Empirical Wavelet Transform And Prony Algorithm Power Harmonic Detection And Recognition

Since the beginning of the new century,the use of non-linear loads represented by power electronic components has increased dramatically,bringing convenience to people's production and life,but it has also injected a great deal of harmonics into the power grid.To deal with harmonics,the first task is to identify and analyze harmonic signals.Aiming at the disadvantages of traditional Prony algorithm which is sensitive to noise and difficult to estimate signal components,this paper proposes an improved Prony algorithm based on empirical wavelet transform(EWT).The main contents and innovations of this article are as follows:The advantages and limitations of traditional time-frequency analysis methods in the detection of harmonic signals are analyzed.The algorithm principles of the empirical mode decomposition algorithm and the empirical wavelet decomposition algorithm and the application in signal recognition are compared.Combining the respective characteristics of empirical wavelet transform and Prony algorithm,they are applied simultaneously in the detection of harmonic signals,and an improved Prony algorithm based on empirical wavelet is proposed.The algorithm firstly uses the EWT algorithm to decompose the noisy harmonic signal,obtains a series of signal components,filters out the noise component and removes the noise component signal,and then performs soft threshold wavelet denoising on other signal component component signals.The harmonic noise signal is processed to achieve enhanced noise reduction,and an accurate number of signal components is obtained.The number of components is used to determine the dimension of the Prony algorithm,and then the parameters of the vibration mode of the harmonic signal after noise reduction are identified.The simulation examples and conditions are set to carry out experimental analysis of the performance of the algorithm to verify the improved algorithm's identification performance,and to analyze the anti-aliasing performance of the EWT algorithm and the improved algorithm's anti-noise performance.The composite signal is set up,and the composite signal is analyzed using the EMD algorithm and the EWT algorithm,respectively.The decomposition results show that the EMD algorithm generates false signal components,that is,modal aliasing occurs,and there is a large error in the decomposition result.The decomposition result of the EWT algorithm coincides with the simulation conditions,indicating that the algorithm can decompose the composite signal correctly.The noise-added electrical signal was set and EMD noise reduction method and EWT combined wavelet noise reduction method were respectively used to compare the noise suppression effects of the two methods.It can be seen from the noise reduction processing effect graph that the EMD algorithm can not completely remove the noise,and there is a shortage of excessive noise residue;the EWT combined wavelet noise reduction method can suppress most of the noise components,so that the signal will not appear obvious distortion.Can achieve effective denoising.Under different noise modes,the accuracy of signal recognition of traditional Prony algorithm,EMD combined with Prony algorithm and EWT combined with Prony algorithm used in this paper was compared.The identification results show that the proposed algorithm has an accurate decomposition result,with less noise reduction distortion and higher identification accuracy.Utilizing the simulation software,a WSCC three-machine nine-node system in accordance with IEEE standards was constructed to test the algorithm.The experimental results show that compared with the traditional Prony algorithm and the joint EMD Prony algorithm,the proposed algorithm is decomposed by EWT.Anti-modal aliasing characteristics,with accurate decomposition results,less noise reduction distortion,identification accuracy and high,suitable for harmonic signal recognition of three-machine nine-node system.