Abstract
High-density artificial intelligence and machine learning workloads introduce rapid and extreme power transients, which traditional data center power distribution and energy storage systems struggle to manage, leading to voltage and frequency oscillations. To address this, a tiered energy storage architecture is described for low-voltage direct current (LVDC) data centers. This architecture precisely maps specific energy storage technologies to distinct physical locations and time scales. Rack-level storage, utilizing high-power capacitors or supercapacitors, is placed immediately adjacent to the payload to mitigate sub-millisecond power fluctuations. Bus and row-level storage, including lithium-ion batteries or supercapacitors, functions as a critical intermediary buffer for seconds to minutes of backup, serving as the primary backup solution with a 1:1 allocated storage power to data center load ratio. Finally, facility-level bulk energy storage, such as BESS, flow batteries, or generators, provides long-duration options for minutes to hours, managing grid services like peak shaving and load shifting. This design ensures voltage stability, smooths power fluctuations, enhances power quality, and offers scalable, flexible energy management for demanding AI/ML environments.
Keywords: Tiered energy storage, LVDC data centers, AI/ML workloads, rack-level storage, row-level storage, facility-level storage, power quality, voltage stability, hybrid energy storage
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
N/A, "Tiered Energy Storage Architecture for Next-Gen Low Voltage Direct Current Data Centers", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/10119