Abstract
Higher order spectra exhibit a powerful detection capability of low-energy fault-related signal components, buried in background random noise. This paper investigates the powerful nonlinear non-stationary instantaneous wavelet bicoherence for local gear fault detection. The new methodology of selecting frequency bands that are relevant for wavelet bicoherence fault detection is proposed and investigated. The capabilities of wavelet bicoherence are proven for early-stage fault detection in a gear pinion, in which natural pitting has developed in multiple pinion teeth in the course of endurance gearbox tests. The results of the WB-based fault detection are compared with a stereo optical fault evaluation. The reliability of WB-based fault detection is quantified based on the complete probability of correct identification. This paper is the first attempt to investigate instantaneous wavelet bicoherence technology for the detection of multiple natural early-stage local gear faults, based on comprehensive statistical evaluation of the industrially relevant detection effectiveness estimate—the complete probability of correct fault detection.
Highlights
Local gear tooth faults generate vibration transients each time a mechanical interaction with a faulty tooth surface occurs
The methodology has a series of advantages over previously used methodologies: (I) it does not require higher order spectra (HOS) estimation, and it does not have a negative impact on calculation time; (II) it has a physical sense allowing it to precisely identify frequency bands in which vibration energies have increased due to gear faults; and (III) it is based on the wavelet transform, which is suitable for gear fault-related transients and matching with wavelet HOS techniques
The relatively high values of Fisher criterion (FC) peaks, which indicate the angular location of integrated WB (IWB) values, to wavelet bicoherence (WB) values, are contained within the range from zero to uni the pinion faults, confirm the relatively high effectiveness of fault detection based on
Summary
Local gear tooth faults generate vibration transients each time a mechanical interaction with a faulty tooth surface occurs. Gelman et al [34] proposed a novel instantaneous WB for the detection of transient signal components The novelty of this approach is that the WB was used without a local average for the evaluation of statistical relations between multiple spectral components, contained in multiple broadband frequency bands and generated by local gear faults. The novelties of the presented research are as follows: Instantaneous wavelet bicoherence is investigated for the first time in worldwide terms for gear fault detection via very comprehensive experimental statistics of naturally developed pitting of gear teeth at a very early stage of fault development, based on the industrially relevant detection effectiveness estimate—the complete probability of correct fault identification; We use a novel methodology of selecting the frequency bands that are relevant for gear fault detection for wavelet HOS integrated features, which enable efficient fault detection and localisation
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
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 callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
