An artificial intelligence (A.I.) model developed by researchers at the University of Texas at Dallas, in collaboration with engineers at the University at Buffalo in New York, may be able to prevent power outages by automatically rerouting electricity at the grid in milliseconds, according to a study published in June 2024 in Nature Communications.

This early example of “self-healing grid” technology uses A.I. to identify and repair problems without human intervention, which could be beneficial in addressing power lines damaged by storms.

“Power grids across the world are being challenged by the growing number of extreme weather events, the likelihood of cyberattacks and projected increases in demand,” said Souma Chowdhury, co-author of the study, associate professor in the UB Department of Mechanical and Aerospace Engineering and co-director of the UB Center for Embodied Autonomy and Robotics (CEAR). “Therefore, it is imperative that we develop tools that modernize the system and make it more resilient against future power outages.”

Researchers leveraged reinforcement learning that chooses the options that achieve the best results. They demonstrated that their system can identify alternative routes that transfer electricity to users even before an outage happens—and it’s done in milliseconds by A.I. instead of the minutes to hours it would take for humans to decide which alternate paths to use.

“Our goal is to find the optimal path to send power to the majority of users as quickly as possible,” said Jie Zhang, co-author of the study and associate professor of mechanical engineering at the University of Texas at Dallas. “But more research is needed before this system can be implemented.”

Using algorithms that apply machine learning to graphs, the researchers mapped the relationships between entities that make up a power distribution network. If line faults block the path of electricity, the system can reconfigure pathways using switches, drawing power from close, available sources.

“These are decisions that the model can make almost instantaneously, which in turn [have] the potential to eliminate or greatly reduce the severity of power outages,” said co-author Steve Paul, then a UB doctoral student, now a postdoctoral research associate at the University of Connecticut.

Furthermore, mapping network topology may aid in applying A.I. to other complex systems, including critical infrastructure and ecosystems, claimed study co-author Yulia Gel, professor of mathematical sciences at the University of Texas at Dallas.

After focusing on preventing outages, researchers hope to develop technology to repair and restore the grid after a power disruption.