Open Access
Subscription or Fee Access
Improved Sensitivity of Condition Monitoring Features via Hölder Exponent Analysis
Abstract
This paper proposes the use of a condition monitoring method for detecting damage and wear in rotating machines. The method makes use of the time-dependent Hölder exponent, a measure of signal regularity/continuity, and its power spectral density to improve upon traditional condition monitoring damage features. The proposed method is used for detecting worn tooth damage in gearboxes. A basic introduction behind the theory of computing the Hölder exponent is provided, followed by a statistical analysis approach to classifying damage and comparing the performance of the proposed method to traditional methods in current practice.