温度控制
理论(学习稳定性)
环境科学
控制(管理)
温度测量
控制系统
工程类
计算机科学
控制理论(社会学)
汽车工程
工作(物理)
作者
Yifan Zhao,Weixing Yuan,Tongzhi Yang,Kexian Ren
出处
期刊:
[Institute of Electrical and Electronics Engineers]
日期:2026-05-01
卷期号:: 12-17
标识
DOI:10.1109/celct69811.2026.11564363
摘要
In energy-retrofit projects for in-service (existing) data centers, a pronounced vertical elevation difference often exists between the heat-rejection end and the server room, and the retrofit must typically be implemented under online (no-downtime) conditions. Taking an in-service existing data center as the case study, this work upgrades the legacy cooling architecture to a pump-driven two-phase backplate cooling system for near-end thermal management while ensuring service continuity, and establishes a pPUE metric and a corresponding measurement boundary suitable for partial-retrofit scenarios. Four months of continuous field measurements show that, under an approximately 20m vertical elevation difference, the system operates stably and maintains the server-room temperature at 24.5±0.5°C under dynamic load disturbances. Benefiting from the high latent heat of vaporization of the working fluid and the incorporation of a vapor–liquid separator, the system mitigates the energy penalty typically associated with long-distance transport, yielding a cumulative pPUE as low as 1.20 over the monitoring period. These results demonstrate the engineering adaptability and superior energy performance of the proposed technology in complex, space-constrained retrofit settings, and provide a practical reference for the green and low-carbon transition of data infrastructure.
科研通智能强力驱动
Strongly Powered by AbleSci AI