Asynchronous Motor Vibration Fault Diagnosis Research

With the development of science and technology,asynchronous motor as a kind of important electric power equipment in industry,agriculture and aerospace and other fields has been widely used.Normal operation of the motor in the event of failure,not only cause losses in the motor itself will also have serious consequences for all aspects of industrial production,bring great inconvenience to the People's Daily life.Asynchronous motor fault diagnosis technology can under the premise that maintain the integrity of the motor,through the analysis of the motor the various operation parameters,the real-time and effective state distinguishing diagnosis of motor,and in a timely manner to prevent motor fault or to prevent further spread of fault caused by damage.So the asynchronous motor fault diagnosis technology in the motor fault early detection and timely maintenance,is of great theoretical significance and social economic benefits,it has become the domestic and foreign scholars have a research hot topic.In this paper,in the second chapter of the asynchronous motor various types of vibration fault,fault mechanism and main characteristics are introduced and summarized,in view of the asynchronous motor vibration fault occurred in the early fault feature frequency components in original signal energy is difficult to extract weak,using the traditional time-frequency analysis method to analyze the diagnosis more difficult situation,put forward a kind of energy entropy based on empirical mode decomposition,the combination of principal component analysis and support vector machine method,the vibration signal EMD energy entropy is extracted as the classification of feature vector,according to the article break fault vibration signal energy entropy difference before and after the common rotor vibration fault diagnosis for identification.The third chapter of the EMD,energy,entropy,PCA and support vector machine method to do the detailed introduction,this paper expounds the characteristics and advantages of several methods,and in view of the vibration fault of two common motor experiment analysis,through comparison with BP neural network method,based on support vector machine method was verified inthe motor vibration fault diagnosis is of great advantage.The fourth chapter,according to the theory of motor vibration fault of a typical rotor broken bar fault,using stator current spectrum analysis and vibration analysis method based on energy entropy is studied,the analysis results show that the early vibration fault is not obvious when the fault features,the original signal using energy entropy based method can accurate to diagnose fault make an accurate diagnosis.Used in this article based on the EMD energy entropy and principal component analysis and support vector machine of the asynchronous motor vibration fault diagnosis method is simple,the recognition rate is high,can effectively extract the fault characteristic signal,accurate identification of diagnosis to fault.