SHANNON ENTROPY A BETTER INDICES FOR LOCAL DEFECT DETECTION AND TO STUDY THE EFFECT OF VARIABLE LOADING CONDITIONS FOR TAPER ROLLER BEARING

Statistical parameters are having great significance in bearing condition monitoring where characteristic signature localization is not possible in time, frequency or time-frequency domain. In this paper, the presence of local defect is analyzed by carrying out statistical analysis on Acoustic and Vibration signal. The variation of statistical parameters for the defective bearing in comparison to healthy bearing at particular loading condition is presented for the purpose of analysis. Among all the parameters Shannon entropy (SE) is exhibiting better variations for defective bearing compared to the healthy bearing for both the signals. It has also been observed that with increase in loading only standard deviation (σ ) and SE have shown downward trend. This might be because of change in characteristics of signature with loading, the main reason is still underlying. To check the sensitive parameter for loading among the responded parameters simple sensitivity index (SSI) is calculated. From the results it has been observed that SE is having better sensitivity for the loading than σ because of its calculation capability in logarithmic scale.