The blanket in a magnetic confinement fusion reactor is a critical component that provides, among other functions, the cooling for the plasma-facing first wall and breeds the tritium required to fuel the reactor operation. The most promising concepts involve the use of a liquid metal as working fluid, i.e. lithium-lead eutectic alloy, thanks to excellent thermal properties and the possibility to double the coolant role as tritium breeder. The interaction between the liquid metal and the plasma-confining magnetic field defines many, if not all, of the relevant thermal-hydraulic features of these components and are not yet completely understood. In the framework of the development of the prototype reactor DEMO, the EUROfusion consortium is funding the study of three liquid metal blanket conceptual designs: Helium-Cooled (HCLL), Dual-Cooled (DCLL) and Water-Cooled (WCLL). This proposal aims to study a key phenomenon for the blanket design: the mixed convective MHD flows. The non-uniform volumetric heating of the liquid metal due to the fusion high energy neutrons creates sharp temperature gradients in the blanket which, in turn, promote significant buoyancy forces. These are expected to at least affect the main pressure-driven flow, i.e. causing local flow reversal in long vertical ducts. Blanket concepts (e.g. WCLL) that rely on the minimization of the molten metal velocity to less than 1 mm/s to reduce the MHD pressure drop to manageable levels are expected to exhibit a buoyancy-dominated flow. Preliminary hydrodynamic CFD studies have found a considerable deviation in the blanket temperature distribution compared with the pure forced convection case. The main purpose of this work is to support these results performing a MHD CFD analysis with ANSYS CFX. The work would be divided in two phases: a validation against numerical data in the literature for flows in 2D and 3D geometries and, finally, the simulation in realistic blanket conditions.
Nuclear fusion is considered the most promising energy source to meet the ever-increasing demand by the world for electricity consumption in the 21th century due to virtual carbon-free emissions, very limited production of only short-lived radioactive wastes and employing as fuel deuterium (cheap, uniformly distributed and virtually inexhaustible). However, the most promising fusion reaction involves the use of tritium that, as opposed as deuterium, as no significant natural source. Therefore, the design of a blanket able to breed tritium from lithium due to neutron capture reactions to an extent required to guarantee the self-sufficiency of the fusion reactor is a non-negotiable condition to develop this source to a commercial level.
This proposal aims to extend the knowledge on relevant liquid metal MHD flows which are expected to be observed in fusion reactor blankets. The interaction between the liquid metal and the plasma-confining magnetic field defines many, if not all, of the relevant thermal-hydraulic features of these components and are not yet completely understood. In particular, a noticeable lack has been detected in the scientific literature for studies considering mixed convective MHD flows in horizontal ducts. This work plans to partially address this issue by providing numerical data for a realistic blanket configuration derived from the elementary cell of a blanket under development (WCLL). The results obtained would be of immediate impact for the support of the reference design and could highlight the existence of interesting MHD phenomena that would be necessary to further investigate with experimental activities.
To perform this activity, a commercial CFD providing a suitable MHD model is employed (ANSYS CFX 15.0). A preliminary validation would be performed to gauge the quality of the obtained numerical results against state-of-the-art 2D and 3D simulations for the considered class of MHD flows. This would allow to extend the range of reliability of the code which, currently, has already being validated for both 2D buoyant and pressure-driven MHD flow. This is believed to be particularly important since the lack of a dedicated computational MHD code restricts to general purpose CFD codes the tools available to support the blanket design.