行级数据中心热环境模拟及性能优化研究

Thermal environment simulation and performance optimization of row-based data centers

以某行级数据中心为研究对象,模拟研究通道封闭策略、空调位置、机柜负载分布三种因素对行级数据中心热环境的影响,并以回风温度指数(Return Temperature Index,RTI)和服务器过热参数(Server Overheating Parameters,SOP)为评价标准,对三种因素进行正交试验研究。结果表明,通道封闭策略是影响数据中心热性能的主要因素,其次是负载分布,空调位置影响最小。数据中心热环境最佳的设计方案是热冷通道同时封闭(Hot-Cold-Hot, H-C-H)、空调正对布置和负载上下分布,此时RTI和SOP分别为91.9%和1.154,相比于冷热通道同时封闭(Cold-Hot-Cold, C-H-C)、空调交叉1布置和负载均匀分布,RTI提升0.9%、SOP优化3.55%。

Taking a row-based data center as the research object,the effects of three factors of aisle enclosed strategy,air conditioner location and rack load distribution on the thermal environment of a row-based data center were simulated and studied,and the Return Temperature Index(RTI)and Server Overheating Parameters(SOP)were used as evaluation criteria,and orthogonal experiments were conducted to study the three factors.The results show that the aisle enclosed strategy is the main factor affecting the thermal performance of the data center,followed by the load distribution,and the air conditioner location has the least influence.The optimal design scheme for the thermal environment of the data center is that Hot and Cold aisle closure(Hot-Cold-Hot,H-C-H),air conditioner face to face arrangement and the load is distributed up and down.Currently,the RTI and SOP are 91.9%and 1.154,respectively.RTI is improved by 0.9%and SOP is optimized by 3.55%compared to Cold and Hot aisle closure(C-H-C),air conditioner cross-1 arrangement and uniform load distribution.