本文選題：大數(shù)據(jù) + 輻射傳輸仿真計算��； 參考：《北方工業(yè)大學(xué)》2017年碩士論文

【摘要】：本文基于風(fēng)云三號(FY3)衛(wèi)星數(shù)據(jù),研究了輻射傳輸仿真計算的結(jié)構(gòu)和流程。在研究過程中發(fā)現(xiàn)了在大數(shù)據(jù)環(huán)境下單點計算無法快速完成指定任務(wù)的問題。根據(jù)輻射傳輸仿真計算的結(jié)構(gòu)、流程以及數(shù)據(jù)等因素,結(jié)合并行計算相關(guān)知識,提出了加快仿真計算效率的方案,并且根據(jù)該計算方案將輻射傳輸仿真計算部署于一個適合該方案的集群系中去。最后,對輻射傳輸仿真的計算數(shù)據(jù)進(jìn)行了實驗分析。本文的主要創(chuàng)新與內(nèi)容如下:1、輻射傳輸仿真的并行計算設(shè)計方案。為了解決輻射傳輸仿真計算在單一節(jié)點下計算慢的問題,大量研究了輻射傳輸仿真計算的流程中應(yīng)用的數(shù)據(jù)的提取和處理的方式,分析了計算的物理結(jié)構(gòu)。通過實驗統(tǒng)計了計算過程的每一個步驟消耗的時間。根據(jù)以上總結(jié)歸納的輻射傳輸仿真的特性,結(jié)合研究的并行計算設(shè)計方法,提出輻射傳輸仿真的并行計算方案。2、輻射傳輸仿真集群系統(tǒng)的實現(xiàn)�；谳椛鋫鬏�?shù)牟⑿杏嬎惴桨?設(shè)計并實現(xiàn)了一個符合計算方案的集群計算系統(tǒng)。此系統(tǒng)結(jié)合集群技術(shù)知識,滿足集群的單一系統(tǒng)映像、可用性高、低成本等優(yōu)點和特性。根據(jù)實驗證明,該集群系統(tǒng)提高了輻射傳輸仿真計算的效率,說明了輻射傳輸仿真計算并行設(shè)計方案的可行性。3、集群管理系統(tǒng)的實現(xiàn)�；谳椛鋫鬏敺抡婕合到y(tǒng)的復(fù)雜結(jié)構(gòu)以及管理困難等問題,設(shè)計了一個界面友好,功能齊全的集群管理系統(tǒng)。該管理系統(tǒng)實現(xiàn)了對集群系統(tǒng)中任務(wù)的監(jiān)控、節(jié)點的監(jiān)控和管理等功能,為用戶提供了一個良好的集群系統(tǒng)的管理平臺。4、本文通過同一組數(shù)據(jù)對集群系統(tǒng)和單計算機(jī)分別進(jìn)行計算,通過對比消耗時間,證明了集群的高計算性能。通過大量風(fēng)云三號衛(wèi)星數(shù)據(jù)的輻射傳輸仿真計算結(jié)果,分析了不同數(shù)據(jù)對同一組數(shù)據(jù)的敏感性。并且根據(jù)誤差的分析,判斷不同的數(shù)據(jù)對仿真數(shù)據(jù)的影響。
[Abstract]:In this paper, based on Fengyun three (FY3) satellite data, the structure and process of simulation calculation of radiation transmission are studied. In the process of research, the problem that the assigned task can not be completed quickly by single point calculation under the large data environment is found. According to the structure, process and data of the simulation calculation of radiation transmission, the related knowledge of parallel computing is proposed. The scheme to speed up the efficiency of simulation calculation is proposed, and the simulation calculation of radiation transmission is deployed in a cluster system suitable for the scheme according to the calculation scheme. Finally, the experimental analysis of the calculated data of the radiation transmission simulation is carried out. The main innovations and contents of this paper are as follows: 1, the parallel computing design scheme of the radiating transmission simulation is understood. The simulation calculation of radiative transmission is slow under a single node. A large number of methods of extracting and processing the data used in the simulation process of radiation transmission are studied. The physical structure of the calculation is analyzed. The time consumed by each step of the calculation process is calculated by the experiment. The radiation transmission is simulated according to the above summarization. Real characteristics, combined with the parallel computing design method of research, a parallel computing scheme of radiation transmission simulation.2, the implementation of the radiation transmission simulation cluster system. A cluster computing system based on the parallel computing scheme of radiation transmission is designed and implemented. This system combines the cluster technical knowledge to satisfy the cluster single. A system image, high availability, low cost and other advantages and characteristics. According to the experiment, it has been proved that the cluster system improves the efficiency of simulation and calculation of radiation transmission, illustrates the feasibility of the parallel design scheme of the simulation and calculation of radiation transmission.3, the realization of the cluster management system, and the complex structure and management difficulties based on the radiation transmission simulation cluster system. A cluster management system with friendly interface and full function is designed. The management system implements the functions of monitoring, monitoring and management of the tasks in the cluster system, providing a good management platform for the users of the cluster system,.4. In this paper, the cluster system and the single computer are calculated by the same group of data. By comparing the consumption time, the high computing performance of the cluster is proved. The sensitivity of different data to the same group of data is analyzed through the simulation results of a large number of wind and cloud three satellite data, and the influence of different data on the simulation data is determined according to the analysis of the error.
【學(xué)位授予單位】：北方工業(yè)大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：P409;TP311.52

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 買買提依明·買買提;塔西甫拉提·特依拜;買買提沙吾提;丁建麗;;基于6S模型的遙感數(shù)據(jù)大氣校正應(yīng)用研究[J];水土保持研究;2011年03期

2 ;Construction of a Small Scale High Performance Cluster[J];Annual Report of China Institute of Atomic Energy;2009年00期

3 朱曉敏;陸佩忠;;Multi-Dimensional Scheduling for Real-Time Tasks on Heterogeneous Clusters[J];Journal of Computer Science & Technology;2009年03期

4 裴亞輝;張兵利;;計算機(jī)集群技術(shù)應(yīng)用[J];中國西部科技;2008年30期

5 喬延利;鄭小兵;王先華;張黎明;易維寧;王樂意;;衛(wèi)星光學(xué)傳感器全過程輻射定標(biāo)[J];遙感學(xué)報;2006年05期

6 ;集群架構(gòu)支撐氣象預(yù)報——山東省搭建氣象數(shù)值預(yù)報系統(tǒng)[J];每周電腦報;2006年38期

7 熊盛武;王魯;楊婕;;構(gòu)建高性能集群計算機(jī)系統(tǒng)的關(guān)鍵技術(shù)[J];微計算機(jī)信息;2006年03期

8 楊曉寧,伍衛(wèi)國,劉愛華,董小社,胡雷鈞;多負(fù)載均衡器集群系統(tǒng)中負(fù)載均衡器故障恢復(fù)機(jī)制[J];計算機(jī)工程;2004年23期

9 鄭躍鵬,曾新平,蔡勁宏;近年遙感技術(shù)新進(jìn)展及幾點思考[J];礦產(chǎn)與地質(zhì);2004年03期

10 申紅芳,羅四維,趙宏;集群計算系統(tǒng)的模型結(jié)構(gòu)[J];計算機(jī)應(yīng)用研究;2004年02期

相關(guān)碩士學(xué)位論文 前9條

1 陳亮;集群負(fù)載均衡關(guān)鍵技術(shù)研究[D];中南大學(xué);2009年

2 汪筱紅;高可用多節(jié)點集群技術(shù)的研究與實現(xiàn)[D];上海交通大學(xué);2006年

3 馮慧玲;實時集群計算機(jī)體系結(jié)構(gòu)的研究[D];西安科技大學(xué);2006年

4 張毅;專用計算集群組環(huán)境中作業(yè)管理調(diào)度系統(tǒng)的設(shè)計與實現(xiàn)[D];國防科學(xué)技術(shù)大學(xué);2005年

5 李福堂;基于EOS/MODIS的森林火災(zāi)監(jiān)測模型及應(yīng)用研究[D];華中科技大學(xué);2005年

6 安喜鋒;高性能計算集群管理系統(tǒng)與作業(yè)調(diào)度技術(shù)研究與實現(xiàn)[D];西北工業(yè)大學(xué);2005年

7 邵強(qiáng);對象存儲文件系統(tǒng)中元數(shù)據(jù)管理集群關(guān)鍵技術(shù)研究與實現(xiàn)[D];國防科學(xué)技術(shù)大學(xué);2005年

8 梁蘇彬;基于集群的高可用HTTP服務(wù)器[D];四川大學(xué);2004年

9 黃翊;集群環(huán)境下TCP連接遷移的研究與實現(xiàn)[D];國防科學(xué)技術(shù)大學(xué);2002年

，

本文編號：2051996