LOCUST-GPU predictions of fast-ion transport and power loads due to ELM-control coils in ITER

Abstract

The graphics processing unit (GPU) version of the Lorentz-orbit code for use in stellarators and tokamaks (LOCUST) has been applied to study the fast-ion transport and loss caused by resonant magnetic perturbations in the high-performance Q = 10 ITER baseline scenario. The unique computational efficiency of the code is exploited to calculate the impact of the application of ITER's edge-localised mode (ELM) control coil system on neutral beam heating efficiency, as well as producing detailed predictions of the resulting plasma-facing component power loads, for a variety of operational parameters—the applied fundamental toroidal mode number n 0, mode spectrum and absolute toroidal phase of the imposed perturbation. The feasibility of continually rotating the perturbations is assessed and shown to be effective at reducing the time-averaged power loads. Through careful adjustment of the relative phase of the applied perturbation in the three rows of coils, peak power loads are found to correlate with reductions in neutral beam injection (NBI) heating efficiency for n 0 = 3 fields. Adjusting the phase this way can increase total NBI system efficiency by approximately 2%–3% and reduce peak power loads by up to 0.43 MW m−2. From the point of view of fast-ion confinement, n 0 = 3 ELM control fields are preferred overall to n 0 = 4 fields. In addition, the implementation of 3D magnetic fields in LOCUST is also verified by comparison with the SPIRAL code for a DIII-D discharge with ITER-similar shaping and n 0 = 3 perturbation.