数据机房热管理的数值模拟与优化分析

发布时间：2018-05-03 22:42

本文选题：数据机房 + 气流组织CFD数值模拟　；参考：《西安电子科技大学》2013年硕士论文

【摘要】：随着3G时代的来临，通信行业的服务器发展迅速，刀片式服务器的应用也使机房内部的发热量越来越大，增加了机房的总能耗。由于数据机房的环境要求较高造成了空调系统每年消耗大量的电能，，约占总电费的54％左右。除了高昂的电费成本支出之外，机房内部环境温度不均一，同一时间测定的不同区域温差有时可达5℃以上，也造成了机房空调高能耗、低效率的现状。因此，如何合理地设计室内气流组织来充分有效地利用空调设备冷却服务器，从而提高其使用寿命也是值得关注的一个问题。本文以某一数据机房热管理实验平台为研究对象，主要开展了三方面的研究工作：首先，进行数据机房热管理实验平台的建模，根据实际机房布局尺寸，采用CFD数值模拟方法，建立热仿真模型，分析在摆放一台服务器和十台服务器的极值工况下的机房的速度场、温度场分布，根据计算结果校验两种极值工况是否符合实验机房热管理设计要求，使得CFD数值模拟在数据机房实验平台的建设中起到辅助设计的作用。并且通过热仿真分析与实际测试的对比分析，验证CFD所建模型的合理性以及求解设置的正确性。其次，考虑到空调的参数性能，机房布局形式对空调系统的利用效率和气流组织分布均有重大影响。本文通过CFD数值模拟分析空调性能参数的改变以及机房布局形式改变对气流组织分布的影响，重点对不同工况下的速度场以及温度场进行分析，分析得出单一因素改变对速度场和温度场产生的变化规律，以及针对典型机房的合理的机房布局参数，为实际机房实验平台的搭建和数据机房的节能改造提供参考。最后，针对内置两台机柜的典型机房，采用响应曲面法进行数据机房热管理的优化设计，选择空调送风温度、空调送风量、环境温度以及机柜间距为四个设计变量，两机柜出风口温度的平均值为响应值，利用响应曲面法中的中心复合实验设计来得到最优工况和两机柜出风口平均值的经验公式，来预测实际工况中的机柜出风口的温度，达到指导工程实践的作用，为多机柜的研究奠定基础。
[Abstract]:With the advent of the 3G era, the server of the communications industry has developed rapidly. The application of the blade server also makes the heating of the engine room more and more large, increasing the total energy consumption of the machine room. Because of the high requirements of the environment of the data room, the air conditioning system consumes a large amount of electricity every year, accounting for about 54% of the total electricity bill. In addition to the cost, the internal temperature of the machine room is not uniform, and the temperature difference measured at the same time can be up to 5 degrees centigrade at the same time. It also causes the high energy consumption and low efficiency of air conditioning in the machine room. Therefore, how to rationally design the air flow organization to fully and effectively use the cooling server of the air conditioning equipment so as to improve its service life is also to be improved. A matter of concern.
In this paper, a thermal management experimental platform of a data room is taken as the research object. Three main research works are carried out:
First, the thermal management experiment platform of the data room is modeled. According to the actual machine room layout and size, the CFD numerical simulation method is used to establish the thermal simulation model. The velocity field and temperature field distribution of the machine room under the extreme working conditions of a server and ten servers are analyzed, and the two extreme value conditions are verified according to the calculation results. The thermal management design requirements of the laboratory room make the CFD numerical simulation play an auxiliary role in the construction of the experimental platform for the data room. And through the comparison and analysis of the thermal simulation analysis and the actual test, the rationality of the model built by CFD and the correctness of the solution setting are verified.
Secondly, considering the parameter performance of the air conditioning, the layout of the machine room has a great influence on the efficiency of the air conditioning system and the distribution of air distribution. In this paper, the influence of the change of air conditioning performance parameters and the change of the layout form of the machine room to the distribution of air distribution is analyzed by CFD numerical simulation, and the speed field and temperature field under different working conditions are focused on. The analysis shows that the change of the single factor changes the velocity field and the temperature field, as well as the reasonable layout parameters of the machine room for the typical machine room, provides a reference for the construction of the experimental platform of the actual machine room and the energy saving transformation of the data room.
Finally, the response surface method is used to optimize the heat management of the data room with the response surface method. The response surface method is used to optimize the heat management of the data room. Four design variables are selected for air conditioning air supply temperature, air conditioning air supply volume, environment temperature and cabinet spacing, and the average value of the outlet temperature of the two cabinet is the response value, and the central compound experiment in the response surface method is used. The experiential formula of the optimal working condition and the average value of the outlet of the two cabinet is designed to predict the temperature of the outlet of the cabinet in the actual working condition, to guide the role of the engineering practice, and to lay the foundation for the research of the multi cabinet.

【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TU831

【参考文献】