面向任务调度的自然水冷数据中心控制策略  

A control strategy of natural water-cooling for data center system based on task scheduling

作　　者：唐辉 张泉[1] 王加强[1] 陈晓明 雷奥君 Tang Hui;Zhang Quan;Wang Jiaqiang;Chen Xiaoming;Lei Aojun(College of Civil Engineering, Hunan University, Changsha 410082, China;College of Ecological Environment and Urban Construction, Fujian University of Technology, Fuzhon 350118, China)

机构地区：[1]湖南大学土木工程学院,长沙410082 [2]福建工程学院生态环境与城市建设学院,福州350118

出　　处：《低温与超导》2018年第4期51-57,共7页Cryogenics and Superconductivity

基　　金："十三五"国家重点研发项目;政府间国际科技创新合作重点专项(2017YFE0105800)

摘　　要：提出了一种面向任务调度的数据中心自然水冷冷却系统控制策略。该方法综合数据机房任务、冷却系统特征,将任务负载进行集中调度分配,同时实时优化了水泵和风机流量、末端精密空调台数等运行参数,有效降低了冷却系统能耗。以东江湖数据中心实际工程项目为研究对象,利用Matlab分别建立数据机房和冷却系统的传热和能耗模型,对比研究了两种传统控制策略:精密空调台数优化+均分调度方案、无台数优化+均分调度方案的节能性,典型一周任务负载下的模拟结果表明:该控制策略能显著降低自然水冷冷却系统能耗,节能率分别提高了约17.8%、25.9%,为数据中心冷却系统优化提供了一种新思路。This paper proposed a novel control strategy of natural water - cooling for data center based on task scheduling. A sever load centralized management framework was introduced for multi - room data center cooling system, then the flow rate of pumps and fans as well as the sequencing control of Computer Room Air Handling （ CRAH ） units were optimized to reduce the energy consumption of cooling system. Simulations for a week containing the typical load dynamic characteristics were performed on the Matlab platform, taking a practical engineering project of the Dongjiang Lake Data Center as a simulation example to devel- op the thermal dynamic and energy consumption models. The results show that the presented control strategy can achieve energy saving of 17.8% over the traditional control strategy with optimal sequencing control and task centralized; while it can obtain a- bout 25.9% compared with the case under non - optimal sequencing control and homogeneous scheduling. It will provide a new view for the optimization of data center cooling system.

关 键 词：数据中心 任务调度 控制策略 自然水冷 

分 类 号：TP308[自动化与计算机技术—计算机系统结构]