Computational Thermal Fluid Dynamic Analysis of Cooling Systems for Fusion Reactor Components

In a fusion reactor during plasma operation, the heat loads on plasma facing components can be as high as 5 MW/m2 [1], which should be removed by a proper mechanism to prevent the damage of reactor components. In order to handle such high heat fluxes a suitable heat sink with proper thermal hydraulics is required. In the recent past several heat sinks have been proposed; among which the Hypervapotron heat sink, operating in the highly subcooled boiling regime, is considered as one of the potential candidates. In order to accurately predict the performance of the system, a thermal hydraulic analysis is required. This thesis employs a Computational Fluid Dynamic (CFD) approach to do the thermal hydraulic analysis of the subcooled flow boiling inside the Hypervapotron channel. For this purpose four boiling models are tested using two commercial CFD codes. The four boiling models tested are Rensselaer Polytechnic Institute (RPI) boiling model [2] available in ANSYS-FLUENT 13, Bergles-Rohsenow (BR) model [3] implemented as an external User Defined Function (UDF) in the FLUENT code, the Rohsenow boiling model [4] extended with the capability of transition to film boiling for high heat fluxes available in STAR-CCM+ 7.02, and finally Transition boiling model [4] available in STAR-CCM+ 7.02. These models are used to test the thermal performance of Hypervapotron using the experimental data (showing the variation of temperature with heat flux) obtained from the experiments conducted at Efremov Institute Russia and Joint European Torus United Kingdom. Simulations were conducted using the above mentioned boiling models, the obtained results were compared against the experimental data and also different boiling models are compared with each other whenever possible to test their applicability. From the simulations conducted on the Hypervapotron geometries it is found that the Transition boiling model can capture the thermal performance (in terms of tracing the experimental data) better than any other model both quantitatively and qualitatively, covering the different boiling regimes shown by the experiments ( that is no boiling, nucleate boiling and hard boiling regimes), than the other models.