Researchers at the University of Texas at Dallas (UT Dallas), in collaboration with engineers from the University at Buffalo in New York, have developed an AI-based system that can automatically repair the power grid, rerouting power within milliseconds.
The AI model aims to autonomously detect and repair problems on the power grid without the need for human intervention, according to a statement from the university.
According to one article Interesting engineering, “Our goal is to find the optimal path to get power to the most users as quickly as possible, but more work is needed to implement this system,” said Jie Zhang, an associate professor of mechanical engineering in the Eric Johnson School of Engineering and Computer Science.
The new system developed by the researchers focuses primarily on North American power grid systems.
The AI-powered system falls into the category of a “self-healing smart grid,” and researchers hope the technology will help reduce power interruptions during power outages.
The article further states that the system uses a machine learning technique called graph reinforcement learning to analyze and manage complex power grid networks.
The study was published on June 4th: Nature Communications.
Interesting Engineering The system has been tested in a variety of network configurations and has demonstrated near-optimal real-time performance, the company noted. The system also significantly reduced energy losses during power outages in a variety of scenarios.
“In this interdisciplinary project, we leveraged our team's expertise in power systems, mathematics and machine learning to explore how graph abstraction can be used to systematically describe the various interdependencies within electrical distribution systems,” the article quoted Dr. Julia Gel, one of the study's co-authors, as saying.
Published June 26, 2024 08:04 IST
