A MDOF Model to Predict Ball-Race Interaction in a Damaged Roller Bearing

In the last few decades, machine diagnostics has undergone considerable developments in the direction of Predictive Maintenance. The state of the art in the field of machine monitoring suggests vibration analysis as the best approach to reach appreciable and reliable results. This technique is based on the identification and surveillance of the vibration source - constituted by elements such as shafts, belts, gears and rolling bearings - picking out machine case vibration, usually on a support. Rolling bearings are the most widely used machine elements, so a great deal of work has been dedicated to the development of effective procedures for their monitoring and diagnostics. However, little work has been focused on determining the actual dynamical behaviour of rolling elements during their impact on default: better knowledge of this behaviour should improve diagnostic technique application. The present work is the outcome of an extensive study carried out by the Authors [3,4], in order to relate the characteristics of a defect – shape, dimension and depth – to the signal picked up by a transducer mounted on the case