Affiliation:
1. USTHB
2. Université Champagne Ardène
Abstract
The monitoring and vibratory analysis of gear transmission allow the prediction of a possible malfunction and breakdowns. As the gear transmission product non-stationary signals its treatment is too difficult with the usual tools of signal processing witch can product errors in its interpretation. As the characteristics of gear frequencies are predetermined, it is proposed to monitor (fault identification) using wavelet analysis. To simulate the signal to be analyzed, we intentionally introduced a spalling defect. We chose the Daubechies wavelet type which are the most used in diagnostic. The aim of this work is to try to control the various parameters related to the wavelet analysis for reliable and inexpensive detection, i.e., the order of the wavelet and level decomposition. The approach witch was previously used for bearings, consists on observing the kurtosis for several orders wavelet based on the default severity..
Publisher
Trans Tech Publications, Ltd.
Reference13 articles.
1. K. Drouiche, S. Menad, Y. Grenier, Détection de défauts d'engrenages par analyse vibratoire. Traitement du Siganl, 1991, Vol. 8, n°5, p.333 – 343.
2. S. Aatola, R. Leskinen, Cepstrum analysis predicts gearbox failure. Noise Control Engineering Journal, 1990, Vol. 34, n°2, pp.53-59.
3. B. Bouzouane, A. Miloudi, N. Hamzaoui, A. Benchaala, Détection de défauts de machines tournantes par la méthode des ondelettes. 2003, 16ème Congrès Français de Mécanique.
4. B. Randall, A new method of modelling gear default. Journal of Mechanical Design. 1982, Vol 104, p.259–267.
5. A. Djebala, N. Ouelaa, T. Kebabsa, Y. Ramdane et M.K. Babouri, Détection des défauts non stationnaires de dentures d'engrenages par application de l'analyse en ondelettes et la transformée de Hilbert. Congrès Algérien de Mécanique Mascara Algérie. (2013).