A numerical model for neoclassical transport under nonaxisymmetric magnetic perturbations in low collisionality plasmas in tokamaks is developed. To take into account bounce-drift resonances and magnetic drift effects, a Fourier decomposition of the drift kinetic equation in new coordinates, rather than bounce average of it, is employed. A pitch angle scattering collisional operator is used to keep accuracy in the nonresonant regimes or resonant plateau regimes with resonant pitch near pitch space boundaries. Full toroidal geometry effects are also included to increase the accuracy in the modeling of neoclassical transport in the resonant regimes. Neoclassical transport in the most important collisionless regimes, including resonant super-banana plateau and bounce-drift resonances regimes, nonresonant 1/ν and ν−ν regimes, and the transitions between them, can be modeled simultaneously in this model by numerically solving the drift kinetic equation. By application to the neoclassical toroidal viscosity modeling in one discharge in the EAST tokamak, it is found that the bounce-drift resonances dominate the contributions near the plasma core where the plasma E→×B→ drift frequency is close to the bounce frequency, while the precessional resonance dominates the contribution near the edge pedestal top where the E→×B→ drift frequency is close to zero.

A: A novel magnetic island structure identification (MISI) system based on Mirnov probes and DDR3 look-up table has been developed on the HL-2A device. The system is dedicated to identify the magnetic island by analyzing the measurement of Mirnov probes and equilibrium fitting (EFIT) code. The applications of high-precision sampling and PCI express (PCIe) platform in the system provides the convenience for Mirnov measurements and the EFIT results transmission. A high performance field programmable gate array (FPGA) is used to realize the real-time identification algorithm. By applying the DDR3 look-up table, 30 current filaments are arranged on a 129 × 129 grid in the calculation of perturbation flux, simplifying the algorithm structure and improving the efficiency. For the advantage of parallel and pipeline processing abilities in the FPGA, the magnetic island structure can be identified with time resolution of 3 ms and spatial resolution of 1.37 cm and 1.02 cm.