ETD

• C++
• CFD
• heat pipe
• nuclear reactor
• numerical modeling
• sodium
• software development
• solver verification
• thermo-fluid dynamics

Heat pipes are widely recognized as highly efficient thermal management devices, capable of transferring large heat loads with minimal temperature gradients and outperforming conventional heat transfer solutions. In recent years, interest in this technology has intensified, driven by demanding applications in advanced energy systems and aerospace engineering, where reliability, robustness, and predictive capability are critical. In these contexts, the accurate numerical modeling of heat pipe behavior is essential, as experimental testing alone is often limited by cost, complexity, and operational constraints. A particularly novel application concerns heat removal in nuclear reactors, for which heat pipes offer a passive heat removal mechanism that does not rely on gravity.

This thesis presents the development of a dedicated C++ solver called HPCM (Heat Pipe Coupled Model) for heat pipe analysis, conceived as a hybrid framework inspired by two state-of-the-art legacy codes, THROHPUT and HPTAM. The solver is constructed as the coupling of three standalone modules, respectively modeling the heat pipe wall, porous wick, and vapor core. Its objective is to capture the coupled thermo-fluid dynamics governing heat pipe operation while maintaining modularity, numerical robustness, and computational efficiency. Particular emphasis is placed on the radial treatment of mass and energy transport in the different regions of the heat pipe, as well as on the implementation of temperature- and pressure-dependent thermophysical properties for the various materials.

Each standalone module of HPCM is verified against analytical solutions and reference numerical simulations, ensuring correctness, stability, and reproducibility. Residuals and global heat and mass balances of the coupled solver are monitored to assess consistency. The results produced by HPCM are qualitatively compared with data reported in the THROHPUT manual, demonstrating its capability to reproduce key heat pipe phenomena and to provide a foundation for more advanced simulation approaches. The developed framework establishes a basis for future extensions toward more complex operating regimes and geometries, contributing to the advancement of predictive numerical tools for heat pipe analysis.

Finally, the thesis outlines the development of two additional solvers: a vapor-core monolithic compressible solver named MCS (Monolithic Compressible Solver) and a solver addressing the transition from the free-molecular regime during frozen startup to the continuous-flow regime. These solvers can be integrated into a conduction-based framework commonly referred to as the Liquid Conduction Vapor Flow (LCVF) model. In this approach, the wall and wick are described solely by heat conduction equations, while the vapor core is modeled using a Bosanquet formulation in the free-molecular regime and a one-dimensional compressible flow model in the continuous regime.