Research On The Features Of Frequency Modulation For Induction Motor Bearing Fault

Three-phase induction motor play an important role in the field of industrial productions.As the main component of motor,bearing has a high incidence of malfunction.Researching on the monitoring of bearing faults is of great significance.At present,the common method to monitor induction motor bearing fault with stator current signal is Power spectral Density which has a few deficiencies.First of all,bearing fault characteristic frequency is modulated with supply power frequency,and the signal is too weak.Secondly,there are many complex noise signals in the low frequency band.Finally,comparing with bearing fault characteristic frequency signal,power source is strong.So it's hard to extract bearing fault characteristic frequency.In order to overcome the deficiencies above,this thesis will calculate and analyze the current expression with cyclostationary theory.Then discusses the feasibility of the extraction and amplification of bearing fault characteristic frequency signal along with reduction of noise by cyclic autocorrelation function.The calculation and programming in MATLAB of cyclic autocorrelation function for discrete data is presented in the next part of this thesis.Simulation of bearing fault has been established in this thesis.Furthermore,the cyclic autocorrelation function of the stator current measured by the laboratory experimental platform for motor bearing fault under various motor condition is calculated.Comparing the result of cyclic autocorrelation function and power spectral density,it verify the correctness of the theory and method mentioned in this thesis.it comes out that the identifiability of stator current bearing fault characteristic frequency can be ascended.This conclusion proves the superiority of cyclostationary theory in terms of extracting the bearing fault characteristic frequency.