Condition monitoring and fault diagnosis of bearings are of practical significance in industry. In order to get a feature containing different fault signatures, this paper uses Wavelet Transform (WT), Wavelet Lifting Scheme (WLS) and Empirical Mode Decomposition (EMD), respectively, to decompose signal into different frequency bands. Then, Singular Value Decomposition (SVD) is utilized to extract intrinsic characteristic of signal from obtained matrix. These singular value vectors are regarded as inputs to Hidden Markov Models (HMM) for identification of machinery health condition. In this research, the fault diagnosis system is validated by motor bearing data, including normal bearings, inner race fault bearings, outer race fault bearings and roller fault bearings. Analysis results show that this method is effective in bearing fault diagnosis and its classification rate is excellent.
Application of Singular Value Decomposition and Deep Learning in Bearing Fault Diagnosis
