Conference Agenda

Valve leakage failures are a common equipment problem in nuclear power plants. Such failures not only lead to economic losses, but also cause radioactive leaks in serious cases, resulting in major safety hazards. Currently, experimental data on valve leakage is very scarce, and the diagnosis of many types of valves is very complex, so it is necessary to supplement experimental data on valve leakage and select the best diagnostic algorithm.

This paper describes a novel experimental system designed to simulate Reactor Coolant Capacity Control System (RCV) upper loading line valve leakage faults, to verify the reasonableness of the experimental setup by analysing the experimental data and to explore the impact of leakage on the system performance. Within the realm of algorithmic analysis, this study evaluates the diagnostic efficacy of various algorithms, including logistic regression, random forest, and support vector machine (SVM) models. The empirical findings of the study reveal that the Random Forest algorithm exhibits the most superior diagnostic precision, achieving a remarkable accuracy of 99.82% in the detection of valve leakage incidents. Algorithms such as Support Vector Machines (SVMs) and Simple Bayes have unsatisfactory performance metrics, with diagnostic accuracies not exceeding the 95% threshold. Therefore, the latter algorithms are considered less suitable for diagnosing valve leakage faults.

This research not only enriches the experimental data but also offers a valuable reference for the selection of appropriate diagnostic algorithms. The in-depth investigation of valve leakage faults can serve as a robust safeguard for the secure operation of nuclear power plants.

One of state-of-the-art pressurized light-water small modular reactors, an innovative small modular reactor (i-SMR), is being developed in Republic of Korea. A steel containment vessel (CV) is adopted not only to prevent release of radioactivity material but also to reduce pressure and temperature of the reactor module. As a newly suggested passive safety system (PSS), a passive emergency core cooling system (PECCS) prevents water level reduction of reactor coolant system (RCS) using natural circulation without additional injection of coolant. The emergency depressurization valve (EDV) and emergency recirculation valve (ERV) which are installed on the wall of reactor vessel (RV) play as the natural circulation flow paths between RV and CV. The pressurized steam from the RV through the EDV is condensed in the CV by heat transfer on heat exchanger of passive containment cooling system (PCCS). The condensed water recirculates to the RV through the ERV. The level of condensed water is important physical variable because the difference of water levels between CV and RV determines recirculation flow rate.
Integral effect test facility is under basic design to conduct thermal-hydraulic validation tests to support the process of standard design approval (SDA) for i-SMR. The basic design of PCCS heat exchanger considering condensation heat transfer was completed and the scaling distortion in integral effect test facility against the prototype was evaluated. Four trains of PCCS heat exchanger with thermocouples to measure fluid and outer wall temperatures will be installed in the integral effect test facility.

The integrated full natural circulation SMR has a high degree of integration, high intrinsic safety, flexible arrangement and can be applied in various scenarios. However, integrated full natural circulation SMR also incorporates some innovative designs, such as the elimination of the Main Circulation Pumps (MCPs), the adoption of helical-coiled tube heat exchanger, the application of the passive safety design, and so on.

Due to these new characteristics of the integrated full natural circulation SMR, the existing system analysis code cannot be directly applied to the integrated full natural circulation SMRs, and the corresponding system analysis code still needs to be developed. RELAP5 code is a widely used system analysis code internationally and has been successfully applied to some SMRs. The RELAP5 code has been improved or added the relevant models that are required for the analysis of integrated full natural circulation SMR by authors.

The improved RELAP5 code is validated by an Integral Effect Test conducted by SNERDI. The comparison results of natural circulation tests at various power levels are shown in this paper. The results show that the improved RELAP5 code compares well with the test results, with a temperature difference of about 5 degrees. Additional test cases will be performed and further validation and evaluation will be conducted in the future.

The recent revision of Emergency Core Cooling System (ECCS) acceptance criteria, which incorporates Design Extension Conditions (DECs) and addresses high-burnup fuel safety concerns, has intensified the need for more accurate loss-of-coolant accident (LOCA) analyses. Previously, thermo-mechanical and thermal-hydraulic behaviors were evaluated separately, resulting in conservative estimates that limited insights into actual coupled phenomena. Implementing an integrated multi-physics approach now enables simultaneous characterization of these behaviors, leading to more realistic analyses and enhanced safety margins. In response to this need, the Korea Atomic Energy Research Institute (KAERI) developed the ICARUS facility to simulate fuel cladding behavior from the post-blowdown stage through the reflood phase of a large-break LOCA (LBLOCA). A Zircaloy-4 cladding and heater assembly, combined with controlled boundary conditions, replicates the thermo-mechanical and thermal-hydraulic environment of a reflood scenario. Real-time measurements of cladding surface temperature, deformation, subchannel fluid temperature, and water level are carried out. By varying heater power, internal cladding pressure, and the reflood initiation time, this study systematically evaluates the coupled phenomena, thereby offering critical insights into the multi-physics behavior of nuclear fuel cladding under LBLOCA conditions. By integrating thermo-mechanical and thermal-hydraulic analyses, this work moves beyond conservative assumptions and provides a more realistic understanding of cladding behavior under accident conditions.

The Molten Salt Reactor Experiment (MSRE) was a key nuclear project in the 1960s that demonstrated the viability of molten salt as a coolant and fuel. In molten salt reactors, understanding complex thermohydraulic behavior is essential for optimizing performance and safety. However, building a full-scale experimental model is often impractical due to high costs and the reactor’s large size. A scaled-down model provides an efficient and cost-effective approach to studying critical aspects of fluid flow, heat transfer, and pressure distribution while capturing key physical phenomena. This study presents the design and computational fluid dynamics (CFD) validation of a 1/5 scaled-down mock-up of the MSRE. The scaled-down model was developed to replicate the geometry of the original MSRE while maintaining fluid behavior, such as Reynolds number. The study also provides a detailed explanation of the scaling laws used to ensure that the down-scaled model accurately reflects the behavior of the full-scale system.

Given the reduced size, the model cannot replicate every detail, such as all the surfaces and channels exposed to molten salt within the reactor core, making validation crucial. CFD simulations were performed using the scaled model to analyze fluid flow and pressure characteristics. The results of the simulations were compared to experimental data from the original full-scale MSRE. This comparison confirmed the accuracy of the scaled mock-up and its reliability for predicting the thermohydraulic behavior of molten salt reactors, making it a valuable tool for further research.

Differentially-heated cavity natural convection is an important phenomenon relevant to the design of thermal energy storage systems, concentrated solar power receivers, building-integrated photovoltaic systems, and nuclear reactor passive safety systems. Particle Image Velocimetry (PIV) measurements are implemented to study the natural convection behavior of molten nitrate salt in a differentially-heated cavity for Rayleigh numbers up to 10⁹ and Prandtl numbers from 22 to 30. A low melting point salt mixture, NaNO₃-KNO₃-LiNO₃-CaNO₃, is selected as a working fluid to provide operating temperatures from 100°C to 500°C. The experimental methodology for PIV measurements in a heated molten salt test section with a transparent optical window is presented along with preliminary test data. 2-D planar PIV measurements of the flow field in molten nitrate salt are compared to measurements of the flow field in water, with matching Rayleigh numbers and Prandtl numbers from 2 to 7. Quantification of measurement uncertainties is described and compared to alternative flow measurement techniques.