A New Method for the Analysis of Alfvén Eigenmodes using Data from Unevenly-Spaced Mirnov Coils
In toroidal plasmas, multiple Alfvén Eigenmodes (AE's) with several toroidal mode numbers can be excited by fast ion populations, and in some circumstances may lead'to rapid loss of fast particle confinement. Although independent of each other, multiple AE's often oscillate with the same or nearly the same frequency. For example, for the case of externally excited AE's, a broad spectrum of modes necessarily is driven at one single frequency. Mode number identification must be accomplished by interpreting signals from a finite number of external Mirnov coils, which typically are unevenly spaced in the toroidal coordinate. However, sensor signals are then the result of a superposition of multiple modes, so the task of spatial mode decomposition becomes a difficult problem. Previous efforts to resolve toroidal mode numbers involve Singular Value Decomposition (SVD) and/or Lomb periodogram techniques. We describe a new approach, based on the SparSpec method, and show that it may be superior in several respects.