Power Quality Disturbances Analysis Based On Synchrosqueezing Wavelet Transform

With the wide use of new energy,a large number of nonlinear power electronics,generation devices of new energy and impact load are connected to grid,power quality are getting worse.The problem of power quality has received wide focus and research.To improve power quality,detecting power quality is the precondition of controlling power quality.The signal processing method is a wide disturbances detection method for power quality,the detection accuracy is relatively high,but traditional analysis methods,such as wavelet,HHT and so on,can't be balanced in noise immunity,stability and the ability of anti mode aliasing.SST has has better ability of anti mode aliasing and better robustness to noises,so its advantages are more prominent.Synchrosqueezing wavelet transform(SST),which is a new time-frequency analysis tool,is introduced to study power quality in this paper,and SST is applied to detect power quality disturbances,the main contents include:1)Aiming at the problem that the traditional time-frequency analysis method of power quality disturbance is susceptible to noise,the anti-noise,applicability and the possibility of improving detection accuracy of SST in power quality analysis are studied.A minimum allowable signal-to-noise ratio of 20 dB is required for power signals supplied to users in the grid according to IEC standard,the detection accuracy is high when the signal-to-noise ratio is more than 35 dB for wavelet transform to detect disturbance signals,the detection accuracy dramatically drops when signal-to-noise ratio drops to 30 dB.SST method introduced in this paper can detect disturbance signals accurately at 20 dB,it has better noise immunity than wavelet and HHT.As an improvement of wavelet transform,SST has better time-frequency focusing ability and reversibility by squeezing coefficients transformed by continuous wavelet transform in frequency domain,the better detection effect for power quality disturbances by using SST than wavelet transform and EMD is verified by simulation and experiment.2)In order to solve the endpoint problem of SST,ARMA is introduced to extend the endpoint data to improve the decomposition error of the endpoint part of SST.ARMA method is used to predict and extend the two ends of the analysis data,simulation examples and actual data show that ARMA extension can improve the SST endpoint problem.3)In order to better detect power quality disturbances,the time-frequency characteristics of SST are studied.The research shows SST is influenced by the selection of wavelet basis functions and parameters of wavelet basis functions as the same as wavelet,the detection results have no relationship with wavelet basis by selecting different wavelet basis and proper wavelet basis parameters.Morlet wavelet is selected as wavelet mother function in this paper.The simulation results show different center frequencies and bandwidth parameters has a big impact on analysis results.It's not reasonable that it has better detection effect with larger bandwidth parameters and center frequencies,the detection effect is best when bandwidth parameters and center frequencies match with signals to be analyzed.4)For the difficult problems that high precision requirements,strong randomness,wide spectral range for time-frequency analysis of interharmonic,to further improve detection accuracy of interharmonic,the time-frequency analysis ability need to be improved.Selecting wavelet basis parameters of SST based on minimum Shannon entropy principle is proposed in this paper,the Shannon entropy values are calculated by parameters transformed by synchrosqueezing wavelet transform and minimum Shannon entropy value is recognized as the best-matched with original signal,the time-frequency analysis ability by this method is better than that without using optimization to Shannon entropy value,the detection effect to interharmonic is better.