Rotating machine fault diagnosis based on denoising source separation

The application of denoising source separation (DSS) technology to the mechanical vibration signal processing provides a new technique of mechanical fault diagnosis. The DSS theory is investigated in this paper. The analog signals of the rotating machine are separated and the performance index and correlation coefficient of DSS are better than those of the joint approximate diagonalization of eigen-matrices (JADE). Applying the DSS method to rotating machine fault diagnosis, the measured fault signals are analyzed and the results are found to be in agreement with practice. The results show that the DSS method is efficient in analyzing the fault diagnosis of rotating machine.