Research On Time-frequency Characterization Method Of Sunspot Oscillation Based On ConceFT

At present,scientists generally believe that the periodic oscillations in sunspots are closely related to the structure of sunspots and the principle of solar heating.However,the sunspot oscillations in different positions and solar atmospheres,even a special sunspot oscillation will show different characteristics.At the same time,there are a lot of white noise and red noise in the observation signal of sunspot oscillations,which makes its signal to noise ratio low.This series of objective factors make it more difficult to study the frequency components of sunspot oscillations.Time-frequency analysis is an important tool for studying frequency components in signals.So far,the timefrequency analysis methods applied to sunspot oscillations mainly include Continuous Wavelet Transform(CWT)and Synchrosqueezed Transforms(SST).However,these methods do not perform well in environments with low signal-to-noise ratio.Therefore,at current stage,before performing the time-frequency analysis of the sunspot oscillation,the high-power low-frequency noise will be filtered out first,and then analyze the time-frequency characteristic of the sunspot oscillation signal.Filtering may introduce false signals,which will have a significant impact on the analysis results of time-frequency spectrum.ConceFT time-frequency analysis method is proposed on the basis of SST.Compared with SST,this method has accurate frequency positioning and a better performance under low signal-to-noise ratio.This method has been widely used in the analysis of industrial signals and medical signals.Applying ConceFT to the timefrequency analysis of sunspot oscillations can improve the accuracy of frequency positioning without filtering low-frequency noise and show the time-frequency characteristics of the oscillating with low power in the observed signals of sunspot oscillations.The research content of this article includes:(1)Design similar analog signals based on the observed signals of real sunspot oscillations and use ConceFT to analyze the time-frequency characteristic of noise-free non-stationary,white noise-containing non-stationary oscillation signals and red noisecontaining non-stationary signals respectively.The results of time-frequency estimation obtained by CWT and SST can evaluate the accuracy of ConceFT time-frequency analysis and provide a strong basis for practical applications.(2)Use the linear fitting method to obtain the average power of the background noise in the global logarithmic spectrum.Then set the corresponding threshold with chi-square test to suppress the low-frequency noise in the result of ConceFT to improve the accuracy of the time-frequency analysis.(3)Apply ConceFT and threshold denoising to the time-frequency analysis of sunspot oscillations.Analyze Time-frequency characteristic of the whole sunspot oscillation signal,umbra oscillation signal and penumbra oscillation signal respectively.In this paper,ConceFT is applied to the time-frequency analysis of sunspots oscillations.Combining with experiments,it is proved that the time-frequency analysis method ConceFT used in this paper has better frequency positioning ability than CWT.At the same time,it performs better than SST in low signal-to-noise ratio environments and has a good performance in time-frequency analysis on the weak oscillation signals in sunspots.