Research on PWR Core Refueling Optimization Method Based on Bayesian Optimization
-
摘要: 压水堆堆芯换料优化是核电站安全高效经济运行的关键环节，属于有约束的非线性非凸整数组合优化问题。传统方法计算效率低，容易陷入局部最优解。本文提出了一种基于变分自动编码器、深度度量学习和贝叶斯优化的换料优化方法。该方法利用变分自动编码器将离散的堆芯布置方案映射到连续的隐变量空间；再通过深度度量学习构建结构化的隐空间，使堆芯物理特性相近的样本在隐空间中距离也相近；然后利用多目标贝叶斯优化方法在隐空间中高效地搜索最优解，并通过解码器将最优隐变量解码成对应的堆芯布置方案。基于某M310堆芯首循环初装料数据进行的实验验证表明，该方法能够有效提高换料优化效率和求解质量，获得优于传统方法的布置方案。
-
关键词:
- 堆芯换料优化 /
- 贝叶斯优化 /
- 变分自动编码器 /
- 深度度量学习 /
- NECP-Bamboo
Abstract: Refueling optimization for pressurized water reactor (PWR) cores is crucial for the safe, efficient, and cost-effective operation of nuclear power plants, which is a constrained, nonlinear, non-convex integer combinatorial optimization challenge. Traditional methods often struggle with low computational efficiency and the risk of getting trapped in local optima. This paper presents a refueling optimization approach based on variational autoencoders, deep metric learning, and Bayesian optimization. The method leverages variational autoencoders to map discrete core layout configurations into a continuous latent space. Deep metric learning is then used to construct the latent space such that samples with similar core physical characteristics are positioned closer together. A multi-objective Bayesian optimization is subsequently applied to efficiently search for optimal solutions in this latent space, and a decoder transforms the optimal latent variables back into corresponding core layouts. Experimental validation using the first-cycle initial loading data of an M310 core demonstrates that this method significantly improves refueling optimization efficiency and solution quality, producing better configurations than traditional methods. -
图 1 基于VAE、深度度量学习和贝叶斯优化的换料优化算法流程图
Figure 1. Flowchart of Refueling Optimization Algorithm Based on VAE, Deep Metric Learning, and Bayesian Optimization
表 1 换料优化迭代120次帕累托前沿（VAE+深度度量学习+贝叶斯优化）
Table 1. Pareto Front of Refueling Optimization after 120 Iterations (VAE + Deep Metric Learning + Bayesian Optimization)
堆芯布置方案 keff fxy 方案 1 1.09562 1.346 方案 2 1.09888 1.351 方案 3 1.09897 1.359 方案 4 1.10259 1.360 方案 5 1.11544 1.385 方案 6 1.11977 1.403 方案 7 1.12053 1.413 方案 8 1.12081 1.428 方案 9 1.12124 1.429 方案 10 1.12209 1.446 方案 11 1.12594 1.470 方案 12 1.12986 1.575 方案 13 1.13514 1.772 方案 14 1.13592 1.891 方案 15 1.14056 2.521 方案 16 1.14518 2.524 方案 17 1.14554 3.689 方案 18 1.15449 4.944 方案 19 1.18588 5.802 
下载: 导出CSV
表 2 换料优化迭代120次帕累托前沿（NSGA-II）
Table 2. Pareto Front of Refueling Optimization after 120 Iterations (NSGA-II)
堆芯布置方案 keff fxy 方案 1 1.08931 1.436 方案 2 1.09031 1.437 方案 3 1.09490 1.450 方案 4 1.10177 1.459 方案 5 1.12210 1.488 方案 6 1.13769 1.496 方案 7 1.13885 1.521 方案 8 1.14249 1.567 方案 9 1.14627 1.584 方案 10 1.15437 1.600 方案 11 1.15509 1.711 方案 12 1.16796 1.771 方案 13 1.16863 1.881 方案 14 1.16907 2.334 方案 15 1.17031 2.448 方案 16 1.17437 2.496 方案 17 1.17641 2.584 方案 18 1.17833 3.819
下载: 导出CSV
-
[1] LI Z, WANG J C, DING M. A review on optimization methods for nuclear reactor fuel reloading analysis[J]. Nuclear Engineering and Design, 2022, 397: 111950. doi: 10.1016/j.nucengdes.2022.111950 [2] KIRKPATRICK S, GELATT JR C D, VECCHI M P. Optimization by simulated annealing[J]. Science, 1983, 220(4598): 671-680. doi: 10.1126/science.220.4598.671 [3] HOLLAND J H. Adaptation in natural and artificial systems: an introductory analysis with applications to biology, control, and artificial intelligence[M]. Cambridge: MIT Press, 1992. [4] KENNEDY J, EBERHART R. Particle swarm optimization[C]//Proceedings of ICNN’95-International Conference on Neural Networks. Perth: IEEE, 1995: 1942-1948. [5] ERDOĞAN A, GEÇKINLI M. A PWR reload optimisation code (XCore) using artificial neural networks and genetic algorithms[J]. Annals of Nuclear Energy, 2003, 30(1): 35-53. doi: 10.1016/S0306-4549(02)00041-5 [6] WAN C H, LEI K H, LI Y S. Optimization method of fuel-reloading pattern for PWR based on the improved convolutional neural network and genetic algorithm[J]. Annals of Nuclear Energy, 2022, 171: 109028. doi: 10.1016/j.anucene.2022.109028 [7] LI Z, WANG J C, HUANG J, et al. Development and research of triangle-filter convolution neural network for fuel reloading optimization of block-type HTGRs[J]. Applied Soft Computing, 2022, 136: 110126. [8] JIN Y C. Surrogate-assisted evolutionary computation: recent advances and future challenges[J]. Swarm and Evolutionary Computation, 2011, 1(2): 61-70. doi: 10.1016/j.swevo.2011.05.001 [9] JONES D R, SCHONLAU M, WELCH W J. Efficient global optimization of expensive black-box functions[J]. Journal of Global Optimization, 1998, 13(4): 455-492. doi: 10.1023/A:1008306431147 [10] WANG X L, JIN Y C, SCHMITT S, et al. Recent advances in Bayesian optimization[J]. ACM Computing Surveys, 2023, 55(13s): 287. doi: 10.1145/3582078 [11] SHAHRIARI B, SWERSKY K, WANG Z Y, et al. Taking the human out of the loop: a review of Bayesian optimization[J]. Proceedings of the IEEE, 2016, 104(1): 148-175. doi: 10.1109/JPROC.2015.2494218 [12] SEEGER M. Gaussian processes for machine learning[J]. International Journal of Neural Systems, 2004, 14(2): 69-106. doi: 10.1142/S0129065704001899 [13] MOČKUS J. On Bayesian methods for seeking the extremum[C]//Proceedings of the IFIP Technical Conference on Optimization Techniques. Novosibirsk: Springer, 1974: 400-404. [14] AUER P. Finite-time Analysis of the Multiarmed Bandit Problem[Z]. Kluwer Academic Publishers, 2002. [15] KINGMA D P, WELLING M. Auto-encoding variational Bayes[EB/OL]. arXiv, 2022. [2024-07-01]. http://arxiv.org/abs/1312.6114. [16] KAYA M, BİLGE H Ş. Deep metric learning: a survey[J]. Symmetry, 2019, 11(9): 1066. doi: 10.3390/sym11091066 [17] ISHFAQ H, HOOGI A, RUBIN D. TVAE: triplet-based variational autoencoder using metric learning[C]//Proceedings of the 6th International Conference on Learning Representations. Vancouver: ICLR, 2018. [18] KOGE D, ONO N, HUANG M, et al. Embedding of molecular structure using molecular hypergraph variational autoencoder with metric learning[J]. Molecular Informatics, 2021, 40(2): e2000203. doi: 10.1002/minf.202000203 [19] DAULTON S, ERIKSSON D, BALANDAT M, et al. Multi-objective Bayesian optimization over high-dimensional search spaces[EB/OL]. arXiv, 2022. [2024-04-22]. http://arxiv.org/abs/2109.10964. [20] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc. , 2017: 6000-6010. [21] YU W H, SI C Y, ZHOU P, et al. MetaFormer baselines for vision[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024, 46(2): 896-912. doi: 10.1109/TPAMI.2023.3329173 [22] GROSNIT A, TUTUNOV R, MARAVAL A M, et al. High-dimensional Bayesian optimisation with variational autoencoders and deep metric learning[EB/OL]. arXiv, 2021. [2024-02-26]. http://arxiv.org/abs/2106.03609. [23] MCKAY M D, BECKMAN R J, CONOVER W J. A comparison of three methods for selecting values of input variables in the analysis of output from a computer code[J]. Technometrics, 2000, 42(1): 55-61. doi: 10.1080/00401706.2000.10485979 [24] SRINIVAS N, DEB K. Muiltiobjective optimization using nondominated sorting in genetic algorithms[J]. Evolutionary Computation, 1994, 2(3): 221-248. doi: 10.1162/evco.1994.2.3.221 [25] TRIPP A, DAXBERGER E, HERNÁNDEZ-LOBATO J M. Sample-efficient optimization in the latent space of deep generative models via weighted retraining[C]//Proceedings of the 34th International Conference on Neural Information Processing Systems. Vancouver: Curran Associates Inc. , 2020: 945. [26] 梁毅琳,李云召,周原成,等.基于核电厂实测数据的NECP-Bamboo软件验证与确认[J].核动力工程, 2024, 45(2):24-34.梁毅琳, 李云召, 周原成, 等.基于核电厂实测数据的NECP-Bamboo软件验证与确认[J].核动力工程, 2024, 45(2):24-34. [27] 李倩倩. 基于压水堆换料优化基准问题的随机优化方法的机理及应用研究[D]. 上海: 上海交通大学,2010. [28] DEB K, PRATAP A, AGARWAL S, et al. A fast and elitist multiobjective genetic algorithm: NSGA-II[J]. IEEE Transactions on Evolutionary Computation, 2002, 6(2): 182-197. doi: 10.1109/4235.996017 [29] BLANK J, DEB K. Pymoo: multi-objective optimization in python[J]. IEEE Access, 2020, 8: 89497-89509. doi: 10.1109/ACCESS.2020.2990567 -
图(4) / 表(2)
计量
- 文章访问数: 7
- HTML全文浏览量: 3
- PDF下载量: 0
- 被引次数: 0
