On the accuracy of CFD modeling of cyclic high heat flux divertor experiment

To investigate the heat removal capability of conceptual divertor cooling fingers, accurate computational fluid dynamics (CFD) analyses are indispensable. Although the cooling performance of the divertor finger has been successfully high-heat-flux tested under real DEMO conditions in a combined helium loop and electron beam facility at the Efremov Institute, Russia, an accompanying numerical simulation of the experiments is of great importance. This might help to better understand the complex thermo-hydraulic conditions with the aim of predicting other different load cases. To accurately reproduce the experimental boundary conditions, the Gaussian-like shape of the absorbed power was taken into account and the heat losses were estimated. Modeling of the structure thermal conductivity was also found to be an important source of modeling uncertainty. In the context of accurate modeling of experimental conditions, the effect of some modeling assumptions was evaluated. Transient simulations of the cyclic heat flux experiment were performed only for the solid part of the cooling finger to avoid excessively long computation times. The helium cooling was taken into account by the heat transfer coefficient (HTC) on the fluid–structure interface, obtained from the steady-state simulations of the full solid–fluid model. The HTC distribution did not vary with time throughout the entire transient simulation. The modeling error associated with such HTC approximation was estimated for the particular cyclic experiment. It is shown that the simulated temperature cycles on the top of experimental mock-up agree well with the measured data.

► CFD analysis of 1-finger mock-up tests under cyclic high-heat-flux loading.
► Uncertainty of CFD modeling assumptions is estimated.
► Gaussian-like shape of absorbed power and experimental heat losses are considered.
► Thermal conductivity modeling essential, highly depends on manufacturing process.
► Transient simulations show good agreement with measured top-surface temperatures.