A sparsity-based method for the analysis of magnetic fluctuations in unevenly-spaced Mirnov coils
Résumé
A new method for identifying toroidal mode numbers in Mirnov data from toroidal plasmas has been found, and benchmarked with simulated and real data from the JET tokamak. Embodied in the SparSpec code, and originally developed for the analysis of unevenly time-sampled astronomical data, this new method fits signals which are unevenly sampled in the toroidal coordinate to a sum of an arbitrarily large number of toroidal modes with integer mode numbers. The method has proven to be extremely robust, and is especially useful for resolving the amplitudes and phases of multiple Alfvén eigenmodes (AEs) which are ringing with the same or nearly the same frequency. The great efficiency with which the SparSpec method detects multiple modes in large datasets suggests that it may be used in real-time applications such as resistive wall mode or tearing mode control, among others. Examples involving unstable AEs as well as stable AEs, (excited by an array of external antennas) in the JET tokamak are used to illustrate the efficacy of the method. Some additional general considerations regarding optimized spatial sampling strategies are also briefly addressed.