Advanced tools, algorithms and strategies for the diagnostics of rolling elements bearings in the E464 locomotive

In the current industrial world, the tendency of trying to optimize the maintenance operations with the aim of reducing the costs is becoming a strategy of greatest importance. The principal aim of this research work, that is part of a more complex research project, is the implementation of new techniques, algorithms and strategies for the diagnostics of faults in rolling elements bearings for a E464 locomotive. A prerogative of this thesis is to allow the timely identification of any possible malfunctioning before its effective onset, in order to avoid serious damages and problems not only to the interested bearing, but also to other components of the locomotive (such as gearbox, shafts and even to the motor) that could generate really dangerous situations. Moreover, in order to be able to quantify the severity of the damages, new indicators based on some properties of the SES2 (squared Square Envelope Spectrum) have been developed. This analysis has been performed thanks to the availability of the effective conditions of the locomotive bearings, evaluated during the maintenance operation of November 2015 in which all the bearings have been replaced with new ones. Indeed, all the analysis has been focused on the vibrational signals of the broken bearings, and through the comparison between the indicators and the actual state of health of the bearings it has been possible to evaluate the reliability of the whole analysis. Another significant goal of this thesis is the identification of the best band pass filtering of the data and operative conditions of the train for the bearing diagnostics, in terms of values of speed, torque and bearing temperature, that allow to highlight the possible onset of damages. At the end of this work, a hint on the prognostics has been presented.