A Machine leaning-based approach to estimate nuclide production cross sections

Iwamoto, Hiroki; Meigo, Shinichiro; Sugihara, Kenta*

JPS Conference Proceedings (Internet), 45, p.011179_1 - 011179_7, 2026/06

https://doi.org/10.7566/JPSCP.45.011179

The nuclide production cross sections are crucial for evaluating the radioactivity of activation products in accelerators and nuclear facilities. Although the production of radionuclides in spallation reactions can be explained using physics models like the nuclear cascade plus evaporation model, accurately and comprehensively reproducing experimental values remains challenging. To address this problem, we have developed a machine learning (ML) model to comprehensively estimate nuclide production cross sections for target materials. The model is trained on experimental data from the nuclear reaction database EXFOR and can estimate nuclide production cross sections even in data-poor regions by utilizing transfer learning. In a previous study, we demonstrated that our model can comprehensively estimate Be ( d) and H ( y) production cross sections for a wide range of targets. In this study, we apply the ML model to nuclides important for accelerator facility design and astrophysics, such as He, Be ( y), Na ( y), and Na ( h), showing that it can comprehensively estimate their production cross sections.

Photoneutron emission process on nuclei around A = 200 for giant dipole resonance energies based on neutron energy and angular distribution

Nguyen, T. H. T.; Sanami, Toshiya*; Yamazaki, Hirohito*; Iwamoto, Nobuyuki; Itoga, Toshiro*; Kirihara, Yoichi; Lee, E.*; Sugihara, Kenta*; Miyamoto, Shuji*; Hashimoto, Satoshi*; et al.

Physics Letters B, 870, p.139900_1 - 139900_8, 2025/11

https://doi.org/10.1016/j.physletb.2025.139900

Overview of PHITS Ver.3.34 with particular focus on track-structure calculation

Ogawa, Tatsuhiko; Hirata, Yuho; Matsuya, Yusuke; Kai, Takeshi; Sato, Tatsuhiko; Iwamoto, Yosuke; Hashimoto, Shintaro; Furuta, Takuya; Abe, Shinichiro; Matsuda, Norihiro; et al.

EPJ Nuclear Sciences & Technologies (Internet), 10, p.13_1 - 13_8, 2024/11

https://doi.org/10.1051/epjn/2024012

The latest updates on PHITS, a versatile radiation transport code, focusing specifically on track-structure models are presented. Track structure calculations are methods used to simulate the movement of charged particles while explicitly considering each atomic reaction. Initially developed for radiation biology, these calculation methods aimed to analyze the radiation-induced damage to DNA and chromosomes. Several track-structure calculation models, including PHITS-ETS, PHITS-ETS for Si, PHITS-KURBUC, ETSART, and ITSART, have been developed and implemented to PHITS. These models allow users to study the behavior of various particles at the nano-scale across a wide range of materials. Furthermore, potential applications of track-structure calculations have also been proposed so far. This collection of track-structure calculation models, which encompasses diverse conditions, opens up new avenues for research in the field of radiation effects.

Bubble flow analysis using multi-phase field method

Sugihara, Kenta; Onodera, Naoyuki; Sitompul, Y.; Idomura, Yasuhiro; Yamashita, Susumu

EPJ Web of Conferences, 302, p.03002_1 - 03002_10, 2024/10

https://doi.org/10.1051/epjconf/202430203002

In simulations of gas-liquid two-phase flows using conventional interface capture methods, we observed that when bubbles come close to each other, they tend to merge numerically, despite experimental evidence indicating that they would repel each other. Given the significant impact of sequential numerical coalescence on flow patterns, it is necessary to regulate the merging behavior of close bubbles. To address this issue, we introduced the Multi- Phase Field (MPF) method, which mitigates bubble coalescence by applying an independent fluid fraction function to each bubble. In this study, we employed the MPF based on the N-phase model to minimize numerical errors associated with surface interactions at triple junction points. Additionally, we implemented the Ordered Active Parameter Tracking (OAPT) method to efficiently store several hundreds of fluid fraction functions. To validate the MPF method, we conducted analysis of turbulent bubbly pipe flows and compared the results against experimental data from Colin et al. The validation results showed reasonable agreements with respect to the bubble distribution and the flow velocity profiles.

Compressible Navier-Stokes formulation for accelerating Poisson solver of gas-liquid two-phase fluid simulations

Onodera, Naoyuki; Sugihara, Kenta; Ina, Takuya; Idomura, Yasuhiro

Keisan Kogaku Koenkai Rombunshu (CD-ROM), 29, 3 Pages, 2024/06

Gas-liquid two-phase flow analysis is one of the most important research topics in nuclear engineering because it is essential for safety evaluation and reactor design. However, it requires large-scale multi-scale simulations, and advanced numerical approaches are needed. To meet this challenge, we have continued to develop the Poisson solver for the multiphase flow analysis code JUPITER. In this study, we aim to improve the convergence of the pressure Poisson solver by formulating the Navier-Stokes equation without using the incompressible approximation. The convergence performance was measured on 8 GPUs for bubbly flow analysis in a circular tube. The results show that the computation time and the number of iterations are reduced by half compared to those using the incompressible approximation, which indicates the usefulness of the formulation in the present study.

Comprehensive estimation of nuclide production cross sections using a phenomenological approach

Iwamoto, Hiroki; Meigo, Shinichiro; Sugihara, Kenta*

Physical Review C, 109(5), p.054610_1 - 054610_12, 2024/05

https://doi.org/10.1103/PhysRevC.109.054610

Nuclide production cross sections are crucial in nuclear research, development, space exploration, and astrophysical investigations. Despite their importance, limited experimental data availability restricts the practicality of phenomenological approaches to comprehensive cross-section estimation. To address this, we propose a Gaussian process-based machine learning (ML) model capable of transferring knowledge from elements with abundant data to those with limited or no experimental data. Our ML model not only enables comprehensive cross-section estimations for various elements but also demonstrates predictive capabilities akin to physics models, even in regions with scarce training data.

Recent improvements of the Particle and Heavy Ion Transport code System; PHITS version 3.33

Sato, Tatsuhiko; Iwamoto, Yosuke; Hashimoto, Shintaro; Ogawa, Tatsuhiko; Furuta, Takuya; Abe, Shinichiro; Kai, Takeshi; Matsuya, Yusuke; Matsuda, Norihiro; Hirata, Yuho; et al.

Journal of Nuclear Science and Technology, 61(1), p.127 - 135, 2024/01

https://doi.org/10.1080/00223131.2023.2275736

The Particle and Heavy Ion Transport code System (PHITS) is a general-purpose Monte Carlo radiation transport code that can simulate the behavior of most particle species with energies up to 1 TeV (per nucleon for ions). Its new version, PHITS3.31, was recently developed and released to the public. In the new version, the compatibility with high-energy nuclear data libraries and the algorithm of the track-structure modes have been improved. In this paper, we summarize the upgraded features of PHITS3.31 with respect to the physics models, utility functions, and application software introduced since the release of PHITS3.02 in 2017.

Neutron-production double-differential cross sections of Pb and Bi in proton-induced reactions near 100 MeV

Iwamoto, Hiroki; Meigo, Shinichiro; Satoh, Daiki; Iwamoto, Yosuke; Ishi, Yoshihiro*; Uesugi, Tomonori*; Yashima, Hiroshi*; Nishio, Katsuhisa; Sugihara, Kenta*; elik, Y.*; et al.

Nuclear Instruments and Methods in Physics Research B, 544, p.165107_1 - 165107_15, 2023/11

https://doi.org/10.1016/j.nimb.2023.165107

The lack of double-differential cross-section (DDX) data for neutron production below the incident proton energy of 200 MeV hinders the validation of spallation models in technical applications, such as research and development of accelerator-driven systems (ADSs). The present study aims to obtain experimental DDX data for ADS spallation target materials in this energy region and identify issues related to the spallation models by comparing them with the analytical predictions. The DDXs for the () reactions of Pb and Bi in the 100-MeV region were measured over an angular range of 30 to 150 using the time-of-flight method. The measurements were conducted at Kyoto University utilizing the FFAG accelerator. The DDXs obtained were compared with calculation results from Monte Carlo-based spallation models and the evaluated nuclear data library, JENDL-5. Comparison between the measured DDX and analytical values based on the spallation models and evaluated nuclear data library indicated that, in general, the CEM03.03 model demonstrated the closest match to the experimental values. Additionally, the comparison highlighted several issues that need to be addressed in order to improve the reproducibility of the proton-induced neutron-production DDX in the 100 MeV region by these spallation models and evaluated nuclear data library.

Optimization of phase field variables in bubble rise analysis

Sugihara, Kenta; Onodera, Naoyuki; Idomura, Yasuhiro; Yamashita, Susumu

Keisan Kogaku Koenkai Rombunshu (CD-ROM), 28, 5 Pages, 2023/06

https://confit-sfs.atlas.jp/customer/jsces28/abstract/D-07-02.pdf

The Modified conservative Allen-Cahn (MCAC) equation, which assumes a spatial distribution for the interface modification strength by the phase field method and weights it using the local velocity distribution and velocity gradient, is proposed. MCAC was applied to bubble rise calculations for four representative shapes, and the results were found to be reasonable in comparison with those of Grace and Annaland et al. In addition, the results obtained by MCAC are equivalent or superior to those of Conservative Allen-Cahn (CAC), and thus MCAC can be considered superior in that it can use appropriate even when large velocity differences occur in a part of the computational domain. Finally, this method was applied to the analysis of a bundle system, and it was confirmed that the gas-liquid interface can be captured even for flows in complex channels.

Measurement of 107-MeV proton-induced double-differential thick target neutron yields for Fe, Pb, and Bi using a fixed-field alternating gradient accelerator at Kyoto University

Iwamoto, Hiroki; Nakano, Keita; Meigo, Shinichiro; Satoh, Daiki; Iwamoto, Yosuke; Sugihara, Kenta; Nishio, Katsuhisa; Ishi, Yoshihiro*; Uesugi, Tomonori*; Kuriyama, Yasutoshi*; et al.

Journal of Nuclear Science and Technology, 60(4), p.435 - 449, 2023/04

https://doi.org/10.1080/00223131.2022.2115423

Double-differential thick target neutron yields (TTNYs) for Fe, Pb, and Bi targets induced by 107-MeV protons were measured using the fixed-field alternating gradient accelerator at Kyoto University for research and development of accelerator-driven systems (ADSs) and fundamental ADS reactor physics research at the Kyoto University Critical Assembly (KUCA). Note that TTNYs were obtained with the time-of-flight method using a neutron detector system comprising eight neutron detectors; each detector has a small NE213 liquid organic scintillator and photomultiplier tube. The TTNYs obtained were compared with calculation results using Monte Carlo-based spallation models (i.e., INCL4.6/GEM, Bertini/GEM, JQMD/GEM, and JQMD/SMM/GEM) and the evaluated high-energy nuclear data library, i.e., JENDL-4.0/HE, implemented in the particle and heavy iontransport code system (PHITS). All models, including JENDL-4.0/HE, failed to predict high-energy peaks at a detector angle of 5. Comparing the energy- and angle-integrated spallation neutron yields at energies of 20 MeV estimated using the measured TTNYs and the PHITS indicated that INCL4.6/GEM would be suitable for the Monte Carlo transport simulation of ADS reactor physics experiments at the KUCA.