A Precision machining equipment fault diagnosis based on CWT and improved ResNeXt: Establishment of a fault diagnosis model for precision equipment

Abstract

In view of the problem that the working environment of precision machining equipment is very complex and the collected signals contain very rich information, a fault diagnosis method of precision machining equipment based on continuous wavelet transform and improved multi-dimensional residual network is proposed. First, the collected original vibration signal is subjected to continuous wavelet transform (CWT), and the feature map is generated and grayscaled. Finally, the improved ResNeXt network is used to diagnose the fault of the precision machining equipment. The best wavelet basis function Cmor 3-3 was selected by comparing 6 different wavelet base types. The test results show that the proposed CWT and improved ResNeXt algorithm can realize the diagnosis of precision machining equipment in a complex environment and achieve an average accuracy of 99.4%.