Research On Bearing Fault Diagnosis Method Based On MCSA

It is the realistic demand of enterprises and society to diagnose the early bearing failure of squirrel cage induction motor with high precision and high reliability.The research on relevant diagnosis methods has both theoretical and economic significance.When the motor runs at low load,the bearing fault characteristics in the stator current signal are weaker than those in the rated load,which makes it difficult to extract fault characteristics and increase the difficulty of fault identification.Therefore,it is a challenging and promising work to diagnose bearing faults in this situation.The research of early bearing fault diagnosis under low load condition has always been the key content in the field of fault diagnosis.Aiming at the problem that the resolution of classical FFT method depends on the selection of data quantity,and it is difficult to detect bearing fault features when the motor is running at low load,this thesis proposes a bearing fault detection method based on Park transform and high resolution spectral analysis.The proposed method can effectively solve the problem that the fault features are weak and difficult to extract.The simulation and experimental results show that the proposed method not only inherits the advantages of high resolution of Root-MUSIC or rotation invariant algorithm,but also has high diagnostic accuracy.In addition,the running time of fault detection using this method is also shorter,and early bearing fault detection under low load can be realized by using shorter data.1)Research on low load bearing fault detection based on Multi-signal classification(MUSIC)method.In order to solve the problem of low frequency resolution of short-term data of FFT method,MUSIC and Root-MUSIC method were applied to the fault detection of bearing outer ring under low load condition.Under the same experimental conditions,the diagnostic performance of FFT,MUSIC and Root-MUSIC methods was compared and analyzed.The results show that the Root-MUSIC method can detect the fault under low load operation,and the diagnosis effect is stable.2)Research on bearing fault detection under low load condition based on rotation invariant algorithm(ESPRIT).In view of the problems of multi-signal classification methods such as MUSIC and Root-MUSIC,such as consuming more computing resources and running time,Simple ESPRIT algorithm based on rotation invariant is adopted to improve the classification methods.The experimental results under different load conditions show that Simple ESPRIT diagnosis method has the advantages of high frequency resolution and less running time,and its results show that the Simple ESPRIT method can detect the fault under low load operation stably.3)Research on bearing fault detection method based on Park transform and high resolution spectral analysis.Aiming at the problems of long running time of Root-MUSIC method and lack of diagnostic accuracy of Simple ESPRIT method,an early bearing fault detection method combining Park transform and high resolution spectral analysis was proposed.The experimental results show that the Simple ESPRIT method after Park transformation not only has the advantage of shorter running time,but also obtains higher accuracy than the original ESPRIT method,which is suitable for early bearing fault detection under low load.