Development of Computational Technique for Labeling Magnetic Flux-Surfaces

In recent Large Helical Device (LHD) experiments, radial profiles of ion temperature, electric field, etc. have been measured in the m/n = 1/1 magnetic island produced by island control coils, where m is the poloidal mode number and n the toroidal mode number. When the plasma transport in radial profiles is numerically analyzed, an average over a magnetic flux-surface in the island is a very useful concept to understand the transport. When averaging, a proper labeling of the flux-surfaces is necessary. In general, it is not easy to label the flux-surfaces in a magnetic field containing the island, compared with the case of a magnetic field configuration having nested flux-surfaces. In the present paper, we have developed a new computational technique to label the magnetic flux-surfaces. This technique uses an optimization algorithm called the simulated annealing method. The flux-surfaces are discerned by using two labels: one is classification of the magnetic field structure, i.e., core, island, ergodic, and outside regions, and the other depends on the value of the toroidal magnetic flux. We have applied this technique to an LHD configuration with the m/n = 1/1 island, and successfully discriminated of the magnetic field structure