American Nuclear Society
Home

Home / Publications / Journals / Nuclear Technology / Volume 210 / Number 6

Modeling Microreactor Requirements for High-Performance Computing

Alvin J. H. Lee, Lucas Wodrich, Dimitri Kalinichenko, Caleb S. Brooks, Tomasz Kozlowski

Nuclear Technology / Volume 210 / Number 6 / June 2024 / Pages 1027-1041

Research Article / dx.doi.org/10.1080/00295450.2023.2276999

Received:April 30, 2023
Accepted:October 22, 2023
Published:May 2, 2024

The potential deployment of microreactors as a zero-emission source for critical applications within integrated energy systems such as microgrids has been gaining interest in recent years owing to the microreactors’ dispatchable nature, modular design, small site footprint, and carbon-free generation. A particularly high-value but challenging application with rapidly growing demand is in the deployment of high-performance computing (HPC) clusters within microgrids. In this work, a model of a HPC cluster in an energy-diverse microgrid is developed to determine the requirements of a technology-agnostic microreactor deployed for such a challenging application. The minute-resolution simulations revealed that the cluster’s electrical load fluctuation of up to 4.1 MW/min required a fast and responsive load-following capability. When the load-following capability of the microreactor was perturbed, the required microgrid storage capacity associated with having a 0.1 MW/min dispatchable microreactor decreased by two orders of magnitude as compared with load-following solely by energy storage devices, indicating that load-following capability in microreactors is of great value in such applications. The analysis methods described in this work can be extended to other microgrids, other HPC clusters, or other types of challenging applications, and can help microgrid planners in determining the storage size, output capacity, and ramping capabilities of the storage devices required for a given microgrid configuration.