Recherche de signatures robustes des défauts dans une machine tournante a travers l’analyse de signaux vibratoires

Abstract

Monitoring of rotating machines by vibration analysis is a topic that has received a great interest in recent years. Moreover, the vibrations from a machine are very affected by the conditions of its operation. An unexpected fault of an installation may cause huge economic losses, and even personal injuries. In this work, two early diagnostic approaches of gear systems failure are proposed. The first approach is based on the Adaptive Morlet Wavelet (AMW). The criterion for the selection of optimal parameters is based on the maximum value of the Morlet wavelet coefficients. The outputs of this procedure are an optimal pair of parameters that give a better time frequency resolution of a gear vibration signal. The problem of fault detection is considered just as a simple signature search in the time-scale domain using scalograms. The second approach is based on the autocorrelation function of the AMW. The prediction procedure of defects in rotating machines is to determine the autocorrelation of the adaptive Morlet wavelet transform based on the maximum value of the wavelet transform. An early indication of the presence of a gear defect is obtained by both approaches.