基于区域失效的数据中心和5G基站冷却设备可靠性预测方法研究  

Reliability prediction method for cooling equipment of data centers and 5G base stations based on regional failure

作　　者：肖劲松 Xiao Jinsong(Guangdong Midea HVAC Equipment Co.,Ltd.)

机构地区：[1]广东美的暖通设备有限公司

出　　处：《制冷与空调》2024年第7期71-75,92,共6页Refrigeration and Air-Conditioning

基　　金：国家重点研发计划项目资助(2020YFF0218603)。

摘　　要：对数据中心和5G基站用冷却设备在不同安装区域的现场失效大数据进行分析,通过极大似然法建立寿命分布模型,其寿命分布最优拟合符合威布尔三参数分布规律,故障复发符合非齐次泊松过程,按区域聚类可靠性呈现显著性差异。对典型区域失效数据按可修复系统和不可修复系统分别进行可靠性预测,按可修复系统的预测结果与实际维修数据相关性良好。建立的基于区域失效的寿命分布、故障复发及可靠性预测方法适用于数据中心和5G基站用冷却设备的使用现场失效大数据分析,故障预测及维修备件、包修策略制定。The field failure data of cooling equipment used in the data center and 5G base station in different installation areas were analyzed.The maximum likelihood method was used to establish the life distribution model.The optimal fitting of the life distribution conformed to the three parameter of Weibull distribution,and the fault recurrence conformed to the non-homogeneous Poisson process.There was significant difference in the reliability of regional clustering.Reliability predictions are made for typical area failure data based on repairable and non repairable systems,and the predicted results based on repairable systems have good correlation with actual maintenance data.The established life distribution,fault recurrence,and reliability prediction method based on regional failures is suitable for data analysis of external failures of cooling equipment used in data centers and 5G base stations,as well as fault prediction and development of maintenance spare parts and repair strategies.

关 键 词：数据中心 5G基站 冷却设备 区域聚类 威布尔三参数 非齐次泊松过程 可靠性预测 

分 类 号：TB6[一般工业技术—制冷工程]