广谱系组件计算程序开发

张晋超; 张乾; 赵强; 邹航; 于嘉蕾; 武世伏; 陈莹

doi:10.13832/j.jnpe.2024.05.0019

广谱系组件计算程序开发

doi: 10.13832/j.jnpe.2024.05.0019

张晋超^1,,
张乾²,
赵强^1, ,,
邹航¹,
于嘉蕾¹,
武世伏¹,
陈莹³

1.
哈尔滨工程大学核科学与技术学院，哈尔滨，150001
2.
浙江大学物理学院，杭州，310030
3.
中国原子能科学研究院核数据重点实验室，北京，102413

基金项目: 国家自然科学基金（12105063）；中核集团“青年英才”项目

详细信息

作者简介:
张晋超（1996—），男，博士研究生，现主要从事反应堆物理研究，E-mail: 13935397912@hrbeu.edu.cn

通讯作者:
赵　强，E-mail: zhaoqiang@hrbeu.edu.cn

中图分类号: TL329
计量
- 文章访问数: 36
- HTML全文浏览量: 16
- PDF下载量: 16
- 被引次数: 0
出版历程
- 收稿日期: 2023-11-08
- 修回日期: 2024-02-07
- 刊出日期: 2024-10-14

Development of Computing Code for Full Spectrum Assembly

1.
College of Nuclear Science and Technology, Harbin Engineering University, Harbin, 150001, China
2.
School of Physics, Zhejiang University, Hangzhou, 310030, China
3.
Key Laboratory of Nuclear Data, China Institute of Atomic Energy, Beijing, 102413, China

摘要

摘要: 为解决先进组件设计中存在的多样几何设计问题和使用慢化材料引入的复杂能谱问题，以进一步提升组件程序设计能力，本文基于非结构网格，设计了基于2164群结构的子群共振计算与特征线输运计算的组合策略，并完成了程序开发。程序应用了高效多核素共振干涉方法、散射源移位算法和千群级别的图形处理器（GPU）特征线并行方案以保证计算效率。对不同能谱、不同几何特征的先进组件设计的验证结果表明：与蒙特卡洛基准解相比，对于快谱组件问题，特征值偏差均低于72pcm（1pcm=10⁻⁵）；对于含慢化材料的快谱组件问题，特征值偏差均低于132pcm。因此，本文设计的计算方案能够处理具有复杂几何和复杂能谱的组件问题。
- 广谱 /
- 组件计算 /
- 共振计算 /
- 特征线方法（MOC） /
- 图形处理器（GPU）
Abstract: In order to address the issues caused by the diverse geometric designs and complex energy spectrum problems introduced by the use of moderator materials in advanced assembly designs, and to further enhance the design capability of the assembly code, this paper designs a combination strategy and develops a corresponding code based on unstructured mesh. The strategy combines the subgroup method, which is based on a group structure including 2164 groups, with the method of characteristic transport. To ensure computational efficiency, the code employs the efficient multi-nuclide resonance interference method, offset algorithm for scattering source and parallel scheme of graphics processing unit (GPU) characteristics at thousand-group level. Advanced assembly designs with different energy spectra and geometric features are selected to validate the proposed method. The results indicate that, compared to the Monte Carlo reference, the deviations in eigenvalue are within 72pcm (1pcm=10⁻⁵) for problems with fast spectra, and within 132pcm for problems utilizing moderators. In conclusion, the calculation scheme proposed in this paper can handle assembly problems characterized by complex geometry and complex energy spectrum.
- Full spectrum /
- Assembly calculation /
- Resonance calculation /
- Method of characteristic (MOC) /
- Graphics Processing Unit (GPU)

HTML全文

图 1 子群参数制作流程

Figure 1. Work Flow for Generation of Subgroup Parameters

下载: 全尺寸图片幻灯片

图 2 不同核素裂变谱对比

Figure 2. Comparision of Fission Spectrum for Different Isotopes

下载: 全尺寸图片幻灯片

图 3 2164群结构

Figure 3. 2164-group Structure

下载: 全尺寸图片幻灯片

图 4 GPU MOC输运扫描流程

Figure 4. MOC Transport Sweep on GPU

下载: 全尺寸图片幻灯片

图 5 程序计算流程

Figure 5. Calculation Scheme of the Code

下载: 全尺寸图片幻灯片

图 6 快谱组件问题

Figure 6. Fast Spectrum Assembly Problems

下载: 全尺寸图片幻灯片

图 7 快谱组件问题注量率对比

Figure 7. Flux Comparison for Fast Assembly Problems

下载: 全尺寸图片幻灯片

图 8 含慢化材料的快谱组件问题

Figure 8. Fast Spectrum Assembly Problems Using Moderator

下载: 全尺寸图片幻灯片

图 9 不同散射阶数下注量率对比

Figure 9. Comparison of Flux with Different Scattering Orders

下载: 全尺寸图片幻灯片

表 1 快谱组件问题特征值对比

Table 1. Eigenvalue Comparison for Fast Spectrum Problems

问题	OpenMC	子群方法	特征值偏差/pcm
常规组件	1.26270	1.26215	−55
半价值组件	1.04132	1.04060	−72

下载: 导出CSV

表 2 含慢化材料的快谱组件问题特征值和EALF对比

Table 2. Comparison of Eigenvalue and EALF for Fast Spectrum Assembly Problems Using Moderator

几何	OpenMC计算特征值	EALF/MeV	子群方法计算特征值				特征值偏差/pcm
几何	OpenMC计算特征值	EALF/MeV	P0	P1	P2	P3	P0	P1	P2	P3
常规组件	1.45504	1.191×10⁻¹	1.45528	1.45521	1.45526	1.45523	24	17	22	19
紧凑组件	1.45181	1.016×10⁻³	1.46121	1.44953	1.45057	1.45049	940	−229	−124	−132
环形组件	1.42411	1.433×10⁻³	1.42728	1.42215	1.42283	1.42283	317	−196	−128	−128
优化组件	1.40462	1.858×10⁻³	1.40527	1.40312	1.40346	1.40344	65	−150	−116	−118

下载: 导出CSV

表 3 不同散射源计算方法下GPU计算时间对比

Table 3. Comparison of GPU Computation Time under Different Scattering Source Calculating Methods

栅元圈数	平源区数	显存/ GB	4090服务器		3090服务器
栅元圈数	平源区数	显存/ GB	原子加技术耗时/s	移位技术耗时/s	原子加技术耗时/s	移位技术耗时/s
3	648	1.95	4.2	3.4	18.3	7.2
4	1272	3.02	6.9	5.3	32.9	12.7
5	2112	4.65	9.5	7.0	50.5	18.1
6	3168	6.70	13.2	9.6	72.4	24.8
7	4440	9.13	17.2	12.4	98.2	32.8
8	5928	12.00	22.0	15.6	128.3	42.3
9	7632	15.20	27.5	19.4	162.6	52.9
10	9552	18.84	33.6	23.6	201.9	64.9

下载: 导出CSV

参考文献(11)

[1]	WEI L F, ZHENG Y Q, DU X N, et al. Development of SARAX code system for full-range spectrum adaptability in advanced reactor analysis[J]. Annals of Nuclear Energy, 2022, 165: 108664. doi: 10.1016/j.anucene.2021.108664
[2]	LEE C, YANG W S. An improved resonance self-shielding method for heterogeneous fast reactor assembly and core calculations[C]//Proceedings of the 2013 International Conference on Mathematics and Computational Methods Applied to Nuclear Science and Engineering. Sun Valley: American Nuclear Society, 2013.
[3]	LEE C, YANG W S. MC²-3: multigroup cross section generation code for fast reactor analysis[J]. Nuclear Science and Engineering, 2017, 187(3): 268-290. doi: 10.1080/00295639.2017.1320893
[4]	JEON B K, YANG W S, JUNG Y S, et al. Extension of MC²-3 for generation of multigroup cross sections in thermal energy range[C]//International Conference on Mathematics and Computational Methods Applied to Nuclear Science and Engineering. Jeju, Korea: American Nuclear Society, 2017.
[5]	MAO L, ZMIJAREVIC I, SANCHEZ R. Resonance self-shielding methods for fast reactor calculations—Comparison of a new tone’s method with the subgroup method in APOLLO3^®[J]. Nuclear Science and Engineering, 2017, 188(1): 15-32. doi: 10.1080/00295639.2017.1332890
[6]	CHOI N, KANG J, JOO H G. Preliminary performance assessment of GPU acceleration module in nTRACER[C]//Transactions of the Korean Nuclear Society Autumn Meeting. Yeosu, Korea: Korean Nuclear Society, 2018.
[7]	MACFARLANE R E, MUIR D W, BOICOURT R M, et al. The NJOY nuclear data processing system, version 2016: LA-UR-17-20093[R]. Los Alamos: Los Alamos National Laboratory, 2017.
[8]	RUGGIERI J M, TOMMASI J, LEBRAT J F, et al. ERANOS 2.1: international code system for GEN IV fast reactor analysis[C]//Proceedings of the International Congress on Advances in Nuclear Power Plants. Reno: American Nuclear Society, 2006.
[9]	ROMANO P K, HORELIK N E, HERMAN B R, et al. OpenMC: a state-of-the-art Monte Carlo code for research and development[J]. Annals of Nuclear Energy, 2015, 82: 90-97. doi: 10.1016/j.anucene.2014.07.048
[10]	FEI T, MOHAMED A, KIM T K. Neutronics benchmark specifications for EBR-II shutdown heat removal test SHRT-45R - revision 1: ANL-ARC-228 Rev. 1[R]. Argonne: Argonne National Laboratory, 2013.
[11]	WANG L J, LU D, YAO L, et al. Study on temperature feedback effect of supercritical CO₂–cooled reactor[J]. Frontiers in Energy Research, 2021, 9: 764906. doi: 10.3389/fenrg.2021.764906