G. Gelle 1 , M. Colas 1 and C. Servière 2

There is a great interest to apply BSS methods in mechanical system signal processing for monitoring or diagnosis purpose. Actually, we show that BSS allows to recover the vibratory information issued from a single rotating machine working in a noisy environment by freeing the sensor signal from the contribution of other working machines. In that way, BSS can be used as a pre­processing step to rotating machine fault detection and diagnosis. In this paper, we compare two possible approaches to solve BSS problem of rotating machine signals, ie. temporal or frequential approach. The first method, initially developed to temporally white signals is used in an experimental context and we show that the results are comparable to frequential domain approach specially developed for rotating machine signals. These two approaches are tested on real signals arisen from a mechanical testing bench, and the implementation of different methods as well as their performances are discussed.