基于权窗的全局减方差方法中“长历史”问题的研究与改进
A solution to “long history” in the method of GVR based on weight window
查看参考文献8篇
文摘
|
在具有全局特性的蒙特卡罗粒子输运计算中,常采用基于权窗的全局减方差方法,由此引起的“长历史”问题使计算效率大大降低。为了提高蒙特卡罗计算效率,减小计算资源的浪费,本文对“长历史”的成因及解决方案进行了研究。通过研究中国聚变工程试验反应堆(Chinese Fusion Engineering Testing Reactor, CFETR)中子输运计算中“长历史”的事件记录文件,提出了“长历史”的成因与粒子连续在轮盘赌中存活有关,针对此成因提出了通过粒子在轮盘赌中存活的次数来限制粒子分裂的解决方法,并通过CFETR模型对这个方法进行了测试。结果显示,次数限取1–4时,并行效率达到80%,无“长历史”出现,品质因子(pFoMG)达到不采用次数限时的约3倍。综上,采用这种方法可以有效克服“长历史”问题,使全局减方差方法的计算效率明显提升。 |
其他语种文摘
|
Background: The “long history” induced by the method of global variance reduction (GVR) based on weight window would reduce the efficiency of the Monte Carlo particle transport calculation with global characteristic. Purpose: To enhance the efficiency and minimize the waste of computing resource, this article proposes a solution to “long history” based on the study of its causes. Methods: According to the event log of “long history” in Chinese fusion engineering testing reactor (CFETR) neutron transport calculation, the cause of the “long history” is brought up, which is in relation to particle's successively surviving in the Russian roulette games. A solution to the “long history” by limiting particle's splitting according to the number of surviving games (weight roulette number, WRN), is came up with, which is tested in CFETR model. Results: The result shows that when the limited number is between 1 and 4, the parallel efficiency reaches 80%, with no “long history” present and global figure of merit (pFoMG) being around 3 times of that in calculation without the limited number. Conclusion: The solution proposed prevented the “long history” and thus greatly enhanced the efficiency of GVR. |
来源
|
核技术
,2017,40(12):120602-1-120602-7 【核心库】
|
DOI
|
10.11889/j.0253-3219.2017.hjs.40.120602
|
关键词
|
全局减方差
;
全局权窗
;
长历史
;
轮盘赌
;
次数限
|
地址
|
1.
中国科学技术大学核科学技术学院, 合肥, 230027
2.
中国科学院等离子体物理研究所, 合肥, 230031
3.
广西防城港核电有限公司, 防城港, 538000
|
语种
|
中文 |
文献类型
|
研究性论文 |
ISSN
|
0253-3219 |
学科
|
原子能技术 |
基金
|
中央高校基本科研业务费专项资金
|
文献收藏号
|
CSCD:6126974
|
参考文献 共
8
共1页
|
1.
X-5 Monte Carlo Team.
A general Monte Carlo N-particle transport code: user's guide. Version 5. Vol. II,2003
|
被引
1
次
|
|
|
|
2.
聂星辰. 蒙特卡罗模拟CFETR中子输运计算中的全局减方差方法应用及对比.
核技术,2016,39(3):030501
|
被引
2
次
|
|
|
|
3.
Van Wijk A J. An easy to implement global variance reduction procedure for MCNP.
Annals of Nuclear Energy,2011,38:2496-2503
|
被引
7
次
|
|
|
|
4.
Davis A. Comparison of global variance reduction techniques for Monte Carlo radiation transport simulations of ITER.
Fusion Engineering and Design,2011,86:2698-2700
|
被引
8
次
|
|
|
|
5.
Naish J. Radiation mapping at JET and ITER using advanced computational acceleration techniques and tools.
Nuclear Technology,2015,192:299-307
|
被引
3
次
|
|
|
|
6.
聂星辰. 深穿透屏蔽计算中MCNP减方差技巧应用及比较.
核电子学与探测技术,2016,36(7):729-733,741
|
被引
3
次
|
|
|
|
7.
Turner A. Improving computational efficiency of Monte-Carlo simulations with variance reduction.
International Conference on Mathematics and Computational Methods Applied to Nuclear Science & Engineering (M&C 2013),2013
|
被引
1
次
|
|
|
|
8.
Wasastjerna F.
Using MCNP for fusion neutronics,2008
|
被引
1
次
|
|
|
|
|