


摘 要:大型数据中心投产后,运维人员需要有一个模拟环境来解决一系列冷源群控系统的问题,如研究节能控制策略为降低PUE贡献价值;调测优化后的自控程序以提高新程序执行的可靠性;升级更友好的运维人机界面以提升运维效率;提供自控系统操作和学习的途径以提升一线运维人员的操作能力等。将以国内某数据中心冷源群控仿真系统为例来阐述此类系统的功能、构成、特点及应用实践,以期为运维管理者提供对应的解决方案。
关键词:运维;数据中心;冷源集群控制;仿真系统
中图分类号:TP391.9 文献标识码:A文章编号:2096-4706(2021)22-0020-04
Abstract: After the large data center is put into operation, the operation and maintenance personnel need a simulation environment to solve a series of problems of cold source cluster control system, such as studying the energy-saving control strategy to contribute value to reducing PUE; adjust and measure the optimized automatic control program to improve the reliability of the implementation of the new program; upgrade more friendly operation and maintenance man-machine interface to improve operation and maintenance efficiency; provide automatic control system operation and learning ways to improve the operation ability of front-line operation and maintenance personnel, etc. Taking the cold source cluster control simulation system of a domestic data center as an example, this paper expounds the function, composition, characteristics and application practice of this kind of system, in order to provide corresponding solutions for operation and maintenance managers.
Keywords: operation and maintenance; data center; cold source cluster control; simulation system
0 引 言
大型數据中心的制冷需求庞大,冷源系统会按照N+X或2N的架构进行设计,单栋楼冷冻站内设置多台冷水机组已是标准配置,跨楼栋各冷冻站联动供冷也已在超大型数据中心中有所应用,与此对应的冷源集群控制(以下简称冷源群控)系统的架构一般会考虑冗余或容错[1]其控制逻辑越发复杂。系统投产后,根据实际自控运行工况和运维人员使用情况的反馈,我们会发现自控程序、人机互动界面等均会产生改进需求,同时还会发现一些程序BUG需要修复。此时如要对原冷源群控系统进行升级,首先需要申请合适的停机窗口期,其次需要在变更阶段进行大量的测试验证工作,这对需要满足容错的A级数据中心[2]来说压力巨大。如何能够缩短停机窗口期、减少变更工作从而有效降低系统变更的风险是我们需要重点考虑的问题。……