Research On Typical Fault Diagnosis Of Asynchronous Motor Based On Boosting Algorithm

In the process of industrial production and in the system of electric locomotive traction,asynchronous motor plays a key role.The operation of industrial equipment and electric locomotive depends on the normal operation of asynchronous motor works or not.Thus,as for industry,the diagnosis of asynchronous motor is also a subject that needs to be studied.It shows that the stator current signal changes when the asynchronous motor is breakdown,and the characteristics of stator current signal are not the same in the different conditions.This paper analyzes the characteristics of the stator current signal,the normal condition of induction motor,the stator winding inter turn short circuit of rotor broken bar,and the bearing.It also studies the algorithm of Boosting's influence on the normal or fault condition.The AdaBoost of normal and asynchronous motor stator interturn short circuit,broken rotor bar case and the bearing classification research are studied in diverse condition,the results are listed as following:(1)Firstly,after the stator current signal of induction motor were collected,and they are transform into Hilbert transform and wavelet and the stator current signal respectively,which can get their Hilbert frequency spectrum and wavelet waveform.(2)Secondly,it should extract the characteristic value of their extraction with the data od Hilbert spectrum and wavelet waveform,according to the MATLAB.Therefore it will find the peak valley difference,peak frequency,peak value,mean value and standard deviation of the Hilbert spectrum.The difference,mean and standard deviation of peak and valley can be extracted in the small wave shape.(3)Thirdly,train the extracted data by using Boosting algorithm and AdaBoost algorithm and compose array.It can get the two classification system of asynchronous motor in normal and fault condition based on the Boosting algorithm,and the classification system of multiple faults in asynchronous motors based on AdaBoost algorithm.Finally,the formation of the Boosting classifier and AdaBoost classifier through the MATLAB simulation analysis,and the error analysis,verify the reliability of the two methods.