Early detection of gear failure by vibration analysis--ii. interpretation of the time-frequency distribution using image processing techniques

Abstract

A system is described which uses image processing techniques to assist in the automatic interpretation of gear vibration signatures for early failure detection and fault diagnosis. The vibration signature of an individual gear in a gearbox is extracted by the time domain synchronous averaging technique from the total vibration signal measured on the gearbox casing. The time-frequency distribution of the gear vibration signature, determined from the spectrogram, is treated as an image, and the early failure detection and fault diagnosis are performed by interpreting the patterns in the image by the procedures of segmentation, connectivity analysis and feature extraction. The implementation of the technique, the criteria for analysis of the image, and experimental examples are described.