本文關鍵詞： 數(shù)值模擬 氣流組織 測量點位置優(yōu)化 正交函數(shù)逼近 聚類分析　出處：《中國海洋大學》2014年碩士論文　論文類型：學位論文

【摘要】：通信機房是網(wǎng)絡服務中至關重要的一部分，所謂數(shù)據(jù)中心機房通常是指在某一物理空間內(nèi)對信息的處理、存儲、傳輸、交換、管理等進行集中的處理和有機組合。電信設備的穩(wěn)定運行對機房環(huán)境有著較高的要求。同時，數(shù)據(jù)中心行業(yè)的發(fā)展越來越快，一臺3P空調每天的耗電量就達到了36.1度，大力開展節(jié)電節(jié)能工作刻不容緩，這對緩解我國能源緊張，建設低碳經(jīng)濟等問題，都具有非常重要的意義。 在通信機房室內(nèi)熱環(huán)境的模擬過程中，涉及到極其復雜的物理過程，要根據(jù)實際需求找到合適的模型與之匹配。本文通過對通信機房的實地測試，，在了解了計算流體力學的理論知識以及應用范圍的基礎上，對其數(shù)值計算方法進行了研究，并將各個方法進行了對比，找到適合本文模型的方法。結合運用k-ε模型及SIMPLEC算法，選擇合適的邊界條件、各項參數(shù)并借助于FLUENT簡化模擬出機房的溫度場和速度場等數(shù)據(jù)，得出因為實際中風管的沿程阻力、局部阻力等問題，各風口的出風量不均最高能達到9.67m/s，最低是3.27m/s。機房速度場分布并不均勻，而一般的送風口測試點分布都是規(guī)律性分布，對不同的模型使用相同的測試方法會產(chǎn)生較大誤差，使得模擬結果與客戶提供數(shù)據(jù)不能很好的對應。通過對流場的分析為后面的測量點位置改進做出鋪墊。本文的研究采用正交函數(shù)逼近法，這種算法在很大程度上能夠保證分布參數(shù)模型的空間分布特性，而且計算量小，算法簡單。在基于CFD技術的分布參數(shù)方法基礎上將偏微分方程轉化為代數(shù)方程組，從隨后的過程中求得未知參數(shù)，獲得較高精度的辨識結果。將從模擬結果得出的數(shù)據(jù)進行分析，借助于正交函數(shù)逼近法的計算和聚類分析中的模糊識別通過MATLAB來得出測量點的最優(yōu)測試結果。實現(xiàn)了測量點位置的改進。 在網(wǎng)絡服務快速發(fā)展的今天，機房各方面都有待優(yōu)化，數(shù)據(jù)中心的成功運作需要長期提供一個高效的散熱環(huán)境，并且期望以最低的運營費用來達到最高的可用性和可靠性。本文僅提出一種測量點位置改進的方案來為以后的通信機房測試點位置提供一種初步思路。該思路對于今后的通信機房改造也具有一定的意義。
[Abstract]:The communication room is a vital part of the network service. The so-called data center computer room usually refers to the processing, storage, transmission and exchange of information in a certain physical space. Management and other centralized processing and organic combination. The stable operation of telecommunications equipment has higher requirements for the computer room environment. At the same time, the data center industry is developing more and more quickly. The energy consumption of a 3P air conditioner is 36.1 degrees per day. It is of great significance to develop the energy saving and energy saving work, which is of great significance to alleviate the energy shortage and build a low-carbon economy in our country. In the process of simulating the indoor thermal environment of the communication computer room, the extremely complex physical process is involved, and a suitable model should be found to match it according to the actual demand. This paper passes the field test of the communication computer room. On the basis of understanding the theoretical knowledge of computational fluid dynamics and its application scope, the numerical calculation methods are studied and compared. Find the method suitable for this model and choose the appropriate boundary condition by using k- 蔚 model and SIMPLEC algorithm. With the help of FLUENT, the data of temperature field and velocity field of the engine room are simulated, and the problems such as the resistance along the path and the local resistance of the air pipe in practice are obtained. The distribution of velocity field in the engine room is not uniform, but the distribution of the test points of the general air supply outlet is regular. 3. The distribution of the uneven air output of the tuyere is 9.67 m / s, and the lowest is 3.27 m / s. The distribution of velocity field in the engine room is not uniform. Using the same test method for different models will result in large errors. The simulation results and customer data can not be very good correspondence. Through the analysis of the flow field to improve the position of the measurement points. This paper uses orthogonal function approximation method. To a large extent, this algorithm can guarantee the spatial distribution characteristics of the distributed parameter model, and the computational complexity is small. The algorithm is simple. On the basis of the distributed parameter method based on CFD technique, the partial differential equations are transformed into algebraic equations, and the unknown parameters are obtained from the subsequent process. High precision identification results are obtained. The data obtained from the simulation results are analyzed. By means of the calculation of orthogonal function approximation method and fuzzy recognition in clustering analysis, the optimal test results of measuring points are obtained by MATLAB, and the improvement of measuring points' position is realized. With the rapid development of network service, all aspects of computer room need to be optimized. The successful operation of data center needs to provide an efficient heat dissipation environment for a long time. It is expected that the highest availability and reliability can be achieved with the lowest operating cost. In this paper, we only propose a scheme to improve the location of the measuring points to provide a preliminary idea for the location of the test points in the future communication rooms. In the future, the transformation of the communication room also has certain significance.
【學位授予單位】：中國海洋大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TN80

【參考文獻】

相關期刊論文 前10條

1 彭軍;許淑惠;;空調房間置換通風數(shù)值模擬的探討[J];北京建筑工程學院學報;2008年03期

2 王琳,馬平;系統(tǒng)辨識方法綜述[J];電力情報;2001年04期

3 賴世能,侯福平,宗凌;IDC機房空調電源技術專題系列之一:IDC機房上下送風方式空調系統(tǒng)分析比較[J];電信技術;2005年11期

4 邢棉;;MATLAB數(shù)值計算在高等數(shù)學教學中的應用探討[J];中國科教創(chuàng)新導刊;2010年29期

5 朱強;周曉;;幾種數(shù)據(jù)流聚類算法分析[J];赤峰學院學報(自然科學版);2013年11期

6 楊蘋;肖瑩;簡棄非;羅學維;劉文飛;;通信機房的氣流組織特性與空調冷量調配方案[J];華南理工大學學報(自然科學版);2009年11期

7 楊新平;夏云安;;數(shù)據(jù)機房改造的問題與探討[J];華南金融電腦;2009年05期

8 范海平;曾小林;;基于計算流體動力學的流量系數(shù)研究[J];機電設備;2007年02期

9 吳靈敏;;IDC機房空調設計淺析[J];中國勘察設計;2008年12期

10 陳實;路建嶺;麥粵幫;潘城巋;;下送風通信機柜進風速度對其出風溫度的影響[J];潔凈與空調技術;2012年02期

相關博士學位論文 前1條

1 丁研;輻射供冷及獨立新風耦合系統(tǒng)送風形式與控制策略研究[D];天津大學;2012年

本文編號：1480609