Research On Asynchronous Motor Fault Feature Extraction Method Based On Time-frequency Analysis

As the main power equipment converting electrical energy into mechanical energy in modern industry.To ensure the normal,safe and efficient operation of the motor is an important guarantee for the normal operation of industry,agriculture,national defense construction and people’s life.Due to the limitation of high load,environmental deterioration,improper operation and other factors,the problem of fault diagnosis of induction motor becomes more and more complicated and difficult.The fault characteristic frequency extracted from the motor fault data and timely diagnosis of motor fault can effectively ensure the stable operation of the motor.This paper takes the study of motor fault mechanism as the starting point,studies the application of time-frequency analysis method to extract motor fault characteristic frequency,and provides a new idea for motor fault diagnosis.In view of the complex electromagnetic relationship caused by motor air gap eccentricity,rotor broken bar and turn-to-turn short circuit,the changes of magnetomotive force and magnetic tension caused by motor faults were deduced,and the characteristic frequencies of each fault were analyzed theoretically.The mathematical model of the motor in the three-phase coordinate system and two-phase coordinate system is established,and the mathematical model of the turn-to-turn short-circuit fault is derived by increasing the fault component of the winding,and finally the validity of the model is verified by using Simulink numerical experiments.Aiming at the problem that the frequency domain methods can only reveal the signal frequency component characteristics,the time-frequency analysis of typical motor faults was studied.Firstly,we analyze the motor fault data using short-time Fourier transform and continuous wavelet transform to reveal the time-frequency characteristics of the signal to obtain the time-frequency spectrum of the fault signal,and then use discrete wavelets to decompose the signal in different frequency bands to extract the fault characteristic frequencies.Numerical verification combined with engineering application shows that the short-time Fourier and continuous wavelet can effectively obtain the time-frequency characteristics of the fault signal,and the decomposed wavelet signal can extract part of the fault characteristic frequencies.Aiming at the problem that the energy of time-frequency results based on continuous wavelet transform is not concentrated enough to extract the effective ridges,Synchrosqueezing Transform with improved time-frequency distribution and ridge extraction method based on Multisynchrosqueezing Transform were studied.By compressing and rearranging the timefrequency coefficients to improve the time-frequency energy distribution,clear time-frequency ridges are extracted and the ridges are reconstructed,and finally the fault characteristic frequencies are extracted based on the ridge reconstruction.Numerical validation combined with engineering application shows that the method can effectively extract the time-frequency ridges containing the signal fault features.Aiming at the problem that the traditional time-frequency method cannot reflect the fault transient characteristics,a feature extraction method based on Transient-extracting Transform(TET)is studied.TET is an improved time-frequency analysis method based on the short-time Fourier transform.By introducing the Dirac function,the aggregation of time-frequency energy is significantly improved with the removal of uncorrelated time-frequency energy,which can effectively characterize and extract the transient components in the fault signal,and rich fault characteristic frequencies can be extracted by relying on TET for the overall reconstruction of the signal.Numerical validation combined with engineering application shows that the method can effectively concentrate the time-frequency energy and can extract rich fault characteristic frequency components compared with other methods.