Diagnostic des engrenages par le calcul de la corrélation spectrale

Abstract

Among the methods applied in the maintenance of industrial installations, vibration analysis is one of the most widespread. In fact, the vibratory signatures appearing on an installation during operation are closely linked to their dynamic behavior and to their functional state. Rotating machines produce cyclic signals according to one or more basic cycles repeating indefinitely. The existence of these cycles naturally leads to exploiting the cyclostationarity of these signals, that is to say the periodicity of their statistical parameters. Gardner et al. Developed the theory of cyclostationary processes, which has been widely used in the fields of communication. The first application of the theory of cyclostationarity for the diagnosis of gears was proposed by C. Capdessus in 1992 and it is followed by other authors in other parts of machines. Spectral correlation is one of the most versatile spectral tools for analyzing cyclostationary signals. The Averaged Cyclic Periodogram (ACP) has been the most popular and widely used spectral correlation estimator for bearing failure detection. However, their traditional techniques have low computational efficiency. As a result, Antoni proposed the Fast Spectral Correlation Method (Fast-SC), which is a new method of spectral correlation estimation. In this study, we want to apply this last algorithm of fast calculation of the spectral correlation to make an early diagnosis of the failures of the gears