Oak Ridge Lab Cut Power Grid Optimization Time
Researchers developed a method that reduces grid scheduling calculations from 10 hours to less than one second.
Updated on Sept. 21, 2026 in Energy
Live Poll
Do you trust that the power grid can reliably accommodate the growth of AI data centers?
Researchers at Oak Ridge National Laboratory have developed a new computational approach to optimize power generation scheduling. The method, which is currently in the research stage, manages complex grid systems by converting lengthy calculations into near-instantaneous operations.
Why it matters
As U.S. data centers prepare for an influx of tens of gigawatts of new electricity demand, current optimization tools struggle to manage the associated complexity. This new method provides a scalable solution for integrating large-scale power generation systems.
The researchers reduced the calculation time for a 276-generator system from 10 hours to less than one second. The method, tested across systems ranging from 46 to 276 generators, is designed to scale to grids with tens of thousands of generators.
The players
Oak Ridge National Laboratory
A Department of Energy research facility that specializes in high-performance computing, materials science, and energy technology.
Eve Tsybina
A lead researcher at Oak Ridge National Laboratory focused on computational methods for power systems.
Slaven Peles
A researcher at Oak Ridge National Laboratory contributing to algorithmic optimization for energy infrastructure.
Shaked Regev
A researcher at Oak Ridge National Laboratory specializing in grid management and power generation scheduling.
The details
The research team utilizes a relax-and-round approach to solve integer variable problems—mathematical puzzles where choices must be whole numbers like on or off. By relaxing these discrete yes-or-no requirements into continuous values, the system solves the optimization quickly. It then performs a rounding process to translate these continuous values back into practical, actionable power generation decisions.
Timeline
September 21, 2026: Research findings were reported and presented.
The Tech Race
This research follows a tradition of grid efficiency studies funded by the Laboratory Directed Research and Development program. It marks a significant departure from traditional optimization methods that previously lacked the efficiency required for managing massive, modern integer variable systems.
This method aims to enable grid operators to manage increasing electricity demand from data centers without long processing delays. While currently in the research phase, the technique offers a pathway for utilities to automate large-scale generation scheduling in real-time.
The takeaway
This method transforms the computational burden of grid management, offering a way to handle thousands of generators instantly. Watch for follow-up testing of the algorithm on real-world utility networks to see if these benchmark gains translate into production stability.
Further reading
For broader context on current grid efficiency efforts, see the latest developments in Energy.
More information
Find more details on grid research via the DOE Office of Science information portal.
Source note: This article includes information reported by HPCwire.
Live Poll
Do you trust that the power grid can reliably accommodate the growth of AI data centers?
