A Rolling Element Bearing Diagnosis Method Based on Singular Value Decomposition and Squared Envelope Spectrum

The vibration caused by an early defect on the rolling element bearing (REB) is very weak and easy to be submerged other signals and noise. Therefore, the performance of a bearing fault diagnosis method mainly depends on two key steps, namely, bearing fault signal component extraction and bearing fault type identification. In this article, authors have proposed a bearing fault diagnosis method that combines the techniques of squared envelop spectrum (SES) analysis and singular value decomposition (SVD). The original vibration signal will be decomposed into several sub-signals through SVD. Then, sub-signals are grouped according to their similarity. Later, the SES of the grouped signal is applied to identify the fault type. The performance of this method is tested through actual vibration signals obtained from the test-rig.