Numerical study of thermal erosion and topographical change of divertor target plates induced by type-I edge-localized modes

Abstract

The high-Z material tungsten (W) is a promising candidate of the plasma facing components (PFCs) for the future tokamak reactors due to its high melting point (3683 K), low tritium retention and low sputtering yield. However, there are still many problems about W PFCs. One of them is the material melting under off-normal transient heat fluxesit is one of the most outstanding open questions associated with the use of W divertor targets in international thermonuclear experimental reactor (ITER). This requires us urgently to understand the W melting behavior under high power flux deposition condition. In this paper, a two-dimensional (2D) fluid dynamic model is employed by solving the liquid hydrodynamic Navier-Stokes equation together with the 2D heat conduction equation for studying the erosion of the divertor tungsten targets and its resulting topographical modification during a type I-like edge-localized mode (ELM) in ITER with a Gaussian power density profile heat load. In the present model, major interaction forces, including surface tension, pressure gradient and magnetic force responsible for melt layer motion, are taken into account. The simulation results are first benchmarked with the calculated results by other code to validate the present model and code. Simulations are carried out in a wide range of fusion plasma performance parameters, and the results indicate that the lifetime of W plate is determined mainly by the evolution of the melt layer. As a consequence of the melt layer motion, melted tungsten is flushed to the periphery, a rather deep erosion dent appears, and at the dent edges two humps of tungsten form during the ELM. The humps at both edges are almost at the same height. Calculated results show the topographical modification becomes noticeable when the W plate is exposed to a heat flux of 2000 MWm-2 for 0.8 ms (in the simulation, the parameter k=ə/əT is taken to be -9.010-5 Nm-1K-1, where is the surface tension coefficient and T is the temperature). The values of the humps are both about 2.1 m, and the surface roughness is about 1.1 m. The longer the duration of the ELM, the more rapidly the humps rise. The melt flow may account for the higher surface temperature at the pool periphery, and for the larger melt thickness. It is found that when the energy flux is under 3000 MWm-2 the surface tension is a major driving force for the motion of melt layer. Under the same heat flux, the bigger the k used in the simulation, the more severe the surface topography of the target becomes; while at the same k, the higher the heat flux, the more severe the surface topography of the target becomes. In addition, a modified numerical method algorithm for solving the governing equations is proposed.