Novel gear diagnosis technique based on spectral kurtosis

In this paper, a new thresholding technique used for diagnosis of gearbox pitting tooth faults is introduced. The diagnosis procedure involves in estimation of the Time Synchronous Average (TSA) signal and the gear residual signal, and then the Spectral Kurtosis optimal filter is estimated using the proposed thresholding procedure. By considering overlapping among the TSA segments, several realizations of the TSA signal are estimated. It is important that the SK estimated over the realizations should be consistent. The statistical SK thresholding procedure presented in literature is used for comparing the performance of the proposed approach. A three stage diagnosis decision making technique based on weighted majority rule is used for final diagnosis.