Keeping grid voltage stable has always required careful coordination—but as data centers multiply and electricity demand grows faster and less predictably, that task is getting harder. Sandia National Laboratories, backed by the Department of Energy, has now moved a potential solution out of the lab and into the field.
Engineers recently completed demonstrations of an AI-driven distributed energy resource management system at two sites in Lubbock, Texas, where the platform coordinated grid-connected devices in real time to regulate voltage. Side-by-side tests—one day with the controller active, one without—showed measurable improvements in voltage stability.
Sandia completes field demonstrations of AI voltage control in Lubbock
The two test sites were the Scaled Wind Farm Technology facility, known as SWiFT, and the Texas Tech University GLEAMM microgrid—both in Lubbock, Texas. At SWiFT, engineers connected the DERMS platform to operating equipment and watched how it responded as real-world grid conditions shifted. The goal was straightforward: confirm that the system could perform reliably at mission-critical sites where extended downtime simply is not an option.
GLEAMM went a step further. That microgrid includes a data center, making it a realistic stand-in for the kind of high-demand environment utilities increasingly have to manage. Engineers ran a full day with the controller active and a full day without. The voltage graphs from those two days told the story clearly.
Voltage at the GLEAMM site was running roughly 5% above the value utilities aim to maintain. With the DERMS controller coordinating grid-connected devices in real time, voltage moved closer to that target—and the improvement was not marginal. It was apparent in the data.
Why voltage management is becoming harder for utilities
Data centers are multiplying rapidly, and their electricity demand is both large and difficult to predict. That growth puts new pressure on distribution grids already managing a more complex mix of resources. Rooftop solar, battery storage, and backup generators have added flexibility, but they have also added coordination challenges—each resource responds differently to changing conditions, and utilities must keep all of them working together without letting voltage drift outside acceptable limits.
Traditional voltage regulation relies on mechanical equipment—capacitor banks and line voltage regulators—that switches on and off in response to grid conditions. These devices work, but they respond more slowly than the fluctuations a modern distribution grid can produce.
Frequency management typically happens at the bulk power system level. Voltage regulation is a local problem, one that demands fast, precise responses to protect sensitive loads. That requirement is becoming harder to meet with conventional tools alone.
How the DERMS platform coordinates grid devices using AI
The software platform forecasts changes in electricity use and available power, then automatically coordinates inverters and other grid-connected devices to respond. AI enables that coordination to happen in near real time while respecting each device’s operating limits.
A key design choice was to work with inverters already installed in the grid rather than requiring utilities to buy new infrastructure. Inverters — which connect resources like solar panels and batteries to the grid — are already present at many sites, and the DERMS platform treats them as active participants in voltage regulation rather than passive hardware.
Before any field deployment, the team validated the controls using power hardware-in-the-loop, or PHIL, testing at Sandia’s Distributed Energy Technologies Laboratory. That method connects real commercial inverters and battery hardware to a real-time grid simulation, letting researchers test how the control software behaves under realistic conditions — including communication delays and fast-changing scenarios — without touching the actual grid. As Sandia researcher Jon Berg noted, testing with real hardware rather than models gives both the team and its partners confidence that the technology is ready for real-world use.
National security applications and path to commercialization
Sandia frames this work as more than a utility operations problem. Senior manager Charlie Hanley has stated directly that adversaries target energy infrastructure in conflict scenarios and that the DERMS system is designed to keep critical systems operating through adversarial disruptions. The platform’s ability to adapt in real time is central to that mission.
The project was accepted into DOE Energy I-Corps Phase III, a competitive program that accelerates the path from research to deployment. Earlier phases involved structured interviews with utilities, microgrid developers, and industry partners—feedback that pointed consistently to the same need: tools that integrate new equipment without adding significant operational burden. The work also aligns with the DOE’s Genesis Mission, which directs AI toward the nation’s most complex science and technology challenges.
What the Lubbock results mean going forward
The Lubbock demonstrations produced a clear picture. An AI-driven DERMS platform can coordinate distributed energy resources in real time to improve voltage stability, and field tests at two representative sites—one of them including a data center—confirmed what laboratory hardware testing had already suggested.
The system reduces reliance on slower mechanical switching, works with inverters already in the grid, and can respond to both routine fluctuations and deliberate disruptions. For utilities managing a more complex distribution grid, that combination addresses a real and growing gap. For national security purposes, it offers a layer of resilience that conventional equipment cannot easily replicate. Broader deployment is the next step, with Energy I-Corps Phase III supporting that push.
Kelly is an experienced writer with 15 years of experience exploring the big stories that shape our world, from tech breakthroughs and space exploration to climate, energy, and the fascinating quirks of science. She has a talent for turning complex ideas into sharp, memorable insights that stay with readers long after they’ve finished reading.
