Signal Processing Diagnostic Tool for Rolling Element Bearings Using EMD & MED

The signal processing techniques developed for the diagnostics of me-chanical components operating in stationary conditions are often not applicable or are affected by a loss of effectiveness when applied to signals measured in transi-ent conditions. In this paper, an original signal processing tool is developed ex-ploiting some data-adaptive techniques such as Empirical Mode Decomposition, Minimum Entropy Deconvolution and the analytical approach of the Hilbert trans-form. The tool has been developed to detect localized faults on bearings of traction systems of high speed trains and it is more effective to detect a fault in non-stationary conditions than signal processing tools based on envelope analysis or spectral kurtosis, which represent until now the landmark for bearings diagnostics.