--
Brent Crude $91.82/bbl ▼ -8.5%WTI Crude $84.25/bbl ▼ -8.2%Henry Hub Gas $2.63/MMBtu ▼ -8.4% Brent Crude $91.82/bbl ▼ -8.5%WTI Crude $84.25/bbl ▼ -8.2%Henry Hub Gas $2.63/MMBtu ▼ -8.4%
← Back to Smart Grid Smart Grid

Sandia AI Controls Stabilize Grid Voltage as Data Center Demand Surges

Sandia AI Controls Stabilize Grid Voltage as Data Center Demand Surges

⚡ AI Executive Summary

Sandia National Laboratories has developed an AI-driven Distributed Energy Resource Management System (DERMS) to regulate voltage in real time as data center electricity demand grows unpredictably. The technology is critical for utilities managing increasing numbers of distributed resources like solar, batteries, and backup generators while maintaining power quality for sensitive loads. The system has progressed from laboratory testing with hardware-in-the-loop validation to successful field demonstrations in Texas, with commercialization efforts now underway through the DOE Energy I-Corps program.

Rapid growth in artificial intelligence and data center operations is creating unprecedented challenges for electric utilities trying to maintain stable voltage across distribution grids. As traditional power generation patterns shift and distributed energy resources proliferate, utilities face tighter margins for error in keeping electricity supply reliable for critical infrastructure.

Sandia National Laboratories has responded by developing an AI-enabled control system that automatically coordinates grid-connected devices—such as inverters, batteries, and solar arrays—to regulate voltage fluctuations in near-real time. The Distributed Energy Resource Management System, or DERMS, represents a departure from conventional mechanical approaches like capacitor banks and line voltage regulators that operate more slowly through discrete switching.

The DERMS platform forecasts electricity supply and demand changes, coordinates device actions across the distribution network, and automatically responds to disturbances. By leveraging inverter capabilities already installed at many grid sites, utilities can improve voltage stability without major infrastructure upgrades.

Sandia's development process followed a rigorous validation pathway. Engineers first tested the system in computer simulations, then moved to power hardware-in-the-loop testing in the laboratory, where real commercial inverters and battery hardware were connected to a real-time digital grid simulator. This approach revealed communication delays and data-link challenges that simulations alone cannot capture.

Field demonstrations followed at two Texas sites: Sandia's SWiFT wind facility and the Texas Tech University GLEAMM microgrid, which includes a data center. Side-by-side testing showed measurable improvements; voltage that typically ran 5 percent above target values moved closer to desired setpoints when the DERMS controller was active.

Beyond utility benefits, the work addresses national security concerns. Military and critical defense infrastructure depend on consistent power quality, and the system's ability to respond agilely to disruptions—whether from natural disturbances or deliberate attacks—strengthens infrastructure resilience.

The project has been selected for Phase III of the DOE Energy I-Corps program, signaling movement toward commercialization. Industry feedback indicates utilities prioritize solutions that simplify device coordination without adding operational complexity, a requirement the current platform design addresses.

#voltage regulation#distributed energy resources#DERMS#AI controls#grid stability#data center demand#microgrid#power quality
Original source: Sandia National Labs ↗

Related in Smart Grid