基于人工智能算法的数据中心机房气流组织温度预测研究  

Research on Air Distribution Temperature Prediction of Data Center Room Based on Artificial Intelligence Algorithm

作　　者：许俊[1] 程序[1] 胡孝俊 姚贵策 祝华[3] 贺晓[1] Xu Jun;Cheng Xu;Hu Xiaojun;Yao Guice;Zhu Hua;He Xiao(China Information Technology Designing&Consulting Institute Co.,Ltd.Zhengzhou Branch,Zhengzhou 450007,China;Beihang University,Beijing 100191,China;China Unicom Shanghai Branch,Shanghai 200082,China)

机构地区：[1]中讯邮电咨询设计院有限公司郑州分公司,河南郑州450007 [2]北京航空航天大学,北京100191 [3]中国联通上海分公司,上海200082

出　　处：《邮电设计技术》2022年第12期52-56,共5页Designing Techniques of Posts and Telecommunications

摘　　要：在数据中心的总能耗中,制冷装置的能耗占50%。在中国提出“双碳”目标的背景下,设计建造智能数据中心,采取局部制冷的方式对降低制冷所需能耗,进而降低数据中心总能耗以及碳排放具有十分重要的意义。为达成这一目标,需要一种能够快速预测局部热点的方法。传统计算流体力学方法虽然也能实现预测,但对算力和运算时间都要求较高。因此,提出一种使数据中心走向智能化的关键技术,即在已知空调温度、机柜功率等参数下,对数据中心的温度场进行快速而准确的预测,它是智慧数据中心的重要组成部分。In the total energy consumption of the data center,the energy consumption of the refrigeration unit accounts for 50%. Under the background of China’s "dual carbon" goal,it is of great significance to design and build an intelligent data center and adopt local refrigeration to reduce the energy consumption required for refrigeration,thereby reducing the total energy consumption and carbon emissions of the data center. To achieve this goal,a method that can quickly predict local hot spots is needed. The traditional computational fluid dynamics method can also achieve prediction,but it requires high computational force and time.Therefore,it proposes a key technology to make the data center intelligent,that is,to quickly and accurately predict the temperature field of the data center when the air conditioning temperature,cabinet power and other parameters are known. It is an important part of forming a smart data center.

关 键 词：数据中心冷却 气流组织温度分布 卷积神经网络 机器学习 智能计算流体力学 

分 类 号：TU834.3[建筑科学—供热、供燃气、通风及空调工程]