Bridging the Gap Between Grid Requirements and Facility Performance
EPRI examines a growing reliability challenge facing the power sector: ensuring that large, power-electronics-dominated loads such as data centers remain connected during grid disturbances. As data centers continue to scale, individual facilities can reach gigawatt-level demand, making their unexpected disconnection a potential threat to bulk power system stability. The paper does not propose new ride-through requirements. Instead, it identifies gaps in how existing requirements are interpreted, implemented, and verified, and outlines areas where additional industry coordination is needed.
Why Ride-Through Capability Matters
Electrical disturbances, such as short circuits and system faults, are unavoidable in power systems. Historically, the temporary loss of individual loads during such events had limited system-wide consequences because loads were relatively small and geographically dispersed. Today, however, large data centers, cryptocurrency mining operations, and green hydrogen facilities can represent hundreds or even thousands of megawatts at a single point of interconnection. The sudden loss of these loads can cause voltage and frequency excursions, compromising grid reliability.
In response, utilities and system operators have begun establishing ride-through requirements that define how large loads should behave during voltage disturbances. These requirements are intended to keep facilities connected through specified voltage sags and support rapid recovery once normal conditions return. Utilities and transmission operators in North America and Europe—including ERCOT, SPP, AESO, RTE, Fingrid, EirGrid, and Energinet—have published or proposed ride-through curves that define the voltage levels and durations at which facilities are expected to remain connected.
Gaps Between Grid-Level Requirements and Facility Behavior
The disconnect between requirements defined at the point of interconnection and the behavior of equipment inside a data center is a core theme of this research. Ride-through requirements are typically expressed in terms of voltage measurements at the facility’s grid connection. Still, compliance ultimately depends on how internal systems—including uninterruptible power supplies (UPS), power distribution units, adjustable-speed drives, and protection schemes—respond to disturbances.
Several interpretation challenges are highlighted. Utilities may define the voltage used for compliance differently, such as using positive-sequence voltage or the most severe phase voltage. These distinctions can significantly affect compliance assessments during unbalanced faults. In addition, the voltage experienced by equipment within a facility may differ from the voltage measured at the point of interconnection due to transformer configurations, cable impedance, and internal distribution systems. As a result, equipment can experience more severe voltage conditions than those implied by system-level measurements.
There is also a distinction between utility ride-through requirements and equipment immunity standards such as CBEMA, ITIC, IEEE 1668, and SEMI F47. While immunity standards are intended to evaluate equipment survivability at low-voltage terminals, utility ride-through requirements are designed to maintain grid reliability. The lack of a consistent mapping between the two creates uncertainty in equipment specification, testing, and procurement.
Coordinating Data Center Systems for Compliance
Meeting ride-through requirements requires coordination across the entire electrical infrastructure of a data center. UPS systems and rack-level power supplies must be configured to continue supporting loads during allowable voltage sags, often through coordinated use of both grid power and battery systems. Automatic transfer switches and backup generation schemes must also be carefully configured to avoid unnecessary transfers to backup power during disturbances that remain within the ride-through envelope.
Cooling systems are a particularly important but often overlooked part of ride-through compliance. Modern hyperscale and AI-focused data centers rely on tightly controlled cooling architectures, and even brief interruptions can threaten thermal limits. While UPS systems may protect pumps and fans, chillers are often not protected. Voltage sag–induced chiller trips can therefore become a limiting factor in overall facility resilience. Mitigation approaches are discussed, including improved control power immunity, optimized voltage-sag settings for adjustable-speed drives, and rapid-recovery chiller features offered by several manufacturers.
Implementation, Verification, and Emerging Challenges
One of the industry’s largest unresolved challenges is demonstrating compliance. Although ride-through requirements are increasingly specified at the utility level, no widely accepted equipment-level testing or certification framework currently exists for large loads. Unlike distributed energy resources, which can demonstrate compliance through standards and certification programs such as IEEE 1547 and UL 1741, comparable frameworks for data center equipment are largely absent.
Manufacturers and developers face uncertainty when translating utility requirements into equipment specifications and procurement decisions. The authors note that requirements continue to evolve and differ across regions, creating additional complications for product design, firmware development, and long-term planning.
Another emerging issue involves facilities that combine large loads with on-site generation, sometimes referred to as “Energy Parks.” Existing ride-through requirements for loads and generators have generally been developed independently. The resulting misalignment can create unintended imports from or exports to the grid during disturbances, introducing new reliability risks unless protection and control systems are carefully coordinated.
DCFlex and the Path Forward
Drawing on lessons learned, the research emphasizes the importance of clear definitions, broad stakeholder engagement, harmonized requirements, and verifiable compliance pathways.
To help address current gaps, EPRI’s DCFlex collaborative is pursuing two complementary pathways. The first focuses on improving consistency and clarity in ride-through requirements by using common terminology, aligned definitions, and enhanced traceability between grid-level obligations and facility-level performance. The second focuses on developing practical testing, certification, and compliance methodologies, drawing on EPRI’s experience in power quality testing, UPS evaluation, and certification programs.
The ultimate goal is to move the industry from fragmented, utility-specific expectations toward a more harmonized framework supported by a common testing language, clear pass/fail criteria, coordinated facility controls, and transparent methods for demonstrating compliance. By bridging the gap between grid reliability objectives and equipment-level performance, DCFlex aims to help data centers support grid stability while providing greater certainty for utilities, developers, and equipment manufacturers.
Learn More
Banner image licensed by EPRI from Adobe Stock
Microsoft Copilot was used to generate a draft of this article from an EPRI publication. AI-generated content was reviewed, edited, and fact-checked by an EPRI expert to ensure accuracy and quality.
