数据中心蓄电池自动预防性维护策略探究

本文分析了现有的数据中心蓄电池使用时间和容量不能达到预期设计的问题进行了探究,针对数据中心电池在多次充放电之出现损耗的问题,对单体蓄电池的性能进行分析,首先建立CPU/内存/IO/PSU使用率能耗模型优化,根据预测出的CPU/内存/IO/PSU使用率服务器恒功率负荷曲线,再利用DCIM蓄电池自动预防性维护调度算法得出24h内的服务器恒功率负荷的不同变化区间内最优蓄电池自动预防性维护充放电恒功率策略,即每个服务器负荷恒功率的不同区间内电池是否可进行恒功率充放电维护。以此达到利用服务器恒功率大小来设定蓄电池维护所需放电恒功率大小的目标,完成数据中心蓄电池自动预防性维护策略。

Research on Automatic Preventive Maintenance Strategy of Battery in Data Center

This paper analyzed the phenomenon that battery life and capacity in data center could not meet the expected design issues explored for the data center. To solve the problem of that the battery capacity lost much after a number of charge and discharge.We analyzed the performance of a single battery.The first step is that we optimized the mode of predicted CPU / memory / IO / PSU utilization and got the server constant power load curve.And then used the DCIM battery automatic preventive maintenance scheduling algorithm to get the server within 24 hours after the CPU / memory / IO / PSU utilization energy consumption model optimization, Constant power load of different changes in the interval of the optimal battery automatic preventive maintenance charge and discharge constant power strategy, that is,each server load constant power in different intervals whether the battery can be constant power charge and discharge maintenance.In this way we can achieve goal of the use of the size of the server constant power to set the size of the battery discharge constant power maintenance,and to complete the data center battery automatic preventive maintenance strategy.