A wavelet-based method to measure the toroidal mode number of ELMs

The high confinement mode regime (H-mode) in tokamaks is accompanied by the occurrence of burst of MHD activity at the plasma edge, so-called edge localized modes (ELMs). Because of the short time scales involved in the ELM crash (on JET typically 0.2 ms), standard Fourier analysis can hardly be used to extract their toroidal mode number. On the other hand, the assessment of linear stability of ELMs with the ion drift effects included, makes the identification of their toroidal mode numbers an important issue, while an accurate comparison with the theory of nonlinear evolution of ELMs requires the knowledge of the nonlinear spectrum. Compared to Fourier analysis, wavelets are suitable to study transient events on time scales comparable to the wave period. Spectral analysis based on sinusoidal wavelet functions has been applied to study the spectral properties of magnetic perturbations associated with ELMs and with their precursors, in JET plasmas with toroidal rotation driven by unbalanced NBI. It is shown that, combining wavelet analysis with statistical two-point correlation techniques, it is possible to get information on the toroidal mode number structure of magnetic perturbations during the phases that immediately precede the ELM and during the ELM crash itself.