Speaker
Description
Following the re-baselining of ITER aimed at operation with full tungsten wall, a renewed interest in studies of the impurity sources during operation of Ion Cyclotron Range of Frequencies (ICRF) antennas emphasizes the importance of validation of the modelling tools simulating the sheath-rectified DC potentials $V_{DC}$. We use the SSWICH-SW code coupled with RAPLICASOL or TOPICA, in order to establish comparisons between calculations and experiments in ASDEX Upgrade (AUG) and JET, increasing the confidence of the code predictions.
In AUG, SSWICH-SW was successfully used in the past to describe the reduction of impurity production by the 3-strap antenna compared to the 2-strap antenna in L-modes. Using recent measurements by a retarding field analyzer to obtain $V_{DC}$ in H-modes with small ELMs, a good agreement with the experimental data is obtained for spatially resolved cases with various 3-strap antenna feedings: optimized dipole, misbalanced power dipole and -90°. This requires a careful coupling between RAPLICASOL (or TOPICA) which calculates near-field $E_{||}$ and 2D SSWICH-SW which calculates $V_{DC}$ taking into account the slow wave dispersion, the RF sheath boundary condition and DC current transport. Defining the interface plane for coupling between the codes at the radial position of the Faraday Screen (FS), as has been used previously, is insufficient in H-mode, despite the plasma density profiles lying above the lower hybrid density as in L-modes: the important $E_{||}$ contributions from the RF image currents on antenna limiters are not treated adequately due to a different current pattern than in L-modes. To properly take these into account, the interface plane radially in front of the antenna limiters should be chosen, or at least 5 mm in front of the AUG antenna FS.
In JET, direct $V_{DC}$ measurements are not available. The RF sheath behavior induced by the 4-strap A2-antenna D is inferred by measuring BeI line intensity at the magnetically-connected outboard limiters, as a function of the power balance between the central straps ($P_{cen}$) and the outer straps ($P_{out}$) in $(0-\pi-0-\pi)$ and $(0-\pi-\pi-0)$ phasings. Trends of dependencies on $P_{cen}/(P_{cen}+P_{out})$ are well described by spatially averaged $|E_{||}|$ , but show a sharper reaction than the calculations of $V_{DC}$ by SSWICH-SW which compare better with the experiments. A heuristic Be-sputtering model by deuterium with the sputtering yield $Y_{Be}=Y_{Be}(V_{DC})$ provides an improved agreement between the calculations and the experiments, but the results become more sensitive to assumptions on $V_{DC}$ averaging.