Bearing fault diagnosis method based on Morlet wavelet transformation and convolutional neural network

Abstract

The invention belongs to the mechanical fault diagnosis field, especially relates to application of Morlet wavelet transformation and convolutional neural network, to be specific, provides a bearing fault diagnosis method based on Morlet wavelet transformation and convolutional neural network. Morlet wavelet transformation coefficient matrixes of vibration signals can be used as the input of the convolutional neural network after the uniformization. In the training phase of the convolutional neural network, the learning algorithm provided with the labels having the monitoring function can be adopted, and the minimization adaption function rule can be adapted, and then the weight and the offset of every layer can be adjusted by using the gradient descent with the momentum term. The trained convolutional neural network is used for the classification of the bearing faults, and the diagnosis of the bearing faults can be realized by explaining the classification result. The Morlet wavelet transformation and the convolutional neural network can be combined together for the diagnosis of the bearing faults, and the processing of the original classification data is simpler than that of the prior art, and after the test, the diagnosis identification rate of the self-built sample database can reach more than 80%.

Description

technical field [0001] The invention relates to the field of mechanical fault diagnosis, in particular to the application of Morlet wavelet transform and convolutional neural network. Background technique [0002] Rolling bearings are widely used in industrial production, especially in rotating machinery, and their health directly affects the working status of the entire equipment. Therefore, the research on bearing fault diagnosis is of great significance. [0003] Vibration signal analysis is the most commonly used and most effective method for bearing fault diagnosis. The window of wavelet transform has good adaptive characteristics and translation function, which can make the time resolution high at high frequency and low frequency resolution; while the time resolution at low frequency is low and high frequency resolution, so it is widely used in mechanical fault diagnosis, etc. in engineering applications. [0004] Bearing fault diagnosis actually belongs to the categ...

AI Technical Summary

Problems solved by technology

There have also been many applications of artificial intelligence methods such as neural networks and support vector machines for automatic analysis ...

Images