Research on Efficient Solution of Neutron Physics Equations Using NAS-Optimized PINN

Yu Caiyang; Jiang Yong; Chen Qilong; Liu Dong; Lyu Jiancheng

doi:10.13832/j.jnpe.2024.090041

Volume 46 Issue 2

Apr. 2025

Turn off MathJax

Article Contents

Article Navigation > Nuclear Power Engineering > 2025 > 46(2): 119-126

Yu Caiyang, Jiang Yong, Chen Qilong, Liu Dong, Lyu Jiancheng. Research on Efficient Solution of Neutron Physics Equations Using NAS-Optimized PINN[J]. Nuclear Power Engineering, 2025, 46(2): 119-126. doi: 10.13832/j.jnpe.2024.090041

Citation:

Yu Caiyang, Jiang Yong, Chen Qilong, Liu Dong, Lyu Jiancheng. Research on Efficient Solution of Neutron Physics Equations Using NAS-Optimized PINN[J]. Nuclear Power Engineering, 2025, 46(2): 119-126. doi: 10.13832/j.jnpe.2024.090041

Citation:

PDF( 18718 KB)

Research on Efficient Solution of Neutron Physics Equations Using NAS-Optimized PINN

doi: 10.13832/j.jnpe.2024.090041

1.
College of Computer Science, Sichuan University, Chengdu, 610065, China
2.
Nuclear Energy Software and Digital Reactor Engineering Research Center of CNNC, Nuclear Power Institute of China, Chengdu, 610213, China

Received Date: 2024-08-15
Rev Recd Date: 2024-09-10

Available Online: 2025-01-15

Publish Date: 2025-04-02

Abstract

Abstract

To quickly and accurately solve the two kinds of equations of neutron diffusion and transport in the core, Physics-Informed Neural Networks (PINN) can be utilized to enhance the speed and efficiency of solving partial differential equation. However, the predefined structure of PINNs is relatively inflexible, limiting their width and depth in practical applications. This study proposes an innovative approach for determining the optimal PINN structure, (NAS-PINN), which employs a Neural Architecture Search (NAS) strategy to dynamically select the most suitable PINN structure for solving neutron diffusion and transport equations of nuclear reactors. The PINN model identified through this search is applied to equation solving, and the experimental verification comparison is made between the true and predicted values. The results show that the NAS-PINN method has higher accuracy in solving reactor equations with different geometries, and provides a more accurate and efficient solution for complex neutron equations.
- Neutron diffusion equation,
- Neutron transport equation,
- Physics-Informed Neural Network (PINN),
- Neural architecture search (NAS)

FullText(HTML)

References(21)

References

[1]	LISZKA T, ORKISZ J. The finite difference method at arbitrary irregular grids and its application in applied mechanics[J]. Computers & Structures, 1980, 11(1-2): 83-95.
[2]	DOUGLAS JR J, RUSSELL T F. Numerical methods for convection-dominated diffusion problems based on combining the method of characteristics with finite element or finite difference procedures[J]. SIAM Journal on Numerical Analysis, 1982, 19(5): 871-885. doi: 10.1137/0719063
[3]	RAISSI M, PERDIKARIS P, KARNIADAKIS G E. Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J]. Journal of Computational Physics, 2019, 378: 686-707. doi: 10.1016/j.jcp.2018.10.045
[4]	刘振海,张涛,齐飞鹏,等. 基于PINN的燃料棒稳态温度分布快速预测方法研究[J]. 核动力工程,2024, 45(S1): 39-44. doi: 10.13832/j.jnpe.2024.S1.0039.
[5]	杜轲,齐婧,高嘉伟,等. 基于物理信息神经网络(PINN)方法的结构动力响应分析[J/OL]. 工程力学,1-12[2024-10-20]. http://kns.cnki.net/kcms/detail/11.2595.o3.20240822.1525.006.html.
[6]	REN P Z, XIAO Y, CHANG X J, et al. A comprehensive survey of neural architecture search: challenges and solutions[J]. ACM Computing Surveys, 2022, 54(4): 76.
[7]	CHEN X, XIE L X, WU J, et al. Progressive DARTs: bridging the optimization gap for NAS in the wild[J]. International Journal of Computer Vision, 2021, 129(3): 638-655. doi: 10.1007/s11263-020-01396-x
[8]	MIRJALILI S. Genetic algorithm[J]. Studies in Computational Intelligence, 2019, 780: 43–55.
[9]	刘东,罗琦,唐雷,等. 基于PINN深度机器学习技术求解多维中子学扩散方程[J]. 核动力工程,2022, 43(2): 1-8.
[10]	曾付林,张小龙,赵鹏程. 基于hp-VPINN的反应堆中子扩散计算方法研究[J]. 核动力工程,2024, 45(2): 53-62.
[11]	刘东,王雪强,张斌,等. 深度学习方法求解中子输运方程的微分变阶理论[J]. 原子能科学技术,2023, 57(5): 946-959. doi: 10.7538/yzk.2023.youxian.0002
[12]	杜书华. 输运问题的计算机模拟[M]. 长沙: 湖南科学技术出版社,1989: 102-104.
[13]	谢仲生,曹良志,张少泓. 核反应堆物理分析[M]. 第五版. 西安: 西安交通大学出版社,2020: 69-90.
[14]	王伟国,李云召,秦浚玮,等. 基于中子电报方程的均匀裸堆瞬态中子输运问题解析解研究[J]. 核动力工程,2024, 45(2): 42-46.
[15]	JAVAHERIPI M, DE ROSA G H, MUKHERJEE S, et al. LiteTransformerSearch: training-free neural architecture search for efficient language models[C]//Proceedings of the 36th International Conference on Neural Information Processing Systems. New Orleans: Curran Associates Inc. , 2022: 24254-24267.
[16]	程金芮,金瑾,张朝龙,等. 自适应策略优化的粒子群优化算法在神经网络架构搜索中的应用[J]. 计算机应用,2024, 44(S1): 60-64.
[17]	JAAFRA Y, LAUREN J L, DERUYVER A, et al. Reinforcement learning for neural architecture search: a review[J]. Image and Vision Computing, 2019, 89: 57-66. doi: 10.1016/j.imavis.2019.06.005
[18]	YU C Y, WANG Y X, TANG C W, et al. EU-Net: automatic U-Net neural architecture search with differential evolutionary algorithm for medical image segmentation[J]. Computers in Biology and Medicine, 2023, 167: 107579. doi: 10.1016/j.compbiomed.2023.107579
[19]	YE P, LI B P, LI Y K, et al. β-DARTS: beta-decay regularization for differentiable architecture search[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans: IEEE, 2022: 10864-10873.
[20]	XUE Y, CHEN C, SŁOWIK A. Neural architecture search based on a multi-objective evolutionary algorithm with probability stack[J]. IEEE Transactions on Evolutionary Computation, 2023, 27(4): 778-786. doi: 10.1109/TEVC.2023.3252612
[21]	华东师范大学数学系. 数学分析下册[M]. 第三版. 北京: 高等教育出版社,2001: 176-243.