Despite its widespread use, the short-time Fourier transform spectrogram may not always be the best approach to analyze some nonstationary fusion plasma signals. The Choi-Williams distribution has been shown to be effective to analyze such signals. Wavelets have also been used in fusion to improve upon the spectrogram. While they are intuitive and their resolution is better than the spectrogram's, wavelets are still hindered by the uncertainty principle. In principle, the Choi-Williams distrbution is superior to wavelets as far as time-frequency resolution is concerned. Still, a comparison for real fusion plasma signals has not yet been performed. Here, such a comparison is made and the advantages of using the Choi-Williams distribution over wavelets are shown. The comparison is for edge phenomena, such as edge localized modes and washboard modes; and core phenomena, such as sawtooth precursors, fishbones and neoclassical tearing modes in ICRH heated discharges. |
Trabalhos Científicos 
Comunicações a Reuniões Científicas