Conference Agenda

^{The safety analysis of nuclear reactor containment is crucial for maintaining the integrity of nuclear power plants during accident scenarios that threaten containment integrity. The Fukushima-Daiichi incident underscored the importance of containment as the ultimate barrier against the release of radioactive materials into the environment. During a loss of coolant accident (LOCA), the release of coolant from the reactor coolant system (RCS) elevates the temperature and pressure within the containment. Investigating these parameters is vital for ensuring the containment wall’s integrity. In this study, an Unsteady Reynolds-Averaged Navier-Stokes (URANS) simulation was conducted to examine the steam-air mixing behavior inside the containment during a LOCA. The steam injection nozzle is located in the lower part of the steam generator SG-2 compartment. The instantaneous temperature profiles of the steam-air mixture, predicted by various turbulence models, were validated against experimental data at different locations within the containment. The numerical predictions showed good agreement with the experimental temperature profiles. Additionally, the impact of LOCA steam injection on the compartment and containment walls was investigated. The numerical investigation revealed a significant impact of steam injection in the SG-2 compartment and the lower section of the containment. Furthermore, steam was found to be uniformly stratified in the upper section of the containment, exhibiting a comparatively lesser impact from the steam injection during the early transient phase."}

The reflood phase of a loss-of-coolant accident in a nuclear power plant is crucial for safety, as it determines the peak cladding temperature. A high accuracy in the prediction of this parameter by thermal-hydraulic system codes like RELAP5 and TRACE is essential to ensure compliance with regulatory safety limits. Experimental programs, such as the Rod Bundle Heat Transfer (RBHT) project, provide benchmark data for evaluating and improving these models.
This paper assesses the ability of TRACE to simulate the reflood phase, focusing on postcritical heat flux phenomena in the upper rod bundle regions. A TRACE model was developed, incorporating key features like the spacer grid model and a droplet field. The results were compared with experimental data and previous RELAP5 simulations. Additionally, a sensitivity analysis examined which parameters have a stronger influence on the PCT and the quench time.
The findings indicate that TRACE performs well in high reflood rate scenarios but struggles with low or variable flooding rates, particularly in capturing heat transfer mode transitions. The sensitivity analysis highlighted key factors affecting simulation accuracy, such as the quench temperature and the heat transfer coefficient during dispersed flow film boiling. Despite some limitations, TRACE demonstrated potential in modeling certain phenomena, including heat transfer enhancement from spacer grids. Further improvements could enhance its predictive capabilities, increasing confidence in its use for safety assessments of nuclear reactor reflood scenarios.

Sodium-cooled fast-neutron reactors are currently considered to be the most mature type of reactor able to closing the fuel cycle. In France and throughout the world, pool-type reactors are selected to build generation IV power plants. In a sodium-cooled pool-type reactor, thermal stratification can occur in the pools in several cases. This phenomenon is monitored closely because it can impact the behaviour of the reactor and might lead to thermal fatigue. In the 1980s, the SUPERCAVNA test facility was operated at the CEA Grenoble research centre. The experimental campaigns investigated the onset of thermal stratification in a rectangular pool. During transient tests, cold sodium was injected in a hot sodium pool. Depending on the inlet flow velocity, thermal stratification would form and erode the hot sodium layer in the pool. The data from these tests constitute a set of CFD-grade experiment that are very useful to assess the capability of CFD codes to capture the relevant phenomena. Code_Saturne was selected to perform calculations of three transient tests from the SUPERCAVNA experimental campaign. Two tests were well captured. A mixed convection test proved more difficult to predict and lead to extensive tests of turbulence and turbulent heat flux models. In this paper the SUPERCAVNA facility and the tests of interest are presented. Then, the CFD model of the facility is described and the results are presented and discussed. Conclusions and recommendations for this type of flows are proposed.

Nuclear thermal hydraulics codes, such as TRACE, frequently exhibit a lack of thorough documentation regarding their input parameters, particularly for crucial elements like heat transfer coefficients, which are often determined through expert judgment and empirical correlations integrated within the code itself. To enhance the reliability and precision of simulations conducted with these codes, it is vital to systematically assess the uncertainties tied to these parameters. TRACE (version 5.0 Patch 8) supports this by enabling users to adjust 43 physical model parameters in the input script using multipliers, starting with a default value of 1. In this investigation, we conducted a tube condensation test at KAIST's Passive Containment Cooling System Facility using TRACE, followed by a detailed Sensitivity Analysis (SA) aimed at identifying and ranking the parameters that significantly affect a key output variable—the saturated steam temperature at the center—critical for Inverse Uncertainty Quantification (IUQ). We developed a robust mathematical framework for Maximum Likelihood Estimation (MLE) based on the Expectation Maximization algorithm and applied it to the tube condensation test. The sensitivity analysis identified several parameters that influence the code’s predictions of the saturated steam temperature, with the vapor-to-interface and liquid-to-wall heat transfer coefficients being the most impactful. Following this, we implemented inverse uncertainty quantification to assess statistical properties like mean and variance for these key parameters. The MLE approach provided reliable estimates of their probability density functions, significantly enhancing our understanding of the uncertainties involved in TRACE simulations.

In the frame of the European McSAFER project, experimental and numerical investigations for the safety evaluation of water-cooled SMRs such as NuScale, SMART, CAREM and F-SMR were performed. At LUT, the MOTEL facility was designed based on the NuScale geometry with buoyancy driven primary circuit flow including. Experiments with asymmetric heated core for crossflow studies were challenging for numerical simulations because of flow instability and correct prediction of mass flow and pressure loss. At KIT, detailed CFD models for the entire vessel with all components of the primary circuit were developed. The best suited CFD model version was resolving all primary loop components like the core heater rods with spacer grids and the helical coiled heat exchanger tubes of the steam generator in detail. Therefore, more than 2*10⁸ cells were used. A detailed analysis of the simulations and experimental data demonstrated the necessity that also even parts of the SG secondary circuit containing two-phase flow has taken into account in order to obtain full agreement with temperature measurements. Furthermore, several CFD models with simplifications such as modelling the SG´s by a porous media or the consideration of the full resolved core region as a standalone part with specified inlet and outlet conditions were created. The deviations between experimental data and the various model simulations clearly demonstrates the disadvantages of model simplifications and justifies the numerical costs of a detailed full vessel CFD model, which provided very good predictions.

In sodium cooled fast reactors (SFRs), transient thermal stratification of the sodium coolant within reactor components must be thoroughly understood for safety analysis and licensing efforts. Computational fluid dynamics (CFD) modeling can be leveraged to simulate the thermal hydraulics within these reactors and study the transient thermal stratification of their coolant media. However, before they can be used for safety analysis and licensure, these models must be experimentally validated to ensure their results are consistent with physical observation. The Gallium Thermal-hydraulic Experiment (GaTE) studied thermal stratification within SFR upper plena, utilizing liquid gallium as a surrogate fluid for the liquid sodium. Cold-shock flow injection tests conducted with GaTE provide an experimental benchmark for validation of CFD in capturing thermal stratification within SFRs. Large-eddy simulation (LES) of the cold-shock tests of GaTE facility was conducted with Nek5000. The velocity response along the plenum height during the isothermal stage prior to the shock is validated against the GaTE experimental benchmark. Then two mixed convection regimes were simulated, one with more dominant effects of forced convection and one with more dominant effects of natural convection. The axial temperature profiles within the plenum during the thermal transient are then compared to those collected with GaTE. When validated, these LES models can be used to augment and extend the current understanding of thermal stratification within SFR plena by exploring a broad range of convection regimes. These validated data can be used to develop reduced-order models and investigate underlying turbulent mechanisms of transient thermal stratification.