Development of a Novel Finite Volume Methodology for Multi-Dimensional Fuel Performance Applications

Fuel performance codes are an essential tool for ensuring the safe and economic opera-tion of nuclear reactors. Traditionally, these codes have been developed following a simple 1.5-D modelling approach, where the fuel behavior is simplified by assuming axisymmetric and plain strain conditions and where phenomenological correlations are largely privileged over the use of accurate physical and mathematical models. However, in recent decades the advent of modern high-performance computing has awakened a growing interest towards higher-fidelity tools with multi-dimensional and multi-physics capabilities. The main development efforts have been carried out employing the popular finite element method, while the alternative approach offered by the finite volume method has not been explored by the fuel performance community.

Interested in the potential of this numerical scheme for fuel analysis and multi-physics applica-tions, this PhD thesis aims to develop a finite volume methodology and associated software for the high-fidelity analysis of fuel behavior. As a result, a novel finite volume fuel performance code named OFFBEAT is developed. The code is based on the open-source C++ library OpenFOAM and it is envisioned as a multi-dimensional, readily-available and flexible tool, which is straightforward to extend and modify, and open to multi-physics simulations.

The thesis presents the development strategy of OFFBEAT, discussing the numerical framework, the treatment of the gap as well as the structure of the code with its main models and overall so-lution scheme. In this context, a novel contact methodology is developed that introduces a semi-implicit discretization of the contact stresses, improving the convergence properties of many con-tact scenarios. To test the robustness and accuracy of the code, extensive efforts are carried out to verify and validate OFFBEAT against numerical benchmark and experimental data.

The novel methodology is applied to the 3-D analysis of the effect of eccentricity on fuel disc irra-diation test campaigns performed in the past in the Halden Boiling Water Reactor and character-ized by rods with large gap and high conductive fuel. This represents an interesting example of how multi-dimensional capabilities can be used to study separate-effect tests, which are becoming increasingly more relevant in nuclear fuel research and are often characterized by unconventional features.

Additionally, the thesis describes a set of methodologies and tools that complement OFFBEAT ena-bling its interaction with other codes relevant for the fuel performance community. First, a cou-pling between OFFBEAT and the Monte Carlo neutron transport code Serpent2 is developed to allow one to obtain a higher-fidelity solution for the neutron flux and for the fuel isotopic composi-tion. This is particularly relevant for fuel rod configurations outside of the range of application of traditional simplified neutronics models such as for new fuel types, experimental reactors, or fuel rods with strong absorbers. Second, a coupling strategy is developed and implemented between the legacy code TRANSURANUS and OFFBEAT. Taking advantage of the strength of both codes, the main aim is to use a validated and fast 1.5-D code to simulate the base irradiation and set accu-rate initial conditions for a multi-dimensional transient performed with OFFBEAT.