Research on Construction Method of a Machine Learning-Based Fuel Rod Temperature Distribution Surrogate Model

Liu Zhenhai; Qi Feipeng; Zhou Yi; Li Yuanming; Li Wenjie; Zeng Wei; Xin Yong; Wang Haoyu; Ma Chao

doi:10.13832/j.jnpe.2023.S2.0001

Volume 44 Issue S2

Dec. 2023

Turn off MathJax

Article Contents

Article Navigation > Nuclear Power Engineering > 2023 > 44(S2): 1-5

Liu Zhenhai, Qi Feipeng, Zhou Yi, Li Yuanming, Li Wenjie, Zeng Wei, Xin Yong, Wang Haoyu, Ma Chao. Research on Construction Method of a Machine Learning-Based Fuel Rod Temperature Distribution Surrogate Model[J]. Nuclear Power Engineering, 2023, 44(S2): 1-5. doi: 10.13832/j.jnpe.2023.S2.0001

Citation:

Liu Zhenhai, Qi Feipeng, Zhou Yi, Li Yuanming, Li Wenjie, Zeng Wei, Xin Yong, Wang Haoyu, Ma Chao. Research on Construction Method of a Machine Learning-Based Fuel Rod Temperature Distribution Surrogate Model[J]. Nuclear Power Engineering, 2023, 44(S2): 1-5. doi: 10.13832/j.jnpe.2023.S2.0001

Citation:

Liu Zhenhai, Qi Feipeng, Zhou Yi, Li Yuanming, Li Wenjie, Zeng Wei, Xin Yong, Wang Haoyu, Ma Chao. Research on Construction Method of a Machine Learning-Based Fuel Rod Temperature Distribution Surrogate Model[J]. Nuclear Power Engineering, 2023, 44(S2): 1-5. doi: 10.13832/j.jnpe.2023.S2.0001

PDF( 2489 KB)

Research on Construction Method of a Machine Learning-Based Fuel Rod Temperature Distribution Surrogate Model

doi: 10.13832/j.jnpe.2023.S2.0001

Science and Technology on Reactor System Design Technology Laboratory, Nuclear Power Institute of China, Chengdu, 610213, China

Received Date: 2023-07-11
Rev Recd Date: 2023-07-22
Publish Date: 2023-12-30

Abstract

Abstract

In order to improve the computational efficiency of large-scale fuel rod performance simulation, the construction method of fuel rod temperature distribution surrogate model (referred to as "surrogate model") is studied by taking fuel rod temperature prediction as an example. The calculation results of the fuel rod performance analysis code COPERNIC are used as the data source, and the representative training data are selected by using the k-means clustering algorithm. Four fully connected feedforward neural networks are trained to predict the outer surface temperature of the cladding considering the effects of coolant flow heat transfer and oxide film growth, the radial temperature distribution of the cladding, the outer surface temperature of the fuel pellet considering the effects of the gap variation between the fuel pellet and cladding, and the radial temperature distribution of the fuel pellet. By combining these neural networks, the temperature distribution of the fuel rod at different times can be quickly predicted based on the input fuel rod power history. Numerical experiments show that the calculation speed of the surrogate model is about 204 times faster than that of the COPERNIC code, and it has high accuracy. The mean deviations of fuel cladding and pellet temperature prediction on the whole data set are about 0.07°C and 0.44°C respectively.
- Neural network,
- Fuel rod temperature,
- Surrogate model,
- Fuel behavior

FullText(HTML)

References(10)

References

[1]	BONNAUD E, BERNARD C, VAN SCHEL E. COPERNIC: a state-of-the-art fuel rod performance code[J]. Transactions of the American Nuclear Society, 1997, 77: 971125.
[2]	王东东,杨红义,王端,等. 中国实验快堆热工参数的自适应BP神经网络预测方法研究[J]. 原子能科学技术,2020, 54(10): 1809-1816.
[3]	洪亮,金鑫,刘虓瀚,等. 机器学习算法在燃料棒温度性能预测中的应用[J]. 深圳大学学报: 理工版,2022, 39(5): 515-520.
[4]	CHE Y F, YURKO J, SEURIN P, et al. Machine learning-assisted surrogate construction for full-core fuel performance analysis[J]. Annals of Nuclear Energy, 2022, 168: 108905. doi: 10.1016/j.anucene.2021.108905
[5]	阎昌琪,曹欣荣. 核反应堆工程[M]. 哈尔滨: 哈尔滨工程大学出版社,2004: 160.
[6]	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
[7]	PASZKE A, GROSS S, MASSA F, et al. PyTorch: an imperative style, high-performance deep learning library[C]//Proceedings of the 33rd International Conference on Neural Information Processing Systems. Vancouver: Curran Associates Inc, 2019.
[8]	LIU D C, NOCEDAL J. On the limited memory BFGS method for large scale optimization[J]. Mathematical Programming, 1989, 45(1-3): 503-528. doi: 10.1007/BF01589116
[9]	SUN L N, GAO H, PAN S W, et al. Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data[J]. Computer Methods in Applied Mechanics and Engineering, 2020, 361: 112732. doi: 10.1016/j.cma.2019.112732
[10]	PEDREGOSA F, VAROQUAUX G, GRAMFORT A, et al. Scikit-learn: machine learning in python[J]. The Journal of Machine Learning Research, 2011, 12: 2825-2830.