Comparative Analysis of Wear Resistance for Bearing Coating Using VMD

The purpose of this research work is to find an onboard, non-destructive technique which can be correlated directly with the level of wear resistance using vibration signature analysis.

Five ball bearings (Model: 6205) and standard specimens were electroplated with five different coatings with unique level of wear resistance. The wear resistance of the standard specimens was measured using Pin-on-disk apparatus and presented in ascending order: Black oxide, Silver, Copper, Zinc phosphate (ZnP) and Nickel respectively. The coated bearings were made to run in customised bearing test rig at five different speeds ranging between 300 rpm to 1500 rpm and vibration signatures were recorded. The acquired signals were decomposed into six modes using Variational mode decomposition (VMD) for band wise statistical analysis by calculating RMS, Crest factor, Variance, Skewness, Kurtosis, Shannon entropy and Log energy, to develop correlation with the level of wear resistance.

All the six bands were analysed by calculating the statistical parameters in reference to different level of wear resistance. Significant variation in variance, log energy and Shannon entropy was found at one of the six bands at all speeds except 900 RPM. With the variation in wear resistance, the corresponding change in responded statistical parameters was found in low energy bands generated by the VMD, which signifies low impact of wear resistance variation on signal’s energy. On further analysis using chain indexing, Shannon entropy and log energy were found to be most relevant parameters at low and high rpm respectively.

The given technique found to be useful to monitor the wear resistance level of bearing surface, while bearing is in running condition. Therefore, surface defects due to variation in wear resistance can be notified at very incipient stage by carrying out statistical analysis of vibration signature after decomposing the signal up to sixth level using Variational mode decomposition method.