ELM resolved energy distribution studies in the JET MKII Gas-Box divertor using infra-red thermography
Using infra-red (IR) thermography, power loads onto the MKII Gas-Box divertor targets have been investigated in Type-I ELMy H-Mode plasmas at JET in medium current discharges (I-p = 2.6MA and B-T = 2.7 T). Heat fluxes are calculated from the measured divertor target tile surface temperatures taking into account the influence of co-deposited surface layers on tile surfaces. This is particularly important when estimating the energy deposition during transient events such as ELMs. Detailed energy balance analysis is used, both from IR and tile embedded thermocouples, to demonstrate an approximately constant ELM-averaged in/out divertor target asymmetry of approximate to 0.55 and to show that the ELM in/out energy deposition ratio ranges from 1 : 1 to 2 : 1. The inter-ELM in/out ratio is close to the ELM-averaged value at low pedestal collisionalities and decreases down to values close to zero when the inner target plasma detaches at the highest pedestal collisionalities. The fraction of ELM transported energy is observed to behave differently for the inner and the outer divertor. At higher pedestal collisionalities nearly the full inner target load is due to the ELMs whereas for the outer target the ELM transported energy never exceeds values of approximate to 0.3 of the total energy deposited there. The fraction of ELM energy arriving at the divertor compared with the pedestal loss energy in JET is found to be in the range of 0.75 for small ELMs down to 0.4 for large ELMs systematically decreasing with normalized ELM size. Since ITER is bound to use small ELMs the corresponding ELM wall load is expected to be small. The latter experimental result is in fair agreement with the observation that larger ELMs tend to travel faster across the SOL than smaller ELMs. However, a comparison of the presented data with models of ELM perpendicular transport is not conclusive due to the large experimental errorbars and uncertainties in the model assumptions.