Modelling of integrated effect of volumetric heating and magnetic field on tritium transport in a U-bend flow as applied to HCLL blanket concept

Under fusion reactor operational conditions, heat deposition might cause a complex buoyant liquid metal flow in the HCLL blanket, what has a direct influence on tritium permeation ratio. In order to characterise the nature of this flow, a simplified HCLL channel, including the U-bend near the reactor first wall, is analysed using a finite volume CFD code, based on OpenFOAM toolbox, following an electric potential based formulation. Code validation results for developed MHD flow and magneto-convective flow are exposed. The influence of the HCLL U-bend on the flow pattern is studied with the validated code, covering the range of possible Reynolds numbers in HCLL-ITER blanket, and considering either electrically insulating or perfectly conducting walls. It can be stated that, despite the very low velocities and the high Hartmann number, flow pattern is complex and unsteady vortices are formed by the action of buoyancy forces together with the influence of the U-bend. Through the analysis, the flow physics is decoupled in order to identify the exact origin of vortex formation. A simplified tritium transport analysis, considering tritium as a passive scalar, has been carried out including a study on boundary conditions influence and a sensitivity analysis of tritium permeation fluxes to diffusivity and solubility parameters. Results show the relevance of Sievert’s coefficient uncertainties, which alters the permeation ratio by an order of magnitude.

► 3D transient CFD code based on OpenFOAM toolbox and accounting for MHD and thermal et al. effects.
► Hydrodynamic instabilities caused by the jet (generated at the gap narrowing) are found at Reynolds 480.
► Hartmann 1740 is able to stabilise the flow.
► A heat deposition corresponding to Gr = 5.21 × 109 is sufficient for buoyancy to be predominant at the bend region. Flow becomes unstable.
► Tritium permeation ratio cannot be accurately predicted due to major uncertainties in Sievert’s coefficient.