Research On Disturbances Detection And Identification For Power Quality

The network structure and load structure of modern power system are undergoing great changes,the wide application of nonlinear,impact and sensitive loads,as well as the distributed power,energy storage devices and new energy sources that are integrated into the power grid,which make power quality problems have become increasingly prominent.Accurate detection and recognition of power quality disturbance is a necessary prerequisite for the control and evaluation of power quality problems.In this paper,the related research and analysis are carried out from the problem of noise removal,location,detection and recognition of power quality disturbance.Firstly,because the transient and nonstationarity of power quality disturbance signals,a denoising method based on complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)and wavelet adaptive threshold is proposed based on empirical mode decomposition and wavelet transform.Making use of the method and four conventional wavelet threshold denoising algorithm to denoise for oscillation transient,the simulation results show that the method achieves better denoising effects,and preserves more informations of the high frequencies,which verifies effectiveness of the new method.At the same time,the result of denoising on the measured signal is good,and lays a foundation for the disturbances detection for power quality.Secondly,the high-frequency coefficient modulus maximal value of transient disturbances and steady-state disturbances are extracted by wavelet multi-resolution analysis,which combined with the time-frequency localization of wavelet transform,the effective detection and accurate positioning of various types of power quality disturbances are realized.The simulation results show that the detection effect of selected db4 wavelet is better than db10 and db40 in the detection of voltage sags,and the positioning results are more accurate;in the detection of voltage flicker and harmonic,the otherness of high frequency modulus value of transient disturbances is analyzed.Thirdly,because the shortcomings of the traditional Hilbert-Huang Transform(HHT),the problems of modal aliasing,endpoint effect and spurious components are analyzed,combined with the Complete Ensemble Empirical Mode Decomposition(CEEMD),adaptive waveform matching extension and correlation coefficient method,which is used as improved HHT to detection the various disturbances,including transient disturbances,voltage flicker,harmonics,interharmonics and compound disturbances containing harmonics and interharmonics.The simulation results show that the detected values such as disturbance amplitude,frequency and time detection obtained by improving HHT are closer to the actual values,and the endpoint effect and modal aliasing has been significantly improved,which verifies the accuracy and effectiveness of the method.At the same time,the analysis result of the measured signal is close to the actual situation,and it provides a guarantee for the identification and classification of power quality disturbance.Finally,according to the detection results of different disturbances based on improved HHT,five characteristic parameters are extracted from instantaneous amplitude and Hilbert spectrum,such as the duration of disturbance,amplitude and frequency compnents of disturbance,etc,with the threshold of each branch of the decision tree classifier determined,a decision tree classifier is designed for the recognition of disturbances.According to the mathematical model of thirteen kinds of power quality disturbances,a large number of testing samples are randomly produced by Matlab,Classification tests are conducted under three kinds of conditions,including SNR of 30 d B,SNR of 40 d B and SNR of 50 d B,the average correct recognition rate under each signal to noise ratio is counted,and the accuracy of the method is verified,it also has a certain degree of noise resistance and stability.