WS3: Grid Planning for Operational Flexibility

Regions across the globe are expecting significant load growth from data centers. A prior EPRI white paper presented a framework for forecasting regional data center load growth. This report provides case studies on using the framework to create forecasts for three regions: New York State, the Pacific Northwest, and the Southeastern United States. The results illustrate how electric system planners can apply the framework using data available from their load interconnection queues and the DCFlex project. Additionally, the results highlight the importance of forecasts accounting for not only uncertainty regarding how projects in the load interconnection queue will enter electric service, but also uncertainty related to differences in requested capacity versus actual demand and load ramping schedules.

The industry is currently facing a significant increase in large load interconnection requests, primarily driven by the rapid expansion of data centers. This surge has sparked considerable interest in understanding the specific interconnection requirements for these substantial loads. This Volume II installment aims to provide reporting on the evolving landscape of interconnection requirements for large loads.

This document summarizes the North American Electric Reliability Corporation (NERC) guideline on Risk Mitigation for Emerging Large Loads, focusing on the reliability challenges posed by data centers, cryptocurrency mining, hydrogen electrolyzers, and other power-intensive facilities. These loads introduce rapid demand variability, oscillations, and operational risks that require proactive measures across interconnection, planning, operations, stability, power quality, and resilience. The guideline emphasizes standardized data collection, formalized interconnection processes, advanced planning methodologies, and collaborative approaches among transmission operators, planners, balancing authorities, equipment manufacturers, and large load entities. EPRI’s initiatives, including DCFlex and resource adequacy modeling, support industry readiness by enhancing forecasting, reserve strategies, and dynamic modeling. Ultimately, the guideline calls for large load entities to transition from passive consumers to active partners in ensuring Bulk Power System reliability and resilience.

This study explores how strategic siting and load flexibility can accelerate data center integration while maintaining system stability. Using a stylized Texas grid model for 2030, we assess hosting capacity under different transmission and siting scenarios and evaluate four flexibility archetypes. Findings underscore the importance of coordinated planning, flexibility-enabled architectures, and transmission enhancements to unlock hosting capacity and support resilient grid operations.

This report provides a library of load shapes for a range of type of data centers. The load shapes were derived from empirical measurements from operational data centers. System planners and load forecasters can use the load shapes directly. Note this is the first version of the load shape library and EPRI plans to refine and expand the library throughout the DCFlex project.

This report considers interconnection requirements regarding modeling, performance, ancillary services, and co-located generation. Of these requirements, the most well established for both the US and Europe are those focused on power quality. Both regions are also seeing focused requirements emerging for data centers and hydrogen electrolysis loads. For those differences, the review indicates that between the two regions, Europe has generally more defined requirements for load interconnection and some of its areas have stricter performance requirements during degraded system voltage and frequency conditions.

This report provides an initial perspective on strategies and data to help electric sector planners project regional data center load growth. This working paper is an internal DCFlex project deliverable. The intent is to improve strategies for forecasting over the three years of the project, leveraging the DCFlex community’s diverse knowledge and perspectives.