Conference Agenda

A coupled neutronics and thermal-hydraulics simulation code is developed at JAEA. In the coupling simulation code, the 3-dimensional two-fluid ACE-3D code, which is the in-house code of JAEA, will be adopted to simulate thermal-hydraulics behavior inside nuclear reactor fuel assemblies. The ACE-3D code calculates void fraction distributions under operational conditions for use in neutron transport simulations. This research aims to validate ACE-3D using data from a two-phase flow experiment. For this purpose, a two-phase flow experiment was conducted in a 4×4 unheated fuel assembly. In the experiment, the time-averaged void fraction distribution was measured using a wire mesh sensor system under high temperatures (373 K-500 K) and high-pressure conditions of up to 2.6 MPa. The experimental results were analyzed, and the data were visualized to understand better the behavior and characteristics of the two-phase flow in the fuel assembly. A two-phase flow data set is being developed, covering a wide range of experimental conditions, including higher-pressure regions, which can be used for validating thermal-hydraulic codes. Finally, the ACE-3D thermal hydraulics code was applied to the two-phase flow experiment. The calculation results were then compared to the experimental ones, and the issues were identified for improving ACE-3D in future simulations.

CFD is considered to be a valuable tool for assessing and improving the performance and safety of nuclear reactors. Verifying or creating CFD models to predict reactor fluid dynamics is crucial for Gen-IV reactors, which are cooled by liquid metals. The thermal boundary layer in liquid metals is of a greater thickness than that of the momentum layer, leading common turbulence models to incorrect predictions: this highlights not only the necessity for the development of new models but also the creation of databases to validate them. Two main methods may be used to collect the data: perform experimental tests or conduct high-fidelity simulations. This study compares these two methods by assesing the results of a benchmark exercise proposed by the EGTHM. The reference system for the thermo-hydraulic exercise is the Advanced LFR European Demonstrator. Particle Image Velocimetry was employed to obtain high-resolution data on the flow around a triangular lattice.* Computational high-fidelity data are obtained via DNS using an original discretisation technique on a periodic domain of four subchannels. The numerical study also considers heat transfer, by setting a Prandtl number Pr=0.031 representative of LBE. Statistics of velocity, thermal fields and main turbulent flow features are presented and compared with experimental data. This approach allows not only the comparison of experimental and numerical results, but also the integration of one with the other where there are deficiencies, with the aim of providing the optimal dataset for the development of future turbulence and heat transfer models.

*Menezes et al., DOI: https://doi.org/10.1063/5.0154590

Spacer grids in PWR fuel assemblies are equipped with mixing vanes for inducing lateral coolant flow and improving thermal performance. Therefore, understanding how shape of mixing vanes has impacts on flow is important in spacer grid design . This paper presents the results of the PIV (Particle Image Velocimetry) experiment, which was performed on 5x5 rod bundles with spacer grids of two different designs. From the results, the shape effects of mixing vanes were evaluated. Additionally, the experimental results were compared to the results of CFD (Computational Fluid Dynamics) and the applicability of CFD in spacer grid design was investigated.

This study presents a novel application of Lagrangian Particle Tracking (LPT) combined with plenoptic imaging to perform three-dimensional flow measurements within a 6x6 nuclear fuel rod bundle. The experiments were conducted in the Shaking Bundle Facility (SBF), a full-scale experimental model of a nuclear fuel assembly featuring 36 acrylic rods arranged in a square lattice. The assembly replicates the rod-to-rod pitch and mechanical properties of prototypical fuel bundles, with spacer grids providing structural support and simulating realistic flow conditions. Para-cymene, a working fluid with the same refractive index as the surrogate rods, was used to achieve optical transparency and accurate flow measurements.
The experimental setup enables detailed characterization of mean flow fields, boundary layer profiles, and turbulence statistics. Flow data were captured using a plenoptic 3.0 imaging system capable of volumetric measurements in a 14x12x12mm domain, with particles tracked at a frequency of 2 kHz. Advanced Radial Basis Function (RBF) interpolation was applied to reconstruct smooth 3D velocity fields while enforcing incompressibility and boundary conditions.
Results are systematically compared to Direct Numerical Simulations (DNS), Large Eddy Simulations (LES), and prior experimental data, providing a robust benchmark for validating thermal-hydraulic models. This work offers a high-fidelity experimental dataset for improving the predictive accuracy of computational models used in nuclear reactor safety analysis and for understanding flow behaviors under reactor-like conditions.

The depressurization process in light water reactors is an important factor for nuclear safety, and there is a need to develop an analysis code for this transient phenomenon and its validation process. Corroborated experimental data are crucial for evaluating the thermal characteristics of transient boiling and its associated uncertainties. In particular, the spatiotemporal distribution of void fraction during the depressurization process remains undetermined. This study conducted a transient flow boiling experiment during a depressurization process with our test facility for 3D thermal hydraulics in light water reactors (SIRIUS-3D). The test section was a 5×5 rod bundle partially simulating the fuel assembly of an actual reactor. Five units of the subchannel void sensor, capable of measuring the local void fraction between electrodes at a high sampling rate, were installed along the axial direction in the test section’s heated region. We evaluated the multi-dimensional void behavior in a 5×5 rod bundle with a linear depressurization rate ranging from 0.5 to 2.0 MPa/s and an initial system pressure of 7.2 MPa. The rod-surface heat flux and inlet mass flux were set to 70 kW/m² and 750 kg/m²/s, respectively, for all cases. The development of the boiling flow during the depressurization process was summarized with the depressurization rate as a parameter. The void fraction growth rate and time-averaged void fraction were quantified. The spatial void fraction distribution was organized and discussed based on the average values obtained by dividing the regions according to the distance from the center of the bundle cross-section.

The United States has the largest operating fleet of nuclear reactors in the world. Operating cost reduction and power uprates are the two major topics that can further bolster the economic and technical sustainability of LWRs. Over the past decade, a large amount of two-phase flow and heat transfer data, especially for the post-CHF regime, has been collected through the NRC/PSU RBHT reflood test facility. The present work summarizes the on-going efforts in large-scale RBHT data compilation and the assessment of heat transfer modeling accuracy for the dispersed flow film boiling and the transition boiling regime under elevated pressure conditions (up to 60 psi). Various existing post-CHF heat transfer models were compared with the rigorously compiled large-scale experimental data sets, and their performances were evaluated. It was found that existing models carry with them large prediction uncertainties, which further leads to an overly conservative safety limit for existing LWRs. Therefore, it is recommended to develop new post-CHF heat transfer models with significantly reduced thermal hydraulic uncertainties that originate from advanced two-phase flow diagnostic and processing data techniques as well as more in-depth understanding of the underlying physics involved. The large-scale dataset is also found to be extremely useful for developing physic-integrated data-driven models that provide superior prediction accuracy with physically realistic projections.