Bearing defect diagnosis based on a complex filter and singular value ratio spectrum

A novel signal processing algorithm was proposed here for vibration signal analysis in condition monitoring and health diagnosis of rolling bearings. Such technique required an envelope being extracted from vibration signal with a complex analytical band-pass filter which was designed for better efficiency in processing than Morlet wavelet and Harmonic wavelet. The principal periodic component in the envelope was subsequently detected, enhanced and reconstructed with sweep frequency method based on singular value ratio (SVR) spectrum. Such signal processing approach was experimentally evaluated by using vibration signals measured on rolling element bearings that contained localized structural defects with proved validity and efficiency.