edison, power outages, grid reliability, critical infrastructure
Tags: AI-powered grid disturbance detectionclimate-adaptive electrical systemsearly wildfire ignition detectionelectrical arcing detectionelectrical fault monitoringgrid emergency decision supportmodern power grid safetyR&D 100 Market Disruptor Awardsmart power grid sensorsutility infrastructure resiliencevegetation-related electrical faultsWildfire prevention technology
