Non-axisymmetric magnetic perturbations contribute to the transport of fast ions in tokamaks. While knowledge of the perturbed magnetic topology suffices for characterizing the displacement of thermal particles (few keV), the trajectories of fast particles (0.1–few keV) vary depending on their charge, mass, kinetic energy and velocity pitch. The computational effort needed to follow the drift orbits for a large number of phase space samples (as done in Monte Carlo simulations) is too high for scenario simulators, so it is necessary to develop computationally efficient reduced models. In this paper, we present a method that allows to estimate the locations, widths and overlaps of resonant drift orbit islands in the guiding center phase space of passing fast ions with small computational effort. The new method constitutes a significant improvement over the method presented in our previous work (Shinohara K. et al 2018 Nucl. Fusion58 082026), both in performance and in accuracy. This is achieved by deriving an expression for the drift orbit island width from the perturbed guiding center velocity in the straight-orbit coordinates of the unperturbed drift surfaces and eliminating the need for computationally expensive Poincaré maps. When applied to resonant magnetic perturbation fields in a KSTAR tokamak plasma, our new estimates for the island widths are shown to agree with Poincaré maps from orbit-following simulations to within 15%. The island positions and the chaotic regions caused by island overlaps also agree well.
Phase-space spectrum analysis of fast-ion resonances with externally applied 3D magnetic perturbations in the EAST tokamak