Abstract

Studies reveal that the most prominent cause of bearing failure is a crack on any of its mating surfaces. When the crack is initiated, the bearing can still be used for some duration, but this is majorly depending upon the loading conditions. This work primarily focuses on the effects of different levels of static loading on the crack propagation after crack initiation. To analyze the effect of static loading, an axial groove defect was seeded on the outer race of a taper roller bearing randomly and bearing run continuously under five different static loading conditions. Initially, the bearing was made to run under loading conditions to initiate the crack naturally but the crack was not initiated even after 800 h of running. Therefore, crack was initiated artificially for the purpose of studying crack propagation. It was observed from the experimentation that in the case of maximum static load of 20 kg, the crack propagates rapidly in terms of area after 109 h of continuous running, whereas in the case of no load, it started propagating quickly after 267.5 h of running. Statistical analysis was also carried out for the recorded signals at different intervals of times, and it was observed that the Shannon entropy value was showing a sudden rise with the edge breakage (visually verified) while the crack was propagating. However, in the statistical analysis, none of the parameters showed a correlation with crack propagation. To develop the correlation of crack propagation, Shannon entropy of high, medium and low frequency bands of continuous wavelet-based (CWT) was carried out using different wavelets. Shannon entropy for high frequency band of CWT using Daubechies 10 as mother wavelet has responded well to the crack propagation as the value showed a sudden rise and an overall increase for edge breakage and crack propagation, respectively. A high frequency band of CWT using Daubechies 10 was found suitable for detecting edge breakage and crack growth at the same time because of its capability to respond to transient characteristics for a large duration of time.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call