Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Uesawa, Shinichiro; Hiramatsu, Natsuki*; Ono, Koji*; Yoshida, Hiroyuki
Dai-53-Kai Kashika Joho Shimpojiumu Koen Rombunshu (Internet), 3 Pages, 2026/08
To realize the early practical application of innovative reactors, it is essential to utilize numerical simulations as alternatives to large-scale mock-up tests. To clarify the validity, we are developing measurement techniques capable of capturing instantaneous and local gas-liquid interface information. In this study, a measurement technique based on deep learning is being developed to obtain instantaneous and local bubble characteristics, such as bubble diameter, bubble velocity, and aspect ratio, for validating detailed two-phase flow simulations. This presentation introduces visualization results in a rod bundle flow channel for acquiring bubble characteristics. To enable visualization of the entire flow channel, a 4
4 rod bundle test section was fabricated using perfluoroalkoxy (PFA) tubes with a refractive index close to that of water as simulated fuel rods. Bubble behavior was captured using two high-speed video cameras arranged orthogonally. The three-dimensional distribution of bubbles was reconstructed from the images obtained from the two viewing directions. In addition, void fraction measurements were conducted under the same conditions using a wire-mesh sensor, and the validity and applicability of the present experimental method were examined through comparisons with the visualization measurement results.
Hiramatsu, Natsuki*; Ono, Koji*; Uesawa, Shinichiro; Yoshida, Hiroyuki
Dai-53-Kai Kashika Joho Shimpojiumu Koen Rombunshu (Internet), 3 Pages, 2026/08
For the early practical deployment of innovative reactors, the use of detailed two-phase flow analysis is being considered as an alternative or complementary approach to large-scale mock-up experiments. In this study, we are developing a technique to calculate bubble centroid positions, diameters, and velocities and to reconstruct the three-dimensional bubble distribution by applying image segmentation, bounding-box-based tracking, and bubble matching between images acquired from two different directions, in order to validate the detailed two-phase flow analysis. This presentation reports the current status of the development of this technique and evaluates the impact of using not only experimental data but also numerical simulation data as training data on the detection accuracy.
Uesawa, Shinichiro; Maeshima, Takahiro*; Okada, Makoto*; Tomita, Hirobumi*; Tate, Naofumi*; Aoki, Kunitomo*; Hiramatsu, Natsuki*; Ono, Koji*; Yoshida, Hiroyuki
Konsoryu Shimpojiumu 2026 Koen Rombunshu (Internet), 2 Pages, 2026/08
To use detailed two-phase flow simulations as an alternative or complement to large-scale mockup experiments, quantitative validation of instantaneous and local gas-liquid interfacial information is required in addition to flow-regime reproducibility. However, conventional wire-mesh sensors (WMSs) cannot measure void fraction distributions at a spatial resolution finer than the wire-crossing interval. To overcome this limitation, we are developing a technique to estimate high accurate, high-resolution void fraction distributions from WMS signals by combining WMS measurements, detailed two-phase flow simulations, electrostatic field analysis, and generative AI. This presentation reports on the development status of a Pix2Pix-based model and its application to experimentally obtain WMS signals in a 33 rod-bundle channel.
Kawabe, Tomosaburo*; Sano, Yoshihiko*; Kuwahara, Fujio*; Uesawa, Shinichiro; Yoshida, Hiroyuki
Dai-63-Kai Nihon Dennetsu Shimpojiumu Koen Rombunshu (Internet), 1 Pages, 2026/05
To estimate the thermal behavior of fuel debris inside the Primary Containment Vessels (PCVs) of TEPCO's Fukushima Daiichi Nuclear Power Station, a simulation method has been developed using the JUPITER code with porous medium models. However, it has been found that the selection of models and parameters according to the internal structure of the porous media significantly affects the accuracy of simulation results, highlighting the need for appropriate model selection strategies. In this study, we focus on packed beds with particles as one form of porous media and investigate a methodology for determining model parameters that characterize their flow behavior. Systematic knowledge of packed beds with heterogeneous particle size ratios and packing structures remains limited, and comprehensive organization accounting for structural differences has not yet been fully established. Therefore, numerical simulations were employed to evaluate the permeability and Forchheimer coefficient of packed beds with different particle size ratios and packing structures. Based on the obtained results, a method for appropriately determining these parameters according to the structural characteristics of packed beds is proposed, and insights are provided that contribute to improving the accuracy of flow analysis models in porous media.
Kamiya, Tomohiro; Kondo, Ryoichi; Fukuda, Takanari; Fukuda, Kodai; Tada, Kenichi; Ono, Ayako; Nagaya, Yasunobu; Yoshida, Hiroyuki
JAEA-Data/Code 2025-021, 28 Pages, 2026/03
JAEA-Data-Code-2025-021.pdf:1.39MB
Japan Atomic Energy Agency has developed a high-fidelity multi-physics platform JAMPAN for connecting single-physics codes such as a neutronics code and a thermal-hydraulics code. It consists of the HDF5 formatted data container and input/output data handler modules to generate the input file and read the output file of the single-physics codes. Users can easily add or exchange the code by implementing input and output data handler modules for this code. JAMPAN is equipped with interfaces compatible with the neutronics code MVP and the thermal-hydraulics codes JUPITER, ACE-3D, and NASCA, enabling neutronics and thermal-hydraulics coupling calculations to provide reference solutions for core analysis codes. Users can select the thermal-hydraulics code depending on the required calculation accuracy. In addition, the fuel rod properties can be calculated using FEMAXI. This report explains the overview of JAMPAN.
Uesawa, Shinichiro; Ono, Ayako; Yoshida, Hiroyuki
Haikan Gijutsu, 68(4, 増刊号), p.52 - 56, 2026/03
This paper introduces a new measurement technique for visualizing the three-dimensional distribution of bubbles in a complex channel such as a nuclear reactor fuel assembly. Bubbly flow is important in many engineering fields, and especially in nuclear engineering, where bubble behavior significantly affects the performance and safety of nuclear reactors, and thus requires detailed understanding. Conventional rule-based image recognition has difficulty identifying bubbles overlapping in the line-of-sight direction, but in this study, deep learning (Mask R-CNN and Swin Transformer) is used to achieve highly accurate bubble detection with a small amount of training data. Furthermore, the tracking technique using ByteTrack made it possible to track many bubbles with complex motions, and by combining images taken from different viewpoints using two high-speed cameras and reconstructing the 3D shape of the bubbles using the ellipsoid approximation, 3D instantaneous local information such as bubble position, diameter, and velocity was obtained. To eliminate the effects of refraction and obstruction of vision by structures in the channel, a simulated fuel rod was made of a transparent material (PFA tube) with a refractive index similar to that of water, enabling distortion-free imaging and measurement even in channels with complex structures. This enabled 3D visualization of bubble behavior in complex channels, which had been difficult to achieve in the past. Since this technology enables highly accurate 3D visualization with a small number of cameras and a small amount of learning, it is expected to be applied to objects other than bubbles.
Fukuda, Takanari; Yamashita, Susumu; Yoshida, Hiroyuki
Journal of Computational Physics, 545, p.114485_1 - 114485_32, 2026/01
Times Cited Count：1 Percentile：0.00(Computer Science, Interdisciplinary Applications)This paper puts forward a novel approach for the evaluation of the geometrical fidelity and the interface sharpness of the VOF advection schemes separately and quantitatively. This new evaluation has elucidated the trade-off relationship of the geometrical fidelity and the interface sharpness between the existing schemes of the original THINC and the THINC/WLIC. By investigating and resolving this trade-off relationship, we have developed a novel THINC-based scheme that exhibits high performance with regard to both geometrical fidelity and interface sharpness, despite employing an algorithm as concise as those of the original THINC and the THINC/WLIC. The novel scheme, designated "THINC/Advanced WLIC (THINC/AWLIC)," has been developed by redefining the weight function of the preceding THINC/WLIC so that the contribution of the first-order upwind flux can be variably blended with the usage of the control parameter. The results of the multiple benchmark tests in two and three dimensions demonstrate that both the geometrical fidelity and the interface sharpness are significantly enhanced if the control parameter is appropriately determined. Furthermore, the associated error of THINC/AWLIC is comparable to that of the geometrical scheme, although the implementation complexity is unchanged from that of the simple THICN/WLIC.
Yamashita, Susumu; Yoshida, Hiroyuki
Journal of Nuclear Science and Technology, 19 Pages, 2026/00
Times Cited Count：0 Percentile：0.00(Nuclear Science & Technology)This study investigates the applicability of the mechanistic CFD code JUPITER to three-dimensional melt relocation phenomena in nuclear reactor cores during severe accidents. The XR2-1 BWR metallic melt relocation experiment was analyzed as an integral-effect test case focusing exclusively on melt relocation behavior under an inert atmosphere. A detailed three-dimensional model of the XR2-1 test section, including fuel assemblies, control blades, and lower core support structures, was constructed, and time-dependent injections of SS/B4C and Zircaloy melts were simulated under experimentally based thermal conditions. The simulation results were evaluated through qualitative comparisons of melt relocation paths and quantitative comparisons of relocated material volumes in key regions of the test section. The analysis successfully reproduced the three experimentally observed melt relocation paths - through the control blade guide tube, nosepiece and inlet nozzle, and along the channel box region - without the formation of internal blockages. Quantitative comparisons showed reasonable agreement with post-test X-ray tomographic measurements for most evaluation regions. These results demonstrate that JUPITER can realistically capture three-dimensional melt relocation behavior in complex core geometries and indicate its potential usefulness as a complementary tool to large-scale experiments for evaluating severe accident melt relocation phenomena. This study identifies major problems in analyzing fuel assembly melt relocation behavior using the current JUPITER code, including the excessive computational cost of radiative heat - transfer calculation, insufficient computational grid resolution, and the absence of fluid-structure interaction modeling.
Uesawa, Shinichiro; Yoshida, Hiroyuki
Journal of Nuclear Science and Technology, 24 Pages, 2026/00
Times Cited Count：0In void-fraction measurements using conductance-type wire-mesh sensors (WMSs) in rod-bundle flows, the measurement accuracy is governed by the non-uniform electric current density formed near the electrodes. This non-uniformity causes the instantaneous WMS signal to be strongly affected by the bubble position and shape, thereby introducing systematic errors specific to rod-bundle geometries. However, these errors have not been sufficiently evaluated. In this study, three-dimensional electrostatic simulations were performed to clarify the mechanisms and magnitudes of these errors. The current density distributions and WMS signal responses were analyzed for the subchannel, inter-subchannel, and corner-subchannel regions while varying the bubble position, bubble shape, and transmitter-receiver layer distance. The results demonstrated that the spatial characteristics of the current-density distribution differ among the three channel types. Consequently, the correlation between the WMS signal and void fraction was not unique, and the channel-dependent variations in the correlation were confirmed to arise from differences in bubble position, bubble shape, and layer distance. By consolidating these correlations, representative void-fraction conversion formulas incorporating quantitative systematic errors were developed. The findings of this study provide a quantitative basis for improving the reliability of WMS measurements and for evaluating experimental uncertainties in the validation of two-phase flow CFD simulation codes.
Ono, Ayako; Okawa, Tomio*; Yoshida, Hiroyuki
Journal of Nuclear Science and Technology, 62(12), p.1231 - 1239, 2025/12
Times Cited Count：0 Percentile：0.00(Nuclear Science & Technology)The development of a precise and reliable prediction method for a departure from nucleate boiling (DNB) is urgently needed to design and develop new-generation reactors that enable us to establish a carbon-neutral society. In this study, we consider that the formation of a large vapor mass on the heating surface is the primary trigger of the DNB. Therefore, we embark on the development of a new model for vapor mass formation under forced convective boiling. We assume that the primary bubbles coalesced with each other, which are generated from the nucleation sites, and the coalesced bubble is formed. The nucleation sites are assumed to be distributed on the basis of the Poisson distribution. The large vapor mass is assumed to be formed by the merging of coalesced bubbles when the diameter of the coalesced bubble satisfies the criteria of the slug formation by Mishima. The analysis using the experimental data showed that the proposed model predicted the heat flux to form the large vapor mass well.
Fukuda, Takanari; Yamashita, Susumu; Yoshida, Hiroyuki
Journal of Nuclear Science and Technology, 62(12), p.1264 - 1278, 2025/12
Times Cited Count：2 Percentile：41.98(Nuclear Science & Technology)This study compared three interface capturing schemes (ICSs) for multi-phase flow simulations based on the VOF method, focusing on bubble volume conservation. The THINC/WLIC scheme showed significant VOF diffusion and underestimated total bubble volume, while the original THINC and PLIC conserved bubble volumes. Moreover, an analysis of THINC/WLIC based on a new visualization approach revealed that VOF fragments were ripped off by shear forces around interface, making it unsuitable for accurate void fraction prediction in boiling water reactors. The original THINC may be a viable alternative to PLIC due to its simplicity.
Zhang, H.*; Umehara, Yutaro*; Horiguchi, Naoki; Yoshida, Hiroyuki; Eto, Atsuro*; Mori, Shoji*
Energy, 335, p.138090_1 - 138090_18, 2025/10
Times Cited Count：0 Percentile：0.00(Thermodynamics)Nuclear power is a key low-carbon energy source for a carbon-neutral future. In boiling water reactors (BWRs), steam-water annular flow near fuel rods is crucial for reactor safety, but its high-temperature, high-pressure conditions (285
C, 7 MPa) make direct measurement challenges. To address this, we used an HFC134a-ethanol system at lower conditions (40C, 0.7 MPa) to simulate BWR annular flow. Using a high-speed camera and the constant electric current method, we analyzed liquid-film characteristics, wave velocity and frequency. We also examined surface tension and interfacial shear stress effects. Furthermore, we proposed a new correlation for base film thickness.
Masaki, Naoto*; Kaneko, Akiko*; Horiguchi, Naoki; Yoshida, Hiroyuki
Extended Abstracts of the 11th JSME-KSME Thermal and Fluid Engineering Conference (J-K TFEC11) (Internet), 6 Pages, 2025/10
In the event of a core meltdown in a nuclear reactor, molten fuel has a possibility to enter a coolant pool in a jet-like manner. Jet atomization promoting heat transfer between molten fuel and coolant is crucial for severe accident management. Previous studies of jet atomization have focused on interfacial shear stress acting on a jet because it contributes to the growth and deformation of interfacial waves and ultimately promote droplet formation. However, it is extremely difficult to observe interfacial wave evolution on a cylindrical jet and quantify interfacial shear stress acting on an interface. Then, the effect of interfacial shear stress acting on a jet is not fully clarified. This study aims to experimentally clarify the effect of interfacial shear stress on jet atomization. We conducted experiments in a liquid-liquid system using the 3D-LIF method and PIV. Herein, a glycerol-water solution was used as a jet phase and silicone oil as a continuous phase to match their refractive indices. A jet entered vertically downward through a 3 mm nozzle into a pool. The injection velocity was 2.6 m/s, and the pool depth was 30 mm. To use the 3D-LIF method and PIV with refractive index matching, Rhodamine 6G and tracer particles were used. As a result, waves on a jet interface were observed to have two modes. The larger waves appeared in a two-dimensional (axisymmetric) form at the interface more than 15 mm away from the nozzle, and three-dimensional smaller waves (wavelength 
= 1 mm) were found to form on the surface of these larger waves. Furthermore, the interfacial shear stress acting on the smaller waves was found to be greater than that acting on the larger waves. Droplets were also observed to form easily from the smaller waves. From these results, we experimentally concluded that high interfacial shear stress acting on small waves promotes droplet formation.
Sano, Yoshihiko*; Ota, Kensuke*; Kuwahara, Fujio*; Uesawa, Shinichiro; Yoshida, Hiroyuki
Nihon Kikai Gakkai Netsu Kogaku Konfuarensu 2025 Koen Rombunshu, 1 Pages, 2025/10
To estimate the thermal behavior of fuel debris inside the Primary Containment Vessels (PCVs) of TEPCO's Fukushima Daiichi nuclear power station, a simulation method has been developed using the JUPITER code with porous medium models. However, it has been found that the selection of models and parameters according to the internal structure of the porous media significantly affects the accuracy of simulation results, highlighting the need for appropriate model selection strategies. In this study, we investigated methods for calculating macroscopic model constants that characterize flow behavior in porous media. These properties are influenced not only by the volume ratio of solid and gas phases but also by the structural features of the porous media. Focusing on packed beds composed of particles with varying diameters, we conducted numerical simulations to evaluate permeability and Forchheimer coefficient across diverse structural configurations. Based on the results, we propose a method for appropriately determining these parameters according to the structure of the packed bed, thereby contributing to the improvement of the porous media heat transfer and flow model.
Kamiya, Tomohiro; Yoshida, Hiroyuki
Physics of Fluids, 37(10), p.103359_1 - 103359_23, 2025/10
Times Cited Count：0 Percentile：0.00(Mechanics)In this study, we developed a conservative scheme based on a volume of fluid (VOF) and a ghost fluid method for liquid-gas two-phase compressible fluid simulations. We treated several one- and two-dimensional numerical problems to investigate the capability and applicability of the proposed method for compressible two-phase fluid simulations. The results agree well with the exact solutions or the numerical results of previous studies. Furthermore, the results also show that the proposed method can almost completely ensure the conservation property. Consequently, we concluded that the proposed method could simulate compressible two-phase flows and conserve mass, momentum, and total energy.
Uesawa, Shinichiro; Ono, Ayako; Yoshida, Hiroyuki
Gazo Rabo, p.1 - 5, 2025/08
This paper introduces a new measurement technique for visualizing the three-dimensional distribution of bubbles in a complex channel such as a nuclear reactor fuel assembly. Bubbly flow is important in many engineering fields, and especially in nuclear engineering, where bubble behavior significantly affects the performance and safety of nuclear reactors, and thus requires detailed understanding. Conventional rule-based image recognition has difficulty identifying bubbles overlapping in the line-of-sight direction, but in this study, deep learning (Mask R-CNN and Swin Transformer) is used to achieve highly accurate bubble detection with a small amount of training data. Furthermore, the tracking technique using ByteTrack made it possible to track many bubbles with complex motions, and by combining images taken from different viewpoints using two high-speed cameras and reconstructing the 3D shape of the bubbles using the ellipsoid approximation, 3D instantaneous local information such as bubble position, diameter, and velocity was obtained. To eliminate the effects of refraction and obstruction of vision by structures in the channel, a simulated fuel rod was made of a transparent material (PFA tube) with a refractive index similar to that of water, enabling distortion-free imaging and measurement even in channels with complex structures. This enabled 3D visualization of bubble behavior in complex channels, which had been difficult to achieve in the past. Since this technology enables highly accurate 3D visualization with a small number of cameras and a small amount of learning, it is expected to be applied to objects other than bubbles.
Kamiya, Tomohiro; Nagatake, Taku; Ono, Ayako; Tada, Kenichi; Kondo, Ryoichi; Nagaya, Yasunobu; Yoshida, Hiroyuki
Mechanical Engineering Journal (Internet), 12(4), p.24-00461_1 - 24-00461_9, 2025/08
JAEA has developed the JAEA Advanced Multi-Physics Analysis platform for Nuclear systems (JAMPAN) to realize high-fidelity neutronics/thermal-hydraulics coupling simulations. We performed a neutronics/thermal-hydraulics coupling simulation for a single BWR fuel assembly in order to confirm that the MVP/JUPITER coupling through JAMPAN is feasible. As a result, we confirmed that the void fraction and the corresponding change in the heat generation distribution are reasonable qualitatively.
Yuki, Kohei*; Horiguchi, Naoki; Yoshida, Hiroyuki; Yuki, Kazuhisa*
Mechanical Engineering Journal (Internet), 12(4), p.24-00451_1 - 24-00451_8, 2025/08
Fuel debris at the Fukushima Daiichi nuclear power station is typically cooled under immersion. However, an unexpected significant drop in water level results in coolant contact with high-temperature fuel debris having porous structure. In such scenarios, rapid cooling is essential, yet the thermal behavior at the liquid-solid interface, including capillary phenomena, is not well understood. This paper presents basic research evaluating the evaporation characteristics of droplets upon contact with metallic porous media featuring small pores under 1 mm. We conducted experiments using bronze or stainless steel porous media with pore diameters of 1, 40, or 100 
m to derive lifetime curves for droplets. Our findings indicate that Leidenfrost effect is mitigated on porous surfaces as the vapor can escape through the pores. Moreover, in bronze porous media, as the temperature increases, oxide film with a fine structure facilitates capillary action. In contrast, the low wettability of stainless steel porous media prevents capillary action, inhibiting droplet absorption and dispersion into the pores. Consequently, rapid cooling via the capillary action is unlikely if the fuel debris shares similar characteristics with steel porous media. Therefore, for risk management, the cooling system should be established assuming that capillary force does not act in the fuel debris.
Zhang, H.*; Umehara, Yutaro*; Horiguchi, Naoki; Yoshida, Hiroyuki; Mori, Shoji*
Proceedings of 12th International Conference of Fluid Flow, Heat and Mass Transfer (FFHMT 2025), P. 222_1, 2025/07
In the past decades, the liquid film characteristics of annular flow, such as film thickness and wave height, have been studied extensively. However, most of experimental data and analyses available in previous studies are limited to the air-water annular flows under near-atmospheric conditions. Thus, the variation in surface tension and gas-liquid density ratio is also limited, and their effects on the characteristics of annular flow are not well-understood. The objective of this study is to clarify the effect of surface tension and gas-liquid density ratio on the wave height and interfacial shear stress in annular flows. We conducted experiments for gas-liquid annular flows and measured the time-varying liquid film thickness. Using water and ethanol as working liquids, whereas nitrogen and HFC134a gas as working gases, we varied the surface tension range from 30.7 to 67.4 mN/m and density ratio range from 27 to 434. The accuracy of sensors for the measurement using the conductance probe method is 
5%. By post-processing for the measurement data, liquid film thicknesses (base, average, and maximum film thickness) and wave height were estimated. As the experimental results, the data of both the maximum film thickness and the height of disturbance wave converged on a single curve when they plotted for the interfacial shear stress. The reason was considered that the shape of the disturbance wave is decided by the balance between the surface tension and the interfacial shear stress. With additional analysis based on a potential method, a direct relationship between the height of disturbance wave and interfacial shear stress was clarified. These findings enhance the understanding of annular flow dynamics and contribute to improving the two-phase flow modelling.
Uesawa, Shinichiro; Yamashita, Susumu; Sano, Yoshihiko*; Yoshida, Hiroyuki
Journal of Nuclear Science and Technology, 62(6), p.523 - 541, 2025/06
Times Cited Count：0 Percentile：0.00(Nuclear Science & Technology)Japan Atomic Energy Agency (JAEA) has developed a numerical method with the JUPITER code with a porous medium model to calculate the thermal behavior in PCVs of 1F. In this study, we performed an experiment and numerical simulation of the natural convective heat transfer with the porous medium to validate JUPITER with the porous medium model. In comparison of the temperature and velocity distributions between the experiment and simulation, the temperature distribution in the simulation was in good agreement with the distribution in the experiment except the temperature near the top surface of the porous medium. The velocity distribution also agreed qualitatively with the experimental result. In addition, we also performed the numerical simulations with various effective thermal conductivity models to discuss the effect of the conductivity based on the internal structure of porous media on the natural convective heat transfer. The result indicated that the temperature distribution in the porous medium and the velocity distribution of the natural convection were significantly different for each model, and thus the conductivity of the fuel debris was one of the key parameters of in the thermal behavior analysis in 1F.
