Research On Machinery Fault Diagnosis Method Based On Wavelet Neural Network

It is out of question that default diagnosis is very important both in safety production and in national economy.In recent years,with the development of computer technology,signal processing,artificial intelligence,mode identification and etc,the fault diagnosis technology has been continuously promoted. Because wavelet neural network has the following merits: high precision and learning rate fast, we use wavelet neural network in the field of fault diagnosis.Based on a multitude of literature,this paper reviewes that the fault diagnosis technology and the active theory and methods in the field of fault diagnosis which include wavelet transform,neural network and wavelet neural network. The wavelet neural network has become a focus in the field of fault diagnosis recently.In view of the many nonlinear vibration of running-machine and Fast Fourier Transform has some limits in the dispose of nonlinear signal,at the same time,because wavelet transform has the time-frequency characteristics and wavelet analysis is an effective tool to process signals,this paper not only introduce the theory of wavelet transform,but also set forth its application in fault diagnosis and its superiority of processing nonlinear signal.Because wavelet neural network has the better diagnosis effect while using wavelet function substitutes network's excitation function,so this paper mainly studies the compact wavelet neural network which using morlet wavelet substitutes network's traditional excitation function(S function) and its application in fault diagnosis of rotating machinery. Classical neural networks mostly train the network with BP algorithm. But the BP algorithm often gets into the minimum value and its constringent speed is slow. In view of these limitations,this paper improves the classieal BP algorithm by means of introducing momentum item.The actual example proves that the speed of this wavelet neural network is very faster than classical BP neural network.Finally, the experiment shows that the fault diagnosis method based on wavelet neural network in this paper can be used well in fault diagnosis of rotating machinery and has a more faster convergence rate.