Information and data technology is booming, and the global demand for data centers is growing rapidly as their services are integrating with nearly every aspect of modern society. This growth is outpacing the supply of available power from the standard electric utility grid, creating significant challenges for grid operators, including regional constraints on capacity and long lead times for establishing new connections. In the presence of these growing challenges and uncertainty from the utility grid, many are considering alternate sources, such as localized microgrids. A central consideration in designing such microgrids is the balance between reliability, availability, and cost. Through its emphasis on reliability metrics, the nuclear industry has developed a reputation for safety and performance and is a top-candidate power source for data centers.

This work investigates the capabilities of a representative nuclear-enabled microgrid for supplying power to a data center. Generation risk assessment (GRA) with fault tree analysis is employed to evaluate the microgrid operation by estimating the probability and duration for a generator availability due to unanticipated failures and anticipated maintenance downtimes. GRA offers a pathway for comparing microgrid configurations and identifying effective microgrid designs capable of meeting data center availability targets. The output of GRA provides reasonable size and capacity constraints for the microgrid power generators. Ultimately, this work provides inputs and considerations for conducting GRA of nuclear-enabled microgrids and offers a reliability basis for future technoeconomic assessments of data centers.