科学技术与工程
中国技术经济学会
数据中心园区的负载迁移和多能源调控策略优化
丁中亮 , 张泉 , 李俊山 , 雷建军 , 廖曙光 , 翟志强 , 马贞俊
科学技术与工程 ›› 2025, Vol. 25 ›› Issue (33) :14267-14276
数据中心园区的负载迁移和多能源调控策略优化
Optimization of Load Migration and Multi-energy Regulation Strategies for Data Center Park
为实现数据中心园区的低成本运营，构建了面向全湖水自然冷却数据中心的多能源系统。针对可再生能源发电的波动性及峰谷电价差异，提出了一种基于特定时空规则控制的粒子群算法优化的负载迁移与蓄／放电精细化调控策略，将延迟容忍算力负载与可再生能源出力、电价时段精准匹配，减少了配置容量冗余与绿电损耗，储能设备容量配置降低了23.1%；全年运行电费节省了19.7%，夏季和过渡季的绿电损耗率分别降低1.5%、0.4%；电源使用效率（power usage effectiveness, PUE）从1.121降至1.115，该策略为多能源与负载调度的优化提供了参考。
To achieve low-cost operation of the data center park, a multi-energy system for the fully lake water natural cooling data center has been constructed. In view of the volatility of renewable energy power generation and the difference in peak-valley electricity prices, a fine-grained regulation strategy for load migration and storage/discharge optimized by particle swarm optimization based on specific spatio-temporal rule control was proposed. This strategy precisely matches the delay-tolerant computing power load with the output and electricity price periods of renewable energy, reducing configuration capacity redundancy and green electricity loss. The capacity configuration of energy storage equipment has decreased by 23.1%. The annual operating electricity cost is saved by 19.7%, and the green electricity loss rate in summer and the transitional season is reduced by 1.5% and 0.4% respectively. The power usage effectiveness (PUE) decreased from 1.121 to 1.115. The strategy provides a reference for the optimization of multi-energy and load scheduling.
数据中心园区 / 负载迁移 / 设备容量优化 / 能源管理 / 制冷系统 / 粒子群优化算法
data center park / load migration / equipment capacity optimization / energy management / refrigeration system / particle swarm optimization algorithm
/
| 〈 |
|
〉 |
PDF
(13231KB)
