Conference Agenda

Nuclear energy provides about 70% of France’s electricity, with 56 pressurized water reactors (PWRs) operated by Electricité de France (EDF). EDF R&D uses advanced fluid mechanics to ensure reactor safety, employing in-house 3D codes like neptune_cfd to study two-phase flows and critical phenomena such as the boiling crisis.

This article focuses on phase change at the interface between liquid and vapour, often referred to as bulk or interfacial condensation/boiling. Although overshadowed by wall-driven condensation/boiling, interfacial phase change is crucial in nuclear applications, particularly in liquid metal flows for sodium-cooled reactors and microchannels used in the nuclear industry.

While single-fluid Volume of Fluid (VOF) codes effectively model interfacial phase change, multi-field computational fluid dynamics (CFD) codes lag behind. This article introduces a new phase change model for multi-field codes, using the gradient method to capture interfacial phase change accurately. The model is validated against both analytical and experimental cases involving bulk boiling, showing excellent agreement. Its mesh convergence aligns with single-fluid codes, and we propose a hybrid approach combining this model’s accuracy with the computational efficiency of dispersed-phase methods for simulating complex two-phase flows.

A novel analytical solution-based thermal resistance network computational method has been proposed to provide a more accurate and reasonable temperature calculation framework for pebble bed reactors. This method generates positional and contact information for large-scale pebble beds based on the analytical solution of multidimensional generalized thermal resistance and results from the discrete element method. It also calculates the generalized thermal resistance between the center points of adjacent particle contact surfaces. The generated data is trained using decision tree and random forest algorithms, constructing multiple weak classifiers (i.e., individual decision trees) and combining them into a strong classifier to reduce overfitting and enhance the model's generalization capability. A random forest model was built on the TreeBagger framework, utilizing 100 decision trees and 3 leaf nodes. The importance of each feature's impact on thermal resistance was analyzed, and the trained values effectively reflected the thermal resistance values, achieving a maximum percentage error of 1.45% in the testing set. Validation was conducted using simple cubic, body-centered cubic, and face-centered cubic packing arrangements, showing good agreement with finite volume method results. The novel thermal resistance network model was applied to compute the temperature field caused by heat conduction in the HTTU, confirming the model's feasibility and providing the temperature distribution at various nodes.

The accumulation of Corrosion-Related Unidentified Deposit (CRUD) in the core of the Pressurized Water Reactor (PWR) poses potential threats to reactor safety. This study investigates the deposition behavior of CRUD on the PWR fuel cladding surface, constructing a high-precision 3D CRUD dynamic growth model within a complex coupling framework. The Discrete Phase Model (DPM) is employed to analyze the transport and deposition processes of particulate corrosion products within the fuel rod bundle channels. By integrating the Zeta potential and Electrical Double Layer (EDL) model, the study systematically examines how the Zeta potential near the fuel cladding influences particle deposition behavior. Dynamic mesh technology is used to visualize the dynamic growth process of CRUD layer. In parallel, species transport equations are employed to analyze the distribution characteristics of metal ion concentrations and pH within the coolant. Finally, a multi-physics coupling mechanism of CRUD deposition behavior with rods channel inclusion, water chemistry, boiling heat transfer, and flow field distribution is revealed. The results show that Zeta potential and local pH affect the deposition behavior of corrosion product particles. Specifically, when the Zeta potential of the fuel cladding wall is positive, deposition becomes more challenging in acidic conditions but easier in alkaline environments as the Zeta potential increases. Furthermore, the rough surface of the CRUD layer induces localized accelerated flow near the cladding wall, which exacerbates electrochemical corrosion. The complex coupling effects of CRUD layer thickness and temperature field, particle deposition behavior with Zeta potential, local accelerated flow and electrochemical corrosion are revealed.

Flows past a solid wall for a well-known region between the wall and the bulk of the flow, the boundary region. The characteristic thickness of this boundary region is defined by the appropriate diffusion coefficients and is a place of high gradients and non-linear behaviour. In simulations, we prefer to avoid the calculation of the flow properties in the inner boundary region, since it increases the computational cost of the simulation greatly, being a very thin region next to the walls. This is possible, as the inner boundary layer exhibits a self-similar behaviour, that can be described with explicit functions, called wall- functions.

For thermal fields, the temperature gradient at the wall-normal direction determines the heat extracted from the wall, therefore its correct representation will determine the overall temperature field in the domain. It is therefore important we accurately compensate for the effect of the wall on the rest of the flow, if not resolved.

In the proposed paper we examine a turbulent boundary layer with a DNS with multiple temperature fields of various Prandtl numbers to design more accurate thermal wall-functions. The simulations are performed by the incompressible Navier-Stokes solver Nek5000 and restricted to forced convection flows. We will use these simulations to establish highly accurate explicit wall function that depends on the Reynolds and Prandtl number, making it applicable for a wider range of fluid flows.

This paper presents the 3D modelling and CFD analysis of a CANDU irradiated fuel bay (IFB). CANDU IFBs are significantly different from light water reactor spent fuel pools in terms of the bundle type and orientation of the assemblies. Following the Fukushima Daiichi accident, the Canadian Nuclear Safety Commission (Canadian regulator) undertook a vigorous re-evaluation of the current safety measures and margins of the CANDU IFBs under various stages of an extreme beyond design basis accident scenario. Amongst a few phenomena and hypothetical scenarios of interest, advancing the knowledge of the air-cooling effect on the fuel assemblies and storage racks during the complete loss-of-coolant accident (LOCA) was deemed significant. Under the pan-Canadian PIRT effort, the air-cooling effect on the fuel assemblies, which is driven by the natural convection and radiation modes of heat transfer, was identified as an area of high-importance with low knowledge-level phenomenon for IFBs. The objective of this study is to simulate the airflow and temperature distribution around the irradiated fuel racks at various decay power under a postulated accident scenario (complete LOCA) using the CFD code Simcenter STAR‑CCM+. A detailed 3D model based on the geometry of the CANDU fuel storage module was developed that was qualitatively analyzed in CFD for its capabilities to predict the sheath temperature of the irradiated fuel; a key parameter to monitor the severity of the IFB LOCA. It is anticipated that the developed CFD model could be leveraged to inform lower-fidelity codes and to guide experiments to develop validation data.