Development of a CFD Model for Analysis of Grooved Heat Pipes for Satellites Applications

Heat pipes are widely used in satellite thermal control systems to transport heat from heat-generating components to the cooler parts of the satellite. Computational Fluid Dynamics (CFD) is widely used in the design and optimization of those components. In the open literature, however, little material is present regarding high-fidelity simulations that can describe both the fluid behaviour inside the grooves and the overall thermal performance. Moreover, information about the fluid behaviour inside the heat exchanger is not easy to obtain experimentally due to the difficulty of obtaining visual data inside the aluminium coating. This thesis aims to propose a computational model that is able to predict fluid behaviour in grooved heat pipes. The CFD model is developed using the open-source software OpenFOAM. In order to achieve a good compromise between the computational cost of the simulations and the physical phenomena taken into account by the equations, a two-phase VOF model was used. Boiling and condensation are described with the Lee model. The ‘interCondensatingEvaporatingFoam’ solver was used in the first place, but the incompressible flow hypothesis was considered too stringent. For this reason, a new solver was developed starting from ‘compressibleInterFoam’. This new solver is shown to provide more accurate interface capturing that can describe better the bubbles’ motion and the pressure difference caused by the heat input. In the released versions of the software, ‘compressibleInterFoam’ generates numerical instabilities at the interface that are being investigated. The solver thus can only be used in specific conditions. The final part of the thesis aims to build the foundation to remove the instabilities and validate the model using experimental data available in the literature. A validation of the capillary rise is proposed, together with a setup to validate separately the phase change and tune the coefficients in the Lee model.