Vibration Processing Techniques for Fault Detection in Gearboxes

Abstract

Vibration analysis is a very important tool in condition monitoring of operating machines. Many signal processing methods have been developed to extract information about incipient faults from externally measured vibration signals. The article presents the laboratory examinations of some faults in spur and helical gears. In case of spur gears two types of progressing local faults of cracked and chipped gear tooth were simulated and the smoothed pseudo Wigner-Ville distribution was used to demonstrate fault advancement via residual vibration signal analysis. Observing changes in the features of the WV distribution in the contour plots and changes of Kurtosis value monitored the progression of a fault. In case of helical gears some signal changes of transverse vibration velocity of shafts during the process of pitting growth in the tooth working surface have been investigated. Some new indices of pitting wear have been suggested and compared with other non-dimensional discriminants.