Rotating Machinery Fault Diagnosis Research Based On The Wavelet Transform And Neural Network

With the rapid development of the modern industry, The machinery equipment about themachinery, petrochemicals, energy, national defense and carrying industry in the national economy isgrowing more and more large-scale, integrated, high-speed and automation. In these industries, therotating machinery applications is very broad,and the most important parts of rotating machinery wasundoubtedly the rotor, so increasingly subject to the attention of the engineering staff of rotor faultdiagnosis.but in the actual, the rotor failure is very complex, it is a very difficult to correct diagnosisits fault. With the development of science and technology, there have been more newtechnologies,new methods advanced by researchers about the rotor fault diagnosis.This paper firstly made a detailed presentation about the contents、development status and trendsabout the rotating machinery fault diagnosis at home and abroad base on reading the literature. Thenintroduce the development of the data processing technology about the rotating machinery faultdiagnosis,and analysis of the advantages and disadvantages of the various data processingmethod,then introduce the wavelet analysis method and application in fault diagnosis. Next, analysisthe common failure of the rotor in practice, and summary the vibration characteristics, cause andgovernance measures of the common faults. Once again, base on the studies of previous, applicationthe wavelet packet decomposition techniques and sample entropy method with the rotor faultdiagnosis,and simulation of the experimental data to prove the correctness of this method, Finally, theneural network technology application in fault diagnosis, training the sample entropy value obtainedin front as the network input value, the results show that the BP neural network can correct diagnosticclassification about the rotor faults.