Abstract:  Aiming at the problem that the vibration signal of variable speed rolling bearing has non-stationary characteristics and the fault characteristics are difficult to effectively extract, an intelligent diagnosis method for rolling bearing faults based on multiple simultaneous compression transforms (MSST) and dual-channel convolutional neural networks (CNN) was proposed. Firstly, multisynchro-squeezing transform on the vibration signal under variable speed conditions was performed to obtain an energy-concentrated time-frequency representation. After that, it was input into two-channel CNN with different sizes of small convolution kernels for supervised learning, and the fault feature information was extracted. Secondly, the deeper fault characteristic information was fused by the Concatenate mechanism. Finally, the experiment results of bearing fault identification were output by SoftMax function. A failure recognition rate of more than 99% was achieved on a set of experimental data of rolling bearing failures under variable speed conditions, which verifies the effectiveness of the method and compares it with the single-channel CNN model. The results indicate that the classification accuracy of this model was higher, and the accuracy can reach 99.67%, which proves that the improved model has better non-linear fitting ability and strong robustness, and can be effectively used in the fault diagnosis of variable speed rolling bearings.

Key words:  varying speed bearings; multisynchro-squeezing transform (MSST); convolution neural network (CNN); fault diagnosis; feature fusion

ZHANG Bing, JIANG Pei-gang, LIN Tian-ran.Varying speed bearing fault diagnosis based on MSST and two-channel CNN technique[J].Journal of Mechanical & Electrical Engineering,2021,38(9):1145-1151