Research on Fault Diagnosis Method of Bearings in the Spindle System for CNC Machine Tools Based on DRSN-Transformer

Abstract

A CNC machine tool is an industry mother machine in manufacturing industry. As its key rotating part, bearings in the spindle system of machine tools usually suffer great noise during operation. This paper puts forward a method based on DRSN-Transformer algorithm which can diagnose rolling bearing faults accurately under conditions with high noise. Firstly, we add Gaussian white noise of different signal-to-noise ratios (SNRs) into the original vibration signal, and continuous wavelet transform (CWT) is performed to the vibration signal with noise to construct wavelet time-frequency map. Then, the soft threshold module in the deep residual shrinkage networks (DRSN) can achieve the better effect of noise reduction. And the multi-head attention in the Transformer can pay attention to the important features on different subspaces, so as to enhance the overall feature extraction ability of the model. Finally, the output of the classification module is utilized to obtain the fault diagnosis results. The experimental results show that the proposed method can reach the accuracy of at least 98% under different SNRs condition, which fully verifies the method proposed in this paper can still obtain higher accuracy in strong noise environments. The method in this paper provides a certain reference for fault diagnosis for the spindle system for machine tools.

Keywords

• CWT
• DRSN
• fault diagnosis
• strong noise
• Transformer