Development and validation of a nuclear data adjustment module based on sensitivity analysis
-
摘要: 随着中子学计算方法的发展和精确建模能力的提高,核反应堆物理计算程序中模型近似和离散方法带来的误差逐渐减小,而核数据因其测量难度高,成为影响计算精度的关键输入参数。因此,基于自主研发的敏感性和不确定性分析平台SUPES,开发了基于敏感性分析和广义线性最小二乘算法的核数据调整模块。首先,由敏感性分析获取响应关于输入参数的变化规律;其次,通过相似性分析筛选中子学层面上相似程度高的实验装置参与核数据调整;最后,采用广义线性最小二乘算法使得计算值与实测值之间的误差最小,获得核数据调整量。根据临界基准题HEU-MET-FAST-078中的22个算例,对ACE格式连续能量数据库进行调整,数值结果表明,有效增殖因子keff的均方根误差从3.10×10−3降低到1.53×10−3。通过数值结果对比分析,验证了所开发的核数据调整模块的正确性。Abstract:
Background With the development of neutron calculation methods and improved modeling capabilities, the errors introduced by model approximations and discretization methods in nuclear reactor physics calculations have gradually decreased. However, nuclear data, due to the challenges in measurement, have become the key input parameter affecting computational accuracy.Purpose In this study, a nuclear data adjustment module based on sensitivity analysis and the generalized linear least squares algorithm was developed within the self-developed sensitivity and uncertainty analysis platform, SUPES.Methods First, sensitivity analysis was used to determine the relationship between system responses and input parameter variations. Next, similarity analysis was applied to select experimental setups with high similarity at the neutron physics level. Finally, the generalized linear least squares algorithm was employed to minimize the error between computed and measured values, resulting in nuclear data adjustments.Results The adjustment of the ACE format continuous energy database was performed on 22 cases from the critical benchmark HEU-MET-FAST-078. The numerical results show that the root mean square error of the effective multiplication factor (keff) was reduced from 3.10×10−3 to 1.53×10−3.Conclusions The comparison and analysis verified the correctness of the developed nuclear data adjustment module.-
Key words:
- sensitivity analysis /
- similarity analysis /
- nuclear data adjustment /
- ACE database
-
表 1 HEU-MET-FAST-078基准题的材料和尺寸信息
Table 1. Material and dimension information of HEU-MET-FAST-078 benchmark cases
case 235U mass
fraction/%fuel density/
(g/cm3)diameter
D/cmfuel height
Hc/cmreflector
materialtop thickness
Ht/cmbottom thickness
Hb/cmreflector density/
(g/cm3)#1 93.3 17.9 38.1 5.7562 water 15.24 0.0 0.9982 #3 93.3 17.9 38.1 6.7331 polyethylene 2.54 0.0 0.925 #5 93.3 17.9 38.1 5.9663 polyethylene 5.08 0.0 0.925 #7 93.3 17.9 38.1 3.8392 polyethylene 5.08 5.08 0.925 #9 93.3 17.9 38.1 5.7404 polyethylene 7.62 0.0 0.925 #11 93.3 17.9 38.1 5.6984 polyethylene 10.16 0.0 0.925 #13 93.3 17.9 38.1 5.7089 polyethylene 15.24 0.0 0.925 #15 93.3 17.9 38.1 5.6984 polyethylene 20.32 0.0 0.925 #17 93.3 17.9 38.1 5.6984 polyethylene 25.40 0.0 0.925 #19 93.3 17.9 38.1 5.5881 Lucite 15.24 0.0 1.18 #21 93.3 17.9 38.1 5.7352 paraffin 15.24 0.0 0.87 #23 93.3 17.9 38.1 7.1165 paraffin 2.54 0.0 1.79 #25 93.3 17.9 38.1 6.5860 paraffin 5.08 0.0 1.79 #27 93.3 17.9 38.1 6.0346 paraffin 15.24 0.0 1.70 #29 93.3 17.9 38.1 3.9600 paraffin 15.24 15.24 1.70 #31 93.3 17.9 38.1 5.9820 paraffin 17.78 0.0 1.71 #33 93.3 17.9 38.1 3.8340 paraffin 17.78 17.78 1.70 #35 93.3 17.9 38.1 5.9453 paraffin 20.32 0.0 1.72 #37 93.3 17.9 38.1 5.9558 paraffin 30.48 0.0 1.70 #39 93.3 17.9 38.1 5.9505 paraffin 35.56 0.0 1.71 #41 93.3 17.9 38.1 8.1564 none / / / #43* 93.3 17.9 38.1 6.7751 polyethylene 15.24 0.0 0.925 Note:*case #43 has a 0.0381 cm thick cadmium sheet between the polyethylene reflector and the fuel cylinder.表 3 核数据调整前后keff计算值与基准值的误差
Table 3. Error of keff calculation values before and after nuclear data adjustment from benchmark values
case benchmark
keff (±1σ)ENDF/B-VII.1 ACE
MCNP result (±1σ)error before
adjustment/10−5after adjustment
MCNP result (±1σ)error after
adjustment/10−5#1 0.9995 ±0.0018 0.99460 ±0.00007 −490 0.99799 ±0.00007 −151 #3 0.9994 ±0.0022 0.99597 ±0.00007 −343 0.99781 ±0.00007 −159 #5 0.9991 ±0.0019 0.99619 ±0.00007 −291 0.99864 ±0.00007 −46 #7 1.0000 ±0.0019 0.99857 ±0.00008 −143 1.00147 ±0.00008 147 #9 0.9997 ±0.0022 0.99580 ±0.00007 −390 0.99815 ±0.00007 −155 #11 0.9995 ±0.0015 0.99582 ±0.00007 −368 0.99845 ±0.00007 −105 #13 1.0000 ±0.0017 0.99735 ±0.00007 −265 0.99996 ±0.00008 −4 #15 0.9991 ±0.0018 0.99670 ±0.00007 −240 0.99925 ±0.00007 15 #17 0.9995 ±0.0018 0.99668 ±0.00007 −282 0.99928 ±0.00007 −22 #19 0.9995 ±0.0016 0.99560 ±0.00008 −390 0.99842 ±0.00007 −108 #21 0.9995 ±0.0020 0.99624 ±0.00007 −326 0.99885 ±0.00008 −65 #23 0.9992 ±0.0022 0.99804 ±0.00006 −116 0.99947 ±0.00006 27 #25 0.9992 ±0.0025 0.99739 ±0.00007 −181 0.99896 ±0.00007 −24 #27 0.9992 ±0.0021 0.99597 ±0.00007 −323 0.99733 ±0.00006 −187 #29 1.0000 ±0.0025 1.00235 ±0.00007 235 1.00292 ±0.00007 292 #31 0.9994 ±0.0020 0.99513 ±0.00007 −427 0.99668 ±0.00007 −272 #33 0.9996 ±0.0026 0.99591 ±0.00007 −369 0.99677 ±0.00007 −283 #35 0.9991 ±0.0022 0.99452 ±0.00007 −458 0.99592 ±0.00007 −318 #37 0.9986 ±0.0021 0.99624 ±0.00007 −236 0.99777 ±0.00006 −83 #39 0.9989 ±0.0021 0.99676 ±0.00007 −214 0.99835 ±0.00007 −55 #41 0.9992 ±0.0025 0.99678 ±0.00006 −242 0.99860 ±0.00006 −60 #43 1.0000 ±0.0019 0.99895 ±0.00007 −105 1.00044 ±0.00008 44 表 2 敏感性分析包含的核素反应道
Table 2. Nuclide reaction channels included in the sensitivity analysis
reaction symbol instruction involved nuclides σelas elastic scattering reaction 1H 12C 16O 234U 235U 238U σinel inelastic scattering reaction 234U 235U 238U σf fission reaction 234U 235U 238U σγ capture reaction 1H 12C 16O 234U 235U 238U νp prompt fission neutron yield 234U 235U 238U χp prompt fission neutron spectrum 235U 表 4 核数据调整前后有效增殖因子keff不确定度的变化
Table 4. The change in keff uncertainty of eigenvalues before and after nuclear data adjustment
case benchmark
keff (±1σ)ENDF/B-VII.1 ACE
MCNP result (±1σ)before adjustment
keff uncertainty/10−5after adjustment
keff uncertainty/10−5#1 0.9995 ±0.0018 0.99460 ±0.00007 1130 482 #3 0.9994 ±0.0022 0.99597 ±0.00007 1145 516 #5 0.9991 ±0.0019 0.99619 ±0.00007 1135 478 #7 1.0000 ±0.0019 0.99857 ±0.00008 1321 462 #9 0.9997 ±0.0022 0.99580 ±0.00007 1128 464 #11 0.9995 ±0.0015 0.99582 ±0.00007 1119 460 #13 1.0000 ±0.0017 0.99735 ±0.00007 1123 464 #15 0.9991 ±0.0018 0.99670 ±0.00007 1126 471 #17 0.9995 ±0.0018 0.99668 ±0.00007 1117 463 #19 0.9995 ±0.0016 0.99560 ±0.00008 1136 471 #21 0.9995 ±0.0020 0.99624 ±0.00007 1117 462 #23 0.9992 ±0.0022 0.99804 ±0.00006 1039 531 #25 0.9992 ±0.0025 0.99739 ±0.00007 1098 533 #27 0.9992 ±0.0021 0.99597 ±0.00007 1199 539 #29 1.0000 ±0.0025 1.00235 ±0.00007 1538 525 #31 0.9994 ±0.0020 0.99513 ±0.00007 1204 534 #33 0.9996 ±0.0026 0.99591 ±0.00007 1574 536 #35 0.9991 ±0.0022 0.99452 ±0.00007 1206 532 #37 0.9986 ±0.0021 0.99624 ±0.00007 1211 534 #39 0.9989 ±0.0021 0.99676 ±0.00007 1212 539 #41 0.9992 ±0.0025 0.99678 ±0.00006 994 567 #43 1.0000 ±0.0019 0.99895 ±0.00007 1141 504 -
[1] 曹良志, 邹晓阳, 刘宙宇, 等. 高保真数值核反应堆不确定度量化方法研究进展[J]. 核动力工程, 2021, 42(2): 1-15 doi: 10.13832/j.jnpe.2021.02.0001Cao Liangzhi, Zou Xiaoyang, Liu Zhouyu, et al. Research progresses of uncertainty quantification methods for high fidelity numerical nuclear reactor[J]. Nuclear Power Engineering, 2021, 42(2): 1-15 doi: 10.13832/j.jnpe.2021.02.0001 [2] 阳文俊, 胡赟. 核数据调整方法研究现状与发展[J]. 科技创新与应用, 2023, 13(27): 24-28 doi: 10.19981/j.CN23-1581/G3.2023.27.006Yang Wenjun, Hu Yun. Research status and development of nuclear data adjustment methods[J]. Technology Innovation and Application, 2023, 13(27): 24-28 doi: 10.19981/j.CN23-1581/G3.2023.27.006 [3] Salvatores M, Palmiotti G, Ishikawa M, et al. Assessment of existing nuclear data adjustment methodologies[R]. NEA/NSC/WPEC/DOC(2010)429, 2011. [4] Ishikawa M, Sugino K, Sato W, et al. Development of a unified cross-section set ADJ2000 based on adjustment technique for fast reactor analysis[J]. Journal of Nuclear Science and Technology, 2002, 39(s2): 1073-1076. doi: 10.1080/00223131.2002.10875287 [5] 吴海成. 基于核数据调整基准例题的235U协方差数据测试[J]. 原子能科学技术, 2020, 54(11): 1985-1990 doi: 10.7538/yzk.2020.youxian.0279Wu Haicheng. Test of 235U covariance data with nuclear data adjustment benchmark[J]. Atomic Energy Science and Technology, 2020, 54(11): 1985-1990 doi: 10.7538/yzk.2020.youxian.0279 [6] Wu Qu, Peng Xingjie, Yu Yingrui, et al. Nuclear cross-section adjustment based on the 2-D/1-D transport code KYADJ[J]. Annals of Nuclear Energy, 2020, 137: 107068. doi: 10.1016/j.anucene.2019.107068 [7] 吴屈. 基于一步法确定论程序的核数据敏感性与不确定度分析[D]. 北京: 清华大学, 2019Wu Qu. Sensitivity and uncertainty analysis for nuclear data based on one-step determistic code[D]. Beijing: Tsinghua University, 2019 [8] 朱帅涛. 快能谱反应堆多群核数据的调整与精度评估方法研究[D]. 北京: 华北电力大学(北京), 2020Zhu Shuaitao. Research on multigroup nuclear data adjustment and accuracy assessment method for fast spectrum reactor[D]. Beijing: North China Electric Power University (Beijing), 2020 [9] 万承辉. 核反应堆物理计算敏感性和不确定性分析及其在程序确认中的应用[D]. 西安: 西安交通大学, 2018Wan Chenghui. Researches on sensitivity and uncertainty analysis for reactor physics computation and its application in code validation[D]. Xi'an: Xi'an Jiaotong University, 2018 [10] 乔梁. 钠冷快堆中子输运并行算法及堆芯物理参数的不确定性分析与核数据调整研究[D]. 西安: 西安交通大学, 2021Qiao Liang. Neutron transport parallel calculation method in sodium-cooled fast reactor applied with uncertainty analysis and nuclear data adjustment research[D]. Xi'an: Xi'an Jiaotong University, 2021 [11] 黄金龙. 基于连续能量核数据的反应堆物理敏感性与不确定性分析及数据同化方法研究[D]. 西安: 西安交通大学, 2024Huang Jinlong. Researches on the sensitivity and uncertainty analysis and data assimilation methos of reactor physics based on continuous energy nuclear data[D]. Xi'an: Xi'an Jiaotong University, 2024 [12] Reed R, Schira N. Heu metal cylinders, partially reflected by water, polyethylene, lucite, and paraffin[R]. OECD Nuclear Energy Agency, 2006. [13] Wieselquist W, Zhu T, Vasiliev A, et al. PSI methodologies for nuclear data uncertainty propagation with CASMO-5M and MCNPX: results for OECD/NEA UAM benchmark phase I[J]. Science and Technology of Nuclear Installations, 2013, 2013: 549793. doi: 10.1155/2013/549793 [14] Pusa M. Adjoint-based sensitivity and uncertainty analysis of lattice physics calculations with CASMO-4[C]//Proceedings of the International Conference on the Physics of Reactors. 2014. -
下载: